Compare commits

...

2373 Commits

Author SHA1 Message Date
mstembera 378c8bdbb8 Stockfish 9
Official release version of Stockfish 9

Bench 5023629
2018-01-31 11:41:09 +01:00
Torsten Franz 2e11388581 Retire lever (#1378)
Retire the lever bonus from pawns evaluation

STC: http://tests.stockfishchess.org/tests/view/5a6aef5c0ebc590d945d59c8
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 88290 W: 19549 L: 19560 D: 49181

LTC: http://tests.stockfishchess.org/tests/view/5a6b70140ebc590d945d59f7
LLR: 3.22 (-2.94,2.94) [-3.00,1.00]
Total: 104603 W: 18105 L: 18108 D: 68390

Bench 5023629
2018-01-30 09:40:56 +01:00
Rocky640 0a5b03af3f Limit the king distance factor when evaluating passed pawns (#1373)
Limit the king distance factor when evaluating passed pawns
Passed STC
http://tests.stockfishchess.org/tests/view/5a6bf7290ebc590d945d5a3a
LLR: 3.31 (-2.94,2.94) [0.00,5.00]
Total: 23987 W: 5550 L: 5281 D: 13156

and LTC
http://tests.stockfishchess.org/tests/view/5a6c57710ebc590297c36af2
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 16926 W: 3014 L: 2820 D: 11092

Bench: 5059457
2018-01-28 14:56:45 +01:00
syzygy1 fd4d800c98 This modifies the in-search TB probing to continue searching for a mate "behind" a TB win (or loss). (#1285)
It does the following:

- If a TB win or loss value allows an alpha or beta cutoff, the cutoff is taken.
- Otherwise, the search of the current subtree continues. In PV nodes, the final value returned is adjusted to reflect that the position is a TB win (or loss).

The patch also fixes a potential problem caused by root_probe() and root_probe_wdl() dirtying the root-move scores.

This patch removes the limitation of current master that a mate is never found if the root position is not yet in the TBs, but the path to mate at some point enters the TBs. The patch is intended to preserve the efficiency and effectiveness of the current TB probing approach.

No functional change (withouth TB)
2018-01-28 14:40:07 +01:00
mibere a5a98d822e Top CPU Contributors as of January 2018 (#1367)
No functional change.
2018-01-24 16:54:04 +01:00
Stéphane Nicolet 254d995e18 Contempt 20
Set the default contempt value of Stockfish to 20 centipawns.

The contempt feature of Stockfish tries to prevent the engine from
simplifying the position too quickly when it feels that it is very
slightly behind, instead keeping the tension a little bit longer.

Various tests in November 2017 have proved that our current imple-
mentation works well against SF7 (which is about 130 Elo weaker than
current master) and than the Elo gain is an increasing function of
contempt, going (against SF7) from +0 Elo when contempt is set at
zero centipawns, to +30 Elo when contempt is 40 centipawns.

See pull request 1325 for details:

https://github.com/official-stockfish/Stockfish/pull/1325

This november discussion left open the decision of which "default"
value for contempt we should use for Stockfish, taking into account
the various uses ofStockfish (opening preparation for humans, computer
online tournaments,analysis tool for web pages, human/computer play,
etc).

This pull request proposes to set the default contempt value of SF
to twenty centipawns, which turns out to be the highest value which
is not a regression against current master, as this seemed to be a
good compromise between risk and safety. A couple of SPRT[-3..1]
tests were done to bisect this value:

Contempt 10: http://tests.stockfishchess.org/tests/view/5a5d42d20ebc5902977e2901 (PASSED)
Contempt 15: http://tests.stockfishchess.org/tests/view/5a5d41740ebc5902977e28fa (PASSED)
Contempt 20: http://tests.stockfishchess.org/tests/view/5a5d42060ebc5902977e28fc (PASSED)
Contempt 25: http://tests.stockfishchess.org/tests/view/5a5d433f0ebc5902977e2904 (FAILED)

Surprisingly, a test at "very long time control" hinted that using
contempt 20 is not only be non-regressive against contempt 0, but
may actually exhibit some small Elo gain, giving a likehood of superio-
rity of 88.7% after 8500 games:

VLTC:
ELO: 2.28 +-3.7 (95%) LOS: 88.7%
Total: 8521 W: 1096 L: 1040 D: 6385
http://tests.stockfishchess.org/tests/view/5a60b2820ebc590297b9b7e0

Finally, there was some concerns that a contempt value of 20 would
be worse than a value of 7, but a test with 20000 games at STC was
neutral:

STC:
ELO: 0.45 +-3.1 (95%) LOS: 61.2%
Total: 20000 W: 4222 L: 4196 D: 11582
http://tests.stockfishchess.org/tests/view/5a64d2fd0ebc590297903868

See the comments in pull request 1361 for the long, nice discussion
(180 entries :-)) leading to the decision to propose contempt 20 as
the default value:

https://github.com/official-stockfish/Stockfish/pull/1361

Whether Stockfish should strictly adhere to the Komodo and Houdini
semantics and add the UCI commands to force the contempt to be White
in the so-called "analysis mode" is still under discussion, and may
be or may not be the object of a future commit.

Bench: 5783344
2018-01-23 14:26:45 +01:00
Tom Vijlbrief 5451687efb Make razor margin depth independent
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 37171 W: 6680 L: 6587 D: 23904

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 11632 W: 1574 L: 1442 D: 8616

bench: 5098576
2018-01-23 14:15:05 +01:00
Fabian Fichter b61759e907 Simplify away redundant SEE pruning condition (#1363)
SEE immediately returns true for promotions,
so excluding them before checking SEE is redundant.

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 56758 W: 10166 L: 10106 D: 36486
http://tests.stockfishchess.org/tests/view/5a645eaf0ebc590297903833

No functional change.
2018-01-23 14:05:48 +01:00
Marco Costalba b508f9561c Fix bench number
Incorrect bench number in master. Fix it.

bench: 4971497
2018-01-13 09:21:23 +01:00
Günther Demetz 1b6459195c Simplify verification search (#1362)
1. avoid recursive call of verification.
   For the interested side to move recursion makes no sense.
   For the other side it could make sense in case of mutual zugzwang,
   but I was not able to figure out any concrete problematic position.
   Allows the removal of 2 local variables.
   
2. avoid further reduction by removing R += ONE_PLY;

Benchmark with zugzwang-suite (see #1338), max 45 secs per position:
Patch  solves 33 out of 37
Master solves 31 out of 37

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 76188 W: 13866 L: 13840 D: 48482
http://tests.stockfishchess.org/tests/view/5a5612ed0ebc590297da516c

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40479 W: 5247 L: 5152 D: 30080
http://tests.stockfishchess.org/tests/view/5a56f7d30ebc590299e4550e

bench: 5340015
2018-01-13 09:01:23 +01:00
IIvec aa88261a8f Revert to old time management (#1351)
As many users reported some problems with new time management,
and recent tests on longer time controls

http://tests.stockfishchess.org/tests/view/5a460e160ebc590ccbb8c35d
http://tests.stockfishchess.org/tests/view/5a462f4d0ebc590ccbb8c37a

are even little in favor of old time management, this revert seems as a logical step.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14060 W: 2562 L: 2430 D: 9068

LTC:
LLR: 3.44 (-2.94,2.94) [-3.00,1.00]
Total: 31611 W: 3958 L: 3827 D: 23826

bench: 5365777 (same as master)
2018-01-13 08:59:20 +01:00
mbootsector 33682bfb98 Use mobility in kingsafety (#1360)
Use mobility in kingsafety

STC: http://tests.stockfishchess.org/tests/view/5a55134d0ebc590296938a28
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 32651 W: 6057 L: 5792 D: 20802

LTC: http://tests.stockfishchess.org/tests/view/5a5618f40ebc590297da516f
LLR: 2.94 (-2.94,2.94) [0.00,5.00]
Total: 17875 W: 2421 L: 2245 D: 13209

bench: 5365777
2018-01-13 08:56:42 +01:00
hxim 66ce8ad5fd Remove ThreatByHangingPawn bonus (#1356)
* no ThreatByHangingPawn

* bench: 4919682
2018-01-04 22:31:30 +01:00
Joost VandeVondele 9afa1d7330 New Year 2018
Adjust copyright headers.

No functional change.
2018-01-01 13:18:10 +01:00
Joost VandeVondele 2ba47416cb Explicitly zero TT upon resize.
as discussed in issue #1349, the way pages are allocated with calloc might imply some overhead on first write.
This overhead can be large and slow down the first search after a TT resize significantly, especially for large TT.
Using an explicit clear of the TT on resize fixes this problem.

Not implemented, but possibly useful for large TT, is to do this zero-ing using all search threads. Not only would this be faster, it could also lead to a more favorable memory allocation on numa systems with a first touch policy.

No functional change.
2018-01-01 13:17:51 +01:00
Guenther Demetz 7d4d3a2c3a Include x-ray attacks through all queens independently of the color.
When calculating attacks from rooks/bishops current master includes
x-rays through own queen. This patch includes also x-rays through
opponent queen.

Credits go to Brian who inspired for this idea
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/Z3APRYpQeMU

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 41549 W: 7544 L: 7244 D: 26761
Elo	2.05 [-0.29,4.19] (95%)
http://tests.stockfishchess.org/tests/view/5a3b5fe50ebc590ccbb8bf9a

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 110010 W: 14208 L: 13739 D: 82063
Elo	1.20 [-0.27,2.55] (95%)
http://tests.stockfishchess.org/tests/view/5a3b8e7c0ebc590ccbb8bfad

bench: 5544445
2017-12-26 10:42:24 +01:00
Joost VandeVondele 1c50d8cbf5 Upon changing the number of threads, make sure all threads are bound
The heuristic to avoid thread binding if less than 8 threads are requested resulted in the first 7 threads not being bound.
The branch was verified to yield a roughly 13% speedup by @CoffeeOne on the appropriate hardware and OS, and an earlier version of this patch tested well on his machine:

http://tests.stockfishchess.org/tests/view/5a3693480ebc590ccbb8be5a
ELO: 9.24 +-4.6 (95%) LOS: 100.0%
Total: 5000 W: 634 L: 501 D: 3865

To make sure all threads (including mainThread) are bound as soon as the total number exceeds 7, recreate all threads on a change of thread number.
To do this, unify Threads::init, Threads::exit and Threads::set are unified in a single Threads::set function that goes through the needed steps.
The code includes several suggestions from @joergoster.

Fixes issue #1312 

No functional change
2017-12-26 10:40:42 +01:00
Joost VandeVondele 2198cd0524 Allow for general transposition table sizes. (#1341)
For efficiency reasons current master only allows for transposition table sizes that are N = 2^k in size, the index computation can be done efficiently as (hash % N) can be written instead as (hash & 2^k - 1). On a typical computer (with 4, 8... etc Gb of RAM), this implies roughly half the RAM is left unused in analysis.

This issue was mentioned on fishcooking by Mindbreaker:
http://tests.stockfishchess.org/tests/view/5a3587de0ebc590ccbb8be04

Recently a neat trick was proposed to map a hash into the range [0,N[ more efficiently than (hash % N) for general N, nearly as efficiently as (hash % 2^k):

https://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modulo-reduction/

namely computing (hash * N / 2^32) for 32 bit hashes. This patch implements this trick and now allows for general hash sizes. Note that for N = 2^k this just amounts to using a different subset of bits from the hash. Master will use the lower k bits, this trick will use the upper k bits (of the 32 bit hash).

There is no slowdown as measured with [-3, 1] test:

http://tests.stockfishchess.org/tests/view/5a3587de0ebc590ccbb8be04
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 128498 W: 23332 L: 23395 D: 81771

There are two (smaller) caveats:

1) the patch is implemented for a 32 bit hash (so that a 64 bit multiply can be used), this effectively limits the number of clusters that can be used to 2^32 or to 128Gb of transpostion table. That's a change in the maximum allowed TT size, which could bother those using 256Gb or more regularly.

2) Already in master, an excluded move is hashed into the position key in rather simple way, essentially only affecting the lower 16 bits of the key. This is OK in master, since bits 0-15 end up in the index, but not in the new scheme, which picks the higher bits. This is 'fixed' by shifting the excluded move a few bits up. Eventually a better hashing scheme seems wise.

Despite these two caveats, I think this is a nice improvement in usability.

Bench: 5346341
2017-12-18 16:32:21 +01:00
Günther Demetz b53239d641 Enhanced verify search (#1338)
by disabling null-move-pruning for the side to move for first part of
the remaining search tree. This helps to better recognize zugzwang.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 18220 W: 3379 L: 3253 D: 11588
http://tests.stockfishchess.org/tests/view/5a2fa6460ebc590ccbb8bc2f

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 41899 W: 5359 L: 5265 D: 31275
http://tests.stockfishchess.org/tests/view/5a2fcf440ebc590ccbb8bc47

For further detail see commit notes and discussion at 
https://github.com/pb00068/Stockfish/commit/6401a80ab91df5c54390ac357409fef2e51ff5bb

bench: 5776193
2017-12-18 16:30:27 +01:00
Gontran Lemaire 83e829c9dc Remove QueenMinorsImbalance array #1340
Remove QMI array and adjust bishop, knight and queen coefficients
in QuadraticOurs and QuadraticTheirs arrays in compensation of
this removal.

STC : http://tests.stockfishchess.org/tests/view/5a21d8350ebc590ccbb8b5fe
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 49659 W: 9029 L: 8957 D: 31673

LTC : http://tests.stockfishchess.org/tests/view/5a33c0dd0ebc590ccbb8bd7e
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 45905 W: 5834 L: 5745 D: 34326

Bench: 5176807
2017-12-17 09:20:17 +01:00
Joost VandeVondele f6981b1321 Make staticEval independent of the search path
Current master can yield different staticEvals depending on the path
used to reach the position. The reason for this is that the evaluation after a
null move is always computed subtracting 2 * Eval::Tempo, while this is not
the case for lazy or specialized evals. This patch always adds tempo to evals,
which doesn't affect playing strength:

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 59911 W: 7616 L: 7545 D: 44750

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 104947 W: 18897 L: 18919 D: 67131

Fixes issue #1335

Bench: 5208264
2017-12-17 09:11:55 +01:00
Rocky640 be6fafd079 Simplify other checks (#1337)
Replace an intricate definition with a more natural one.

Master was excluding squares occupied by a pawn which was blocked by a pawn.
This version excludes any squares occupied by a pawn which is blocked by "something"

Passed STC
http://tests.stockfishchess.org/tests/view/5a2f557b0ebc590ccbb8bc0d
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 44211 W: 8009 L: 7928 D: 28274

and LTC
http://tests.stockfishchess.org/tests/view/5a301d440ebc590ccbb8bc80
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 31958 W: 4108 L: 4002 D: 23848

Bench 5000136
2017-12-17 08:50:45 +01:00
Alain SAVARD 020dd69a35 Simplify other checks #1334
Simplify the other check penalty computation. Compared to current master,

a) it uses a 143 kingDanger penalty instead of S(10, 10) for the "otherCheck"
(credits to ElbertoOne for finding a suitable kingDanger range to replace the score
and to Guardian for showing this could also be a neutral change at LTC).
This makes our king safety model more consistent and simpler.

b) it might also score more than one "otherCheck" penalty for a given piece type instead of just one

c) it might score many pinned penalties instead of just one.

d) It also remove 3 conditionals and uses simpler expressions.
So it was tested as a SPRT[-3, 1]

Passed STC
http://tests.stockfishchess.org/tests/view/5a2b560b0ebc590ccbb8ba6b
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 11705 W: 2217 L: 2080 D: 7408

And LTC
http://tests.stockfishchess.org/tests/view/5a2bfd0d0ebc590ccbb8bab0
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 26812 W: 3575 L: 3463 D: 19774

Trying to improve on b) another attempt was made to score also the
"otherchecks" for piece types which had some safe checks, but this
failed STC http://tests.stockfishchess.org/tests/view/5a2c79e60ebc590ccbb8badd

bench: 5149133
2017-12-11 15:27:44 +01:00
Ben Koshy 70262f2027 Add Resources to understand code base (#1332)
No functional change.
2017-12-10 13:46:43 +01:00
Günther Demetz 196ae7c38b Don't consider defending queen as check blocker (#1328)
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 22636 W: 4212 L: 3990 D: 14434
http://tests.stockfishchess.org/tests/view/5a2506140ebc590ccbb8b75a

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 63448 W: 8287 L: 7965 D: 47196
http://tests.stockfishchess.org/tests/view/5a253a610ebc590ccbb8b776

bench: 5767699
2017-12-06 14:23:41 +01:00
Stéphane Nicolet be382bb0cf A better contempt implementation for Stockfish (#1325)
* A better contempt implementation for Stockfish

The round 2 of TCEC season 10 demonstrated the benefit of having a nice contempt implementation: it gives the strongest programs in the tournament the ability to slow down the game when they feel the position is slightly worse, prefering to stay in a complicated (even if slightly risky) middle game rather than simplifying by force into a drawn endgame.

The current contempt implementation of Stockfish is inadequate, and this patch is an attempt to provide a better one.

Passed STC non-regression test against master:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 83360 W: 15089 L: 15075 D: 53196
http://tests.stockfishchess.org/tests/view/5a1bf2de0ebc590ccbb8b370

This contempt implementation is showing promising results in certains situations. For instance, it obtained a nice +30 Elo gain when playing with contempt=40 against Stockfish 7, compared to current master:

• master against SF 7 (20000 games at LTC): +121.2 Elo
• this patch with contempt=40 (20000 games at LTC): +154.11 Elo

This was the result of real cooperative work from the Stockfish team, with key ideas coming from Stefan Geschwentner (locutus2) and Chris Cain (ceebo) while most of the community helped with feedback and computer time.

In this commit the bench is unchanged by default, but you can test at home with the new contempt in the UCI options. The style of play will change a lot when using contempt different of zero (I repeat: not done in this version by default, however)!

The Stockfish team is still deliberating over the best default contempt value in self-play and the best contempt modeling strategy, to help users choosing a contempt value when playing against much weaker programs. These informations will be given in future commits when available :-)

Bench: 5051254

* Remove the prefetch

No functional change.
2017-12-05 07:25:42 +01:00
Joost VandeVondele d193482213 Pawn endgames directly skip early pruning.
Instead of checking individual steps. Idea by @Stefano80.

passed STC
http://tests.stockfishchess.org/tests/view/5a23e5d20ebc590ccbb8b6d5
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 37445 W: 6866 L: 6773 D: 23806

passed LTC
http://tests.stockfishchess.org/tests/view/5a24260c0ebc590ccbb8b716
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 38780 W: 4946 L: 4848 D: 28986

Bench: 5466219
2017-12-04 17:57:36 +01:00
Stéphane Nicolet 54f6ce70fd Avoid warnings by the Clang compiler
Clang gave a couple of warnings for unused parameters after the recnet commit "Use constexpr when makes sense".

No functional change.
2017-12-04 17:53:42 +01:00
syzygy1 822695d4d3 Use a Direction enum for Square deltas
Currently the NORTH/WEST/SOUTH/EAST values are of type Square, but conceptually they are not squares but directions. This patch separates these values into a Direction enum and overloads addition and subtraction to allow adding a Square to a Direction (to get a new Square).

I have also slightly trimmed the possible overloadings to improve type safety. For example, it would normally not make sense to add a Color to a Color or a Piece to a Piece, or to multiply or divide them by an integer. It would also normally not make sense to add a Square to a Square.

This is a non-functional change.
2017-12-04 17:52:31 +01:00
Joost VandeVondele 2acda1fde3 Use bool(Bitboard b) instead of !!b (#1321)
The idiom !!b is confusing newcomers (e.g. Stefan needs explaining here https://groups.google.com/d/msg/fishcooking/vYqnsRI4brY/Gaf60QuACwAJ).

No functional change.
2017-12-03 18:29:55 +01:00
Guy Vreuls 28b6a457c2 Use constexpr when makes sense
No functional change.
2017-12-03 12:44:24 +01:00
Stéphane Nicolet ccd6bad512 Compile without exceptions
Add the -fno-exceptions flag to the Makefile to avoid the unecessary exceptions support in the executable (we do not use any exception in Stockfish at the moment).

This change gives a 9.2% reduction in size for the executable binary.

Before : executable size = 376956 bytes
After: executable size = 347652 bytes

No functional change.
2017-12-03 12:30:09 +01:00
syzygy 8a5a64eac5 Minor cleanup of search.cpp
Four very minor edits. Note that tte->save() uses posKey and
not pos.key() in other places.

Originally I also added a futility_move_counts() function to
make things more consistent with the futility_margin() and
reduction() functions. But then razor_margin[] should probably
also be turned into a function, etc. Maybe a good idea, maybe not.
So I did not include it.

Non functional change.
2017-12-03 12:24:46 +01:00
Stefan Geschwentner a87a1005ad Attack threats
Give bonus for safe attack threats from bishops and rooks on opponent queen

STC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 8629 W: 1599 L: 1438 D: 5592
http://tests.stockfishchess.org/tests/view/5a1ad4490ebc590ccbb8b30d

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 7093 W: 987 L: 846 D: 5260
http://tests.stockfishchess.org/tests/view/5a1aec5d0ebc590ccbb8b317

Bench: 5051254
2017-12-03 10:41:48 +01:00
basepr1me 7dd1f4a7c0 OpenBSD friendly start. 2017-11-18 16:45:33 +01:00
Stefan Geschwentner d64ffd9621 Simplify good/bad capture detection. bench 5336313 2017-11-18 16:27:44 +01:00
Stefano Cardanobile c769d4df84 Fix comments. Bench: 5109559. 2017-11-18 14:08:06 +01:00
Stéphane Nicolet 03a9b3bd8d Simplify away the PawnSet[] imbalance array (#1308)
Simplify away the PawnSet[] imbalance array

STC:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 52977 W: 9550 L: 9484 D: 33943
http://tests.stockfishchess.org/tests/view/5a06b4780ebc590ccbb8a833

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 83717 W: 10599 L: 10559 D: 62559
http://tests.stockfishchess.org/tests/view/5a0aa36a0ebc590ccbb8aa99

Bench: 5340212
2017-11-18 11:24:23 +01:00
Rocky640 53239d7d3d Simplify some kingring penalties expressions
The new "weak" expression helps simplify the safe check calculations for rooks or minors, (but the end result for all the safe checks is the exactly the same as in current master)

The only functional change is for the "outer king ring" (for example, squares f3 g3 h3 when white king is on g1). In current master, there was a 191 penalty if any of these was not defended at all.
With this pr, there is this 191 penalty if any of these is not defended at all or is only defended by a white queen.

Tested as a simplification
STC
http://tests.stockfishchess.org/tests/view/59fb03d80ebc590ccbb89fee
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 66167 W: 12015 L: 11971 D: 42181
(against master (Update Copyright year inMakefile))

LTC
http://tests.stockfishchess.org/tests/view/5a0106ae0ebc590ccbb8a55f
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 15790 W: 2095 L: 1968 D: 11727
(against master (Handle BxN trade as good capture when history scor))

same as #1296 but rebased on latest master
bench: 5109559
2017-11-11 13:37:29 +01:00
ceebo 3f44f5303b Add comments to pos.see_ge()
In terms of technical changes this patch eliminates the return
statements from the main loop of pos.see_ge() and replaces two conditional
computations with a single bitwise negation.

No functional change
2017-11-10 12:14:53 +01:00
VoyagerOne 87452f3a8c Capture Stat Simplification- Bench: 5363761 2017-11-10 12:12:58 +01:00
Gian-Carlo Pascutto 8cfcca12d1 Always do MaxCardinality checks.
Stockfish currently relies on the "filter_root_moves" function also
having the side effect of clamping Cardinality against MaxCardinality
(the actual piece count in the tablebases). So if we skip this function,
we will end up probing in the search even without tablebases installed.

We cannot bail out of this function before this check is done, so move
the MultiPV hack a few lines below.
2017-11-08 13:45:14 +01:00
Joost VandeVondele 0a74c16ffe Simplify Null Move Search condition
Removes depth condition, adjust parameters.

passed STC:
http://tests.stockfishchess.org/tests/view/5a008cbc0ebc590ccbb8a512
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 29282 W: 5317 L: 5210 D: 18755

passed LTC:
http://tests.stockfishchess.org/tests/view/5a00d8530ebc590ccbb8a541
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 26893 W: 3458 L: 3345 D: 20090

Bench: 5015773
2017-11-08 13:44:24 +01:00
Günther Demetz 652199d840 Handle BxN trade as good capture when history score is good
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 19374 W: 3499 L: 3294 D: 12581
http://tests.stockfishchess.org/tests/view/59fc23f50ebc590ccbb8a0bf

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 91030 W: 11680 L: 11274 D: 68076
http://tests.stockfishchess.org/tests/view/59fc43ad0ebc590ccbb8a0d0

Bench: 5482249
2017-11-05 22:05:51 +01:00
Stefan Geschwentner 4bc11984fc Introduce capture history table for capture move sorting
Introduce capture move history table indexed by moved piece,
target square and captured piece type for sorting capture moves.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 11374 W: 2096 L: 1924 D: 7354
http://tests.stockfishchess.org/tests/view/59fac8dc0ebc590ccbb89fc5

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 24791 W: 3196 L: 3001 D: 18594
http://tests.stockfishchess.org/tests/view/59fae4d20ebc590ccbb89fd9

Bench: 5536775
2017-11-03 13:57:18 +01:00
Joost VandeVondele 486c8175c4 Replace easyMove with simple scheme
Reduces time for a stable bestMove, giving some of the won time for the next move.

the version before the pvDraw passed both STC and LTC

passed STC:
http://tests.stockfishchess.org/tests/view/59e98d5a0ebc590ccbb896ec
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 78561 W: 13945 L: 13921 D: 50695
elo =    0.106 +-    1.445 LOS:   55.716%

passed LTC:
http://tests.stockfishchess.org/tests/view/59eb9df90ebc590ccbb897ae
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 29056 W: 3640 L: 3530 D: 21886
elo =    1.315 +-    1.982 LOS:   90.314%

This version, rebased on pvDrawPR with the obvious change, was verified again on STC:

http://tests.stockfishchess.org/tests/view/59ee104e0ebc590ccbb89899
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19890 W: 3648 L: 3525 D: 12717
elo =    2.149 +-    2.895 LOS:   92.692%

and LTC:
http://tests.stockfishchess.org/tests/view/59f9673a0ebc590ccbb89ea0
Total             :    17966
Win               :     2273 (  12.652%)
Loss              :     2149 (  11.961%)
Draw              :    13544 (  75.387%)
Score             :   50.345%
Sensitivity       :    0.014%
2*(W-L)/(W+L)     :    5.608%

LLR  [-3.0,  1.0] :     2.95

BayesElo range    : [  -1.161,   4.876,  10.830] (DrawElo:  341.132)
LogisticElo range : [  -0.501,   2.105,   4.677]
LOS               :   94.369 %

LTC again:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17966 W: 2273 L: 2149 D: 13544
LogisticElo range : [ -0.501, 2.105, 4.677]
LOS : 94.369 %

unchanged bench: 5234652
2017-11-03 13:51:53 +01:00
Cooffe e0d2fdc843 Update Copyright year inMakefile
No functional change.
2017-10-28 12:35:44 +02:00
Joost VandeVondele e50af36a00 Extra thinking before accepting draw PVs.
If the PV leads to a draw (3-fold / 50-moves) position
and we're ahead of time, think a little longer, possibly
finding a better way.

As this is most likely effective at higher draw rates,
tried speculative LTC after a yellow STC:

STC:
http://tests.stockfishchess.org/tests/view/59eb173a0ebc590ccbb8975d
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 56095 W: 10013 L: 9902 D: 36180
elo =    0.688 +-    1.711 LOS:   78.425%

LTC:
http://tests.stockfishchess.org/tests/view/59eba1670ebc590ccbb897b4
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 59579 W: 7577 L: 7273 D: 44729
elo =    1.773 +-    1.391 LOS:   99.381%

bench: 5234652
2017-10-28 12:33:48 +02:00
IIvec 287e2e2f74 Fix premature using of all available time in x/y TC
In x/y time controls there was a theoretical possibility
to use all available time few moves before the clock will
be updated with new time. This patch fixes that issue.

Tested at 60/15 time control:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 113963 W: 20008 L: 20042 D: 73913

The test was done without adjudication rules!

Bench 5234652
2017-10-22 07:43:37 +02:00
Gian-Carlo Pascutto 86ac50403d Don't filter root moves if MultiPV mode is enabled
A band-aid patch to workaround current TB code
limitations with multi PV.

Hopefully this will be removed after committing the
big update of TB impementation, now under discussion.

No functional change.
2017-10-22 07:18:48 +02:00
ceebo 0dc3b0978d Add initiative to trace
No functional change
2017-10-22 07:00:46 +02:00
Joost VandeVondele 9d79138682 Fix issue #1268
If the search is quit before skill.pick_best is called,
skill.best_move might be MOVE_NONE.

Ensure skill.best is always assigned anyhow.

Also retire the tricky best_move() and let the underlying
semantic to be clear and explicit.

No functional change.
2017-10-11 11:47:50 +02:00
Alain SAVARD 43c186c645 Simplify bonus for bishop on long diagonal
Removing 2 conditions, and increase the ThreatbyPawn to compensate.

Passed STC
http://tests.stockfishchess.org/tests/view/59dbde900ebc5916ff64be6d
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14236 W: 2615 L: 2483 D: 9138

Passed LTC
http://tests.stockfishchess.org/tests/view/59dc26470ebc5916ff64be92
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 16552 W: 2136 L: 2010 D: 12406

Bench: 5234652
2017-10-11 11:13:44 +02:00
Niklas Fiekas b36489742b WLDEntryPiece -> WDLEntryPiece for consistency
No functional change.
2017-10-05 11:00:01 +02:00
Stéphane Nicolet 452e5154cf Good bishops on the main diagonals
Bonus in midgame for bishops on long diagonals when the central squares are not occupied by pawns.

Author: ElbertoOne

Passed STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 10801 W: 1955 L: 1786 D: 7060
http://tests.stockfishchess.org/tests/view/59cf5c1d0ebc5916ff64b9da

and LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 83978 W: 10685 L: 10303 D: 62990
http://tests.stockfishchess.org/tests/view/59cf6f6c0ebc5916ff64b9e4

Bench: 5620312
2017-10-02 07:53:28 +02:00
VoyagerOne 07b5a28a68 Decrease reduction for exact PV nodes
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 59004 W: 10621 L: 10249 D: 38134

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 25801 W: 3306 L: 3108 D: 19387

Bench: 5742466
2017-09-30 20:56:27 +02:00
Joost VandeVondele c33af32dad Measure nodes after search finished.
Only affects nmpsec in the multithreaded case.

No functional change.
2017-09-29 16:39:12 +02:00
GuardianRM 0e949ac2c9 Tweak statScore condition
The first change (ss->statScore >= 0) does nothing.

The second change ((ss-1)->statScore >= 0 ) has a massive change.
(ss-1)->statScore is not set until (ss-1) begins to apply LMR to moves.
So we now increase the reduction for bad quiets when our opponent is
running through the first captures and the hash move.

STC
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 57762 W: 10533 L: 10181 D: 37048

LTC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 19973 W: 2662 L: 2480 D: 14831

Bench: 5037819
2017-09-22 16:48:08 +02:00
syzygy ba4e215493 Let ss->ply denote the number of plies from the root to the current node
This patch lets ss->ply be equal to 0 at the root of the search.

Currently, the root has ss->ply == 1, which is less intuitive:

- Setting the rootNode bool has to check (ss-1)->ply == 0.

- All mate values are off by one: the code seems to assume that mated-in-0
  is -VALUE_MATE, mate-1-in-ply is VALUE_MATE-1, mated-in-2-ply is VALUE_MATE+2, etc.
  But the mate_in() and mated_in() functions are called with ss->ply, which is 1 in
  at the root.

- The is_draw() function currently needs to explain why it has "ply - 1 > i" instead
  of simply "ply > i".

- The ss->ply >= MAX_PLY tests in search() and qsearch() already assume that
  ss->ply == 0 at the root. If we start at ss->ply == 1, it would make more sense to
  go up to and including ss->ply == MAX_PLY, so stop at ss->ply > MAX_PLY. See also
  the asserts testing for 0 <= ss->ply && ss->ply < MAX_PLY.

The reason for ss->ply == 1 at the root is the line "ss->ply = (ss-1)->ply + 1" at
the start for search() and qsearch(). By replacing this with "(ss+1)->ply = ss->ply + 1"
we keep ss->ply == 0 at the root. Note that search() already clears killers in (ss+2),
so there is no danger in accessing ss+1.

I have NOT changed pv[MAX_PLY + 1] to pv[MAX_PLY + 2] in search() and qsearch().
It seems to me that MAX_PLY + 1 is exactly right:

- MAX_PLY entries for ss->ply running from 0 to MAX_PLY-1, and 1 entry for the
  final MOVE_NONE.

I have verified that mate scores are reported correctly. (They were already reported
correctly due to the extra ply being rounded down when converting to moves.)

The value of seldepth output to the user should probably not change, so I add 1 to it.
(Humans count from 1, computers from 0.)

A small optimisation I did not include: instead of setting ss->ply in every invocation
of search() and qsearch(), it could be set once for all plies at the start of
Thread::search(). This saves a couple of instructions per node.

No functional change (unless the search searches a branch MAX_PLY deep), so bench
does not change.
2017-09-17 10:44:10 +02:00
Stéphane Nicolet 043a469f83 Score unopposed weak pawns only if majors
Do not use the opposed flag for scoring backward and isolated pawns
in pawns.cpp, instead give a S(5,25) bonus for each opponent unopposed
weak pawns when we have a rook or a queen on the board.

STC run stopped after 113188 games:
LLR: 1.63 (-2.94,2.94) [0.00,5.00]
Total: 113188 W: 20804 L: 20251 D: 72133
http://tests.stockfishchess.org/tests/view/59b58e4d0ebc5916ff64b12e

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 66673 W: 8672 L: 8341 D: 49660
http://tests.stockfishchess.org/tests/view/59b902580ebc5916ff64b231

This is Alain Savard's idea, just with a different bonus.
Original patch there:

green STC, http://tests.stockfishchess.org/tests/view/597dcd2b0ebc5916ff64a09b
yellow LTC, http://tests.stockfishchess.org/tests/view/597ea69e0ebc5916ff64a0e6

Bench: 6259498
2017-09-17 09:52:27 +02:00
IIvec 21926ce2d8 Higher Move Overhead
This shoudl reduce time losses experienced by
users after new time management code.

Verified for no regression in very short TC (4sec + 0.1)
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 35262 W: 7426 L: 7331 D: 20505

Bench 5322108
2017-09-12 12:31:53 +02:00
ianfab ed8286eb1b Extend ShelterWeakness array by dimension isKingFile
Use different penalties for weaknesses in the pawn shelter
depending on whether it is on the king's file or not.

STC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 71617 W: 13471 L: 13034 D: 45112

LTC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 48708 W: 6463 L: 6187 D: 36058

Bench: 5322108
2017-09-09 12:23:34 +02:00
VoyagerOne 3ac47c84d3 Streamlline reduction based on movecount
Use MoveCount History only at quiet moves and simply reduce
reduction by one depth instead of increasing moveCount in formula.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 27511 W: 5171 L: 4919 D: 17421

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 92337 W: 12135 L: 11748 D: 68454

Bench: 6351176
2017-09-05 11:03:50 +02:00
syzygy 741523eda8 Small simplication of see_ge()
Two simplifications:

- Remove the initialisation to 0 of occupied, which is now unnecessary.
- Remove the initial check for nextVictim == KING

If nextVictim == KING, then PieceValue[MG][nextVictim] will be 0, so that
balance >= threshold is true. So see_ge() returns true anyway.

No functional change.
2017-09-05 10:57:10 +02:00
Marco Costalba 04eb87fd08 Travis CI: Make all warnings into errors
Compile with -Werror flag. To make debugging easier
also show compile ourput.

This flag is enabled only in Travis CI, not in the shipped
Makefile becuase we can't test on every possible platform.
2017-09-05 10:40:34 +02:00
Joost VandeVondele 323925b91c Remove unneeded compile options.
In light of issue #1232, a test was performed about the value of '-fno-exceptions' and a second one of the combination '-fno-exceptions -fno-rtti'. It turns out these options are can be removed without introducing slowdown.

STC for removing '-fno-exceptions'
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 13678 W: 2572 L: 2439 D: 8667

STC for removing '-fno-exceptions -fno-rtti' (current patch)
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32557 W: 6074 L: 5973 D: 20510

No functional change.
2017-09-02 16:58:23 +02:00
syzygy 5ba4373522 Prevent Stockfish from exiting if DTZ table is not present
During TB initialisation, Stockfish checks for the presence of WDL
tables but not for the presence of DTZ tables. When attempting to probe
a DTZ table, it is therefore possible that the table is not present.
In that case, Stockfish should neither exit nor report an error.

To verify the bug:
$ ./stockfish
setoption name SyzygyTable value <path_to_WDL_dir>
position fen 8/8/4r3/4k3/8/1K2P3/3P4/6R1 w - -
go infinite
Could not mmap() /opt/tb/regular/KRPPvKR.rtbz
$

(On my system, the WDL tables are in one directory and the DTZ tables
in another. If they are in the same directory, it will be difficult
to trigger the bug.)

The fix is trivial: check the file descriptor/handle after opening
the file.

No functional change.
2017-09-02 09:59:04 +02:00
Marco Costalba 3fd7e37868 Fix a warning with MSVC
warning C4244: '*=': conversion from 'double' to 'int', possible loss of data

No functional change.
2017-09-01 20:23:31 +02:00
Joost VandeVondele 9d95d43c57 Multi-threaded search testing with valgrind
Also check with valgrind the multi-threaded search.

On top of the fix for issue #1227 (PR #1235).

No functional change.
2017-09-01 20:19:43 +02:00
Joost VandeVondele e385f194e9 Fix uninitialized memory usage
After increasing the number of threads, the histories were not cleared,
resulting in uninitialized memory usage.

This patch fixes this by clearing threads histories in Thread c'tor as
is the idomatic way.

This fixes issue 1227

No functional change.
2017-09-01 20:16:56 +02:00
VoyagerOne 7b4c9852e1 Adjust moveCount history only at LMR
STC:
LLR: 3.32 (-2.94,2.94) [-3.00,1.00]
Total: 17584 W: 3277 L: 3131 D: 11176

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26412 W: 3447 L: 3334 D: 19631

Bench: 5417521
2017-08-31 08:53:37 +02:00
Joost VandeVondele bf485f4aff Simplify away non-normal moves in SEE
credit goes to @mstembera for suggesting this approach.
SEE now deals with castling, promotion and en passant in a similar way.

passed STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32902 W: 6079 L: 5979 D: 20844

passed LTC
LLR: 3.92 (-2.94,2.94) [-3.00,1.00]
Total: 110698 W: 14198 L: 14145 D: 82355

Bench: 5713905
2017-08-30 15:02:40 +02:00
Marco Costalba a2b8f91cfa Appveyor: do a Debug and Release build
And set x86 and x64 platforms for real.

Currently this is broken and the same binary is compiled for all platforms.

This is becuase we use a custom build step. OTH the default
build step seems not compatible with cmake generated *sln file.

No functional change.
2017-08-26 11:50:27 +02:00
Joost VandeVondele d5f883ab29 Improve multi-threaded mate finding
If any thread found a 'mate in x' stop the search. Previously only
mainThread would do so. Requires the bestThread selection to be
adjusted to always prefer mate scores, even if the search depth is less.

I've tried to collect some data for this patch. On 30 cores, mate finding
seems 5-30% faster on average. It is not so easy to get numbers for this,
as the time to find a mate fluctuates significantly with multi-threaded runs,
so it is an average over 100 searches for the same position. Furthermore,
hash size and position make a difference as well.

Bench: 5965302
2017-08-26 09:53:34 +02:00
Stéphane Nicolet 92c39522b1 Count all weak squares in the king ring with a single popcount
Passed STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 26966 W: 4993 L: 4745 D: 17228
http://tests.stockfishchess.org/tests/view/599e798a0ebc5916ff64aa8c

and LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 39570 W: 5104 L: 4857 D: 29609
http://tests.stockfishchess.org/tests/view/599ee5230ebc5916ff64aabe

Bench: 5965302
2017-08-26 09:34:27 +02:00
Joost VandeVondele 5ef94eb970 Use moveCount history for reduction
Use less reduction for moves with larger moveCount if your
opponent did an unexpected (== high moveCount) move in the
previous ply... unexpected moves might need unexpected answers.

passed STC:
http://tests.stockfishchess.org/tests/view/599f08cc0ebc5916ff64aace
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 9638 W: 1889 L: 1720 D: 6029

passed LTC:
http://tests.stockfishchess.org/tests/view/599f1e5c0ebc5916ff64aadc
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 28308 W: 3742 L: 3533 D: 21033

Bench: 5747429
2017-08-26 09:30:42 +02:00
Stéphane Nicolet 002bf4d8db Avoid constructing an empty tuple in qsearch
Avoid constructing, passing as a parameter and binding a useless empty tuple of pointers in the qsearch move picker constructor.

Also reformat the scoring function in movepicker.cpp and do some cleaning in evaluate.cpp while there.

No functional change.
2017-08-22 10:16:19 +02:00
Marco Costalba 5ea327d924 Improve appeyor build
Check bench number and do not
hard-code *.cpp file names.

No functional change.
2017-08-20 19:22:11 +02:00
lucasart fe60caba94 Restore safety margin of 60ms
What this patch does is:
* increase safety margin from 40ms to 60ms. It's worth noting that the previous
  code not only used 60ms incompressible safety margin, but also an additional
  buffer of 30ms for each "move to go".
* remove a whart, integrating the extra 10ms in Move Overhead value instead.
  Additionally, this ensures that optimumtime doesn't become bigger than maximum
  time after maximum time has been artificially discounted by 10ms. So it keeps
  the code more logical.

Tested at 3 different time controls:

Standard 10+0.1
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 58008 W: 10674 L: 10617 D: 36717

Sudden death 16+0
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 59664 W: 10945 L: 10891 D: 37828

Tournament 40/10
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 16371 W: 3092 L: 2963 D: 10316

bench: 5479946
2017-08-20 20:00:31 +08:00
Marco Costalba 7aa7dfd4df Fix some Clang warnings
Found by Clang in extra verbose mode :-)

No functional change.
2017-08-19 14:32:31 +02:00
Marco Costalba fa5b0936ee Wide bench coverage
Add tests for:

- Positions with move list
- Chess960 positions

Now bench covers almost all cases, only few endgames
are still out of reach (verified with lcov)

It is a non functionality patch, but bench
changed because we added new test positions.

bench: 5479946
2017-08-19 13:58:12 +02:00
Marco Costalba 45e254a0a0 Restore perft
Rewrite perft to be placed naturally inside new
bench code. In particular we don't have special
custom code to run perft anymore but perft is
just a new parameter of 'go' command.

So user API is now changed, old style command:

$perft 5

becomes

$go perft 4

No functional change.
2017-08-18 09:04:38 -07:00
Marco Costalba 444d99b6d2 Rewrite benchmark
First step in improving bench to handle
arbitrary UCI commands so to test many
more code paths.

This first patch just set the new code
structure.

No functional change.
2017-08-18 09:04:38 -07:00
Marco Costalba e10255339f Reformat time manager code
In particular clarify that 'sd'
parameter is used only in !movesToGo
case.

Verified with Ivan's check tool it is
equivalent to original code.

No functional change.
2017-08-18 08:54:38 -07:00
Joost VandeVondele daf0fe1f57 Collect more corrections to optimum/maximum
The only call site of Time.maximum() corrected by 10.
Do this directly in remaining().

Ponder increased Time.optimum by 25% in init(). Idem.
Delete unused includes.

No functional change.
2017-08-18 08:38:40 -07:00
Marco Costalba 4d511512d2 Speed up Trevis CI
Avoid a couple of redundant rebuilds and compile
with 2 threads since travis gives 2vCPUs.

Also enable -O1 optimization for valgrind and
sanitizers, it should be safe withouth false
positives and it gives a very sensible speed
up, especially with valgrind.

The spee dup allow us to increase testing to
depth 10, useful for thread sanitizer.

No functional change.
2017-08-18 03:07:41 -07:00
Marco Costalba 9c35b9365d Clarify stats range
Current update formula ensures that the
possible value range is [-32 * D, 32 * D].

So we never overflow if abs(32 * D) < INT16_MAX

Thanks to Joost and mstembera to clarify this.

No functional change.
2017-08-18 02:02:35 -07:00
IIvec 01d97521fd Time management simplification
STC (http://tests.stockfishchess.org/tests/view/598188a40ebc5916ff64a21b):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 25363 W: 4658 L: 4545 D: 16160

LTC (http://tests.stockfishchess.org/tests/view/5981d59a0ebc5916ff64a229):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 75356 W: 9690 L: 9640 D: 56026

40/10 TC (http://tests.stockfishchess.org/tests/view/5980c5780ebc5916ff64a1ed):
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19377 W: 3650 L: 3526 D: 12201

15+0 TC (http://tests.stockfishchess.org/tests/view/5982cb730ebc5916ff64a25d):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 5913 W: 1217 L: 1069 D: 3627

This time management handles base time and movestogo cases separatelly. One can test one case without affecting the other. Also, increment usage can be tested separately without (necessarily) affecting sudden death or x moves in y seconds performance.

On stable machines there are no time losses on 0.1+0.001 time control (tested on i7 + Windows 10 platform).

Bench 5608839
2017-08-17 14:42:22 -07:00
Marco Costalba d482e3a890 Fix involuntary conversions of ExtMove to Move
The trick is to create an ambiguity for the
compiler in case an unwanted conversion to
Move is attempted like in:

    ExtMove m1{Move(17),4}, m2{Move(4),17};

    std::cout << (m1 < m2) << std::endl; // 1
    std::cout << (m1 > m2) << std::endl; // 1(!)

This fixes issue #1204

No functional change.
2017-08-17 02:04:00 -07:00
Marco Costalba 9001f55147 Unify stats update()
Now that is a bit bigger makes sense to
unify some duplicated code.

No functional change.
2017-08-17 01:52:26 -07:00
lucasart ae6a4ebf1f Use int16_t in History values
Reduces memory footprint by ~1.2MB (per thread).

Strong pressure: small but mesurable gain
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 258430 W: 46977 L: 45943 D: 165510

Low pressure: no regression
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 73542 W: 13058 L: 13026 D: 47458

Strong pressure + LTC: elo gain confirmed
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 31489 W: 4532 L: 4295 D: 22662

Tested for crashing on overflow and after 70K
games at STC we have only 4 time losses,
possible candidate for an overflow.

No functional change.
2017-08-17 00:32:44 -07:00
Marco Costalba 232c50fed0 Fix incorrect StateInfo
We use Position::set() to set root position across
threads. But there are some StateInfo fields (previous,
pliesFromNull, capturedPiece) that cannot be deduced
from a fen string, so set() clears them and to not lose
the info we need to backup and later restore setupStates->back().
Note that setupStates is shared by threads but is accessed
in read-only mode.

This fixes regression introduced by df6cb446ea

Tested with 3 threads at STC:
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 14436 W: 2304 L: 2196 D: 9936

Bench: 5608839
2017-08-14 23:01:58 -07:00
Marco Costalba c3e964f35e Run clang-tidy 'modernize'
Some warnings after a run of:

$ clang-tidy-3.8 -checks='modernize-*' *.cpp syzygy/*.cpp -header-filter=.* -- -std=c++11

I have not fixed all suggestions, for instance I still prefer
to declare the type instead of a spread use of 'auto'. I also
perfer good old 'typedef' to the new 'using' form.

I have not fixed some warnings in the last functions of
syzygy code because those are still the original functions
and need to be completely rewritten anyhow.

Thanks to erbsenzaehler for the original idea.

No functional change.
2017-08-13 05:46:21 -07:00
Marco Costalba df6cb446ea Thread code reformat
Simplify out low level sync stuff (mutex
and friends) and avoid to use them directly
in many functions.

Also some renaming and better comment while
there.

No functional change.
2017-08-13 04:41:59 -07:00
Marco Costalba bdeda52efd Retire States global variable
And other small touches in uci.cpp

No functional change.
2017-08-12 23:54:48 -07:00
tthsqe12 5837228aa0 Fix the handling of opposite bishops in KXK endgame evaluation
The case of three or more bishops against a long king must look at all of the
bishops and not just the first two in the piece lists. This patch makes sure
that the position is treated as a win when there are bishops on opposite
colors. This functional change is very small because bench remains the same.

LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 24249 W: 4349 L: 4275 D: 15625
http://tests.stockfishchess.org/tests/view/598186530ebc5916ff64a218

Bench: 5608839
2017-08-12 02:47:11 -07:00
Joost VandeVondele 2783203428 Simplify finished search in ponder/infinite mode.
In this rare case (e.g. go infinite on a stalemate),
just spin till ponderhit/stop comes.

The Thread::wait() is a renmant of the old YBWC
code, today with lazy SMP, threads don't need to
wait when outside of their idle loop.

No functional change.
2017-08-10 22:42:52 -07:00
Marco Costalba 66c5eaebd8 Re-apply the fix for Limits::ponder race
But this time correctly set Threads.ponder

We avoid using 'limits' for passing pondering
flag because we don't want to have 2 ponder
variables in search scope: Search::Limits.ponder
and Threads.ponder. This would be confusing also
because limits.ponder is set at the beginning of
the search and never changes, instead Threads.ponder
can change value asynchronously during search.

No functional change.
2017-08-10 12:47:31 -07:00
Marco Costalba 44236f4ed9 Revert "Fix a race on Limits::ponder"
This reverts commit 5410424e3d.

After the commit pondering is broken, so revert for now. I will
resubmit with a proper fix.

The issue is mine, Joost original code is correct.

No functional change.
2017-08-10 10:59:38 -07:00
Joost VandeVondele 5410424e3d Fix a race on Limits::ponder
Limits::ponder was used as a signal between uci and search threads,
but is not an atomic variable, leading to the following race as
flagged by a sanitized binary.

Expect input:
```
 spawn  ./stockfish
 send "uci\n"
 expect "uciok"
 send "setoption name Ponder value true\n"
 send "go wtime 4000 btime 4000\n"
 expect "bestmove"
 send "position startpos e2e4 d7d5\n"
 send "go wtime 4000 btime 4000 ponder\n"
 sleep 0.01
 send "ponderhit\n"
 expect "bestmove"
 send "quit\n"
 expect eof
```

Race:
```
WARNING: ThreadSanitizer: data race (pid=7191)
  Read of size 4 at 0x0000005c2260 by thread T1:

  Previous write of size 4 at 0x0000005c2260 by main thread:

  Location is global 'Search::Limits' of size 88 at 0x0000005c2220 (stockfish+0x0000005c2260)
```

The reason of teh race is that ponder is not just set in UCI go()
assignment but also is signaled by an async ponderhit in uci.cpp:

      else if (token == "ponderhit")
          Search::Limits.ponder = 0; // Switch to normal search

The fix is to add an atomic bool to the threads structure to
signal the ponder status, letting Search::Limits to reflect just
what was passed to 'go'.

No functional change.
2017-08-10 10:46:46 -07:00
Marco Costalba 750dfa0521 Fix some races and clarify the code
Better split code that should be run at
startup from code run at ucinewgame. Also
fix several races when 'bench', 'perft' and
'ucinewgame' are sent just after 'bestomve'
from the engine threads are still running.

Also use a specific UI thread instead of
main thread when setting up the Position
object used by UI uci loop. This fixes a
race when sending 'eval' command while searching.

We accept a race on 'setoption' to allow the
GUI to change an option while engine is searching
withouth stalling the pipe. Note that changing an
option while searchingg is anyhow not mandated by
UCI protocol.

No functional change.
2017-08-10 10:19:56 -07:00
AndyGrant dbc984d9f8 Make variable naming consistent
moved_piece is the only variable in search not using camel case
2017-08-10 10:04:30 -07:00
Joost VandeVondele f731bcadb7 Unify scoring functions in MovePicker
No functional change.
2017-08-10 02:06:21 -07:00
Joost VandeVondele b40e45c1cc Remove Stack/thread dependence in movepick
as a lower level routine, movepicker should not depend on the
search stack or the thread class, removing a circular dependency.
Instead of copying the search stack into the movepicker object,
as well as accessing the thread class for one of the histories,
pass the required fields explicitly to the constructor (removing
the need for thread.h and implicitly search.h in movepick.cpp).
The signature is thus longer, but more explicit:

Also some renaming of histories structures while there.

passed STC [-3,1], suggesting a small elo impact:

LLR: 3.13 (-2.94,2.94) [-3.00,1.00]
Total: 381053 W: 68071 L: 68551 D: 244431
elo =   -0.438 +-    0.660 LOS:    9.7%

No functional change.
2017-08-06 01:45:54 -07:00
snicolet 53c2d9df5e Tweak connected pawns seed[] array values
Raise a little bit the values in the connected pawns seed[] array.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 99033 W: 17939 L: 17448 D: 63646
http://tests.stockfishchess.org/tests/view/597355630ebc5916ff649e3e

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 48044 W: 6371 L: 6099 D: 35574
http://tests.stockfishchess.org/tests/view/597596610ebc5916ff649eba

Bench: 5608839

Closes #1182
2017-08-01 18:41:29 -07:00
Rocky640 b24bd762b2 Rework the "unsupported" penalty into a "supported" bonus
A pawn (according to all the searched positions of a bench run) is not supported 85% of the time,
(in current master it is either isolated, backward or "unsupported").

So it made sense to try moving the S(17, 8) "unsupported" penalty value into the base pawn value hoping for a more representative pawn value, and accordingly
a) adjust backward and isolated so that they stay more or less the same as master
b) increase the mg connected bonus in the supported case by S(17, 0) and let the Connected formula find a suitable eg value according to rank.

Tested as a simplification SPRT(-3, 1)

Passed STC
http://tests.stockfishchess.org/tests/view/5970dbd30ebc5916ff649dd6
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 19613 W: 3663 L: 3540 D: 12410

Passed LTC
http://tests.stockfishchess.org/tests/view/597137780ebc5916ff649de3
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 24721 W: 3306 L: 3191 D: 18224

Bench: 5581946

Closes #1179
2017-08-01 18:37:54 -07:00
VoyagerOne 722e1e0da6 Remove redundant if-statements
No functional change

Closes #1173
2017-07-27 02:14:18 -07:00
mstembera 973ede008a Tuned PSQT using a custom tuner.
bench: 5878420

Closes #1177
2017-07-23 17:35:44 -07:00
VoyagerOne a6ae2d3a31 Simplify aspiration window
Don't modify alpha window on fail-high

Bench: 5875983

Closes #1172
2017-07-23 17:25:23 -07:00
Joost VandeVondele 272e4d1ac7 Faster travis checks
in the last month a couple of timeouts have been seen in travis valgrind testing, leading to undesired false positives. The precise cause of this is unclear: a normal valgrind instrumented run is about 6min, the timeout is 10min. Either there are rare hangs (not reproduced locally), or maybe the actual runtime fluctuates on the travis infrastructure (which uses VMs on AWS as far as I know). This patch leads to roughly a 2x speedup of the instrumented testing by reducing the depth from 10 to 9. If timeouts persist, it needs further analysis.

No functional change.

Closes #1171
2017-07-23 17:23:14 -07:00
Marco Costalba e551afbab7 Move game_phase() to material.cpp
For some reason, although game phase is used
only in material, it is computed in Position.

Move computation to material, where it belongs,
and remove the useless call chain.

No functional change.
2017-07-15 07:28:38 +02:00
Joona Kiiski d31f068312 Revert "Remove questionable gcc flags from profile-build"
This reverts commit 0371a8f8c4.
2017-07-13 16:36:27 -07:00
joergoster 377d77dbe9 Provide selective search depth info for each pv move
No functional change

Closes #1166
2017-07-13 16:30:03 -07:00
Joost VandeVondele 36a93d90f7 Move stop signal to Threads
Instead of having Signals in the search namespace,
make the stop variables part of the Threads structure.
This moves more of the shared (atomic) variables towards
the thread-related structures, making their role more clear.

No functional change

Closes #1149
2017-07-13 16:08:37 -07:00
Joona Kiiski 0371a8f8c4 Remove questionable gcc flags from profile-build
Optimization options for official stockfish should be
consistent, easy, future proof and simple.

We don't want to optimize for any specific version of gcc

No functional change

Closes #1165
2017-07-08 14:20:46 -07:00
GuardianRM c8e5384c3a Queen vs. Minors imbalance
Addition of correction values in case of Imbalance of queens,
depending on the number of light pieces on the side without a queen.

Passed patch:

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 29036 W: 5379 L: 5130 D: 18527

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 13680 W: 1836 L: 1674 D: 10170

Bench: 6258930

Closes #1155
2017-07-08 14:14:14 -07:00
Marco Costalba 802fca6fdd Don't uselessy share rootDepth
It is not needed becuase the only case is a real special
one (bench on depth with many threads) and can be easily
rewritten to avoid sharing rootDepth.

Verified with ThreadSanitizer.

No functional change.

Closes #1159
2017-07-02 22:06:47 -07:00
Marco Costalba 01b6cdb76b Fix some warnings with clang static analyzer
Only one remains (also in tbprobe.cpp), but is bougus.

As a side note, tbprobe.cpp is almost clean, only the last 3
functions probe_wdl(), root_probe() and root_probe_wdl()
are still the original ones and are quite hacky.

Somewhere in the future we will reformat also the last 3
ones. The reason why has not been done before it is because
these functions are really wrong by design and should be
rewritten entirely, not only reformatted.

No functional change.

Closes #1160
2017-07-02 22:02:11 -07:00
Marco Costalba c0cb713a00 Indentation fix in index()
No functional change.

Closes #1158
2017-07-02 22:00:29 -07:00
Alain SAVARD 6d24ef8585 Tidy up
No functional change

Closes #1148
2017-07-02 21:53:45 -07:00
mstembera 69eb391cd7 Magic::index()
Make magic_index() a member of Magic since it uses all it's members
and keep us from having to pass the function pointer around to
init_magics().

No functional change

Closes #1146
2017-06-28 17:11:17 -07:00
Joost VandeVondele 7e897a64f2 Remove race suppression.
Pull #1134 fixed another race, so that can be removed from the thread sanitizer suppressions.

No functional change.

Closes #1150
2017-06-28 17:06:52 -07:00
Marco Costalba 05513a6641 Only main thread checks time
The main change of the patch is that now time check
is done only by main thread. In the past, before lazy
SMP, we needed all the threds to check for available
time because main thread could have been blocked on
a split point, now this is no more the case and main
thread can do the job alone, greatly simplifying the logic.

Verified for regression testing on STC with 7 threads:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 11895 W: 1741 L: 1608 D: 8546

No functional change.

Closes #1152
2017-06-28 17:03:35 -07:00
Marco Costalba fa1e3427bd Simplify pos_is_ok()
Now we don't need anymore the tricky pointer to
show the failed test. Added some few tests too.

Also small rename in see_ge() while there.

No functional change

Closes #1151
2017-06-28 16:54:59 -07:00
VoyagerOne 77342126d8 Increase reduction if tt-move is a capture
The idea is that chances are the tt-move is best and will be difficult to raise alpha when playing a quiet move.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 7582 W: 1415 L: 1259 D: 4908

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 59553 W: 7885 L: 7573 D: 44095

Bench: 5725676

Closes #1147
2017-06-21 14:06:05 -07:00
snicolet 612d93234b Improve readability of evaluation functions
This patch puts the evaluation helper functions inside EvalInfo struct, which simplifies a bit their signature and (most importantly, IMHO) makes their C++ code much cleaner and simpler to read (by removing the "ei." qualifiers all around in evaluate.cpp).

Also rename the EvalInfo struct into Evaluation class to get a natural invocation v = Evaluation(p).value() to evaluation position p.

The downside is an increase of 20 lines in evaluate.cpp (for the prototypes of the helper functions). The upsides are better readability and a speed-up of 0.6% (by generating all the helpers for the NO_TRACE case together, which helps the instruction cache).

No functional change

Closes #1135
2017-06-21 14:01:59 -07:00
VoyagerOne 0149a4c3d6 Update Top CPU - Bench: 6599721
Closes #1145
2017-06-21 13:47:10 -07:00
Joona Kiiski 336901fdb0 Revert "Prefetch earlier in qsearch()"
This reverts commit b73016bb41.

No functional change

Closes #1144
2017-06-21 13:45:03 -07:00
Joost VandeVondele 3cb0200459 Fix four data races.
the nodes, tbHits, rootDepth and lastInfoTime variables are read by multiple threads, but not declared atomic, leading to data races as found by -fsanitize=thread. This patch fixes this issue. It is based on top of the CI-threading branch (PR #1129), and should fix the corresponding CI error messages.

The patch passed an STC check for no regression:

http://tests.stockfishchess.org/tests/view/5925d5590ebc59035df34b9f
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 169597 W: 29938 L: 30066 D: 109593

Whereas rootDepth and lastInfoTime are not performance critical, nodes and tbHits are. Indeed, an earlier version using relaxed atomic updates on the latter two variables failed STC testing (http://tests.stockfishchess.org/tests/view/592001700ebc59035df34924), which can be shown to be due to x86-32 (http://tests.stockfishchess.org/tests/view/592330ac0ebc59035df34a89). Indeed, the latter have no instruction to atomically update a 64bit variable. The proposed solution thus uses a variable in Position that is accessed only by one thread, which is copied every few thousand nodes to the shared variable in Thread.

No functional change.

Closes #1130
Closes #1129
2017-06-21 13:37:58 -07:00
Alain SAVARD 2c237da546 Misc coding style fixes
a few comment and blank fixes.

No functional change

Closes #1141
2017-06-16 19:55:30 -07:00
snicolet b73016bb41 Prefetch earlier in qsearch()
Closes #1139
2017-06-16 19:52:38 -07:00
Marco Costalba 27ba611a3d Better naming in endgame code
And small clean-up of magic bitboards code.

No functional change.

Closes #1138
2017-06-16 19:33:44 -07:00
Brian Sheppard f907d5b7d9 Move depth calculation in probCut
The change passed an STC regression:

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 59350 W: 10793 L: 10738 D: 37819

I verified that there was no change in performance on my machine, but of course YMMV:

Results for 40 tests for each version:

                Base      Test      Diff
        Mean    2014338   2016121   -1783
        StDev   62655     63441     3860

p-value: 0.678
speedup: 0.001

No functional change.

Closes #1137
2017-06-16 19:30:19 -07:00
Joost VandeVondele ebc563059c Call TT.new_search() earlier.
TT.new_search() was being called by mainThread in Thread::search(). However, mainThread is the last to start searching, and helper threads could reach a measured rootDepth 10 (on 64 cores) before mainThread increments the TT generation. Fixed by moving the call to MaintThread::search() before helper threads start searching.

No functional change.

Closes #1134
2017-06-16 19:20:01 -07:00
mstembera 659990b43f Reordering magic data
Gather all magic relevant data into a struct.

This changes memory layout putting everything necessary for processing a single square
in the same memory location thus speeding up access.

Original patch by @snicolet

No functional change.

Closes #1127
Closes #1128
2017-06-06 10:22:12 -07:00
atumanian 6d89d0b64a Don't score as an immediate draw 2-fold repetitions of the root position
In the current version a search stops when the current position is the same as
any position earlier in the search stack,
including the root position but excluding positions before the root.
The new version makes an exception for repeating the root position.

This gives correct scores for moves in the MultiPV > 1 mode.

Fixes #948 (see it for the detailed description of the bug).

LTC: http://tests.stockfishchess.org/tests/view/587910bc0ebc5915193f754b
ELO: 0.38 +-1.7 (95%) LOS: 66.8%
Total: 40000 W: 5166 L: 5122 D: 29712

STC: http://tests.stockfishchess.org/tests/view/5922e6230ebc59035df34a50
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 94622 W: 17059 L: 17064 D: 60499

 LTC: http://tests.stockfishchess.org/tests/view/59273a000ebc59035df34c03
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 61259 W: 7965 L: 7897 D: 45397

Bench: 6599721

Closes #1126
2017-06-06 10:15:13 -07:00
Joost VandeVondele 1781439fc2 use auto& for histories
No functional change.

Closes #1113
2017-06-06 09:56:13 -07:00
Marco Costalba ecd3218b6b History code rewrite (#1122)
Rearrange and rename all history heuristic code. Naming
is now based on chessprogramming.wikispaces.com conventions
and the relations among the various heuristics are now more
clear and consistent.

No functional change.
2017-05-26 08:42:50 +02:00
Nathan Rugg 24df0f72c0 Changed spelling back to "Bishops" in eval output
No functional change.

Closes #1124
2017-05-23 20:27:30 -07:00
VoyagerOne 1d31065e1d Evasion Pruning Tweak
Use moveCount to decide when to prune for evasion pruning

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 24476 W: 4518 L: 4289 D: 15669

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 18362 W: 2476 L: 2298 D: 13588

Bench: 6469989

Closes #1120
2017-05-21 18:27:57 -07:00
snicolet c216dcbe7b Do check analysis later in the game
The previous patch has added a fraction of the king danger score to the
endgame score of the tapered eval, so it seems natural to perform the
king danger computation later in the endgame.

With this patch we extend the limit of such check analysis down to the
material of Rook+Knight, when we have at least two pieces attacking the
opponent king zone.

Passed STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 7446 W: 1409 L: 1253 D: 4784
http://tests.stockfishchess.org/tests/view/591c097c0ebc59035df3477c

and LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 14234 W: 1946 L: 1781 D: 10507
http://tests.stockfishchess.org/tests/view/591c24f10ebc59035df3478c

Bench: 5975183

Closes #1121
2017-05-17 18:24:43 -07:00
snicolet cf893bcded Use a fraction of king danger in endgame score
When SF has an attack on the opponent king in one flank, the huge
midgame -> endgame gradient of the tapered eval prevents us to properly
evaluate neutral exchanges on the other flank as the current king
danger score is a pure midgame term. This may affect SF's ability to
switch to defense in some positions. We add a small contribution
of the king danger to the endgame score to limit this
effect.

Again suggested in the following forum thread:
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/xrUCQ7b0ObE

Passed STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 12719 W: 2371 L: 2192 D: 8156
http://tests.stockfishchess.org/tests/view/5919761a0ebc59035df3468f

And LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 31293 W: 4194 L: 3974 D: 23125
http://tests.stockfishchess.org/tests/view/591980450ebc59035df34695

Bench: 5961548

Closes #1118
2017-05-17 18:19:47 -07:00
Joost VandeVondele 732aa34e3d Fix memory access in Search::clear()
Fixes a bug in Search::clear, where the filling of CounterMoveStats&, overwrote (currently presumably unused) memory because sizeof(cm) returns the size in bytes, whereas elements was needed.

No functional change

Closes #1119
2017-05-17 18:15:01 -07:00
snicolet 862934d7ae Limit king ring to eight squares
In current master the size of the king ring varies abruptly from eight
squares when the king is in g8, to 12 squares when it is in g7. Because
the king ring is used for estimating attack strength, this may lead to
an overestimation of king danger in some positions. This patch limits
the king ring to eight squares in all cases.

 Inspired by the following forum thread:
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/xrUCQ7b0ObE

Passed STC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 9244 W: 1777 L: 1611 D: 5856

and LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 87121 W: 11765 L: 11358 D: 63998

Bench: 6121121

Closes #1115
2017-05-15 19:28:37 -07:00
Joost VandeVondele 7edd1f7ccd Execute an implied ucinewgame at startup
execute an implied ucinewgame upon entering the UCI::loop,
to make sure that searches starting with and without an (optional) ucinewgame
command yield the same search.

This is needed now that seach::clear() initializes tables to non-zero default values.

No functional change

Closes #1101
Closes #1104
2017-05-15 18:54:13 -07:00
Marco Costalba 0c1f119069 Default argument for see_ge()
No functional change.

Closes #1111
2017-05-10 18:20:45 +02:00
Joost VandeVondele 99d914985f Remove int to int conversion, unused include.
No functional change.

Closes #1112
2017-05-09 18:36:32 -07:00
FauziAkram 6b4959e3e0 Linear Protector bonus by distance
Replacing the old Protector table with a simple linear formula which takes into account a different slope for each different piece type.

The idea of this simplification of Protector is originated by Alain (Rocky)

STC: LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 70382 W: 12859 L: 12823 D: 44700

LTC: LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61554 W: 8098 L: 8031 D: 45425

Bench: 6107863

Closes #1099
2017-05-07 21:11:51 -07:00
IIvec ae97941628 King safety and rook mobility parameters tweak
STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 19280 W: 3595 L: 3373 D: 12312

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 221405 W: 28940 L: 28220 D: 164245

Bench: 6506664

Closes #1105
2017-05-07 20:57:29 -07:00
Stefan Geschwentner 69ec09bd4b Bonus for pawn scrifice which create passed pawn
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16752 W: 3141 L: 2944 D: 10667

LTC:
LLR: 3.34 (-2.94,2.94) [0.00,5.00]
Total: 33928 W: 4544 L: 4300 D: 25084

Bench: 5639223

Closes #1092
2017-05-07 20:51:52 -07:00
Marco Costalba 25296547d0 Move Pieces[] out of global visibility
It is an helper array used only in position.cpp

Also small code tidy up while there.

No functional change.

Closes #1106
2017-05-07 20:20:02 -07:00
mstembera 321a27fbe3 Avoid *begin always being included in the sorted list regardless of its value.
This was a minor criticism by @zamar in the original pull request
https://github.com/official-stockfish/Stockfish/pull/1065
necessitating a comment explanation.

No functional change.

Closes #1091
2017-05-07 20:15:56 -07:00
joergoster 8b15961349 Fix multiPV issue #502
In general, this patch handles the cases where we don't have a valid score for each PV line in a multiPV search. This can happen if the search has been stopped in an unfortunate moment while still in the aspiration loop. The patch consists of two parts.

Part 1: The new PVIdx was already part of the k-best pv's in the last iteration, and we therefore have a valid pv and score to output from the last iteration. This is taken care of with:

      bool updated = (i <= PVIdx && rootMoves[i].score != -VALUE_INFINITE);

Case 2: The new PVIdx was NOT part of the k-best pv's in the last iteration, and we have no valid pv and score to output. Not from the current nor from the previous iteration. To avoid this, we are now also considering the previous score when sorting, so that the PV lines with no actual but with a valid previous score are pushed up again, and the previous score can be displayed.

  bool operator<(const RootMove& m) const {
    return m.score != score ? m.score < score : m.previousScore < previousScore; } // Descending sort

I also added an assertion in UCI::value() to possibly catch similar issues earlier.

No functional change.

Closes #502
Closes #1074
2017-05-03 19:46:40 -07:00
Joost VandeVondele e9f26cccdd gcc 7 port
Testing the release candidate revealed only one minor issue, namely a new warning -Wimplicit-fallthrough (part of -Wextra) triggers in the movepicker. This can be silenced by adding a comment, and once we move to c++17 by adding a standard annotation [[fallthrough]];.

No functional change.

Closes #1090
2017-04-30 08:43:43 -07:00
VoyagerOne a18c2c2c3f Don't do InCheck Pruning at the root of QS
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 34603 W: 6441 L: 6167 D: 21995

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 24474 W: 3274 L: 3076 D: 18124

Bench: 5934421

Closes #1089
2017-04-28 20:40:45 -07:00
Rocky640 b948b037a5 Remove cap in kingDanger initialization
Passed STC
http://tests.stockfishchess.org/tests/view/58fd53be0ebc59035df33eb5
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 52048 W: 9397 L: 9329 D: 33322

Passed LTC
http://tests.stockfishchess.org/tests/view/58ff9e0a0ebc59035df33f5c
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 35382 W: 4650 L: 4549 D: 26183

Bench: 5872717

Closes #1087
2017-04-28 20:38:03 -07:00
Marco Costalba e06a117d5e Retire the misdesigned StepAttacks[] array.
StepAttacks[] is misdesigned, the color dependance is specific
to pawns, and trying to generalise to king and knights, proves
neither useful nor convinient in practice.

So this patch reformats the code with the following changes:

- Use PieceType instead of Piece in attacks_() functions

- Use PseudoAttacks for KING and KNIGHT

- Rename StepAttacks[] into PawnAttacks[]

Original patch and idea from Alain Savard.

No functional change.

Closes #1086
2017-04-28 20:33:30 -07:00
Joost VandeVondele b1b19343cd Copy killers in the movepicker
ss->killers can change while the movepicker is active.
The reason ss->killers changes is related to the singular
extension search in the moves loop that calls search<>
recursively with ss instead of ss+1,
effectively using the same stack entry for caller and callee.
By making a copy of the killers,
the movepicker does the right thing nevertheless.

Tested as a bug fix

STC:
http://tests.stockfishchess.org/tests/view/58ff130f0ebc59035df33f37
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 70845 W: 12752 L: 12716 D: 45377

LTC:
http://tests.stockfishchess.org/tests/view/58ff48000ebc59035df33f3d
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 28368 W: 3730 L: 3619 D: 21019

Bench: 6465887

Closes #1085
2017-04-28 20:29:04 -07:00
snicolet 49a9d4cf99 Avoid misuse of StepAttacksBB for pawns
Make it explicit that first index of StepAttacksBB is a piece, not a piece type.

No functional change

Closes #1083
2017-04-25 17:57:49 -07:00
Joost VandeVondele d06a8d0c18 Zero unused constant
No functional change

Closes #1081
2017-04-25 17:21:36 -07:00
Marco Costalba b48439e906 Assorted code style issues
I have removed the check for

 pieceCount[PAWN] > FILE_NB

because totally useless.

No functional change.
2017-04-24 09:49:44 +02:00
Joost VandeVondele 6b9a22b40d Sort moves partially: linear depth dependence
STC: http://tests.stockfishchess.org/tests/view/58f98d260ebc59035df33d5e
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 58958 W: 10862 L: 10485 D: 37611

LTC: http://tests.stockfishchess.org/tests/view/58fa45d40ebc59035df33d86
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 18607 W: 2427 L: 2251 D: 13929

Bench: 6065528

Closes #1079
2017-04-23 08:37:55 -07:00
IIvec 0868de705d King safety parameters improved
STC:
LLR: 2.97 (-2.94,2.94) [0.00,4.00]
Total: 58648 W: 10883 L: 10524 D: 37241

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 52546 W: 7131 L: 6844 D: 38571

Bench 6121479

Closes #1078
2017-04-23 08:04:03 -07:00
Joost VandeVondele 9da3b44ddc Use int instead of Value for history related stats.
history related scores are not related to evaluation based scores.
For example, can easily exceed the range -VALUE_INFINITE,VALUE_INFINITE.
As such the current type is confusing, and a plain int is a better match.

tested for no regression:

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43693 W: 7909 L: 7827 D: 27957

No functional change.

Closes #1070
2017-04-23 07:59:28 -07:00
Joost VandeVondele ced29248c9 simplify logic for history based pruning
STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 34255 W: 6292 L: 6194 D: 21769

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 89914 W: 11769 L: 11739 D: 66406

Bench: 6581936

Closes #1066
2017-04-20 11:28:11 -07:00
Joost VandeVondele 2d96e8fbe9 Partial insertion sort
the order of elements returned by std::partition is implementation defined (since not stable) and could depend on the version of libstdc++ linked.
As std::stable_partition was tested to be too slow (http://tests.stockfishchess.org/tests/view/585cdfd00ebc5903140c6082).
Instead combine partition with our custom implementation of insert_sort, which fixes this issue.
Implementation based on a patch by mstembera (http://tests.stockfishchess.org/tests/view/58d4d3460ebc59035df3315c), which suggests some benefit by itself.
Higher depth moves are all sorted (INT_MIN version), as in current master.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33116 W: 6161 L: 6061 D: 20894

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 88703 W: 11572 L: 11540 D: 65591

Bench: 6256522

Closes #1058
Closes #1065
2017-04-20 11:19:01 -07:00
Stefano Cardanobile d4b9ee0f1d Update Readme.md
Update number of threads.

Closes #1072
2017-04-17 10:38:37 -07:00
Joost VandeVondele 57a3334322 Prefer std::find over a hand-coded loop
tested for no regression.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 37378 W: 6649 L: 6556 D: 24173

No functional change.

Closes #1071
2017-04-17 09:25:27 -07:00
VoyagerOne 9d3ed9ed2e Move-Count Formula Tweak
STC:
LLR: 3.18 (-2.94,2.94) [0.00,4.00]
Total: 55004 W: 10289 L: 9930 D: 34785

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 48184 W: 6401 L: 6128 D: 35655

Bench: 5960754
2017-04-17 09:22:10 -07:00
Stefano80 06175c6055 Remove cap from space score contribution and increase bonus
STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 58462 W: 10615 L: 10558 D: 37289

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 65061 W: 8539 L: 8477 D: 48045

It is worth noting that an attempt to only increase the bonus passed STC but failed LTC, and
an attempt to remove the cap without increasing the bonus is still running at STC, but will probably fail after more than 100k.

Bench: 6188591

Closes #1063
2017-04-17 09:19:36 -07:00
Stéphane Nicolet 1454831220 Doubled and supported pawns
Do not give the doubled pawn penalty when the frontmost pawn is
supported, for instance f2-g2-g3

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 55927 W: 10418 L: 10052 D: 35457
http://tests.stockfishchess.org/tests/view/58eb9fc20ebc59035df33858

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 32078 W: 4257 L: 4035 D: 23786
http://tests.stockfishchess.org/tests/view/58ec48420ebc59035df3388b

Bench: 5995472

Closes #1062
2017-04-16 06:22:48 -07:00
Stefano80 b258b4fee7 Remove minimum to contribution from king danger to score.
STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 24858 W: 4559 L: 4445 D: 15854

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40789 W: 5338 L: 5244 D: 30207

Bench: 7027489

Closes #1059
2017-04-09 07:48:58 -07:00
joergoster 72a501c6fe Fix zugzwang pruning issues
By adding pos.non_pawn_material(pos.side_to_move()) as a precondition in step 13,
which is already in use in Futility Pruning (child node) and Null Move Pruning for similar reasons.

Pawn endgames, especially those with only 1 or 2 pawns, are simply heavily influenced by zugzwang situations.

Since we are using a bitbase for KPK endgames, I see no reason to accept buggy evals as shown in #760

Patch looks neutral at STC
LLR: 2.32 (-2.94,2.94) [-3.00,1.00]
Total: 79580 W: 10789 L: 10780 D: 58011

and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 27071 W: 3502 L: 3390 D: 20179

Bench: 6259071

Closes #1051
Closes #760
2017-04-07 17:15:00 -07:00
VoyagerOne 35b77b120e Standardize stat penalty
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 90631 W: 16325 L: 16323 D: 57983

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 97679 W: 12779 L: 12759 D: 72141

Bench: 6340591

Closes #1053
2017-04-07 17:02:31 -07:00
VoyagerOne ebe021f6a5 Don't update TT at excluded move ply
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 38906 W: 7125 L: 6835 D: 24946

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 28600 W: 3752 L: 3543 D: 21305

Bench:  6861050

Closes #1048
2017-04-02 20:32:54 -07:00
Daniel Dugovic 06eba14dc9 Add assertion for the maximum number of pawns
No functionl change

Closes #1039
2017-03-27 15:55:48 -07:00
Joost VandeVondele c5de4080db Introduce assert for stats update
Make sure updates to the stats are done in a stable way.

No functional change

Closes #1038
Closes #1037
2017-03-25 17:57:07 -07:00
Joost VandeVondele 3b7c1a17e4 Increase maximum number of threads
a single Xeon Phi can present itself as a single numa node with up to 288 threads (4 threads per hardware core).
Tested to work as expected with a Xeon Phi CPU 7230 up to 256 threads.

No functional change

Closes #1045
2017-03-25 10:35:17 -07:00
joergoster afe75571d8 Simplify ThreatBySafePawn scoring
Bench: 6197938

Closes #1047
2017-03-25 10:22:20 -07:00
VoyagerOne 30c583204f Singular extension and check extension tweak
If singular extension fails to trigger extension then don't consider check extension.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 69428 W: 12663 L: 12271 D: 44494

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 44023 W: 5875 L: 5612 D: 32536

Bench: 6170444

Closes #1043
2017-03-25 10:13:25 -07:00
VoyagerOne 352bd6f5aa Skip quiet moves based on moveCount pruning threshold and history stats
If we can moveCountPrune and next quiet move has negative stats,
then go directly to the next move stage (Bad_Captures).

Reduction formula is tweaked to compensate for the decrease in move count that is used in LMR.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 6847 W: 1276 L: 1123 D: 4448

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 48687 W: 6503 L: 6226 D: 35958

Bench: 5919519

Closes #1036
2017-03-18 15:44:49 -07:00
Joost VandeVondele c80d52c845 History stat bonus: Move condition to bonus calculation
about 0.5% speedup.

No functional change

Closes #1034
2017-03-17 14:46:47 -07:00
joergoster c076216a32 Pawns count imbalance table
Instead of having a continuous increasing bonus for our number of pawns when calculating imbalance, use a separate lookup array with tuned values.
Idea by GuardianRM.

STC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16155 W: 2980 L: 2787 D: 10388

LTC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 100478 W: 13055 L: 12615 D: 74808

Bench: 6128779

Closes #1030
2017-03-17 14:41:08 -07:00
Marco Costalba a6d6a2c2fa Assorted code style fixes
No functional change

Closes #1029
2017-03-14 21:02:21 -07:00
mstembera d01b66ae8f Fix pawn entry prefetch
No functional change

Closes #1026
2017-03-14 20:56:26 -07:00
snicolet c3d2e6aba9 Helper functions to count material for both sides
Syntactic sugar: helper functions to count material or pieces for both sides.

No functional change

Closes #1025
2017-03-08 18:45:38 -08:00
Joost VandeVondele d490bb9973 Always have counterMoves associated
Simplifies away all associated checks, leading to a ~0.5% speedup.
The code now explicitly checks if moves are OK, rather than using nullptr checks.

Verified for no regression:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32218 W: 5762 L: 5660 D: 20796

No functional change

Closes #1021
2017-03-08 18:35:23 -08:00
pb00068 cc76524c2e Further simplify skipping of plies with threads
No functional change

Closes #1020
2017-03-08 18:03:01 -08:00
VoyagerOne 3627348e2b Allow pruning advance pawn pushes if not near endgame
STC:
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 101088 W: 18016 L: 17717 D: 65355

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 61194 W: 8108 L: 7791 D: 45295

Bench: 5803228

Closes #1023
2017-03-05 18:58:06 -08:00
snicolet 728ce2195c Speed-up some arrays reading
This patch removes the empty rows at the beginning and at the end of
MobilityBonus[] and Protector[] arrays:

• reducing the size of MobilityBonus from 768 bytes to 512 bytes
• reducing the size of Protector from 1024 to 512 bytes

Also adds some comments and cleaner code for the arrays in pawns.cpp

No speed penalty (measured speed-up of 0.4%).

No functional change.

Closes #1018
2017-03-05 18:20:27 -08:00
Joost VandeVondele 1810c4d758 Simplify skipping of plies with helper threads
Replaces the HalfDensity array with an equivalent, compact implementation.
Includes suggestions by mcostalba & snicolet.

No functional change

Closes #1004
2017-02-26 16:41:58 -08:00
snicolet 8f7e032b8c Change definition of "weak" in threats calculation
By defining "strongly protected" as "protected by a pawn, or protected
by two pieces and not attacked by two enemy pieces".

Passed STC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 17050 W: 3128 L: 2931 D: 10991

Passed LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 120995 W: 15852 L: 15343 D: 89800

Bench : 6269229

Closes #1016
2017-02-25 17:43:54 -08:00
mstembera f1e3dfea74 Reorder members of Material::Entry
This eliminates alignment padding and reduces size from 48 to 40 bytes.
This makes the material HashTable smaller and more cache friendly.

No functional change

Closes #1013
2017-02-23 21:33:03 -08:00
GuardianRM 9f48e1ec15 Pieces protecting king
Initial protective idea by Snicolet for knight, for other pieces too
Patch add penalties and bonuses for pieces, depending on the distance from the own king

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 21192 W: 3919 L: 3704 D: 13569

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26177 W: 3642 L: 3435 D: 19100

Bench : 6687377

Closes #1012
2017-02-23 21:26:59 -08:00
snicolet eefbe967c2 Keep pawns on both flanks
Positions with pawns on only one flank tend to be more drawish. We add
a term to the initiative bonus to help the attacking player keep pawns
on both flanks.

STC: yellowish run stopped after 257137 games
LLR: -0.92 (-2.94,2.94) [0.00,5.00]
Total: 257137 W: 46560 L: 45511 D: 165066

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 15602 W: 2125 L: 1956 D: 11521

Bench : 6976310

Closes #1009
2017-02-19 14:27:03 -08:00
FauziAkram c243cd5f4a Variable tuning
A tuning patch which cover the following changes:

increase the importance of queen and rook mobility in endgame and
decrease it in mg, since if we use the heavy pieces too early in the game
we will just make opponent develop their pieces by threatening ours.

King Psqt:
1)King will be encouraged more to stay in the first ranks in the MG
2)and will be encouraged more to go to the middle of the board/last ranks in the EG

Bishop scale better in EG
Logical changes on various psqt tables
1/6 of the changes of the last tuning session on mobility tables

STC: LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 227879 W: 41240 L: 40313 D: 146326
LTC : LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 167047 W: 21871 L: 21291 D: 123885

Bench: 5695960

Closes #1008
2017-02-19 14:00:44 -08:00
VoyagerOne 05cf45f2d1 Razor Simplification
Remove code that restrict using tt-moves for razoring.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 67442 W: 12039 L: 11997 D: 43406

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 38541 W: 5044 L: 4947 D: 28550

Bench: 5667216

Closes #1002
2017-02-18 22:50:37 -08:00
torfranz faedcf08a8 Retire loose enemies bonus
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 44727 W: 7943 L: 7862 D: 28922

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 148343 W: 19044 L: 19088 D: 110211

Bench: 5669076

Closes #1005
2017-02-14 21:26:08 -08:00
VoyagerOne 83fb4547f6 search(): Move nullValue variable into local scope
No functional change

Closes #1003
2017-02-14 21:22:58 -08:00
Joost VandeVondele 1e814e0ca0 Fix makefile: 32 bit builds without optimization.
Fixes failing build for

make ARCH=x86-32 clean && make ARCH=x86-32 optimize=no build

by passing -m32 also to the link step.

Extend travis testing accordingly.

No functional change.

Closes #999
2017-02-14 21:11:44 -08:00
torfranz e0d91f4c44 Retire small bonus in passed pawn evaluation
STC: http://tests.stockfishchess.org/tests/view/5899824d0ebc59099759f3ee
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 33370 W: 6061 L: 5961 D: 21348

LTC: http://tests.stockfishchess.org/tests/view/5899e3820ebc59099759f415
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 36206 W: 4664 L: 4563 D: 26979

Bench: 6072262

Closes #998
2017-02-10 16:52:34 -08:00
Joona Kiiski a753e20bd4 A small tweak in doEasyMove()
Time.elapsed() > Time.optimum() * 5 / 44
instaed of:
Time.elapsed() > Time.optimum() * 5 / 42

This was yellow on STC:
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 156856 W: 28317 L: 27942 D: 100597

Passed on LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 36909 W: 4926 L: 4682 D: 27301

Note: Patch was originally submitted by user GuardianRM.
However his repo was deleted before merge.

No functional change

Closes #995
2017-02-10 16:44:13 -08:00
Stefano80 5205d44f87 Simplify scale factor computation
Minor non-functional simplifications in computing the scale factor.

In my opinion, the code is now slightly more readable:

- remove one condition which can never be satisfied.
- immediately return instead of assigning the sf variable.

Tested for non-regression:

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 62162 W: 11166 L: 11115 D: 39881

No functional change

Closes #992
2017-02-05 16:06:37 -08:00
VoyagerOne 0553b46829 Simplify Queen Mobility
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 85997 W: 15550 L: 15540 D: 54907

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 25333 W: 3370 L: 3256 D: 18707

Bench: 6459194

Closes #991
2017-02-05 15:40:30 -08:00
Stéphane Nicolet ddecdc97d7 Simplify away QueenContactChecks
Changing the definition of safe checks to include
squares protected only by the king, but twice
attacked by the opponent.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 8691 W: 1683 L: 1541 D: 5467
http://tests.stockfishchess.org/tests/view/588f53b50ebc5915193f7dc7

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 31266 W: 4150 L: 4043 D: 23073
http://tests.stockfishchess.org/tests/view/588f68ab0ebc5915193f7dda

Bench : 5885815
2017-01-31 09:54:38 +01:00
VoyagerOne fa24cc25a4 Simplify TT penalty stat (#980)
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 20251 W: 3692 L: 3570 D: 12989

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 16432 W: 2155 L: 2029 D: 12248

Bench: 5941174
2017-01-29 08:54:58 +01:00
VoyagerOne 5254a6040c Penalty for a quiet ttMove that fails low
Also the penalty/bonus function is misleading, we
should simply change it to stat_bonus(depth) for
bonus and -stat_bonus(depth+ ONE_PLY) for extra
penalty.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 11656 W: 2183 L: 2008 D: 7465

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 11152 W: 1531 L: 1377 D: 8244

Bench: 6101931
2017-01-28 09:32:07 +01:00
Stefan Geschwentner 471f7a1b5c Candidate passed pawns
Detect safe candidate passers.

STC: http://tests.stockfishchess.org/tests/view/5882395c0ebc5915193f78b3
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 53569 W: 9925 L: 9570 D: 34074

LTC: http://tests.stockfishchess.org/tests/view/5882b4fb0ebc5915193f78e2
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 77576 W: 10387 L: 10014 D: 57175

Bench: 5325829
2017-01-28 09:04:24 +01:00
pb00068 58c181de9a Simplify away pinnedPieces bitboard in EvalInfo (#975)
Results for 20 tests for each version (pgo-builds):

            Base      Test      Diff      
    Mean    2110519   2118116   -7597     
    StDev   8727      4906      10112     

p-value: 0,774
speedup: 0,004

Further verified for no regression:
http://tests.stockfishchess.org/tests/view/5885abd10ebc5915193f79e6
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 21786 W: 3959 L: 3840 D: 13987

No functional change
2017-01-28 08:43:54 +01:00
Alain SAVARD cf4a38e0cb Simplification of lazy threshold
Passed STC
http://tests.stockfishchess.org/tests/view/587846c10ebc5915193f74ec
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 217236 W: 39041 L: 39254 D: 138941

Passed LTC
http://tests.stockfishchess.org/tests/view/587e157a0ebc5915193f76e7
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 52396 W: 6883 L: 6804 D: 38709

This submitted version (using if (abs(mg + eg) > 1500) )
seems more logical than the following other green simplification (using if (abs(mg)>1500))
since it can happen than mg_value is > eg_value (about 20% of the time)
and the submitted version seems stronger at LTC

STC
http://tests.stockfishchess.org/tests/view/5879702d0ebc5915193f7585
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39958 W: 7315 L: 7227 D: 25416

LTC
http://tests.stockfishchess.org/tests/view/5879af3e0ebc5915193f7592
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 204322 W: 26529 L: 26648 D: 151145

bench: 6406285
2017-01-21 10:47:54 +01:00
Alain SAVARD 9eed183489 Outpost array simplification
The ReachableOutpost values were almost exactly half the Outpost values.

Passed STC
http://tests.stockfishchess.org/tests/view/588020510ebc5915193f781e
LLR: 3.86 (-2.94,2.94) [-3.00,1.00]
Total: 119238 W: 21462 L: 21460 D: 76316

Passed LTC
http://tests.stockfishchess.org/tests/view/5880ae090ebc5915193f7843
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 23540 W: 3097 L: 2980 D: 17463

Curiously, using a division by 2, with slightly different values, did not passed
http://tests.stockfishchess.org/tests/view/587fece00ebc5915193f780a

bench: 5828283
2017-01-21 10:36:46 +01:00
Alain SAVARD 243a9f5484 Reformat eval_init()
Move more code into eval_init, removing some
clutter in the main routine.

Write eval_init only from "our" point of view
(do not init the attackedBy[Them] bitboards).

Add mobilityArea to the evalinfo

A few edits while being there

tested for non-regression at STC
http://tests.stockfishchess.org/tests/view/587fab230ebc5915193f77d9
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39585 W: 7183 L: 7094 D: 25308

Non functional change.
2017-01-21 10:13:49 +01:00
Stéphane Nicolet 9f8f093fd6 Update some comments (#973)
Use somewhat more precise comments in a couple of places.

No functional change.
2017-01-17 14:50:03 +01:00
Stefano Cardanobile de02768af7 Introduce lazy evaluation
After we have taken into account all cheap evaluation
terms, we check whether the score exceeds a given threshold.
If this is the case, we return a scaled down evaluation.

STC:
LLR: 3.35 (-2.94,2.94) [0.00,5.00]
Total: 12575 W: 2316 L: 2122 D: 8137

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 67480 W: 9016 L: 8677 D: 49787

Current version is the one rewritten by ceebo
further edited by me.

Bench: 5367704
2017-01-13 09:17:48 +01:00
loco-loco 99cd513264 Removing CM parameter from the Stats struct. (#970)
After the history simplifications, we are only using Value Stats for CounterMoveHistory table. Therefore the parameter CM is not necessary.

No functional change.
2017-01-12 08:46:46 +01:00
Stéphane Nicolet d2971f3fca Check for overflow in Score * int multiplication (#969)
Add asserts to check for overflow in Score * int multiplication.

There is no overflow in current master, but it would be easy to
create one as the scale of the current eval does not leave many
spare bits. For instance, adding the following unused variables
in master at the end of evaluate() (line 882 of evaluate.cpp)
overflows:

Score s1 = score * 4;  // no overflow
Score s2 = score * 5;  // overflow

Assertion failed: (eg_value(result) == (i * eg_value(s))),
function operator*, file ./types.h, line 336.

Same md5 checksum as current master for non debug compiles.

No functional change.
2017-01-11 18:11:17 +01:00
Rocky640 d40351243b StormDanger CleanUp (#964)
Order the enum and the array the same way they appear around line 250.
Makes it much easier to follow.

Add comments in the array definition and critical rows.
Use same terminology as elsewhere in pawns.cpp

No functional change.
2017-01-11 08:56:38 +01:00
Joost VandeVondele d8f683760c Adjust copyright headers to 2017 (#965)
No functional change.
2017-01-11 08:46:29 +01:00
Marco Costalba 332b5013b5 Travis: fix bench fetch in case of PR (#968)
When Travis tests a PR, a commit merge is created
but master branch is not updated, although HEAD is.

No functional change.
2017-01-11 08:44:06 +01:00
lucasart 34e47ca87d Rename FromTo -> History (#963)
Previously, we had duplicated History:

- one with (piece,to) called History
- one with (from,to) called FromTo

Now that we have only one, rename it to History, which is the generally accepted
name in the chess programming litterature for this technique.

Also correct some comments that had not been updated since the introduction of CMH.

No functional change.
2017-01-10 08:47:56 +01:00
lucasart e0504ab876 Remove HistoryStats
STC:
LLR: 3.44 (-2.94,2.94) [-3.00,1.00]
Total: 120831 W: 21572 L: 21594 D: 77665

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 26565 W: 3519 L: 3406 D: 19640

bench 5920493
2017-01-09 15:50:12 +01:00
Marco Costalba d9dd520896 Fix previous patch for OS X (#961)
Use posix version of sed that is available on all
platforms.

No functional change.
2017-01-09 15:37:09 +01:00
Marco Costalba 394e9cd892 Trevis CI: use commit bench number as a reference
No functional change.
2017-01-09 12:51:28 +01:00
Joost VandeVondele ba15781be8 New shell scripts for testing, used for travis CI (#957)
Perform more complex verification and validation.

- signature.sh : extract and optionally compare Bench/Signature/Node count.
- perft.sh : verify perft counts for a number of positions.
- instrumented.sh : run a few commands or uci sequences through valgrind/sanitizer instrumented binaries.
- reprosearch.sh : verify reproducibility of search.

These script can be used from directly from the command line in the src directory.

Update travis script to use these shell scripts.

No functional change.
2017-01-09 10:30:57 +01:00
ElbertoOne d39ffbeea6 Simplified select best thread (#958)
Only select best thread if score is better and depth equal or larger.

STC (7 threads): http://tests.stockfishchess.org/tests/view/586a4d090ebc5903140c64b2
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 120297 W: 18652 L: 18682 D: 82963

LTC (7 threads): http://tests.stockfishchess.org/tests/view/586e31b30ebc5903140c663d
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 57187 W: 7035 L: 6959 D: 43193

bench: 4940355
2017-01-09 10:26:50 +01:00
Marco Costalba 3ab3e55bb5 Drop Stats c'tors
Now taht we correctly value-initialize Thread objects,
we don't need c'tors anymore because tables will be
zero-initialized by the compier when Thread object
is instanced.

Verified that we have no errors with Valgrind.

No functional change.
2017-01-07 10:14:39 +01:00
Joost VandeVondele 1c316c41bb Correctly zero-initialize MainThread
It can be used uninitialized in time management.
Fixes all valgrind errors on './stockfish go wtime 8000 btime 8000 winc 500 binc 500'

This is one (of the many) quirks of C++. There is a subtle difference between:

new Foo
new Foo()

The first statement calls the default constructor (that in case of a POD leaves data members
uninitialized), the second one performs a value-initialization (that in case of POD is
equivalent to a zero-initialization)

See:
http://stackoverflow.com/questions/620137/do-the-parentheses-after-the-type-name-make-a-difference-with-new
http://stackoverflow.com/questions/5116541/difference-between-creating-object-with-or-without

No functional change.
2017-01-07 10:02:43 +01:00
Joost VandeVondele 90b052462c Zero init fromToStats in constructor. (#953)
Extend commit fe99de to fromToStats, which fixes the last valgrind errors on 
a simple 'go depth 12' at startup.

No functional change.
2017-01-06 10:43:18 +01:00
pb00068 8b2c81d3ea Rejoin lines that belong to HalfDensity map (#952)
No functional change.
2017-01-05 09:00:41 +01:00
Marco Costalba fe99de20ff Correct zero-init of Thread data members
If not explicitly initialized in a class constructor,
then all data members are default-initialized when
the corresponing struct/class is instanced.

For array and built-in types (int, char, etc..)
default-initialization is a no-op and we need to
explicitly zero them.

No functional change.
2017-01-05 08:50:17 +01:00
Joost VandeVondele 6b16ebc825 Use consistent variable names for counterMoveStats (#949)
Unify naming in movepick and search, by adopting the latter convention (cmh,fmh,fmh2).

No functional change.
2017-01-02 09:36:43 +01:00
Stéphane Nicolet e3a8b8bcff Simplify unstoppable again (#950)
Assign a small bonus for our passed pawns when the opponent has no
pieces left.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 51124 W: 9036 L: 8966 D: 33122

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 114560 W: 14604 L: 14604 D: 85352

Bench: 4940355
2017-01-02 09:33:40 +01:00
Jonathan Calovski 1052ce74f6 Tweak best thread selection logic
STC 7 threads:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26881 W: 4161 L: 3941 D: 18779
http://tests.stockfishchess.org/tests/view/58667a830ebc5903140c632f

LTC 7 threads:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 22988 W: 2767 L: 2583 D: 17638
http://tests.stockfishchess.org/tests/view/586722690ebc5903140c636d

bench: 5468995
2017-01-01 11:28:50 +01:00
lucasart e258c5a779 WDL: rename WDLCursedLoss into WDLBlessedLoss
Tested using syzygy bench method:

- 2016 random positions ranging between 3 and 10 pieces
- each searched using bench at depth=10

Same node count (and no speed regression).

No functional change.
2017-01-01 11:11:52 +01:00
Aram Tumanian b7b9d7c9c7 Don't clear EasyMove in search()
EasyMove is cleared after every iteration of the
search if the 3rd move in the PV of the main thread
changes from the previous iteration. Therefore,
clearing EasyMove during a search iteration may be
excessive. The tests show that this is indeed unnecessary.
In the new version the EasyMove variable is used only in
the Thread::search function.

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 47719 W: 8438 L: 8362 D: 30919

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 122841 W: 15448 L: 15457 D: 91936

bench: 5468995
2017-01-01 11:04:24 +01:00
Sergei Antonov 881a9dfb0a Threefold repetition detection
Implement a threefold repetition detection. Below are the examples of
problems fixed by this change.

    Loosing move in a drawn position.
    position fen 8/k7/3p4/p2P1p2/P2P1P2/8/8/K7 w - - 0 1 moves a1a2 a7a8 a2a1
    The old code suggested a loosing move "bestmove a8a7", the new code suggests "bestmove a8b7" leading to a draw.

    Incorrect evaluation (happened in a real game in TCEC Season 9).
    position fen 4rbkr/1q3pp1/b3pn2/7p/1pN5/1P1BBP1P/P1R2QP1/3R2K1 w - - 5 31 moves e3d4 h8h6 d4e3
    The old code evaluated it as "cp 0", the new code evaluation is around "cp -50" which is adequate.

Brings 0.5-1 ELO gain. Passes [-3.00,1.00].

STC: http://tests.stockfishchess.org/tests/view/584ece040ebc5903140c5aea
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 47744 W: 8537 L: 8461 D: 30746

LTC: http://tests.stockfishchess.org/tests/view/584f134d0ebc5903140c5b37
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 36775 W: 4739 L: 4639 D: 27397

Patch has been rewritten into current form for simplification and
logic slightly changed so that return a draw score if the position
repeats once earlier but after or at the root, or repeats twice
strictly before the root. In its original form, repetition at root
was not returned as an immediate draw.

After retestimng testing both version with SPRT[-3, 1], both passed
succesfully, but this version was chosen becuase more natural. There is
an argument about MultiPV in which an extended draw at root may be sensible.
See discussion here:

   https://github.com/official-stockfish/Stockfish/pull/925

For documentation, current version passed both at STC and LTC:

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 51562 W: 9314 L: 9245 D: 33003

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 115663 W: 14904 L: 14906 D: 85853

bench: 5468995
2017-01-01 10:56:46 +01:00
Alain SAVARD 43f6b33e50 Small eval cleanup and renaming
Non-functional changes

a) splitting the threat array to avoid using an enum
b) reorder the scores according to functions where they are used.
c) declarations in evaluate_pieces after the const(s) like elsewhere
d) more compact definitions of KingFlank,
now that we need it also for the PanwLessFlank penalty.
e) reuse CenterFiles in evaluate_space
f) move one line inside next popcount

No functional change.
2016-12-31 14:15:57 +01:00
lucasart ab4f498bbc Remove SafeCheck (#946)
It was a bit of a hack, without intrinsic value, but rather compensating for the
fact that checks were mistuned.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 88308 W: 15553 L: 15545 D: 57210

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 53115 W: 6741 L: 6662 D: 39712

bench 5468995
2016-12-31 13:52:42 +01:00
Jörg Oster 8765f9ce16 Further simplify unstoppable (#938)
By finally moving it into passed pawns eval.

Tested for no regression:
STC
LLR: 3.25 (-2.94,2.94) [-3.00,1.00]
Total: 57109 W: 10023 L: 9947 D: 37139

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21511 W: 2800 L: 2680 D: 16031

Bench: 5255881
2016-12-25 10:44:56 +01:00
hxim 3728e833aa Fix psqt format and use smaller numbers in king psqt (#940)
Fix minus sign in pawn psqt and use smaller numbers in king psqt.

No functional change.
2016-12-25 10:40:17 +01:00
Joost VandeVondele af7412e58a Explicitly use alpha+1 for beta in NonPV search (#939)
Fixes the only exception, in razoring.

The code already does assert(PvNode || (alpha == beta - 1)), and it can be verified by studying the program flow that this is indeed the case, also for the modified line.

No functional change.
2016-12-25 10:34:48 +01:00
Joost VandeVondele 1ceaea701b Simplify threshold handling for probcut. (#936)
Just use greater equal as this is what see_ge does now.

passed STC
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 226506 W: 39755 L: 39978 D: 146773

passed LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 138483 W: 17450 L: 17479 D: 103554

Bench: 5212921
2016-12-22 16:02:32 +01:00
Stefano Cardanobile f72b7dc99a piecesCount (#932)
All counts in search.cpp are of the form xxxCount. Conform piecesCnt to this unwritten rule.

No functional change.
2016-12-20 11:18:19 +01:00
Joost VandeVondele ee22b61f5e Use DEPTH_ZERO initializer for depth in qsearch (#931)
Simplifies the main search function.

No functional change.
2016-12-20 11:17:38 +01:00
VoyagerOne 8c61bbda54 Another simplification for SEE pruning
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 29310 W: 5225 L: 5118 D: 18967

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 128454 W: 16424 L: 16442 D: 95588

Bench: 4556848
2016-12-17 08:37:58 +01:00
Stefan Geschwentner 847bc0e80f Simplify pruning
STC: http://tests.stockfishchess.org/tests/view/5842be140ebc5903140c5619
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 16014 W: 2839 L: 2710 D: 10465

LTC: http://tests.stockfishchess.org/tests/view/584316a50ebc5903140c5638
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 182360 W: 22830 L: 22914 D: 136616

Retested at LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 45502 W: 5821 L: 5732 D: 33949

Bench: 4684146
2016-12-12 12:01:16 +01:00
Joost VandeVondele 1b62d413c1 Clean-up skipEarlyPruning (#921)
make skipEarlyPruning a search argument instead of managing this by hand.

Verified for no regression at STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 96754 W: 17089 L: 17095 D: 62570

No functional change.
2016-12-11 20:05:25 +01:00
Jonathan Calovski 589049a0e5 Simplify unstoppable condition
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 86389 W: 15165 L: 15153 D: 56071

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 98143 W: 12311 L: 12288 D: 73544

Bench: 5437987
2016-12-11 19:57:02 +01:00
Joost VandeVondele a47bbca0ea Refactor bonus and penalty calculation (#917)
* Refactor bonus and penalty calculation

Compute common terms in a helper function.

No functional change.

* Further refactoring

Remove some parenthesis that are now useless.
Define prevSq once, use repeatedly.

No functional change.

bench: 5884767 (bench of previous patch is wrong)
2016-12-05 18:58:12 +01:00
Stéphane Nicolet 0d33466bcd Pawn flank attacks
This patch tweaks some pawn values to favor flank attacks.

The first part of the patch increases the midgame psqt values of external pawns to launch more attacks (credits to user GuardianRM for this idea), while the second part increases the endgame connection values for pawns on upper ranks.

Passed STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 34997 W: 6328 L: 6055 D: 22614

and LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 13844 W: 1832 L: 1650 D: 10362

Bench: 5884767
2016-12-05 18:49:07 +01:00
ElbertoOne 46d066b041 Remove piece condition in decrease lmr reduction check
STC: http://tests.stockfishchess.org/tests/view/584154780ebc5903140c55cf
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40866 W: 7251 L: 7164 D: 26451

LTC: http://tests.stockfishchess.org/tests/view/5841e6e50ebc5903140c5605
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 62413 W: 7948 L: 7881 D: 46584

Bench: 5807913
2016-12-04 09:28:55 +01:00
Andrey Neporada bf8b45fe63 Help GCC to optimize msb() to single instruction
GCC compiles builtin_clzll to “63 ^ BSR”. BSR is processor instruction "Bit Scan Reverse".
So old msb() function is basically 63 - 63 ^ BSR.
Unfortunately, GCC fails to simplify this expression.

Old function compiles to

    bsrq    %rdi, %rdi
    movl    $63, %eax
    xorq    $63, %rdi
    subl    %edi, %eax
    ret

New function compiles to

    bsrq    %rdi, %rax
    ret

BTW, Clang compiles both function to the same (optimal) code.

No functional change.
2016-12-03 09:37:07 +01:00
goodkov e70da0d2eb Simplify pruning rule
STC: http://tests.stockfishchess.org/tests/view/583df86d0ebc5903140c5481
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43283 W: 7761 L: 7678 D: 27844

LTC: http://tests.stockfishchess.org/tests/view/583f42670ebc5903140c5525
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 56587 W: 7232 L: 7157 D: 42198

bench: 5084980
2016-12-02 09:04:45 +01:00
Joost VandeVondele 1e76ba7cec WeakQueen Parameter tweak
New tuned values.

passed STC
http://tests.stockfishchess.org/tests/view/5834573c0ebc5903140c507b
LLR: 3.16 (-2.94,2.94) [0.00,4.00]
Total: 157415 W: 27917 L: 27227 D: 102271

passed LTC
http://tests.stockfishchess.org/tests/view/58388d2b0ebc5903140c523b
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 95405 W: 12350 L: 11959 D: 71096

Bench: 4912054
2016-12-01 14:55:00 +01:00
ElbertoOne 535435b7fc TrappedRook simplification
Just remove rank checks for rook and king for TrappedRook evaluation.

STC: http://tests.stockfishchess.org/tests/view/5833fdfc0ebc5903140c5050
LLR: 3.03 (-2.94,2.94) [-3.00,1.00]
Total: 34474 W: 6088 L: 5986 D: 22400

LTC: http://tests.stockfishchess.org/tests/view/58392f3f0ebc5903140c5276
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 48358 W: 6221 L: 6136 D: 36001

bench: 5536128
2016-11-27 17:42:20 +01:00
theo77186 7a3844e6ef Fix PGO build with GCC (#904) 2016-11-27 14:43:52 +01:00
Michael Byrne fbb2ffacfd Fix PGO Build for clang
This fixes https://github.com/official-stockfish/Stockfish/issues/167.

Additional improvments by Joost VandeVondele.
2016-11-27 10:03:52 +01:00
joergoster 8f30d233f8 Fix trace in case of space evaluation
We only compute space eval during the opening/early midgame.
Apply the same logic for DoTrace.

No functional change.
2016-11-27 09:28:32 +01:00
Marco Costalba ec83e8a72c Fix regression: print const position
Fix a regression introduced with new TB code.

No functional change.
2016-11-27 09:11:56 +01:00
mbootsector e7289465b9 Rank based threats
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 19404 W: 3581 L: 3374 D: 12449

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 16204 W: 2194 L: 2023 D: 11987

Bench: 5757843
2016-11-27 09:05:57 +01:00
Joost VandeVondele 8ceb1ff53b Fix undefined behavior
This fixes #892. Undefined behavior as seen with
clang -fsanitize=undefined.

No functional change.
2016-11-26 16:49:59 +01:00
Marco Costalba 2ec626ddae Fix compile under Windows XP
The needed Windows API for processor groups could be missed from old Windows
versions, so instead of calling them directly (forcing the linker to resolve
the calls at compile time), try to load them at runtime. To do this we need
first to define the corresponding function pointers.

Also don't interfere with running fishtest on numa hardware with Windows.
Avoid all stockfish one-threaded processes will run on the same node

No functional change.
2016-11-26 07:04:17 +01:00
Aram Tumanian 9eccba7761 Fix the pawn hash failure when the pawn key is 0
This patch fixed bugs #859 and #882.
At initialization we generate a new random key (Zobrist::noPawns).
It's added to the pawn key of all positions, so that the pawn key
of a pawnless position is no longer 0.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 21307 W: 3738 L: 3618 D: 13951

LTC:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 45270 W: 5737 L: 5648 D: 33885

No functional change.
2016-11-25 08:48:35 +01:00
erbsenzaehler ca464fc89e Cleanup Makfile for MacOs
1) Explicitly setting the default lib to the system-default is not
   needed on a Mac. See:
   http://libcxx.llvm.org/docs/UsingLibcxx.html

2) We do no longer need to exclude bmi2-builds from LTO. See:
   https://llvm.org/bugs/show_bug.cgi?id=19416

Changes tested and discussed on FishCooking:
   https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/acUQtKtEzMM

No functional change.
2016-11-25 08:46:20 +01:00
Marco Costalba 0d9a9f5e98 Handle Windows Processors Groups
Under Windows it is not possible for a process to run on more than one
logical processor group. This usually means to be limited to use max 64
cores. To overcome this, some special platform specific API should be
called to set group affinity for each thread. Original code from Texel by
Peter sterlund.

Tested by Jean-Paul Vael on a Xeon E7-8890 v4 with 88 threads and confimed
speed up between 44 and 88 threads is about 30%, as expected.

No functional change.
2016-11-22 07:56:04 +01:00
Joost VandeVondele 6036303bb6 Avoid touching source files in profile-build
This refines the profile-build target to avoid 'touch'ing the sources,
keeping meaningful modification dates and avoiding editor warnings like vi's:

WARNING: The file has been changed since reading it!!!
Do you really want to write to it (y/n)?

Instead of touching sources, the (instrumented) object files are removed,
which has the same effect of rebuilding them in the next step.

As a side effect, this simplifies the Makefile a bit.

No functional change.
2016-11-20 10:51:42 +01:00
Fabian Beuke b5d10d17c2 Reduce variable scope in swap_byte
Added a specialization to remove the 'if' condition

No functional change.
2016-11-19 21:31:24 +01:00
Aram Tumanian 797602938d Start searching for a repetition from the 4th ply behind
A position can never repeat the one on the previous move.
Thus we can start searching for a repetition from the 4th
ply behind. In the case:

 std::min(st->rule50, st->pliesFromNull) < 4

We don't need to do any more calculations. This case happens
very often - in more than a half of all calls of the function.

No functional change.
2016-11-19 10:20:28 +01:00
Alain SAVARD 76d113f5f0 Pawn shelter and pawn storm tuned
Based on SPSA tuned values

Passed STC
http://tests.stockfishchess.org/tests/view/582363b30ebc5910626b9ca8
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 40628 W: 7380 L: 7087 D: 26161

and passed LTC
http://tests.stockfishchess.org/tests/view/5823b73b0ebc5910626b9cb5
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 273312 W: 35991 L: 35131 D: 202190

bench: 5773672
2016-11-19 09:42:11 +01:00
Marco Costalba 18df1698f4 Fix compile error from previous patch
Due to different types on some platforms.

No functional change.
2016-11-19 09:35:57 +01:00
Stéphane Nicolet 7f4de0196b Do not use GCC extension for anonymous unions
Anonymous struct inside anonymous unions are a GCC extension.
This patch uses named structs to stick to the C+11 standard.

Avoids a string of warnings on the Clang compiler.

Non functional change (same bench and same MD5 signature,
so compiled code is exactly the same as in current master)
2016-11-19 09:22:49 +01:00
Joost VandeVondele cddc8d4546 More accurate 'go nodes' searches at low count
Makes the actual number of nodes searched match closely
the number of nodes requested, by increasing the frequency
of checking the number of nodes searched at low node count.
All other searches retain the default checking frequency of
once per 4096 nodes, and are thus unaffected.

Passed STC as non-regression
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26643 W: 4766 L: 4655 D: 17222

No functional change.
2016-11-19 08:47:41 +01:00
Stefan Geschwentner a90fc4c877 Non-quiet pruning tweak
Count in the difference of static evaluation
and alpha for pruning threshold.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 16885 W: 3061 L: 2866 D: 10958

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 33566 W: 4428 L: 4202 D: 24936

Bench: 5513149
2016-11-19 08:37:52 +01:00
Aram Tumanian e6c2899020 Make a version of Position::do_move() without the givesCheck parameter
In 10 of 12 calls total to Position::do_move()the givesCheck argument is
simply gives_check(m). So it's reasonable to make an overload without
this parameter, which wraps the existing version.

No functional change.
2016-11-12 09:55:12 +01:00
joergoster de269ee18e FEN parsing: add a second check for correctly setting e.p. square
Currently, we only check if there is a pawn in place
to make the en-passant capture. Now also check that
there is a pawn that could just have advanced two
squares. Also update the corresponding comment.

This makes the parsing of FENs a bit more robust, and
now correctly handles positions like the one reported by Dann Corbit.

position fen rnbqkb1r/ppp3pp/3p1n2/3P4/8/2P5/PP3PPP/RNBQKB1R w KQkq e6
d

 +---+---+---+---+---+---+---+---+
 | r | n | b | q | k | b |   | r |
 +---+---+---+---+---+---+---+---+
 | p | p | p |   |   |   | p | p |
 +---+---+---+---+---+---+---+---+
 |   |   |   | p |   | n |   |   |
 +---+---+---+---+---+---+---+---+
 |   |   |   | P |   |   |   |   |
 +---+---+---+---+---+---+---+---+
 |   |   |   |   |   |   |   |   |
 +---+---+---+---+---+---+---+---+
 |   |   | P |   |   |   |   |   |
 +---+---+---+---+---+---+---+---+
 | P | P |   |   |   | P | P | P |
 +---+---+---+---+---+---+---+---+
 | R | N | B | Q | K | B |   | R |
 +---+---+---+---+---+---+---+---+

Fen: rnbqkb1r/ppp3pp/3p1n2/3P4/8/2P5/PP3PPP/RNBQKB1R w KQkq - 0 1

No functional change.
2016-11-10 11:45:51 +01:00
Mira dc4655e1f9 Stack offset changed from -5 to -4
Non functional change, tests under sanitizers OK.

Rationales for change

- Offset in code is in range -4 ... 2
- There was an error by (pathological) corner case MAX_PLY=0

No functional change.
2016-11-10 11:43:21 +01:00
atumanian 0fa80c9ba3 Update comments related after new see_ge()
Update comments according to changes from my patch: #822

No functional change.
2016-11-10 11:40:31 +01:00
Joost VandeVondele 61c727fdcb Allow benches with more than 2G nodes.
./stockfish bench 128 1 4000000000 default nodes

    crashes before, works after.

    No functional change.
2016-11-07 13:35:28 +01:00
Joost VandeVondele 876f07cbee Fix undefined behaviour with unaligned loads in syzygy code
Casting a pointer to a different type with stricter alignment
requirements yields to implementation dependent behaviour.
Practicaly everything is fine for common platforms because the
CPU/OS/compiler will generate correct code, but anyhow it is
better to be safe than sorry.

Testing with dbg_hit_on() shows that the unalignment accesses are
very rare (below 0.1%) so it makes sense to split the code in a
fast path for the common case and a slower path as a fallback.

No functional change (verified with TB enabled).
2016-11-06 11:48:07 +01:00
ppigazzini f5d3f0ded6 Update AUTHORS for SF8
And format top contriutor (space instead of tabs, proper Unix
line endings).
2016-11-06 10:28:17 +01:00
Marco Costalba e4659693de Fix a warning with debug=no
Warning in TB code due to unused variable.

Verified same bench with TB code enabled.

No functional change.
2016-11-06 09:45:25 +01:00
Joost VandeVondele 52cb348023 Reproducible searches after ucinewgame
Fixes issue #859.

thisThread->callsCnt in search<>() was different (by 1) for the first and second game played.

No functional change.
2016-11-05 22:25:55 +01:00
Marco Costalba 0d669be76c Restore development version
No functional change.
2016-11-05 09:32:39 +01:00
Marco Costalba 23f384cac3 Add explicit braces and fix a warning
Warning under both gcc and clang.

No functional change.
2016-11-05 09:29:22 +01:00
VoyagerOne b915fdc889 Reduction Simplification
Simplify reduction formula by removing a parameter.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 35798 W: 6368 L: 6272 D: 23158

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 125375 W: 15827 L: 15839 D: 93709

Bench: 4735038
2016-11-05 08:17:42 +01:00
Joost VandeVondele 67d19447f4 Makefile fix for sanitize
Small fixes for compilation with sanitize=yes optimize=no,
by always adding -fsanitize=undefined to the LDFLAGS as required.
Updates config-sanity to check&report the status of the flag.

No functional change.
2016-11-05 08:15:56 +01:00
VoyagerOne d4abf54247 Top CPU Contributors
Give some well needed credit...

No functional change.
2016-11-05 08:12:22 +01:00
Miroslav Fontán 445aade39c Simplify code, delete else after return 2016-11-05 08:08:25 +01:00
Miroslav Fontán f3cd7002aa Sync variable names in decl vs def 2016-11-05 08:05:22 +01:00
Joost VandeVondele c43146edb7 Travis-ci: integrate perft checking
makes verifying perft numbers for a few positions
part of travis-ci. Adds <5s testing time.

No functional change.
2016-11-05 08:03:34 +01:00
Marco Costalba c0bb041539 Rewrite syzygy in C++
Rewrite the code in SF style, simplify and
document it.

Code is now much clear and bug free (no mem-leaks and
other small issues) and is also smaller (more than
600 lines of code removed).

All the code has been rewritten but root_probe() and
root_probe_wdl() that are completely misplaced and should
be retired altogheter. For now just leave them in the
original version.

Code is fully and deeply tested for equivalency both in
functionality and in speed with hundreds of games and
test positions and is guaranteed to be 100% equivalent
to the original.

Tested with tb_dbg branch for functional equivalency on
more than 12M positions.

stockfish.exe bench 128 1 16 syzygy.epd

Position: 2016/2016
Total 12121156 Hits 0 hit rate (%) 0
Total time (ms) : 4417851
Nodes searched : 1100151204
Nodes/second : 249024

Tested with 5,000 games match against master, 1 Thread,
128 MB Hash each, tc 40+0.4, which is almost equivalent
to LTC in Fishtest on this machine. 3-, 4- and 5-men syzygy
bases on SSD, 12-moves opening book to emphasize mid- and endgame.

Score of SF-SyzygyC++ vs SF-Master: 633 - 617 - 3750  [0.502] 5000
ELO difference: 1

No functional change.
2016-11-05 07:55:08 +01:00
Marco Costalba 369eff437c Stockfish 8
Bench: 5926706

No functional change
2016-11-01 10:19:17 +01:00
Stefan Geschwentner 344616e917 Bonus for attacked passed pawn promotion path
Bonus for each attacked square on the promotion path
of an enemy passed pawn.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 141511 W: 25295 L: 24579 D: 91637

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 23161 W: 3022 L: 2831 D: 17308

Bench: 5926706
2016-10-30 13:48:03 +01:00
Joost Vandevondele 40b1b27178 Fix a series of undefined behaviours
Avoid shifting negative signed integers and use typed
enum to avoids decrementing a variable beyond its defined
range, like:
       for (Rank r = RANK_8; r >= RANK_1; --r)

Changes were tested individually and passed SPRT[-3, 1].

With this patch gcc --sanitize builds cleanly.

No functional change.
2016-10-27 06:44:41 +02:00
Joost VandeVondele bf51b4796a travis-ci: Enable undefined behavior checking 2016-10-27 06:29:24 +02:00
syzygy e18e557e77 Output PV if last iteration does not complete
Instead of outputting "info nodes ... time ..." when the last
iteration is interrupted, simply call UCI::pv() to output the PV.

I thought about calling UCI:pv() with bounds -VALUE_INFINITE, VALUE_INFINITE
to avoid "lowerbound" or "upperbound" appearing in it, but I'm not sure that
would be any better.

This patch fixes rare inconsistencies between the first move of
the last PV output and the bestmove played. It also makes sure
that all the latest statistics are sent to the GUI (not only nodes
and time but also nps, tbhits, hashfull).

No functional change.
2016-10-27 06:26:22 +02:00
Stéphane Nicolet 818b4a126d Endgame malus for having a king in a pawnless flank
Original idea by "ElbertoOne", while "FauziAkram" suggested to put a
small midgame penalty too.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 71808 W: 13038 L: 12610 D: 46160

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 150874 W: 19828 L: 19221 D: 111825

Bench: 6077005
2016-10-25 06:57:29 +02:00
VoyagerOne e77f38c431 History Stat Comparison
Adjust LMR by comparing history stats
with opponent (prior ply).

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 27754 W: 5066 L: 4824 D: 17864

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 216596 W: 28157 L: 27343 D: 161096

Bench: 5437729
2016-10-25 06:44:19 +02:00
Marco Costalba e18321f55a Correcty resey TB hit counter
Restore original behaviour to reset
the counter before a new move search.

Also fixed some warnings and added const
qualifier to a couple of functions, as
suggested by m_stembera.

Thanks to Werner Bergmans for reporting
the regression.

No functional change.
2016-10-22 08:22:13 +02:00
syzygy ca67752645 Per-thread TB hit counters
Use a per-thread counter to reduce contention
with many cores and endgame positions.

Measured around 1% speed-up on a 12 core and 8%
on 28 cores with 6-men, searching on:
 7R/1p3k2/2p2P2/3nR1P1/8/3b1P2/7K/r7 b - - 3 38

Also retire the unused set_nodes_searched() and fix
a couple of return types and naming conventions.

No functional change.
2016-10-21 06:15:45 +02:00
Joost Vandevondele 3686e719a1 Simplify next_move by always scoring evasions
For a default bench, this fixes the last valgrind
error (jump on uninitialised value).

Passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 187869 W: 33303 L: 33463 D: 121103

No functional change.
2016-10-20 17:17:14 +02:00
Joost Vandevondele 9893e7fd53 Make valgrind testing part of travis ci. 2016-10-19 21:29:00 +02:00
ajithcj 99f3ad6858 Remove useless assignments to currentMove
We reference (ss-1)->currentMove, i.e. we peek
current move of the parent node, so currentMove
should be valid in the main move loop, when we
search() the subtree, but outside of main loop
it is useless.

No functional change.
2016-10-18 09:00:52 +02:00
VoyagerOne e27d3bb884 Use explicit logic for pruning
Also a speedup since we don't need to recalculate SEE
for extensions...as it already determined to be positive.

Results for 12 tests for each version:

        Base      Test      Diff
Mean    2132395   2191002   -58607
StDev   128058    85917     134239
p-value: 0.669
speedup: 0.027

Non functional change.
2016-10-18 08:53:51 +02:00
Jacques 16e1881126 Fixes for ARM compilation: take 2
The target:

Odroid U3 (http://www.hardkernel.com/main/products/prdt_info.php?g_code=g138745696275)
Debian Jessie
As listed in #550 and #638 three modifications are needed for compilation to work:

float-abi flag for GCC If an FPU is present and supported by the installed os then passed value need to be hard.
I didn't find any better solution than using readelf to check for the availibilty of Tag_ABI_VFP_args which sould indicate support for the FPU. The check is only done if the arch is arm and if readelf is not present
on the system, there will be an error (/bin/sh: 1: readelf: not found) but it will not break and will continue with the default softfp value. Outputing the error is not really acceptable but I wanted some feedback on the
check itself.

-lpthread is needed on armv7 outside of Android
I replaced UNAME with KERNEL and OS to allow to differentiate Android.

m32 flag
My understanding is that outside of Android the flag is generating errors on armv7.

These modifications should introduce change only for non Android armv7 build.

No functional change.
2016-10-14 08:58:07 +02:00
Marco Costalba e1f600f186 Revert "Fixes for ARM compilation"
This reverts commit a3fe80c36a.

Break compilation on mingw for me.
2016-10-13 08:36:30 +02:00
Jacques a3fe80c36a Fixes for ARM compilation
The target:

Odroid U3 (http://www.hardkernel.com/main/products/prdt_info.php?g_code=g138745696275)
Debian Jessie
As listed in #550 and #638 three modifications are needed for compilation to work:

float-abi flag for GCC If an FPU is present and supported by the installed os then passed value need to be hard.
I didn't find any better solution than using readelf to check for the availibilty of Tag_ABI_VFP_args which sould indicate support for the FPU. The check is only done if the arch is arm and if readelf is not present
on the system, there will be an error (/bin/sh: 1: readelf: not found) but it will not break and will continue with the default softfp value. Outputing the error is not really acceptable but I wanted some feedback on the
check itself.

-lpthread is needed on armv7 outside of Android
I replaced UNAME with KERNEL and OS to allow to differentiate Android.

m32 flag
My understanding is that outside of Android the flag is generating errors on armv7.

These modifications should introduce change only for non Android armv7 build.

No functional change.
2016-10-13 08:34:04 +02:00
Marco Costalba fdf3a51c68 AppVeyor: run bench after build
And show resulting bench signature.

The run is very slow becuase optimizations
are all disabled by default /Od /RTC1

No functional change.
2016-10-10 21:00:59 +02:00
Marco Costalba e61f7b1e6d Add AppVeyor integration
It is like Trevis CI but for Windows platform.

Currently just compile builds, wthouth benching
the resulting executable.

No functional change.
2016-10-10 16:29:29 +02:00
ajithcj f799610d4b Simplify futility pruning return value
Return eval as it is while doing futility pruning.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 167687 W: 29778 L: 29904 D: 108005

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 26905 W: 3503 L: 3390 D: 20012

Bench: 5936728
2016-10-09 09:54:43 +02:00
atumanian 073eed590e Optimisation of Position::see and Position::see_sign
Stephane's patch removes the only usage of Position::see, where the
returned value isn't immediately compared with a value. So I replaced
this function by its optimised and more specific version see_ge. This
function also supersedes the function Position::see_sign.

bool Position::see_ge(Move m, Value v) const;

This function tests if the SEE of a move is greater or equal than a
given value. We use forward iteration on captures instread of backward
one, therefore we don't need the swapList array. Also we stop as soon
as we have enough information to obtain the result, avoiding unnecessary
calls to the min_attacker function.

Speed tests (Windows 7), 20 runs for each engine:
Test engine: mean 866648, st. dev. 5964
Base engine: mean 846751, st. dev. 22846
Speedup: 1.023

Speed test by Stephane Nicolet

Fishtest STC test:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26040 W: 4675 L: 4442 D: 16923
http://tests.stockfishchess.org/tests/view/57f648990ebc59038170fa03

No functional change.
2016-10-08 06:38:36 +02:00
Stéphane Nicolet 1e586288ca Do not use SEE in evasion scoring
Idea by Aram Tumanian (atumanian)

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43889 W: 7849 L: 7767 D: 28273

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 29333 W: 3809 L: 3700 D: 21824

Bench: 6421663
2016-10-06 00:00:27 +02:00
Stefano Cardanobile 0162fb83c2 Retire implicit malus for stonewalls
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 75864 W: 13466 L: 13437 D: 48961

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 99050 W: 12472 L: 12451 D: 74127

bench: 6098474
2016-10-05 09:32:08 +02:00
VoyagerOne ab26c61971 Allow inCheck pruning
This is a bit tricky because we don't want
to prune the only legal evasions, even if
with negative SEE. So add an assert to avoid
this subtle bug to slip in later.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 14140 W: 2625 L: 2421 D: 9094

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 11558 W: 1555 L: 1379 D: 8624

bench: 5256717
2016-10-03 16:18:53 +02:00
Marco Costalba eccccba0ce Remove useless razoring condition
Condition is always true! For any value of the
array index! Even an out of bound array, like
razor_margin[120]!!!!

No functional change.
2016-09-29 15:24:36 +02:00
HiraokaTakuya b77bae0529 Make razor_margin[4] ONE_PLY value independent
No functional change.
2016-09-29 15:20:07 +02:00
Stéphane Nicolet 7ae3c05795 Rename shift_bb() to shift()
Rename shift_bb() to shift(), and DELTA_S to SOUTH, etc.
to improve code readability, especially in evaluate.cpp
when they are used together:

    old b = shift_bb<DELTA_S>(pos.pieces(PAWN))
    new b = shift<SOUTH>(pos.pieces(PAWN))

While there fix some small code style issues.

No functional change.
2016-09-25 10:45:10 +02:00
joergoster 351844061e Allowing singular extension in mate positions
Drop useless condition

abs(ttValue) < VALUE_KNOWN_WIN

And extend singular extension search to cases when ttValue
stores a mate score. This improves mate finding and does
not introduce any regression.

Yery tested this patch against current master on the 6500+
Chest mate suite with 200K fixed nodes:

    shortest mates found: master: 1206 patch:1205
    any mate found: master: 1903 patch: 2003

with 1 sec time:

    shortest mates found: master: 2667 patch: 2628
    any mate found: master: 3585 patch: 3646

Verified for no regression:

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25655 W: 4578 L: 4465 D: 16612

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 66247 W: 8618 L: 8557 D: 49072

bench: 6335042
2016-09-24 19:56:02 +02:00
Marco Costalba 8662bdfa12 Fix crash when passing a mate/stalemate position
Both Tablebases::filter_root_moves() and
extract_ponder_from_tt(9 were unable to handle
a mate/stalemate position.

Spotted and reported by Dann Corbit.

Added some mate/stalemate positions to bench so
to early catch this regression in the future.

No functional change.
2016-09-24 07:37:52 +02:00
Stéphane Nicolet 28240d375c Simplify pinners conditions in SEE()
Use the following transformations:

- to check that A is included in B, testing "(A & ~B) == 0" is faster
than "(A & B) == A"

- to remove the intersection of A and B from A, doing "A &= ~B;" is as
fast as "if (A & B) A &= ~B;" but is simpler.

Overall, the simpler patch version is 0.3% than current master.

No functional change.
2016-09-22 08:31:23 +02:00
Guenther Demetz 943ae89be1 Fix pin-aware SEE
Correct pinners calculation and fix bug with pinned
pieces giving check. With this patch 'pinners' only
returns sliders with exactly one defensive piece between
the slider and the attacked square (in other words, pinners
returns exact pinners).

This was a co-operation between Marco Costalba,
Stphane Nicolet and me.

Special thanks to Ronald de Man for reporting the bug with
pinned pieces giving check, discussed here:
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/S_4E_Xs5HaE

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 132118 W: 23578 L: 23645 D: 84895

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36424 W: 4770 L: 4670 D: 26984

bench: 6272231
2016-09-21 08:42:25 +02:00
Joost Vandevondele 4b0043ae7c Use fixed depth bench to make PGO builds more reproducible
Discussed on fishcooking

proposal and objdump verification:
https://groups.google.com/d/msg/fishcooking/4_ausUwMXP0/EGPsMYqOFAAJ

verified no significant speed difference between depth and time:
https://groups.google.com/d/msg/fishcooking/4_ausUwMXP0/KazW5QZmFgAJ

stockfish_time - stats:
mean = 2207232.56        std = 7079.51        std/mean = 0.003207

stockfish_depth - stats:
mean = 2201783.57        std = 6356.69        std/mean = 0.002887

No functional change
2016-09-18 08:13:34 +02:00
Marco Costalba 92f01aa2bd Fix a warning with MSVC
warning C4706: assignment within conditional expression

No functional change.
2016-09-17 10:14:28 +02:00
Stéphane Nicolet ea41f18e6e Swap mg and eg in internal representation of Score
Instrumentation shows that in make_score(mg, eg) calls, the mg value is
zero in 25,9% of the calls while the eg value is zero in 36,8% of the
calls.

Swapping the internal fields of mg and eg in the internal
representation of Score allows the compiler to optimize away the shift
in (eg << 16) + mg in more cases, thus resulting in a 0.3% speed-up
overall.

No functional change
2016-09-17 09:56:36 +02:00
Marco Costalba 057d710fc2 Fix indentation in struct FromToStats
And other little trivial stuff.

No functional change.
2016-09-17 09:51:20 +02:00
Stéphane Nicolet 01f2466f6e Retire KingDanger array
Rescales the king danger variables in evaluate_king() to
suppress the KingDanger[] array. This avoids the cost of
the memory accesses to the array and simplifies the non-linear
transformation used.

Full credits to "hxim" for the seminal idea and implementation,
see pull request #786.
https://github.com/official-stockfish/Stockfish/pull/786

Passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 9649 W: 1829 L: 1689 D: 6131

Passed LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 53494 W: 7254 L: 7178 D: 39062

Bench: 6116200
2016-09-16 08:30:06 +02:00
Marco Costalba 5c58d1f5cb Use per-thread counterMoveHistory
Drops a scalability bottleneck due to memory contention
of a single shared table across threads. The effect starts
to be sensible with a high number of threads. Specifically
we have a small regression with 7 threads both at 60 and
180 seconds TC:

10000 @ 60+0.6 th 7
ELO: -2.46 +-3.2 (95%) LOS: 6.5%
Total: 9896 W: 1037 L: 1107 D: 7752

5000 @ 180+0.6 th 7
ELO: -1.95 +-4.1 (95%) LOS: 17.7%
Total: 5000 W: 444 L: 472 D: 4084

We have a regression because counterMoveHistory table is
quite big and it takes time for a single thread to fill it.
Sharing the table yields to a higher fill rate and better
quality of moves and up to 7 threads the benefits of sharing
more then compensate the loss in speed due to contention.
Interestingly even with a 3X longer TC, so with more time
for the single thread to catch up, the improvment is quite
limited and below noise level. It seems we really need much
longer TC to saturate the table.

When we move to high threads number it's another story:

5000 @ 60+0.6 th 22
ELO: 3.49 +-4.3 (95%) LOS: 94.6%
Total: 4880 W: 490 L: 441 D: 3949

2000 @ 60+0.6 th 32
ELO: 8.34 +-6.9 (95%) LOS: 99.1%
Total: 2000 W: 229 L: 181 D: 1590

As expected the speed-up more than compensates the filling
rate, and we expect that with tournament TC, where single
thread is able to saturate the table, the difference will
be even stronger. For instance for TCEC 9 super-final time
control will be 180 minutes + 15 seconds and this scalability
improvement seems definitely the way to go.

So, summarizing:

GOOD:

Measured big improvement in high core scenario

Suitable for TCEC 9 superfinal (big hardware, very long TC)

Consistent and natural patch that extends to counterMoveHistory
what we already do for remaining history tables, that are all per-thread

Non functional change for the common case of a single core

Very simple (just 6 lines modified, no added ones)

BAD:

Small regression (within 2-3 ELO) with few threads and short TC

bench: 5341477
2016-09-16 08:15:07 +02:00
Marco Costalba b96dd754ed Renaming in MovePicker
Rename stages and simplify a bit the code.

No functional change.
2016-09-15 09:07:49 +02:00
Marco Costalba 01ee509a5c Retire MovePicker::see_sign()
No more used after last patch.

No functional change.
2016-09-14 15:43:56 +02:00
VoyagerOne 95ad2b51b7 Tweak SEE margin in pruning conditions
Use 35 * depth^2 to calculate see_margin.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 22636 W: 4212 L: 3990 D: 14434

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 47241 W: 6314 L: 6041 D: 34886

The Movepick SEE is now dead code, retire it.

Bench: 5341477
2016-09-14 15:38:38 +02:00
syzygy 438805aee8 Integrate next_stage() logic into next_move()
Measured bench speed up goes from 0,7% to 2%,
given the unreliable measure a reverse simmplification
test was done on fishtest:

master vs patch
LLR: -2.94 (-2.94,2.94) [-3.00,1.00]
Total: 15499 W: 2685 L: 2867 D: 9947

Test result is positive, master is weaker.

No functional change.
2016-09-13 07:14:09 +02:00
Guenther Demetz ace8e951d7 Simplify code for pinaware SEE
This is the most compact and neatest version
is was able to produce.

On normal builds I have a small slowdown:
normal builds base vs. simplification (gcc 4.8.1 Win7-64 i7-3770 @ 3.4GHz x86-64-modern)
Results for 20 tests for each version:

        Base      Test      Diff
Mean    1974744   1969333   5411
StDev   11825     10281     5874
p-value: 0,178
speedup: -0,003

On pgo-builds however I measure a nice 1.1% speedup

pgo-builds base vs. simplification
Results for 20 tests for each version:

        Base      Test      Diff
Mean    1974119   1995444   -21325
StDev   8703      5717      4623
p-value: 1
speedup: 0,011

No functional change.
2016-09-12 15:45:00 +02:00
Guenther Demetz 90ce24b11e Pinned aware SEE
Don't allow pinned pieces to attack the exchange-square as long all
pinners (this includes also potential ones) are on their original
square.
As soon a pinner moves to the exchange-square or get captured on it, we
fall back to standard SEE behaviour.

This correctly handles the majority of cases with absolute pins.

bench: 6883133
2016-09-12 09:31:09 +02:00
Stefano Cardanobile 4c95edddbf Reorder evaluation start
In evaluate, we start by initializing the pos.psq_score
and adding the material imbalance. After that, we check
whether a specialized eval exists and if yes we return
that value and discard whatever we have computed until now.

It sounds more logical to first probe material entry and
return if we have a specialized eval, and only if it is
not the case initialize eval with some values. There is
no measurable speed-difference on my computer.

Non functional change.
2016-09-11 07:42:12 +02:00
Marco Costalba 602d7fbb07 Use Movepick SEE value in search
This halves the calls to the costly pos.see_sign(),
speed up is about 1-1.3%

Non functional change.
2016-09-09 17:11:54 +02:00
Marco Costalba d909d10f33 Refactor previous patch
No functional change.
2016-09-08 06:02:42 +02:00
ajithcj 38428ada54 Prune dangerous moves at low depth
At very low depths prune captures,
promotions and checks if see is negative.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 6772 W: 1328 L: 1173 D: 4271

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 8917 W: 1270 L: 1122 D: 6525

bench: 6024713
2016-09-08 05:55:10 +02:00
Marco Costalba e340ce221c Syntactic sugar to loop across pieces
Also add some comments to the new operator~(Piece).

No functional change.
2016-09-04 15:33:17 +02:00
syzygy ca6c9f85a5 Change from [Color][PieceType] to [Piece]
Speed up of almost 1% in both normal and
pgo builds.

No functional change.
2016-09-04 09:22:09 +02:00
Marco Costalba c5828c4eba Fix syzygy with partial TB
In case we have installed a not complete set of 6-men tables and
there is 6 piece position on board, but no corresponding
tablebase engine is not using any syzygy at all.

Reported by Jouni Uski, fix by Peter Österlund,
confirmed as a bug by Ronald de Man.

bench: 7591630
2016-09-03 08:21:05 +02:00
Stéphane Nicolet d37dfe9ae4 Space bonus in presence of open files
If the opponent has a cramped position, opening a file often
helps him/her to exchange pieces, so it makes sense to reduce
the space bonus if there are open files.

Credits: Leonardo Ljubičić for the strategic idea, Alain Savard for the
implementation of the open files calculation, "CrunchyNYC" for the
compensation of the numerator.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 49112 W: 9239 L: 8900 D: 30973

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 89415 W: 12014 L: 11601 D: 65800

Bench: 7591630
2016-09-03 00:04:20 +02:00
lucasart 13b4444d9e Change exclusion key setup
Should depend on which move is excluded. This
allow us to remove the dedicated Position::exclusion_key().

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 59814 W: 11136 L: 11083 D: 37595

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 31023 W: 4187 L: 4080 D: 22756

bench 7553379
2016-09-02 08:37:01 +02:00
Stefano80 7f2eb10e93 Retire linear imbalance
Retire linear imbalance and compensate
in piece values enumeration.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 43596 W: 8105 L: 8023 D: 27468

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 24482 W: 3352 L: 3237 D: 17893

Bench: 7777707
2016-09-02 08:25:17 +02:00
ajithcj 5cffc032da Optimize order of a few conditions in search
Also fix size of KingDanger array to reduce memory footprint.

Small speed up of around 0.5%

No functional change.
2016-08-31 13:47:45 +02:00
VoyagerOne 2731bbaf6b Remove condition on killers in history pruning
Now allows main killer to be history prune.

STC:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 15852 W: 2910 L: 2781 D: 10161

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 56428 W: 7610 L: 7537 D: 41281

Bench: 8032058
2016-08-30 09:09:55 +02:00
Stefan Geschwentner 6aa9308f08 Tweak probcut threshold
Use better threshold for capture move generation.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23265 W: 4415 L: 4188 D: 14662

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 36618 W: 5083 L: 4836 D: 26699

bench: 7030088
2016-08-29 22:15:03 +02:00
Andrew Grant b4f6728e61 Removed an extra space
No functional change.
2016-08-28 09:47:30 +02:00
Alain SAVARD 2b57b61cb1 Move king tropism to evaluate_king
No functional change.
2016-08-28 08:49:40 +02:00
Marco Costalba 1ee2838214 Retire CheckInfo
Move its content directly under StateInfo.

Verified for no speed regression.

No functional change.
2016-08-28 08:08:13 +02:00
Marco Costalba 0b944c7186 Silence some warnings with MSVC 2013
No functional change.
2016-08-27 12:16:13 +02:00
Stéphane Nicolet 805afcbf3d Move CheckInfo under StateInfo
This greately simplifies usage because hides to the
search the implementation specific CheckInfo.

This is based on the work done by Marco in pull request #716,
implementing on top of it the ideas in the discussion: caching
the calls to slider_blockers() in the CheckInfo structure,
and simplifying the slider_blockers() function by removing its
first parameter.

Compared to master, bench is identical but the number of calls
to slider_blockers() during bench goes down from 22461515 to 18853422,
hopefully being a little bit faster overall.

archlinux, gcc-6
make profile-build ARCH=x86-64-bmi2
50 runs each

bench:
base = 2356320 +/- 981
test = 2403811 +/- 981
diff = 47490 +/- 1828

speedup = 0.0202
P(speedup > 0) = 1.0000

perft 6:
base = 175498484 +/- 429925
test = 183997959 +/- 429925
diff = 8499474 +/- 469401

speedup = 0.0484
P(speedup > 0) = 1.0000

perft 7 (but only 10 runs):
base = 185403228 +/- 468705
test = 188777591 +/- 468705
diff = 3374363 +/- 476687

speedup = 0.0182
P(speedup > 0) = 1.0000

$ ./pyshbench ../Stockfish/master ../Stockfish/test 20
run base     test     diff
...

base = 2501728 +/- 182034
test = 2532997 +/- 182034
diff = 31268 +/- 5116

speedup = 0.0125
P(speedup > 0) = 1.0000

No functional change.
2016-08-27 09:53:26 +02:00
Marco Costalba 4c5cbb1b14 Make engine ONE_PLY value independent
This non-functional change patch is a deep work to allow SF to be independent
from the actual value of ONE_PLY (currently set to 1). I have verified SF is
now independent for ONE_PLY values 1, 2, 4, 8, 16, 32 and 256.

This patch gives consistency to search code and enables future work, opening
the door to safely tweaking the ONE_PLY value for any reason.

Verified for no speed regression at STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 95643 W: 17728 L: 17737 D: 60178

No functional change.
2016-08-27 09:12:25 +02:00
gamander 133808851d Fixed wrong definition of WhiteCamp and BlackCamp
No functional change.
2016-08-27 08:48:07 +02:00
hxim 9bd856e28e Simplify stats update
Simplify code by moving countermove and follow-up move
history update into procedure.

No functional change.
2016-08-27 08:39:32 +02:00
Marco Costalba 3e739f883f Reformat stats update
Rewritten in a way to have explicit in the search
the bonus/penalty we apply: hopefully this will lead
to further simplification/fix of current rather messy
stats update code.

No functional change.
2016-08-25 09:51:50 +02:00
VoyagerOne 0b4f6e562b Refutation penalty on captures
Apply refutation penalty for prior PV quiet move on captures

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 8208 W: 1153 L: 1008 D: 6047
http://tests.stockfishchess.org/tests/view/57bc5a9f0ebc59030fbe47b5

Only LTC because a very similar patch already passed STC + LTC

bench: 7038730
2016-08-24 08:26:32 +02:00
VoyagerOne 5596492f6e Simplify IID
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 30468 W: 5687 L: 5582 D: 19199
http://tests.stockfishchess.org/tests/view/57b1ddd80ebc591c761f63e2

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 87406 W: 11756 L: 11725 D: 63925
http://tests.stockfishchess.org/tests/view/57b212590ebc591c761f63f9

bench: 6554900
2016-08-19 08:59:18 +02:00
VoyagerOne 8493c8c6b8 Do LMR on captures
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 5361 W: 1086 L: 936 D: 3339
http://tests.stockfishchess.org/tests/view/57b31b0f0ebc591c761f643d

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 54694 W: 7591 L: 7287 D: 39816
http://tests.stockfishchess.org/tests/view/57b3442b0ebc591c761f6450

bench: 6881120
2016-08-19 08:50:14 +02:00
Marco Costalba e3af492142 Remove a stale assignment
No more used after previous patch.

Spotted by Jekaa .

No functional change.
2016-08-18 10:40:56 +02:00
Stefano80 9585f8ef58 Retire pawn span
Retire pawn span and replace with pawn count in evaluate_scale_factor.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26482 W: 4929 L: 4818 D: 16735

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61938 W: 8400 L: 8335 D: 45203

Bench: 7662861
2016-08-18 10:16:56 +02:00
VoyagerOne 7396c08b79 Use predicted depth for history pruning
STC: (Yellow)
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 69115 W: 12880 L: 12797 D: 43438

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 124163 W: 16923 L: 16442 D: 90798

Note: Note based off past experiments / patches... history pruning
is quite TC sensitive. I believe the reason for this TC dependency
is that the CMH/FMH is a very large table that takes time to fill
up with. In addition having more time for will increase the accuracy
of the stats' value.

Bench: 7351698
2016-08-15 14:45:40 +02:00
Luca Brivio 3d10cfcdd2 Cap space evaluation bonus
When computing space evaluation, limit the bonus square count to 16.

STC @ 10+0.1 th 1:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 30793 W: 5910 L: 5648 D: 19235

LTC @ 60+0.6 th 1:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 31361 W: 4410 L: 4184 D: 22767

Bench: 7165385
2016-08-13 10:12:09 +02:00
Alain SAVARD 8abb98455f Simplify space formula
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 16868 W: 3260 L: 3132 D: 10476

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 16910 W: 2381 L: 2255 D: 12274

bench: 6663531
2016-08-10 12:49:24 +02:00
VoyagerOne b3525fa9ea Use Color-From-To history stats to help sort moves
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 33502 W: 6498 L: 6223 D: 20781
http://tests.stockfishchess.org/tests/view/578abb940ebc5972faa169e2

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 50782 W: 7124 L: 6832 D: 36826
http://tests.stockfishchess.org/tests/view/578b8e5d0ebc5972faa169fd

LTC: (Sanity test against latest master)
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 32759 W: 4600 L: 4370 D: 23789
http://tests.stockfishchess.org/tests/view/5798b7d30ebc591c761f5b72

bench: 6985912

P.S. Thanks @mstembera for rewriting my code to make it smp compatible. A BIG thank you!
2016-08-02 09:17:14 +02:00
VoyagerOne 85924db496 Futility tweak
Use a different margin for pruning child nodes.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16692 W: 3251 L: 3051 D: 10390
http://tests.stockfishchess.org/tests/view/579b95d10ebc591c761f5c03

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 24140 W: 3501 L: 3297 D: 17342
http://tests.stockfishchess.org/tests/view/579bb15d0ebc591c761f5c0b

Bench: 7927017
2016-08-01 21:47:42 +02:00
ajithcj f2f3a06a1a Allow null pruning at depth 1
This removes a check that prevents null pruning at depth 1 PLY.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 23445 W: 4638 L: 4521 D: 14286

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61416 W: 8627 L: 8563 D: 44226

bench: 8145304
2016-07-24 09:54:07 +02:00
Stefano80 714329dbdc See prune at higher depth
Allow SEE pruning at higher depths in shallow depth
pruning using a threshold increasing with depth.

STC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 35366 W: 7011 L: 6724 D: 21631

LTC
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 15578 W: 2243 L: 2070 D: 11265

Bench: 8417887
2016-07-24 08:55:58 +02:00
ajithcj 76971d8acb Gradually relax the NMP staticEval check
Gradually relax the NMP staticEval check as we go to
higher depths.

Use tuned values.

STC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16745 W: 3371 L: 3168 D: 10206

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 5906 W: 875 L: 736 D: 4295

bench: 8548212
2016-07-16 07:35:31 +02:00
Lyudmil Antonov f619f1d38c Workaround gcc stack alignment bug
GCC uses SSE instructions to move data but in 32-bit gcc version used
by abrok the stack is not 16-byte aligned due to a bug.

This patch workaround teh bug by not using the stack
to store KingFlank[]

Fixes issue #721.

No functional change.
2016-07-16 07:06:13 +02:00
joergoster 82d02a3133 Fix extract_ponder_from_tt()
Checking for legality of a possible ponder move
must be done before we undo the first pv move,
of course. (spotted by mohammed li.)

This obviously only has any effect when playing in ponder mode.

No functional change.
2016-07-12 08:47:21 +02:00
Stéphane Nicolet 1d09ee70f7 King tropism
Bonus for each square that we attack in the flank where the opponent
king is. Squares that we attack twice and are not protected by an enemy
pawn count double.

Passed STC:
http://tests.stockfishchess.org/tests/view/577dfca60ebc5972faa166b8
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 48373 W: 9832 L: 9481 D: 29060

And LTC:
http://tests.stockfishchess.org/tests/view/577e77870ebc5972faa166df
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 8881 W: 1408 L: 1255 D: 6218

Bench: 7577046
2016-07-08 09:35:22 +02:00
Alain SAVARD eb20a87c67 More safe checks
Consider a check given by a rook or a minor to be a "safe check"
also in the case where supported by another piece,
and given on a square only defended by a queen

Was yellow STC
http://tests.stockfishchess.org/tests/view/576fcbc80ebc5972faa163e8
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 55453 W: 10431 L: 10315 D: 34707

Passed LTC
http://tests.stockfishchess.org/tests/view/57733a0b0ebc5972faa164b7
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 54550 W: 7671 L: 7365 D: 39514

bench: 7398346
2016-07-07 08:39:58 +02:00
ajithcj ade3bb9a4e Use staticEval in null prune condition
Don't null prune at depth < 12 if staticEval < beta

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 62858 W: 12035 L: 11632 D: 39191

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 49784 W: 7009 L: 6720 D: 36055

bench: 8054611
2016-07-02 16:43:49 +02:00
loco-loco 716a145a6c Removing inCheck condition for counter move bonus
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 20206 W: 3946 L: 3823 D:

LTC:
LLR: 3.10 (-2.94,2.94) [-3.00,1.00]
Total: 25004 W: 3512 L: 3390 D: 18102

Bench: 8172428
2016-06-26 10:25:05 +02:00
Marco Costalba 190d2ea4bc Restore standard passed pawn definition
Use the usual and accepted passed pawn semantic
instead of a non-standard one and remove a FIXME.

STC (http://tests.stockfishchess.org/tests/view/576401350ebc5972faa1608d):
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 29646 W: 5663 L: 5557 D: 18426

LTC (http://tests.stockfishchess.org/tests/view/5764e4e90ebc5972faa160c3):
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 40224 W: 5578 L: 5484 D: 29162

bench: 7543902
2016-06-25 11:46:17 +02:00
ElbertoOne c2c0e6b07d Remove redundant PvNode condition
After commit 6d58bf777c we always call PvNodes
with cutNode set to false.

No functional change.
2016-06-24 08:46:36 +02:00
VoyagerOne c94145b65c Comment out a redundant condition
Take advantage that VALUE_NONE = 32002 to remove
the condition.

Commented out and not removed becuase it is tricky
to rely on the hidden value of VALUE_NONE and code
can break in case we change VALUE_NONE in the future.

No functional change.
2016-06-24 08:26:39 +02:00
Jonathan Calovski 09efbf915e Remove scalefactor dependency
STC
http://tests.stockfishchess.org/tests/view/5764539e0ebc5972faa160a4
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 43878 W: 8289 L: 8208 D: 27381

LTC
http://tests.stockfishchess.org/tests/view/5764f0130ebc5972faa160c9
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39338 W: 5408 L: 5313 D: 28617

bench: 7977279
2016-06-21 09:01:39 +02:00
ElbertoOne 6d58bf777c On IID do not always search with cutNode = true
On IID now search with cutNode value instead of fixed value true.

STC (http://tests.stockfishchess.org/tests/view/575fa3860ebc5972faa15f67):
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 163974 W: 30744 L: 30874 D: 102356

LTC (http://tests.stockfishchess.org/tests/view/5763b0640ebc5972faa16075):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 48363 W: 6611 L: 6528 D: 35224

Bench: 7806393
2016-06-19 18:04:22 +02:00
VoyagerOne 1c0c4db677 Simplify Check Extension
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32704 W: 6146 L: 6045 D: 20513

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 146622 W: 19967 L: 20017 D: 106638

Bench: 8245662
2016-06-18 08:32:25 +02:00
ajithcj 8e45e70e55 Don't insert pv back into tt
This code was added before the accurate pv patch, when
we retrieved PV directly from TT.

It's not required for correct (and long) PVs any more and
should be safe to remove it.

Also, allowing helper threads to repeatedly over-write
TT doesn't seem to make sense(that was probably an un-intended
side-effect of lazy smp). Before Lazy SMP only Main Thread used
to run ID loop and insert PV into TT.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 74346 W: 13946 L: 13918 D: 46482

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 47265 W: 6531 L: 6447 D: 34287

bench: 8819179
2016-06-15 08:55:40 +02:00
lucasart 126036abb0 Do not hardcode Debug Log File
Allow to specifiy the log file name, this comes
handy in case of self-matches so that each SF
instance writes into a different log file.

No functional change.
2016-06-15 08:47:08 +02:00
Marco Costalba ca14345ba2 Filter root moves filter before copy to threads
Currently root moves are copied to all teh threads
but are DTZ filtered only in main thread at the
beginning of teh search.

This patch moves the TB filtering before the
copy of root moves fixing issue #679

https://github.com/official-stockfish/Stockfish/issues/679

No bench change.
2016-06-11 09:24:40 +02:00
VoyagerOne 7d2a79f037 Stat Formula Tweak
bonus = d * d + 2 * d - 2

STC:
LLR: 2.94 (-2.94,2.94) [0.00,4.00]
Total: 99444 W: 18274 L: 17778 D: 63392

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 89757 W: 12285 L: 11896 D: 65576

bench: 8276130
2016-06-10 07:27:47 +02:00
VoyagerOne 7c5d724724 Tweak check extension condition
There are two concepts with this patch:

Limit check extensions by using move count.
The idea is to limit search explosion.

Always extend check if the first move gives check.
The idea is to save expensive SEE calls, since the vast
majority of first move will have SEE value >= 0, also
first move may still be strong even if the SEE is negative.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16503 W: 3068 L: 2873 D: 10562

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 37202 W: 5261 L: 5014 D: 26927

bench: 8543366
2016-06-10 07:15:56 +02:00
Alain SAVARD 6fed8ff22a Small Queen simplification
Moving a few lines from evaluate_threats to evaluate_pieces allows to
a) Remove a condition pos.count<QUEEN>(Them) == 1
b) use precalculated s instead of pos.square(Them)
c) do not check the condition at all in queenless endings

Passed STC
http://tests.stockfishchess.org/tests/view/5752e0410ebc59029919b1f4
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 67175 W: 12194 L: 12152 D: 42829

and LTC
http://tests.stockfishchess.org/tests/view/57587b140ebc59029919b2f4
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 20276 W: 2774 L: 2653 D: 14849

bench: 7907962
2016-06-10 06:53:07 +02:00
mstembera 0c076f1136 Avoid some redundant scaling function calls
Posted by Mohammed Li here:
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/N-PHfN0O79o

No functional change.
2016-06-10 06:43:37 +02:00
ElbertoOne e48c7547c5 LMR reduction parameter tweak
More reduction for cut nodes, less for moves that escape a capture:

STC (http://tests.stockfishchess.org/tests/view/57548c1e0ebc59029919b247):
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 60165 W: 11519 L: 11149 D: 37497

LTC (http://tests.stockfishchess.org/tests/view/57555b570ebc59029919b260):
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 10353 W: 1493 L: 1317 D: 7543

Bench: 8902859
2016-06-07 15:15:49 +02:00
Marco Costalba 6e2ca97d93 Fix syzygy DTZ bug
In this position: 3K4/8/3k4/8/4p3/4B3/5P2/8 w - - 0 5

Current DTZ probe returns 1 instead of 15

What happens is that the double push f4 is erroneously detected as a win move.

After the push we have:

[D]3K4/8/3k4/8/4pP2/4B3/8/8 b - f3 0 5

And here the code misses the possible ep capture exf3.

The bug is in probe_dtz_no_ep() where is used probe_ab() that is
blind to ep captures so it returns v == 2 (win) for position

3K4/8/3k4/8/4pP2/4B3/8/8 b - f3 0 5

Note that at the caller site the original position did not have any
possible ep capture, so probe_dtz() returns immediately after calling
probe_dtz_no_ep().

The fix is to call the ep-aware probe_wdl() instead of probe_ab()

I have verified that DTZ is correct now and also there are no more
mistmatches compared to the new 'syzygy' branch. Tested on a set of
more than 600 endgame positions, included some tricky ones.

For people interested to redo the test or doing additional tests
please pull branch tb_dbg from https://github.com/mcostalba/Stockfish repo.

bench: 8450534 (bench unaffected because syzygy is not exercized during bench)
2016-06-07 15:06:26 +02:00
VoyagerOne 5f096e9bef Simplify Futility Pruning
Don't update bestValue when futility pruning.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 21933 W: 4031 L: 3912 D: 13990

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 46225 W: 6115 L: 6028 D: 34082

Bench: 8450534
2016-06-03 19:58:42 +02:00
ElbertoOne 20023ac9b8 LMR Simplification
LMR simplification that also gives a slight ELO gain, especially at LTC:

STC (http://tests.stockfishchess.org/tests/view/574ec8e20ebc59029919b147):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32402 W: 5967 L: 5866 D: 20569

LTC (http://tests.stockfishchess.org/tests/view/574fbebf0ebc59029919b16d):
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 15103 W: 2103 L: 1975 D: 11025

Bench: 8248133
2016-06-03 19:53:04 +02:00
joergoster 3549d98d07 Tuned values for piece check and attack unit factors
A middle ground patch of two successful tuning patches,
one at STC, the other at LTC, which now passed both.

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 67893 W: 12777 L: 12384 D: 42732

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 30165 W: 4189 L: 3960 D: 22016

bench: 9209507
2016-06-03 07:39:27 +02:00
Stéphane Nicolet 07b247f943 Pins or discovered attacks on the opponent's queen
Bonus for pins or discovered attacks on the opponent's queen

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 32020 W: 5914 L: 5652 D: 20454

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 10946 W: 1530 L: 1375 D: 8041

Bench: 7031649
2016-05-28 14:57:58 +02:00
Stéphane Nicolet abac509ccb Teach check_blockers to check also non-king pieces
This is a prerequisite for next patch

No functional change.
2016-05-28 14:52:21 +02:00
Stéphane Nicolet ab0f4c0353 Simplify doubled pawn
Only use doubled pawn malus when the doubled pawns are on consecutive squares.

Passed STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7678 W: 1469 L: 1325 D: 4884

And LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 26739 W: 3562 L: 3449 D: 19728

Bench: 8211685
2016-05-26 12:33:44 +02:00
Leonid Pechenik 71bfbb22fc More detailed dependence of time allocation on the magnitude of score change
10+0.1:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 5657 W: 1130 L: 979 D: 3548

60+0.6:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 36884 W: 5002 L: 4762 D: 27120

bench: 8428997
2016-05-20 19:44:50 +02:00
loco-loco 7cb8cbb403 Assorted pruning tweaks
LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 38257 W: 5206 L: 4961 D: 28090

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16550 W: 3110 L: 2914 D: 10526

Bench: 8428997
2016-05-20 19:34:49 +02:00
mstembera 0784bd542b Fix a multiPV bug in lazy SMP
Where the helper threads were not doing multiPV at all.

Regression tested sprt @ 5+0.05 th 7

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 73918 W: 11891 L: 11853 D: 50174

bench: 8716243
2016-05-14 21:34:55 +02:00
Stéphane Nicolet 16c603ce9b Double pawn simplification
Try doubled pawn simplification, with psq
table compensation.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36094 W: 6558 L: 6463 D: 23073

LTC:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 102352 W: 13417 L: 13404 D: 75531

Bench: 8716243
2016-05-13 13:01:59 +02:00
loco-loco 969982406c Merge good and bad quiets
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 58613 W: 10779 L: 10723 D: 37111

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33608 W: 4539 L: 4436 D: 24633

Bench: 9441294
2016-05-10 18:15:16 +02:00
Alain SAVARD 4b9ed6566a Unsafe checks
Introducing a new multi-purpose penalty related to King safety, which
includes all kind of potential checks (from unsafe or unavailable
squares currently occupied by some other piece)

This will indirectly detect and reward some pins, discovered checks, and
motifs such as square vacation, or rook behind its pawn and aligned with
King (example Black Rg8, g7 against Kg1),
and penalize some pawn blockers (if they move, it allows a discovered
check by the pawn).

And since it looks also at protected squares, it detects some potential
defense overloading.

Finally, the rook contact checks had been removed some time ago. This
test will give a small bonus for them, as well as for bishop contact
checks.

Passed STC
http://tests.stockfishchess.org/tests/view/5729ec740ebc59301a354b36
LLR: 2.94 (-2.94,2.94) [0.00,5.00]
Total: 13306 W: 2477 L: 2296 D: 8533

and LTC
http://tests.stockfishchess.org/tests/view/572a5be00ebc59301a354b65
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 20369 W: 2750 L: 2565 D: 15054

bench: 9298175
2016-05-06 20:04:57 +02:00
Marco Costalba 5e4cd3fc0d Retire __popcnt64 intrinsic
Just use _mm_popcnt_u64() that is available
both for MSVC abd Intel compiler.

Verified on MSVC that the produced assembly
has the hardware 'popcnt' instruction.

No functional change.
2016-05-05 09:09:07 +02:00
VoyagerOne 5486911e01 Simplify History LMR Formula
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 41713 W: 7589 L: 7504 D: 26620

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 41353 W: 5484 L: 5391 D: 30478

Bench: 8946983
2016-05-05 08:53:50 +02:00
Marco Costalba 3487eb9f9e Fix a warning with MSVC
Introduced by 2dd24dc4e6 ("Use popcount intrinsic with Intel")

No functional change.
2016-05-01 15:10:33 +02:00
joergoster dc0030de4f Fix LazySMP when searching to a fixed depth.
Currently, helper threads will only search up to the
specified depth limit. Now let them search until the
main thread has finished the specified depth.

On the other hand, we don't want to pick a thread with
a higher search depth.

This may be considered cheating. ;-)

No functional change.
2016-05-01 14:30:50 +02:00
erbsenzaehler 2dd24dc4e6 Use popcount intrinsic with Interl compiler
It seems that icc used our fallback version of popcount.
Now use intrinsics.

icc version 16.0.2 (gcc version 5.3.0 compatibility)
bmi2 compile
uname -r 4.5.1-1-ARCH

20xbench gives a nice speedup
./stockfish-icc-master 2161515 +- 34462
./stockfish-icc-sse42 2260857 +- 50349
2016-05-01 14:18:16 +02:00
Krgp 8f934dff9a Remove useless -mbmi flag in Makefile
I could not find anything documented that is necessary that prepending -mbmi to -mbmi2 gives some benefit.
Instead at
https://gcc.gnu.org/onlinedocs/gcc/x86-Built-in-Functions.html#x86-Built-in-Functions

The following built-in functions are available when -mbmi is used. All of them generate the machine instruction that is part of the name.
unsigned int __builtin_ia32_bextr_u32(unsigned int, unsigned int);
unsigned long long __builtin_ia32_bextr_u64 (unsigned long long, unsigned long long);

The following built-in functions are available when -mbmi2 is used. All of them generate the machine instruction that is part of the name.
unsigned int _bzhi_u32 (unsigned int, unsigned int)
unsigned int _pdep_u32 (unsigned int, unsigned int)
unsigned int _pext_u32 (unsigned int, unsigned int)
unsigned long long _bzhi_u64 (unsigned long long, unsigned long long)
unsigned long long _pdep_u64 (unsigned long long, unsigned long long)
unsigned long long _pext_u64 (unsigned long long, unsigned long long)

and at
https://gcc.gnu.org/ml/gcc/2014-02/msg00204.html

( "... The real optimization comes from being able to use pext
(parallel bit extract), which can implement several bextr expressions in
parallel.")

Apart from that we don't use all -msse -msse2 -msse3 -msse4.2 etc. but just -msse3 (or -msse4.2) only.

As regards to the speedup within noise level - this pull request is actually reversal of mcostalba#198 wherein prepending -mbmi to -mbmi2 was claimed to be 0.3% faster and here (removing -mbmi) gives 0.4% speed gain.
2016-05-01 14:11:28 +02:00
Stéphane Nicolet 2694ef23c3 Isolated pawn simplification
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 117822 W: 21697 L: 21744 D: 74381

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 92307 W: 12330 L: 12305 D: 67672

Bench: 8813983

Resolves #659
2016-04-30 22:23:22 +01:00
VoyagerOne e082112cfe Use FMHs to assist with LMR formula.
STC:
LLR: 2.99 (-2.94,2.94) [0.00,5.00]
Total: 52232 W: 9654 L: 9304 D: 33274

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 115988 W: 15550 L: 15049 D: 85389

Bench: 7890808

Resolves #651
2016-04-24 01:00:58 +01:00
erbsenzaehler 4048bae47b Use -O3 for all compilers (including ICC)
There seems to be no benefit from using -fast over -O3 with icc.
So use -O3 everywhere.

No functional change

Resolves #652
2016-04-24 00:55:56 +01:00
DU-jdto c737062436 Remove some pointless micro-optimizations
Seems to give around 1% speed-up for CPUs with popcnt support.
Seems to give a very minor speed-up for CPUs without popcnt.

No functional change

Resolves #646
2016-04-23 02:04:28 +01:00
Marco Costalba 94e41274bb Fix incorrect draw detection
In this position we should have draw for repetition:

position fen rnbqkbnr/2pppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1 moves g1f3 g8f6 f3g1
go infinite

But latest patch broke it.

Actually we had two(!) very subtle bugs, the first is that Position::set()
clears the passed state and in particular 'previous' member, so
that on passing setupStates, 'previous' pointer was reset.

Second bug is even more subtle: SetupStates was based on std::vector
as container, but when vector grows, std::vector copies all its contents
to a new location invalidating all references to its entries. Because
all StateInfo records are linked by 'previous' pointer, this made pointers
go stale upon adding more element to setupStates. So revert to use a
std::deque that ensures references are preserved when pushing back new
elements.

No functional change.
2016-04-18 00:13:16 +02:00
loco-loco ec6aab0136 Add a second level of follow-up moves
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 6438 W: 1229 L: 1077 D: 4132

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 4000 W: 605 L: 473 D: 2922

bench: 7378965

Resolves #636
2016-04-17 15:19:20 +01:00
Marco Costalba 7eaea3848c StateInfo is usually allocated on the stack by search()
And passed in do_move(), this ensures maximum efficiency and
speed and at the same time unlimited move numbers.

The draw back is that to handle Position init we need to
reserve a StateInfo inside Position itself and use at
init time and when copying from another Position.

After lazy SMP we don't need anymore this gimmick and we can
get rid of this special case and always pass an external
StateInfo to Position object.

Also rewritten and simplified Position constructors.

Verified it does not regress with a 3 threads SMP test:
ELO: -0.00 +-12.7 (95%) LOS: 50.0%
Total: 1000 W: 173 L: 173 D: 654

No functional change.
2016-04-17 08:29:33 +02:00
Niklas Fiekas ee7a68ea5f Fix last search info carried over to mate position
When starting search in a mate or stalemate position, Stockfish does not
even care to reinitialize and start worker threads. However after search
all threads are checked for the best move.

This can lead to bestmove and info beeing carried over from the last
search.

Example session:

    setoption name threads value 7
    go movetime 4000
    position startpos moves f2f3 e7e5 g2g4 d8h4
    go movetime 4000

Actual output is like (almost always):

    [...]
    bestmove e2e4
    info depth 0 score mate 0
    info depth 20 seldepth 29 multipv 1 score cp 28 [...] pv e2e4
    bestmove e2e4

Expected output / output after fix:

    [...]
    bestmove e2e4 ponder e7e6
    info depth 0 score mate 0
    bestmove (none)

Resolves #623
2016-04-16 10:22:36 +01:00
Marco Costalba d30994ecd5 Hide global visibility when not needed
Also move PieceValue definition in psqt.cpp,
where it is initialized.

Fix a warning in popcount16() with Intel compiler

No functional change.
2016-04-09 10:42:04 +02:00
Marco Costalba bd04f9a0f1 Fix Travis Cl
Broken after "32-bit/64-bit Makefile fix" commit.

Ubuntu "Precise" 12.04.5 supports multilib only until
g++ 4.6 that is not enough to compile Stockfish.

So move to Ubuntu 14.04.4 LTS (Trusty Tahr)

No functional change.
2016-04-09 09:35:19 +02:00
DU-jdto 1cbba8d6fa Small passed pawn simplification
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 21993 W: 4197 L: 4078 D: 13718

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 67213 W: 9135 L: 9077 D: 49001

Bench: 7482426

Resolves #622
2016-04-08 19:48:03 +01:00
Alain SAVARD e9e5f72c74 Undefended King Ring
There was already a penalty for squares only defended by King (undefended)

This test records a penalty for completely undefended squares in the so called extended king-ring
(so if we exclude squares defended by a Kg8 for example, we only look at h6 g6 and f6)

We also exclude squares occupied by opponent pieces in this computation,
based on the following results

Was yellow at STC
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 112499 W: 20649 L: 20293 D: 71557

and passed LTC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 36805 W: 5100 L: 4857 D: 26848

Bench: 8430233

Resolves: #619
2016-04-08 19:22:33 +01:00
Alain SAVARD 29b5842da8 Backward simplication
On top of the usual conditions
a) some opponent in front (but no lever)
b) some neighbours (in front) (but no neighbour behind or same rank)
c) < rank_5

to find out if a pawn is backward we look at the squares in front of this pawn to reach the same rank as the next neighbour.

In current master, a pawn is backward if any of those squares is controlled by an enemy pawn on an adjacent file

In this version, a pawn is ALSO backward if any of those squares is occupied by an enemy pawn.

STC:
http://tests.stockfishchess.org/tests/view/56fe7efd0ebc59301a3541f1
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 19051 W: 3557 L: 3433 D: 12061

LTC:
http://tests.stockfishchess.org/tests/view/56febc2d0ebc59301a354209
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40810 W: 5619 L: 5526 D: 29665

Bench: 7525245

Resolves #614
2016-04-08 19:12:55 +01:00
mstembera 8fb45caade Simplify popcnt
Also a speedup(about 1%) on 64-bit w/o hardware popcnt

Retire Max15 and Full template parameters
(Contributed by Marco Costalba)

Now that we have just SW and HW versions, use
template default parameter to get rid of explicit
template parameters.

Retire bitcount.h and move the only defined
function to bitboard.h

No functional change

Resolves #620
2016-04-08 18:52:15 +01:00
lucasart 900279a06f 32-bit/64-bit Makefile fix
Counter intuitively, make build ARCH=x86-32 does NOT produce a 32-bit compile
when running a 64-bit OS. Nor would ARCH=x86-64 produce a 64-bit compile when
running a 32-bit OS (assuming it compiled w/o errors).

No functional change

Resolves #621
2016-04-08 18:47:31 +01:00
NicklasPersson d5b24ad77b A combo patch of two tuning patches
STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 14223 W: 2700 L: 2494 D: 9029

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 66294 W: 9065 L: 8739 D: 48490

Bench: 7607385

Resolves #612
2016-03-31 13:26:00 +01:00
lucasart f256388e08 Guard against UB in lsb/msb
lsb(b) and msb(b) are undefined when b == 0. This can lead to subtle bugs, where
the resulting code behaves differently on different configurations:
- It can be the home grown software LSB/MSB
- It can be the compiler generated software LSB/MSB (when using compiler
  intrinsics without the right compiler flags to allow compiler to use hardware
  LSB/MSB). Which of course depends on the compiler.
- It can be hardware LSB/MSB generated by the compiler.
- Not to mention that hardware LSB/MSB can return different value on different
  hardware when b == 0.

No functional change

Resolves #610
2016-03-31 13:22:37 +01:00
Marco Costalba db4b0d8b7d Rewrite bsfq management
Use compiler intrinsics when possible to
avoid writing platform specific asm code.

Tested on Windows 7 with MSVC 2013 and mingw 4.8.3 (32 and 64 bit)
and on Linux Mint with g++ 4.8.4 and clang 3.4 (32 and 64 bit).

No functional change

Resolves #609
2016-03-28 15:46:55 +01:00
snicolet 24dac5ccd3 Bonus for loose enemies
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 30504 W: 5743 L: 5485 D: 19276

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 11936 W: 1651 L: 1493 D: 8792

Bench: 8880041

Resolves #606
2016-03-27 20:56:54 +01:00
mbootsector 8788612828 Raise endgame passed pawn and material values
STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 136149 W: 25213 L: 24588 D: 86348

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 54637 W: 7533 L: 7238 D: 39866

Bench: 8546808

Resolves #608
2016-03-27 20:22:48 +01:00
fanon 5d1644ba69 Simplify pawns King Safety calculation
STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 130209 W: 23516 L: 23581 D: 83112

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33541 W: 4563 L: 4460 D: 24518

Bench: 8644370

Resolves #604
2016-03-27 20:08:56 +01:00
VoyagerOne 60590577f2 A small simplification in movepick.h
No functional change

Resolves #597
2016-03-14 20:54:49 -07:00
mstembera 04be84e0e2 Simplify Safe Checks
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 11796 W: 2211 L: 2074 D: 7511

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14324 W: 1935 L: 1806 D: 10583

Bench: 8075202

Resolves #600
2016-03-14 20:51:16 -07:00
Marco Costalba 647402ff79 Assorted cleanup of latest commits
No functional change.

Resolves #601
2016-03-14 20:42:44 -07:00
Stefan Geschwentner a273b6ef8c Add followup moves history for move ordering
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 7955 W: 1538 L: 1378 D: 5039

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 5323 W: 778 L: 642 D: 3903

Bench: 8261839

Resolves #599
2016-03-10 14:26:06 -08:00
mbootsector e1a7d135b2 Passed pawn bonus simplification
STC: (yellow)

LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 86114 W: 16063 L: 15921 D: 54130

LTC:

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14347 W: 2025 L: 1896 D: 10426

Bench: 8576437

Resolves #595
2016-03-02 19:10:56 +00:00
snicolet c1be0c68c7 Tweak initiative formula
Give more weight to the pawns number and
the vertical king distance in evaluate_initiative()

Passed STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26729 W: 5067 L: 4825 D: 16837

and LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 60480 W: 8338 L: 8016 D: 44126

Bench: 8295162

Resolves #594
2016-03-02 00:02:03 +00:00
ElbertoOne c13052f344 Clean up depth reduction calculation
Might also be a slight speed up

No functional change

Resolves #593
2016-02-28 13:40:47 +00:00
joergoster 8de29390f2 Pass endgame value to evaluate_scale_factor()
No functional change

Resolves #592
2016-02-28 13:35:34 +00:00
VoyagerOne 45a309d92e Simplify Reduction Formula
Formula now only contains one coefficient. Making it much easier to tune.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 187443 W: 34858 L: 35028 D: 117557

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 88329 W: 11982 L: 11953 D: 64394

Bench: 7521394

Resolves #591
2016-02-28 13:31:25 +00:00
Leonid Pechenik d5ba8e827d Revert "Remove slowMover"
This reverts commit 77fa960f89.

Resolves #590
2016-02-28 13:25:05 +00:00
IIvec 77fa960f89 Remove slowMover
Removes a slowMover and one paramater from move_importance function.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 77023 W: 14456 L: 14433 D: 48134

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 37175 W: 5190 L: 5092 D: 26893

Resolves #589
2016-02-21 20:16:28 +00:00
VoyagerOne bfe9044ad9 History Stat Formula Simplification
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 67476 W: 12561 L: 12521 D: 42394

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 111923 W: 15147 L: 15149 D: 81627

Bench: 8430465

Resolves #588
2016-02-21 20:12:33 +00:00
VoyagerOne 744ed85a4d Fix futility pruning bug
PredictedDepth can be negative, causing the futility_margin to be negative.
It will be very difficult to tweak moveCount pruning and reduction formula, as they are tuned to prevent this behavior.

No functional change

Resolves #587
2016-02-14 19:48:46 +00:00
hxim 56dd58e6f9 Remove Weights
Removed remaining redundant weights for pawn structure,
passed pawns, space and king safety by redistributing them
into individual evaluation terms.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 15173 W: 2790 L: 2659 D: 9724

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 43433 W: 5936 L: 5846 D: 31651

Bench: 7156237

Resolves #586
2016-02-07 21:00:24 +00:00
Marco Costalba 4f6aa15228 Document HalfDensityMap
No functional change.

Resolves #584
2016-02-07 20:54:25 +00:00
Leonid Pechenik aedebe35cf Time management simplification
10+0.1:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 41963 W: 7967 L: 7883 D: 26113

60+0.6:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 132314 W: 17939 L: 17969 D: 96406

Resolves #580
2016-01-29 00:47:07 +00:00
Guenther Demetz 9a10313a9d rotating symmetric patterns with increasing skipsize
STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00] sprt @ 5+0.1 th 21
Total: 7068 W: 1121 L: 975 D: 4972

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00] sprt @ 12+0.12 th 21
Total: 26691 W: 3594 L: 3481 D: 19616

No functional change with a single thread

Resolves #574
2016-01-24 14:08:01 +00:00
Joona Kiiski 8c3a5bbc52 Do not probe syzygy bases when castling is possible
Almost no functional change. Bench is unchanged.

Resolves #230
Resolves #573
2016-01-20 15:24:21 +00:00
lucasart 28933a580e Retire RootNode template
There is no reason to compile 3 different copies of search(). PV nodes are on
the cold path, and PvNode is a template parameter, so there is no cost in
computing:

const bool RootNode = PvNode && (ss-1)->ply == 0;

And this simplifies code a tiny bit as well.

Speed impact is negligible on my machine (i7-3770k, linux 4.2, gcc 5.2):

            nps   +/-
test    2378605  3118
master  2383128  2793
diff      -4523  2746

Bench: 7751425

No functional change.

Resolves #568
2016-01-18 22:21:42 +00:00
Guenther Demetz 12eb345ebd Depth margin parameter-tweak in TT-save
Verified that is improvement with multiple threads:

LLR: 2.95 (-2.94,2.94) [0.00,4.00] sprt @ 30+0.3 th 3
Total: 14817 W: 2103 L: 1915 D: 10799

LLR: 2.96 (-2.94,2.94) [0.00,4.00] sprt @ 15+0.15 th 7
Total: 10264 W: 1498 L: 1321 D: 7445

Verified that is not a significant regression with a single thread:

LLR: 2.96 (-2.94,2.94) [-4.00,0.00] sprt @ 60+0.6 th 1
Total: 23975 W: 3294 L: 3210 D: 17471

Resolves #575
2016-01-18 22:04:38 +00:00
Joona Kiiski ebec2fa48e Remove redundant -std=c++0x flag
This flag is functionally identical to '-std=c++11' flag which
is part of standard flags.

No functional change

Resolves #571
2016-01-18 21:54:40 +00:00
Joona Kiiski 552b3ccb66 Makefile: Allow specifying compiler executable
No functional change

Resolves #570
2016-01-18 21:47:52 +00:00
Marco Costalba 356147d99a Rewrite time formula
Time management is really too complex, our aim is
to simplify it, but for time being at least rewrite
in an understandable way.

No functional change.
2016-01-18 17:12:18 +01:00
Lyudmil Antonov 89723339d9 Assorted English grammar changes
No functional change

Resolves #567
2016-01-16 21:34:29 +00:00
Stefano80 74e2fa97b7 Adjust reductions based on history and cmh tables
STC:
LLR: 4.06 (-2.94,2.94) [0.00,5.00]
Total: 149395 W: 28029 L: 27208 D: 94158

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 9628 W: 1368 L: 1217 D: 7043

bench: 8076724

Resolves #565
2016-01-13 16:18:35 +00:00
Stefano80 dcd8ce7094 Update comments in LMR step
No functional change

Resolves #564
2016-01-13 16:03:53 +00:00
Leonid Pechenik 4d1220d672 Tune time management for LTC
60+0.6:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 102533 W: 14270 L: 13842 D: 74421

Resolves #558
2016-01-09 09:21:52 +00:00
lucasart aa31f7f096 Retire CenterBind
And compensate in the PSQT.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 27714 W: 5161 L: 5052 D: 17501

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36354 W: 5008 L: 4909 D: 26437

Bench: 8603285

Resolves #556
2016-01-03 17:33:36 +00:00
Alain SAVARD 1de97337c2 Fine tuning of unsupported pawn penalty
Adjust the unsupported pawn penalty when the pawn is supporting 2 pawns
(for example g7 in f6-g7-h6)

Passed STC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23833 W: 4384 L: 4158 D: 15291

Passed LTC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 42711 W: 5918 L: 5655 D: 31138

Bench: 8390233

Resolves #549
2016-01-03 14:15:29 +00:00
Leonid Pechenik 9eceb894ac Adjust time used for move based on previous score
Use less time if evaluation is not worse than for previous move and even less time if in addition no fail low encountered for current iteration.

STC: 10+0.1
ELO: 5.37 +-2.9 (95%) LOS: 100.0%
Total: 20000 W: 3832 L: 3523 D: 12645

STC: 10+0.1
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 17527 W: 3334 L: 3132 D: 11061

LTC: 60+0.6
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 28233 W: 3939 L: 3725 D: 20569

LTC: 60+0.6
ELO: 2.43 +-1.4 (95%) LOS: 100.0%
Total: 60000 W: 8266 L: 7847 D: 43887

LTC: 60+0.06
LLR: 2.95 (-2.94,2.94) [-1.00,3.00]
Total: 38932 W: 5408 L: 5207 D: 28317

Resolves #547
2016-01-03 14:01:15 +00:00
Joona Kiiski 5972c4a678 Restore development version 2016-01-03 13:52:01 +00:00
Joona Kiiski dd9cf30581 Stockfish 7
Bench: 8355485

No functional change
2016-01-02 20:26:02 +00:00
ppigazzini d4af15f682 Update AUTHORS and copyright notice
No functional change

Resolves #555
2016-01-02 09:43:51 +00:00
Marco Costalba 9742fb10fd Update Copyright year
No functional change.

Resolves #554
2016-01-01 10:17:36 +00:00
Joona Kiiski 7cafbab529 Stockfish 7 Beta 2
Bench: 8355485

No functional change
2015-12-30 11:32:45 +00:00
Alain SAVARD 6a79d5d10e Correct Pawn Trace Score + Code Clean up
No functional change

Resolves #542
2015-12-30 11:30:28 +00:00
Marco Costalba 3e4fed3a91 Fix assert with very high score position
In case of a very high material score, we can
overflow VALUE_INFINITE.

This patch fixes an assert with:

position fen 7k/QQQQR3/2B5/4KN1Q/3QQ3/8/8/4R3 b - - 0 1
go depth 1

No functional change.

Resolves #546
2015-12-30 11:16:48 +00:00
Joona Kiiski a5c76d69c3 Stockfish 7 Beta 1
Bench: 8355485

No functional change
2015-12-27 19:36:56 +00:00
Marco Costalba 1b5b900a29 Move some globals into main thread scope
Make it explicit that those variables are not globals, but
are used only by main thread. I think it is a sensible
clarification because easy move is already tricky enough
and current patch makes the involved actors explicit.

No functional change.

Resolves #537
2015-12-27 19:29:16 +00:00
Marco Costalba 90f5937373 Revert "Fix compiling of 32 bit binary on 64-bit Windows"
This reverts commit 1e8836d921

Broken compile on mingw under Windows:

Config:
debug: 'yes'
optimize: 'yes'
arch: 'i386'
bits: '32'
prefetch: 'yes'
bsfq: 'no'
popcnt: 'no'
sse: 'yes'
pext: 'no'

Flags:
CXX: i686-w64-mingw32-c++
CXXFLAGS: -Wall -Wcast-qual -fno-exceptions -fno-rtti -std=c++11  -Wextra -Wshadow -g -O3 -msse
LDFLAGS:  -static

Testing config sanity. If this fails, try 'make help' ...

mingw32-make[1]: Leaving directory 'C:/stockfish/src'
c:/MinGw/bin/mingw32-make ARCH=x86-32 COMP=mingw all
mingw32-make[1]: Entering directory 'C:/stockfish/src'
sh: C:\Program: No such file or directory
i686-w64-mingw32-c++ -Wall -Wcast-qual -fno-exceptions -fno-rtti -std=c++11  -Wextra -Wshadow -g -O3 -msse   -c -o benchmark.o benchmark.cpp
<builtin>: recipe for target 'benchmark.o' failed
process_begin: CreateProcess(NULL, i686-w64-mingw32-c++ -Wall -Wcast-qual -fno-exceptions -fno-rtti -std=c++11 -Wextra -Wshadow -g -O3 -msse -c -o benchmark.o benchmark.cpp, ...) failed.
make (e=2): Impossibile trovare il file specificato.

mingw32-make[1]: *** [benchmark.o] Error 2
mingw32-make[1]: Leaving directory 'C:/stockfish/src'
makefile:401: recipe for target 'build' failed
mingw32-make: *** [build] Error 2

No functional change.
2015-12-23 09:41:56 +01:00
braich 1e8836d921 Fix compiling of 32 bit binary on 64-bit Windows
Two versions of mingw-w64 (targeting Win64 and Win32)
can be installed on Windows too.

No functional change

Resolves #532
2015-12-22 19:07:10 +00:00
DU-jdto e3c85c314d Remove another unnecessary Search::Stack field
No functional change

Resolves #535
2015-12-22 17:16:15 +00:00
Alain SAVARD da34023cfc New mobility bonus
Tuned the global mobility factor for each piece, as well as some +- delta,

The master mobility factor was {266,334} and tuning gave
{267, 362} +S(-2,-2) for the Knight
{249, 328} +S( 0,-2) for the Bishop
{298, 353} +S(1,1) for the Rook
{265, 358} +S(2,-1) for the Queen

Passed STC
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 49402 W: 9367 L: 9037 D: 30998

and LTC
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 26831 W: 3871 L: 3658 D: 19302

Bench: 8355485

Resolves #536
2015-12-22 10:02:37 +00:00
VoyagerOne ed72a1e9ba Remove killer move conditions from LMR
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 8459 W: 1619 L: 1477 D: 5363

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32239 W: 4404 L: 4299 D: 23536

Bench: 7597031

Resolves #534
2015-12-20 20:41:17 +00:00
loco-loco 5bbd944099 Remove unused field SearchStack::ttMove
No functional change

Resolves #533
2015-12-20 20:37:18 +00:00
Guenther Demetz 38adb487ca Distinct iteration paths for Lazy SMP threads
STC 5+0.1, threads 7
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 6026 W: 1047 L: 901 D: 4078

LTC: 20+0.2, threads 7
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 19739 W: 2910 L: 2721 D: 14108

STC 5+0.1, threads 20
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 2493 W: 462 L: 331 D: 1700

LTC 30+0.3, threads 20
ELO: 8.86 +-3.7 (95%) LOS: 100.0%
Total: 8000 W: 1076 L: 872 D: 6052

Bench: 8012530

Resolves #525
2015-12-18 21:59:09 +00:00
mstembera 2c1797ab81 Fix easy move bug in SMP mode
Fix a bug where we could stop the search after only 10% of time used due to a matching easy move but later switch to a different move that was never pre-screened as easy due to SMP thread select.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 27227 W: 4910 L: 4800 D: 17517

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 40368 W: 5826 L: 5733 D: 28809

Resolves #521
2015-12-14 18:47:12 +00:00
Alain SAVARD 9fac5774f4 Threats retuned
STC:

LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 45239 W: 8913 L: 8591 D: 27735

LTC:

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 21046 W: 3200 L: 2989 D: 14857

Bench: 8012530

Resolves #526
2015-12-14 18:11:13 +00:00
Leonid Pechenik 69240a982d Simplify time management and fix 'ponder on' bug
Simplify time management code by removing hard stops for unchanging first root moves.
Search is now stopped earlier at the end iteration if it did not have fail-lows at root.

This simplification also fixes pondering bug. Ponder flag was true by default
and cutechess-cli doesn't change it to false even though no pondering is possible.
Fix the issue by setting the default value of 'Ponder' flag to false.

10+0.1:
ELO: 3.51 +-3.0 (95%) LOS: 99.0%
Total: 20000 W: 3898 L: 3696 D: 12406

40+0.4:
ELO: 1.39 +-2.7 (95%) LOS: 84.7%
Total: 20000 W: 3104 L: 3024 D: 13872

60+0.06:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 37231 W: 5333 L: 5236 D: 26662

Stopped run at 100+1:
LLR: 1.09 (-2.94,2.94) [-3.00,1.00]
Total: 37253 W: 4862 L: 4856 D: 27535

Resolves #523
Fixes #510
2015-12-14 18:00:52 +00:00
mstembera 7904a7d930 Fix MultiPv and Skill in SMP.
7 threads, 5+0.1:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 55460 W: 9665 L: 9601 D: 36194

No functional change in normal playing mode
2015-12-12 10:32:28 +00:00
absimaldata 8315620522 New Tuned Weights
More accurate evaluation weights

Performed better at STC

LLR: 1.32 (-2.94,2.94) [0.00,4.00]
Total: 190043 W: 37433 L: 36675 D: 115935

Passed LTC

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 30157 W: 4540 L: 4303 D: 21314

Bench: 9264977

Resolves #515
2015-12-05 10:31:00 +00:00
Marco Costalba 52eac1e535 Simplify outpost code
Also inline defintions of SpaceMask and CenterBindMask.

Verified from assembly that compiler computes the values
at compile time, so it is also theoretical faster.

While there factor out scale factor evaluation.

No functional change.
2015-12-05 11:16:19 +01:00
mstembera 5d077bb482 Proper Makefile for cross compiling 64 or 32 bit PGO + LTO + static Windows binaries under Linux.
No functional change

Resolves #511
2015-12-03 14:01:38 +00:00
Stéphane Nicolet 5a42427cb8 Introduce new Threats weights = {350, 256}
Raise the midgame threats weight by 37%.

Passed STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 8165 W: 1675 L: 1487 D: 5003

and LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 28181 W: 4141 L: 3912 D: 20128

Bench: 7824961

Resolves #512
2015-12-02 11:18:26 +00:00
Joona Kiiski 5339ab57d8 Revert "Allow cross compilation of Windows binaries on a Linux system"
This reverts commit 388630ae28.

Confuses fishtest build system
2015-11-24 17:24:45 -08:00
mstembera 388630ae28 Allow cross compilation of Windows binaries on a Linux system
that are PGO, LTO, and statically linked.
Credit: pasquale....@gmail.com

No functional change

Resolves #505
2015-11-23 17:06:14 -08:00
mstembera 79f393084a Clean up RootMove less operator
This is used by std::stable_sort() to sort moves from highest score to lowest score.

1) The comment is incorrect since highest to lowest means descending.
2) It's more natural to implement a less operator using another less operator rather than a greater operator.

No functional change.

Resolves #504
2015-11-20 23:30:07 -08:00
lucasart 328098d027 Fix TT comment and static_assert()
Comment is based on a misunderstanding of what unaligned memory access is. Here
is an article that explains it very clearly:
https://www.kernel.org/doc/Documentation/unaligned-memory-access.txt

No matter how we define TTEntry or TTCluster, there will never be any unaligned
memory access. This is because the complier knows the alignment rules, and does
the necessary adjustments to make sure unaligned memory access does not occur.

The issue being adressed here has nothing to do with unaligned memory access. It
is about cache performance. In order to achieve best cache performance:
- we prefetch the cacheline as soon as possible.
- we ensure that TT clusters do not spread across two cachelines. If they did,
  we would need to prefetch 2 cachelines, which could hurt cache performance.

Therefore the true conditions to achieve this are:
1/ start adress of TT is cache line aligned. void TranspositionTable::resize()
enforces this.
2/ TT cluster size should *divide* the cache line size. Currently, we pack 2
clusters per cache lines. It used to be 1 before "TT sardines". Does not matter
what the ratio is, all we want is to fit an integer number of clusters per cache
line.

No functional change.

Resolves #506
2015-11-20 23:23:53 -08:00
Marco Costalba 93195555ed Rewrite how threads are spawned
Instead of creating a running std::thread and
returning, wait in Thread c'tor that the native
thread of execution goes to sleep in idle_loop().

In this way we can simplify how search is started,
because when main thread is idle we are sure also
all other threads will be idle, in any case, even
at thread creation and startup.

After lazy smp went in, we can simpify and rewrite
a lot of logic that is now no more needed. This is
hopefully the final big cleanup.

Tested for no regression at 5+0.1 with 3 threads:
LLR: 2.95 (-2.94,2.94) [-5.00,0.00]
Total: 17411 W: 3248 L: 3198 D: 10965

No functional change.
2015-11-21 07:48:50 +01:00
VoyagerOne 07e0741dfb History Pruning: Don't prune the main killer move.
Also increased pruned depth to 4.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23380 W: 4581 L: 4350 D: 14449

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 28934 W: 4329 L: 4105 D: 20500

Bench: 8369743

Resolves #498
2015-11-15 20:40:09 -08:00
Ken T Takusagawa 27ca89d479 Do not conceal the invocation of the benchmark program
It is better to be able to see what arguments it is being called with.

No functional change

Resolves #497
2015-11-15 20:34:37 -08:00
Stefan Geschwentner cccf3c815c Bonus for reachable outpost
Give a bonus for outpost squares which in reach of a bishop or knight.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 22725 W: 4570 L: 4339 D: 13816

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 15019 W: 2333 L: 2157 D: 10529

Bench: 8503181

Resolves #495
2015-11-14 06:31:45 -08:00
Marco Costalba 76ed0ab501 Retire ThreadBase
Now that we don't have anymore TimerThread, there is
no need of this long class hierarchy.

Also assorted reformatting while there.

To verify no regression, passed at STC with 7 threads:
LLR: 2.97 (-2.94,2.94) [-5.00,0.00]
Total: 30990 W: 4945 L: 4942 D: 21103

No functional change.
2015-11-13 08:22:44 +01:00
Marco Costalba ce84ab6e9d Fix broken UCI 'wait for stop'
When we reach the maximum depth, we can finish the
search without a raise of Signals.stop. However, if
we are pondering or in an infinite search, the UCI
protocol states that we shouldn't print the best move
before the GUI sends a "stop" or "ponderhit" command.

It was broken by lazy smp. Fix it by moving the stopping
of the threads after waiting for GUI.

No functional change.
2015-11-13 08:13:59 +01:00
lucasart fca8dbc029 Avoid friend
operator<<(os, pos) does not need to access any private members of pos.

No functional change.

Resolves #492
2015-11-10 21:46:02 +00:00
lucasart e6eeb17aa6 Ensure that rootDepth < DEPTH_MAX
Indeed, if we use a depth >= DEPTH_MAX, we start having negative depth in the
TT (due to int8_t cast).

No functional change in single thread mode

Resolves #490
2015-11-10 21:41:42 +00:00
Marco Costalba 9c9205860c Get rid of timer thread
Unfortunately std::condition_variable::wait_for()
is not accurate in general case and the timer thread
can wake up also after tens or even hundreds of
millisecs after time has elapsded. CPU load, process
priorities, number of concurrent threads, even from
other processes, will have effect upon it.

Even official documentation says: "This function may
block for longer than timeout_duration due to scheduling
or resource contention delays."

So retire timer and use a polling scheme based on a
local thread counter that counts search() calls and
a small trick to keep polling frequency constant,
independently from the number of threads.

Tested for no regression at very fast TC 2+0.05 th 7:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32969 W: 6720 L: 6620 D: 19629

TC 2+0.05 th 1:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 7765 W: 1917 L: 1765 D: 4083

And at STC TC, both single thread
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 15587 W: 3036 L: 2905 D: 9646

And with 7 threads
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 8149 W: 1367 L: 1227 D: 5555

bench: 8639247
2015-11-03 11:27:00 +01:00
mbootsector 27c5cb5912 Pick bestmove from the deepest thread.
STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26930 W: 4441 L: 4214 D: 18275

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 7783 W: 1017 L: 876 D: 5890

No functional change in single thread mode

Resolves #485
2015-11-02 10:05:43 +00:00
Marco Costalba 86f04dbcc0 Assorted trivia in search.cpp
The only interesting change is the moving of
stack[MAX_PLY+4] back to its original position
in id_loop (now renamed Thread::search).

No functional change.
2015-10-31 19:26:35 +01:00
VoyagerOne aa242d2f84 New History Bonus Formula
bonus = d^2 + d - 1

Bench: 8639247

Resolves #484
2015-10-31 18:17:00 +00:00
Ajith 8d858783e1 Reduce variation in rootDepth between different threads
Reduce the variation in Root Depth between different threads. This
prevents threads from searching at a depth much higher than Main Thread.

Performed well at STC 24 Threads:
ELO: 3.44 +-3.8 (95%) LOS: 96.1%
Total: 10000 W: 1627 L: 1528 D: 6845

And LTC 24 Threads
LLR: 1.43 (-2.94,2.94) [0.00,4.00]
Total: 3804 W: 500 L: 420 D: 2884
ELO : +7.31
p-value: 73.16%

Passed no regression at STC 3 Threads:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40457 W: 7148 L: 7060 D: 26249

And LTC 3 Threads:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 17704 W: 2489 L: 2364 D: 12851

Raising a pull request early as 24 Thread tests are very expensive and
this is clearly a positive gain at high thread counts and high time
controls. The change is a small parameter tweak with no additional
logic.

No functional change for single thread mode.

Resolves #481
2015-10-29 17:06:13 -07:00
Stéphane Nicolet 80d7556af7 Some code and comment cleanup
- Remove all references to split points
- Some grammar and spelling fixes

No Functional change

Resolves #478
2015-10-29 15:28:59 +00:00
lucasart 00d9e9fd28 Use atomics instead of volatile
Rely on well defined behaviour for message passing, instead of volatile. Three
versions have been tested, to make sure this wouldn't cause a slowdown on any
platform.

v1: Sequentially consistent atomics

No mesurable regression, despite the extra memory barriers on x86. Even with 15
threads and extreme time pressure, both acting as a magnifying glass:

threads=15, tc=2+0.02
ELO: 2.59 +-3.4 (95%) LOS: 93.3%
Total: 18132 W: 4113 L: 3978 D: 10041

threads=7, tc=2+0.02
ELO: -1.64 +-3.6 (95%) LOS: 18.8%
Total: 16914 W: 4053 L: 4133 D: 8728

v2: Acquire/Release semantics

This version generates no extra barriers for x86 (on the hot path). As expected,
no regression either, under the same conditions:

threads=15, tc=2+0.02
ELO: 2.85 +-3.3 (95%) LOS: 95.4%
Total: 19661 W: 4640 L: 4479 D: 10542

threads=7, tc=2+0.02
ELO: 0.23 +-3.5 (95%) LOS: 55.1%
Total: 18108 W: 4326 L: 4314 D: 9468

As suggested by Joona, another test at LTC:

threads=15, tc=20+0.05
ELO: 0.64 +-2.6 (95%) LOS: 68.3%
Total: 20000 W: 3053 L: 3016 D: 13931

v3: Final version: SeqCst/Relaxed

threads=15, tc=10+0.1
ELO: 0.87 +-3.9 (95%) LOS: 67.1%
Total: 9541 W: 1478 L: 1454 D: 6609

Resolves #474
2015-10-25 09:15:45 +00:00
Stefano80 3428a28751 KRPPKRP endgame: Simplify ugly switch statement
No functional change

Resolves #470
2015-10-24 23:17:08 +01:00
Marco Costalba 307a5a4f63 Cleanup history stats
And other assorted trivia.

No functional change.
2015-10-24 17:29:12 +02:00
Alain SAVARD 55758344d3 Simplify threats
Using less parameters and code to compute Threats
Includes also a few spacing edits.

Run as a simplification.

Passed STC 10+0.1
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 18879 W: 3725 L: 3600 D: 11554

Passed LTC 60+0.4
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 74116 W: 11001 L: 10958 D: 52157

bench: 8004751
2015-10-24 10:08:26 +02:00
VoyagerOne 4eca981e55 History pruning
Prune moves with negative History
and CMH scores at low depth.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 24182 W: 4672 L: 4439 D: 15071

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 12579 W: 1959 L: 1792 D: 8828

bench: 8907701
2015-10-24 09:30:31 +02:00
Marco Costalba 5066e17eba Update authors
Fishtest is a key factor of SF success.

Thanks to Fishtest we have not only greately
improved ELO but, even more important, we
have enabled a kind of joint development and
testing that it is the herat of on open
source project like SF.

Open sourcing is not just about open code, it is
about commuity development. In case of a chess engine
this has never been possible before due to missing
a strong and strict testing environment that allows
many people to contribute in a safe and coordinate way.

Fishtest is a new way of developing chess engines,
something that has never exsisted before.

No functional change.
2015-10-22 07:03:03 +02:00
Stefano80 328d314f2f Almost passed tuning attempts
Collect and give a second try to some almost passed tuning attempts and
one-line tweaks from the last month.

Passed STC

LLR: 3.07 (-2.94,2.94) [0.00,4.00]
Total: 15124 W: 2974 L: 2756 D: 9394

And LTC

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 21577 W: 3507 L: 3289 D: 14781

Bench: 8855226

Resolves #464
2015-10-20 19:49:01 -07:00
mbootsector ecc5ff6693 Lazy SMP
Start all threads searching on root position and
use only the shared TT table as synching scheme.

It seems this scheme scales better than YBWC for
high number of threads.

Verified for nor regression at STC 3 threads
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40232 W: 6908 L: 7130 D: 26194

Verified for nor regression at LTC 3 threads
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 28186 W: 3908 L: 3798 D: 20480

Verified for nor regression at STC 7 threads
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 3607 W: 674 L: 526 D: 2407

Verified for nor regression at LTC 7 threads
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 4235 W: 671 L: 528 D: 3036

Tested with fixed games at LTC with 20 threads
ELO: 44.75 +-7.6 (95%) LOS: 100.0%
Total: 2069 W: 407 L: 142 D: 1520

Tested with fixed games at XLTC (120secs) with 20 threads
ELO: 28.01 +-6.7 (95%) LOS: 100.0%
Total: 2275 W: 349 L: 166 D: 1760

Original patch of mbootsector, with additional work
from Ivan Ivec (log formula), Joerg Oster (id loop
simplification) and Marco Costalba (assorted formatting
and rework).

Bench: 8116244
2015-10-20 06:58:08 +02:00
Stéphane Nicolet 7ea5659c5f Asymmetry bonus for the attacking side
Use asymmetry in the position (king separation, pawn structure) to
compute an "initiative bonus" for the attacking side.

Passed STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 14563 W: 2826 L: 2636 D: 9101

And LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 14363 W: 2317 L: 2141 D: 9905

Bench: 8116244

Resolves #462
2015-10-15 21:27:52 -07:00
VoyagerOne 8fd34d7763 Combination of two ideas:
Apply bonus for the prior CMH that caused a fail low.

Balance Stats: CMH and History bonuses are updated differently.
This eliminates the "fudge" factor weight when scoring moves. Also
eliminated discontinuity in the gravity history stat formula. (i.e. stat
scores will no longer inverse when depth exceeds 22)

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 21802 W: 4107 L: 3887 D: 13808

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 46036 W: 7046 L: 6756 D: 32234

Bench: 7677367
2015-10-12 14:00:54 -07:00
Jonathan Calovski 55b46ffa90 Retire rook contact checks
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 34114 W: 6363 L: 6265 D: 21486

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61776 W: 9349 L: 9289 D: 43138

LTC (after rebasing):
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 15261 W: 2343 L: 2214 D: 10704

Bench: 7523382

Resolves #442
2015-10-07 19:43:20 +01:00
Marco Costalba 08d6465d90 Travis CI: add gcc 4.8 for osx
This setup was still missing.

Suggested by Stéphane Nicolet.

No functional change.
2015-10-07 20:33:37 +02:00
Marco Costalba 5f4d9309c8 Travis CI: add clang and osx
Extend builds to clang and osx platforms.

And check bench numbers.

No functional change.
2015-10-06 12:35:42 +02:00
Stefano80 436c303731 Tuning of assorted values
Passed STC

LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 45401 W: 8590 L: 8274 D: 28537

Passed LTC

LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 36089 W: 5589 L: 5331 D: 25169

Bench: 8397672

Resolves #445
2015-10-05 19:49:52 -07:00
Alain SAVARD 540b49a152 Remove queen threat evaluation
Threats by queen seem to be worthless.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 13627 W: 2607 L: 2473 D: 8547

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19146 W: 2950 L: 2827 D: 13369

Bench: 8222484

Resolves #439
2015-10-06 03:35:17 +01:00
Marco Costalba 2e45447957 Add Trevis CI support
Add Travis CI support to GitHub repo.

After every push to master, Travis will build
the sources directly from GitHub repo according
to .travis.yml and verify everything is ok.

No functional change.
2015-10-05 12:39:34 +02:00
Marco Costalba dc3508d157 Fix a comment in TTEntry::save
Comment was slightly incorrect.

No functional change.
2015-10-05 09:16:16 +02:00
Marco Costalba ca38358574 Run PVS-STUDIO analyzer
Fix issues after a run of PVS-STUDIO analyzer.
Mainly false positives but warnings are anyhow
useful to point out not very readable code.

Noteworthy is the memset() one, where PVS prefers ss-2
instead of stack. This is because memeset() could
be optimized away by the compiler when using 'stack',
due to stack being a local variable no more used after
memset. This should normally not happen, but when
it happens it leads to very sublte and difficult
to find bug, so better to be safe than sorry.

No functional change.
2015-10-05 09:13:33 +02:00
Stefan Geschwentner 83e19fbed5 File based passed pawn bonus
Add file based bonus for passed pawns. Values tuned by SPSA.

STC:
LLR: 3.33 (-2.94,2.94) [0.00,5.00]
Total: 36889 W: 6805 L: 6507 D: 23577

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 32301 W: 5101 L: 4858 D: 22342

Bench: 8073614

Resolves #436
2015-10-03 03:50:06 -07:00
Jonathan Calovski 9f5b31c21d Bonus for checking moves
STC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 14531 W: 2765 L: 2576 D: 9190

LTC:
LLR: 3.20 (-2.94,2.94) [0.00,5.00]
Total: 52518 W: 8107 L: 7782 D: 36629

Bench: 7556477

Resolves #435
2015-10-03 03:39:21 -07:00
Marco Costalba 3c0fe1d9b2 Rework lock protecting
When changing 'search' and 'splitPointsSize' we have to
use thread locks, not split point ones, because can_join()
is called under the formers.

Verified succesfully with 24 hours toruture tests with 20
cores machine by Louis Zulli: it does not hangs.

Verifyed for no regressions with STC, 7 threads:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 52804 W: 8159 L: 8087 D: 36558

No functional change.
2015-09-30 10:47:20 +02:00
Jonathan Calovski 77b4f4c2e7 Refine ranks and increase resulting bonus.
STC:
LLR: 2.94 (-2.94,2.94) [0.00,4.00]
Total: 272379 W: 51773 L: 50658 D: 169948

LTC:
LLR: 3.06 (-2.94,2.94) [0.00,4.00]
Total: 41504 W: 6555 L: 6273 D: 28676

bench: 7658406

Resolves #430
2015-09-19 09:14:35 -07:00
mstembera 68fbb1e052 Reduce writes in TT::probe().
Only refresh TT entry when it's really necessary.
This should give a small speed boost for some machines.
And it's a risk-free change.

No functional change.

Resolves #429
2015-09-17 17:33:40 -07:00
mstembera 01fab4d432 Remove unnecessary generation check in TT save
Checking for generation is unnecessary because if the key matches then the entry was probed and refreshed earlier.

STC 2MB
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 57391 W: 10671 L: 10613 D: 36107
http://tests.stockfishchess.org/tests/view/55ef59fa0ebc5976a2d6da5d

LTC 8MB
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 60732 W: 9260 L: 9199 D: 42273
http://tests.stockfishchess.org/tests/view/55ef8fe60ebc5976a2d6da6b

STC 16MB
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 23443 W: 4369 L: 4293 D: 14781
http://tests.stockfishchess.org/tests/view/55ef8fe60ebc5976a2d6da6b

No functional change

Resolves #427
2015-09-17 17:13:45 -07:00
Stefan Geschwentner 660c38f781 Scales the endgame score by the number of pawns.
Credits goes also to Stephane Nicolet for his great idea of scaling by pawns.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 9994 W: 1929 L: 1760 D: 6305

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 11240 W: 1789 L: 1626 D: 7825

bench 7298564

Resolves #423
2015-09-15 05:55:09 -07:00
Joona Kiiski 613dc66c12 Careful SMP locking - Fix very occasional hangs
Louis Zulli reported that Stockfish suffers from very occasional hangs with his 20 cores machine.

Careful SMP debugging revealed that this was caused by "a ghost split point slave", where thread
was marked as a split point slave, but wasn't actually working on it.

The only logical explanation for this was double booking, where due to SMP race, the same thread
is booked for two different split points simultaneously.

Due to very intermittent nature of the problem, we can't say exactly how this happens.

The current handling of Thread specific variables is risky though. Volatile variables are in some
cases changed without spinlock being hold. In this case standard doesn't give us any kind of
guarantees about how the updated values are propagated to other threads.

We resolve the situation by enforcing very strict locking rules:
- Values for key thread variables (splitPointsSize, activeSplitPoint, searching)
can only be changed when the thread specific spinlock is held.
- Structural changes for splitPoints[] are only allowed when the thread specific spinlock is held.
- Thread booking decisions (per split point) can only be done when the thread specific spinlock is held.

With these changes hangs didn't occur anymore during 2 days torture testing on Zulli's machine.

We probably have a slight performance penalty in SMP mode due to more locking.

STC (7 threads):
ELO: -1.00 +-2.2 (95%) LOS: 18.4%
Total: 30000 W: 4538 L: 4624 D: 20838

However stability is worth more than 1-2 ELO points in this case.

No functional change

Resolves #422
2015-09-10 19:15:43 +01:00
mstembera 3e2591d83c Minor clean up of some function parameters
No function change

Resolves #416
2015-09-07 20:17:39 +01:00
mstembera 46b5a5f0db Fix syzygy en passant issue
v = value without ep capture being considered
v1 = value of the ep capture

The correct logic is:
if without e.p. capture we are losing, and the value of e.p is either draw, or win or "loss, but 50 move rule saves us", then we should use the value of ep capture.

Credit and thanks to syzygy1 and lantonov !

No functional change (except with syzygy bases)

Resolves #415
Resolves #394
2015-09-06 22:19:33 +01:00
gguliash 84a641b8bb A small code simplification
No functional change

Resolves #411
2015-08-30 19:58:32 +01:00
Joona Kiiski 7f300a7698 History gravity
Instead of using hard coded Min and Max values for history,
always adjust the old value slightly downwards before adding a new value.

The adjustment acts like gravity that prevents the value escaping too
far from zero.

Bench: 8020484

Resolves #407
2015-08-29 15:09:00 +01:00
Marco Costalba 087b638f6c Reformat trace code
Apart from usual renaiming, take advantage of
C++11 function template default parmeter to
get rid of Eval trampoline functions.

Some triviality fixes while there.

No functional change.
2015-08-29 08:28:01 +02:00
lucasart 7ad85fca6d Prune castling moves
Align the behaviour with reductions. Initially castling moves had to be
treated differently, because the SEE did not handle them correctly. But now it
does.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 83750 W: 15722 L: 15711 D: 52317

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 97183 W: 15120 L: 15115 D: 66948

bench 7759837

Resolves #403
2015-08-28 20:33:12 +01:00
mstembera bf6bc1b37b Better document entry age calculation used in TT replace.
No functional change.

Resolves #401
Resolves #400
2015-08-20 20:27:06 +01:00
lucasart c052e03426 Retire dangerous flag
Replace by its value where it is used. Code is more clear that
way.

No functional change.

Resolves #402
2015-08-20 20:18:23 +01:00
Alain SAVARD 69a1a808c8 Retire PawnSafePush bonus
PawnSafePush, with the value S(5,5) proved not "necessary"
possibly due to recent changes to MobilityArea and other changes to Connected bonus.

STC:
LLR: 3.22 (-2.94,2.94) [-3.00,1.00]
Total: 98528 W: 18757 L: 18759 D: 61012

LTC:
LLR: 5.30 (-2.94,2.94) [-3.00,1.00]
Total: 204194 W: 31698 L: 31734 D: 140762

Bench: 7620871

Resolves #396
2015-08-17 19:32:43 +01:00
Marco Costalba 600234f2e2 Reformat PassedPawnsBonus
Align to SF coding standards.

No functional change.
2015-08-15 16:04:16 +02:00
mstembera 9dbb3ae8b2 TT entry value based on depth and relative age
Calculate TT replace value as depth minus eight times relative age.

STC 2MB
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 45258 W: 8595 L: 8279 D: 28384

LTC 8MB
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 56401 W: 8809 L: 8489 D: 39103

STC 16MB
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 34764 W: 6565 L: 6529 D: 21670

Bench: 9069474

Resolves #395
2015-08-15 12:24:48 +01:00
DiscanX f4ace94f91 Tuned values for mid and end game passed pawns.
STC :
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 22691 W: 4468 L: 4228 D: 13995

LTC :
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 13620 W: 2216 L: 2023 D: 9381

Bench: 8384669

Resolves #391
2015-08-10 01:26:59 +08:00
mstembera 1aae35b8b7 Revert TT replacement strategy changes (#380)
It could cause problems with high depths and long time controls

Bench: 8626315

Resolves #390
2015-08-08 17:49:53 +01:00
Marco Costalba e6310b3469 Rename Position::list
Use Position::square and Position::squares instead.

This allow us to remove king_square(), simplify
endgames and to have more naming uniformity.

Moreover, this is a prerequisite step in case
in the future we decide to retire piece lists
altoghter and use pop_lsb() to loop across
pieces and serialize the moves. In this way
we just need to change definition of Position::square
to something like:

template<PieceType Pt> inline
Square Position::square(Color c) const {
  return lsb(byColorBB[c]);
}

No functional change.
2015-08-04 09:51:06 +02:00
Marco Costalba 68d61b80c6 Simplify IID depth formula
Restore original formula messed up during
half-ply removal.

STC
LLR: 4.11 (-2.94,2.94) [-3.00,1.00]
Total: 21349 W: 4091 L: 3909 D: 13349

LTC
LLR: 5.42 (-2.94,2.94) [-3.00,1.00]
Total: 52819 W: 8321 L: 8122 D: 36376

bench: 8040572
2015-07-30 09:54:25 +02:00
VoyagerOne a2410227cc PV refutation penalty
Extra penalty for PV move in previous ply when it gets refuted.

STC:
LLR: 4.49 (-2.94,2.94) [-1.50,4.50]
Total: 41094 W: 7889 L: 7620 D: 25585

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 12304 W: 1967 L: 1811 D: 8526

Bench: 8373608

Resolves #386
2015-07-30 00:51:49 +08:00
Alain SAVARD 4a6d59c6c7 MobilityArea (simplified)
Based off of Pull request #383:

Include squares occupied by some pawns in the MobilityArea
a) not blocked
b) on rank 4 and above
c) or captures

Passed STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8157 W: 1644 L: 1516 D: 4997

And LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 26086 W: 4274 L: 4051 D: 17761

-----------

Then, a simplification test failed, trying to remove b and c)
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 6048 W: 1117 L: 1288 D: 3643

Another simplification test, was run to remove just (c)
Passed STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 28073 W: 5364 L: 5255 D: 17454

And LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 34652 W: 5448 L: 5348 D: 23856

A parameter tweak test showed that changing b) for "on rank 3 and above"
does not work
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 5233 W: 937 L: 1077 D: 3219

Finally, a small rewrite, and we have this version

Include squares occupied by some pawns in the MobilityArea which are
a) not blocked
b) on rank 4 and above

Bench: 8977899

Resolves #385
2015-07-30 00:47:51 +08:00
mstembera 51330414c4 Tuned version of TT replacement policy
If the used multiplier of 8 was any number larger than DEPTH_MAX
this would be a non functional patch.

STC 2MB
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16353 W: 3216 L: 3066 D: 10071
http://tests.stockfishchess.org/tests/view/55a6d0630ebc590abbe1babd

LTC 8MB
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 134618 W: 21276 L: 20716 D: 92626
http://tests.stockfishchess.org/tests/view/55a74d760ebc590abbe1bad6

STC 16MB
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 22549 W: 4257 L: 4178 D: 14114
http://tests.stockfishchess.org/tests/view/55a9a2f90ebc590abbe1bb16

Bench: 7372460

Resolves #380
2015-07-25 00:33:53 +08:00
Marco Costalba 42b77ea7ee Tidy up in movepick.cpp
Some formattng after recent changes.

No functional change.
2015-07-19 12:38:33 +02:00
Marco Costalba 4095ff0ee5 Fix formatting of previous patch
No functional change.
2015-07-16 08:17:40 +02:00
VoyagerOne 1d977aa79e LMR Simplification: Remove countermove condition
Removed countermove condition for decreasing reduction.

LTC:
LLR: 3.01 (-2.94,2.94) [-3.00,1.00]
Total: 32410 W: 5092 L: 4986 D: 22332

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 24450 W: 4632 L: 4517 D: 15301

Bench: 6943812

Resolves #378
2015-07-15 20:21:14 +01:00
mstembera d8aeaab79c Consistent TT replace policy
This fixes an inconsistency bug where TT entries were valued differently
depending on which pointer they were accessed through.

STC 2MB
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 11301 W: 2176 L: 2038 D: 7087

LTC 8MB
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17732 W: 2870 L: 2745 D: 12117

STC 16MB
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 17401 W: 3324 L: 3227 D: 10850

Bench: 8248164

Resolves #377
2015-07-15 20:17:16 +01:00
joergoster 3054db989d Use distance<file>() function in endgame.cpp
This one occurance of distance function was most likely overlooked.

No functional change.

Resolves #376
2015-07-12 20:48:50 +01:00
Marco Costalba b2edac7075 Remove useless razoring condition
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 20626 W: 3977 L: 3855 D: 12794

LTC:
LLR: 3.10 (-2.94,2.94) [-3.00,1.00]
Total: 87334 W: 13675 L: 13648 D: 60011

Retire also the now unused pawn_on_7th() helper.

bench: 8248166
2015-07-09 08:13:30 +02:00
VoyagerOne aa0dc16b75 CMH Fix: Exclude captures for TT move refutation penalty
This will make sure we store only quiet moves for TT Penalty.

STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 34748 W: 6617 L: 6420 D: 21711

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19975 W: 3259 L: 3137 D: 13579

Bench: 8063826

Resolves #373
2015-07-04 14:22:41 +01:00
Marco Costalba 112607bf49 Correctly check for no-makefile compile
Under Windows with MSVC we use the IDE to compile,
in this case we infer some compiler flags usually
set by Makefile.

The condition to check this was wrong, namely when compiling
with mingw under Windows 64 bit we always set IS_64BIT and
USE_BSFQ even if compiled with ARCH=x86-32 (this is how I
found it).

Small code style touches while there.

No functional change.
2015-06-28 10:56:42 +02:00
VoyagerOne 7e9f776fca LMR Tweak: Decrease reduction if cmh>0 && history>0.
STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 9627 W: 1879 L: 1748 D: 6000

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 21543 W: 3433 L: 3234 D: 14876

Bench: 8646407

Resolves #370 #371
2015-06-27 09:45:59 +01:00
Marco Costalba 3f1bc9e013 Fix compile on icc
Error is:

  a value of type "int" cannot be assigned to an entity of type "Value"

Also retire the now unused squares_of_color() function.

No functional change.
2015-06-25 05:29:56 +02:00
joergoster e8cc5c9968 Small coding style fix for Outpost array
No functional change

Resolves #367
2015-06-17 15:02:57 +01:00
lucasart 00563be5cc Retire -Wno-long-long
long long is part of the C++11 standard.

No functional change.

Resolves #364
2015-06-13 16:59:50 +01:00
lucasart ad87d707ff Simplify outpost evaluation
STC

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 114149 W: 21986 L: 22032 D: 70131

LTC

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 52232 W: 8468 L: 8396 D: 35368

Bench: 6716940

No functional change

Resolves #363
2015-06-07 09:49:01 +01:00
Vince Negri 7b8ffe0f5d Remove intermediate re-search in LMR
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 20149 W: 3830 L: 3707 D: 12612

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 45384 W: 7089 L: 7006 D: 31289

Bench: 8110365

Resolves #361
2015-06-05 21:03:10 +01:00
Marco Costalba c1d3bd2dba Rename stages
Hopefully more clear.

No functional change.
2015-06-04 15:51:45 +02:00
lucasart 5ebdf2f8c8 Tune pawn shelter/storm
STC
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 31679 W: 6183 L: 5912 D: 19584

LTC
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 43550 W: 6885 L: 6602 D: 30063

Bench: 9219343

Resolves #360
2015-06-02 12:09:19 -07:00
Stéphane Nicolet 4502a1934f Tweak backward pawns definition
Advanced pawns cannot be backward. Also lower the backward penalty in
endgame.

Passed STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 18534 W: 3588 L: 3433 D: 11513

and LTC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 21319 W: 3415 L: 3217 D: 14687

Bench: 7271152

Resolves #359
2015-06-02 12:07:02 -07:00
Marco Costalba 2795aedbc3 Checking for rook color when setting castling
In Chess960 we can have legal positions with
opponent rook in A or H file and with castling
available, for instance:

4k3/pppppppp/8/8/8/8/PPPPPPPP/rR2K3 w Q - 0 1

In those cases we pick up the wrong rook when
setting castling.

Fix it by checking the color of the rook.

Bug reported by Matthew Lai.

No functional change.
2015-05-29 05:38:40 +02:00
lucasart 8e95c39bfa Simplify backward pawn scoring
STC

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 52322 W: 10011 L: 9945 D: 32366

LTC

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14143 W: 2334 L: 2203 D: 9606

Bench: 7976423

Resolves #354
2015-05-27 20:05:20 +01:00
Joona Kiiski 80a759d544 Resolve build failure for Mac
Remove '-Wl' switch from gcc arguments when compiling for Mac

No functional change

Resolves #353
2015-05-24 13:07:15 +01:00
lucasart 2e0ec4ad6f Fix merge error for Tuned PSQT
Fall-out from 411e704f

Bench: 7907776

Resolves #352
2015-05-21 19:56:01 +01:00
Stefan Geschwentner e14046517e Remove Gain Stats
Additionally in futility pruning the margin is raised for compensation.

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 48481 W: 9229 L: 9156 D: 30096

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32058 W: 5134 L: 5031 D: 21893

Bench: 8098149

Resolves #350
2015-05-18 13:59:30 -07:00
lucasart 411e704fdf Tuned PSQT
STC:
LLR: 3.11 (-2.94,2.94) [-0.50,4.50]
Total: 58764 W: 11530 L: 11185 D: 36049

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.50]
Total: 282710 W: 46339 L: 45209 D: 191162

Bench: 8512947

Resolves #349
2015-05-18 13:54:45 -07:00
Marco Costalba ee0371f86e Cleanup work in misc.cpp
Also some code style tidy up of latest patches.

Also renamed checkSq -> checkSquares because it
is a bitboard and not a square.

No functional change.
2015-05-10 09:42:26 +02:00
mstembera eaeb63f1d0 Smart TT save
Don't overwrite more valuable data with less valuable data

STC 2MB
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 21132 W: 4108 L: 3946 D: 13078
http://tests.stockfishchess.org/tests/view/5547d59f0ebc5940ca5d6883

LTC 8MB
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 13381 W: 2149 L: 1987 D: 9245
http://tests.stockfishchess.org/tests/view/5549b5a80ebc5940ca5d68b9

STC 16MB regression w/ zero effective hash pressure
LLR: 2.96 (-2.94,2.94) [-5.00,0.00]
Total: 18944 W: 3607 L: 3564 D: 11773
http://tests.stockfishchess.org/tests/view/554b0fda0ebc5940ca5d68ea

Bench: 8787152

Resolves #347
2015-05-09 17:43:57 +01:00
lucasart 92dbee7875 Edge distance
Instead of crafting a clever formula to calculate the array offset, simply use a
3 dimensional array. Remove the comment while at it, because now the code is
self-documenting.

No functional change.

Resolves #344
2015-05-09 14:06:19 +01:00
lucasart 2e86d1febc Restore deterministic search state
Introduce helper function Search::reset() which clears all kind of search
memory, in order to restore a deterministic search state.

Generalize TT.clear() into Search::reset() for the following use cases:
- bench: needed to guarantee deterministic bench (ie. if you call bench from
interactive command line twice in a row you get the same value).
- Clear Hash: restore clean search state, which is the purpose of this button.
- ucinewgame: ditto.

No functional change.

Resolves #346
2015-05-07 14:20:32 -07:00
lucasart 946fa47625 Never clear stats
Based on an idea and patch by VoyagerOne.

Small simplification, but was tedted for an ELO gain anyway.

STC:
LLR: 2.95 (-2.94,2.94) [-1.00,4.00]
Total: 5375 W: 1119 L: 977 D: 3279

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 17893 W: 2984 L: 2792 D: 12117

bench 8322847
2015-05-06 19:17:33 +08:00
Marco Costalba 7231b18af1 Halve PSQT row data
Use symmetry along vertical middle axis of the board
to reduce the number of parameters.

For instance psqt value of SQ_A5 == SQ_A4 and value of
SQ_F8 == SQ_F1.

This is always true, at least until now nobody came in
with an asymmetric psqt table that worked.

Original patch by Lucas.

No functional change.
2015-05-03 20:07:52 +02:00
Marco Costalba 578b21bbee Split PSQT init from Position init
Easier for tuning psq tables:

TUNE(myParameters, PSQT::init);

Also move PSQT code in a new *.cpp file, and retire the
old and hacky psqtab.h that required to be included only
once to work correctly, this is not idiomatic for a header
file.

Give wide visibility to psq tables (previously visible only
in position.cpp), this will easy the use of psq tables outside
Position, for instance in move ordering.

Finally trivial code style fixes of the latest patches.

Original patch of Lucas Braesch.

No functional change.
2015-05-03 20:07:15 +02:00
Stefano80 59f64fda4f Improve ordering of good captures using rank term
Rank based term improved approximation of pos.see() for scoring good
captures.

STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 4632 W: 945 L: 827 D: 2860

LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 25770 W: 4184 L: 3964 D: 17622

Bench: 7593704

Resolves #342
2015-05-03 10:19:34 -07:00
Stefano80 cc54a91ca5 Replace MVV/LVA by MVV for good captures
Passed STC

LLR: 3.71 (-2.94,2.94) [-3.00,1.00]
Total: 64363 W: 12299 L: 12214 D: 39850

and LTC

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 69976 W: 11056 L: 11011 D: 47909

Bench: 8012532

Resolves #340
2015-04-28 22:07:48 +01:00
VoyagerOne b7063ef65b Change extra ply LMR condition to: cmh <= 0 && hist < 0
Extra ply LMR condition is now cmh <= 0 && h < 0
Instead of cmh + h < 0

STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 55210 W: 10812 L: 10557 D: 33841

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 13212 W: 2239 L: 2045 D: 8928

Bench: 8420865

Resolves #339
2015-04-26 20:12:25 +01:00
Alain SAVARD 078625e333 Exclude queen from Rook Contact Check computation
In ei.attackedBy, Queen does not x-ray through Rook, but the Rook does
X-ray through the Queen.

So most of the rook contact checks supported by queen are, in fact,
Queen Contact Checks and they are already scored separately.

Bench: 7762189

Resolves #338
2015-04-18 22:17:46 +01:00
lucasart 6c040c821a Retire FORCE_INLINE
No speed regression on my machine (i7-3770k, gcc 4.9.1, linux 3.16):

        stat        test     master   diff
        mean   2,482,415  2,474,987  7,906
        stdev      4,603      5,644  2,497

        speedup        0.32%
        P(speedup>0)  100.0%

Fishtest 9+0.03:

ELO: 0.26 +-1.8 (95%) LOS: 61.2%
Total: 60000 W: 12437 L: 12392 D: 35171

No functional change.

Resolves #334
2015-04-15 21:21:45 +01:00
VoyagerOne 20e92895af Removed extra condition (history < 0) in LMR to help sync up with move ordering.
LMR condition is now cmh+history<0
Instead of history<0 OR cmh+history<0

STC:
LLR: 2.96 (-2.94,2.94) [-3.00, 1.00]
Total: 26446 W: 5092 L: 4980 D: 16374

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00, 1.00]
Total: 14129 W: 2340 L: 2209 D: 9580

Bench: 7815183

Resolves #331
2015-04-12 20:05:59 +01:00
Marco Costalba fb03188fc7 Assorted cleanup of last patches
No functional change.
2015-04-11 23:24:43 +02:00
Marco Costalba a66c73deef Allow Position::init() to be called more than once
Currently Zobrist::castling[] are not properly zeroed
and rely on the compiler to do this at startup, but this
makes Position::init() to set different values every time
it is called!

This is a bit odd, and although not impacting normal usage,
can yield to subtle misbehaviour, very difficult to track
down, in case we happen to call it more than once for some
reason. I found this while developing tuning support and
it took me a while to track it down.

So properly init Zobrist::castling[]

No functional change.

Resolves #329
2015-04-10 20:39:15 +01:00
Stefan Geschwentner 27efc5ac99 Update stats at pv nodes
If a quiet best move is found at a pv node then always update stats.

STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 41485 W: 8047 L: 7830 D: 25608

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 14351 W: 2420 L: 2250 D: 9681

Bench: 6985247

Resolves #330
2015-04-10 20:34:55 +01:00
mstembera 36f2133df3 New formula for quiet move scoring: 3 * cmh + 1 * hist
STC

LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 45363 W: 8759 L: 8532 D: 28072

LTC

LLR: 3.51 (-2.94,2.94) [0.00,4.00]
Total: 125092 W: 20032 L: 19468 D: 85592

Bench: 7058819

Resolves #328
2015-04-10 20:14:11 +01:00
Stefan Geschwentner ef4d89c9bd update stats also in check
Update stats also if in check (drop condition).

STC:
LLR: 3.22 (-2.94,2.94) [-3.00,1.00]
Total: 87472 W: 16929 L: 16913 D: 53630

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 39971 W: 6436 L: 6345 D: 27190

Bench: 7086031

Resolves #327
2015-04-09 20:41:08 +01:00
lucasart aaf17326e2 Prune evasions when we can castle
A minor simplification.

STC:

LLR: 2.95 (-2.94,2.94) [-3.50,0.50]
Total: 67877 W: 12882 L: 12904 D: 42091

STC:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 20677 W: 4023 L: 3901 D: 12753

LTC:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 12221 W: 2022 L: 1888 D: 8311

Bench: 7911336

Resolves #326
2015-04-09 20:34:06 +01:00
Stéphane Nicolet 2ca142a5b4 Use minimumSplitDepth = 5
Using minimumSplitDepth = 5 seems to be the best compromise in the
current SMP implementation

STC, 11 threads:

ELO: 14.87 +-4.1 (95%) LOS: 100.0%
Total: 8509 W: 1497 L: 1133 D: 5879

STC, 4 threads:

ELO: 0.30 +-2.8 (95%) LOS: 58.2%
Total: 20000 W: 3365 L: 3348 D: 13287

STC, 2 threads:

ELO: -1.02 +-2.0 (95%) LOS: 16.4%
Total: 40000 W: 7087 L: 7204 D: 25709

Resolves #324
2015-04-09 20:32:36 +01:00
Marco Costalba ffbaa1c5a8 Fix MSVC warning from previous patch
No functional change.
2015-04-03 19:29:29 +02:00
Marco Costalba 57dfe41ddb Fix elapsed()
Messed up during merge.

No functional change.
2015-04-03 04:53:53 +02:00
Marco Costalba 926f215061 Add support for playing in 'nodes as time' mode
When running more games in parallel, or simply when running a game
with a background process, due to how OS scheduling works, there is no
guarantee that the CPU resources allocated evenly between the two
players. This introduces noise in the result that leads to unreliable
result and in the worst cases can even invalidate the result. For
instance in SF test framework we avoid running from clouds virtual
machines because are a known source of very unstable CPU speed.

To overcome this issue, without requiring changes to the GUI, the idea
is to use searched nodes instead of time, and to convert time to
available nodes upfront, at the beginning of the game.

When nodestime UCI option is set at a given nodes per milliseconds
(npmsec), at the beginning of the game (and only once), the engine
reads the available time to think, sent by the GUI with 'go wtime x'
UCI command. Then it translates time in available nodes (nodes =
npmsec * x), then feeds available nodes instead of time to the time
management logic and starts the search. During the search the engine
checks the searched nodes against the available ones in such a way
that all the time management logic still fully applies, and the game
mimics a real one played on real time. When the search finishes,
before returning best move, the total available nodes are updated,
subtracting the real searched nodes. After the first move, the time
information sent by the GUI is ignored, and the engine fully relies on
the updated total available nodes to feed time management.

To avoid time losses, the speed of the engine (npms) must be set to a
value lower than real speed so that if the real TC is for instance 30
secs, and npms is half of the real speed, the game will last on
average 15 secs, so much less than the TC limit, providing for a
safety 'time buffer'.

There are 2 main limitations with this mode.

1. Engine speed should be the same for both players, and this limits
the approach to mainly parameter tuning patches.

2. Because npms is fixed while, in real engines, the speed increases
toward endgame, this introduces an artifact that is equivalent to an
altered time management. Namely it is like the time management gives
less available time than what should be in standard case.

May be the second limitation could be mitigated in a future with a
smarter 'dynamic npms' approach.

Tests shows that the standard deviation of the results with 'nodestime'
is lower than in standard TC, as is expected because now all the introduced
noise due the random speed variability of the engines during the game is
fully removed.

Original NIT idea by Michael Hoffman that shows how to play in NIT mode
without requiring changes to the GUI. This implementation goes a bit
further, the key difference is that we read TC from GUI only once upfront
instead of re-reading after every move as in Michael's implementation.

No functional change.
2015-04-03 04:40:55 +02:00
Marco Costalba df722521ba Rename of TimeMgr and friends
More natural naming IMO.

No functional change.
2015-04-03 04:19:29 +02:00
Marco Costalba 5d1b92e8f9 Introduce elapsed_time()
And reformat a bit time manager code.

Note that now we set starting search time in think() and
no more in ThreadPool::start_thinking(), the added delay
is less than 1 msec, so below timer resolution (5msec) and
should not affect time lossses ratio.

No functional change.
2015-04-03 04:19:26 +02:00
mstembera 6661a31541 Simplification to use only one counter move.
STC http://tests.stockfishchess.org/tests/view/5518dca30ebc5902160ec5d0
LLR: 2.95 (-2.94,2.94) [-3.50,0.50]
Total: 18868 W: 3638 L: 3530 D: 11700

LTC http://tests.stockfishchess.org/tests/view/5518f7ed0ebc5902160ec5d4
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 69767 W: 11019 L: 10973 D: 47775

Extracted from http://tests.stockfishchess.org/tests/view/5511028a0ebc5902160ec40b

Original patch by hxim.  All credit goes to him.

Bench: 7664249

Resolves #320
2015-04-03 01:16:15 +08:00
Marco Costalba 6c42575208 Assorted code style of latest commits
No functional chnage.
2015-03-29 10:16:10 +02:00
Joona Kiiski 61a6e46664 Fix indentations for hanging pawns code
No functional change
2015-03-29 08:54:25 +01:00
lucasart 60beb18efc Remove some difficult to understand C++11 constructs
Code like this is more a case of showing off one's C++ knowledge, rather than
using it adequately, IMHO.

**First loop (std::generate)**

Iterators are inadequate here, because they lose the key information which is
idx. As a result, we need to carry a redundant idx variable, and increment it
along the way. Very clumsy.
Usage of std::generate and a lambda function only obfuscate the code, which is
merely a simple and stupid loop over the elements of a vector.

**Second loop (std::accumulate)**

This code is thoroughlly incomprehensible. Restore the original, which was much
simpler to understand.

**Third loop (range based loop)**

Again, a range based loop is inadequate, because we lose idx! To resolve this
artificially created problem, the data model was made redundant (idx is a data
member of db[] elements!?), which is ugly and unjustified. A simple and stupid
for loop with idx does the job much better.

No functional change.

Resolves #313
2015-03-29 08:40:24 +01:00
Alain SAVARD 8463fa479e PSV3_1
Small speed-up in pawn.cpp
Results for 10 tests for each version:

Base      Test      Diff
Mean    1435636   1445238   -9602
StDev   22576     23189     1848

p-value: 1
speedup: 0.007

No functional change

Resolves #295
2015-03-29 07:30:46 +08:00
Ajith 63500274db Give a reduced bonus for threats by hanging pawns
Passed STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 105539 W: 20389 L: 20001 D: 65149

and LTC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 9629 W: 1577 L: 1432 D: 6620

Bench: 7658627

Resolves #317
2015-03-29 07:28:47 +08:00
mbootsector 1d5eaba573 Retire follow-up move heuristic
STC: http://tests.stockfishchess.org/tests/view/5501d0f30ebc5902160ec0fd
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 34891 W: 6904 L: 6808 D: 21179

LTC: http://tests.stockfishchess.org/tests/view/550328540ebc5902160ec133
LLR: 3.10 (-2.94,2.94) [-3.00,1.00]
Total: 182653 W: 29866 L: 29993 D: 122794

Bench: 8396161

Resolves #310
2015-03-28 22:12:06 +00:00
VoyagerOne ac8e6ff000 Use CounterMoveHistory when calculating LMR for cut nodes
If the sum of CounterMoveHistory heuristic and History heuristic is below zero,
then reduce an extra ply in cut nodes

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 6479 W: 1099 L: 967 D: 4413

Bench: 7773299

Resolves #315
2015-03-28 21:15:49 +00:00
Marco Costalba 3a6753328c Clean up previous patch
No functional change.
2015-03-25 07:05:27 +01:00
joergoster 50182f1100 Tuned values for the pawn piece square table
Quick measure at very fast tc:
ELO: 4.77 +-2.2 (95%) LOS: 100.0%
Total: 40124 W: 8711 L: 8160 D: 23253

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 52284 W: 8880 L: 8559 D: 34845

Bench: 8865736

Resolves #311
2015-03-25 07:39:44 +08:00
VoyagerOne e51965aa57 Introduce a new counter move history penalty
Extra penalty for TT move in previous ply when it gets refuted

STC:

LLR: 2.94 (-2.94,2.94) [-1.50,4.50]
Total: 31303 W: 6216 L: 6025 D: 19062

LTC:

LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 6950 W: 1189 L: 1054 D: 4707

Bench: 8191926

Resolves #309
2015-03-24 23:04:08 +00:00
Joona Kiiski a7381d5e81 Fully yielding locks, no spinning
7 threads:

ELO: 2.00 +-2.7 (95%) LOS: 92.4%
Total: 20000 W: 3276 L: 3161 D: 13563

There is no functional change in single thread mode

Resolves #304
2015-03-24 21:34:19 +00:00
Marco Costalba dc3a5f791e Allow Bitbases::init() to be called more than once
Currently if we call it more than once, we crash.

This is not a real problem, because this function is
indeed called just once. Nevertheless with this small fix,
that gets rid of a hidden 'static' variable, we cleanly
resolve the issue.

While there, fix also ThreadPool::exit to return in a
consistent state. Now all the init() functions but
UCI::init() are reentrant and can be called multiple
times.

No functional change.
2015-03-23 17:14:31 +01:00
Marco Costalba 35a082064f Double magics generation speed
Profiling shows that resetting attacks table after
a failed candidate magic attempt is the biggest
time consumer, so rewrite the logic avoiding the
memset()

Magics init for rook+bishop goes from 200msecs to
under 100msec.

No functional change.
2015-03-23 17:10:38 +01:00
Marco Costalba be77406a55 Get rid of nativeThread
No functional change.
2015-03-23 09:02:52 +01:00
Marco Costalba 26dabb1e6b Use only one ConditionVariable to sync UI
To sync UI with main thread it is enough a single
condition variable because here we have a single
producer / single consumer design pattern.

Two condition variables are strictly needed just for
many producers / many consumers case.

Note that this is possible because now we don't send to
sleep idle threads anymore while searching, so that now
only UI can wake up the main thread and we can use the
same ConditionVariable for both threads.

The natural consequence is to retire wait_for_think_finished()
and move all the logic under MainThread class, yielding the
rename of teh function to join()

No functional change.
2015-03-21 07:55:33 +01:00
lucasart 35b6079852 Fix comment
We always probe, but we do not prune at PV nodes.

No functional change.

Resolves #300
2015-03-20 22:40:03 +00:00
Marco Costalba ebf3735754 Retire ConditionVariable
Now that we use spinlocks everywhere and don't put
threads to sleep while idle, we can use the slower
(but no more in hot path) std::condition_variable_any
instead of our homwgrown ConditionVariable struct.

Verified fo rno regression at STC with 7 threads:
ELO: -0.66 +-2.7 (95%) LOS: 31.8%
Total: 20000 W: 3210 L: 3248 D: 13542

No functional change
2015-03-20 03:05:25 +01:00
joergoster 966bc477af Tuned mobility with another SPSA run
Further improved mobility values after another SPSA session, 50k
iterations.

Elo measure at very fast 9+0.05":
ELO: 3.40 +-2.2 (95%) LOS: 99.9%
Total: 40000 W: 8434 L: 8042 D: 23524

and LTC SPRT[0, 4]:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 11052 W: 1874 L: 1687 D: 7491

Bench: 8226843

Resolves #301
2015-03-20 08:08:21 +08:00
joergoster c6f987d1ad Fix the comment for Position::is_draw()
We no longer check for insufficient material.

No functional change

Resolves #299
2015-03-18 20:30:50 +00:00
lucasart d477a0d611 connected should be bool, not Bitboard
There's no reason to define it as a Bitboard, so for consistency, use bool.

This is even a speedup on my machine: i7-3770k, using gcc 4.9.1 (linux):

    stat        test     master    diff
    mean   2,341,338  2,327,998  13,134
    stdev     15,765     14,717   5,405

    speedup       0.56%
    P(speedup>0) 100.0%

No functional change.

Resolves #298
2015-03-18 20:21:41 +00:00
Marco Costalba 54889618c2 Reformat FastMove
Align to SF coding style.

Verified no regression:

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 55938 W: 10893 L: 10835 D: 34210

No functional change.
2015-03-18 08:12:59 +01:00
Marco Costalba 9a6cfee73b Simplify nosleep logic
Avoid redundant 'while' conditions. It is enough to
check them in the outer loop.

Quick tested for no regression 10K games at 4 threads
ELO: -1.32 +-3.9 (95%) LOS: 25.6%
Total: 10000 W: 1653 L: 1691 D: 6656

No functional change.
2015-03-18 08:01:50 +01:00
Marco Costalba 2e8552db76 Fix a bogus use of mutex
Spinlock must be used instead.

Tested for no regression at 15+0.05 th 4:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25928 W: 4303 L: 4190 D: 17435

No functional change.

Resolves #297
2015-03-17 08:19:29 +00:00
Marco Costalba a4b2eeea75 Re-arrange history update code
Unify the quites moves loop for both cases,
the compiler optimizes away the

 if (is_ok((ss-1)->currentMove))

inside loop, so that the result is same
speed as original.

No functional change.
2015-03-16 15:14:09 +01:00
Marco Costalba 13d4df95cd Use acquire() and release() for spinlocks
It is more idiomatick than lock() and unlock()

No functional change.
2015-03-16 08:14:08 +01:00
Vince Negri bae4679de2 Fix dependency generation for MacOSX
No functional change

Resolves #290
2015-03-15 20:31:43 +00:00
Joona Kiiski a3943b2aa7 Fix dependency generation for C++11
No functional change

Resolves #291
2015-03-15 20:26:08 +00:00
Joona Kiiski f04f50b368 Do not sleep, but yield
During the search, do not block on condition variable, but instead use std::this_thread::yield().

Clear gain with 16 threads. Again results vary highly depending on hardware, but on average it's a clear gain.

ELO: 12.17 +-4.3 (95%) LOS: 100.0%
Total: 7998 W: 1407 L: 1127 D: 5464

There is no functional change in single thread mode

Resolves #294
2015-03-15 19:45:30 +00:00
joergoster a4b98a052e New values for Mobility and Outposts.
Both are the result of a SPSA tuning session with a custom book, 50k iterations each.

After an additional tuning session of the mobility values, tuning the delta values, with following result.

40k games at 9+0.05:
ELO: 4.13 +-2.2 (95%) LOS: 100.0%
Total: 40000 W: 8581 L: 8106 D: 23313

and LTC
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 36518 W: 6049 L: 5782 D: 24687

Bench: 8567402

Resolves #284
2015-03-14 21:35:52 +00:00
Marco Costalba e5da0e4b79 Link with -static in mingw
Fixes reported startup error about missing libwinpthread-1.dll
when the dll is not in the path.

The current -static-xxxx flags, introduced with:

https://github.com/official-stockfish/Stockfish/commit/373503f4a9a990054b5

Only take in account standard libraries, but not thread
library.

No functional change.

Resolves #289
2015-03-14 19:23:54 +00:00
Joona Kiiski d71f707040 Introduce yielding spin locks
Idea and original implementation by Stephane Nicolet

7 threads 15+0.05
ELO: 3.54 +-2.9 (95%) LOS: 99.2%
Total: 17971 W: 2976 L: 2793 D: 12202

There is no functional change in single thread mode
2015-03-14 19:14:52 +00:00
Joona Kiiski 558b0c848c Remove check for gcc version from Makefile.
This check is obsolete.
very old gcc versions can't compile c++11 code.

No functional change

Resolves #285
2015-03-13 20:29:28 +00:00
mstembera 062ca91db5 New easy move implementation
Spend much less time in positions where one move is much better than all other alternatives.
We carry forward pv stability information from the previous search to identify such positions.
It's based on my old InstaMove idea but with two significant improvements.

1) Much better instability detection inside the search itself.
2) When it's time to make a FastMove we no longer make it instantly but still spend at least 10% of normal time verifying it.

Credit to Gull for the inspiration.
BIG thanks to Gary because this would not work without accurate PV!

20K
ELO: 8.22 +-3.0 (95%) LOS: 100.0%
Total: 20000 W: 4203 L: 3730 D: 12067

STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 23266 W: 4662 L: 4492 D: 14112

LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 12470 W: 2091 L: 1931 D: 8448

Resolves #283
2015-03-12 19:49:30 +00:00
Stefan Geschwentner 13c11f4048 Introduce Counter Move History tables
Introduce a counter move history table which additionally is indexed by the last move's piece and target square.
For quiet move ordering use now the sum of standard and counter move history table.

STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 4747 W: 1005 L: 885 D: 2857

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 5726 W: 1001 L: 872 D: 3853

Because of reported low NPS on multi core test
STC (7 threads):
ELO: 7.26 +-3.3 (95%) LOS: 100.0%
Total: 14937 W: 2710 L: 2398 D: 9829

Bench: 7725341

Resolves #282
2015-03-12 07:29:57 +00:00
Joona Kiiski 81c7975dcd Use thread specific mutexes instead of a global one.
This is necessary to improve the scalability with high number of cores.

There is no functional change in a single thread mode.

Resolves #281
2015-03-11 21:59:34 +00:00
Marco Costalba 4b59347194 Retire spinlocks
Use Mutex instead.

This is in preparaation for merging with master branch,
where we stilll don't have spinlocks.

Eventually spinlocks will be readded in some future
patch, once c++11 has been merged.

No functional change.
2015-03-11 21:20:47 +01:00
Marco Costalba 6027652773 Cleanup thread_win.h
No functional change.
2015-03-10 17:13:52 +01:00
Marco Costalba 04372316b3 Disable spinlocks
To allow testing on fishtest.

No functional change.
2015-03-10 12:47:49 +01:00
Marco Costalba 8725494966 Add thread_win32.h header
Workaround slow std::thread implementation in mingw
and gcc for Windows with our own old low level thread
functions.

No functional change.
2015-03-10 12:42:40 +01:00
Marco Costalba be50952533 Sync with master
bench: 8285241
2015-03-07 10:56:01 +01:00
Marco Costalba e53774bc49 Sync with master
No functional change.
2015-03-07 10:45:58 +01:00
hxim 3017e8c604 Transform minKingPawnDistance into a local variable
minKingPawnDistance is used only as local variable in one place so we don't need it to be part of "Pawns::Entry" structure.

No functional change.

Resolves #277
2015-03-07 09:03:19 +00:00
Stéphane Nicolet 8fcaa78a04 Update Makefile for Mac OS X compilation
This change in the Makefile restores the possibility to compile
Stockfish on Mac OS X 10.9 and 10.10 after the C++11 has been merged.

To use the default (fastest) settings, compile with:

make build ARCH=x86-64-modern

To test the clang settings, compile with

make build ARCH=x86-64-modern COMP=clang

Beware that the clang settings may provide a slightly slower (6%)
executable.

Backported from master.

No functional change

Resolves #275
2015-03-07 08:39:21 +01:00
Marco Costalba a590d1d52d Re-enable spinlocks
For branch C++11, that doe snot run on fishtest,
there is no need of this kludge, let only master
have it.

No functional change.
2015-03-07 08:38:26 +01:00
Joona Kiiski 856a5f3aaa Revert C++11 merge
Restore the state of repo back to commit 'Simplify pawn code a bit' (1e6d21dbb6)

No functional change
2015-03-07 07:38:22 +00:00
Stéphane Nicolet 6fa6da3ee1 Update Makefile for Mac OS X compilation
This change in the Makefile restores the possibility to compile
Stockfish on Mac OS X 10.9 and 10.10 after the C++11 has been merged.

To use the default (fastest) settings, compile with:

make build ARCH=x86-64-modern

To test the clang settings, compile with

make build ARCH=x86-64-modern COMP=clang

Beware that the clang settings may provide a slightly slower (6%)
executable.

No functional change

Resolves #275
2015-03-05 04:18:59 +00:00
Marco Costalba cb2111f0b6 Disable spinlocks
Now that c++11 branch has been merged in master,
disable unconditionally the spinlocks and use mutex
instead. This will allow to run fishtest even on HT
machines withouth changes.

In the future we will reintorduce spinlocks, once
we will have took care of fishtest.

No functional change.
2015-03-02 08:11:39 +01:00
Marco Costalba 6645115377 Allow to disable spinlocks
And use mutex instead. You may never want to do this.
It is a workaround to run c++11 on fishtest where many
machiens have HTenabled and this can be a problem when
number of cores set is higher than number of physical cores.

To disable spinlocks, just compile with -DNO_SPINLOCK flag

No functional change.
2015-03-01 17:16:05 +01:00
Marco Costalba 63a5fc2366 Rename available_to()
Change this API to be more natural and simple.

Inspired by a patch by Joona.

No functional change.
2015-03-01 12:33:05 +01:00
Marco Costalba 0da7295795 Sync with master
bench: 8285241
2015-02-28 20:22:28 +01:00
Stéphane Nicolet 1e6d21dbb6 Simplify pawn code a bit
Simplify a bit the number of bitwise operators used to calculate the
pawn evaluation in pawns.cpp

No functional change.

Resolves #269
2015-03-01 01:01:37 +08:00
Stéphane Nicolet 9369f4963d Raise penalty for knight attacked by pawn
Raise a bit the penalty for knight attacked by pawn.

STC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 27744 W: 5563 L: 5380 D: 16801

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 87571 W: 14747 L: 14314 D: 58510

Bench: 8285241

Resolves #270
2015-03-01 00:58:32 +08:00
Marco Costalba a459709fcc Retire apply_weight()
Use the more natural operator*() instead.

No functional change.
2015-02-27 10:38:12 +01:00
Marco Costalba 2dbb1adf2a Sync with master
bench: 8069601
2015-02-26 20:50:39 +01:00
Marco Costalba 8a2c4134a6 Normalize twice supported pawns
Align codying style to current conventions and move
formula for twice supported pawns to Pawns::init()
where it should be.

No functional change.
2015-02-27 01:56:16 +08:00
Alain SAVARD fce799294c Apex Pawns
Pawns which are supported already have a bonus. Apex are pawns which are
supported twice.
This patch gives an additional 50% bonus for them.

STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 6549 W: 1333 L: 1209 D: 4007

LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 18002 W: 3037 L: 2850 D: 12115

Bench: 8069601

Resolves #267
2015-02-27 01:52:23 +08:00
Marco Costalba 0b36ba74fc Don't assume the type of Time::point
But instead use the proper definition. Also
rewrite chrono functions while there.

No functional change.
2015-02-24 14:08:14 +01:00
Marco Costalba 153fb216a1 Small tweaks in pawns.cpp
No functional change.
2015-02-24 11:45:29 +01:00
Marco Costalba d3d26a94b3 Improve spinlock implementation
Calling lock.test_and_set() in a tight loop creates expensive
memory synchronizations among processors and penalize other
running threads. So syncronize only only once at the beginning
with fetch_sub() and then loop on a simple load() that puts much
less pressure on the system.

Reported about 2-3% speed up on various systems.

Patch by Ronald de Man.

No functional change.
2015-02-23 19:48:46 +01:00
Marco Costalba 38112060dc Use spinlock instead of mutex for Threads and SplitPoint
It is reported to be defenitly faster with increasing
number of threads, we go from a +3.5% with 4 threads
to a +15% with 16 threads.

The only drawback is that now when testing with more
threads than physical available cores, the speed slows
down to a crawl. This is expected and was similar at what
we had setting the old sleepingThreads to false.

No functional change.
2015-02-23 13:47:07 +01:00
Marco Costalba 775f8239d3 Introduce Spinlock class
Initialization is more complex than what I'd like due
to MSVC compatibility that for some reason does not like:

std::atomic_flag lock = ATOMIC_FLAG_INIT;

No functional change.
2015-02-23 13:37:46 +01:00
Marco Costalba 098f645d26 Sync with master
bench: 8253813
2015-02-23 13:36:15 +01:00
Marco Costalba 29ff925fd8 Fix build under OS X
Reported by Vince Negri

No functional change.
2015-02-22 16:33:24 +01:00
Marco Costalba 8d16111ffd Sync with master
bench: 8253813
2015-02-22 14:20:23 +01:00
Marco Costalba e2226cbb20 Use only 'level' as late join metric
It seems other metric are useless, this allow us
to simplify the code and to prune useless stuff.

STC 20K games 4 threads
ELO: -0.76 +-2.8 (95%) LOS: 29.9%
Total: 20000 W: 3477 L: 3521 D: 13002

STC 10K games 16 threads
ELO: 1.36 +-3.9 (95%) LOS: 75.0%
Total: 10000 W: 1690 L: 1651 D: 6659

bench: 8253813
2015-02-22 12:59:34 +01:00
Marco Costalba 5fd5453e59 Further refine SMP code
Backported from C++11 branch:

https://github.com/official-stockfish/Stockfish/commit/7ff965eebfbc17d2b
https://github.com/official-stockfish/Stockfish/commit/e74c2df907d5336d3d2b

Fully verified it is equivalent to master (see log msg
of individual commits for details).

No functional change.
2015-02-21 11:33:03 +01:00
Marco Costalba e74c2df907 Use sp->master instead of bestThread
Verified with:

dbg_hit_on(th != sp->master);

It is 100% equivalent on more than 200K hits.

No functional change.
2015-02-21 10:40:59 +01:00
Stéphane Nicolet 41ccc885ec Fix comment for kingAdjacentZoneAttacksCount
The comment for kingAdjacentZoneAttacksCount[] was bogus, using
reversed semantics for color.

No functional change

Resolves #262
2015-02-20 19:59:26 +00:00
snicolet 2f46592736 Mobile phalanxes
Try to create mobile phalanxes

STC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 52393 W: 10912 L: 10656 D: 30825

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 30398 W: 5315 L: 5063 D: 20020

Bench: 8253813

Resolves #261
2015-02-20 19:56:57 +00:00
Marco Costalba 7ff965eebf Improve comments in SMP code
No functional change.
2015-02-20 12:38:54 +01:00
Marco Costalba a6f873cd8d Use range-based-for in late join
No functional change.
2015-02-20 10:50:47 +01:00
Marco Costalba 40548c9153 Sync with master
bench: 7911944
2015-02-20 10:37:29 +01:00
Marco Costalba 667f350737 Clarify we don't late join with only 2 threads
Thanks to Gary for pointing this out.

No functional change.
2015-02-19 23:12:59 +01:00
Marco Costalba 950c8436ed Use size_t consistently across thread code
No functional change.
2015-02-19 10:43:28 +01:00
Marco Costalba 8d47caa16e Retire redundant sp->slavesCount field
It should be used slavesMask.count() instead.

Verified 100% equivalent when sp->allSlavesSearching:

dbg_hit_on(sp->allSlavesSearching, sp->slavesCount != sp->slavesMask.count());

No functional change.
2015-02-19 10:36:15 +01:00
Marco Costalba b9d4e6f7fd Fix a warning under MSVC
Assignment of size_t to int.

No functional change.
2015-02-19 10:18:24 +01:00
Marco Costalba 193a7ae35b Add a couple of asserts to late join
Document and clarify that we cannot rejoin on ourselves
and that we never late join if we are master and all
slaves have finished, inded in this case we exit idle_loop.

No functional change.
2015-02-19 10:08:29 +01:00
Marco Costalba 4f906a2589 Remove useless condition in late join
In case of Threads.size() == 2 we have that sp->allSlavesSearching
is always false (because we have finished our search), bestSp is
always NULL and we never late join, so there is no need to special
case here.

Tested with dbg_hit_on(sp && sp->allSlavesSearching) and
verified it never fires.

No functional change.
2015-02-19 09:53:39 +01:00
Marco Costalba dccaa145d2 Compute SplitPoint::spLevel on the fly
And retire a redundant field. This is important also
from a concept point of view becuase we want to keep
SMP structures as simple as possible with the only
strictly necessary data.

Verified with

dbg_hit_on(sp->spLevel != level)

that the values are 100% the same out of more 50K samples.

No functional change.
2015-02-18 21:50:35 +01:00
Marco Costalba 6656ed8904 Simplify attackUnits formula
Use '/ 8' instead of '* 31 / 256'

Passed STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 55077 W: 10999 L: 10940 D: 33138

And LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14751 W: 2530 L: 2400 D: 9821

bench: 7911944
2015-02-17 08:23:35 +01:00
Joona Kiiski d65f75c153 Improve smp performance for high number of threads
Balance threads between split points.

There are huge differences between different machines and autopurging makes it very difficult to measure the improvement in fishtest, but the following was recorded for 16 threads at 15+0.05:

    For Bravone (1000 games): 0 ELO
    For Glinscott (1000 games): +20 ELO
    For bKingUs (1000 games): +50 ELO
    For fastGM (1500 games): +50 ELO

The change was regression for no one, and a big improvement for some, so it should be fine to commit it.
Also for 8 threads at 15+0.05 we measured a statistically significant improvement:
ELO: 6.19 +-3.9 (95%) LOS: 99.9%
Total: 10325 W: 1824 L: 1640 D: 6861

Finally it was verified that there was no (significant) regression for

4 threads:
ELO: 0.09 +-2.8 (95%) LOS: 52.4%
Total: 19908 W: 3422 L: 3417 D: 13069

2 threads:
ELO: 0.38 +-3.0 (95%) LOS: 60.0%
Total: 19044 W: 3480 L: 3459 D: 12105

1 thread:
ELO: -1.27 +-2.1 (95%) LOS: 12.3%
Total: 40000 W: 7829 L: 7975 D: 24196

Resolves #258
2015-02-16 20:36:13 +00:00
lucasart f8f5dcbb68 Compute checkers from scratch
This micro-optimization only complicates the code and provides no benefit.
Removing it is even a speedup on my machine (i7-3770k, linux, gcc 4.9.1):

stat        test     master    diff
mean   2,403,118  2,390,904  12,214
stdev     12,043     10,620   3,677

speedup       0.51%
P(speedup>0) 100.0%

No functional change.
2015-02-16 09:34:26 +08:00
Marco Costalba 686b45e121 Retire one do_move() overload
After Lucas patch it is almost useless.

No functional change.
2015-02-15 12:23:03 +01:00
lucasart dc13004283 Compute checkers from scratch
This micro-optimization only complicates the code and provides no benefit.
Removing it is even a speedup on my machine (i7-3770k, linux, gcc 4.9.1):

stat        test     master    diff
mean   2,403,118  2,390,904  12,214
stdev     12,043     10,620   3,677

speedup       0.51%
P(speedup>0) 100.0%

No functional change.
2015-02-15 12:11:05 +01:00
Marco Costalba 901bfb1f55 Revert "Delayed killers checking"
It seems a slowdown when run with fishbench.

No functional change.
2015-02-15 11:32:53 +01:00
Marco Costalba 20a5c07472 Further simplify KingDanger init
And remove a tale whitespace while there.

No functional change.
2015-02-14 15:56:54 +01:00
Marco Costalba e38eb2302d Further simplify KingDanger init
And remove a tale whitespace while there.

No functional change.
2015-02-14 15:55:11 +01:00
Marco Costalba 0af24a1445 Sync with master
Bench: 7369224
2015-02-14 15:30:16 +01:00
snicolet a8f9c7a790 Small bonus for all safe pawn pushes
Pawn flexibility: add a small bonus for all safe pawn pushes

STC:
LLR: 2.70 (-2.94,2.94) [-1.50,4.50]
Total: 18233 W: 3705 L: 3557 D: 10971

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 17684 W: 3042 L: 2854 D: 11788

Bench: 7369224

Resolves #253
2015-02-13 21:33:00 +00:00
Marco Costalba c0a80afe89 Reformat tracing functions
No functional change.
2015-02-13 11:54:46 +01:00
hxim 44a571c1c7 Fix KingDanger[] array initialization
Use integer arithmetic instead of floating point arithmetic.
Floating point arithmetic was causing different results for some 32-bit compiles

No functional change

Resolves #249
Resolves #250
2015-02-09 22:02:35 +00:00
Marco Costalba ce8ac7997c Sync with master
bench: 7699138
2015-02-08 21:32:14 +01:00
Joona Kiiski e118570038 Pawn Center Bind Bonus
Bonus for two pawns controlling the same central square

STC:

LLR: 3.14 (-2.94,2.94) [-1.50,4.50]
Total: 15974 W: 3291 L: 3133 D: 9550

LTC:

LLR: 3.24 (-2.94,2.94) [0.00,6.00]
Total: 10449 W: 1837 L: 1674 D: 6938

Idea from Lyudmil Tsvetkov.

Bench: 7699138

Resolves #248
2015-02-08 19:28:01 +00:00
Marco Costalba 8f10f6c9cd Shuffle put_piece() and friends signatures
It is more consistent with the others member functions.

No functional change.
2015-02-08 18:17:08 +01:00
Marco Costalba 3184852bdc Small tweaks in do_move and friends
Also remove useless StateCopySize64 optimization:
compiler uses SSE movups instruction anyhow and
does not need this trick (verified with fishbench).

No functional change.
2015-02-08 13:09:29 +01:00
Marco Costalba 99c9cae586 Avoid casting to char* in prefetch()
Funny enough, gcc __builtin_prefetch() expects
already a void*, instead Windows's _mm_prefetch()
requires a char*.

The patch allows to remove ugly casts from caller
sites.

No functional change.
2015-02-07 19:13:41 +01:00
Marco Costalba 152a4dc5cd Rewrite pos_is_ok()
No functional change.
2015-02-07 15:02:28 +01:00
Marco Costalba 47a0768102 Micro-optimize SEE
Results for 10 tests for each version (gcc 4.8.3 on mingw):

            Base      Test      Diff
    Mean    1502447   1507917   -5470
    StDev   3119      1364      4153

p-value: 0,906
speedup: 0,004

Results for 10 tests for each version (MSVC 2013):

            Base      Test      Diff
    Mean    1400899   1403713   -2814
    StDev   1273      2804      2700

p-value: 0,851
speedup: 0,002

No functional change.
2015-02-07 12:21:39 +01:00
Marco Costalba 170bdf40cd Rename dbg_hit_on_c() to dbg_hit_on()
Use an overload instead of a new named function.

I have found this handier and easier when adding
some quick debug code.

No functional change.
2015-02-07 11:18:06 +01:00
Marco Costalba 8b0fee9998 Rename dbg_hit_on_c() to dbg_hit_on()
Use an overload instead of a new named function.

I have found this handier and easier when adding
some quick debug code.

No functional change.
2015-02-07 11:15:38 +01:00
Marco Costalba 1277a42823 Sync with master
bench: 7696257
2015-02-07 10:32:28 +01:00
lucasart 35aa21c1fe Removes useless templates, some of which lead to code duplication: is_K*() functions.
No functional change

Resolves #245
2015-02-07 09:12:04 +00:00
Stefan Geschwentner 18b0809639 Add bonus for pawn attack threats
Latent pawn attacks: Add a bonus to safe pawn pushes which attacks an
enemy piece.  Based on an idea of Lyudmil Tsvetkov.

STC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 7925 W: 1666 L: 1537 D: 4722

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 40109 W: 6841 L: 6546 D: 26722

Bench: 7696257

Resolves #240
2015-02-03 11:19:33 +08:00
mstembera f4136c5434 Profile build options
I went through all the individual compile options that differ between
-fprofile-generate/-fprofile-use  and  -fprofile-arcs/-fbranch-probabilities
and distilled the speed difference down to only turning off
-fno-peel-loops and -fno-tracer.  Using this we still get the full speedup
(maybe a bit more because other optimizations stay on) and it's also much cleaner
because we can get rid of the "@rm -f ucioption.gc*" hack for all versions of gcc.

No functional change.

Resolves #237
2015-02-03 11:09:37 +08:00
NicklasPersson ddccb5355c Improved King Safety values
From an SPSA-session on king safety.

STC:
ELO: 3.21 +-2.1 (95%) LOS: 99.8%
Total: 40000 W: 8181 L: 7812 D: 24007

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 47765 W: 8091 L: 7785 D: 31889

Bench: 8589262

Resolves #241
2015-02-03 04:00:52 +08:00
Marco Costalba 1c7a727795 Use move assignment in movegen.h
No functional change and same speed (tested with perft)
2015-02-01 16:42:31 +01:00
Marco Costalba e56e5045dd Allow to assign a Move to an ExtMove
After defining ExtMove::operator Move(), this is a
natural extension.

No fnctional change.
2015-02-01 13:31:49 +01:00
Marco Costalba 20d6a8e57f Delay checking for duplicated killer moves
Follow the usual approach to delay computation
as far as possible, in case an earlier killer
cut-offs we avoid to do useless work.

This also greatly simplifies the code.

No functional change.
2015-02-01 13:17:42 +01:00
Marco Costalba 0dc6f16992 Small tweaks in movepick.cpp
No functional change.
2015-02-01 12:01:35 +01:00
Marco Costalba 21120288a3 Silence a warning under MSVC
warning C4100: 'ci' : unreferenced formal parameter

It is a silly and wrong one, but just silent it.

No functional change.
2015-02-01 10:40:22 +01:00
Marco Costalba 519b2fe849 More readable score<CAPTURES>()
No functional change.
2015-01-31 20:13:38 +01:00
Marco Costalba f3189bdc9a Use C++11 loops in MovePicker
No functional change.
2015-01-31 20:02:08 +01:00
Marco Costalba 65f46794af Implicit conversion from ExtMove to Move
Verified with perft there is no speed regression,
and code is simpler. It is also conceptually correct
becuase an extended move is just a move that happens
to have also a score.

No functional change.
2015-01-31 19:22:07 +01:00
Marco Costalba 81d6c4a0d6 Another small tweak to skills
No functional change.
2015-01-31 18:24:39 +01:00
Marco Costalba 60c121f3b1 Sync with master
bench: 7374604
2015-01-31 13:05:51 +01:00
Marco Costalba 45eac9507c Use C++ loops in insert_pv_in_tt
Also small tweak to extract_ponder_from_tt

No functional change.
2015-01-31 12:57:52 +01:00
Marco Costalba bfd0f95f06 Move uci_pv under UCI namespace
That's the correct place.

No functional change.
2015-01-31 12:18:24 +01:00
Marco Costalba 1b947aafbf Convert Reductions[] from int8_t to Depth
This is the type anyhow. Assorted cleanup while there.

No functional change.
2015-01-31 11:44:35 +01:00
Marco Costalba a7592e69d7 Fix a MSVC warning
warning C4805: '|' : unsafe mix of type 'Bitboard' and type 'bool' in operation

No functional change.
2015-01-31 10:19:00 +01:00
Jean-Francois Romang a3b4e9e23c Ressurrect hashfull patch
This is an old patch from Jean-Francois Romang to send
UCI hashfull info to the GUI:
https://github.com/mcostalba/Stockfish/pull/60/files

It was wrongly judged as a slowdown, but it takes much
less than 1 ms to run, indeed on my core i5 2.6Ghz it
takes about 2 microsecs to run!

Regression test is good:

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7352 W: 1548 L: 1401 D: 4403

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 61432 W: 10307 L: 10251 D: 40874

I have set the name of the author to the original
one.

No functional change.
2015-01-30 18:07:20 +01:00
Marco Costalba f6f1d24223 Sync with master
bench: 7374604
2015-01-30 17:58:18 +01:00
Marco Costalba ce0a95c2c0 Simplify skill level and reduce ELO
This patch has two positive effects:

- Retire a hackish formula and leave
  just a natural, simple and plain one.

- Reduce strenght at very low level, but
  don't impact medium/high levels.

Actually even at level 0, SF is still too
strong for many beginners (this was reported
many times for instance on Droidfish user
comments on Google Play).

Test on fishtest shows that ELO drop is around
170 ELO at level 0 (good!), 130 ELO at level 1
and smoothly reduces (as expected) until level
10 where the drop is just of 8 ELO.

No functional change.
2015-01-29 22:35:24 +01:00
Alain SAVARD 8aa8608c2a Simplify backward pawn definition
Make use of 'lever' attribute

No functional change

Resolves #234
2015-01-28 21:29:04 +00:00
NicklasPersson 7837fb2aca King safety tuning with values obtained by SPSA.
Part I:

LTC:

LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 11529 W: 2075 L: 1882 D: 7572

Part II:

LTC:

ELO: 2.07 +-2.1 (95%) LOS: 97.3%
Total: 34859 W: 5967 L: 5759 D: 23133

Bench: 7374604

Resolves #228
2015-01-28 21:05:21 +00:00
Joona Kiiski 9f0d5241bf Restore development version
No functional change
2015-01-28 20:53:50 +00:00
Joona Kiiski 5b555525d2 Stockfish 6
Stockfish bench signature is: 8918745
2015-01-27 20:27:38 +00:00
Joona Kiiski c4518e395e Stockfish 6 Release Candidate 3
- Fix a skill level problem: Don't allow move pruning at root node
- Revert "Fix profile build for gcc on Mac OSX". Results for a faster binary in x86-64.
- Fix a MSVC warning

Bench: 8918745
2015-01-25 22:03:57 +00:00
Stefan Geschwentner d8b3ad2208 Fix a skill level problem: Don't allow move pruning at root node
Bench: 8918745

Resolves #231
2015-01-25 21:57:13 +00:00
Joona Kiiski ec36b8dea9 Revert "Fix profile build for gcc on Mac OSX"
Seems to be a performance regression for standard build.

For SF6 people compiling on Mac OSX using profile-build option
just need to make necessary adjustments manually...

No functional change

Resolves #223
2015-01-25 21:51:09 +00:00
Marco Costalba 14f7d9a629 Re-arrange Skill struct
Instead of swapping sub-optimal move in Skill
d'tor, make it explicitly at the end of the search.

Also streamline and clarify relation with multiPV
and pass it directly instead of relying on the hacky
'candidates' member.

No functional change.
2015-01-25 11:16:54 +01:00
Marco Costalba 8ebf30d44a Fix a MSVC warning at W4
Warning is C4512 (assignment operator could not be generated)

Now, apart the foreign syzygy code, everything compiles
without warnings at warning level 4.

Backported from C++11 branch.

No functional change.
2015-01-25 09:36:05 +01:00
Joona Kiiski 5154ac9cff Stockfish 6 Release Candidate 2
- Fix a compilation issue related to BMI2 PEXT instruction
- Retrieve a ponder move from TT if PV is only one move long

Bench: 8080602

No functional change
2015-01-24 19:42:49 +00:00
Marco Costalba 54b5b528d9 Don't use _pext_u64() directly
This intrinsic to call BMI2 PEXT instruction is
defined in immintrin.h. This header should be
included only when USE_PEXT is defined, otherwise
we define _pext_u64 as 0 forcing a nop.

But under some mingw platforms, even if we don't
include the header, immintrin.h gets included
anyhow through an include chain that starts with
STL <algorithm> header. So we end up both defining
_pext_u64 function and at the same time defining
_pext_u64 as 0 leading to a compile error.

The correct solution is of not using _pext_u64 directly.

This patch fixes a compile error with some mingw64
package when compiling with x86-64.

No functional change.

Resolves #222
2015-01-24 19:38:06 +00:00
Marco Costalba 44643c2770 Try hard to retrieve a ponder move
In case we stop the search during a fail-high
it is possible we return to GUI without a ponder
move. This patch try harder to find a ponder move
retrieving it from TT. This is important in games
played with 'ponder on'.

bench: 8080602

Resolves #221
2015-01-24 19:35:31 +00:00
Marco Costalba 69407ae9b8 Additional work in bitbases
Verified the generated bitbases are unchanged.

No functional change.
2015-01-24 09:29:35 +01:00
Marco Costalba 18375e3bfb Rearrange bitbases C++11 way
No functional change.
2015-01-22 11:05:31 +01:00
Marco Costalba 6390a3da94 Document how to enable PEXT with MSVC
When not using Makefile, e.g. with MSVC, if hardware
supports BMI2 instructions, then USE_PEXT should be
added in project configuration to enable pext support.

No functional change.
2015-01-21 19:54:15 +01:00
Marco Costalba 36f8814aa6 Rearrange Endgames
Remove references to EndgameBase and use instead
Value and ScaleFactor as template parameters of
the endgames maps.

No functional change.
2015-01-21 17:35:53 +01:00
Marco Costalba 96e36a7897 Explicitly defaulted and deleted members
Better than a bit obscure implicit ones.

No functional change.
2015-01-21 13:18:19 +01:00
Marco Costalba 2ca2c3f35b Fun with lambdas
Use lambda functions instead of has_positive_value()
and toggle_case()

No functional change.
2015-01-21 11:33:53 +01:00
Marco Costalba f54c44e6be Don't use _pext_u64() directly
This intrinsic to call BMI2 PEXT instruction is
defined in immintrin.h. This header should be
included only when USE_PEXT is defined, otherwise
we define _pext_u64 as 0 forcing a nop.

But under some mingw platforms, even if we don't
include the header, immintrin.h gets included
anyhow through an include chain that starts with
STL <algorithm> header. So we end up both defining
_pext_u64 function and at the same time defining
_pext_u64 as 0 leading to a compile error.

The correct solution is of not using _pext_u64 directly.

This patch fixes a compile error with some mingw64
package when compiling with x86-64.

No functional change.
2015-01-20 22:17:22 +01:00
Marco Costalba ca3622f8e8 Fix a MSVC warning at W4
Warning is C4512 (assignment operator could not be generated)

Now, apart the foreign syzygy code, everything compiles
without warnings at warning level 4.

No functional change.
2015-01-19 08:11:43 +01:00
Joona Kiiski 97a034ad3e Stockfish 6 Release Candidate 1
Bench: 8080602

No functional change

Resolves #218
2015-01-18 14:53:11 +00:00
Joona Kiiski 1b62b47b62 Fix profile build for gcc on Mac OSX
Switch back to using -fprofile-generate and
-fprofile-use flags

No functional change

Resolves #219
Resolves #210
2015-01-18 14:42:30 +00:00
Joona Kiiski 7f51610103 Don't print fail-high or fail-lows in MultiPV mode
Supposed to give a better user experience when using MultiPV mode

No functional change

Resolves #217
2015-01-18 14:38:46 +00:00
Marco Costalba 712bbb052e Fix compile error with MSVC 2013
Without this std::min chokes with windows 'min' macro.

No functional change.
2015-01-18 15:14:05 +01:00
Marco Costalba fe4fda474f Fun with initializer lists in UCI::square
No functional change.
2015-01-18 11:04:51 +01:00
Marco Costalba f7d8ea3866 Fix a coverity scan warning
Coverity scan warns about uninitialized 'sf' argument when
calling probe(). Actually it is a false positive because
argument is passed by reference and assigned inside
probe(). Nevertheless it is a hint that fucntion signature
is a bit tricky, so rewrite it in a more conventional way,
assigning 'sf' from probe() return value.

No functional change.
2015-01-18 10:41:56 +01:00
Marco Costalba f53aea45e3 Add syzygy support
bench: 8080602
2015-01-18 08:27:46 +01:00
Marco Costalba 3c07603dac Import C++11 branch
Import C++11 branch from:

https://github.com/mcostalba/Stockfish/tree/c++11

The version imported is teh last one as of today:
https://github.com/mcostalba/Stockfish/commit/6670e93e50a7a4e739ac2ac8b6026ffa3a12150a

Branch is fully equivalent with master but syzygy
tablebases that are missing (but will be added with
next commit).

bench: 8080602
2015-01-18 08:00:50 +01:00
Marco Costalba c73f33f37e Fix syzygy warnings with Intel compiler
Quick hack, a better job would require to
first move all syzygy code up to SF standard.

No functional change.
2015-01-18 07:29:51 +01:00
Marco Costalba 05cb58f4fc Fix some missing rename from previous patch
No functional change.
2015-01-17 22:15:15 +01:00
Marco Costalba 595fc342cf Fix a possible overflow in TT resize
On platforms where size_t is 32 bit, we
can have an overflow in this expression:

(mbSize * 1024 * 1024)

Fix it setting max hash size of 2GB on platforms
where size_t is 32 bit.

A small rename while there: now struct Cluster
is definied inside class TranspositionTable so
we should drop the redundant TT prefix.

No functional change.
2015-01-17 21:42:47 +01:00
mstembera 58fdb84b0d Simplify and optimize value extractors
Speed up results by Joona:

gcc-4.7 (1.5%)
gcc-4.8 (0.5%)
gcc-4.9 (1.0%)

Speed up results by mstembera:

gcc 474
p-value: 0.719

gcc 482
p-value: 1

gcc 492
p-value: 0.859

No functional change

Resolves #211
2015-01-16 19:03:49 +00:00
lucasart ab276357d7 Enable futility pruning for PV nodes in qsearch
STC:

LLR: 4.20 (-2.94,2.94) [-3.00,1.00]
Total: 85573 W: 17195 L: 17125 D: 51253

LTC:

LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 43385 W: 7298 L: 7214 D: 28873

Bench: 8080602

Resolves #206
2015-01-14 20:21:54 +00:00
Stefan Geschwentner 4abe333e1f Enable Futility pruning in PV nodes
STC:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21553 W: 4342 L: 4221 D: 12990

LTC:

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7675 W: 1351 L: 1209 D: 5115

Bench: 8668014

Resolves #205
2015-01-14 20:13:48 +00:00
hxim 7b20bb6e1a Use VALUE_ZERO instead of 0 when comparing with Value
This is the correct practice.

No functional change

Resolves #208
2015-01-13 21:07:18 +00:00
Marco Costalba 4eb2d8ce09 Assorted headers cleanup
Mostly comments fixing and other small things.

No functional change.
2015-01-11 22:56:35 +01:00
Marco Costalba b97df4c236 Fix a crash of syzygy on Android
On Android-ARM current TB code crashes at
random times even in single thread mode.

Reported, debugged, fixed and verified
by Peter Osterlund.

No functional change.

Resolves #201
2015-01-11 20:10:31 +00:00
Stefano80 17bcdb9212 Higher movetime resolution for benchmarks
Use a millisecond movetime precision for benchmarks

No functional change

Resolves #198
2015-01-11 20:07:11 +00:00
Marco Costalba 202a78e8e4 Fix compile for Android 5
Android 5 can only run position independent executables.

Note that this breaks Android 4.0 and earlier.

See here for more info:
http://stackoverflow.com/questions/24818902/running-a-native-library-on-android-l-error-only-position-independent-executab

Thanks to Peter Osterlund for the support.

No functional change
2015-01-11 10:52:42 +01:00
Marco Costalba 42b48b08e8 Update copyright year
No functional change.
2015-01-10 11:46:28 +01:00
Marco Costalba aea2fde611 Assorted formatting and comment tweaks in position.h
No functional change.
2015-01-07 09:09:41 +01:00
lucasart b73ae56ee1 Tune null search reduction
STC

LLR: 2.96 (-2.94,2.94) [-0.50,4.50]
Total: 107289 W: 21851 L: 21325 D: 64113

LTC

LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 83837 W: 14378 L: 13916 D: 55543

Bench: 7604776

Resolves #194
2015-01-05 22:09:12 +00:00
Marco Costalba 3fda064a66 Retire one implementation of pop_lsb()
We have two implementations that are equivalent,
so retire one.

Plus usual tidy up of comments and code reshuffle.

No functional change.
2015-01-03 22:16:30 +01:00
lucasart a6e292034a Remove a useless optimization
This optimization is aimed at old hardware only (withouth popcount), and even on
non popcount compile (ARCH=x86-64), it provides no mesurable speedup:

stat        test     master     diff
mean   2,341,779  2,354,699  -12,920
stdev     12,910     14,770   18,150

speedup      -0.55%
P(speedup>0)  23.8%

No functional change.

Resolves #187
2015-01-03 20:40:57 +00:00
Joona Kiiski c7332d5610 Smoother king safety
STC:

LLR: 4.03 (-2.94,2.94) [-1.50,4.50]
Total: 35707 W: 7352 L: 7106 D: 21249

LTC:

LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 24563 W: 4330 L: 4095 D: 16138

Bench: 8411409

Resolves #190
2015-01-03 20:34:20 +00:00
Marco Costalba 9d1e2c0e76 Assorted work in uci.cpp
- Change UCI::value() signature

This function should only return the value,
lowerbound and upperbound info is up to the
caller because it requires external knowledge,
out of the scope of this little helper.

- Retire 'key' command

It is not an UCI command and is absolutely
useless: never used.

- Comments fixing and other trivia

No functional change.
2015-01-03 18:46:59 +01:00
Marco Costalba c150f07291 Fix a silly warning with Intel compiler
Intel compiler cries for an explicit cast

warning #2259: non-pointer conversion from "double" to "int"
may lose significant bits

No functional change.
2015-01-03 10:17:04 +01:00
Marco Costalba 62f531254e Fix comments in thread.cpp
And reshuffle a bit the functions to place
them in a consistent order.

To be on the safe side, patch has been
validated for no regression/crashes with
a small 8K games test with 3 threads:

ELO: 3.98 +-4.4 (95%) LOS: 96.3%
Total: 8388 W: 1500 L: 1404 D: 5484

No functional change.
2015-01-03 09:34:58 +01:00
Marco Costalba 91cc82aa25 Let material probing to access per-thread table
It is up to material (and pawn) table look up
code to know where the per-thread tables are,
so change API to reflect this.

Also some comment fixing while there

No functional change.
2015-01-02 21:31:02 +01:00
Matthew Lai 19b8249ff4 Disable the default copy constructor for Position class
No functional change

Resolved #183
2015-01-02 20:19:11 +00:00
Stefan Geschwentner 25b492ab58 Big King Safety tuning
All king safety related terms (shelterweakness, stormdanger,
attackunits, ..) was tuned together. Additionally for attack units a
finer granularity (factor 4) is used.

STC
ELO: 9.22 +-3.1 (95%) LOS: 100.0%
Total: 19514 W: 4340 L: 3822 D: 11352

LTC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 6399 W: 1192 L: 1056 D: 4151

Bench: 8224782

Resolves #184
2015-01-03 03:33:02 +08:00
Marco Costalba 4c9b423161 Retire Material::space_weight()
Move all in evaluation.

Simplify the code and concentrate in a single place
all the logic behind space evaluation, making it much
more clear.

Verified also at STC it does not regress due to a possible
slow down:

LLR: 3.91 (-2.94,2.94) [-3.00,1.00]
Total: 65744 W: 13285 L: 13194 D: 39265

No functional change.
2015-01-02 10:02:28 +01:00
Marco Costalba 2416242c96 Additional tidy up in timeman.cpp
Fixed some comments and moved/renamed some
variables.

No functional change.
2014-12-30 09:49:54 +01:00
Marco Costalba 6933f05f4b Use score and value consistently
And other assorted small fixing, code style
tweaks and reshuffles in evaluate.cpp

No functional change.
2014-12-28 19:06:56 +01:00
Arjun Temurnikar f9571e8d57 Remove some comments in timeman.cpp
Commenst are obsolete now, an updated description
would be quite obscure, so better let the code
to talk and remove them all together.

No functional change.
2014-12-28 11:58:29 +01:00
Marco Costalba 2bfacf136c Reformat promotion generation
Use the same template of other pawns moves generation,
make the code more uniform, simplify generate_promotions
that has now been renamed.

No functional change (verified also with perft).
2014-12-25 11:09:07 +01:00
Marco Costalba 296534f234 Prefer names to numbers in storm code
Also replaces some tabs with spaces and
change StormDanger order to reflect
ShelterWeakness one.

No functional change.
2014-12-22 08:33:07 +01:00
Joona Kiiski e5c7b44f7a Use "rm -f" instead of "rm" for gcc profiling hack in Makefile
In some UNIX systems "rm" prompts user for confirmation.
However "rm -f" is always a guaranteed forced deletion.

Also move gcc profiling hack under the correct target

No Functional change

Resolves #168
2014-12-20 17:55:18 +00:00
Stefan Geschwentner 3231038262 Big King Safety tuning
ShelterWeakness and Stormdanger array are now indexed additionally by
file pair (a/h,b/g,c/f,d/e). The special case of king blocking a pawn
is incorporated in the StormDanger array.  Finally the 93 parameters
are tuned by SPSA on LTC.

STC
ELO: 3.46 +-2.2 (95%) LOS: 99.9%
Total: 40000 W: 8275 L: 7877 D: 23848

LTC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 10311 W: 1876 L: 1721 D: 6714

Bench: 9498821

Resolves #163
2014-12-21 01:53:44 +08:00
Marco Costalba 9cae6e66ce Don't account for Tempo in specialized endgames
The evaluation is already done by the specialized
function, don't need to add something elese later.

With this patch following positions are evaluated
correctly as draws:

8/6p1/1Pkp1p1p/2nNn2P/2P1K1P1/8/8/3B4 w - - 7
8/1k4p1/1P1p1p1p/3NnK1P/2P3P1/1n6/4B3/8 w - -

Verified it not regress with an STC test:
LLR: 3.15 (-2.94,2.94) [-3.00,1.00]
Total: 49812 W: 10095 L: 10016 D: 29701

Reported by Arjun Temurnikar.

bench: 8289983
2014-12-19 11:06:40 +01:00
mstembera 46d5fff01f Change profile-build options to produce 1% to 2% faster executables.
The "@rm ucioption.gc*" line is necessary to avoid a gcc 4.7.x bug.
Confirmed for gcc 4.7.4, 4.8.1, and 4.9.1
Suggested by Kiran Panditrao on fishcooking forum.
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/AY8gN53nG18

No functional change.

Resolves #160
2014-12-19 03:57:04 +08:00
Marco Costalba b8fd1a78dc Improve comments in UCI
And simplify naming while there.

No functional change.

Resolves #159
2014-12-14 23:50:33 +00:00
Marco Costalba 413b243809 Coding style in TT code
In particular seems more natural to return
bool and TTEntry on the same line, actually
we should pass and return them as a pair,
but due to limitations of C++ and not wanting
to use std::pair this can be an acceptable
compromise.

No functional change.

Resolves #157
2014-12-14 23:49:00 +00:00
Gary Linscott 0edb6348d2 Fix compile for some versions of mingw
The bswap intrinsics are specific to the compiler, not the
host platform.

No functional change.

Resolves #155
2014-12-14 14:45:43 -05:00
mstembera 14cf27e6f6 Avoid searching TT twice for the same key/position during probe() and store().
Just keep the pointer and remove code from tt.cpp

STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 13620 W: 2810 L: 2665 D: 8145

LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 13021 W: 2238 L: 2073 D: 8710STC http://tests.stockfishchess.org/tests/view/548436860ebc59331739b90c

STC 4MB
ELO: 2.41 +-2.2 (95%) LOS: 98.6%
Total: 40000 W: 8175 L: 7897 D: 23928

LTC 16MB
ELO: 1.78 +-2.0 (95%) LOS: 96.1%
Total: 39683 W: 6763 L: 6560 D: 26360

Resolves #151

Bench: 8116521
2014-12-13 07:22:37 +00:00
Gary Linscott 7b4828b68c Only use _ReadWriteBarrier on MSVC
It was causing compile errors when cross-compiling using mingw.

No functional change.
2014-12-11 14:56:24 -05:00
joergoster f6d220ab14 Halve StormDanger bonus for blocked pawn on A/H file
STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 3410 W: 758 L: 641 D: 2011

LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 33225 W: 5708 L: 5445 D: 22072

Bench: 8465811

Resolves #153
2014-12-11 13:08:29 -05:00
Gary Linscott afafdf7b73 MSVC compiling fixes
No functional change.

Resolves #150
2014-12-11 13:03:44 -05:00
Joona Kiiski b15dcd9774 Fix profile build for syzygy
Touch source files under syzygy directory to force recompilation
after collecting profile data.

No functional change

Resolves #149
2014-12-10 17:59:41 +00:00
Joona Kiiski 94dd204c3b Retire 'os' flag from Makefile
Appears to be unused

No functional change

Resolves #147
2014-12-10 17:57:55 +00:00
Marco Costalba 5943600a89 Assorted nitpicking code-style
No functional change.
2014-12-10 12:38:13 +01:00
Marco Costalba 589c711449 Clarify when forcing the moves loop
In some cases we want to go direcly to the moves loop
without checking for early return. The patch make this
logic more clear and consistent.

Tested for no regression, passed STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25282 W: 5136 L: 5022 D: 15124

and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 72007 W: 12133 L: 12095 D: 47779

bench: 9316798
2014-12-10 12:35:21 +01:00
Ernesto Gatti 158864270a Simpler PRNG and faster magics search
This patch replaces RKISS by a simpler and faster PRNG, xorshift64* proposed
by S. Vigna (2014). It is extremely simple, has a large enough period for
Stockfish's needs (2^64), requires no warming-up (allowing such code to be
removed), and offers slightly better randomness than MT19937.

Paper: http://xorshift.di.unimi.it/
Reference source code (public domain):
http://xorshift.di.unimi.it/xorshift64star.c

The patch also simplifies how init_magics() searches for magics:

- Old logic: seed the PRNG always with the same seed,
  then use optimized bit rotations to tailor the RNG sequence per rank.

- New logic: seed the PRNG with an optimized seed per rank.

This has two advantages:
1. Less code and less computation to perform during magics search (not ROTL).
2. More choices for random sequence tuning. The old logic only let us choose
from 4096 bit rotation pairs. With the new one, we can look for the best seeds
among 2^64 values. Indeed, the set of seeds[][] provided in the patch reduces
the effort needed to find the magics:

64-bit SF:
Old logic -> 5,783,789 rand64() calls needed to find the magics
New logic -> 4,420,086 calls

32-bit SF:
Old logic -> 2,175,518 calls
New logic -> 1,895,955 calls

In the 64-bit case, init_magics() take 25 ms less to complete (Intel Core i5).

Finally, when playing with strength handicap, non-determinism is achieved
by setting the seed of the static RNG only once. Afterwards, there is no need
to skip output values.

The bench only changes because the Zobrist keys are now different (since they
are random numbers straight out of the PRNG).

The RNG seed has been carefully chosen so that the
resulting Zobrist keys are particularly well-behaved:

1. All triplets of XORed keys are unique, implying that it
   would take at least 7 keys to find a 64-bit collision
   (test suggested by ceebo)

2. All pairs of XORed keys are unique modulo 2^32

3. The cardinality of { (key1 ^ key2) >> 48 } is as close
   as possible to the maximum (65536)

Point 2 aims at ensuring a good distribution among the bits
that determine an TT entry's cluster, likewise point 3
among the bits that form the TT entry's key16 inside a
cluster.

Details:

     Bitset   card(key1^key2)
     ------   ---------------
RKISS
     key16     64894   = 99.020% of theoretical maximum
     low18    180117   = 99.293%
     low32    305362   = 99.997%

Xorshift64*, old seed
     key16     64918   = 99.057%
     low18    179994   = 99.225%
     low32    305350   = 99.993%

Xorshift64*, new seed
     key16     65027   = 99.223%
     low18    181118   = 99.845%
     low32    305371   = 100.000%

Bench: 9324905

Resolves #148
2014-12-08 08:18:26 +08:00
Gary Linscott a87da2c4b3 Add some tablebase positions to bench
This makes it easier to check for regressions in the tablebase code.

Bench: 9489202
5-man bench: 8943906

Resolves #145
2014-12-08 07:58:05 +08:00
hxim fbb53524ef Rename some variables for more clarity.
No functional change.

Resolves #131
2014-12-08 07:53:33 +08:00
Marco Costalba ba1464751d Explicitly pass RootMoves to TB probes
Currently Search::RootMoves is accessed and even
modified by TB probing functions in a hidden
and sneaky way.

This is bad practice and makes the code tricky.
Instead explicily pass the vector as function
argument so to clarify that the vector is modified
inside the functions.

No functional change.
2014-12-06 15:08:21 +00:00
Marco Costalba eeb6d923fa Move TB stuff under Tablebases namespace
Simplified also some logic while there.

TBLargest needs renaming too, but itis for
a future patch because touches also syzygy
directory stuff.

No functional change.
2014-12-06 14:58:00 +00:00
Marco Costalba c30eb4c9c9 Refactor syzygy code in search
Move to SF coding style.

Also skip calculating piece count in search()
when TB are not found (!TBCardinality)

No functional change.
2014-12-06 14:35:50 +00:00
Joona Kiiski 35c1ccef39 Retire support for Haiku installation directory from Makefile
- It is out of the scope of the project.
- It is the responsibility of Haiku package maintainer to
  configure this.

No functional change

Resolves #143
2014-12-06 14:23:08 +00:00
Joona Kiiski 0935dca9a6 Retire hackish support for aCC and HP-UX from Makefile
- It is out of scope of the project.
- We have no way to verify that it even works anymore

No functional change

Resolves #142
2014-12-06 14:19:39 +00:00
Marco Costalba 314d446518 Retire total_piece_count()
We really don't need to uglify in this way
our nice count() API with this ad-hoc hack.

So remove the hack and use the already
existing infrastructure.

No functional change.

Resolves #134
2014-11-30 20:37:24 +00:00
hxim c014444f09 Remove CONNECTED_KINGS from Syzygy code
No functional change

Resolves #140
2014-11-30 20:24:32 +00:00
Rodrigo Exterckötter Tjäder 9b4e123fbe Cleaning Syzygy profiling data
Updating the makefile so that the clean and gcc-profile-clean targets also
remove the profiling data files in the syzygy directory.

No functional change.

Resolves #136
2014-11-30 19:53:04 +00:00
Marco Costalba a43f633c19 Rewrite TBScore in uci_pv()
Streamline the code and make
it understandable.

No functional change.

Resolves #135
2014-11-30 19:35:35 +00:00
Marco Costalba 66f5cd3f9d Retire #ifdef SYZYGY macro
It just clutters the code for no
real reason.

No functional change.

Resolves #139
2014-11-30 19:23:17 +00:00
mstembera fe07ae4cb4 Bitbase index() from ADD to OR.
No functional change.

Resolves #132
2014-11-26 07:56:48 +08:00
lucasart 2c52147dbf Introduce ratio operation
Just like in Physics, the ratio of 2 things in the same unit, should be
without unit.

Example use case:
- Ratio of a Depth by a Depth (eg. ONE_PLY) should be an int.
- Ratio of a Value by a Value (eg. PawnValueEg) should be an int.

Remove a bunch of useless const while there.

No functional change.

Resolves #128
2014-11-26 07:55:57 +08:00
Ronald de Man 7caa6cd338 Syzygy tablebases
Adds support for Syzygy tablebases to Stockfish.  See
the Readme for information on using the tablebases.

Tablebase support can be enabled/disabled at the Makefile
level as well, by setting syzygy=yes or syzygy=no.

Big/little endian are both supported.

No functional change (if Tablebases are not used).

Resolves #6
2014-11-26 07:49:58 +08:00
Gary Linscott 4509eb1342 Fix out-of-bound array access printing ponder move
It is possible that we won't have a ponder move if our PV
is too short.  In that case, just don't print a ponder move.

No functional change

Resolves #130
2014-11-24 08:53:00 +08:00
Gary Linscott 7ad59d9ac9 Fix pondering
The UCI specification states that an engine can never exit the search
while pondering.

No functional change.

Resolves #118
2014-11-24 08:50:36 +08:00
Jonathan Calovski 48127fe5d3 Amend defended
Amend defended to remove now redundant condition.

No functional change.

Resolves #125
2014-11-22 05:46:59 +08:00
Marco Costalba 84408e5cd6 Fix doubled pawns asymmetry
When evaluating double pawns we use always
lsb() to extract the frontmost square.

This breaks evaluation color symmetry as is
possible to verify with an instrumented evaluate()

  Value evaluate(const Position& pos) {

    Value v = do_evaluate<false>(pos);
    Position p = pos;
    p.flip();
    assert(v == do_evaluate<false>(p));
    return v;
  }

Passed no regression test:

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21035 W: 4244 L: 4122 D: 12669

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39839 W: 6662 L: 6572 D: 26605

bench: 8255966
2014-11-21 20:40:25 +01:00
Marco Costalba 79232be02a Further tweak accurate pv
It is a non functional change, but because
we now skip copying of pv[] in SpNode, patch
has been tested for regression with 3 threads:

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 54668 W: 9873 L: 9809 D: 34986

No functional change.
2014-11-21 20:37:45 +01:00
hxim 0a1f54975f Fix some comments
No functional change.

Resolves #123
2014-11-19 06:39:17 +08:00
Gary Linscott bffe32f4fe Fix fen output for castling rights
This is a regression from 428962a

We have to cast to char here, otherwise the compiler
interprets it as an integer, and writes a number.

No functional change

Resolves #122
2014-11-19 06:37:59 +08:00
Marco Costalba 4aca11ae2a Codying style in accurate PV
This is the first of a patch series to
rearrange and simplify accurate PV.

In this patch there is simple coding
style and reformatting stuff.

Verified with fishtest it does not crash
with MAX_PLY = 8

No functional change.
2014-11-18 11:57:57 +01:00
Marco Costalba 1a939cd8c8 Fix a warning on Intel C++
warning #2259: non-pointer conversion from "int" to
"uint8_t={unsigned char}" may lose significant bits

No functional change.
2014-11-17 12:56:48 +01:00
sf-x d65c9a3262 Use PHASE_MIDGAME in game_phase()
No functional change

Resolves #117
2014-11-17 07:50:33 +08:00
Rodrigo Exterckötter Tjäder 99f2c1a2a6 Clear token before reading from input
Previously token would keep its value from the previous line when an empty
line was input, leading to unexpected behaviour.

No functional change

Resolves #119
2014-11-17 07:48:30 +08:00
lucasart 3b1f552b08 Half History Max
STC
LLR: 3.35 (-2.94,2.94) [-0.50,3.50]
Total: 17993 W: 3740 L: 3508 D: 10745

LTC
LLR: 3.25 (-2.94,2.94) [0.00,4.00]
Total: 21346 W: 3691 L: 3453 D: 14202

Bench: 7694316

Resolves #120
2014-11-17 07:04:58 +08:00
Marco Costalba 4840643fed Use DEPTH_MAX instead of MAX_PLY
When comparing to a Depth it is more
consistent to use DEPTH_MAX instead
of a int.

This is a subtle difference because we use
ply and depth almost interchangably in SF,
but they are different. FOr counting plies
makes ense to continue using ints, while
for Depth we have our specific enum.

This cleanly fixes a new Clang 3.5 warning:

No functional change.
2014-11-15 05:36:49 +01:00
Gary Linscott 4739037f96 100% accurate PV display
This gives SF accurate PVs, such that the evaluation of the leaf node in
the PV matches the score backed up to the root (99% of the time.
q-search will use the value stored in the hash table instead of the eval
value sometimes).

One drawback is that fail-high/low only get a minimal 2 move PV.

It doesn't add any additional overhead to the non-PV codepath except an
extra eight bytes to the SearchStack structure in multi-threaded
searches.

A core part of this is not pruning based on TT score in PV nodes. This
was measured as not being a regression at multiple TCs, except for one
exception, fast TC with huge hash, which is not realistic for longer
searches.

STC - 1 thread, 128 mb hash
ELO: 1.42 +-3.1 (95%) LOS: 81.9%
Total: 20000 W: 4078 L: 3996 D: 11926

STC - 3 thread, 128 mb hash
ELO: -3.60 +-2.9 (95%) LOS: 0.8%
Total: 20000 W: 3575 L: 3782 D: 12643

STC - 3 thread, 8 mb hash
ELO: 0.12 +-2.9 (95%) LOS: 53.3%
Total: 20000 W: 3654 L: 3647 D: 12699

LTC - 3 thread, 32mb hash
ELO: 2.29 +-2.0 (95%) LOS: 98.8%
Total: 35740 W: 5618 L: 5382 D: 24740

Bench: 6984058

Resolves #102
2014-11-12 16:16:33 -05:00
lucasart 234344500f Use quiet ttMove in qsearch() (7962287)
Daniel Jose reported that it was an elo gain in his engine:
http://www.talkchess.com/forum/viewtopic.php?t=54290

STC: Hash=16
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33067 W: 6670 L: 6571 D: 19826

LTC: Hash=64
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 41181 W: 7008 L: 6920 D: 27253

And another one to verify no regression with hash pressure:

STC: Hash=4
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 25085 W: 5059 L: 4991 D: 15035

Verified that qsearch does not explode after this patch (recapture threshold).

Bench 7962287

Resolves #112
2014-11-12 21:06:14 +00:00
Marco Costalba b777b17f6f Use if/else instead of goto
Real men jump/branch by hand...but
we prefer the humble way.

Moved also some uci info code where it
belongs, while there.

No functional change.

Resolves #110
2014-11-12 21:02:20 +00:00
Marco Costalba db4b8ee000 Use Depth instead of int in search
And make it more ONE_PLY value independent,
although we are not there yet.

No functional change.

Resolves #111
2014-11-12 21:00:16 +00:00
lucasart c6d45c60b5 Profile Build with Hash=16
16MB for 1" searches is more comensurate with the average use case.

And 16 is the default used by stockfish bench, so it makes sense to be
consistent, if only to have the same minimum memory requirement for using
SF and compiling it with PGO.

No functional change.

Resolves #109
2014-11-10 23:06:12 +00:00
lucasart d709a5f1c5 Fix bounds of FutilityMoveCounts
This is a left-over from ONE_PLY == 2.

No functional change.

Resolves #107
2014-11-09 20:13:56 +00:00
Marco Costalba 1b0df1ae14 Retire pvMove in search()
Now we can directly replace it with
the definition resulting in simpler
and possibly faster code because
PvNode is evaluated at compile time.

No functional change.
2014-11-09 20:36:28 +01:00
Marco Costalba 57fdfdedcf Assorted code-style triviality
No functional change.
2014-11-09 20:17:29 +01:00
Marco Costalba 6fb0a1bc40 Introduce distance() and unify some API
Original work by Lucas.

No functional change.
2014-11-09 10:27:04 +01:00
lucasart 8631b08d97 Codestyle massage Search::init()
* remove some erroneous comments, that were based on the ONE_PLY == 2.
* rename hd to d, because there's no more half-depth in SF.
* rescope variables into the for loops.
* reindent the for loops correctly.
* add a comment to explain the eval improving part (not so obvious to read
this code as array has 4 dimensions).

No functional change.
2014-11-08 10:56:51 -05:00
lucasart 3d2aab11d8 Be more optimistic in aspiration window
Be more optimistic wrt search instability, and set the unviolated bound
half window.

STC
LLR: 2.96 (-2.94,2.94) [-1.00,4.00]
Total: 16362 W: 3371 L: 3197 D: 9794

LTC
LLR: 2.94 (-2.94,2.94) [0.00,5.00]
Total: 21666 W: 3780 L: 3569 D: 14317

Bench: 6807896

Resolves #105
2014-11-08 10:47:56 -05:00
lucasart 7ebb872409 Prune ttMove like any other
This should reduce search inconsistencies, and doesn't seem to have a measurable ELO Impact:

STC with Hash=16
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 49264 W: 10076 L: 10007 D: 29181

LTC with Hash=64
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 82149 W: 14044 L: 14023 D: 54082

Plus an extra test, to make sure it doesn't regress with strong hash pressure:

STC with Hash=4
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 21498 W: 4327 L: 4246 D: 12925

Bench: 7302735

Resolves #100
2014-11-07 21:40:24 +00:00
lucasart 375797d51c Retire CACHE_LINE_ALIGNMENT
Speed tests showed no benefit.

No functional change.

Resolves #97
2014-11-07 14:27:04 -05:00
lucasart 8e98bd616e Apply King Safety later in the endgame
Idea is to apply king safety later in the endgame. Previously, we didn't
apply KS in a RR vs. Q ending for example, which causes poor play.
Now we calculate king attacks when the attacking side has a queen or more.

STC with 8moves_v3
LLR: 3.06 (-2.94,2.94) [0.00,4.00]
Total: 38481 W: 6228 L: 5952 D: 26301

LTC with 2moves_v1
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 51053 W: 8670 L: 8353 D: 34030

Bench: 7514010

Resolves #98
2014-11-06 13:01:47 -05:00
mstembera bc83515c9e Removing some superfluous extern declarations
No functional change.

Resolves #93
2014-11-05 21:17:19 +00:00
Marco Costalba bcbab19376 Assume UCI 'nodes' is int64_t instead of int
UCI specification is not clear on the size of
integers that are exchanged in the protocol, so
instead of a simple int, assume 'nodes' is a
int64_t because we need a bigger size to store
this value in many real cases, especialy with
very long searches.

No functional change.

Resolves #75
2014-11-05 21:11:05 +00:00
Marco Costalba d29a68f585 Rearrange check_time()
Remove an ugly workaround for a gcc
warning while there.

No functional change.
2014-11-05 21:09:21 +00:00
Ajith 0608d6aaec Add bonuses for each threat instead of max threat value.
Use SPSA tuned values for all threat bonuses

STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 6074 W: 1284 L: 1160 D: 3630

LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 13563 W: 2402 L: 2232 D: 8929

Bench: 6981908

Resolves #94
2014-11-04 23:50:54 +08:00
mstembera 2fd075d1ea Optimize TranspositionTable::store() and TranspositionTable::probe() for speed.
No functional change.

Resolves #85
2014-11-03 18:40:49 +00:00
lucasart d12378497c Do not assume that enum are signed
Clang 3.5 issues warning on constructs like: abs(f1 - f2). The thing is that
f1 and f2 are enum types, and the range given (all positive) allows the
compiler to choose an unsigned type (efficiency being one reason to prefer
unsigned arithmetic). If f1 < f2 are unsigned, then f1 - f2 wraps around zero
and the abs() becomes a no-op. It's the reinterpretation of the unsigned
result (large value) as a signed int that happens to give the correct result,
thanks to 2's complement. This is all tricky and dangerous!

In the spirit of the standard, we assume nothing on the signedness of enums,
and simply calculate the rank and file distances as:
- rank_dist(r1, r2) = r1 < r2 ? r2 - r1 : r1 - r2
- file_dist(f1, f2) = f1 < f2 ? f2 - f1 : f1 - f2
this logic can in fact be applied to any enum we may use, so for better
generality and to avoid code duplication, we use a template function diff()
here.

No functional change.

Resolves #95
2014-11-04 00:35:02 +08:00
lucasart 8ab9c2511a Cleanup MAX_PLY
This area has become obscure and tricky over the course of incremental
changes that did not respect the original's consistency and clarity. Now,
it's not clear why we use MAX_PLY = 120, or why we use MAX_PLY+6, among
other things.

This patch does the following:

* ID loop: depth ranges from 1 to MAX_PLY-1, and due to TT constraint (depth
must fit into an int8_t), MAX_PLY should be 128.

* stack[]: plies now range from 0 to MAX_PLY-1, hence stack[MAX_PLY+4],
because of the extra 2+2 padding elements (for ss-2 and ss+2). Document this
better, while we're at it.

* Enforce 0 <= ply < MAX_PLY:
  - stop condition is now ss->ply >= MAX_PLY and not ss->ply > MAX_PLY.
  - assert(ss->ply < MAX_PLY), before using ss+1 and ss+2.
  - as a result, we don't need the artificial MAX_PLY+6 range. Instead we
  can use MAX_PLY+4 and it's clear why (for ss-2 and ss+2).

* fix: extract_pv_from_tt() and insert_pv_in_tt() had no reason to use
MAX_PLY_PLUS_6, because the array is indexed by plies, so the range of
available plies applies (0..MAX_PLY before, and now 0..MAX_PLY-1).

Tested with debug compile, using MAX_PLY=16, and running bench at depth 17,
using 1 and 7 threads. No assert() fired. Feel free to submit to more severe
crash-tests, if you can think of any.

No functional change.
2014-11-03 23:36:24 +08:00
Marco Costalba fc0733087a Restore std::dec after std::hex
Code is leaking a std::hex, and causes subsequent
sync_cout output to be in hexadecimal.

Spotted by Lucas

No functional change.
2014-11-02 08:03:52 +01:00
Marco Costalba 42a20920e5 Retire ScalePawnSpan[]
Obscure, undocmented and misnamed array. Replace with
the direct formula: it is much more clear what the
code does.

No functional change.

Resolves #90
2014-11-01 22:10:25 +00:00
Marco Costalba 79fa72f392 Merge pull request #89 from official-stockfish/pull_no_pretty
Prefer operator<<() to pretty()

No functional change.
2014-11-01 22:24:33 +01:00
Marco Costalba d3091971b7 Retire PawnsFileSpan
It is useless. Tested as no regression:

STC
LLR: 4.06 (-2.94,2.94) [-3.00,1.00]
Total: 140718 W: 28527 L: 28568 D: 83623

LTC
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 60034 W: 10359 L: 10303 D: 39372

bench: 6564212

Resolves #88
2014-11-01 20:50:52 +00:00
lucasart d9caede324 Correctly describe POPCNT compile
SSE4.2 has nothing to do with POPCNT. We must dispell this myth, because
Stockfish is a reference and many will copy this mistake if they see it in Stockfish:
* SSE is an SIMD instruction set, relative to vectorization (using special 128-bit registers).
* POPCNT/LZCNT work on normal registers (eg. AL, AX, EAX, RAX).

The confusion comes from the fact that, in the Intel product line, it just
so happens that SSE4.2 and POPCNT/LZCNT came out at the same time. But this
is not true for AMD. For example, all AMD Pheniom II have SSE3 but no
POPCNT/LZCNT, and that is why the modern compile uses  -msse3 -popcnt and not -msse4.2.

No functional change.

Resolves #86
2014-11-01 20:43:57 +00:00
lucasart 8a7876d48d Consistent use of anonymous namespace
Objects that are only accessible at file-scope should be put in the anonymous namespace.
This is what the  C++ standard recommends, rather than using static, which is really C-style and results in static linkage.

Stockfish already does this throughout the code. So let's weed out the few exceptions,
because... they have no reason to be exceptional.

No functional change.

Resolves #84
2014-11-01 20:35:10 +00:00
lucasart 2ee1250294 Remove dead code
No functional change.

Closes #87
2014-11-01 20:16:29 +00:00
Marco Costalba 5644e14d0e Prefer operator<<() to pretty()
It is more idiomatic, we didn't used it
in the past because Position::pretty(Move)
had a calling argument, but now we can.

As an added benefit, we avoid a lot of string
copies in the process because now we avoid
std::ostringstream ss.

No functional change.
2014-11-01 19:02:35 +01:00
Marco Costalba d07a875398 Merge pull request #82 from official-stockfish/clean_up_bishop_code
Code style clean-up

No functional change.
2014-11-01 18:05:03 +01:00
Marco Costalba 460892382a Code style clean-up
This piece of code needs some love.

No functional change.
2014-10-30 12:32:43 +01:00
mstembera 5605cc7684 max_piece_type cleanup, and slight speed increase.
No functional change.

Resolves #81
2014-10-28 22:23:01 +08:00
Joona Kiiski 7d42752158 A small clean up of previous patch suggested by Marco
No functional change
2014-10-27 20:26:12 +00:00
Joona Kiiski fefb27bab4 Speed up max_piece_type()
Write code in the way that allows compiler to perform loop unrolling.

My measurement (32 cycles each):

Orig:

Time (Mean: 2466.59375, Trimmed mean: 2464.25, Std: 12.6869487803348)
Nodes (Mean: 4294458, Trimmed mean: 4294458, Std: 0)
Speed (Mean: 1741.09247987678, Trimmed mean: 1742.72879715475, Std: 8.93612608292678)

Time (Mean: 2470.15625, Trimmed mean: 2468.75, Std: 12.7484581610433)
Nodes (Mean: 4294458, Trimmed mean: 4294458, Std: 0)
Speed (Mean: 1738.58176151341, Trimmed mean: 1739.54618465403, Std: 8.95585822316946)

Mod:

Time (Mean: 2449.90625, Trimmed mean: 2445.9375, Std: 12.1000116635508)
Nodes (Mean: 4294458, Trimmed mean: 4294458, Std: 0)
Speed (Mean: 1752.94829372932, Trimmed mean: 1755.75934908231, Std: 8.61478453124504)

Time (Mean: 2442.78125, Trimmed mean: 2441.1875, Std: 8.17839157228837)
Nodes (Mean: 4294458, Trimmed mean: 4294458, Std: 0)
Speed (Mean: 1758.03872783803, Trimmed mean: 1759.16825356261, Std: 5.81131316346191)

No functional change
2014-10-27 20:25:57 +00:00
snicolet 7de40076ac Tune PawnsFileSpan
Passed the following SPRT tests:

STC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 24428 W: 5056 L: 4880 D: 14492

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 26590 W: 4715 L: 4472 D: 17403

Bench: 6615949

Resolves #78
2014-10-27 11:48:44 +00:00
Pascal Romaret f1359845de Improve compatibility with Shredder Classic GUI
This commit fixes two issues:

1) Don't print PVs after the search has been interrupted

    This solves the "mate 0 upperbound" scores that sometimes
    creep up when a multi-PV analysis gets interrupted with
    the `stop` command.

2) Print multipv before score

    Shredder Classic fails to identify the main PV
    (the one with multipv 1) if `score` comes first.
    This leads to an eval graph that doesn't reflect
    the scores actually reported by Stockfish when
    doing a multiPV analysis.

No functional change

Closes #76
2014-10-27 11:07:35 +00:00
Marco Costalba 5ab55827b8 Fix an obscure gcc warning
warning: narrowing conversion from ‘int’ to ‘char’ inside { }
is ill-formed in C++11 [-Wnarrowing]

When pedantic meets esoteric!

No functional change.
2014-10-26 19:40:20 +00:00
Marco Costalba 9ba391c5cb Retire notation.cpp
Now we can finally retire notation.cpp
and move UCI helpers under uci.cpp

No functional change.
2014-10-26 19:40:13 +00:00
Marco Costalba 2469daebb1 Final UCI helpers renaming
To reflect new changes, specifically that
now are all under UCI namespace.

No functional change.
2014-10-26 19:40:04 +00:00
Marco Costalba aa60c80ade Retire notation.h
And move the few remaining content
under UCI namespace where they belong.

No functional change.
2014-10-26 19:39:56 +00:00
Marco Costalba 5cbcff55cc Rename ucioption.h to uci.h
We are going to add all UCI related
functions here, so first rename it
to a more proper name.

No functional change.
2014-10-26 19:39:46 +00:00
Marco Costalba 428962a2e8 Retire to_char() helpers
Remove some useless wrappers and make
the conversion explicit and starightforward.

No functional change.
2014-10-26 19:39:37 +00:00
Marco Costalba 4f6b1bf3be Reformat max_threat()
Helper function should just know how to find the
biggest piece type in a bitboard. All the threat
logic and data shoud be in evaluate_threats().

This nicely separates the scope of the two functions
in a more consistent way and simplifies the code.

No functional change.
2014-10-26 09:12:56 +00:00
snicolet 8d62ece945 Calculate maximum threat for hanging pieces
Use the max_threat() helper function to estimate more precisely the
best hanging piece threat.  Also retunes the Threat array using SPSA.

STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 7598 W: 1596 L: 1468 D: 4534

LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 7896 W: 1495 L: 1350 D: 5051

Bench: 6816504

Resolves #73
2014-10-24 01:10:11 +08:00
Marco Costalba 480682b670 Document why initing eval tables
Instead of hard-code the weights in a big table,
we prefer to calculate them out of few parameters
at startup. This allows to keep low the number of
independent parameters and hence is good for tuning
and for a better insight in the meaning of the numbers.

No functional change.
2014-10-15 19:36:22 +01:00
Marco Costalba 89d9db2979 Rename Tracing methods
Easier to understand and more in line with
standard Trace classes naming like:

http://msdn.microsoft.com/en-us/library/system.diagnostics.trace.aspx

No functional change.
2014-10-12 20:35:19 +01:00
Marco Costalba 907f912463 Account for Tempo in do_evaluate()
This is more correct because we let evaluate()
to be a pure dispatcher and also now evaluate
and tracing outputs are consistent.

No functional change.
2014-10-12 20:35:07 +01:00
Marco Costalba c97b702f4d Fix some warnings with Intel C++ compiler
No functional change.
2014-10-12 20:34:54 +01:00
lucasart 91de6b0f37 Further streamline connected pawn evaluation
Make even more clear what are the terms that
contribute to evaluate connected pawns, and
completely separate them from the weights
that are now fully looked up in a table.

For future tuning makes sense to init the table with
a formula instead of hard-code it. This allows to
reduce problem space cardinality and makes tuning
easier.

And fix a MSVC warning while there:
warning C4804: '>>' : unsafe use of type 'bool' in operation

No functional change.
2014-10-12 20:03:49 +01:00
lucasart 069073ea68 Merge Connected and Candidate
These two notions are very correlated. Since connected has the most
generality, it makes sense to generalize it to encompass what is
covered by candidate.

STC:
LLR: 4.03 (-2.94,2.94) [-3.00,1.00]
Total: 11970 W: 2577 L: 2379 D: 7014

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 13194 W: 2389 L: 2255 D: 8550

bench 7328585
2014-10-07 07:42:17 +01:00
joergoster 15e2191111 Remove the now redundant TT prefetch call from Position::do_move.
Tested for no regression and passed both
STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 48250 W: 9757 L: 9686 D: 28807

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 51412 W: 8887 L: 8816 D: 33709

No functional change.

Resolves #66
2014-10-06 23:59:56 +08:00
Luca Brivio 8a9b9ec96a Simplify futility move count array formula
No functional change
2014-10-05 12:49:18 +01:00
Marco Costalba 74829342ef Reformat and rename hash_after_move()
Align to standard coding style and properly use
enum types. Rename while there.

No functional change.
2014-10-04 10:36:29 +01:00
Ajith 4b926f227d Extend King Threats to all pieces (other than pawns).
STC
LLR: 2.99 (-2.94,2.94) [-1.50,4.50]
Total: 20559 W: 4261 L: 4095 D: 12203

LTC
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 75232 W: 13097 L: 12696 D: 49439

Bench: 7543790

Resolves #63
2014-10-04 03:54:12 +08:00
joergoster 82d065b011 Speculative prefetch
Idea by Peter Oesterlund.
Implemented and tested by Joerg Oester

STC 3 threads
ELO: 3.19 +-2.1 (95%) LOS: 99.9%
Total: 40000 W: 7576 L: 7209 D: 25215

LTC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 22026 W: 3829 L: 3619 D: 14578

STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 7291 W: 1531 L: 1404 D: 4356

No functional change

Resolves #61
2014-10-02 22:19:14 +01:00
lucasart e60cdca9b0 Convert TT depth to int8_t
Now that half plies have been removed from the engine, we can encode
TT depth into an int8_t.

Range is -128 to +127, so it goes still further than the previous
limit of 121 plies (with ONE_PLY == 2 where depth - DEPTH_NONE was
encoded as an uint8_t).

No functional change.

Resolved #60
2014-10-01 20:51:32 +01:00
Marco Costalba a1b62d68ec Trivial code style fixes
Mainly to sync mine and official repo.

No functional change.
2014-09-30 09:05:20 +02:00
Marco Costalba 222f59b9c1 Move ONE_PLY to be 1 instead of 2: search()
Now that half-plies are no more used we can simplify
the code assuming that ONE_PLY is 1 and no more 2.

Verified with a SMP test:
LLR: 2.95 (-2.94,2.94) [-4.50,0.00]
Total: 8926 W: 1712 L: 1607 D: 5607

No functional change.
2014-09-29 15:17:12 +02:00
lucasart 27a1877299 Clean up VALUE_KNOWN_WIN conditions
A patch (+ some extra changes) passed with:

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14575 W: 3101 L: 2967 D: 8507

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 42579 W: 7580 L: 7496 D: 27503

Bench: 6545733

Resolves #52
2014-09-28 17:45:49 +01:00
Uri Blass ea9c424bba Remove use of half-ply reductions from LMR, Null-move, IID and
Singular extensions.

STC:
ELO: 3.80 +-3.1 (95%) LOS: 99.2%
Total: 19727 W: 4190 L: 3974 D: 11563

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7647 W: 1356 L: 1214 D: 5077

Bench: 6545733

Resolves #55
2014-09-28 04:33:28 +08:00
Joona Kiiski 7ed15af371 Cap evaluation based null move extra reduction to three plies
It's a zero elo patch, and a reasonable safeguard against uncontrolled extreme reductions.

STC:
ELO: -0.08 +-2.0 (95%) LOS: 46.9%
Total: 40000 W: 6728 L: 6737 D: 26535

LTC:
ELO: -0.14 +-1.8 (95%) LOS: 44.0%
Total: 40000 W: 5557 L: 5573 D: 28870

Bench: 7201830
2014-09-25 20:42:25 +01:00
Uri Blass d6613b7589 Change history reduction in LMR to be a full ply.
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 9829 W: 2142 L: 1998 D: 5689

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 27162 W: 4802 L: 4692 D: 17668

Bench: 7284120

Resolves #53
2014-09-26 02:22:39 +08:00
uriblass fed3e752ae Evaluate king safety when no queen is present.
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 16657 W: 3547 L: 3391 D: 9719

LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 31258 W: 5664 L: 5403 D: 20191

Bench: 8331165

Resolves #51
2014-09-22 01:32:47 +08:00
Gary Linscott 766fb9c67d Fix spacing.
No functional change.
2014-09-22 01:27:34 +08:00
Marco Costalba b652720903 Rearrange evaluation constants definitions
Make them more uniform and consistent.

No functional change.

Conflicts:
	src/evaluate.cpp
2014-09-21 08:26:25 +08:00
Marco Costalba 4d0a6c5a6f Rename time variable to reflect UCI parameters
On top of previous patch, rename time variables to
reflect the simplification of UCI parameters.

It is more correct to use as varibales directly the
corresponding UCI option, without intorducing redundant
intermediate variables.

This allows also to simplify the code.

No functional change.
2014-09-21 08:23:12 +08:00
mbootsector 2eec710318 King-pawn threat bonus for endgames.
STC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 10224 W: 1765 L: 1638 D: 6821

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 6923 W: 1027 L: 899 D: 4997

bench 7818100

Resolves #49
2014-09-19 22:36:05 +08:00
Marco Costalba aa10d0aaa6 Rename "Contempt Factor" to "Contempt"
Suggested by Ronald and Lucas on talkchess.

No functional change.

Conflicts:
	src/ucioption.cpp
2014-09-16 21:14:00 +01:00
Marco Costalba ff480bdb83 Retire struct Log
No more used now that we have removed
"Write Search Log" UCI option.

No functional change.
2014-09-16 21:13:50 +01:00
lucasart b66552fc27 Simplify Time Management UCI options
No functional change
2014-09-15 20:24:51 +01:00
Marco Costalba cd065dd584 Small tweak to idle_loop()
In case of a succesful late join we set again
'searching' flag, so we can restart search
immediately without an useless lock/unlock
cycle.

No functional change.
2014-09-04 20:19:03 +01:00
Joona Kiiski 95b24083fb Simplify idle_loop()
No functional change
2014-08-30 21:03:41 +01:00
lucasart 877313a413 Retire Search Log
No functional change

Bench: 7461881
2014-08-24 16:28:51 +01:00
Marco Costalba 8b8885ab07 Fix perft 1
Compute correct number of moves for this corner case.

A smal bug crept in after recent perft rework.

No functional change.
2014-08-10 07:36:22 +08:00
lucasart a903ed07e0 Retire move_to_san()
Now "Write Search Log" will pring moves in UCI format, consistent with all the rest. This functionality is
not aimed at end-users anyway. It's hardly useful at all, in fact. Also, pretty-printing SAN moves is
something that better belongs in the GUI than in the engine.

No functional change.
2014-08-10 07:28:00 +08:00
lucasart 880e3cd7c8 Move to_char() and to_string() to notation
Where they better belong.

Also, this removes '#include <string>' from types.h, which reduces the amount of code to compile (every
translation unit includes types.h).

No functional change.
2014-08-09 13:25:05 +08:00
Marco Costalba a67c22611a Rework perft implementation
Unify various perft functions and move all the code
to search.cpp.

Avoid perft implementation to be splitted between
benchmark.cpp (where it has no reason to be) and
search.cpp

No functional and no speed change (tested).
2014-08-09 13:00:59 +08:00
lucasart 2efeded6e3 Write perft(N-1) into cout
So that one can redirect cout to /dev/null and only print print cerr in the terminal (for more accurate speed
tests).

Suggested by Marco.

No functional change.
2014-08-07 21:15:05 +08:00
lucasart 6044f25d71 Fix Hash in bench
The compiler tries to cast Options["Hash"] into a string, using:

Option::operator std::string() const {
  assert(type == "string");
  return currentValue;
}

And, as expected, the assert() fails.

std::to_string() would be the right solution, but it's C++11. And using a stringstream is too much code to
achieve so little. Let's keep it the way it was: hardcoded (ie. default hash defined in two places).

No functional change.
2014-08-07 18:51:07 +08:00
joergoster 9da015517c Remove insufficient material rule
The eval already returns zero in KK, KBK, KNK (see material.cpp). The difference is:
- we lose the "TB pruning" benefit of the draw rule (ie. search goes on even if eval is zero)
- we gain some speed by removing a useless test from the hot path

STC:
LLR: 0.05 (-2.94,2.94) [-3.00,1.00]
Total: 128000 W: 21357 L: 21560 D: 85083

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 33023 W: 4613 L: 4509 D: 23901

bench 7461881
2014-08-06 18:42:10 +08:00
lucasart 888a1d3445 Remove useless code in Position::pretty()
First, remove some dead code (function never called with a Move argument).

Then, remove printing of legal moves, which does not belong here. Let's keep commands orthogonal and minimal:
- the "d" command should display the board, nothing more, or less.
- "perft 1" will display the list of legal moves.

No functional change.
2014-08-06 07:04:27 +08:00
lucasart 8b88ca9017 Reduce minimum memory requirement by 16MB
Stockfish allocates the default hash (32MB) in main(), before entering UCI::loop(). If there is not enough
memory, the program will crash even before UCI::loop() is entered and the GUI is given a change to specify a
lower Hash value.

This defective design could be resolved by doing a lazy allocation upon "isready" command, as the UCI protocol
guarantees that "isready" will be sent at least once before any search. But it's a bit cumbersome when using
Stockfish "manually" to have to remember to type "isready" everytime.

So leave the current design, but reduce the default hash to 16MB instread of 32MB. In order to perform such
quick searches (depth=13), there is no reason to use so much Hash anyway. Another benefit is to introduce a
bit of hash pressure in bench, which increases chances to detect rare bugs related to TT replacement, for
example.

This is not a functional change, although it obviously changes the bench.

bench 7461879
2014-08-05 11:42:48 +08:00
lucasart 94fe366779 Default Hash defined in a single place
Instead of defining it both in ucioption.cpp and benchmark.cpp. Obviously changing the default Hash will
change the bench as a result.

No functional change.
2014-08-05 11:40:25 +08:00
lucasart 535f70088e Retire divide command
The main purpose of perft is to help debugging. But without the breakdown in sum of perft(N-1), it is a
completely useless debugging tool.

So perft now displays the breakdown, and divide is therefore removed.

No functional change.
2014-08-04 13:54:09 +08:00
Marco Costalba 1b69910865 Ensure printing UCI info in multi-pv case
After commit 94b1bbb68b, in case available root moves are less than multiPV, we
could never reach condition:

PVIdx + 1 == multiPV

and as a consequence UCI output is not printed.

Fixed suggested by Joerg Oster.

No functional change.
2014-08-02 22:16:45 +08:00
Oskar Werkelin Ahlin 29451de874 Correct bench timing
No functional change.
2014-08-01 19:32:53 +08:00
lucasart 5adc678628 Document Threat[] indices
From Marco's repo.

No functional change.
2014-07-30 07:06:48 +08:00
Marco Costalba f2053ba19f Fix a warning with MSVC 2010
Warning C4267: 'argument' : conversion from 'size_t' to 'int', possible loss of data

No functional change.
2014-07-29 06:53:09 +08:00
David Zar 68171ecaca Small code style reformatting
No functional change.
2014-07-29 06:52:32 +08:00
lucasart 6b82a234d0 Aspiration: widen slower
STC:
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 182323 W: 30664 L: 30234 D: 121425

LTC:
LLR: 3.51 (-2.94,2.94) [0.00,4.00]
Total: 59841 W: 8345 L: 8006 D: 43490

bench 7962536
2014-07-27 09:18:04 +08:00
David Zar 4758fd31b1 Outpost tuning
double mg bonus and half eg bonus.

STC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 18142 W: 3094 L: 2948 D: 12100

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 83561 W: 11706 L: 11329 D: 60526

bench 7831429
2014-07-23 07:05:10 +08:00
lucasart 7f68fc611c Revert Contempt = 20
Despite being neutral at STC, it turned out to be regressive at LTC:

40k games at LTC with Hash=8
ELO: -2.06 +-1.9 (95%) LOS: 1.4%
Total: 39720 W: 5740 L: 5976 D: 28004

40k games at LTC with Hash=128
ELO: -2.69 +-1.9 (95%) LOS: 0.2%
Total: 39149 W: 5702 L: 6005 D: 27442

bench 7477963
2014-07-20 08:50:54 +08:00
Marco Costalba 896bd917f8 Simplify evaluate_passed_pawns
From a suggestion by David Zar.

No functional change.
2014-07-15 20:59:32 +08:00
lucasart 67a5e1ecf9 Contempt = 20
Also raise the admissible bounds to (-100,100), as there is no reason to prevent users from using high
values if they want to.

Does not regress in self play:
ELO: 0.10 +-2.0 (95%) LOS: 53.7%
Total: 40000 W: 7084 L: 7073 D: 25843

master vs SF 3
ELO: 182.86 +-2.7 (95%) LOS: 100.0%
Total: 40000 W: 21843 L: 2541 D: 15616

Contempt = 20 vs SF 3
ELO: 189.25 +-2.8 (95%) LOS: 100.0%
Total: 40000 W: 22721 L: 2859 D: 14420

Diff is therefore 6.4 +/- 3.9 elo against a 180-190 elo weaker engine, which is significantly positive,
as expected. This elo difference is likely understated, because of FishTest aggressive draw adjudication
though.

We could push Contempt further, but after 20cp, it would get in the way of FishTest draw adjudication
rule, and is likely to reduce the testing throughput as a result.

bench 8198667
2014-07-15 07:14:58 +08:00
David Zar a9b8e8b931 Bonus passed pawn blocked by our pieces
passed STC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 63965 W: 10950 L: 10692 D: 42323

and LTC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 19237 W: 2740 L: 2562 D: 13935

bench: 7477963
2014-07-14 21:19:07 +08:00
Marco Costalba 94b1bbb68b Small reformat to Skill class
Steamline a bit the implementation of
skill levels. As a side effect we can
retire MultiPV global and use a local
variable instead.

No functional change.
2014-07-11 18:08:27 +08:00
lucasart f133f61e3f Remove useless condition
Small simplification, suggested by Uri Blass.

passed STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25839 W: 4464 L: 4351 D: 17024

and LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 220039 W: 29981 L: 30131 D: 159927

No functional change.
2014-07-10 18:25:20 +08:00
Joona Kiiski eb50793cff Retire FakeSplit
- Currently broken
    - Never been really useful
    - Does not work well with new splitting model

Verified for no regression at STC with 3 threads:
LLR: 2.96 (-2.94,2.94) [-6.00,0.00]
Total: 81905 W: 12122 L: 12381 D: 57402

No functional change
2014-07-09 07:19:06 +08:00
Marco Costalba 9b30913996 Avoid 'double assigments' tricks
Bitboard init code is already noteasy to follow,
so don't make it even harder using 'smart' code.

Also reindent a while loop in standard way.

No functional change.
2014-07-07 21:53:25 +08:00
joergoster ac7780bd35 Tune trapped rook penalty
Passed STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 15687 W: 3352 L: 3199 D: 9136

and LTC (parameter tweaks)
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 27983 W: 5046 L: 4797 D: 18140

bench: 8330705
2014-07-03 19:38:54 +08:00
lucasart c192b692cf size_t cast in TranspositionTable::first_entry()
32-bit truncation would make this function bogus when clusterCount >= 2^33 (ie. Hash >= 256 GB).

No function change.
2014-07-03 18:23:56 +08:00
lucasart 85b08ce3ad Use compiler intrinsic instead of assembly for popcnt
This time, do not break compatibility with some AMD machines that have SSE3 and popcnt, but do
not have SSE4.2.

No functional change.
2014-07-03 18:22:53 +08:00
Gary Linscott b9a88da4ab Revert "Use compiler intrinsic instead of assembly for popcnt"
This reverts commit a69f1d7c20.
2014-07-01 17:01:54 -04:00
lucasart a69f1d7c20 Use compiler intrinsic instead of assembly for popcnt
No functional change.
2014-07-01 20:50:33 +08:00
lucasart 24ba204931 Raise max Hash to 1TB
And use size_t where appropriate, as suggested on FishCooking.

No functional change.
2014-07-01 18:37:18 +08:00
Ajith 6b354305e1 Add bonuses for Minors attacking enemy pieces(except pawns) even when they are protected by enemy pawns.
Patch passed STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8206 W: 1426 L: 1304 D: 5476

and LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 19534 W: 2821 L: 2640 D: 14073

Bench: 9942172
2014-06-30 10:55:10 -04:00
joergoster ffedfa3354 Fix Singular extension condition to handle mate scores
With Eelco's patch "Don't special case for abs(beta) >= VALUE_MATE_IN_MAX_PLY" condition "abs(ttValue) < VALUE_KNOWN_WIN" has been removed from singular extension search, and condition "abs(beta) < VALUE_KNOWN_WIN" was added to the SingularExtensionNode definition.
This might lead to problems, especially in positions, where a mate is due.
For example, this position 5rk1/4K1pp/8/5PPP/8/8/8/1R6 w - - 12 1 triggers an assert.
stockfish: search.cpp:434: Value {anonymous}::search(Position&, Search::Stack*, Value, Value, Depth, bool) [with {anonymous}::NodeType NT = (<unnamed>::NodeType)2u; bool SpNode = false]: Assertion `-VALUE_INFINITE <= alpha && alpha < beta && beta <= VALUE_INFINITE' failed.

So let's re-insert the removed condition.
First spotted by Uri Blass, fix by me.

Bench: 8759675
2014-06-29 20:17:40 +01:00
Joseph R. Prostko e4e423bb05 Change the install prefix for Haiku
* /boot/common was removed from Haiku
* The equivalent path now that package management has been implemented is /boot/system/non-packaged

No functional change

Bench: 8759681
2014-06-29 15:15:10 +01:00
Ron Britvich ccd823a4ff Pack 3 TT entries in 32 bytes cluster
Idea from Ron Britvich

Code reworked by Marco Costalba and Joona Kiiski

Bench: 8095369

Resolves #3
Resolves #10
2014-06-28 14:06:32 -04:00
Gary Linscott 7ff865b924 Merge pull request #9 from glinscott/pawnspan
Scale down endgames with pawns on one or two adjacent files

Passed STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 16081 W: 2745 L: 2604 D: 10732

Passed LTC
LLR: 2.95 (-2.94,2.94) [1.00,6.00]
Total: 123832 W: 17292 L: 16584 D: 89956

128k games to measure ELO at 15+0.05:
ELO: 2.07 +-1.1 (95%) LOS: 100.0%
Total: 128000 W: 21632 L: 20869 D: 85499

New bench: 8028792
2014-06-26 12:21:06 -07:00
Gary Linscott ec627911f0 A bit more cleanup 2014-06-26 15:20:30 -04:00
Gary Linscott 22ffb588e5 Merge pull request #8 from glinscott/revert_ce1c260
Revert "Check for an available slave early on"
2014-06-26 08:20:45 -07:00
Gary Linscott e6a8d03dd8 Revert "Check for an available slave early on"
This reverts commit ce1c260ea9.
2014-06-26 11:16:36 -04:00
Gary Linscott 747e035c7d Original version of shane's patch 2014-06-25 16:45:14 -04:00
shane31 6c9f4cf36f Scale down endgames with pawns on one or two adjacent files 2014-06-25 16:01:00 -04:00
Gary Linscott ab580106fd Merge pull request #5 from glinscott/authors
Add AUTHORS
2014-06-23 09:14:03 -07:00
Marco Costalba aa23eb4440 Some reformatting in pawns.cpp
No functional change.
2014-06-21 15:08:31 +02:00
Marco Costalba e8baf2b772 Ensure ttValue != VALUE_NONE in singular extension search
The assert:

  assert(ttValue != VALUE_NONE);

Could fire for multiple reasons (although is very rare),
for instance after an IID we can have ttMove != MOVE_NONE
while ttValue is still set at VALUE_NONE.

But not only this, actually SMP is a source of corrupted
ttValue and anyhow we can detect the condition:

 ttMove != MOVE_NONE && ttValue == VALUE_NONE

even north of IID.

Reported by Ronald de Man.

It is so rare that bench didn't change.

bench: 7710548
2014-06-21 13:07:29 +02:00
Gary Linscott 58bb23d0c3 Remove some duplicates 2014-06-21 00:49:49 -04:00
Gary Linscott bbacff5e01 Add AUTHORS 2014-06-20 21:35:09 -04:00
Marco Costalba 43efd7fad7 Better value clipping in game_phase()
No functional change.
2014-06-21 00:05:14 +02:00
Marco Costalba 3b315c9ada Move game_phase() to Position
It seems a more natural to place this
function there.

No functional change.
2014-06-20 23:44:46 +02:00
Marco Costalba f7926ea41e Small renaming in material weights
Also dropped some temporary variable: compiler
is more than able to push on stack temp values
by itself (verified).

No functional change.
2014-06-20 23:15:31 +02:00
Marco Costalba 264c8637a3 Simplify a condition in is_KXK()
No functional change.
2014-06-19 15:44:42 +02:00
Eelco de Groot 55a3e0af8d Don't special case for abs(beta) >= VALUE_MATE_IN_MAX_PLY
Remove from the search this special case and apply
null search and razoring also in mate positions.

Tested in no-regression mode and passed both

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 65431 W: 10860 L: 10810 D: 43761

and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 34928 W: 4814 L: 4713 D: 25401

This patch kicks in only in mate positions and in
these cases it seems beneficial in finding mates
faster as Yery Spark measured on the Chest mate suite:

Total number of positions 6425
Fixed nodes 200K per position

master: 1049
new:    1154

And also the 5446 'hard' positions again with 2000K nodes
(those not found by both engines in 200K nodes):

master: 1069
new:    1395

bench: 7710548
2014-06-16 21:50:14 +02:00
Leonid Pechenik 66c93245e0 Simplify unstoppable pawns
Tested in no-regression mode and passed both

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17919 W: 3103 L: 2978 D: 11838

and LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 31460 W: 4414 L: 4308 D: 22738

bench: 7709279
2014-06-16 12:39:50 +02:00
Marco Costalba 73ca93f3c0 Remove -ansi flag for Clang
It seems this flag is only for gcc and
yields a warning under OSX Mavericks:

clang: warning: argument unused during compilation: '-ansi'

No functional change.
2014-06-15 10:31:57 +02:00
Marco Costalba c6fc51c5ea Fix a warning with MSVC in 'analyze' mode
Here MSVC is worried that

StepAttacksBB[PAWN][psq]

could overflow, so change psq initialization
to clarify psq is always less than 64.

No functional change.
2014-06-14 12:46:58 +02:00
Marco Costalba 2cb4c7052e Triviality in UCI::loop
Code style paranoid in action here :-)

No functional change.
2014-06-14 12:26:08 +02:00
Marco Costalba 3c1201c20c Factor out pawn attacks in 'lever'
Improves readibility and possibly speed.

No functional change.
2014-06-14 12:20:54 +02:00
Ajith 84dabe5982 Simplify pawn threats and merge into ThreatenedByPawn[]
Tested in no-regression mode,

passed STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14477 W: 2493 L: 2362 D: 9622

and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 51964 W: 7091 L: 7013 D: 37860

bench: 7875814
2014-06-12 08:32:16 +09:00
Reuven Peleg e10276f45d Tweak outpost name
This name is more accurate, since that function evaluates only one
outpost in every call.
No functional change.
2014-06-11 23:42:47 +09:00
Reuven Peleg 2312c26722 Add bonus for advanced lever
Passed both STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 15413 W: 2670 L: 2530 D: 10213

And LTC:
LLR: 2.95 (-2.94,2.94) [1.00,6.00]
Total: 66908 W: 9398 L: 8960 D: 48550

Bench : 7859385
2014-06-11 23:38:25 +09:00
Marco Costalba b6cd89aeaf Small renaming in Tracing
No functional change.
2014-06-09 05:30:18 +09:00
Marco Costalba ce1c260ea9 Check for an available slave early on
Don't take the split lock if we don't have
available slaves (about 30-40% of times).

This new condition allows to retire the now
redundant one on number of threads.

No functional change.
2014-06-09 04:56:31 +09:00
Marco Costalba 4d30126e4b Use unsigned char as argument of std::isspace
Although signature allows an int:

int isspace( int ch );

The behavior is undefined if the value of ch is
not representable as unsigned char and is not
equal to EOF.

See

http://en.cppreference.com/w/cpp/string/byte/isspace
http://www.greenend.org.uk/rjk/tech/cfu.html

This is really a tricky corner case of C standard!

Spotted and reported by Ron Britvich.

No functional change.
2014-06-07 01:21:47 +02:00
Marco Costalba 7f56d2949d Avoid to use nullChild
Use instead:

(ss-1)->currentMove == MOVE_NULL

No functional change.
2014-06-06 11:12:05 +02:00
Marco Costalba d10ae90dea Remove 'update gains' hack
Use (move != MOVE_NONE) condition to
filtering out updating gains at root.

bench: 8454456
2014-06-06 11:10:40 +02:00
Marco Costalba 2f75639485 Temporary revert previous patch
Split previous patch in 2 steps: first remove
the MOVE_NULL hack, then retire nullChild.

The first step is a prerequisite
for second one and affects bench.

The second step (next patch) just removes nullChild
without affecting bench.

bench: 8205159
2014-06-06 11:08:35 +02:00
Marco Costalba ad1167c482 Avoid to use nullChild
Should be a non functional change, but
for some reason bench is changed.

bench: 8454456
2014-06-06 10:07:48 +02:00
Marco Costalba 08753771fc Move Tempo to evaluation
No functional change.
2014-06-06 09:40:01 +02:00
Marco Costalba 69ac45d903 Revert "Score extractors"
Are broken for big-endian case and
I have verified with MSVC 2013 Premium
bench is correct and there is no
miscompilation, so the main reason
to change the original code drops.

No functional change.
2014-06-05 23:55:18 +02:00
Marco Costalba 323a006666 Fix a warning with MSVC Premium 2013
Reported by Ron Britvich.

No functional change.
2014-06-05 23:46:58 +02:00
pellanda 16e170d105 Update Readme.md
Retire line about polyglot.ini and
book as they are not used anymore.

No functional change.
2014-06-04 22:31:59 +02:00
Lucas Braesch adeded29fb Symmetric King Safety: take 2
Another attempt at retiring current asymmetric
king evaluation and use a much simpler symmetric
one. As a good side effect we can avoid recalculating
eval after a null move.

Tested in no-regression mode and passed

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21580 W: 3752 L: 3632 D: 14196

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 18253 W: 2593 L: 2469 D: 13191

And a LTC regression test against SF DD to
verify we don't have regression against
weaker engines due to some kind of 'contempt'
effect:

ELO: 54.69 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 11072 L: 4827 D: 24101

bench: 8205159
2014-06-03 20:01:36 +02:00
Marco Costalba 187a9fe5e7 Use see() instead of see_sign() in previous patch
Before it was working by accident in case of
see_sign() and failing with see() due to how
castle moves are coded (king captures the rook).

Better to explicitly filter out castling moves
and use see() without any surprise/trick.

No functional case.
2014-06-03 12:38:58 +02:00
kinderchocolate 6f48367094 Add some const qualifier
No functional change.
2014-06-03 11:43:52 +02:00
Reuven Peleg 83a574ff27 Decrease reduction for moves that escape a capture
Passed both STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 10094 W: 1833 L: 1704 D: 6557

and LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 27738 W: 4147 L: 3928 D: 19663

bench : 8599236
2014-06-03 11:33:19 +02:00
Marco Costalba f4dcec0b94 Retire Polyglot Book management
Book handling belongs to GUI, we kept this code
for historical reasons, but nowdays there is
really no need of this old, (mostly) unused
and especially incorrect designed functionality.

It is up to the GUI to choose the book (far easier for
the user) and to select the book parameters. In no
place, including fishtest, TCEC, rating lists, etc.
the "own book" is used, moreover currently SF is
released without any book and even if in the future we
bundle a book in the release package, it will be the GUI
that will take care of it.

This corrects a wrong design decision that Galurung
and later Stockfish inherited from what was common
practice many yeas ago.

No functional change.
2014-06-01 15:29:34 +02:00
Marco Costalba e4fc9d84d7 Retire eval weights UCI options
There is really little that user can achieve (apart
from a weakened engine) tweaking these parameters
that are already tuned and have no immediate or visible
effect.

So better do not expose them to the user and avoid the
typical "What is the best setup for my machine?" kind of
question (by far the most common, by far the most useless).

No functional change.
2014-06-01 13:37:21 +02:00
Marco Costalba 07a525cdfd Retire polyglot.ini
All actively maintained GUI are able
to understand UCI protocol, so there
is no more need for this obsolete file.

No functional change.
2014-05-31 23:48:23 +02:00
Marco Costalba f5622cd5ec Restore development version
bench: 8732553
2014-05-31 23:34:36 +02:00
Marco Costalba 54f8a9cb13 Stockfish 5
Stockfish bench signature is: 8732553
2014-05-31 09:16:54 +02:00
Marco Costalba ad937d0b2d Revert "Symmetric king safety"
Regression test of 40K games at 60 secs shows
this commit to be a 2-3 ELO regression.

So revert to original king safety.

bench: 8732553
2014-05-26 21:39:48 +02:00
Marco Costalba 495a0fa699 Fix a warning with Intel compiler
warning #2259: non-pointer conversion from
"int" to "int16_t={short}" may lose significant
bits.

No functional change.
2014-05-25 00:21:46 +02:00
Marco Costalba 585655b16e Tidy up tt.h
Backport some non-functional changes
found working on 'dense TT' patch.

No functional change.
2014-05-25 00:02:09 +02:00
Marco Costalba e49eb67119 Add perft 'divide' command
To show perft numbers for each move. Just
use 'divide' instead of 'perft', for instance:

position startpos moves e2e4 e7e5
divide 4

Inspired by Ronald de Man.

No functional change.
2014-05-24 09:56:32 +02:00
Marco Costalba 88b5100e29 Update polyglot.ini after last patch
No functional change.
2014-05-19 21:19:47 +02:00
Lucas Braesch 40f5abba10 Symmetric king safety
Retire current asymmetric king evaluation
and use a much simpler symmetric one.

As a side effect retire the infamous
'Aggressiveness' and 'Cowardice' UCI
options.

Tested in no-regression mode,

Passed both STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33855 W: 5863 L: 5764 D: 22228

And LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 40571 W: 5852 L: 5760 D: 28959

bench: 8321835
2014-05-19 14:24:39 +02:00
Marco Costalba 5e03734eac Fix an off-by-one bug in extract_pv_from_tt
At root we start counting plies from 1,
instead pv[] array starts from 0. So
the variable 'ply' we use in extract_pv_from_tt
to index pv[] is misnamed, indeed it is
not the real ply, but ply-1.

The fix is to leave ply name in extract_pv_from_tt
but assign it the correct start value and
consequentely change all the references to pv[].
Instead in insert_pv_in_tt it's simpler to rename
the misnamed 'ply' in 'idx'.

The off-by-one bug was unhidden when trying to use
'ply' for what it should have been, for instance in
this position:

position fen 8/6R1/8/3k4/8/8/8/2K5 w - - 0 1

at depth 24 mate line is erroneusly truncated due
to value_from_tt() using the wrong ply.

Spotted by Ronald de Man.

bench: 8732553
2014-05-17 22:59:07 +02:00
Marco Costalba e46a72dd1d Extract a reliable PV line
Truncate the extracted PV from the point where
the score stored in hash starts to deviate from
the root score.

Idea from Ronald de Man.

bench: 8732553
2014-05-17 12:49:52 +02:00
Michel Van den Bergh 5ec63eb6b6 Drop to qsearch at low depth in razoring
If razoring conditions are satisfied and
depth is low, then directly drop in qsearch.

Passed both STC
LLR: 2.98 (-2.94,2.94) [-1.50,4.50]
Total: 12914 W: 2345 L: 2208 D: 8361

And LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 50600 W: 7548 L: 7230 D: 35822

bench: 8739659
2014-05-13 22:37:28 +02:00
Arjun Temurnikar a3c8c4b70d Remove undefended minors
Tested in "no regression" mode.

Passed both STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 68026 W: 12277 L: 12236 D: 43513

And LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 85682 W: 12861 L: 12836 D: 59985

bench: 7311935
2014-05-13 21:37:42 +02:00
Marco Costalba 696d6cedb9 Save stalemates in TT
When there aren't legal moves after
a search, instead of returning imediately,
save bestValue in TT as in the usual case.

There is really no reason to special case
this one.

With this patch is fully fixed (again) follwing
position:

    7k/6p1/6B1/5K1P/8/8/8/8 w - - 0 1

Also in SMP case.

bench: 8802105
2014-05-11 10:56:25 +02:00
Marco Costalba 9f843adf89 Retire "Idle Threads Sleep" UCI option
After last Joona's patch there is no measurable
difference between the option set or unset.

Tested by Andreas Strangmüller with 16 threads
on his Dual Opteron 6376.

After 5000 games at 15+0.05 the result is:

1 Stockfish_14050822_T16_on   : 3003  5000 (+849,=3396,-755), 50.9 %
2 Stockfish_14050822_T16_off  : 2997  5000 (+755,=3396,-849), 49.1 %

bench: 880215
2014-05-11 10:29:56 +02:00
Arjun Temurnikar bfd8704a7d Make imbalance table more clear
No functional change.
2014-05-10 08:54:31 +02:00
Reuven Peleg f89a8f0769 Pass Position as const ref in update_stats()
No functional change.
2014-05-08 22:36:30 +02:00
Marco Costalba 6ba1d3ead6 Clarify some comments in SMP code
Spotted by Joona.

No functional change.
2014-05-08 09:09:35 +02:00
Marco Costalba 7e3dba4f4c Reformat and simplify previous patch
No functional change.
2014-05-07 08:56:16 +02:00
Joona Kiiski f6e98a924a Allow a slave to 'late join' another splitpoint
Instead of waiting to be allocated, actively search
for another split point to join when finishes its
search. Also modify split conditions.

This patch has been tested with 7 threads SMP and
passed both STC:

LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 2885 W: 519 L: 410 D: 1956

And a reduced-LTC at  25+0.05
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 4401 W: 684 L: 566 D: 3151

Was then retested against regression in 3 thread case
at standard LTC of  60+0.05:

LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 40809 W: 5446 L: 5406 D: 29957

bench: 8802105
2014-05-07 08:38:56 +02:00
Ron Britvich 8f6a494ad7 Rewrite Score extractors
Less tricky and even a bit faster. With this
version Visual Studio Ultimate 2013 Update 2 RC
runs fine even in O2 optimization.

No functional change.
2014-05-05 09:05:29 +02:00
Marco Costalba b8e6f83cfb Change search() signature
Pass SpNode as template parameter.

No functional change.
2014-05-04 13:35:30 +02:00
Marco Costalba 5413fda739 Revert dynamic contempt
On a final fixed game number test it failed
to prove better than standard version.

STC 15+0.05

ELO: -0.86 +-1.7 (95%) LOS: 15.8%
Total: 57578 W: 10070 L: 10213 D: 37295

bench: 8802105
2014-05-04 09:58:49 +02:00
Marco Costalba 145d293142 Revert stalemate detection in evaluation
Unfortunatly we have a slow down that causes
a regression in STC with no-regression mode:

LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 22454 W: 3836 L: 4029 D: 14589

bench: 8678654
2014-05-04 09:42:32 +02:00
Marco Costalba f1240483fb Revert back KBPsK to latest Gary's version
The bug was found to be elsewhere. This version
is correct and also is able to detect as draw
positions like:

8/8/5b2/8/8/4k1p1/6P1/5K2 b - - 6 133

bench: 8678654
2014-05-04 09:34:22 +02:00
Ronald de Man c0d3010438 Fix KXK endgame
Position is win also if strong side has a bishop
and a knight (plus other material, otherwise
KBNK would be triggered instead of KXK).

This fixes a subtle bug where a search on position

k7/8/8/8/8/P7/PB6/K7 b - - 6 1

Instead of returning a draw score, suddendly returns
a big score. This happens because at one point in
search we reach this position:

8/Pk6/8/8/8/4B3/P7/K7 w - - 3 8

Where white can promote. In case of rook promotion (and also in case of
queen promotion) the resutling position gets a huge static eval that is
above VALUE_KNOWN_WIN (from the point of view of white). So for rook
promotion it is

          &&  futilityBase > -VALUE_KNOWN_WIN

that prevents futility pruning in qsearch. (Removing this condition indeed
lets the problem occur). Raising the static eval for K+B+N+X v K to a value
higher than VALUE_KNOWN_WIN fixes this particular problem without having to
introduce an extra futility pruning condition in qsearch.

I just checked and it seems K+R v K, K+2B v K and even K+B+N v K already get
a huge static eval. Why not K+B+N+P v K?

I think this fix corrects an oversight. There is special code for KBNK, but
KBNXK is handled by KXK, so the test for sufficient material should also test
for B+N.

bench: 8678654
2014-05-03 21:40:09 +02:00
Marco Costalba a9e93fa6a5 Fully correct stealmate detection
In the (rare) cases when the two conditions
are true, then fully check again with a slow
but correct MoveList<LEGAL>(pos).size().

This is able to detect false positives like
this one:

8/8/8/Q7/5k1p/5P2/4KP2/8 b - - 0 17

When we have a possible simple pawn push that
is not stored in attacks[] array. Because the
third condition triggers very rarely, even if
it is slow, it does not alters in a measurable
way the average speed of the engine.

bench: 8678654
2014-05-03 12:12:22 +02:00
Marco Costalba 9e8bf82350 Add stealmate detection to evaluation
Currently a stealmate position is misevaluated
in a negative/positive score, this leads qsearch(),
that does not detects stealmates too, to return the
wrong score and this yields to some kind of endgames
to be completely misevaluated.

With this patch is fully fixed follwing position

7k/6p1/6B1/5K1P/8/8/8/8 w - - 0 1

Also in SMP case.

Correct root cause analysys by Ronald de Man.

bench: 8678654
2014-05-03 11:47:49 +02:00
Marco Costalba dc87ec7258 Revert to Galurung's KBPsK endgame
After reverting to the original Tord's
endgame, a search on position

7k/6p1/6B1/5K1P/8/8/8/8 w - - 0 1

Reports, correctly, a draw score instead of
an advantage for white.

Issue reported by Uri Blass.

bench: 8678654
2014-05-02 11:04:18 +02:00
Marco Costalba 43973f43c6 Use only standard conforming eg_value()
Remove the optimization for Intel, is not
standard and can break at any time, moreover
our release build is not done with Intel C++
anymore so we don't need to sqeeze the extra
speed out from this compiler.

No functional change.
2014-05-01 23:08:07 +02:00
Marco Costalba bee4f1cf09 Don't save stale value in TT after split
If we return from split with a stale value
due to a stop or a cutoff upstream occurred,
then we exit moves loop and save a stale value
in TT before returning search().

This patch, from Joona, fixes this.

bench: 8678654
2014-05-01 16:26:18 +02:00
Marco Costalba da91a34c09 Better document search stop condition
Handling a stop or a cutoff in the search is
a tricky business, so better document this
difficult part of the code.

No functional change.
2014-05-01 08:48:59 +02:00
Marco Costalba 626dc8a03b Remove dead code in search
We can never have bestValue == -VALUE_INFINITE at
the end of move loop because if no legal move exists
we detect it with previous condition on !moveCount,
if a legal move exists we never prune it due to
futility pruning condition:

bestValue > VALUE_MATED_IN_MAX_PLY

So this code never executes, as I have also verified
directly.

Issue reported by Joona.

No functional change.
2014-05-01 07:46:44 +02:00
Marco Costalba cf50e265fa Fix a compile error with Intel C++
Intel compiler is very picky:
"error: this operation on an enumerated type requires an
applicable user-defined operator function"

Reported by Tony Gaor.

No functional change.
2014-04-30 08:55:45 +02:00
snicolet d3ffd0ffca Penalize hanging pieces
Passed both STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8519 W: 1565 L: 1440 D: 5514

And LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 60618 W: 9141 L: 8777 D: 42700

bench: 8678654
2014-04-30 08:36:44 +02:00
Marco Costalba db229504e2 Rearrange interpolation formula
Put the division at the end to reduce
rounding errors. This alters the bench
due to different rounding errors, but
should not alter ELO in any way.

bench: 7615217
2014-04-28 17:27:49 +02:00
mstembera 918c29f83a Minor stuff scattered around
Just random minor stuff I found while browsing the code.

No functional change.
2014-04-28 17:07:43 +02:00
Marco Costalba a1f39c1ef9 Remove other useless floor()
No functional change.
2014-04-27 19:17:40 +02:00
Marco Costalba 9f2a64abd2 Don't need floor() in timeman.cpp
For positive numbers result is equivalent:
http://stackoverflow.com/questions/3300290/cast-to-int-vs-floor

Spotted by Joseph Ellis.

No functional change.
2014-04-27 19:10:00 +02:00
Marco Costalba 93e3b06fe2 Fix Intel compiler warnings
Fallout from previous patch: Intel compiler
is very noisy.

No functional change.
2014-04-27 12:02:36 +02:00
Marco Costalba 86c20416c8 Remove some useless casts
No functional change.
2014-04-27 11:44:16 +02:00
Marco Costalba c9e396b542 We can add an integer to a Value
We have defined corresponding operators,
so rely on them to streamline the code
and increase readibility.

No functional change.
2014-04-27 11:25:42 +02:00
Arjun Temurnikar fe23f27086 Remove rook passers eval completely
Tested in no-regression mode.

Passed STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17727 W: 3248 L: 3122 D: 11357

And (a very long!) LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 106327 W: 16258 L: 16265 D: 73804

bench: 7396783
2014-04-27 09:51:48 +02:00
Marco Costalba de6fe2cb4e Reformatting in material.h
No functional change.
2014-04-27 09:31:22 +02:00
Marco Costalba 057c3d60cd Move game phase constants to enum Value
No functional change.
2014-04-27 09:23:45 +02:00
Marco Costalba cb4ac4221e Speed up by almost 3%
This apparentely silly tweak allows
to speed up the bench by almost 3%.

Not clear why, repeating with perft,
the speed up vanishes.

Suggested by Jonathan Calovski.

No functional change.
2014-04-27 00:25:47 +02:00
Marco Costalba 55604f156b Fix issues detected by Coverity Scan
Most of Coverity Scan reports are false
positives, but in rare cases we have
confirmed (very small) issues.

No functional change.
2014-04-26 09:33:50 +02:00
Marco Costalba 7ddbcf7e87 Speed up picking of killers
Changing the order of the conditions gives
about 1% speed up!

No functional change.
2014-04-25 12:53:51 +02:00
Arjun Temurnikar 3705559fdb Remove RookOn7th and merge values into psqt
Tested in no-regression mode:

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 55678 W: 9954 L: 9892 D: 35832

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 9238 W: 1448 L: 1311 D: 6479

bench: 7905850
2014-04-24 08:53:46 +02:00
Arjun Temurnikar 6579a65bbb Remove penalty for knight when few enemy pawns
Tested in standard mode at STC and no-regression
mode at LTC:

STC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 19503 W: 3502 L: 3349 D: 12652

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 67474 W: 9974 L: 9921 D: 47579

bench: 8331217
2014-04-24 08:47:13 +02:00
Jonathan Calovski f70cef3b79 Shuffle movepicker score
Believed to be a speed optimization as benched
on Windows with bench realtime affinity 0x1 deleting
highest and lowest runs:

Base	Test
1549259	1608202
1538115	1583934
1543168	1556938
1536365	1554179
1533026	1582010

Signature remains unchanged and gives anywhere from 1-2% nps
boost in analysis depending on number of cores used.

No functional change.
2014-04-24 08:38:11 +02:00
Marco Costalba a89b26bedd Correctly apply previous patch
Apply the correct values from previous patch.

bench: 8082049
2014-04-24 08:35:13 +02:00
joergoster 8bfb53efe2 Move queen vs. 3 minors rule to imbalance tables
Tuned with CLOP after 57k games.

Simplification: tested in no-regression mode.

Passed both STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17254 W: 3159 L: 3032 D: 11063

And LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32413 W: 4967 L: 4864 D: 22582

bench: 8082049
2014-04-23 08:51:34 +02:00
Marco Costalba a66e6e5ad9 Revert "Generalize shift_bb() to handle double pushes"
Seems to intorduce some compiler warning as
reported by Gary. Warning seems bogus, but
revert anyhow.

No functional change.
2014-04-21 20:56:12 +02:00
Marco Costalba 56273fca1e Reset DrawValue[] before new search
Spotted by Ronald de Man

bench: 7384368
2014-04-21 14:30:27 +02:00
Leonid Pechenik ef43e6b05d Rise contempt when in advantage
This is a very discussed patch with many
argumentations pro and against. The fact is
it passed both STC:

LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16305 W: 3001 L: 2855 D: 10449

And LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 34273 W: 5180 L: 4931 D: 24162

Although it is true that a correct test should
include foreign engines, we commit it anyhow so
people can test it out in the wild, under broader
conditions.

bench: 7384368
2014-04-21 12:23:03 +02:00
Marco Costalba 223ebe7b40 Streamline implementation of Position::pretty()
Simpler and more in line with Bitboards::pretty()

No functional change.
2014-04-21 11:47:01 +02:00
Stefan Geschwentner 0e8ad40ef0 Raise penalty blocked enemy pawn on 6th rank
Idea from Lyudmil Tsvetkov.

The value seems to be raised a bit abruptly, but as
Gary said, a blocked pawn on the sixth rank has been
instrumental in limiting king mobility in multiple
losses that I've seen from SF. A blocked pawn on fifth
rank is much less serious on the king safety impact.

Passed both STC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 14551 W: 2750 L: 2607 D: 9194

and LTC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 43595 W: 6917 L: 6618 D: 30060

And even a retest at 60" fixed games 40K
ELO: 1.79 +-1.9 (95%) LOS: 97.0%
Total: 39889 W: 6018 L: 5813 D: 28058

bench: 7154916
2014-04-21 07:47:50 +02:00
Marco Costalba eced15fe36 Generalize shift_bb() to handle double pushes
And use it in evaluate_space.

No functional change.
2014-04-20 15:52:37 +02:00
Arjun Temurnikar a4d058bca2 Small simplification to passed pawns
Tested in no-regression mode.

Passed both STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 28521 W: 5066 L: 4958 D: 18497

And LTC
LLR: 3.04 (-2.94,2.94) [-3.00,1.00]
Total: 21939 W: 3261 L: 3138 D: 15540

bench: 8165681
2014-04-20 10:06:51 +02:00
Joseph Hellis 619d66b7ab Remove supporting pawns
Tested in no-regression mode

Passed both STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 44957 W: 7984 L: 7903 D: 29070

and LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 12396 W: 1916 L: 1783 D: 8697

Bench: 7907885
2014-04-17 08:45:31 +02:00
Jean-Francois Romang 9073866491 Enable BMI1 too when using BMI2 ARCH
Adding BMI1 allows the compiler to use _blsr_u64
automatically (the advertised 0.3% speed gain).
I verified that the compiler does not use this
instruction with the -mbmi2 flag only. Also, all
processors supporting BMI2 is also supporting BMI1.

No functional change
2014-04-14 07:54:24 +02:00
Marco Costalba 7bce8831d3 More readable trapped rook condition
Prefer

file_of(s) < file_of(ksq)

to the inidrect

file_of(ksq) < FILE_E

To evaluate if semiopen side to check is the left side.

Also other small touches while there.

No functional change.
2014-04-13 14:29:42 +02:00
Daylen Yang 81a8c1118b Simplify Makefile by removing OS X targets
Right now the Makefile is cluttered with OS X equivalents
of all the x86 targets. We can get rid of all of them and
just check UNAME against "Darwin" for the few OS X-specific
things we need to do.

We also disable Clang LTO when using BMI2 instructions. For
some reason, LLVM cannot find the PEXT instruction when using
LTO. I don't know why, but disabling LTO for BMI2 fixes it.

No functional change.
2014-04-13 09:05:21 +02:00
Marco Costalba b2c0634d48 Move args parsing to UCI::loop
This leaves a very clean main.cpp

No functional change.
2014-04-12 13:51:52 +02:00
Marco Costalba ada55c5d0a Reshuffle in uci.cpp
Move function definitions before call site.

No functional change.
2014-04-12 12:05:25 +02:00
Marco Costalba 17cb7e7fa3 Reshuffle in timeman.cpp
Move template definitions before call site.

No functional change.
2014-04-12 12:00:37 +02:00
Marco Costalba 800ba28e83 Fix a typo in evaluate
Spotted by Lyudmil Antonov.

No functional change.
2014-04-12 10:17:41 +02:00
Marco Costalba 0b2794ae12 Retire signature-build
Does not seem used.

No functional change.
2014-04-12 09:18:55 +02:00
Jean-Francois Romang 226bbc1e63 Add ARCH x86-64-bmi2 support
Intel Haswell and newer CPUs can calculate sliders
attacks using special PEXT asm instructions instead
of magic bitboards. This gives a +3% speed up.

To enable it just compile with ARCH=x86-64-bmi2

No functional change.
2014-04-12 09:15:14 +02:00
Marco Costalba da2f8880b9 Switch to hardware PEXT
Retire software pext and introduce hardware
call when USE_PEXT is defined during compilation.

This is a full complete implementation of sliding
attacks using PEXT.

No functional change.
2014-04-12 08:55:30 +02:00
Marco Costalba c556fe1d71 Implement PEXT based attacks
According to:

https://chessprogramming.wikispaces.com/BMI2#PEXTBitboards

No functional change.
2014-04-12 08:55:30 +02:00
Marco Costalba 2bfe61c33b Add PEXT software implementation
For development/debug purposes.

No functional change.
2014-04-12 08:55:30 +02:00
Marco Costalba 2f92e3b525 Big reshuffle in evaluate.cpp
Reshuffle functions to define them in reverse
calling order (C style).

This allow us to define templates before they are
used. Currently it is not like this, for instance
evaluate_pieces is defined after do_evaluate that
calls it. This happens to work for some strange
reason (two phase lookup?) but we want to avoid
code that works 'by magic'.

As a nice side-effect we can now remove the function
prototypes.

No functional change.
2014-04-12 08:39:18 +02:00
Gary Linscott 0510112f91 Move LowMobPenalty into psq/mobility tables
Tested in no-regression mode.

Passed both STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36705 W: 6537 L: 6443 D: 23725

and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 9358 W: 1495 L: 1358 D: 6505

bench: 6921356
2014-04-12 07:15:51 +02:00
Gary Linscott 4597324572 Reduce penalty for doubled pawns further away from each other
Passed both STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 38339 W: 6849 L: 6649 D: 24841

and LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 22693 W: 3455 L: 3256 D: 15982

bench: 7508468
2014-04-12 07:05:02 +02:00
Marco Costalba ce6b7a1b85 Further simplification in evaluate
No functional change
2014-04-11 08:38:09 +02:00
Marco Costalba e3b54235ad Get rid of an evaluate_pieces() overload
Rewrite and greatly simplify that part of code.

No functional change.
2014-04-10 19:45:18 +02:00
Gary Linscott 8863afeb84 Add a penalty for low mobility pieces
Passed both STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 81857 W: 14652 L: 14342 D: 52863

and LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 45400 W: 6999 L: 6697 D: 31704

bench: 7716978
2014-04-10 08:35:10 +02:00
Gary Linscott 5c75455c8e Restrict queen mobility to safe squares
Passed both STC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 16188 W: 3119 L: 2971 D: 10098

and LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 6336 W: 1010 L: 882 D: 4444

bench: 7533692
2014-04-08 22:25:54 +02:00
Lucas Braesch 0d8a4c7565 Rescale UCI scores to PawnValueEg
This is more consistent with what other engines are doing.
Often people thinks that SF's scores are overblown. In the
end, it just boils down to the arbitrary way of rescaling them.

No functional change.
2014-04-06 11:53:28 +02:00
Marco Costalba 64d29a6330 Sync some common names
No functional change.
2014-04-06 11:26:12 +02:00
Marco Costalba fcf2a34080 Some more work in pretty_pv
No functional change.
2014-04-06 10:15:11 +02:00
Marco Costalba 698b645e10 Small tidy up in move_to_san
No functional change.
2014-04-05 19:05:37 +02:00
mstembera d28ea7b518 Smaller and faster profile-build binaries
I have noticed that increasing the bench depth produces
progressively smaller and slightly faster executables at
the cost of longer compile times.  Also using bench "time"
instead of "depth" seems to produce slightly smaller/faster
executables  given comparable compile times.

I have made a new Makefile that generates smaller and
about 1% to 2% faster profile executables at only a
little extra compile time.  On  my mobile 2GHz i7 a
full profile build time goes from 3'48" to 4'13" and
the exe goes down by 5% from 416,310 bytes to 395,567
bytes.

No functional change.
2014-04-05 12:48:20 +02:00
Lucas Braesch be641e881f Remove QueenOn7th and QueenOnPawn
Small simplification.

Passed SPRT(-3,1) both at STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17051 W: 3132 L: 3005 D: 10914

and LTC:
LLR: 4.55 (-2.94,2.94) [-3.00,1.00]
Total: 24890 W: 3842 L: 3646 D: 17402

The rationale behind this is that I've never managed to add a
Queen on 7th rank bonus in DiscoCheck, because it never showed
to be positive (evne slightly) in testing. The only thing that
worked is Rook on 7th rank.

In terms of SF code, it seemed natural to group it with QueenOnPawn
as well as those are done together. I know you're against groupping
in general, but when it comes to non regression test, you are being
more conservative by groupping. If the group passes SPRT(-3,1) it's
safer to commit, than test every component in SPRT(-3,1) and end up
with the risk of commiting several -1 elo regression instead of just
one -1 elo regression.

In chess terms, perhaps it's just easier to manouver a Queen (which
can more also diagonaly) than a Rook. Therefore you can let the search
do its job without needing eval ad-hoc terms to guide it. For the Rook
which takes more moves to manouver such eval terms can be (marginally)
useful.

bench: 7473314
2014-04-05 11:26:44 +02:00
mstembera 3b19ea6ae5 Speed up apply_weight
Speed up by about 2% this hot path function pre-calculating
midgame and endgame values of the weight.

No functional change.
2014-04-05 11:12:18 +02:00
Marco Costalba ccf59b0228 Update Readme to 128 max threads
No functional change.
2014-04-03 11:33:42 +02:00
Joerg Oster 299afcd886 Queen vs. 3 pieces imbalance
Passed both STC
LLR: 7.32 (-2.94,2.94) [-1.50,4.50]
Total: 98108 W: 18087 L: 17576 D: 62445

And LTC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 15082 W: 2417 L: 2248 D: 10417

bench: 7717336
2014-04-03 10:42:06 +02:00
Marco Costalba c15b132f03 Make operator<< to return void
This should help preventing misuse.

No functional change.
2014-04-03 10:34:25 +02:00
Stefan Geschwentner 0ba814b3ca Drop not defended by pawn condition
Passed no-regression test both at STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 96554 W: 17563 L: 17572 D: 61419

and at LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 31891 W: 4905 L: 4801 D: 22185

bench: 7720896
2014-04-01 11:59:25 +02:00
Marco Costalba d2caba1f66 Do not workaround function argument evaluation
Rewrite options initialization to do not trying to
hack upon the undefined evaluation order of function
arguments.

No functional change.
2014-04-01 09:39:41 +02:00
Marco Costalba 1efc19ade0 Fix a compile error with Intel C++
Error: a value of type "int" cannot be assigned
to an entity of type "Value"

No functional change.
2014-03-30 14:20:12 +02:00
Marco Costalba 678425f274 Fix a warning with MSVC 2010
Warning C4804: '<' : unsafe use of type 'bool' in operation

No functional change.
2014-03-30 14:25:57 +02:00
Marco Costalba 422c9c2acd Show evaluation from white POV in trace
We chose this instead of negamax sign convention
(ie. from the point of view of the side to move)
because it is more in line to how the eval
table is presented.

Also some tweak to formatting while there.

No functional change.
2014-03-30 10:45:46 +02:00
Marco Costalba 9350d0dce5 Raise VALUE_INFINITE
In some legal positions like this one:
R6R/3Q4/1Q4Q1/4Q3/2Q4Q/Q4Q2/Np1Q4/kB1N1KB1  b -- 0 1

We can have a very high score, in this case 30177 and 29267
for midgame and endgame respectively, and because
VALUE_INFINITE = 30001 we have an assert in interpolate()

Midgame and endgame scores are stored in 16 bit signed integers
so we can rise VALUE_INFINITE a little bit. This does not fix
the possibility of overflow in general case, just makes the
condition more difficult to trigger and anyhow better uses all
the score width.

Raising VALUE_INFINITE to 32000 seems to fix the problem for this
particular case.

No functional change.
2014-03-29 11:13:42 +01:00
Jean-Francois Romang 4833887842 Further simplification of TT replace strategy
No functional change
2014-03-29 10:05:02 +01:00
Marco Costalba f811a5693e Restore old aspiration window to 16
Tested directly at LTC because previous long
test series on this topic shows it is TC dependant.

Tested with no-regression mode because gets rid of
an ugly and ad-hoc rule.

Test at LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 67918 W: 10590 L: 10541 D: 46787

bench: 7926803
2014-03-29 08:45:40 +01:00
Stefan Geschwentner af0c13ba6a Pinned pieces affect king safety
Here the new idea is to link pinned pieces
with king safety.

Passed both STC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 10047 W: 1867 L: 1737 D: 6443

And LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 10419 W: 1692 L: 1543 D: 7184

bench: 8325087
2014-03-29 08:37:55 +01:00
Marco Costalba c7cf45241c Fix a bug in pawns eval tracing
Instead of totals we were showing white and
black values.

Spotted by Sven Schüle

No functional change.
2014-03-26 07:06:29 +01:00
Marco Costalba 865b71309c Simplify TT replace strategy
Tested for no-regression with SPRT[-3, 1] at STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32046 W: 6020 L: 5918 D: 20108

No functional change.
2014-03-24 08:32:08 +01:00
snicolet e7362dae78 Introduce penalty for weak (=unsupported) pawns.
We add a penalty for each pawn which is not protected by another pawn
of the same color.

Passed both short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 12107 W: 2411 L: 2272 D: 7424

And long TC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 9204 W: 1605 L: 1458 D: 6141

bench: 7682173
2014-03-24 08:27:40 +01:00
Marco Costalba 9e72e35942 Fix an incorrect 'friend' declaration
Spotted by Lee David.

No functional change.
2014-03-23 11:17:38 +01:00
Marco Costalba bc183b0c04 Retire last usage of operator|(File f, Rank r)
This for some reason was missed.

No functional change.
2014-03-23 10:42:37 +01:00
mstembera ffdf63ff7c Refresh TT entries generation automatically on probe
And other assorted simplifications, tested with SPRT[-3, 1]

Passed both short TC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 18814 W: 3600 L: 3475 D: 11739

And long TC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 20731 W: 3217 L: 3096 D: 14418

No functional change.
2014-03-23 09:46:15 +01:00
Marco Costalba f12449d492 Rename score to value in ExtMove
We use 'score' for the value mid-endgame pair.

No functional change.
2014-03-23 08:54:10 +01:00
Marco Costalba 441a34a1c6 Trivial formatting in ucioption.cpp
No functional change.
2014-03-23 00:06:26 +01:00
Marco Costalba 5f12069cbf Retire operator|(File f, Rank r)
Use make_square() instead. Less fancy but
more clear.

No functional change.
2014-03-22 23:54:18 +01:00
Marco Costalba 4eee603433 Fix a subtle bug in UCI options printing
We want all the UCI options are printed in the order in which are
assigned, so we use an index that, depending on Options.size(),
increases after each option is added to the map. The problem is
that, for instance, in the first assignment:

o["Write Debug Log"] = Option(false, on_logger);

Options.size() can value 0 or 1 according if the l-value (that
increments the size) has been evaluated after or before the
r-value (that uses the size value).

The culprit is that assignment operator in C++ is not a
sequence point:

http://en.wikipedia.org/wiki/Sequence_point

(Note: to be nitpick here we actually use std::map::operator=()
 that being a function can evaluate its arguments in any order)

So there is no guarantee on what term is evaluated first and
behavior is undefined by standard in this case. The net result
is that in case r-value is evaluated after l-value the last
idx is not size() - 1, but size() and in the printing loop
we miss the last option!

Bug was there since ages but only recently has been exposed by
the removal of UCI_Analyze option so that the last one becomes
UCI_Chess960 and when it is missing engine cannot play anymore
Chess960.

The fix is trivial (although a bit hacky): just increase the
last loop index.

Reported by Eric Mullins that found it on an ARM and MIPS
platforms with gcc 4.7

No functional change.
2014-03-22 11:30:06 +01:00
Marco Costalba c714b90594 Fix a typo
Spotted by Isaac Haïk Dunn.

No functional change.
2014-03-18 18:36:39 +01:00
Marco Costalba aab5863dd4 Increase max threads to 128
Thanks to std::bitset we can easily increase
the limit of active threads above 64.

Thanks to Lucas Braesch for pointing at the
correct solution of using std::bitset.

No functional change.
2014-03-18 12:07:26 +01:00
Marco Costalba fa3f6dcbea Fix a crash under MSVC
Using memset on a std::vector is undefined behavior,
so manually init all the data memebers of LimitsType.

Bug intorduced in 41641e3b1e

No functional change.
2014-03-16 10:55:58 +01:00
Marco Costalba a091ae4cc8 Split also if no slaves are found
Because we test for available slaves before
entering split(), we almost always allocate a
slave, only in the rare case of a race (less
then 2% of cases) this is not true, but to
special case this occurrence is not worth
the added complexity.

bench: 7451319
2014-03-15 23:43:35 +01:00
Marco Costalba a1a7bc84da Remove "Max Threads per Split Point" UCI option
Experimental patch to verify if drop of nps
in endgames at very long TC is due to this.

Suggested by Ronald de Man.

bench: 7451319
2014-03-15 21:26:04 +01:00
Marco Costalba 6e4b4c42ed Merge default tests in pos_is_ok
No functional change.
2014-03-15 15:34:31 +01:00
Marco Costalba 142874b058 Microptimize castling in undo_move()
We don't need to set 'captured' and 'pt' after we
castle back.

No functional change.
2014-03-15 11:47:30 +01:00
Mysseno 36c381154b Depth dependant aspiration window delta
Split delta value in aspiration window so that when
search depth is less than 24 a smaller delta value
is used. The idea is that the search is likely to
be more accurate at lower depths and so we can exclude
more possibilities, 25% to be exact.

Passed STC
LLR: 2.96 (-2.94, 2.94) [-1.50, 4.50]
Total: 20430 W: 3775 L: 3618 D: 13037

And LTC
LLR: 2.96 (-2.94, 2.94) [0.00, 6.00]
Total: 5032 W: 839 L: 715 D: 3478

Bench: 7451319
2014-03-14 19:47:41 +01:00
Marco Costalba d0587f2c7f Use legal positions for endgame initialization
During endgame initialization we get the material
hash key of each endgame forging and ad-hoc position
that in same cases is illegal (leaves teh king under
capture). This is not a problem for the material key,
but rises an assert when SF is run in debug mode with
'testKingCapture' set in pos_is_ok().

So rewrite the code to always produce legal positions.

No functional change.
2014-03-14 09:57:34 +01:00
Marco Costalba 1f4798a173 Further work in pos_is_ok()
No functional change.
2014-03-14 09:43:19 +01:00
Daylen Yang 61e4443342 Re-add link time optimization on OS X
In the new version of clang, -O4 does not imply -flto, so we set the
flag
2014-03-14 08:49:08 +01:00
Marco Costalba 6571acffaa Reformat do_castling()
No functional change.
2014-03-13 12:53:03 +01:00
Marco Costalba 6f3d787692 Further merge StateInfo setup functions
No functional change.
2014-03-12 22:46:17 +01:00
Marco Costalba 1e032ece92 Merge hash key computation functions
No functional change.
2014-03-12 09:14:38 +01:00
Marco Costalba cc3002ff04 Use std::count in pos_is_ok()
No functional change.
2014-03-11 23:19:47 +01:00
Marco Costalba c40d4e0133 Small simplification in gives_check
Use switch statement also for normal case.

No speed regression.

No functional change.
2014-03-11 22:58:08 +01:00
Marco Costalba ca0804dfe4 Print dbg counters after bench
Print last debug counters update just
before to exit benchmark.

Suggested by Stephane Nicolet.

No functional change.
2014-03-11 22:19:14 +01:00
Marco Costalba d2a8ba3299 Simplify pseudo_legal()
Big simplification of pawn move check.

Code has been tested with a brute force approach: for
every position reached during a bench search, the function
has been called for each combinations of Move(from, to)
and verified the result is the same of old code.

Actually this function is very critical becuase is the
one that ensures corrupted TT moves are discarded, so
to properly test it a simple bench is not enough.

Verified also speed is not changed.

No functional chnage.
2014-03-10 08:38:23 +01:00
Marco Costalba 20ff12e1be Simplify generate<EVASIONS>
No speed regression, tested with both perft and
bench.

No functional change.
2014-03-09 12:16:27 +01:00
Marco Costalba 5cf9e0b254 Retire SERIALIZE macros
Explicitly write the 'while' loops. This adds some
code but makes clear what's the code behind the
macro.

No functional change.
2014-03-09 11:10:33 +01:00
Marco Costalba 1d1b7df7c6 Rename castling flag to castling right
This is a more conventional naming as
reported also in:

http://chessprogramming.wikispaces.com/Castling+rights

No functional change.
2014-03-08 15:08:55 +01:00
Stephane Nicolet 13b9e1e098 Fix dbg_mean_of() for negative numbers
Type should be int64_t instead of uint64_t

No functional change.
2014-03-05 08:55:12 +01:00
mstembera 553ead429d Some minor cleanup stuff
I came across while browsing the code.

No functional change.
2014-03-03 08:57:20 +01:00
Marco Costalba bbd69c0260 Revert dynamic draw value
When tested with weaker engines did not
performed as expected, actually it was even
a regression from standard version.

bench: 8430785
2014-03-03 08:39:34 +01:00
Marco Costalba 708cb311a0 Pass CastlingFlag argument only
Instead of Color and CastlingSide. Change functions API
accordingly.

No functional change.
2014-03-02 13:21:19 +01:00
Marco Costalba 3e5470d88f Remove limit of minimumSplitDepth
There is no reason why an user cannot set
it at a value less than 4.

No functional change.
2014-03-01 23:22:14 +01:00
Marco Costalba 3d8c0f16c2 Rename xxx_to_char() -> to_char()
No functional change.
2014-03-01 22:07:41 +01:00
Marco Costalba de2ba70830 Simplify Bitboards::pretty
No functional change.
2014-03-01 13:10:05 +01:00
Marco Costalba 9f0485e343 Retire UCI_AnalyseMode option
It has been obsoleted out already some time ago
and currently there is no point in changing eval
score according to if we are in game or analyzing.

So retire the option.

No functional change.
2014-03-01 12:10:42 +01:00
Joerg Oster b917cd275e Dynamic draw value
Try to avoid repetition draws at early midgame,
this should give an edge against weaker opponents
and reduce draw rate.

Tested for regressions with SPRT[-3, 1] and
passed both short TC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 68498 W: 12928 L: 12891 D: 42679

And long TC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 40212 W: 6386 L: 6295 D: 27531

bench: 7990513
2014-02-26 19:33:52 +01:00
Marco Costalba 0949f06a60 Fix a warning with Intel compiler
warning #2259: non-pointer conversion from "int" to
"uint8_t={unsigned char}" may lose significant bits

No functional change
2014-02-22 12:00:14 +01:00
Marco Costalba 012f20d66e Fix an assert in Probcut
When running the following position:

8/kPp5/2P3p1/p1P1p1P1/2PpPp2/3p1p2/3P1P2/5K2 w - - 0 1

An assert is raised at depth 92:

assert(-VALUE_INFINITE <= alpha && alpha < beta && beta <= VALUE_INFINITE);

This is because it happens that beta = 29832,
so rbeta = 30032 that is > VALUE_INFINITE

Bug spotted and analyzed by Uri, fix suggested by Joerg.

Other fixes where possible but this one is pointed
exactly at the source of the bug, so it is the best
from a code documentation point of view.

bench: 8430785
2014-02-22 10:38:21 +01:00
Leonid Pechenik b8cfc255d4 Distribute part of first move time to other moves
Passed both short TC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 18907 W: 3475 L: 3322 D: 12110

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 19044 W: 2997 L: 2811 D: 13236

bench: 8430785
2014-02-20 08:39:00 +01:00
Marco Costalba 9fcefb2760 Avoid recalculating CheckInfo in generate_castling()
No functional change.
2014-02-20 08:27:13 +01:00
Marco Costalba 64a71c3c02 Don't update pieceCount for ALL_PIECES
It is currently unused and only adds
overhead for nothing.

Also set proper type of chess960.

No functional change.
2014-02-18 16:41:24 +01:00
Marco Costalba 7b0a2f2a90 Update SEE to return a Value
It seems more natural because the actual returned
value is from PieceValue[] array.

No functional change.
2014-02-16 13:06:31 +01:00
Marco Costalba 17ffc22279 Sync code style in material.cpp
Update to use common code style.

No functional change.
2014-02-16 12:56:54 +01:00
Marco Costalba 52ebf87238 Increase MAX_PLY from 100 to 120
Under some very rare case 100 plies of search
could be not enough. Increasing more could lead
to crashes due to reached stack size limit on
some platforms.

Strongly requested by Uri.

bench: 8430785
2014-02-16 12:20:37 +01:00
Marco Costalba d91079d4b0 Fix material key for King
Currently king has no material key associated because
it can never happen to find a legal position without
both kings, so there is no need to keep track of it.

The consequence is that a position with only the two
kings has material key set at zero and if the material
hash table is empty any entry will match and this is
wrong.

Normally bug is hidden becuase the checking for a draw
with pos.is_draw() is done earlier than evaluate() call,
so that we never check in gameplay the material key of a
position with two kings.

Nevertheless the bug is there and can be reproduced setting
at startup a position with only two kings and typing
'eval' from prompt.

The fix is very simple: add a random key also for the king.

Also fixed the condition in material.cpp to avoid asserting
when a 'just 2 kings' postion is evaluated.

No functional change.
2014-02-16 11:51:30 +01:00
Marco Costalba 62c0dc5dea Restore PorbCut name
Actually MultiCut is too different from current scheme.
Note that neither ProbCut is exactly what we do because
we try just a handful of captures instead of all moves,
nevertheless it seems more in line with what we do.

Suggested by Joona.

No functional change.
2014-02-15 22:21:39 +01:00
Marco Costalba 7bdb8c9c5c Reformat of eval tracing code
Also assorted rename while there.

No functional change.
2014-02-15 14:11:26 +01:00
Marco Costalba 3336963185 Rename ProbCut to Multicut
The teqnique used is actually MultiCut:

https://chessprogramming.wikispaces.com/Multi-Cut

And not ProbCut:

https://chessprogramming.wikispaces.com/ProbCut

No functional change.
2014-02-15 09:37:55 +01:00
Marco Costalba 4808d15a85 Assorted renaming in search
Inspired by DON.

No functional change.
2014-02-15 09:20:27 +01:00
Joerg Oster 67f88f5e3e Return static eval when reaching MAX_PLY
Makes more sense than returning a draw score. Tested
with reduced MAX_PLY = 30 and passed both short TC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 17434 W: 3345 L: 3194 D: 10895

And long TC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 2610 W: 488 L: 373 D: 1749

With current limit of MAX_PLY = 100 the patch should not
introduce any measurable change, nevertheless is the correct
approach.

Idea of returning eval is from  Michel Van den Bergh.

bench: 8430785
2014-02-15 08:20:33 +01:00
Marco Costalba b3470d7ab1 Fix magic boosters conversion
Fix small overflow error while converting
magic boosters from right rotate to left rotate,
in particular booster 38 was converted to 4122
instead of the corrcet value 26.

Formula used was:

s1 = original & 63, s2 = (original >> 6) & 63;
new = (64 - s1) | ((64 - s2) << 6);

Instead of:

s1 = original & 63, s2 = (original >> 6) & 63;
new = ((64 - s1) & 63) | (((64 - s2) & 63) << 6);

This has no impact in number of cycles needed, but
just in the resultig number that yields to a rotate
amount bigger than 63.

Spotted by Ehsan Rashid.

No functional change.
2014-02-14 10:43:37 +01:00
Marco Costalba e4695f15bc Additional renaming from DON
Assorted renaming and triviality.

No functional change.
2014-02-14 09:42:50 +01:00
joergoster ffd6685f79 Fix a compiler warning
Latest master triggers a compiler warning due
to comparing int64_t to uint64_t.

notation.cpp: In Funktion »std::string pretty_pv(Position&, int, Value, int64_t, Move*)«:
notation.cpp:230:30: Warnung: Vergleich zwischen vorzeichenbehafteten und vorzeichenlosen Ganzzahlausdrücken [-Wsign-compare]

This patch should fix it.

No functional change.
2014-02-12 21:52:21 +01:00
Leonid Pechenik 72e8640f4d Simplify time management
Tested with simplification mode SPRT[-4, 0]

Passed both short TC
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 34102 W: 6184 L: 6144 D: 21774

And long TC
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 16518 W: 2647 L: 2545 D: 11326

And also 40/10 TC
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 22406 W: 4390 L: 4312 D: 13704

bench: 8430785
2014-02-12 20:01:11 +01:00
Marco Costalba e6523e56b8 Move magic random to RKISS
When initializing the magic numbers used to
compute sliding attacks, we endless generate a
random and test it as a possible magic.

In the general case this takes a lot of iterations,
but here, insteaad of picking a casual random, we
rotate it a couple of times and generate a number that
we know has a good probability to be a magic candidate.

This is becuase the quantities by which we rotate the
number are known in advance to produce quickly a good
canidate.

The patch, inspired by DON, just moves the shuffle to RKISS
changing the boosters to take in account a left rotation
instead of a right rotation as in the original.

No functional change.
2014-02-12 14:47:36 +01:00
Marco Costalba b534176d4a Revert "Retire null search verification"
Although does not change ELO level, it seems
verification is useful in many zugzwang positions
as reported by many sources.

So revert this simplification.

bench: 8430785
2014-02-12 14:16:21 +01:00
Henri Wiechers ddeb01612b Faster handling of king captures in Position::see
Another SEE speed up that passed the SPRT short TC test!

LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 81337 W: 15060 L: 14745 D: 51532

No functional change.
2014-02-09 19:42:24 +01:00
Marco Costalba 41641e3b1e Assorted tweaks from DON
Mainly renames and some little code style improvment,
inspired by looking at DON sources:

https://github.com/erashid/DON

No functional change.
2014-02-09 17:31:45 +01:00
Marco Costalba 2f5aaf7de6 Rewrite previous patch removing the macro
No functional change.
2014-02-08 13:27:44 +01:00
Thanar2 d48a304262 Inline common path of pos.gives_check
Test for common case which, when running default
stockfish bench, avoids 96% of 19 million calls to
pos.gives_check().

Yields to a 2-4% speed up according to platform.

Passed fishtest at STC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 12441 W: 2333 L: 2196 D: 7912
http://tests.stockfishchess.org/tests/view/52f02d790ebc5901df50f887

Passed fishtest at LTC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 42175 W: 6702 L: 6409 D: 29064
http://tests.stockfishchess.org/tests/view/52f0dbe00ebc5901df50f8a0

No functional change.
2014-02-08 13:17:29 +01:00
Marco Costalba 0a1092f64e Don't fear races when are harmless
Actually race conditions do exist in an engine, just
think for a moment to TT concurrent access. Racy code
is not a problem per se, if the consequences are well
known and correctly handled.

In case of TT access we ensure that the TT move is validated
before to be tried, here we just retry the same move in less
that 1 case out of a million: this is totally harmless considering
that very probably the second time the move is tried we get
immediately a TT hit and search quickly returns.

So we simplify the code for no harm.

No fuctional change (in single thread case)
2014-02-08 13:07:57 +01:00
Lucas Braesch 399968f1d0 Retire null search verification
Tested with SPRT in simplification mode [-4.00,0.00],
this ensures that the patch is (very probably) not
a regression.

Passed both short TC
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 27543 W: 4278 L: 4209 D: 19056

And long TC
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 39483 W: 7325 L: 7305 D: 24853

bench: 8347121
2014-02-08 08:25:08 +01:00
Lucas Braesch e5c3effdb1 Better document null search window
Hopefully this patch makes the code more:

* Self-documenting: Null search is always a zero window search,
  because it is testing for a fail high. It should never be done
  on a full window! The current code only works because we don't
  do it at PV nodes, and therefore (alpha, beta) = (beta-1, beta):
  that's the kind of "clever" trick we should avoid.

* Idiot-proof: If we want to enable null search at PV nodes, all we
  need to do now is comment out the !PvNode condition. It's that simple!

In theory, null search should not be done at PV nodes, because PV nodes
should never fail high. But in practice, they DO fail high, because of
aspiration windows, and search inconsistencies, for example. So it makes
sense to keep that flexibility in the code.

No functional change.
2014-02-04 08:24:46 +01:00
Lucas Braesch e88ef801af Better document razoring
Use ralpha instead of rbeta

* rbeta is confusing people. It took THREE attempts to code razoring
at PV nodes correctly in a recent test, because of the rbeta trick.
Unnecessary tricks should be avoided.

* The more correct and self-documenting way of doing this, is to say
that we use a zero window around alpha-margin, not beta-margin.
The fact that, because we only do it at PV nodes, alpha happens to be
beta-1 and that the current stuff with rbeta works, may be correct,
but is confusing.

Remove the misleading and partially erroneous comment about returning
v + margin:

* comments should explain what the code does, not what it could have done.

* this comment is partially wrong in saying that v+margin is "logical",
  and that it is "surprising" that is doesn't work.

From a theoretical perspective, at least 3 ways of doing this are equally
defendable:

1/ fail hard: return alpha: The most conservative. We bet that the search
will fail low, but we don't know by how much and don't want to take risks.

2/ aggressive fail soft: return v (what the current code does). This
corresponds to normal fail soft, with the added assumption that we don't
care about the reduction effect (see below point 3/)

3/ conservative fail soft: return v + margin. If the reduced search (qsearch)
gives us a score <= v, we bet that the non reduced search will give us a
score <= v + margin.

* Saying that 2/ is "logical" implies that 1/ and 3/ are not, which is
arguably wrong. Besides, experimental results tell us that 2/ beats 3/,
and that's not something we can argue against: experimental results are
the only trusted metric.

* Also, with the benefit of hindsight, I don't think the fact that 2/ is
better than 3/ is surprising at all. The point is that it is YOUR turn to
move, and you are assuming that by NOT playing (and letting the opponent
capture your hanging pieces in QS) you cannot generally GAIN razor_margin(depth).

No functional change.
2014-02-03 21:37:14 +01:00
Joona Kiiski d8796fe7e9 H-file penalty and center bonus
After almost 50K at long TC it seems
slightly positive:

ELO: 1.73 +-1.8 (95%) LOS: 97.3%
Total: 47122 W: 7507 L: 7273 D: 32342

bench: 8430785
2014-02-03 21:23:49 +01:00
Joona Kiiski 77341f67f3 Tweak bishop PSQT tables
Tuned after 49K iterations of SPSA.

Passed both short TC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 14231 W: 2684 L: 2542 D: 9005

And long TC:
LLR: 2.94 (-2.94,2.94) [0.00,4.00]
Total: 87556 W: 13757 L: 13342 D: 60457

bench: 6875743
2014-01-29 18:02:13 +01:00
Uri Blass 93f95cc936 Reduce VALUE_KNOWN_WIN to 10000
With some positions like

8/8/8/2p2K2/1pp5/br1p1b2/2p2r2/qqkqq3 w - -

The eval score is higher than VALUE_INFINITE because
is the sum of VALUE_KNOWN_WIN plus a big material
advantage. This leads to an assert. Here are the
steps to reproduce:

Compile SF with debug=yes then do

./stockfish
position fen 8/8/8/2p2K2/1pp5/br1p1b2/2p2r2/qqkqq3 w - -
go depth 1

This patch fixes the issue in this case, but do exsist
other positions for which the patch is not enough and
we will need to limit the eval score to be sure not
overflow the limit.

Note that is not possible to increase the value of
VALUE_INFINITE because should remain within int16_t
type to be stored in a TT entry.

bench: 7356053
2014-01-27 20:09:51 +01:00
Marco Costalba f434cea287 Fix null reduction formula
Depth is already dependent on the actual value
of ONE_PLY, in particular can be expressed like:

Depth = n * ONE_PLY

And because formula is used to calculate R that is
also dependent on the value of ONE_PLY and can be
expressed like:

R = x * ONE_PLY

We don't want to divide depth by a 'ply' value but
directly by an integer number.

Spotted by sf-x

No functional change.
2014-01-27 08:18:48 +01:00
Marco Costalba 216972186e Grammar fix in MovePicker::next_move
Thanks to Lyudmil Antonov and Michel Van den Bergh
for spotting this and suggesting the fix.

No functional change.
2014-01-26 23:09:22 +01:00
Stefan Geschwentner 074c7a3c30 Variable null-move value based reduction
Instead of a fixed reduction of ONE_PLY, now
Null move dynamic reduction based on value can
grow larger in case we are above beta of a value
much higher then PawnValueMg.

Note that now an eval returning VALUE_KNOWN_WIN
makes null search to drop in qsearch.

Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 26141 W: 4871 L: 4699 D: 16571

And long TC:
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 33695 W: 5309 L: 5056 D: 23330

bench: 7356053
2014-01-26 10:23:31 +01:00
Joona Kiiski cf95a55d76 Do not set default value for architeture in Makefile
Fixes a regression that ARCH parameter was not properly validated.
Invalid value would default to generic 32-bit build.

No functional change.
2014-01-25 11:29:32 +01:00
Henri Wiechers a08a21d5a0 Small simplification to Position::see
Verified there are no hidden bugs and is
actually a speed optimization:

Fixed games at 15+0.05 TC
ELO: 1.72 +-2.9 (95%) LOS: 87.5%
Total: 20000 W: 3741 L: 3642 D: 12617

No functional change
2014-01-19 11:16:34 +01:00
Joona Kiiski cabd512916 Fix +M0 score when low on time
When time remaining is less than Emergency Move Time,
we won't even complete one iteration and engine reports
a stale +M0 score.

To reproduce run "go wtime 10"

info depth 1 seldepth 1 score mate 0 upperbound nodes 2 nps 500 time 4 multipv 1 pv a2a3
info nodes 2 time 4
bestmove a2a3 ponder (none)

This patch fixes the issue.

Tested by Binky at very short TC: 0.05+0.05
ELO: 5.96 +-12.9 (95%) LOS: 81.7%
Total: 1458 W: 394 L: 369 D: 695

And at a bit longer TC:
ELO: 1.56 +-3.7 (95%) LOS: 79.8%
Total: 16511 W: 3983 L: 3909 D: 8619

bench: 7804908
2014-01-19 11:09:44 +01:00
Marco Costalba 40c863d41a Increase max hash size to 16GB
TCEC season 3, which is due to start in a few weeks, just
had its server upgraded to 64GB RAM and will therefore allow
16GB hash to be used per engine.

This is almost the upper limit without changing the
type of size and hashMask. After this we need to
move to uint64_t instead of uint32_t.

No functional change.
2014-01-18 18:22:32 +01:00
Joona Kiiski be61edbd19 Tweak knight PSQT tables
Passed both short TC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 31765 W: 6103 L: 5913 D: 19749

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 38867 W: 6268 L: 5988 D: 26611

bench: 7804908
2014-01-18 17:52:37 +01:00
Chris Cain df201175c6 Simplify pawnless endgame evaluation
Retire KmmKm evaluation function. Instead give a very drawish
scale factor when the material advantage is small and not much
material remains.

Retire NoPawnsSF array. Pawnless endgames without a bishop will
now be scored higher. Pawnless endgames with a bishop pair will
be scored lower. The effect of this is hopefully small.

Consistent results both at short TC (fixed games):
ELO: -0.00 +-2.1 (95%) LOS: 50.0%
Total: 40000 W: 7405 L: 7405 D: 25190

And long TC (fixed games):
ELO: 0.77 +-1.9 (95%) LOS: 78.7%
Total: 39690 W: 6179 L: 6091 D: 27420

bench: 7213723
2014-01-18 17:22:54 +01:00
Stefan Geschwentner 53ab32ef0b Introduce 'follow up' moves
When we have a fail-high of a quiet move, store it in
a Followupmoves table indexed by the previous move of
the same color (instead of immediate previous move as
is in countermoves case).

Then use this table for quiet moves ordering in the same
way we are already doing with countermoves.

These followup moves will be tried just after countermoves
and before remaining quiet moves.

Passed both short TC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 10350 W: 1998 L: 1866 D: 6486

And long TC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 14066 W: 2303 L: 2137 D: 9626

bench: 7205153
2014-01-14 09:24:35 +01:00
Ralph Stößer 402a7ae151 Ad-hoc shelter rule
This hacky rule allows to get an about right eval out of this position:
r2qk2r/ppp2p2/2npbn2/2b1p3/2P1P1P1/2NB1PPp/PPNP3K/R1BQ1R2 b kq - 0 13

And, more importantly, passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 6239 W: 1249 L: 1127 D: 3863

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 38371 W: 6165 L: 5888 D: 26318

bench: 8183238
2014-01-12 22:48:08 +01:00
Marco Costalba 3d58092ec1 Retire KBBKN endgame
As pointed out by Joona, Lucas and otehr people in
the forum, this endgame is not a known, there are many
positions where it takes more than 50 moves to claim the
win and becasue exact rules is not possible better to
retire and allow the search to workout the endgame for us.

bench: 8502826
2014-01-12 22:32:41 +01:00
Henri Wiechers db05b1f9f5 Rename Position::hidden_checkers to check_blockers
No functional change.
2014-01-12 04:10:46 +09:00
Henri Wiechers 262c380c4b Position::gives_check - use ci.ksq
Also remove a couple of local variables while
there.

No functional change.
2014-01-09 07:38:04 +09:00
Marco Costalba 6a6fd0b5f5 Fix early stop condition
While editing original Uri's messy patch
I have incorrectly simplified the logic
condition. Here is the correct original
version, as it was tested.

bench: 8502826
2014-01-09 06:58:25 +09:00
Ralph Stoesser f14cd1bb89 Retire easy move
Remove the easy move code and add the condition to
play instantly if only one legal move is available.

Verified there is no regression at 60+0.05
ELO: 0.17 +-1.9 (95%) LOS: 57.0%
Total: 40000 W: 6397 L: 6377 D: 27226

bench: 8502826
2014-01-08 23:57:06 +09:00
Uri Blass a5869d8d25 Stop earlier if iteration is taking too long
If we are still at first move, without a fail-low and
current iteration is taking too long to complete then
stop the search.

Passed short TC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 26030 W: 4959 L: 4785 D: 16286

Long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 18019 W: 2936 L: 2752 D: 12331

And performed well at 40/30
ELO: 4.33 +-2.8 (95%) LOS: 99.9%
Total: 20000 W: 3480 L: 3231 D: 13289

bench: 8502826
2014-01-08 23:45:55 +09:00
renouve 45dbd9cd03 Retire grain size code
Seems useless at long TC.

Tested at 60+0.05
ELO: 2.98 +-2.2 (95%) LOS: 99.6%
Total: 30440 W: 4934 L: 4673 D: 20833

And at 120+0.05
ELO: 2.50 +-2.6 (95%) LOS: 97.1%
Total: 19633 W: 2848 L: 2707 D: 14078

bench: 8502826
2014-01-08 23:22:17 +09:00
Marco Costalba a646f74e6a Assorted grammar fixes
Mainly from Lyudmil Antonov and
one from Henri Wiechers and Louis Zulli.

No functional change.
2014-01-07 14:30:41 +09:00
Joona Kiiski 1fee23a598 Tweak King PST tables
First tested with 50K games at very short TC of 5+0.05
ELO: 3.11 +-2.0 (95%) LOS: 99.9%
Total: 49665 W: 10941 L: 10497 D: 28227

Then retested with usual SPRT at short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16875 W: 3198 L: 3049 D: 10628

And at long TC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 5890 W: 985 L: 857 D: 4048

bench: 7800379
2014-01-07 13:41:16 +09:00
Marco Costalba 1574428f64 Fix a typo
Spotted by Isaac H. Dunn.

No functional change.
2014-01-06 01:22:44 +01:00
Pablo Vazquez 0118623495 Remove duplicated code
Introduce update_stats() and remove correspondng
duplicated code.

No functional change.
2014-01-05 14:10:29 +01:00
H. Felix Wittmann 8454d871ec Ensure move_importance() is non-zero
In case ply is very high, function will round
to zero (although mathematically it is always
bigger than zero). On my system this happens at
movenumber 6661.

Although 6661 moves in a game is, of course,
probably impossible, for safety and to be locally
consistent makes sense to ensure returned value
is positive.

Non functional change.
2014-01-02 13:01:24 +01:00
shane31 153309e287 Scale eval when down to only one pawn
Passed both short TC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 11921 W: 2346 L: 2208 D: 7367

And long TC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 21002 W: 3395 L: 3197 D: 14410

bench: 7602383
2014-01-02 12:44:46 +01:00
Marco Costalba c9dcda6ac4 Update copyright year
No functional change.
2014-01-02 01:49:18 +01:00
Marco Costalba 29e0d8caa7 Simplify move_importance(): take 3
Use pow() of a negative number instead of 1/x

No functional change.
2014-01-01 13:43:58 +01:00
Marco Costalba a795187c50 Further simplify move_importance()
Function move_importance() is already always
positive, so we don't need to add a constant
term to ensure it.

Becuase move_importance() is used to calculate
ratios of a linear combination (as explained in
previous patch), result is not affected. I have
also verified it directly.

No functional change.
2014-01-01 13:35:11 +01:00
H. Felix Wittmann 92faa74dfa Simplify move_importance()
Drop a useless parameter. This works because ratio1 and ratio2
are ratios of linear combinations of thisMoveImportance and
otherMovesImportance and so the yscale cancels out.

Therefore the values of ratio1 and ratio2 are independent
of yscale and yscale can be retired.

The same applies to yshift, but here we want to ensure
move_importance() > 0, so directly hard-code this safety
guard in function definition.

Actually there are some small differences due to rounding errors
and usually are at most few millisecond, that's means below 1% of
returned time, apart from very short intervals in which a difference
of just 1 msec can raise to 2-3% of total available time.

No functional change.
2014-01-01 12:58:10 +01:00
Marco Costalba c5d478b923 Rename pawn chain to connected
The flag raises also in case of a pawn duo, i.e.
when we have two adjacent pawns on the same rank,
and not only in case of a chain, i.e. when the two
pawns are on a diagonal line.

See this for a reference:
http://en.wikipedia.org/wiki/Connected_pawns

Renaming suggested by Ralph.

No functional change.
2014-01-01 10:56:57 +01:00
Gary Linscott 9b1d594456 Remove bishop pin bonus
Shows no regression at LTC after 20K games:

ELO: 0.03 +-2.7 (95%) LOS: 51.0%
Total: 20608 W: 3252 L: 3250 D: 14106

bench: 7516178
2013-12-31 15:27:52 +01:00
Arjun Temurnikar 71440cf77b Retire KingExposed[] array
And merge its values into KPSQT table.

Passed blazingly fast both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 5348 W: 1091 L: 971 D: 3286

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 3029 W: 530 L: 415 D: 2084

bench: 8702197
2013-12-31 12:05:22 +01:00
Henri Wiechers 082fbf4264 Remove asymmThreshold stale comment
No functional change.
2013-12-30 10:34:20 +01:00
Marco Costalba f7742669cb Retire asymmThreshold
Verified with 40K games at long TC does not regress:
ELO: 1.74 +-1.9 (95%) LOS: 96.2%
Total: 39624 W: 6402 L: 6203 D: 27019

bench: 7762310
2013-12-30 01:27:57 +01:00
Matt Sullivan 59a9bc9351 Retire MoveImportance[]
Use a skew-logistic function to replace the
MoveImportance[] array.

Verified it does not regress at fixed number
of games both at short TC:
LLR: -2.91 (-2.94,2.94) [-1.50,4.50]
Total: 39457 W: 7539 L: 7538 D: 24380

And long TC:
ELO: -0.49 +-1.9 (95%) LOS: 31.0%
Total: 39358 W: 6135 L: 6190 D: 27033

bench: 7335588
2013-12-29 10:33:39 +01:00
Stefan Geschwentner 8ca80cb0f1 Fine tune previous patch
Passed short TC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 18331 W: 3608 L: 3453 D: 11270

And scored above 50% on a very long test in long TC
LLR: -2.97 (-2.94,2.94) [0.00,6.00]
Total: 51533 W: 8181 L: 8047 D: 35305

bench: 7335588
2013-12-27 18:46:49 +01:00
Marco Costalba 14aebe2b7c Further simplify previous patch
Use a single XOR instead of NEGATE + AND

No functional change.
2013-12-26 12:08:23 +01:00
Stefan Geschwentner cf0a2a26a9 Bonus for file distance of the outermost pawns
In endgame it's better to have pawns on both wings.
So give a bonus according to file distance between left
and right outermost pawns.

Passed both short TC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 39073 W: 7749 L: 7536 D: 23788

And long TC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 6149 W: 1040 L: 910 D: 4199

bench: 7665034
2013-12-25 21:31:41 +01:00
Ralph Stößer 899a2c033e Loosened trigger condition for king safety
Reduce eval discontinuity becuase now we kick in
king safety evaluation in many more cases.

Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8708 W: 1742 L: 1613 D: 5353

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 6743 W: 1122 L: 990 D: 4631

bench: 6835416
2013-12-23 20:55:30 +01:00
Chris Caino 8aa4f3fde4 Increase pawn king attack weight
Tighter lower bound for pawn attacks so to
activate king safety in some cases like here:

6k1/2B3p1/2Pp1p2/2nPp3/2Q1P2K/P2n1qP1/R6P/1R6 w

Original patch by Chris, further simplified by
Jörg Oster.

Passed both short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 30171 W: 5887 L: 5700 D: 18584

And long TC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 20706 W: 3402 L: 3204 D: 14100

bench: 7607562
2013-12-23 20:52:12 +01:00
Gary Linscott 26689d8c2a Faster and simplified threat eval
Add a bonus according if the attacking
pieces are minor or major.

Passed both short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 13142 W: 2625 L: 2483 D: 8034

And long TC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 18059 W: 3031 L: 2844 D: 12184

bench: 7425809
2013-12-19 18:52:34 +01:00
Marco Costalba f196c1bad8 Further simplify Makefile
No functional change.
2013-12-17 10:16:00 +01:00
Marco Costalba e460ab74ad Reformat Makefile
No functional change.
2013-12-17 10:02:14 +01:00
Lucas Braesch f5727deee3 Remove threat move stuff
A great simplification that shows no regression
and it seems even a bit scalable.

Tested with fixed number of games:

Short TC
ELO: 0.60 +-2.1 (95%) LOS: 71.1%
Total: 39554 W: 7477 L: 7409 D: 24668

Long TC
ELO: 2.97 +-2.0 (95%) LOS: 99.8%
Total: 36424 W: 5894 L: 5583 D: 24947

bench: 8184352
2013-12-15 09:43:29 +01:00
Marco Costalba b96079f86b Sync history and counter moves updating
Change updating rule after a TT hit to match
the same one at the end of the search.

Small change in functionality, but we want to
have uniform rules in the code.

bench: 7767864
2013-12-10 07:05:06 +01:00
Lucas Braesch 86347100a5 Update History and Counter move on TT hit
We already update killers so it is natural to extend to
history and counter move too.

Passed both short TC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 52690 W: 9955 L: 9712 D: 33023

And long TC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 5555 W: 935 L: 808 D: 3812

bench: 7876473
2013-12-10 06:57:06 +01:00
Ralph Stößer 8e9d4081ee Research at intermediate depth if LMR is very high
After a fail high in LMR, if reduction is very high do
a research at lower depth before teh full depth one.
Chances are that the re-search will fail low and the
full depth one is skipped.

Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 11363 W: 2204 L: 2069 D: 7090

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 7292 W: 1195 L: 1061 D: 5036

bench: 7869223
2013-12-09 08:03:47 +01:00
Marco Costalba 4630ab5743 More work on Bitboards::init()
No functional change.
2013-12-07 12:32:18 +01:00
Marco Costalba 1ff5ce8863 More readable init of MS1BTable[]
Because now it uses lsb(), the BSFTable[] must be
already initialized.

No functional change.
2013-12-07 11:35:26 +01:00
Marco Costalba b71418defa Further simplify Bitboards init()
No functional change.
2013-12-07 10:57:05 +01:00
Marco Costalba 47b89f2e37 Clarify definition of capture_or_promotion()
No functional change.
2013-12-06 10:43:17 +01:00
Arjun Temurnikar 431c3ac485 Even more spelling fixes
No functional change.
2013-12-06 09:03:24 +01:00
Arjun Temurnikar 190aea4cdc Assorted spelling/grammar/captitalization
No functional change.
2013-12-04 23:30:37 +01:00
Chris Caino f026517e5e Micro-optimise dangerous condition
Since all ENPASSANT moves are now considered dangerous, this
change of order should give a slight speedup.

Also simplify futilityValue formula.

No functional change.
2013-12-04 17:51:25 +01:00
Marco Costalba b71cedb2b0 Retire TheirHalf[]
We avoid to use an ad-hoc table at the cost of a
relative_rank() call in advanced_pawn_push().

On my 32 bit system it is even slightly faster (on 64bit
may be different). This is the speed in nps alternating
old and new bench runs:

new

368890
368825
369972

old

367798
367635
368026

No functional change.
2013-12-04 17:45:09 +01:00
Chris Caino 69a14554ee Broader condition for dangerous pawn moves
Instead of a passed pawn now we just require the pawn to
be in the opponent camp to be considered a dangerous
move. Added some renaming to reflect the change.

Passed both short TC test
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 10358 W: 2033 L: 1900 D: 6425

And long TC
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 21459 W: 3486 L: 3286 D: 14687

bench: 8322172
2013-12-04 17:19:45 +01:00
Marco Costalba 2408243cf4 Shrink Position::is_draw()
No functional change.
2013-12-03 11:40:31 +01:00
Marco Costalba 342fd6385b Remove redundant argument from hidden_checkers()
No functional change.
2013-12-03 11:15:39 +01:00
Marco Costalba e6c9ce6358 Small improvment to Position::fen()
No functional change.
2013-12-03 10:53:21 +01:00
Jerry Donald c7e7d9217b Re-fix a comment
No functional change.
2013-12-03 08:50:12 +01:00
Jerry Donald a8af78c833 Another round of spelling fixes
And also renamed a loop variable while there.

No functional change.
2013-12-02 23:51:29 +01:00
Richard Lloyd 13a73f67c0 Big assorted spelling fixes
No functional change.
2013-12-02 20:29:35 +01:00
Jerry Donald 500b9b0eb3 Assorted spelling fixes
No functional change.
2013-12-02 18:41:30 +01:00
Marco Costalba f99cb3dc27 Rename CASTLE to CASTLING
It is call 'castling move', not 'castle move'

Noticed while reading DiscoCheck sources.

No functional change.
2013-12-01 11:16:47 +01:00
Marco Costalba dd4e5db2be Simplify a condition in gives_check()
Now that aligned() is quite fast we can skip
some logic.

No functional change.
2013-12-01 10:09:30 +01:00
Marco Costalba 5f2bf91ad1 Rename Bitboards print to pretty
To align to same named Position function and
avoid using std::cout directly.

Also remove some stale <iostream> include while
there.

No functional change.
2013-11-30 11:32:49 +01:00
Marco Costalba 034a2b04f2 Rewrite some bitboard init code
And move the static function Position::attacks_from() to
bitboard code renaming it attacks_bb()

No functional change.
2013-11-30 11:02:56 +01:00
Daylen Yang 6ea5dc294c Makefile improvements for compiling on OS X
Add a Mac SSE4.2 target. Also change the Mac OS X minimum version to
10.6. Rationale: 97% of Macs run at least 10.6, version 10.9 is now
free, and using 10.6 as the minimum version gives a small 5% boost in
benchmark speed over versions using 10.0 as the minimum version.

Finally, enable Clang’s Link Time Optimization when compiling for the
Mac.

No functional change.
2013-11-30 00:09:17 +01:00
Marco Costalba 8f5deaea61 Restore development version
bench: 8596156
2013-11-29 10:50:43 +01:00
Marco Costalba c5bb9b9da9 Stockfish DD
Stockfish bench signature is: 8596156
2013-11-29 10:23:14 +01:00
Kelly Wilson 67e5581e37 Add support for PPC 64bit on Linux
In particular Debian Linux-3.9.8-1- PPC64

No functional change.
2013-11-29 10:06:54 +01:00
Joona Kiiski 3ede7daab3 Generate Qsearch checks only at depth 0
An old idea retested at SPRT(0, 3) with 60+0.05 TC:
LLR: 2.95 (-2.94,2.94) [0.00,3.00]
Total: 98872 W: 15549 L: 15123 D: 68200

This is a very small elo increase patch so it really
stresses the limits of fishtest.

bench: 8596156
2013-11-24 10:26:49 +01:00
Marco Costalba dbd6156fce Revert previous fix
It seems to intorduce a regression when tested
with 3 threads at 15+0.05:

ELO: -2.26 +-2.2 (95%) LOS: 2.4%
Total: 30000 W: 4813 L: 5008 D: 20179

bench: 8331357
2013-11-19 07:20:50 +01:00
Hongzhi Cheng 691ed425ba Get correct excluded moves for split nodes
Tested setting FakeSplit to true and running

./stockfish bench 128 2

There is a different signature with and without
the patch so it affects functionality but
only in SMP case.

bench: 8331357
2013-11-18 16:41:49 +01:00
Marco Costalba c376ffce0f Revert previous patch
It seems a regression at 15+0.05:
ELO: -4.82 +-2.1 (95%) LOS: 0.0%
Total: 40000 W: 7181 L: 7736 D: 25083

bench: 8331357
2013-11-17 23:47:18 +01:00
Marco Costalba 917944e9c5 Fix an assert in SMP case
SMP case is very tricky and raises an assert in stage_moves():

assert(stage == KILLERS_S1 || stage == QUIETS_1_S1 || stage == QUIETS_2_S1)

So rewrite the code to just return moves[] when we are sure
we are in quiet moves stages.

Also rename stage_moves to quiet_moves to reflect that.

No functional change (but needs testing in SMP case)
2013-11-17 10:24:25 +01:00
Marco Costalba d9c7cad630 Retire quietsSearched[]
Use MovePicker moves[] to access already tried
quiet moves. A bit of care shall be taken
to avoid calling stage_moves() when we are still
at ttMove stage, because moves are yet to be
generated. Actually our staging move generation
makes this code a bit more tricky than what I'd
like, but removing an ausiliary redundant
array like quietsSearched[] is a good thing.

Idea by DiscoCheck

bench: 9355734
2013-11-17 09:51:04 +01:00
Marco Costalba 9763c69fa5 Simplify generate<EVASIONS>
Use the newly introduced LineBB[] to simplify this
super hot-path function.

Verified with perft we don't have any speed regression, although
the number of squares removed is less than before in case of
contact check.

Insipred by DiscoCheck implementation.

Perft numbers are the same, but we have an harmless functional
change due to reorder of moves, because now some illegal moves
are no more detected at generation time, but in the search.

bench: 8331357
2013-11-11 19:53:19 +01:00
Gregor Cramer 555d9a8711 Faster castling in Chess960 case
Only rook attackers has to be considered, all other attackers are
already handled in the lines above.

No functional change.
2013-11-11 15:55:08 +01:00
Joona Kiiski b9768b8bc5 Reintroduce gains
This seems a die hard idea :-)

Passed both short TC
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 17485 W: 3307 L: 3156 D: 11022

And long TC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 38181 W: 6002 L: 5729 D: 26450

bench: 8659830
2013-11-11 08:46:11 +01:00
Jörg Oster 4ef6b2c32a Remove opposed flag for doubled pawns
Actually, it is not used, as both arrays have the
same values. Some local tests in either direction
showed no improvement.

Also some minor corrections in the comments.

No functional change.
2013-11-10 17:21:59 +01:00
Marco Costalba 1d18647e73 Rename squares_aligned()
Rename to the shorter but still
clear aligned()

No functional change.
2013-11-10 17:14:46 +01:00
Marco Costalba a518d5d3ad Simplify squares_aligned()
Use newly introduced LineBB[]

No functional change.
2013-11-10 12:05:19 +01:00
Chris Caino 091aff0445 Evaluate mobility of pinned pieces exactly
Previously some squares could be "incorrectly" awarded
to a pinned piece.

e.g. in 3k4/1q6/3b4/3Q4/8/5K2/B7/8 b - - 0 1 the black
bishop get 4 squares too many and the white queen gets 6.

Passed both short TC.
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 4871 W: 934 L: 817 D: 3120

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 38968 W: 6113 L: 5837 D: 27018

bench: 9282549
2013-11-10 11:52:38 +01:00
Chris Caino 3ed86ed3f9 Remove RedundantMajor
But compensate by reducing rook and queen
value by 53 = (160 / 3)

Material imbalances are affected as follows:

       Red. Major   Rook  Queen  Total
QRR      +160      -2*53    -53     +1
QR       +160        -53    -53    +54
RR       +160      -2*53      0    +54
R           0        -53      0    -53
Q           0          0    -53    -53

so that the imbalance changes by at most 54 + 53 = 107 units.
This corresponds to appromximately 3.5cp in the final evaluation.

Verified with fixed number 40000 games at both short
and long TC it does not regress.

Short TC 15+0.05
ELO: 1.93 +-2.1 (95%) LOS: 96.6%
Total: 40000 W: 7520 L: 7298 D: 25182

Long TC 60+0.05
ELO: -0.33 +-1.9 (95%) LOS: 36.5%
Total: 39663 W: 6067 L: 6105 D: 27491

bench: 6703846
2013-11-10 09:48:49 +01:00
Marco Costalba e4d34e1815 Fix printing of incorrect PV in some cases
As, Gary (that analyzed the bug) says:

SF does not print a PV when the original best move fails low,
we hit our time allowance, and stop the search.  The output from
the SF search is below.  It was failing low on Ne1 at depth 34.
Then, we get bestmove Qd3, but no PV change.

info depth 34 seldepth 45 score cp 38 upperbound nodes 483484489 nps 15464575 time 31264 multipv 1 pv f3e1 h5h4 e1d3 h4g3 f2g3 a6f6 f1f6 e7f6 d1a4 f6e7 a1f1 d8f8 a4b3 b7b6 b3c2 f7f6 c2a4 h3g5 b2b3 g5f7 a4c6 f7d6 h1g2 f6f5 e4f5 d6f5
info depth 34 seldepth 45 score cp 38 upperbound nodes 483484489 nps 15464575 time 31264 multipv 1 pv f3e1 h5h4 e1d3 h4g3 f2g3 a6f6 f1f6 e7f6 d1a4 f6e7 a1f1 d8f8 a4b3 b7b6 b3c2 f7f6 c2a4 h3g5 b2b3 g5f7 a4c6 f7d6 h1g2 f6f5 e4f5 d6f5
info depth 34 seldepth 47 score cp 30 upperbound nodes 2112334132 nps 17255517 time 122415 multipv 1 pv f3e1 h5h4 d1a4 a6f6 e1d3 d8f8 a4c2 h4g3 f2g3 f6f1 a1f1 h7g8 b2b3 f7f6 a3a4 b7b6
info depth 34 seldepth 47 score cp 30 upperbound nodes 2112334132 nps 17255517 time 122415 multipv 1 pv f3e1 h5h4 d1a4 a6f6 e1d3 d8f8 a4c2 h4g3 f2g3 f6f1 a1f1 h7g8 b2b3 f7f6 a3a4 b7b6
info nodes 18235667001 time 969824
bestmove e2d3 ponder c8d7

Looking at the code, if we hit Signals.stop, we return from id_loop
before printing any PV.  It is possible for us to have resorted the
RootMove list though, which will change the move that is actually
played.

No functional change.
2013-11-09 19:05:43 +01:00
Marco Costalba 42caebfaa5 Fix compile in debug mode
No functional change.
2013-11-09 18:41:51 +01:00
Lucas Braesch eed508b444 Futility pruning simplification
1/ eval margin and gains removed:
16bit are now free on TT entries, due to the removal of eval margin. may be useful
in the future :) gains removed: use instead by Value(128). search() and qsearch()
are now consistent in this regard.

2/ futility_margin()
linear formula instead of complex (log(depth), movecount) formula.

3/ unify pre & post futility pruning
pre futility pruning used depth < 7 plies, while post futility pruning used
depth < 4 plies. Now it's always depth < 7.

Tested with fixed number of games both at short TC:
ELO: 0.82 +-2.1 (95%) LOS: 77.3%
Total: 40000 W: 7939 L: 7845 D: 24216

And long TC
ELO: 0.59 +-2.0 (95%) LOS: 71.9%
Total: 40000 W: 6876 L: 6808 D: 26316

bench 7243575
2013-11-09 10:17:27 +01:00
Marco Costalba 343544f3f7 Revert "Retire eval margin and gains"
This reverts commit ecd07e51d0.

Patch was incorrect and partial. It will be reapplied in
the correct form.

bench: 9189063
2013-11-07 22:32:13 +01:00
Gary Linscott 13d1f0ae43 Restrict mobility of pinned pieces
Passed both short TC:
LLR: 3.00 (-2.94,2.94) [-1.50,4.50]
Total: 54342 W: 10950 L: 10692 D: 32700

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 61976 W: 10654 L: 10251 D: 41071

This patch introduces a slowdown of 3.5 % !!!!!

bench: 7911558
2013-11-07 22:26:03 +01:00
Lucas Braesch ecd07e51d0 Retire eval margin and gains
1/ eval margin and gains removed:
 - gains removed by Value(128): search() and qsearch() now behave consistently!

2/ futility_margin()
 - testing showed that there is no added value in this weird (log(depth), movecount)
   formula, and a much simpler linear formula is just as good. In fact, it is most
   likely better, as it is not yet optimally tuned.
 - the new simplified formula also means we get rid of FutilityMargins[], its
   initialization code, and more importantly ss->futilityMoveCount, and the hacky
   code that updates it throughout the search().
 - the current formula gives negative futility margins, and there is a hidden interaction
   between the move coutn pruning formula and the futility margin one: what happens is
   that MCP is supposed to be triggered before we use the non-sensical negative futility
   margins.

3/ unify pre & post futility pruning
 - pre futility pruning (what SF calls value based pruning) used depth < 7 plies,
   while post futility pruning (what SF calls static null move pruning) used depth < 4 plies.
 - also the condition depth < 7 in pre futility pruning was not obvious, and it seemd
   to be depth < 16 (futility_margin() returns an infinite value when depth >= 7).

Tested with fixed number of games both at short TC:
ELO: 0.82 +-2.1 (95%) LOS: 77.3%
Total: 40000 W: 7939 L: 7845 D: 24216

And long TC
ELO: 0.59 +-2.0 (95%) LOS: 71.9%
Total: 40000 W: 6876 L: 6808 D: 26316

bench: 10206576
2013-11-07 19:46:51 +01:00
Chris Caino 52ae0efccf Two more parameters eliminated
RedundantRook and RedundantQueen replaced by simple
variable RedundantMajor. Also the SameColor coefficient
for Queen<->Queen has been set by definition to 0.

The remaining 5 parameters:

LinearCoefficients[ROOK]
LinearCoefficients[QUEEN]
QuadraticCoefficientsSameColor[ROOK][ROOK]
QuadraticCoefficientsSameColor[QUEEN][ROOK]
RedundantMajor

are sufficient to equate the material imbalances for the
5 common material configurations of R, RR, Q, QR and QRR
to any desired values simultaneously.

With the chosen parameters there should be no functional
change unless one side has more than 2 rooks or more
than 1 queen. For example bench from the start position
using the commands:

./stockfish
go depth 16

produces identical output except for one extra node
in the last iteration.

bench: 8198094
2013-11-07 19:20:24 +01:00
Chris Caino 53c04c0429 Zero more redundant coefficients
Coefficients for Bishop<->BishopPair and Bishop<->Bishop
are also pretty much redundant. By altering the values
in LinearCoefficients[] these coefficients can be zeroed
without changing the imbalance calculations in any position
with less than 3 bishops for one side.

bench: 7995098
2013-11-05 20:08:39 +01:00
Chris Caino 1064288b38 Zero redundant material imbalance terms
First coefficient in the SameColor array does an
equivalent job when folded into the LinearCoefficients
array.

All of the diagonal terms in the OppositeColor array
are redundant due to cancellation.

No functional change.
2013-11-05 20:08:39 +01:00
Joona Kiiski d34bb889b1 Test Easy Move if no BestMoveChanges
In case we find a very good move after a
troubled start, we don't return immediately
anymore.

Tested directly at long TC where it passed:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 13910 W: 2397 L: 2228 D: 9285

bench: 7995098
2013-11-02 11:34:42 +01:00
Marco Costalba a3a0df92a3 Set timer to a fixed interval
And remove a complex (and broken) formula.

Indeed previous code was broken in case of TC with big
time increments where available_time() was too similar
to total time yielding to many time losses, so for instance:

go wtime 2600 winc 2600
info nodes 4432770 time 2601 <-- time forfeit!

maximum search time = 2530 ms
available_time = 2300 ms

For a reference and further details see:

https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/dCPAvQDcm2E

Speed tested with bench disabling timer alltogheter vs timer set at
max resolution, showed we have no speed regressions both in single
core and when using all physical cores.

No functional change.
2013-11-01 08:56:15 +01:00
Ralph Ster e8f9447b11 Use a formula for chain membership bonus
Passed both short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 5087 W: 1072 L: 951 D: 3064

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 28620 W: 5042 L: 4798 D: 18780

bench: 7995098
2013-10-31 06:13:30 +01:00
Marco Costalba 3cc47edf62 Tweak bishop pair and knight weight
A combo of two patches that failed SPRT with score
higher than 50% but togheter they succeed:

SPRT at 60+0.05
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 7312 W: 1276 L: 1139 D: 4897

bench: 8029334
2013-10-28 19:31:25 +01:00
Matthew Sullivan d454cd4216 Fix divide by zero bug in late game
If the game got late enough that move_importance(currentPly) * slowMover / 100
rounds to 0, then we ended up dividing 0 by 0 when only looking 1 move ahead.

This apparently caused the search to almost immediately abort and Stockfish
would blunder in long games. So convert thisMoveImportance to a double.

No functional change.
2013-10-27 08:03:58 +01:00
Marco Costalba 48f38f3092 Retire mirror()
Inline the only caller site.

No functional change.
2013-10-24 20:40:26 +02:00
Marco Costalba 281472e50e Prefer file_bb() to FileBB[]
No functional change.
2013-10-24 20:34:23 +02:00
Jörg Oster f011a5af11 Penalty for Knight when enemy pawns are few
This seems more a material imbalance topic,
anyhow test is good and so patch is applied
as is.

Passed both short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 17391 W: 3548 L: 3393 D: 10450

And long TC:
LLR: 3.00 (-2.94,2.94) [0.00,6.00]
Total: 34660 W: 5972 L: 5700 D: 22988

bench: 8291883
2013-10-24 20:11:33 +02:00
Marco Costalba 67b0da83da Further smplify pawn endgames
Dumb down a bit the code and trade some possible
speed (but this is far from hot path anyhow) for
some added readability for the layman.

No functional change.
2013-10-23 19:56:35 +02:00
Chris Caino 3674f18b97 Use flip_sq idea in endgame.cpp
The normalising transformation is computed all at
once by the helper function get_flip_sq and then
applied immediately to the relevant squares as soon
as they are loaded from the position class.

bench: 8350690
2013-10-23 16:00:49 +02:00
Chris Caino 72f7282ad4 Simplify futility move count formula
Simpler formula but introduces some slight changes if d >= 10

Original code grows like  0.225 * d^1.8
New code grows like       0.222 * d^1.8

Full list of values:

d old new diff
--------------
0 2 2 0
1 2 2 0
2 3 3 0
3 4 4 0
4 5 5 0
5 6 6 0
6 7 7 0
7 9 9 0
8 11 11 0
9 13 13 0
10 15 16 1
11 18 19 1
12 21 21 0
13 24 24 0
14 27 28 1
15 31 31 0
16 35 35 0
17 39 38 -1
18 42 42 0
19 47 46 -1
20 51 51 0
21 55 55 0
22 60 60 0
23 65 65 0
24 70 70 0
25 75 75 0
26 81 80 -1
27 87 86 -1
28 92 91 -1
29 98 97 -1
30 104 103 -1
31 111 109 -2

Test code:

int main() {

  for(int d=0; d<32; d++)
  {
     int a = int(3 + 0.3 * pow(double(d), 1.8)) * 3/4 + (2 < d && d < 5);
     int b = int(2.4 + 0.222 * pow(d + 0.0, 1.8));

     std::cout << d << " " << a << " " << b << " " << b-a << std::endl;
  }

  return 0;
}

bench: 8350690
2013-10-22 23:09:40 +02:00
Chris Caino fbfce2132a Simplify futility margins formula
New formula mathces the old formula until d = 45

Test code:

int main() {

  for(int d=1; d<=45; d++)
  {
     int a = int(log(double(d * d) / 2) / log(2.0) + 1.001);
     int b = int(2.9 * log(double(d)));

     if (a != b) std::cout << d << std::endl;
  }

  return 0;
}

bench: 8455956
2013-10-22 23:06:06 +02:00
Marco Costalba 2c825294ec Tweak again chain pawn bonus
This is the first chain bonus version
from Ralph that also passed both

Short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 23460 W: 4727 L: 4556 D: 14177

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 31858 W: 5497 L: 5240 D: 21121

And performed better against current
committed version, always at 60secs:

LLR: -2.94 (-2.94,2.94) [-3.00,3.00]
Total: 26301 W: 4477 L: 4580 D: 17244

This test was done by Leonid.

bench: 8455956
2013-10-22 17:47:16 +02:00
Marco Costalba f86d2aee29 Re-add "Further increase safe checks bonus"
After 40K games at 60 secs, result is still
not clear, but not a regression against SF 4

After
ELO: 50.11 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 10547 L: 4817 D: 24636

Before
ELO: 49.51 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 10483 L: 4821 D: 24696

So re-apply the patch to avoid to
special-case this one.

bench: 7403882
2013-10-22 17:33:11 +02:00
Marco Costalba 35ea39bed2 Restore behaviour after count<ALL_PIECES> fix
Because pos.count<ALL_PIECES>(Us) was always zero,
rewrite the formula as if this would still be
the case.

bench: 8510004
2013-10-22 17:27:58 +02:00
Ralph Stößer 97015afce8 Further improve chain pawn evaluation
Passed both short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 28299 W: 5854 L: 5667 D: 16778

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 9738 W: 1797 L: 1644 D: 6297

bench: 9294116
2013-10-20 23:41:40 +02:00
Marco Costalba f22a63ce67 Fix pos.count<ALL_PIECES>()
It was never updated !

Currently it only affects evaluate_passed_pawns()
and in particularly the rule to increase the bonus
if we have more non-pawn pieces. We could simply use
popcount() instead and avoid the little slowdown
in put_piece() and remove_piece(), but this would
leave a very subtle and tricky hole where people
are forced to remember that pos.count<ALL_PIECES>()
does not work. This is not obvious and so dangerous.

Thanks to Ronald de Man for spotting this.

bench: 7931424
2013-10-20 23:36:46 +02:00
Marco Costalba c08e7419a0 Fix build on Intel compiler
Due to a strange issue (bug?) the ternary
operator does not return a BitCountType for
icc, so revert to the expression.

The same patch was already applied in
9749f1f14c

Thanks to NssY Wanyonyi for pointing out
this.

No functional change.
2013-10-20 23:19:08 +02:00
Marco Costalba 67f91bc5ea Revert "Further increase safe checks bonus"
This reverts commit 4bc2374450 for
two reasons.

First regression testing shows almost equal
score:

Before the patch:
ELO: 49.75 +-2.5 (95%) LOS: 100.0%
Total: 27205 W: 7113 L: 3244 D: 16848

After the patch:
ELO: 48.87 +-2.9 (95%) LOS: 100.0%
Total: 20860 W: 5478 L: 2563 D: 12819

Second, and more sensible to me, this patch
increases safe check bonuses to 4 times their
original value (!) and considering:

- Values were already well tuned

- Values are highly critical

- King safety is highly critical, very TC
  dependent and very difficult to test

- Our testing coverage is partial (self-testing,
  blitz times)

I think is better to be safe than sorry and so
I revert the patch.

bench: 8440524
2013-10-20 10:04:43 +02:00
Ralph Stößer 4bc2374450 Further increase safe checks bonus
Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 10466 W: 2087 L: 1953 D: 6426

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 26334 W: 4540 L: 4310 D: 17484

And also proved stronger than a slightly
different patch, also succesful against master:

https://github.com/mcostalba/Stockfish/commit/dc6830a3b4ed12

But losing against current one in a match
at 60secs with SPRT [-3, 3]:

LLR: -2.96 (-2.94,2.94) [-3.00,3.00]
Total: 44484 W: 7360 L: 7463 D: 29661

bench: 9160831
2013-10-19 12:19:36 +02:00
Marco Costalba f5e872a0e3 Some evaluation code reshuffle
No functional change.
2013-10-18 09:49:38 -07:00
Jörg Oster 25cb851f8a Score chain pawn also by rank
Use the (rescaled) CandidatePassed[] table
that is already rank based.

Passed both short TC
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 11048 W: 2272 L: 2135 D: 6641

And long TC
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 4116 W: 769 L: 645 D: 2702

bench: 8440524
2013-10-18 10:27:55 +02:00
Chris Caino 3cddb0c076 Simplification of KPsK function
Also the drawing criteria has been slightly loosened.
It now detects a draw if the king is ahead of all the
pawns and on the same file or the adjacent file.

bench: 7700683
2013-10-15 07:36:01 +02:00
Chris Caino 2bf18bfc63 Bug fix for KQKRPs endgame
This lost position 8/8/3q4/8/5k2/2P1R3/2K2P2/8 w - - 0 1
was previously evaluated as a draw.

The king and rook need to be correctly placed with
respect to the _same_ pawn.

(Note also that the check for the pawn being on RANK_2
in the old version is redundant: it must be on RANK_2 if
it hopes to protect a rook on RANK_3)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-10-15 07:34:29 +02:00
Ralph Stößer 5aeb907fa1 Double king safety weights
Good both at short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 5448 W: 1133 L: 1012 D: 3303

And at long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 40509 W: 6836 L: 6541 D: 27132

bench: 7700683
2013-10-14 23:24:29 +02:00
Chris Caino 0c68971c13 Remove a drawing rule from KBPsK function
The rule can be incorrect if the attacking king is
well placed e.g. 8/6K1/8/8/7k/1B6/7P/8 w - - 0 1

bench: 8279065
2013-10-14 19:55:07 +02:00
Marco Costalba d9be00342c Massive stronger/weaker renming
No functional change.
2013-10-14 19:38:08 +02:00
Chris Caino 027d85e82a Add helper function verify_material
Allows to remove a lot of assert code in endgames.

No functional change.
2013-10-14 19:23:57 +02:00
ceebo 3bc3c069f1 Add some knowledge for KRPKB endgame
bench: 8279065
2013-10-14 07:44:02 +02:00
ceebo e9366fa155 Improve KBPsK endgame
Better endgame with bishop and blocked g-pawn

bench: 8279065
2013-10-14 07:44:01 +02:00
Marco Costalba 549b5c478f Remove unuseful optimization in RKISS
Don't need a struct here. Speed test shows
result is teh same. Moreover RKISS is used
mainly at startup to compute magics, so
prefer to keep it simple...RKISS ;-)

Also some assorted triviality while there.

No functional change.
2013-10-13 03:35:17 -07:00
Joona Kiiski b15e148b5e Smoother transition for LMR
Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 12376 W: 2596 L: 2454 D: 7326

And long TC:
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 14798 W: 2584 L: 2409 D: 9805

bench: 8279065
2013-10-09 19:13:41 +02:00
Uri Blass bb83a417cb Increase slowmover and reduce instability
These two changes go in opposite directions and it
seems that the combination is stronger than original.

Here are the positive tests at various TC:

15+0.05
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 24561 W: 4946 L: 4772 D: 14843

60+0.05
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 15259 W: 2598 L: 2423 D: 10238

40/30
LLR: 2.96 (-2.94,2.94) [-3.00,3.00]
Total: 2570 W: 527 L: 422 D: 1621

Unfortunately there is also a bad result
with one sec time increment that needs
to be further investigated:

12+1
LLR: -2.97 (-2.94,2.94) [-3.00,3.00]
Total: 2694 W: 438 L: 543 D: 1713

bench: 8340585
2013-10-08 21:24:21 +02:00
Lucas Braesch 984ee9d05b Use TT refined value to stand pat
Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 17811 W: 3520 L: 3366 D: 10925

And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 30255 W: 5070 L: 4825 D: 20360

bench: 8340585
2013-10-08 19:51:08 +02:00
Lucas Braesch a0cc15ccbc Use double everywhere
Rationale:

- Speed of double and float is about the same (not on the hot path anyway)

- Double makes code prettier (no need to write 1.0f, just 1.0)

- Only practical advantage of float is to use less memory, but since we never
  store large arrays of double, we don't care.

No functional change.
2013-10-05 18:12:52 +02:00
Lucas Braesch 7f142d6817 Use prefix operators wherever possible
No functional change.
2013-10-05 18:10:43 +02:00
Marco Costalba bd1c3ed7e3 Add more depth/positions to bench
Increase bench default depth from 12 to 13 and
add 15 new endgame positions to have broader
coverage and also more reliable nps calulcation
used for fishtest framework.

Due to the new endgame positions, where nps is higher,
the total nps is increased of about 15%.

Thanks to Lucas and Jörg for the suggestions.

No functional change, but bench number is now:

bench: 8336338
2013-09-29 09:43:10 +02:00
Marco Costalba cca34e234c Drop 'is' prefix from query functions
Most but not all.

No functional change.
2013-09-28 06:47:59 -07:00
Marco Costalba ed95ad1c0e Fix build on Mac OS X
For some users -stack_size,0x4000 does not work,
so revert for now.

osX 10.6.8
gcc version 4.7.3 (MacPorts gcc47 4.7.3_2)

g++: error: unrecognized command line option '-stack_size,0x4000'
make[2]: *** [stockfish] Error 1
make[1]: *** [gcc-profile-make] Error 2
make: *** [profile-build] Error 2

No functional change.
2013-09-28 04:16:16 -07:00
Marco Costalba c65d67feb5 Revert "Use a per-thread array"
This reverts commit 800410eef1 and instead increases
stack size.

I went through the old emails with Daylen that reported the
crash issue on Mac OS X and was fixed by 0049d3f337.

It was reported default stack size for a thread in Mac OS X is 8
megabytes while the patch that we are reverting allows to reduce
stack size at max of about 217KB, so the reason for the crash was
only marginal in MAX_MOVES value. On those emails Daylen also
hinted how to increase stack size for Mac OS X to 16MB.

So prefer to increase stack size to 16MB instad of re-inventing
the wheel and do our home grown stack as we did with the patch
that we are now reverting (it will remain anyhow in git history
for documentation purposes).

No functional change.
2013-09-28 10:10:51 +02:00
Lucas Braesch bc6faf633e Simplify extensions
Unify extensions between PV and not PV nodes
and remove all but check extensions.

This is a simplification so tested at fixed number
of games where proved to not regress.

About 45k games at 15+0.05
ELO: 1.23 +-2.0 (95%) LOS: 88.5%
Total: 45643 W: 9107 L: 8946 D: 27590

About 45k games at 60+0.05
ELO: 1.07 +-1.8 (95%) LOS: 87.8%
Total: 46786 W: 7728 L: 7584 D: 31474

bench: 3172206
2013-09-28 09:54:22 +02:00
Reuven Peleg 8d1c1074d5 Simplify tte use condition
No functional change.
2013-09-27 09:40:48 +02:00
Raminder Singh e654209211 Fix best move lookup bug
If the uci option 'Best Book Move' is set to true the lookup still
returns a move at random instead of the move with the highest
weight.

No functional change.
2013-09-27 09:04:24 +02:00
Marco Costalba b742a3f29a Increase MAX_MOVES to 256
This should be enough for any legal position, even
the handcrafted ones, like the one presented by Reuven:

1Q5R/4Q1K1/B1Q5/B4Q2/N2Q4/pQ4Q1/pn2Q3/krQ4R w - -

Where currently we crash. This reverts the patch
0049d3f337 of 8/4/2012 where stack
was shrinked due to crashes while in deep analysys.

No functional change.
2013-09-27 08:59:03 +02:00
Marco Costalba 800410eef1 Use a per-thread array for generated moves
This greately reduces stack usage and is a
prerequisite for next patch.

Verified with 40K games both in single and SMP
case that there are no regressions.

No functional change.
2013-09-27 08:44:36 +02:00
Jean-Francois Romang 7b2cda95d9 Update disabled warnings for Intel compiler
No functional change.
2013-09-23 08:11:43 +02:00
Jean-Francois Romang c2cefa6de0 Replace -O3 with -fast for intel compiler
No functional change.
2013-09-23 08:11:14 +02:00
Ralph Stößer d7f5f15d69 Reduce negative quiets by ONE_PLY / 2
Passed both short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 3402 W: 708 L: 593 D: 2101

And long TC:
LLR: 2.93 (-2.94,2.94) [0.00,6.00]
Total: 23379 W: 3972 L: 3759 D: 15648

bench: 3529630
2013-09-23 08:08:52 +02:00
Marco Costalba 84906b83ec Final time management setup
This is an even safer setup proposed and tested
by Alexandre Meirelles.

Regression testing of 40K games at 10+0.05 show
result is stable both against current master:

ELO: -0.29 +-2.2 (95%) LOS: 39.7%
Total: 40000 W: 8010 L: 8043 D: 23947

and again original master (the one with smallest
time parameters):

ELO: 1.71 +-2.2 (95%) LOS: 93.8%
Total: 40000 W: 8325 L: 8128 D: 23547

Alexandre verified with LittleBlitzer time losses are
greately reduced with this setup:

Games Completed = 2100 of 3000 (Avg game length = 35.745 sec)

Settings = RR/128MB/15000ms+50ms/M 1000cp for 12 moves, D 150 moves/
Time = 39200 sec elapsed, 16800 sec remaining
 1.  Stockfish 190913             1091.5/2100    803-720-577      (L: m=313 t=1 i=0 a=406)    (D: r=278 i=91 f=136 s=8 a=64)    (tpm=212.5 d=14.75 nps=925427)
 2.  Houdini 2.0 w32              1008.5/2100    720-803-577      (L: m=250 t=299 i=0 a=254)    (D: r=278 i=91 f=136 s=8 a=64)    (tpm=204.1 d=12.04 nps=1326351)

No functional change.
2013-09-23 07:59:51 +02:00
Marco Costalba 274079990a Increase Emergency Move Time to 20
Goes in the direction of avoiding time losses and seems
equivalent after almost 40K games at super fast TC of 10+0.05

ELO: 2.61 +-2.2 (95%) LOS: 99.1%
Total: 39869 W: 8258 L: 7959 D: 23652

No functional change.
2013-09-19 07:26:36 +02:00
Marco Costalba 10cb19d534 Increase Emergency Move Time to 10
Goes in the direction of avoiding time losses and seems
equivalent after almost 40K games at super fast TC of 10+0.05

ELO: 2.41 +-2.3 (95%) LOS: 98.1%
Total: 37222 W: 7843 L: 7585 D: 21794

No functional change.
2013-09-17 16:32:39 +02:00
Marco Costalba d50b33cacd Fix a silly unstoppable eval bug
The logic is broken for black side because we get more bonus
for pawn in 7th rank than for pawn in 2nd rank!

Spotted by Reuven Peleg

bench:3884409
2013-09-16 23:59:37 +02:00
Joona Kiiski 77b5ee0117 Fix time parameters for blitz games
The ideal setting for super-blitz might be something like:

    "Emergency Base Time" = 50
    "Emergency Move Time" = 5

This would give a total emergency time buffer of:

    50 + 40 * 5 = 250 ms

This setup replaces the previous half cooked hack
"Don't blunder under extreme time pressure".

Test results are very good at super blitz, but keep good even
at 60 secs.

At 5+0.05
ELO: 24.30 +-2.4 (95%) LOS: 100.0%
Total: 37802 W: 10060 L: 7420 D: 20322

At 15+0.05
ELO: 13.41 +-2.9 (95%) LOS: 100.0%
Total: 22271 W: 4853 L: 3994 D: 13424

At 60+0.05
ELO: 5.30 +-3.2 (95%) LOS: 99.9%
Total: 16000 W: 2897 L: 2653 D: 10450

No functional change.
2013-09-16 09:07:47 +02:00
Marco Costalba af750bd2ef Rewrite unstoppable pawns evaluation
Instead of current code, give a bonus according to the frontmost
square among candidate + passed pawns.

This is a big simplification that removes a lot of accurate code
substituting it with a statistically based one using the common
'bonus' scheme, leaving to the search to sort out the details.

Results are equivalent but code is much less and, as an added bonus,
we now store candidates bitboard in pawns hash and allow this
info to be used in evaluation. This paves the way to possible
candidate pawns evaluations together with all the other pieces,
as we do for passed.

Patch passed short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16927 W: 3462 L: 3308 D: 10157

Then failed (quite quickly) at long TC
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 8451 W: 1386 L: 1448 D: 5617

But when ran with a conclusive 40K fixed games at 60 secs it proved
almost equivalent to original one.

ELO: 1.08 +-2.0 (95%) LOS: 85.8%
Total: 40000 W: 6739 L: 6615 D: 26646

bench: 3884003
2013-09-16 08:57:37 +02:00
Reuven Peleg 21cbfafc03 Code style at passed pawn eval
No functional change.
2013-09-15 21:49:06 +02:00
Reuven Peleg d3947b2f3e Nicer operator declerations
No functional change.
2013-09-15 21:45:18 +02:00
Kojirion a71209868b Use pre-increment also for native types
Now that we use pre-increment on enums, it
make sense, for code style uniformity, to
swith to pre-increment also for native types,
although there is no speed difference.

No functional change.
2013-09-15 09:17:21 +02:00
Marco Costalba 7a1ff6d8ff Fix operator++ definition
ENABLE_OPERATORS_ON has incorrect definitions of
post-increment and post-decrement operators.

In particularly the returned value is the variable
already incremented/decremented, while instead they
should return the variable _before_ inc/dec.

This has no real effect because are only used in loops
and where the returned value is never used, neverthless
it is wrong. The fix would be to copy the variable to a
dummy, then inc/dec the variable, then return the dummy.

So instead, rename to pre-increment that can be implemented
without the dummy, actually the current implementation
it is already the correct pre-increment, with the only change
to return a reference (an l-value) and not a copy, so
to properly mimic the pre-increment on native integers.

Spotted by Kojirion.

No functional change.
2013-09-15 09:09:06 +02:00
Marco Costalba 82f6779c2e Don't blunder under extreme time pressure
We always attempt to keep at least this emergencyBaseTime
at clock. But if available time is very low it means that
we will force ourself to play immediately to satisfy the
emergencyBaseTime constrain and so leading to blunders.

Patch is good at short and very short TC (15secs and 5secs respectively)
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 26590 W: 5426 L: 5245 D: 15919

LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 5767 W: 1397 L: 1268 D: 3102

Instead seems has no influence at longer TC (60 secs)
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 79862 W: 13623 L: 13339 D: 52900

So it is committed to have a broader testing but is
to be consider still EXPERIMENTAL and can be reverted
easily.

No functional change.
2013-09-15 07:59:09 +02:00
Marco Costalba 3abccdc82d Move classify_leaf() to c'tor in bitbases
No functional change.
2013-09-14 13:08:37 +02:00
Marco Costalba 8d6d0223bf Small touches to bitbase.cpp
Inspired by Lucas's code:

https://github.com/lucasart/chess/blob/master/src/kpk.cc

No functional change.
2013-09-14 11:19:12 +02:00
Reuven Peleg fc17d0de77 Increase passed bonus for having more pieces
Passed both short TC:
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 36463 W: 7575 L: 7365 D: 21523

And long TC:
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 2953 W: 564 L: 446 D: 1943

bench: 3846852
2013-09-13 18:08:22 +02:00
Marco Costalba 27f2ce8f6e Revert "Move draw by material check"
Possible regression

bench: 4554579
2013-09-12 08:44:11 +02:00
Marco Costalba 45b0aea875 Revert "Fix random moves when time < 10ms"
Possible regression.

No functional change.
2013-09-12 08:38:19 +02:00
Marco Costalba 4803d5772c Extend checks more when below alpha
Passed both short TC:
LLR: 2.97 (-2.94,2.94) [-1.50,4.50]
Total: 8739 W: 1830 L: 1698 D: 5211

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 6716 W: 1238 L: 1101 D: 4377

bench: 4554576
2013-09-11 19:15:28 +02:00
Uri Blass 738c5595ad Extend checks more in losing positions
Passed both short TC:
LLR: 2.98 (-2.94,2.94) [-1.50,4.50]
Total: 3974 W: 860 L: 741 D: 2373

And long TC:
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 16807 W: 2917 L: 2733 D: 11157

bench: 3767999
2013-09-11 09:15:47 +02:00
Reuven Peleg bebd6e16f6 Simplify unstoppable pawns evaluation
No functional change
2013-09-10 23:02:05 +02:00
Marco Costalba 49e110c52b Fix random moves when time < 10ms
In case we have less then 10ms to think as soon as
we wake up the timer, it immediately fires and calls
check_time() where due to condition:

elapsed > TimeMgr.maximum_time() - 2 * TimerResolution

the stop flag is set and search returns immediately, without
actually search anything.

Here the somewhat hacky fix is to start the timer after
at least one iteration as been completed.

No functional change.
2013-09-10 21:23:20 +02:00
Reuven Peleg 4d90aeb0ab More readable space mask
No functional change.
2013-09-10 19:18:10 +02:00
homoSapiensSapiens 03cd049c68 Change condition to use relative rank
No functional change
2013-09-10 00:47:31 +02:00
Marco Costalba 6ab8b9b6c6 Fix some comments in position.cpp
No functional change.
2013-09-08 06:28:53 -07:00
Marco Costalba 490f67a3f8 Move draw by material check
It is more natural to test this case among
others material distributions.

No functional change.
2013-09-08 06:11:35 -07:00
Marco Costalba 0515ad0fb0 Remove unreachable values in mobility table
The possible maximum mobility cardinality (plus one in case of
zero squares available) is:

- Knights: max. 8  squares -> max. 9  entries
- Bishops: max. 13 squares -> max. 14 entries
- Rooks:   max. 14 squares -> max. 15 entries
- Queen:   max. 27 squares -> max. 28 entries

So remove the extra entries in the table.

Spotted by Dariusz Orzechowski.

No functional change.
2013-09-07 18:25:24 +02:00
Marco Costalba bf51db2526 Fix warning: double to float truncation
MSVC 2013 says:
warning C4305: '*=' : truncation from 'double' to 'float'

No functional change.
2013-09-07 12:30:44 +02:00
Lucas Braesch 59702aca0d Singular extension at 8 plies also for PV nodes
Passed both short TC:
LLR: 2.96 (-2.94,2.94)
Total: 11451 W: 2455 L: 2282 D: 6714

And long TC
LLR: 2.96 (-2.94,2.94)
Total: 15813 W: 2907 L: 2723 D: 10183

bench: 3864419
2013-09-07 09:34:22 +02:00
Lucas Braesch c86eee3918 Union of 2 changes
Union of

- LMR >= 3 plies from Gary tests.stockfishchess.org/tests/view/522522960ebc595d328fcafd

- allows() tweak from Reuven tests.stockfishchess.org/tests/view/5225fa1c0ebc595d328fcb53

Both passed Step I and failed Step II.

Instead this union passed both short TC:
LLR: 2.95 (-2.94,2.94)
Total: 14525 W: 3063 L: 2874 D: 8588

And long TC
LLR: 2.94 (-2.94,2.94)
Total: 31075 W: 5566 L: 5308 D: 20201

bench: 4238160
2013-09-07 09:25:45 +02:00
Lucas Braesch 10b53e1c5e Do not prune useless checks in QS
Passed both SPRT tests in "simplification mode", so with
elo0: -3.00 alpha: 0.05 elo1: 3.00 beta: 0.05

Short TC:
LLR: 2.96 (-2.94,2.94)
Total: 6243 W: 1302 L: 1195 D: 3746

Long TC
LLR: 2.96 (-2.94,2.94)
Total: 22972 W: 4124 L: 4020 D: 14828

bench: 4633330
2013-09-05 18:50:16 +02:00
Marco Costalba a30d3571ca Revert "Fix check for bishop pair in material imbalance"
Idea is sound but implementation is partial. Ryan and Joona noticed that
    we leave an hole in material table. Also we got another report by an user
    of an odd behaviour. Namely, if you start stockfish and from the prompt
    give 'bench' you get 3453941, then if you run again bench you get 3453940.

    The reason is that two different positions with the same number of pieces,
    but one with a bishop pair and another without have the same material key.
    But after Eelco patch also different material imbalance and this yields
    to this issue.

    Restesting at long TC shows the patch does not really contribute at
    ELO improvement. Actually patch failed at long TC.

    LLR: -2.97 (-2.94,2.94)
    Total: 23109 W: 4104 L: 4092 D: 14913

    So revert.

    bench: 3453945
2013-09-05 06:34:48 +02:00
Reuven Peleg 457ac26de5 Rewrite backward pawn detection
Use the new backmost_sq() instead of a loop.

No functional change.
2013-09-03 20:11:00 +02:00
Eelco de Groot 679c2ea227 Fix check for bishop pair in material imbalance
Prefer pos.bishop_pair() to pos.count<BISHOP>(WHITE) > 1
because the first checks that the two bishops are on
different color squares.

Although the change seems to kick in only in very rare cases,
quite surprisingly it was able to pass SPRT test at short TC.

LLR: 2.95 (-2.94,2.94)
Total: 39818 W: 8174 L: 7956 D: 23688

bench: 3453941
2013-09-03 19:40:34 +02:00
Marco Costalba 9ff594c3a9 Rewrite KBBKN endgame
This was thought to be a draw but the bishops generally win. However,
it takes up to 66 moves. The position in the diagram was thought to be
a draw for over one hundred years, but tablebases show that White wins
in 45 moves. All of the long wins go through this type of semi-fortress
position. It takes several moves to force Black out of the temporary
fortress in the corner; then precise play with the bishops prevents Black
from forming the temporary fortress in another corner (Nunn 1995:265ff).

Before computer analysis, Speelman listed this position as unresolved,
but "probably a draw" (Speelman 1981:109).

bench: 3453945
2013-09-02 11:03:01 +02:00
Marco Costalba 849b089a63 Don't use lpthread for Android
Thanks to Peter Osterlund for the feedback.

No functional change.
2013-09-01 13:48:09 -07:00
Marco Costalba aee404f532 Improve ARM compatibility
STANDALONE-TOOLCHAIN.html in Android NDK says:

It is recommended to use the -mthumb compiler flag to force the generation
of 16-bit Thumb-1 instructions (the default being 32-bit ARM ones).

If you want to target the 'armeabi-v7a' ABI, you will need ensure that the
following two flags are being used:

  CFLAGS='-march=armv7-a -mfloat-abi=softfp'

Note: The first flag enables Thumb-2 instructions, and the second one
      enables H/W FPU instructions while ensuring that floating-point
      parameters are passed in core registers, which is critical for
      ABI compatibility. Do *not* use these flags separately!

Thanks to Peter Osterlund for pointout this doc and for showing me
an example Makefile to follow.

No functional change.
2013-09-01 09:18:37 -07:00
Uri Blass 0915f85895 Union of 2 changes that failed with good score
This is a union of 2 changes:

A tweak of recaptures limit from Joona Kiiski
http://tests.stockfishchess.org/tests/view/52166d7c0ebc59319a242400

and a tweak of move count pruning from Leonid Pechenik
http://tests.stockfishchess.org/tests/view/5217c7e60ebc59319a242456

The set passed both short TC at 30+0.05
LLR: 2.96 (-2.94,2.94)
Total: 18936 W: 3723 L: 3566 D: 11647

And the usual long TC at 60+0.05
LLR: 2.95 (-2.94,2.94)
Total: 48962 W: 8837 L: 8487 D: 31638

bench: 3453945
2013-09-01 08:07:21 -07:00
Marco Costalba 5e8bc6ac2a Assorted clean up in endgames
No functional change.
2013-09-01 07:39:04 -07:00
Marco Costalba 3e4dcaa06e Fix a bogus assert in allows()
Becuase castle is coded as "king captures the rook"
the to_sq(move), A1/8 or H1/8 is empty after the move,
leading to assert assert(p != NO_PIECE) in color_of().

Teach allows() asserts about castle and fix the crash.

Bug reported by Ryan Takker and tracked down by Tom Vijlbrief.

No functional change.
2013-08-30 16:42:18 +02:00
Marco Costalba 14f47c8ac6 Use frontmost_sq() and backmost_sq helpers
Should easier to read than the lsb() / msb() low
level functions.

No functional change.
2013-08-30 16:22:22 +02:00
Gary Linscott 5d90c149b5 Enable LMR for dangerous moves
Passed both short TC
LLR: 2.96 (-2.94,2.94)
Total: 5598 W: 1250 L: 1125 D: 3223

And long TC
LLR: 2.97 (-2.94,2.94)
Total: 16441 W: 3102 L: 2912 D: 10427

bench: 4620975
2013-08-29 23:02:18 +02:00
homoSapiensSapiens 4b9e338541 Bonus for rook behind a passed
If our rook is behind a passed pawn, all
squares are defended.

One of the longest tests to pass !

Passed both short TC
LLR: 2.97 (-2.94,2.94)
Total: 44560 W: 9518 L: 9281 D: 25761

And long TC
LLR: 2.96 (-2.94,2.94)
Total: 61348 W: 11618 L: 11192 D: 38538

bench: 3787694
2013-08-29 22:53:21 +02:00
homoSapiensSapiens a0cf424cfc Replace hardcoded 128 by constant
No functional change.
2013-08-29 13:59:49 +02:00
Gary Linscott aecdbfc4a0 Add lsb() overload
Helper to find least significant bit relative to
the given color.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-08-28 19:50:11 +02:00
Uri Blass e6482b7d97 Time management: move faster if PV is stable
Move faster but compensate by allocating more
time when the best move changes.

Passed short TC at 15+0.05
LLR: 2.93 (-2.94,2.94)
Total: 13895 W: 3030 L: 2882 D: 798

Long TC at 60+0.05
LLR: 2.96 (-2.94,2.94)
Total: 9266 W: 1777 L: 1624 D: 5865

At time increment 30+0.5
LLR: 2.96 (-2.94,2.94)
Total: 6703 W: 1238 L: 1134 D: 4331

And at fixed game number, longer TC 120+0.05
ELO: 5.17 +-2.8 (95%) LOS: 100.0%
Total: 19306 W: 3378 L: 3091 D: 12837

bench: 4728533
2013-08-26 10:29:58 -07:00
homoSapiensSapiens 6e6c5b6103 Simplify chain detection
No functional change.
2013-08-25 18:36:49 +02:00
homoSapiensSapiens 04fd7efdfa Simplify kf definition in shelter_storm()
No functional change.
2013-08-25 18:31:47 +02:00
Chris Caino 5e331f9618 Fix KBPK bug
With

position fen 7k/8/8/8/8/7P/6K1/7B w - - 0 1
go depth 25

The evaluation at depth 22 is not draw as it should be. The reason is that
when search reaches the position 8/6kP/8/8/8/3B4/6K1/8 w - - 0 1 if white plays
h8R or h8N then we get a position that is a "KNOWN_WIN" and is _not_ a check, so
futility pruning in qsearch kicks in and black may think that it is "futile"
to reply Kxh8 since, according to the logic of the code, it cannot raise the score
back towards a draw.

bench: 4728533
2013-08-24 10:08:03 +02:00
homoSapiensSapiens b9f5d1c6ff Simplify condition for backward pawn
No functional changes.
2013-08-22 14:39:08 +02:00
Marco Costalba c4533e0d94 Retire redundant endgames
The case of two lone kings on the board is already considered
by the "No pawns" scaling factor rules in material.cpp as is
KBK and KNK.

Moreover we had a small leak in endgames map because for
KK endgame it happens white and black material keys are the
same (both equal to zero), so when adding the black endgame in
Endgames::add() we were overwriting the already exsisting
white one, leading to a memory leak found by Valgrind.

So remove the endgames althogheter and rely on scaling
to correctly set the endgames value to a draw.

No functional change.
2013-08-22 13:13:06 +02:00
Joona Kiiski f39cf1b008 Use null move when depth >= 2 plies
Passed both short TC:
LLR: 2.96 (-2.94,2.94)
Total: 23725 W: 5031 L: 4855 D: 13839

And long TC:
LLR: 2.96 (-2.94,2.94)
Total: 15730 W: 2939 L: 2754 D: 10037

bench: 4729333
2013-08-22 09:06:48 +02:00
Marco Costalba c6baefb79d Restore development version
No functional change.
2013-08-21 08:41:47 +02:00
Marco Costalba 4d120ee02e Stockfish 4
Stockfish bench signature is: 4132374
2013-08-20 09:01:25 +02:00
Tom Vijlbrief f45eee318b Fix crash when reaching max ply
Bug introduced in 1b7223a53c that
updated the ss base stack without increasing
the dimension.

No functional change.
2013-08-19 16:53:46 +02:00
Tom Vijlbrief 8c2fd2170a Remove useless condition in KXK endgame
Because eval is never called when in check.

No functional change.
2013-08-19 08:55:17 +02:00
Leonid Pechenik 91c2c44fb1 Further tweak movecount pruning
Passed both short TC
LLR: 2.95 (-2.94,2.94)
Total: 15140 W: 3125 L: 2976 D: 9039

And long TC
LLR: 2.95 (-2.94,2.94)
Total: 17118 W: 3165 L: 2974 D: 10979

bench: 4132374
2013-08-18 09:13:57 +02:00
Marco Costalba 27e9fc1067 Normalize "pawn in front of minor" patch
No functional change.
2013-08-17 11:05:55 +02:00
homoSapiensSapiens e005270fb6 Use constants arguments where possible
No functional changes.
2013-08-16 09:57:21 +02:00
Marco Costalba 4f55ed14d3 Revert using exceptions
Due to crashes. It will be reapplied once
we understand what's happening.

No functional change.
2013-08-15 09:36:26 +02:00
homoSapiensSapiens a6c0ba2100 Simplify DistanceRingsBB init
Verified by same benchmark and picking some random values.

No functional change.
2013-08-14 10:53:43 +02:00
homoSapiensSapiens fc316cbca9 Some renaming in TT store()
No functional change.
2013-08-14 09:38:35 +02:00
Marco Costalba 11b1a76f35 Use exceptions to stop the search
Instead of classical flags, throw an
exception when we want to immediately halt
the search. Currently only one type
is used for both UCI stop and threads
cut off.

No functional change.
2013-08-14 08:29:57 +02:00
Tom Vijlbrief bd8f463b7e Bonus for a pawn in front of knight/bishop
Idea originated from a post of Don Dailey
on talkchess and reported by Eelco.

This is the last succesful attempt of a long
series of trials (as usually happens, the
'idea' alone is not enough).

Passed both short 15secs TC
LLR: 2.97 (-2.94,2.94)
Total: 7629 W: 1645 L: 1515 D: 4469

And long 60secs TC
LLR: 2.96 (-2.94,2.94)
Total: 10218 W: 1932 L: 1775 D: 6511

bench: 4944581
2013-08-13 14:20:02 +02:00
Ryan Takker 4d14f97482 Remove Now Unneeded Help Text
With the new automatic setting of split depth
instead of a default, the user no longer needs
guidance on setting the split point.

Also threads now defaults to one.

No functional change.
2013-08-13 07:36:33 +02:00
Marco Costalba 15616ad199 Don't set Search::RootColor in Eval::trace
Search::RootColor is a global parameter set
before to start a search, it is not something
trace() should change.

This patch allows to add trace() calls, for
debugging, inside search itself without altering
the bench, and also ensures that the values
returned by trace() and evaluate() are fully
equivalent.

No functional change.
2013-08-11 07:02:50 +02:00
Marco Costalba 94a3608ab9 Fix GrainSize rounding error
The rounding formula is different between
positive and negative scores due to the
GrainSize/2 term that is asymmetric.

So use truncation instead of rounding. This
guarantees that evaluation is rounded to zero
in the same way for both positive and negative
scores.

Found with position's flip

bench: 4634244
2013-08-10 17:11:13 +02:00
Marco Costalba 5769509d72 Fix 'improving' condition
Because VALUE_NONE is 30002, it happens that
after a check the next move is never an improving
one.

After this patch bench signature is independent from
VALUE_NONE actual value.

bench: 4303194
2013-08-09 08:21:55 +02:00
Marco Costalba fff6b9f061 Increase LMR when not improving
Apply to LMR the same Eelco's idea
applied to move count pruning.

This is the result of a series of
attempts started by Thomas Kolarik.

Passed both short TC
LLR: 2.95 (-2.94, 2.94)
Total: 5675 W: 1241 L: 1117 D: 3317


And long TC:
LLR: 2.95 (-2.94, 2.94)
Total: 8748 W: 1689 L: 1539 D: 5520

bench: 4356801
2013-08-08 10:28:48 +02:00
Marco Costalba 56d2c3844a Further tweak Position::flip
No functional change.
2013-08-05 14:44:06 +02:00
Marco Costalba 23b6809f3d Rewrite flip() to use FEN string manipulation
Instead of dealing directly with internal parameters
just "flip" the FEN string and set the position from
that.

No functional change.
2013-08-05 12:58:14 +02:00
Marco Costalba f31847302d Streamline time computation
No functional change.
2013-08-03 18:30:43 +02:00
Marco Costalba b1a4a18d63 Update polyglot.ini with new "Min Split Depth" default
No functional change.
2013-08-03 16:41:18 +02:00
Dan Schmidt f7096ea7ce Refactor do_castle()
Not a real functional change, but bench changed due to different piecelist
reordering. To verify it a temporary my canonicalize_rooks function was
written as follows. It just ensures that the rook on the "smaller" square
is listed first.

void Position::canonicalize_rooks(Color c)
{
   if (pieceCount[c][ROOK] == 2)
   {
      Square s0 = pieceList[c][ROOK][0];
      Square s1 = pieceList[c][ROOK][1];
      if (s0 > s1)
      {
         pieceList[c][ROOK][0] = s1;
         pieceList[c][ROOK][1] = s0;
         index[s0] = 1;
         index[s1] = 0;
      }
   }
}

With this both bench and the test on Chess960 positions

./stockfish bench 128 1 8 Chess960.epd file > /dev/null

Gives same result.

bench: 4424151
2013-08-03 16:18:28 +02:00
Joona Kiiski a16ba5bbd1 Retire cpu_count()
Set threads number always to 1 at startup and let the
user explicitly to chose the number of threads.

Also preserve the useful behavior of automatically set
"Min Split Depth" according to the requested threads,
indeed this parameter is too technical for a casual user,
so, when left to zero, we set it on a sensible value.

No functional change
2013-08-02 16:48:25 +02:00
Marco Costalba 408e6ee9b6 Further factor out position update code
Along the lines of previous patch.

No functional change
2013-08-01 16:32:46 +02:00
Dan Schmidt 7b4f5c8f72 Factor out pieceList updating code
The new Position methods add_piece, move_piece, and remove_piece
now manage the member variables pieceList, pieceCount, and index,
and 9 blocks of code in Position that used to manipulate those
data structures by hand now call the new methods.

There is a slightly slowdown (< 1%) on Clang and on perft,
but the cleanup compensates the little speed loss.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-08-01 15:50:19 +02:00
Marco Costalba 55948623e7 Rework Thread hierarchy
Introduce ThreadBase struct that is search
agnostic and just handles low level stuff,
and derive all the other specialized classes
form here.

In particular TimerThread does not hinerits
anymore all the search related stuff from Thread.

Also some renaming while there.

Suggested by Steven Edwards

No functional change.
2013-07-31 18:35:52 +02:00
Marco Costalba 4d46d29efe Fix a race at thread creation
At thread creation start_routine() is called
and from there the virtual function idle_loop()
because we do this inside Thread c'tor, where the
virtual mechanism is disabled, it could happen that
the base class idle_loop() is called instead.

The issue happens with TimerThread and MainThread
where, at launch, start_routine calls
Thread::idle_loop instead of the derived ones.

Normally this bug is hidden because c'tor finishes
before start_routine() is actually called in the
just created execution thread, but on some platforms
and in some cases this is not guaranteed and the
engine hangs.

Reported by Ted Wong on talkchess

No functional change.
2013-07-31 18:35:32 +02:00
Marco Costalba cc608a7aba Tidy up Position::pretty
No functional change.
2013-07-29 19:33:30 +02:00
Marco Costalba 1f40cd6d02 Small renaming
No functional change.
2013-07-29 19:32:59 +02:00
Marco Costalba 28bc8ed462 Speed up move generation
Pass the color as template parameter
to generate_all()

Speedup of 1,3% in perft and 2,5% in bench !

No functional change.
2013-07-29 19:01:50 +02:00
Eelco de Groot 5ee16a180a Increase pruning if evaluation is not improving
Add an additional set of margins to movecount pruning
to be used when static evaluation is getting worse
than previous move.

Here are the margins table with changing
depth (fm0 not improving, fm1 improving):

    d: 0, fm0: 3, fm1: 3
    d: 1, fm0: 4, fm1: 4
    d: 2, fm0: 6, fm1: 6
    d: 3, fm0: 7, fm1: 10
    d: 4, fm0: 11, fm1: 15
    d: 5, fm0: 15, fm1: 21
    d: 6, fm0: 21, fm1: 29
    d: 7, fm0: 27, fm1: 37
    d: 8, fm0: 35, fm1: 47
    d: 9, fm0: 42, fm1: 57
    d: 10, fm0: 51, fm1: 68
    d: 11, fm0: 60, fm1: 81
    d: 12, fm0: 70, fm1: 94
    d: 13, fm0: 81, fm1: 108
    d: 14, fm0: 92, fm1: 123
    d: 15, fm0: 104, fm1: 139

Good at both short TC

LLR: 2.97 (-2.94,2.94)
Total: 11502 W: 2503 L: 2361 D: 6638

And long TC

LLR: 2.98 (-2.94,2.94)
Total: 7189 W: 1421 L: 1277 D: 4491

bench: 4364793
2013-07-29 01:21:21 +02:00
Marco Costalba d30dfc084c Annotate an unlikely condition
No functional change.
2013-07-27 11:34:15 +02:00
Marco Costalba 6373e88b5b Fix an assert in KBK endgame
The endgame king + minor vs king is erroneusly
detected as king + minor vs king + minor

Here the fix is to detect king + minor earlier,
in particular to add these trivial cases to
endgame evaluation functions.

Spotted by Reuven Peleg

bench: 4727133
2013-07-27 08:25:45 +02:00
Tom Vijlbrief 7487eb0dca Rewrite pawn shield and storm code
Passes quickly both short TC:
LLR: 2.95 (-2.94,2.94)
Total: 5755 W: 1349 L: 1222 D: 3184

And long TC:
LLR: 2.95 (-2.94,2.94)
Total: 2744 W: 628 L: 505 D: 1611

bench: 4727133
2013-07-26 00:12:46 +02:00
Marco Costalba 2067a99c07 Fix a typo in bitboard.h
Introduced by previous patch.

Spotted by Joerg Oster

No functional change.
2013-07-25 07:44:27 +02:00
homoSapiensSapiens 002062ae93 Use #ifndef instead of #if !defined
And #ifdef instead of #if defined

This is more standard form (see for example iostream file).

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-24 19:49:17 +02:00
Marco Costalba 4064ee5406 Simplify captures ordering
A big simplification and removing of useless code.

Finished at 50% both at short TC (with SPRT) than
at long TC at fixed number of games:
ELO: -0.14 +-3.4 (95%) LOS: 46.8%
Total: 15206 W: 2836 L: 2842 D: 9528

bench: 5059948
2013-07-24 07:53:32 +02:00
Marco Costalba b0fd2b6b98 Revert "Halve king eval margin"
This reverts commit 4b3a0fdab0.

As Gary says: " It failed when I tried it at long TC previously, and only
barely passed this time.  Some anecdotal evidence is that it hurts vs other
engines as well (the Lightspeed rating list showed a 16 elo drop from previous
best version - still +- 5 error bars on both, but that's still significant)"

I also agree that if we have some doubts (like in this case) it is better to
be safe than sorry.

bench: 4615572
2013-07-24 07:46:25 +02:00
Ryan Schmitt d0bc951835 Tune pawn PSQT values
Reduces the influence of PSQT for entries such as
the extended center and the h-file.

Passed both short TC test:
LLR: 2.95 (-2.94,2.94)
Total: 23919 W: 5207 L: 5029 D: 13683

And long TC one:
LLR: 2.96 (-2.94,2.94)
Total: 5762 W: 1108 L: 974 D: 3680

Bench: 4617880

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-24 07:41:51 +02:00
Reuven Peleg 73131e7c78 Use arrow operator instead of * and .
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-22 20:22:40 +02:00
Reuven Peleg 1cc18d8a7a Better condition in is_pseudo_legal()
Simplify occupied destination condition.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-22 20:02:31 +02:00
Marco Costalba c45c6d308f Small touches in move generation
No functional change.
2013-07-21 11:01:24 +02:00
Marco Costalba f73bb438aa Some renaming in MovePicker
No functional change.
2013-07-21 09:55:08 +02:00
Marco Costalba 71dd8a333f Rewrite and simplify SEE
This very speed critical code was full of clever (!)
tricks and subtle details.

So I have rewritten it in a more straithforward way
and, as very often happens, result is even faster
than original.

No functional change.
2013-07-21 01:04:29 +02:00
Marco Costalba 9207baed65 Revert "Fix critical SEE bug (take 2)"
This reverts commit 3e95800814

For some reason it fails the short TC test:
LLR: -2.96 (-2.94,2.94)
Total: 20033 W: 4214 L: 4265 D: 11554

bench: 4769737
2013-07-20 18:45:38 +02:00
Marco Costalba b191df5ebe Revert "Yet another attempt at signature-build"
Still broken on OS X

No functional change.
2013-07-20 15:15:31 +02:00
Marco Costalba 3e95800814 Fix critical SEE bug (take 2)
It is somewhat unbilievable but our SEE is broken !

    If the first SEE move is a king capture and square is
    defended then SEE continues instead of breaking.

    The bug shows only on normal SEE, not see_sign() so
    probing with a:

    dbg_hit_on_c(slIndex==1, captured == KING);

    reports just a tiny:

    Total 3465656 Hits 6646 hit rate (%) 0

    Bug was there since Retire seeValues[] and move PieceValue[] out of Position of 26/6/2011 (!)
    although for some reason didn't show immediately, indeed the
    bougous patch was a "No functional change" (!!)

    bench: 4699504
2013-07-20 14:24:47 +02:00
Marco Costalba 2ed56f4d5f Revert all the SEE stuff
The speed up seems to introduce some
functionality change.

Revert to original master for now.

bench: 4769737
2013-07-20 14:20:33 +02:00
Marco Costalba 7b0b463720 Yet another attempt at signature-build
This one should work on all flavours of sed

Suggested by by Louis Zulli

No functional change.
2013-07-20 14:05:04 +02:00
Marco Costalba a6c5b4c6fb Fix critical SEE bug
It is somewhat unbilievable but our SEE is broken !

If the first SEE move is a king capture and square is
defended then SEE continues instead of breaking.

The bug shows only on normal SEE, not see_sign() so
probing with a:

dbg_hit_on_c(slIndex==1, captured == KING);

reports just a tiny:

Total 3465656 Hits 6646 hit rate (%) 0

Bug was there since 351ef5c85b of 26/6/2011 (!)
although for some reason didn't show immediately, indeed the
bougous patch was a "No functional change" (!!)

bench: 4793754
2013-07-20 13:37:12 +02:00
Marco Costalba 0504a6975d Speedup see()
And rename next_attacker() SEE helper

This very simple patch is able to speed up
bench run of almost 2% !

No functional change.
2013-07-20 12:34:35 +02:00
Marco Costalba a5b5a91512 Fix signature-build under OSX
On OS X when you use -i an extension for the in-place
substitution a backup files is required.

http://stackoverflow.com/questions/4247068/sed-command-failing-on-mac-but-works-on-linux

So rewrite to make sed flushing sign.txt in one go and avoid
using -i option.

Reported by Louis Zulli

No functional change.
2013-07-20 02:38:14 +02:00
Reuven Peleg 1a8f63a896 Microptimize gives_check() for castling case
Without patch we have 333198 nps, with patch 334249.

A very small +0.3%, not a lot manily becuase this is a
side path that is taken very few times.

Anyhow idea is correct becuase first 'quick' condition
has an hit rate of about 95%.

No functional change.
2013-07-19 17:07:54 +02:00
Marco Costalba ee5514b8fd Small simplification in space eval scoring
No functional change.
2013-07-19 11:13:18 +02:00
Marco Costalba 99e547f4cb Rename MoveStack to ExtMove
Stack has no meaning here, while ExtMove (extended move),
better clarifies that we have a move + a score.

No functional change.
2013-07-19 10:27:58 +02:00
Marco Costalba 110644d918 Better document what we skip when in check
No functional change.
2013-07-19 09:37:31 +02:00
homoSapiensSapiens 4b3a0fdab0 Halve king eval margin
But still keep the same original
margin for score.

Passed both short TC test
LR: 2.95 (-2.94,2.94)
Total: 3710 W: 845 L: 726 D: 2139

And long TC
LLR: 2.95 (-2.94,2.94)
Total: 57859 W: 10939 L: 10532 D: 36388

bench: 4769737

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-19 08:16:39 +02:00
Marco Costalba 05e31c5e5f Drop grep and tr dependency in Makefile
Use only sed to get the bench signature.

No functional change.
2013-07-15 21:39:06 +02:00
Marco Costalba 46fdb14b2f Don't use __builtin_expect
Partially revert previous patch and use
unlikey() just as code annotation.

Actually it is better to rely on a profiler for branch prediction:

http://blog.man7.org/2012/10/how-much-do-builtinexpect-likely-and.html

"In fact, even when only one in ten thousand values is nonzero,
we're still at only roughly the break-even point"

No functional change,
2013-07-15 21:09:06 +02:00
Marco Costalba cbb1a8ed31 Better annotate unlikely conditions
And in case of gcc we win also a small
speed optimization due to better branch
prediction.

No functional change.
2013-07-15 21:01:02 +02:00
Reuven Peleg a6c5f60caa Simplify a condition in refutes()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-07-15 20:40:49 +02:00
Marco Costalba 9518cc3254 Update 'make help'
No functional change.
2013-07-14 12:23:28 +02:00
Marco Costalba b0a177bc67 Add signature-profile-build make target
Extend patch 3f64a2af6a to profile builds.

here the make command is:

make signature-profile-build ARCH=xxx COMP=xxx

No functional change.
2013-07-14 11:57:06 +02:00
Marco Costalba 5b7b330616 Retire engine Tag
It is somewhat redundant and could make SF
name too long, so use just Version, in case
of a signature build Version will be set to
'sig-xxx' otherwise, if left empty, we fall
back on usual date stamp.

No functional change.
2013-07-14 11:13:13 +02:00
Marco Costalba 3f64a2af6a Add signature-build make target
When compiling with:

make signature-build ARCH=xxx COMP=xxx

After binary has been roduced, it will be run to
get the signature 'stockfish bench' and this
number will be used as Version, so that it
will be easy to track the original sources
from a binary.

No functinal change.
2013-07-14 10:59:25 +02:00
Marco Costalba 9749f1f14c Fix build on Intel compiler
Due to a strange issue (bug?) the ternary
operator does not return a BitCountType for
icc, so revert to the expression used
before bcbc9bfd1f

No functional change.
2013-07-13 23:13:30 +02:00
Marco Costalba 4ede49cd85 Fully qualify memset and memcpy
And other trivial touches.

Ispired by Lucas's DiscoCheck

No functional change.
2013-07-13 18:01:13 +02:00
Tom Vijlbrief 6960f41e03 Retire enoughMaterial + lower trapped rook threshold
Here speed up is the name of the game.

Speed up is gained:

- Removing the useless enoughMaterial code

- Limiting trapped rook evaluation to where it counts

Tested at long TC:
LLR: 2.97 (-2.94,2.94)
Total: 10061 W: 1948 L: 1790 D: 6323

bench: 4558173
2013-07-13 18:01:03 +02:00
Marco Costalba bf90499fc3 A useless assignment found by Clang’s static analyzer
Warning is: "Value stored to 'xxx' is never read" and
it is raised in SpNode case.

No functional change.
2013-07-13 16:57:03 +02:00
Marco Costalba 404c4122ce Fix build with MSVC 2013
Also add an assert hinted by MSVC code analysis tool.

No functional change.
2013-07-13 13:03:09 +02:00
Marco Costalba 2a0bbb9faa Fix printing of PV info: take 2
Now last PV line is printed twice, fix that.

No functional change.
2013-07-13 07:43:50 +02:00
Marco Costalba 128e097d03 Fix printing of PV info
It was erroneusly skipped after the
aspiration window rework.

Reported by Eelco.

No functional change.
2013-07-12 23:42:42 +02:00
Marco Costalba c1264e46d0 Rename some UCI options
Thanks to Don, Miguel, Louis and the other people
of talkchess forum for the suggestion:

http://www.talkchess.com/forum/viewtopic.php?t=48612

Also sync polyglot.ini with current UCI options

No functional change.
2013-07-11 16:07:10 +02:00
Marco Costalba 366f6b0dab Fix a crash with depth 1 perft
Bug recently introduced in e215a88cdd

No functional change.
2013-07-11 07:23:48 +02:00
Marco Costalba 58aee9a9ea Don't IID when in check also in PvNodes
This tiny functional change allows to
nicely simplify things.

Performed at 50% in short TC:
LLR: -0.43 (-2.94,2.94)
Total: 46406 W: 9681 L: 9565 D: 27160

And succesfully passed long TC reverse test:
LLR: -2.95 (-2.94,2.94)
Total: 4945 W: 858 L: 937 D: 3150

bench: 4507230
2013-07-09 19:03:11 +02:00
Marco Costalba 4b703b1429 Revert previous patch
Unfortunatly a reverse test at long TC failed:

master^ vs master
LLR: 1.37 (-2.94,2.94)
Total: 33682 W: 6294 L: 6071 D: 21317

So becuase short TC score is 50% there is a good
possibility patch is not scalable.

So revert it.

bench: 4507288
2013-07-09 08:04:12 +02:00
Marco Costalba 62d38f0196 Simplify "fail high upon reduction" in null search
Do not use threat move to detect the condition. This
let us to retire the big allows() function.

Test at short TC was within 50% score:
LLR: -2.95 (-2.94,2.94)
Total: 38272 W: 7941 L: 7940 D: 22391

To be verified with reverse long TC

bench: 4191565
2013-07-08 07:23:30 +02:00
Marco Costalba 7e575512ae Skip node-level cut-off tests when in check
No functional change.
2013-07-07 13:39:58 +02:00
Marco Costalba a55fb76dcc Simplify aspiration window code
Here the main difference is that now we center
aspiration window on last returned score. This allows
to simplify handling of mate scores.

We have done a reversed SPRT tests, where we wanted to
verify if master is stronger than this patch.

Long TC: master vs this patch (reverse test)
LLR: -2.95 (-2.94,2.94)
Total: 37992 W: 7012 L: 6920 D: 24060

bench: 4507288
2013-07-03 18:59:23 +02:00
Marco Costalba 838255ef91 Workaround github issue
Temporary revert aspiration window patch
so to be visible to everybody: it will be
re-applied with next patch

No functional change (together with next one)
2013-07-03 18:58:23 +02:00
Marco Costalba 6fbe027da0 Simplify aspiration window code
Here the main difference is that now we center
aspiration window on last returned score. This allows
to simplify handling of mate scores.

We have done a reversed SPRT tests, where we wanted to
verify if master is stronger than this patch.

Long TC: master vs this patch (reverse test)
LLR: -2.95 (-2.94,2.94)
Total: 37992 W: 7012 L: 6920 D: 24060

bench: 4507288
2013-07-03 09:06:03 +02:00
Marco Costalba ddcb572c41 Disable flto when debugging
Link-time optimization does not work well with
generation of debugging information:

http://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html

Reported by Louis Zulli

No functional change.
2013-07-03 08:21:21 +02:00
Marco Costalba 12c0dfc113 Revert "Remove confusing optimization"
This reverts commit e05c80a088.

we gain a speed up of 1.5% under gcc !

No functional change.
2013-07-02 20:12:57 +02:00
Marco Costalba db53883f06 Merge branch 'master' into aspiration
bench: 4507288
2013-07-02 07:27:08 +02:00
Marco Costalba 13ffb08136 Revert "Increase earlier aspiration window size"
This reverts commit b88bc7b766.
2013-07-02 07:25:39 +02:00
Marco Costalba 2d82db1d14 Entering a pawn endgame is no more dangerous
A simplification of the 'dangerous' definition.

Seems neutral at reverse test at long TC

master vs patch
LLR: -2.96 (-2.94,2.94)
Total: 16974 W: 3122 L: 3139 D: 10713

bench: 4689029
2013-07-02 07:24:17 +02:00
Marco Costalba b88bc7b766 Increase earlier aspiration window size
bench: 4377851
2013-07-01 19:29:38 +02:00
Marco Costalba e074a19f5c Merge branch 'master' into aspiration 2013-07-01 19:25:23 +02:00
Marco Costalba b8930d0c26 Fix a stale comment
No functional change.
2013-06-30 13:12:04 +02:00
Marco Costalba 4fc7734547 Simplify search results update
Also some rename while there.

No functional change.
2013-06-30 12:49:39 +02:00
Marco Costalba 92dcbfa658 Reorder conditions according to their frequency
This should minimize useless tests.

No functional change.
2013-06-30 11:35:53 +02:00
Marco Costalba b50eb6bea8 Center aspiration window on last returned score
bench: 4428212
2013-06-30 11:14:02 +02:00
Marco Costalba 6f079ae720 Simplify aspiration window loop
Don't open the window in case we find a mate score: this
will be takes care with next patch.

No functional change.
2013-06-30 10:54:09 +02:00
Marco Costalba 203fdc9ac1 Use calloc() in TranspositionTable::set_size()
Function calloc() already initializes memory to
zero, so avoid calling clear() afterwards.

Also some renaming while there (inspired by DiscoCheck).

No functional change.
2013-06-29 11:23:07 +02:00
Marco Costalba 17d41b3861 Fix some stale comments
No functional change.
2013-06-23 13:19:03 +02:00
Marco Costalba 8cff4862a6 Move SquareDistance[] to bitboard.cpp
No functional change.
2013-06-23 13:13:13 +02:00
Marco Costalba 908d98820b Don't explicitize enum values when not needed
Compiler will chose the correct values in sequential
order for you.

Also move file and rank bitboards definitions to
bitboard.h

No functional change.
2013-06-23 11:30:40 +02:00
Marco Costalba a4c11b71ac Retire in_front_bb(Color c, Square s) overload
Explciitly call rank_of() in the few places where
it is used.

No functional change.
2013-06-23 10:16:43 +02:00
Marco Costalba b2fadf32aa Retire ThisAndAdjacentFilesBB[]
It is unused. Also renamed attack_span_mask to
pawn_attack_span

No functional change.
2013-06-23 10:09:24 +02:00
Marco Costalba 378bcfe760 Simplify hidden_checkers()
De-templetize and pass color as function argument.
No speed change.

No functional change.
2013-06-23 09:03:40 +02:00
Marco Costalba fe2ed42661 Name functions along corresponding UCI commands
No functional change.
2013-06-22 12:46:03 +02:00
Marco Costalba e215a88cdd Micro-optimize perft
Avoid to call perft function when we just need to count
moves, at leaf nodes.

Speed up of almost 2%

No functional change.
2013-06-21 09:10:03 +02:00
Ryan Schmitt e95e69515a Include file attacks in 'major on pawn'
Passed both short TC:
LLR: 2.97 (-2.94,2.94)
Total: 57846 W: 12248 L: 11974 D: 33624

And long one:
LLR: 2.95 (-2.94,2.94)
Total: 9181 W: 1732 L: 1581 D: 5868

bench: 4609948
2013-06-19 07:22:10 +02:00
Reuven Peleg e05c80a088 Remove confusing optimization
Here we skip the call to pos.attacks_from<ROOK>(s) in the 98%
of cases, testing the first 2 members first. Unfortunatly
code is a bit triky and not clear. So we give up to the
speed optimization in exchange of more code clarity.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-06-17 18:12:15 +02:00
Reuven Peleg 7b31e81d77 Merge some if statements in pos_is_ok()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-06-17 18:02:59 +02:00
Marco Costalba c0964fc70f Remove redundant condition in probcut
When !ss->skipNullMove it is assured that excludedMove == MOVE_NONE

No functional change.
2013-06-17 09:30:59 +02:00
Marco Costalba cd782c11ec Rename piece_count and piece_list
No functional change.
2013-06-16 13:21:10 +02:00
Marco Costalba 5ea984ac35 Don't calculate pawnsOnSquares twice
And reformat some code while there.

No functional change.
2013-06-16 11:32:10 +02:00
Marco Costalba 1fd020a8ba Use move_pawns() in Pawns::probe
And rename some stuff.

No functional change.
2013-06-16 10:40:36 +02:00
Marco Costalba 02420d4670 Revert "Reduce more CUT nodes only if parent node is reduced"
This reverts commit d54e8a5955.

It was not proved with SPRT this tweak is stronger. So revert it
for now to follow fishtest guidelines.

bench: 5108393
2013-06-14 08:27:06 +02:00
Marco Costalba 2c7ab488a8 Fix description of TT entry
It was way outdated and wrong !

No functional change.
2013-06-14 08:21:02 +02:00
Marco Costalba d54e8a5955 Reduce more CUT nodes only if parent node is reduced
So when we are doing a LMR search at the parent ALL node.

This patch didn't prove stronger at 60" TC
LLR: -2.97 (-2.94,2.94)
Total: 22398 W: 4070 L: 4060 D: 14268

But, first, it scores at 50%, second (and most important for me) the opposite,
i.e. normal reduction when parent node is not reduced, seems very bad:
LLR: -2.95 (-2.94,2.94)
Total: 7036 W: 1446 L: 1534 D: 4056

According to Don, this idea of increased reduction of CUT nodes
works because if parent node is reduced, missing a cut-off due to
reduced depth search (meaning position is somehow tricky) forces
a full depth research at parent node, giving due insight in this
set of sensible positions.

IOW if we expect a node to fail-high at depth n, then we assume it
should fail-high also at depth n-1, if this doesn't happen it means
position is tricky enough to deserve a research at depth n+1.

bench: 4687419
2013-06-13 20:05:02 +02:00
Marco Costalba 4bebb15e94 Reduce more CUT nodes
We got a good result from this tweak, in line with
what was already found by Don Dailey.

At short TC:
LLR: 2.95 (-2.94,2.94)
Total: 13097 W: 2742 L: 2598 D: 7757

At long TC:
LLR: 2.97 (-2.94,2.94)
Total: 7281 W: 1408 L: 1265 D: 4608

bench: 5108393
2013-06-13 19:50:32 +02:00
Marco Costalba 3b8f66f8ac Introduce Cut/All node definitions
Follow Don Dailey definition of cut/all node:

"If the previous node was a cut node, we consider this an ALL node.
The only exception is for PV nodes which are a special case of ALL nodes.
In the PVS framework, the first zero width window searched from a PV
node is by our definition a CUT node and if you have to do a re-search
then it is suddenly promoted to a PV nodes (as per PVS search) and only
then can the cut and all nodes swap positions. In other words, these
internal search failures can force the status of every node in the subtree
to swap if it propagates back to the last PV nodes."

http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=519741&t=47577

With this definition we have an hit rate higher than 90% on:

    if (!PvNode && depth > 4 * ONE_PLY)
        dbg_hit_on_c(cutNode, (bestValue >= beta));

And an hit rate of just 28% on:

    if (!PvNode && depth > 4 * ONE_PLY)
        dbg_hit_on_c(!cutNode, (bestValue >= beta));

No functional change.
2013-06-13 19:46:49 +02:00
Marco Costalba b6e9d901b0 Don't use std::vector::data()
It is a C++11 only function.

Reported by Eelco.

No functional change.
2013-06-13 07:42:43 +02:00
Marco Costalba 1b7223a53c Fix again early stop ss pointer
Fix was wrong becuase search starts from ss+1,
code is a bit tricky here, so rewrite in a way
to be more easy to read and understand.

Spotted by Eelco.

No functional change.
2013-06-09 23:36:46 +02:00
Marco Costalba de1dc4f2de Don't need to expose namespace Zobrist
It can be local to position.cpp

No functional change.
2013-06-09 23:27:07 +02:00
Marco Costalba a6e0f62a4f Zobrist::init() should be Position::init()
No functional change.
2013-06-09 13:54:38 +02:00
Marco Costalba 81e242a96d Convert pieceSquareTable to 3 dimensions
No functional change.
2013-06-09 13:10:21 +02:00
Marco Costalba dd1855eb2f More consistent 'piece' variable naming
No functional change.
2013-06-09 12:56:05 +02:00
Marco Costalba db4cd89cb8 Introduce operator~(Piece c)
Small syntactic sugar to reverse piece color.

No functional change.
2013-06-09 12:44:04 +02:00
Marco Costalba 7e95495b35 Retire psq_delta()
No functional change.
2013-06-09 12:32:16 +02:00
Marco Costalba 55eb7dd1e9 Use alpha instead of beta-1
It is more directly related to a fail-low.

No functional change.
2013-06-09 11:52:39 +02:00
Marco Costalba 902c0566a6 Fix incorrect 'ss' pointer in early stop check
The exclusion search used to verify one move is much
better than other shall be called with 'ss' and not
'ss+1'

No functional change.
2013-06-09 11:01:11 +02:00
Dariusz Orzechowski bc02cc0c8a Fix a typo
No functional change.
2013-06-08 11:01:28 +02:00
Marco Costalba 05c6f7a40b Fix search log when using skills
In case of we pick a sub-optimal move be
sure to print this, and not the best one
on seach log file.

Bug spotted by Guenther Demetz.

No functional change.
2013-06-08 10:56:20 +02:00
Marco Costalba 2a98042c21 Fix a crash when 'go' multiple times
Search is started after setting a position and
issuing UCI 'go' command. Then if we stop the search
and call 'go' again without setting a new position it
is assumed that the previous setup is preserved, but
this is not the case because what happens is that
SetupStates is reset to NULL, leading to a crash as
soon as RootPos.is_draw() is called because st->previous
is now stale.

UCI protocol is not very clear about requiring that a
position is setup always before launching a search,
so here we easy the life of GUI developers assuming
that the current state is preserved after returning
from a 'stop' command.

Bug reported by Gregor Cramer.

No functional change.
2013-06-01 16:19:42 +02:00
Marco Costalba 46409a7852 Assorted renaming in evaluation
And some reshuffle too.

No functional change.
2013-06-01 13:17:39 +02:00
jundery d8b266af8b Passed pawn tuning
A small number of tests with simulated
annealing at 15s indicated these values
may be better

And this is verified at long 60+0.05 TC
LLR: 2.95 (-2.94,2.94)
Total: 40658 W: 7821 L: 7501 D: 25336

bench: 4931544
2013-05-31 09:17:48 +02:00
Marco Costalba abb40777bf Shrink engine UCI name
Some GUI have problems with long names.

Reported by George Speight.

No functional change.
2013-05-27 17:43:38 +02:00
Marco Costalba 90abcac1c7 Add Pawn Structure also to polyglot.ini
No functional change.
2013-05-25 13:14:41 +02:00
Marco Costalba 7222f47350 Re-add "Pawn Structure" UCI option
And reshuffle the code to not special case
this parameter.

No functional change.
2013-05-25 12:38:14 +02:00
Marco Costalba eafb66e1aa More uniform tracing code
No functional change.
2013-05-25 12:01:50 +02:00
Uri Blass d4a02b135d Bunch of 3 small patches
This patch is the sum of:

- Grainsize of 4 instead of 8

- Removing "depth < DEPTH_ZERO"

- Change DEPTH_QS_RECAPTURES = -5 to -7

All the patches individually failed to pass SPRT but scored
around 50%.

Together they pass easily short TC:
LLR: 2.96 (-2.94,2.94)
Total: 4429 W: 964 L: 844 D: 2621

And with some difficult long TC of 60+0.05:
LLR: 2.95 (-2.94,2.94)
Total: 64133 W: 11968 L: 11532 D: 40633

bench: 4821467
2013-05-23 17:59:39 +02:00
Marco Costalba d3608c4e79 Microptimize MoveList loop
Add MOVE_NONE at the tail, this allows to loop
across MoveList checking for *it != MOVE_NONE,
and because *it is used imediately after compiler
is able to reuse it.

With this small patch perft speed increased of 3%

And it is also a semplification !

No functional change.
2013-05-19 22:00:49 +02:00
Marco Costalba 38cfbeeb50 Delay killers[] initialization
Most of the time we cut-off earlier, at captures, so this
results in useless work.

There is a small functionality change becuase 'ss' can change
from MovePicker c'tor to when killers are tried due, for
instance, to singular search.

bench: 4603795
2013-05-19 21:41:56 +02:00
Marco Costalba 77547a4ef1 Reduce countermoves less in LMR
Passed SPRT for both short TC 15+0.05:
LLR: 2.95 (-2.94,2.94)
Total: 17724 W: 3756 L: 3598 D: 10370

And long TC 60+0.05:
LLR: 2.95 (-2.94,2.94)
Total: 22672 W: 4232 L: 4011 D: 14429

bench: 4418832
2013-05-19 21:36:23 +02:00
Marco Costalba 8ceef92266 Mimic an iterator for looping across MoveList
Seems more conventional.

No functional change.
2013-05-19 13:28:25 +02:00
Joona Kiiski f7c013edd0 Use two counter moves instead of one
Very good at long 60"+0.05 TC
LLR: 2.95 (-2.94,2.94)
Total: 5954 W: 1151 L: 1016 D: 3787

[edit: slightly changed form original patch to avoid useless loop
 across killers when killer is MOVE_NONE]

bench: 4327405
2013-05-16 16:20:50 +02:00
Marco Costalba 148490f04c Rename Refutation to Countermove
Use proper naming according to:

http://chessprogramming.wikispaces.com/Countermove+Heuristic

The name of this idea is "Countermove Heuristic" and was
first introduced by Jos Uiterwijk in 1992

No functional change.
2013-05-15 20:59:56 +02:00
Uri Blass 7c6f346c90 Increased mobility array
Performed more or less well at short TC
LLR: 2.95 (-2.94,2.94)
Total: 50517 W: 9815 L: 9574 D: 31128

And a bit better at long TC
LLR: 2.96 (-2.94,2.94)
Total: 15564 W: 2805 L: 2624 D: 10135

bench: 4375253
2013-05-15 00:34:05 +02:00
Marco Costalba 1a34496761 Revert trapped rook bug fix
It seems that do  not limiting checking the
trapped rook only on rank 1 improves the
score.

At long TC
LLR: 2.97 (-2.94,2.94)
Total: 6581 W: 1346 L: 1204 D: 4031

bench: 4985012
2013-05-15 00:06:11 +02:00
Gary Linscott 049e5ca191 Minor bugfixes to refutation table
Don't update refutation table in case of
previous move is MOVE_NULL or MOVE_NONE
and don't try refutation if is already
a killer move.

Pass both short TC
LLR: 2.96 (-2.94,2.94)
Total: 4310 W: 953 L: 869 D: 2488

And long one
LLR: 2.95 (-2.94,2.94)
Total: 6707 W: 1254 L: 1184 D: 4269

bench: 4785954
2013-05-14 23:52:44 +02:00
Marco Costalba 19dd0de4ff Reformat previous patch
No functional change.
2013-05-13 20:42:44 +02:00
Joona Kiiski 8a61b030a6 Enable refuation table
Very good result both at short TC 15+0.05
LLR: 2.95 (-2.94,2.94)
Total: 2803 W: 596 L: 483 D: 1724

And at long TC 60+0.05
LLR: 2.95 (-2.94,2.94)
Total: 2862 W: 548 L: 431 D: 1883

bench: 4329221
2013-05-13 19:48:41 +02:00
Joona Kiiski c7e31d5aa8 Simple always overwrite Refutation table 2013-05-12 21:21:46 +01:00
Marco Costalba 818a3537a7 Use Them instead of ~Us
Unortunatly we have no guarantee that the call to
operator~(Color c) is resolved at compile time.

Perhaps the solution would be to use C++11 const_expr,
but for now simply use the good old-style ternary operator
that works as expected.

No functional change.
2013-05-11 11:49:44 +02:00
Marco Costalba bcbc9bfd1f Some code reformat in evaluate_pieces
No functional change.
2013-05-11 11:13:06 +02:00
Marco Costalba 7eda7335fd Simplify previous patch
No functional change.
2013-05-09 19:09:51 +02:00
Marco Costalba 02606a8c83 Merge 'passed_pawns' tweaks
Good at both short and long TC

15+0.05
LLR: 2.96 (-2.94,2.94)
Total: 28220 W: 5531 L: 5349 D: 17340

TC 60+0.05
LLR: 2.95 (-2.94,2.94)
Total: 12612 W: 2221 L: 2057 D: 8334

bench: 4857939
2013-05-08 23:04:11 +02:00
Reuven Peleg e7505324f6 Avoid explicit bitwise operators
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-05-05 22:51:18 +02:00
jundery 653c0527a7 Passed pawn eval
Use a stepped function to evaluate bonuses and add the bonus to the
middle game

bench: 4857939
2013-05-05 11:12:04 -06:00
Marco Costalba 7f4c7cd785 Merge increased 'movecount' pruning
Good at both short and long TC

15+0.05
LLR: 2.95 (-2.94,2.94)
Total: 13814 W: 2731 L: 2588 D: 8495

TC 60+0.05
LLR: 2.95 (-2.94,2.94)
Total: 18013 W: 3136 L: 2946 D: 11931

bench: 4306557
2013-05-05 13:46:26 +02:00
Marco Costalba 0958e5c6d3 Simplify previous condition
No functional change.
2013-05-05 12:31:32 +02:00
Marco Costalba 9fc77bc414 Fix trapped rook condition
A rook is trapped if on rank 1 as is the king.
Currently the condition aloows for the rook
to be also in front of the pawns as long
as king is on first rank.

Verified with short TC test:
LLR: -1.71 (-2.94,2.94)
Total: 23234 W: 4317 L: 4317 D: 14600

Here what it counts is that after 23K games
result is equal.

bench: 4696542
2013-05-05 12:27:11 +02:00
Marco Costalba 3b92159908 Further simplify previous patch
No functional change.
2013-05-04 12:15:31 +02:00
homoSapiensSapiens e00bb13e85 Merge some conditions
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-05-04 11:13:22 +02:00
Marco Costalba 3b41e62666 Drop some redundant defined(_WIN64)
When it is already defined(_WIN32).

According to Microsoft documentation:
http://msdn.microsoft.com/en-us/library/b0084kay.aspx

_WIN32 Defined for applications for Win32 and Win64. Always defined.

_WIN64 Defined for applications for Win64.

Patch suggested by Joona.

No functional change.
2013-05-03 15:24:54 +02:00
Marco Costalba 37c91aa94c Print time and node count before search ends
This info is normally printed together with
PV info in uci_pv() but when search is stopped,
for instance when max search time is reached,
uci_pv is not called and we miss this bits.

Suggested by gravy_train

No functional change.
2013-05-03 10:26:03 +02:00
Marco Costalba 43f67eab5f Merge mobility area tweak
A nice improvment.

Was good at 15+0.05
LLR: 2.96 (-2.94,2.94)
Total: 10731 W: 2176 L: 2040 D: 6515

And at 60"+0.05
LLR: 2.96 (-2.94,2.94)
Total: 10601 W: 1968 L: 1810 D: 6823

bench: 4676606
2013-05-03 10:12:31 +02:00
Gary Linscott 11d30b6298 Fix rounding issue 2013-05-02 14:37:55 -04:00
Gary Linscott 3edb15d183 More aggressive move count pruning 2013-05-02 09:47:34 -04:00
Marco Costalba d44ac0a485 Another take at TT alignment
This time revert to original version but using
uintptr_t instead of size_t

Suggested by Lucas.

No functional change.
2013-05-02 09:38:23 +02:00
Marco Costalba 481eda4ca0 Re-add "Cache line aligned TT"
But this time do not play with pointers, in
particular do not assume that size_t is an
unsigned type of the same width as pointers.

This code should be fully portable.

No functional change.
2013-05-01 23:42:16 +02:00
jhellis3 7323231786 Tweak Mobility Area
Only consider pawns and the king as restricting.
2013-05-01 02:37:50 -05:00
Marco Costalba e381951a24 Restore development version
No functional change.
2013-04-30 20:01:07 +02:00
Marco Costalba aa2368a687 Stockfish 3
Stockfish bench signature is: 4176431
2013-04-30 19:42:43 +02:00
Marco Costalba 293c44bc09 Revert "Cache line aligned TT"
This reverts commit 083fe58124

It seems to break Android build

No functional change.
2013-04-30 19:42:21 +02:00
Marco Costalba 06b9140e5c Temporary revert "Expose EvalInfo struct to search"
It is not needed for the release and introduces
a slowdown, although very small.

Probably it will be readded after the release.

No functional change.
2013-04-29 00:55:32 +02:00
Marco Costalba 156635749b Fix a 'value > VALUE_INFINITE' assert
This fixes an assert while testing with debug on.

Assert was due to static null pruning returning value

eval - futility_margin(depth, (ss-1)->futilityMoveCount)

That was sometimes higher than VALUE_INFINITE triggering
an assert at the caller site.

Because eval con be equal to ttValue and anyhow is read from
TT that can be corrupted in SMP case, we need to sanity
check it before to use.

bench: 4176431
2013-04-27 13:08:11 +02:00
Marco Costalba 083fe58124 Cache line aligned TT
Let TT clusters (16*4=64 bytes) to hold on a singe cache line.
This avoids the need for the double prefetch.

Original patches by Lucas and Jean-Francois that has also tested
on his AMD FX:

BIG HASHTABLE

./stockfish bench 1024 1 18 > /dev/null

Before:
1437642 nps
1426519 nps
1438493 nps

After:
1474482 nps
1476375 nps
1475877 nps

SMALL HASHTABLE

./stockfish bench 128 1 18 > /dev/null

Before:
1435207 nps
1435586 nps
1433741 nps

After:
1479143 nps
1471042 nps
1472286 nps

No functional change.
2013-04-26 19:38:11 +02:00
Marco Costalba e508494a99 Fix a crash introduced few days ago
Crash is due to uninitialized ss->futilityMoveCount that
when happens to be negative, yields to an out of range
access in futility_margin().

Bug is subtle because it shows itself only in SMP case.
Indeed in single thread mode we only use the

Stack ss[MAX_PLY_PLUS_2];

Allocated at the begin of id_loop() and due to pure
(bad) luck, it happens that for all the MAX_PLY_PLUS_2
elements, ss[i].futilityMoveCount >= 0

Note that the patch does not prevent futilityMoveCount
to be overwritten after, for instance singular search
or null verification, but to keep things readable and
because the effect is almost unmeasurable, we here
prefer a slightly incorrect but simpler patch.

bench: 4311634
2013-04-26 12:14:01 +02:00
Marco Costalba 2ef53ee368 Store Eval::Info in Search::Stack
Instead of a pointer. This should fix the issue of
remaining with a stale pointer when for instance calling
IID, but also null search verification, singular search
and razoring where we call search with the same ss
pointer. In this case ss->ei is overwritten in the
search() call and upon returning remains stale.

This patch could have a performance hit because Eval::Info
is big (176 bytes) and during splitting we copy 4 ss entries.

On the good side, this patch is a clean solution.

Proposed by Gary.

No functional change.
2013-04-25 21:52:26 +02:00
Marco Costalba d810441b35 Expose EvalInfo struct to search
Allow to use EvalInfo struct, populated by
evaluation(), in search.

In particular we allocate Eval::Info on the stack
and pass a pointer to this to evaluate().

Also add to Search::Stack a pointer to Eval::Info,
this allows to reference eval info of previous/next
nodes.

WARNING: Eval::Info is NOT initialized and is populated
by evaluate(), only if the latter is called, and this
does not happen in all the code paths, so care should be
taken when accessing this struct.

No functional change.
2013-04-25 12:57:37 +02:00
Ryan Schmitt 79bcb2ca54 Increase rook/queen on 7th bonus
Shows an increase at 15+0.05
LLR: 3.01 (-2.94,2.94)
Total: 20450 W: 4091 L: 3927 D: 12432

And at 60+0.05
LLR: 2.97 (-2.94,2.94)
Total: 61432 W: 10849 L: 10441 D: 40142

bench: 4493356

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-04-25 12:13:17 +02:00
Marco Costalba 289a767ab3 Merge Joona's increased static null pruning
The idea is to fail high more easily in static
null test if in the parent node we are already
very deep in the move list, so the propability
to fail high there is very low.

[edit: I have slightly changed the functionality
moving

ss->futilityMoveCount = moveCount;

At the end of the pruning code, this should not affect
ELO in anyway, but makes code more natural and logic]

Test with SPRT is good at 15+0.05
LLR: 2.96 (-2.94,2.94)
Total: 50653 W: 10024 L: 9780 D: 30849

And at 60+0.05
LLR: 2.97 (-2.94,2.94)
Total: 40799 W: 7227 L: 6921 D: 26651

bench: 4530093
2013-04-25 12:05:00 +02:00
Marco Costalba 4381ea23fe Fix cpu_count() on some platforms
When we use sysconf(_SC_NPROCESSORS_ONLN) to get number of
cores, we have to include sysconf library that is unistd.h

Sometimes it happens to work just becuase unistd.h indirectly
included by some other libraries, but not always.

Reported and fixed by Eyal BD

No functional change.
2013-04-25 10:56:56 +02:00
Joona Kiiski 817ca1820b Fix potential overflow 2013-04-23 07:26:36 +01:00
Joona Kiiski 8df17204f4 More aggressive post-futility pruning 2013-04-21 14:53:27 +01:00
Marco Costalba f84f04742a Skip a couple of popcount in previous patch
And some little tidy up

No functional change.
2013-04-19 10:31:18 +02:00
Marco Costalba cc40d1c46a Merge Joona's bishop+pawn tweak
The idea is to penalize a bishop in case of
its pawns are on the same colored squares.

Good at short 15"+0.05 TC
LLR: 2.95 (-2.94,2.94)
Total: 4252 W: 925 L: 806 D: 2521

And at longer 60"+0.05 TC
LLR: 2.95 (-2.94,2.94)
Total: 15006 W: 2743 L: 2564 D: 9699

bench: 5274705
2013-04-19 09:36:00 +02:00
Joona Kiiski 818e0b2d2b Try more aggressive version 2013-04-17 07:58:59 +01:00
Joona Kiiski 3e4dfb49a7 Give a small penalty for bishop for each pawn on the same colored square 2013-04-16 21:20:31 +01:00
Marco Costalba 87436e5544 Skip a redundant check
Spotted by Joona

No functional change.
2013-04-14 23:29:25 +02:00
Marco Costalba fe72c93141 De-templetize Position::is_draw()
Now that we always check for repetition we don't
need a template anymore.

No functional change.
2013-04-10 22:23:48 +02:00
Joona Kiiski 75221fcf5e Always check repetition
It seems stronger both at fast 15+0.05 TC with fixed game number test:
ELO: 2.74 +-2.7 (95%) LOS: 97.6%
Total: 24000 W: 4698 L: 4509 D: 14793

And also at long 60+0.05 TC with SPRT
LLR: 3.05 (-2.94,2.94)
Total: 38986 W: 6845 L: 6547 D: 25594

bench: 5157061
2013-04-10 22:20:40 +02:00
Marco Costalba a95cbca568 Simplify and speed up previous patch
Use an optinal argument instead of a template
parameter. Interestingly, not only is simpler,
but also faster, perhaps due to less L1 instruction
cache pressure because we don't duplicate the very
used SEE code path.

No functional change.
2013-04-09 23:32:06 +02:00
Joona Kiiski d23454854e Document asymmetric SEE pruning trick
Here are the tests:

sprt @ 60+0.05
ELO: 3.53 +-2.8 (95%) LOS: 99.3%
Total: 18794 W: 3098 L: 2907 D: 12789

16000 @ 60+0.05
ELO: 1.39 +-3.1 (95%) LOS: 81.0%
Total: 16000 W: 2689 L: 2625 D: 10686

16000 @ 15+0.05
ELO: 2.82 +-3.3 (95%) LOS: 95.1%
Total: 16000 W: 3148 L: 3018 D: 9834

No functional change

Signature: 4969307
2013-04-09 23:31:57 +02:00
Joona Kiiski c2902112e5 Don't treat king safety differently in AnalysisMode
Rationale:

- Current settings seem to make engine *significantly* weaker in analysis mode.
- In practice this setting only has effect when king safety scores are high.
- Even in analysis mode its far more important to know if one side is getting mated,
rather than get evaluation correct with 1cp accuracy.

No functional change
2013-04-09 23:29:58 +02:00
Marco Costalba 9498b2af82 Rescale UCI parameters to 100
And correspondingly modify internal ones
to compensate it.

No functional change.
2013-04-09 23:29:58 +02:00
Marco Costalba 2a5ae34bb2 Tweak some UCI parameters
According to Jean-Paul this setup should be stronger
than default.

And SPRT test seems to confirm it:

At fast TC 15"+0.05
ELO: 3.33 +-2.7 (95%) LOS: 99.2%
Total: 25866 W: 5461 L: 5213 D: 15192

At longer TC 60"+0.05
ELO: 7.27 +-5.0 (95%) LOS: 99.8%
Total: 6544 W: 1212 L: 1075 D: 4257

bench: 5473339

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-04-09 23:29:58 +02:00
Marco Costalba d3fe153fe6 Re-add 'Cowardice' and 'Aggressiveness' UCI options
I have lost my bet with Jean-Paul, so now I re-add
the two options...and I am glad of it :-)

No functional change.
2013-04-09 23:29:58 +02:00
Marco Costalba 889922041b Increase null verification threshold to 12 plies
Increasing depth limit to 10 plies seems stronger
after 16K games at 15"+0.05 (ELO: +3.56) and also
repeating the test at 60"+0.05 TC:

ELO: 2.08 +-3.1 (95%) LOS: 90.9%
Total: 16000 W: 2641 L: 2545 D: 10814

Moreover setting the limit to 12 is proved stronger
then limit set to 10 by direct SPRT test at 15"+0.05:

ELO: 2.56 +-2.0 (95%) LOS: 99.5%
Total: 46568 W: 9240 L: 8897 D: 28431

So we directly set the limit to 12, the strongest setup.

bench: 4361224
2013-04-09 23:29:58 +02:00
Joona Kiiski 7bad50773a Make use of asymmetric SEE 2013-04-05 19:05:32 +01:00
Joona Kiiski d3c3c4f8e7 Fix a silly bug 2013-04-05 19:05:32 +01:00
Joona Kiiski 2097cd1221 Introduce asymmetric SEE.
No functional change
2013-04-05 19:05:32 +01:00
Marco Costalba 0c1b40c5e2 Lower minimum allowed TT size to 1 MB
Setting a very low TT size could be used
for some specific testing.

No functional change.
2013-04-05 16:56:01 +02:00
Marco Costalba 8c10029df1 Revert "Double Impact of Gain tables"
This reverts commit 36c82b751c

Seems a regression against 2.3.1 tested with 20K games at 60"+0.05

With patch applied
ELO: 15.44 +-2.8 (95%) LOS: 100.0%
Total: 20000 W: 3928 L: 3040 D: 13032

Without patch applied
ELO: 18.76 +-2.8 (95%) LOS: 100.0%
Total: 20000 W: 3903 L: 2824 D: 13273

bench: 4781239
2013-04-05 08:59:38 +02:00
Hiraoka Takuya 6e2371a86b Don't early stop if we have a mated score
No functional change.
2013-04-04 21:39:48 +02:00
Joona Kiiski 36c82b751c Double Impact of Gain tables
Very unorthodox idea.

After 16000 games at 60"+0.05
ELO: 3.14 +-3.4 (95%) LOS: 96.6%
Total: 13407 W: 2278 L: 2157 D: 8972

bench: 4705335
2013-04-03 19:14:59 +02:00
Marco Costalba 7d42d02ec7 Set IID half way between d/2 and d-4
Master IID formula is depth / 2
Previous patch is depth - 4 * ONE_PLY

This one is the middle way:

(dept/2 + depth-4*ONE_PLY)/2  -> depth-2*ONE_PLY-depth/4

After 16000 games at 60+0.05 th 1
ELO: 4.08 +-3.1 (95%) LOS: 99.5%
Total: 16000 W: 2742 L: 2554 D: 10704

bench: 4781239
2013-03-30 22:26:30 +01:00
Marco Costalba c89274d8fb Merge branch 'master' into increase_iid 2013-03-29 22:50:04 +01:00
Marco Costalba f2638816bf Raise Min Split Depth
Raise the limit to 12 so to allow people to test
on many cores machines.

Suggested by Gary and Martin.

No functional change.
2013-03-25 20:04:49 +01:00
Marco Costalba 0b4ea54da9 Update bestValue when futility pruning (2)
Same idea of 5af8179647
in qsearch() but applied to search()

After 15500 games at 15+0.05
ELO: 4.48 +-3.4 (95%) LOS: 99.5%
Total: 15500 W: 3061 L: 2861 D: 9578

bench: 4985829
2013-03-24 23:32:21 +01:00
Joona Kiiski 09f1fdf52f Fix bogus mate scores in some positions
Always before pruning the move, it's important to check that:

bestValue > VALUE_MATED_IN_MAX_PLY

See example position:

8/2p1p3/P1NpP3/3k4/1P1BN3/2P1P3/2Q5/6K1 w - - 0 1
http://support.stockfishchess.org/discussions/problems/268-wrong-declaring-a-forced-mate-in-3-moves

This problem was present in 2.3.1, then it was fixed by my patch.

After 24000 games at 15+0.05
ELO: 2.40 +-4.4 (95%) LOS: 95.7%
Total: 24000 W: 4774 L: 4608 D: 14618

bench: 4465997
2013-03-23 21:28:51 +01:00
jundery d39b22927e Use ALL_PIECES value to reference attackedBy
No functional change
2013-03-23 12:17:12 +01:00
Marco Costalba 81e9cf043a Increase non-PV IID search depth
bench: 5146380
2013-03-22 20:57:31 +01:00
Marco Costalba 077e32efc9 Better document bitbase loop
Thanks to Lucas to spot the weak comment and
to Jundery to suggest a better one.

No functional change.
2013-03-19 19:08:41 +01:00
Gary Linscott 52f3e717fa Add KNPKB endgame
In a game vs Junior, SF had the option to trade into a winning
pawn endgame, and failed to do so. PGN at bottom.

This FEN was one key position: 8/2Nb1k2/6pp/4Pp2/5K1P/5PP1/8/8 w - - 5 62.

SF master chooses h5 here, with a fail high, which goes into the drawn KNPKB
ending. With the patch, SF correctly chooses Ke3, which maintains chances to win.

[Event "nTCEC - Stage 2a - Season 1"]
[Site "http://www.tcec-chess.net"]
[Date "2013.03.05"]
[Round "11.3"]
[White "Stockfish 210213"]
[Black "Junior 13.3"]
[Result "1/2-1/2"]
[Variant "normal"]

1. d4 f5 2. g3 Nf6 3. Bg2 e6 4. c4 d5 5. Nh3 c6 6. O-O Bd6 7. Bf4 Be7 8. Nd2 O-O
9. Qb3 a5 10. Rfd1 Ne4 11. Be3 Nd7 12. Nf4 Ndf6 13. f3 a4 14. Qc2 Nxd2 15. Bxd2
dxc4 16. Qxc4 b5 17. Qc3 Qb6 18. Rac1 e5 19. Nd3 exd4 20. Qxc6 Qxc6 21. Rxc6
Bd7 22. Rcc1 Be6 23. Bb4 Rae8 24. Bxe7 Rxe7 25. Nb4 Bc4 26. Bf1 Rd7 27. Rd2 Re8
28. Rcd1 Rc7 29. Ra1 Rd8 30. Rc1 Rdd7 31. Rcd1 Re7 32. Ra1 Nd5 33. Nc2 Ne3
34. Nxd4 Bxa2 35. Nxb5 Rc5 36. Nd4 Bf7 37. Kf2 g6 38. Rd3 Nxf1 39. Kxf1 Rc4
40. b3 axb3 41. Nxb3 Kf8 42. Rd8+ Re8 43. Rxe8+ Bxe8 44. Kf2 Ke7 45. Ra7+ Bd7
46. Ke1 Rc3 47. Rb7 Rc2 48. Kd1 Rc4 49. Kd2 Kd6 50. Kd3 Rc7 51. Rxc7 Kxc7
52. Kd4 Kd6 53. Nc5 Bb5 54. e4 Be2 55. e5+ Ke7 56. Ke3 Bd1 57. Kf4 h6 58. h3
Kf7 59. h4 Bc2 60. Na6 Ba4 61. Nc7 Bd7 62. h5 g5+ 63. Ke3 Ba4 64. f4 Bd1
65. fxg5 hxg5 66. h6 Kg6 67. e6 f4+ 68. gxf4 gxf4+ 69. Kxf4 Bh5 70. Ke5 Kh7
71. Kf6 Kxh6 72. Na6 Bg4 73. e7 Bh5 74. Nc7 Bg6 75. Nd5 Be8 76. Ne3 Kh7 77. Nc4
Kh6 78. Nd2 Kh5 79. Nf3 Kg4 80. Nd4 Bh5 81. Ne6 Be8 82. Nc5 Kf3 83. Kf5 Ke3
84. Ke5 Ke2 85. Kf4 Kd2 86. Ne4+ Kd3 87. Ke5 Ke3 88. Nf6 Bf7 89. Nd5+ Kf3
90. Kf5 Ke2 91. Ke4 Be8 92. Nc3+ Kd2 93. Kd4 Kc2 94. Nd5 Kd1 95. Nf6 Bf7
96. Ne4 Be8 97. Ke3 Kc2 98. Nd6 Bd7 99. Kd4 Kd1 100. Kd3 Ba4 101. Nc4 Bb5
102. Kc3 Be8 103. Nb2+ Ke1 104. Kd3 Kf2 105. Nd1+ 1/2-1/2

No functional change (just because bench does not change)
2013-03-18 20:40:06 +01:00
Marco Costalba 0586b51f9c Further increase SEE prune depth
After 16000 games at 60+0.05
ELO: 2.89 +-5.4 (95%) LOS: 96.5%
Total: 16000 W: 2775 L: 2642 D: 10583

bench: 5442365
2013-03-16 11:20:03 +01:00
Marco Costalba d2eeef89aa Revert "Check for easy move just once"
This reverts commit a24da071f0

Seems a regression when tested against 2.3.1

With this patch, have after 20000 games at 60+0.05, we have

ELO: 13.42 +-4.8 (95%) LOS: 100.0%
Total: 20000 W: 3746 L: 2974 D: 13280

Instead with the patch reverted:

ELO: 16.62 +-4.8 (95%) LOS: 100.0%
Total: 20000 W: 3816 L: 2860 D: 13324

Although we are within error bounds here we take the conservative
approach of not introducing changes that are not proved stronger
It doesn't mean that the change shall be weaker, simply that we
don't want to take any risk.

No functional change.
2013-03-16 11:12:35 +01:00
RyanTaker 70d20326b0 Improved Readability of Material
This is a non-functional change that simply changes the look
of the code to help clarity.

No functional change.
2013-03-15 09:14:00 +01:00
jundery ccf4ec6768 Do more work between prefetch and querying transposition table
More time to load the cache line before access

No functional change.
2013-03-12 19:58:32 +01:00
Gary Linscott a24da071f0 Check for easy move just once
Here the rational seems to be that if after one try easy
move detection fails then the easy move is not so easy :-)

After 15563 games at 60+0.05
ELO: 3.04 +-5.5 (95%) LOS: 97.0%
Total: 15563 W: 2664 L: 2528 D: 10371

No functional change.
2013-03-11 22:23:19 +01:00
Gary Linscott 3698d9aa55 Be more aggressive on trying to finish iterations
Increase MaxRatio to use more time when in trouble.

After 16000 games at 60+0.05
ELO: 4.89 +-5.4 (95%) LOS: 99.9%
Total: 16000 W: 2700 L: 2475 D: 10825

No functional change.
2013-03-11 19:07:55 +01:00
Marco Costalba 10429dd616 Increase see prune depth
This seems good at short TC controls.

After 10000 games at 20+0.05
ELO: 9.56 +-6.8 (95%) LOS: 100.0%
Total: 10000 W: 1949 L: 1674 D: 6377

Testing at long TC and regression testing is still
ongoing. So this is a bit speculative commit and
could be reverted in the future.

Also re-testing at long TC the SEE pruning in PV nodes
seems less effective (perhaps even a regression, but
still ongoing) so disabled for now.

bench: 4968764
2013-03-04 09:41:56 +01:00
Marco Costalba db322e6a63 Revert "Store moves sent with "position" UCI command"
This reverts commit 0d68b523a3.

After easy move semplification this machinery is not
needed anymore (because of we don't need to know if a
root move is a recapture)

No functional change.
2013-03-04 09:29:46 +01:00
Marco Costalba 45dba12c5b Simplify "easy move" detection
Detect a move as easy only if it is the only one ;-)
or if is much better than remaining ones after we
have spent 20% of search time.

Tests are ongoing, but it seems this semplification
stands. Anyhow it is experimental for now and could
be reverted/improved with further work Gary is
testing right now.

No functional change.
2013-03-04 09:27:00 +01:00
Marco Costalba ccad601389 Avoid locking/unlocking in a tight loop
After previous patch if split point master is
waiting for job and "Use Sleeping Threads" is
false (our condition for official releases) then
it will lock/unlock splitPoint mutex in a super
tight loop badly affecting performance.

Rewrite the code to lock only when we are about
to finish. Note that race condition on slavesMask
is anyhow fixed.

No functional change.
2013-03-04 09:07:48 +01:00
jundery d165d5af91 Fix race condition where idle_loop() gets called from Split()
SplitPoint member slavesMask wasn't read under lock

No functional change.
2013-03-04 08:52:24 +01:00
jhellis3 3ce43c20de Stop search if only 1 legal move
There is no point searching a move that is forced.
It wastes time while allowing computer opponents to
fill hash with 100% accuracy.

[edit: Condition moved together with "easy move" ones]

Bench identical: 4922272
2013-03-04 08:30:55 +01:00
Marco Costalba def50020ad Fix easy re-capture case
We detect an easy move as a recapture with an
high margin on the second best move.

Unfortunatly the recapture detection is broken
becuase we identify as a recapture any move that
follows an opponent's previous capture !

This patch fix the logic to correctly detect a
real re-capture.

No functional change.
2013-03-02 13:20:40 +01:00
Marco Costalba 0d68b523a3 Store moves sent with "position" UCI command
Store all the game moves until current position.

This will be used by next patch.

No functional change.
2013-03-02 13:08:50 +01:00
Marco Costalba 0e1ad3ad33 Rename sp to splitPoint
Still keep 'sp' name when used as local
variable with limited scope.

From Jundery.

No functional change.
2013-03-01 09:44:19 +01:00
jundery 0fc9a01933 Remove strange use of the ternary operator
Note that we read shared data without lock
protection, so code is theoretically prone to
torn reads. But, first splitPoint pointer
never changes, and alpha is of integer type so
it is read in a single DWORD access.

No functional change.
2013-03-01 08:05:47 +01:00
jundery 68d1bebd8e Split() clean up locking
Only unlock and relock when idle_loop() is actually called

No functional change
2013-03-01 07:57:18 +01:00
Marco Costalba 57b6df4874 Merge Lucas's "SEE pruning at PV nodes"
bench: 4922272
2013-02-27 08:14:00 +01:00
Marco Costalba bc38efd128 Remove pruning condition on alpha
Further simplifying on Lucas's idea, seems reliable
in tests:

ELO: 2.15 +-7 (95%) LOS: 84.9%
Total: 9999 W: 1831 L: 1769 D: 6399
2013-02-27 08:07:26 +01:00
Lucas Braesch 335b57b5ef Prune negative SEE moves also in PV nodes
This patch is actually the sum of two contributions that
have been tested independently:

1) Pruning of negative SEE moves in PV

After 10000 games at 20+0.05
ELO: 5.18 +-7 (95%) LOS: 99.2%
Total: 10000 W: 1952 L: 1803 D: 6245

2) Remove of bestValue > VALUE_MATED_IN_MAX_PLY condition

After 23000 games at 20+0.05
ELO: 1.63 +-4 (95%) LOS: 88.1%
Total: 23000 W: 4232 L: 4124 D: 14644

The whole patch as been re-tested at long TC with positive results:

After 10000 games at 60+0.05
ELO: 4.31 +-7 (95%) LOS: 98.3%
Total: 10000 W: 1765 L: 1641 D: 6594
2013-02-27 08:04:32 +01:00
Marco Costalba dbd28bc7f8 Avoid a tricky line in shelter_storm()
kf = (kf == FILE_A) ? kf++ : ....

is tricky becuase kf is updated twice and it happens
to do the right thing just by accident.

Rewrite in a better way.

Spotted by pdimov

No functional change.
2013-02-23 19:27:32 +01:00
Marco Costalba ccf21f5595 Convert Readme to markdown format
Looks better on GitHub, that supports this format.

No functional change.
2013-02-23 17:05:28 +01:00
Marco Costalba 1d1fcf80a0 Use DD-MM-YY as date format
In engine name and version number.

No functional change.
2013-02-21 07:19:48 +01:00
Marco Costalba 373503f4a9 Statically link std libraries under mingw
Allows for easier redistribution.

No functional change.
2013-02-21 05:51:04 +01:00
Marco Costalba 9a1d5f0f1d Merge Gary's bishop_pin patch
Give a bonus if a bishop can pin a piece or can
give a discovered check through an x-ray attack.

Seems good after 24000 games at 15"+0.05 (single thread):

ELO: 12.30 +- 99%: 5.79 95%: 4.40 LOS: 100.00%
Total: 24000 W: 4931 L: 4082 D: 14987

bench: 4917064
2013-02-20 12:39:09 +01:00
jundery d5e49a3ad4 Print leading zeroes in hash keys
And convert to uppercase. Reset the stream to dec too.

[Edit: Also fixed the hash key in Position::pretty()]
2013-02-19 20:06:01 +01:00
Gary Linscott ea4e22be1d Merge branch 'master' into bishop_pin_clop 2013-02-19 10:31:52 -05:00
Gary Linscott dc8c6ea2d1 Bring back original bonus 2013-02-19 10:31:50 -05:00
Marco Costalba c5ec94d0f1 Update copyright year
No functional change.
2013-02-19 07:54:14 +01:00
Gary Linscott d570260a28 Back to CLOP average values 2013-02-16 22:36:58 -05:00
Marco Costalba 76caef8ba1 Account for gamePly after each move
Rename startPosPly to gamePly and increment/decrement
the variable after each do/undo move. This adds a little
overhead in do_move() but we will need to have the
game ply during the search for the next patches now
under test.

Currently we don't increment gamePly in do_null_move()
becuase it is not needed at the moment. Could change
in the future.

As a nice side effect we can now remove an hack in
startpos_ply_counter().

No functional change.
2013-02-16 12:44:17 +01:00
Marco Costalba 5a156df719 Merge Gary's king safety tweak
Still well within error bars, so probably not a big improvement,
but may be worthwhile. I will let the test keep running. The idea
for the tweak came from the TCEC game against Houdini. Stockfish saw
it as a draw, well past when it should have seen problems. King safety
was off, since it was QN and pawns only, but in fact the king was
quite vulnerable.

After 8000 games at 60/1 (Gary's test)

ELO: -0.43 +- 99%: 10.02 95%: 7.62
Wins: 1235 Losses: 1245 Draws: 5520 Total: 8000

PGN of game against houdini. Moves 55-59 it was seeing a draw, and
Houdini was seeing a good sized advantage for black. With the change,
Stockfish now recognizes that moving the king there is a bad idea.

[Event "nTCEC - Stage 1 - Season 1"]
[Site "http://www.tcec-chess.net"]
[Date "2013.02.07"]
[Round "4.2"]
[White "Stockfish 2.31"]
[Black "Houdini 3.0"]
[Result "0-1"]
[Variant "normal"]

1. Nf3 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 O-O 5. a3 Bxc3 6. Qxc3 b6 7. b4 a5 8. Bb2
axb4 9. axb4 Rxa1+ 10. Bxa1 Na6 11. e3 Qe7 12. b5 Nc5 13. Qc2 Bb7 14. Be2 d6
15. O-O Ra8 16. Bb2 h6 17. Ra1 Rxa1+ 18. Bxa1 e5 19. Qa2 Be4 20. Ne1 Bg6
21. Bb2 Kh7 22. f3 Nfd7 23. Bc3 h5 24. Qa1 h4 25. Kf1 Qf6 26. Qa7 Qd8 27. Qa2
e4 28. Qa1 h3 29. g3 exf3 30. Nxf3 Bf5 31. Bd4 g6 32. Kf2 Kg8 33. Qa3 Bg4
34. d3 Qc8 35. Bxc5 bxc5 36. Ke1 Qb7 37. e4 Bxf3 38. Bxf3 Ne5 39. Be2 Qc8
40. Qa6 Qd8 41. Qa5 Kh7 42. Kd1 Kg7 43. Qa7 Kg8 44. Qa5 Kh7 45. Qa7 Kg7 46. Qa5
Qb8 47. Kd2 Kh7 48. Kc2 Kg8 49. Qa6 Qd8 50. Qa5 Qf6 51. Qe1 Kg7 52. Qf1 Qe6
53. Qe1 Qd7 54. Qc1 Qe8 55. Kc3 Qa8 56. Kb3 Kh7 57. Qa3 Qd8 58. Qc1 c6 59. Qc3
Qb6 60. Ka4 Qb7 61. Qd2 Nd7 62. Qc3 Qa7+ 63. Kb3 Kg8 64. Bg4 cxb5 65. cxb5 Nb6
66. Bxh3 Qa4+ 67. Kb2 Qd1 68. Ka3 Qe2 69. Kb3 Qh5 70. Bg2 Qxh2 71. Qf6 Qxg3
72. Bf1 Qe3 73. Kc2 Na4 74. b6 Nxb6 75. Qd8+ Kg7 76. Qxb6 Qf2+ 77. Kc3 Qxf1
78. Qxd6 Qf6+ 79. Qxf6+ Kxf6 80. Kc4 g5 81. Kxc5 Ke5 82. d4+ Kxe4 83. d5 g4
84. d6 g3 85. d7 g2 86. d8=Q g1=Q+ 87. Kb5 Qb1+ 88. Ka4 Qa2+ 89. Kb4 f5
90. Qe8+ Kf4 91. Qg6 Qd5 92. Qg1 Qe4+ 93. Ka5 Qe2 94. Qg8 Kf3 95. Qd5+ Qe4
96. Qd1+ Kg2 97. Qd2+ Kf1 98. Qh2 f4 99. Qh3+ Ke2 100. Qg4+ Kd2 101. Qg5 Kc2
102. Qh4 0-1

bench: 5518286
2013-02-15 16:25:33 +01:00
Marco Costalba e0dfb0bc34 Further speed up bitbase generation
Another trick, along the same lines of previous
patch. This time we first check positions with
white side to move that, becuase we start with
pawn on rank 7, are easily classified as wins,
then black ones.

Number of cycles reduced to 15 !

Becuase now it is faster we can remove a lot of
code to detect theoretical draws. We will calculate
them anyhow, although a bit slower, but the speed
up trick more than compensates it.

Verified that generated bitbases match original ones.

No functional change.
2013-02-15 11:58:33 +01:00
Gary Linscott 3922ac2fe4 Add new clop tuned value 2013-02-14 20:29:24 -05:00
Gary Linscott 66a1a77487 Merge branch 'master' into bishop_pin_clop 2013-02-13 21:41:15 -05:00
Gary Linscott cd0ecace19 Revert "Use CLOP mean value instead of max"
This reverts commit d0c2faa5fd.
2013-02-13 21:40:38 -05:00
Marco Costalba 10d29add18 Speedup KPK bitbase of 25%
Change the way the index is coded so that
now looping from 0 to IndexMax generates
the pawns from RANK_7 down to RANK2.

Becuase positions with pawns at RANK_7
are easily classified as wins/draws, this
small trick allows to reduce the number
of needed iterations from 30 down to 26!

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-02-13 20:13:56 +01:00
Marco Costalba f2950ae206 Simplify bitbase.cpp
Use a std::vector to store positions and
rearrange KPKPosition.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-02-13 20:13:44 +01:00
Gary Linscott d0c2faa5fd Use CLOP mean value instead of max 2013-02-13 12:21:16 -05:00
Gary Linscott 671f24ff35 Merge branch 'simplify_eval' into bishop_pin_clop 2013-02-13 12:19:54 -05:00
Gary Linscott f32f899467 CLOP tuned 2013-02-13 00:12:02 -05:00
Gary Linscott 17b71fe51d Add clop parameters 2013-02-12 00:10:21 -05:00
Gary Linscott 30c2e3828a Merge branch 'master' into simplify_eval 2013-02-11 10:26:25 -05:00
Gary Linscott a7cdf7299a Bishop pins only 2013-02-11 10:26:18 -05:00
Marco Costalba 733d0099b2 Rename and de-templetize sort()
Rename to insertion_sort so to avoid confusion
with std::sort, also move it to movepicker.cpp
and use the bit slower std::stable_sort in
search.cpp where it is used in not performance
critical paths.

No functional change.
2013-02-11 00:09:21 +01:00
Marco Costalba 0be7b8c542 Further simplify first_entry()
We can encode the ClusterSize directly in the
hashMask, this allows to skip the left shift.

There is no real change, but bench number is now
different because instead of using the lowest order
bits of the key to index the start of the cluster,
now we don't use the last two lsb bits that are
always set to zero (cluster size is 4). So for
instance, if 10 bits are used to index the cluster,
instead of bits [9..0] now we use bits [11..2].
This changes the positions that end up in the same
cluster affecting TT hits and so bench is different.

Also some renaming while there.

bench: 5383795
2013-02-09 16:37:20 +01:00
Marco Costalba c698362680 Microptimize first_entry() for 32bits
Do a 32bit bitwise 'and' instead of a 64bit
subtract and bitwise 'and'.

This is possible because even in the biggest
hash table case (8GB) the number of entries
is 2^29 so storable in an unsigned int.

No functional change.
2013-02-09 10:55:38 +01:00
Marco Costalba fe3352665b Retire TTCluster and simplify TT
Also some renaming while there.

No functional change.
2013-02-09 10:55:20 +01:00
Marco Costalba 3f44be9baa Simplify move_to_san()
Nicely simplify disambiguation code.

No functional change.
2013-02-09 07:43:53 +01:00
Gary Linscott 0c06b4509d Slight tweak to king safety. Bench: 5534531 2013-02-08 08:53:13 -05:00
Marco Costalba e5bc79fb9c Retire slavesPositions
Save the current active position in each Thread
instead of keeping a centralized array in struct
SplitPoint.

This allow to skip a memset() call at each split.

No functional change.
2013-02-08 11:45:33 +01:00
Marco Costalba 880726c13a Add const qualifer to go()
Obsolete renmant of when position was directly
passed to the search instead of being copied
for the main thread as is now.

From Jundery.

No functional change.
2013-02-08 10:07:09 +01:00
Marco Costalba 50a7200b18 Workaround value-initialization in MSVC
The syntax splitPoints() should force the compiler to
value-initialize the array and because there is no
user defined c'tor it falls back on zero-initialization.

Unfortunatly this is broken in MSVC compilers, because
value initialization for non-POD types is not supported,
so left splitPoints un-initialized and add in split()
initialization of slavesPositions, that is the only
member not already set at split time.

This fixes an assert under MSVC when running with
more than one thread.

Spotted and reported by Jundery.

No functional change.
2013-02-08 09:20:40 +01:00
Gary Linscott c67fb8ef04 Add KBPKP endgame
It is a draw if pawns are on G or B files, weaker pawn is
on rank 7 and bishop can't attack the pawn.

No functional change (because it is very rare and does not appear in bench)
2013-02-07 06:51:55 +01:00
Marco Costalba 14c2c1395b Change slave_available() API
To return a pointer to the available
thread instead of a bool. This allows
to simplify the core loop in split().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-02-06 20:48:26 +01:00
jundery 88c3670edf Rename posKey stored in the transposition table
[Edit: Slightly extended by me]

No functional change.
2013-02-06 08:03:37 +01:00
jundery f69c185e02 Add const qualifer to check_is_dangerous
No functional change.
2013-02-06 07:49:40 +01:00
Marco Costalba bf706c4a4f Slightly change split() API
This function "returns" two values: bestValue and bestMove

Instead of returning one and passing as pointer the other
be consistent and pass as pointers both.

No functional change.
2013-02-05 06:35:38 +01:00
Marco Costalba 1a414cd9cb Derive ThreadPool from std::vector
Prefer sub-classing to composition in this case.

No functional change.
2013-02-04 22:59:20 +01:00
Marco Costalba 91427c8242 Move split() under Thread
Previous renaming patch suggested this reformat:
when a better naming leads to a better code!

No functional change.
2013-02-04 22:17:04 +01:00
Marco Costalba b8c5ea869c Some renaming in split()
Naming suggested by jundery.

No functional change.
2013-02-04 22:00:41 +01:00
Marco Costalba 6560e4cc5b Be clear about not LMR the ttMove
Currently a ttMove is reduced with ss->reduction = DEPTH_ZERO,
so it is actually not reduced (as it should be), but the
trick works just becuase it happens that ttMove is the first
to be tried and

reduction(depth, 1)

Always returns zero. So explicitly forbid reduction of ttMove
in the LMR condition. This is much clear and self-documented.

No functional change.
2013-02-03 11:14:21 +01:00
Marco Costalba 08d615cc95 Templetize score_xxx() functions
So to be style-wise aligned with the corresponding
generate() functions.

No functional change.
2013-02-03 10:11:12 +01:00
Marco Costalba 1c4e6d7ea2 Rename prevents_move() to refutes()
Better! From DiscoCheck.

No functional change.
2013-02-03 09:23:04 +01:00
Marco Costalba 5f58db8c99 Correctly score enpassant captures
Surprisingly this rare case was not considered
when scoring a capture.

Also take in account that in the promotion case
we gain a new piece (typically a queen) but we
lose the promoting pawn.

These small issues were present since Glaurung times!

Found while browsing DiscoCheck sources

bench: 5400063
2013-02-03 09:08:32 +01:00
Marco Costalba ddbe6082c4 Unify History and Gains under a single Stats class
Handling of History and Gains is almost the same, with
the exception of the update logic, so unify both
classes under a single Stats struct.

No functional change.
2013-02-02 17:45:09 +01:00
Marco Costalba 53051eefc7 Retire history.h
And move the contents to movepick.cpp, where they are
mostly used.

Idea from DiscoCheck.

No functional change (bench 5379503)
2013-02-02 16:54:35 +01:00
Gary Linscott a72710c660 Simplify eval take 2. Bench 5097444 2013-02-02 10:19:59 -05:00
Marco Costalba 9f94d22801 Restore "fail-low of reduced" and close regression
This reverts "Threat Extensions" and is the last of
this revert series.

In single-thread tests we should now be on par with 2.3.1
2013-02-02 07:16:33 +01:00
Marco Costalba 0901e12102 Revert "Simplify Evaluation"
This reverts commit 496c7497cb
2013-02-02 06:44:04 +01:00
Marco Costalba 58c9fbacc7 Revert "Extend full 3 fold detection to PvNodes" 2013-02-02 06:41:05 +01:00
Marco Costalba 483c98a69e Rewrite do_castle_move()
And handle the castle directly in do/undo_move().
This allow to greatly simplify the code.

Here the beast is the nasty Chess960 that is
really tricky to get it right because could be
that 'from' and 'to' squares are the same or
that king's 'to' square is rook's 'from' square.

Anyhow should work: verified on all Chess960
starting positions.

No functional and no speed change also in Chess960.
2013-01-28 13:40:47 +01:00
Marco Costalba 2218a5836a Rewrite do_null_move()
Use a more traditional approach, along the same lines
of do_move().

It is true that we copy more in do_null_move(), but we
save the work in undo_null_move(). Speed test shows the
new code to be even a bit faster.

No functional change.
2013-01-27 12:15:02 +01:00
Marco Costalba 76a0d3c05a Get rid of some locals in do_castle_move()
Rewrite the logic to get rid of kBefore and rBefore.

No functional change.
2013-01-27 11:03:55 +01:00
Marco Costalba 52cab06fff Don't prefetch if not needed
Prefetch access to hash tables only in case we
have changed pawn or material hash keys.

No functional change.
2013-01-27 10:19:11 +01:00
Marco Costalba bd87ab9ff5 Retire generate_king_moves()
We have only one call place so inline its content.
BTW, function is already declared as FORCE_INLINE.

Also some small refactoring while there.

No functional change.
2013-01-26 22:43:58 +01:00
Gary Linscott 57797822f8 Bring back just bishop pins 2013-01-26 15:35:00 -05:00
Marco Costalba 7062db7cb2 Clarify slavesMask usage
When a thread is allocated a bit is set in slavesMask.
This bit corresponds to the thread's index field that,
because it happens to be the position in the threads
array, eventually it is equal to the loop index 'i'.

But instead of relying on this 'coincidence', explicitly
use the 'idx' field so to clarify slavesMask usage.

Backported from c++11 branch.

No functional change.
2013-01-26 14:38:51 +01:00
Marco Costalba 166cc0292c Revert "Further push singular extension"
This reverts commit 4c91dbc28e

Seems a regression on extended test by both Gary and me.
2013-01-26 10:20:46 +01:00
Marco Costalba 496c7497cb Merge branch 'simplify_eval' of https://github.com/glinscott/Stockfish
Test results are looking good after 12500 games.

ELO: 6.55 +- 99%: 8.02 95%: 6.09
LOS: 99.99%
Wins: 1968 Losses: 1732 Draws: 8813

Also, here are the noise.py results, which seem to have stabilized:
Games: 12526 , result: [1969, 1734, 8823]
Estimated ELO: 6.94963842777
Noise as function of number of games:
['81.89', '565.26', '110.87', '104.39', '38.22', '49.98', '18.56', '16.76',
'11.02', '8.90', '17.36', '9.84', '10.81', '5.13', '6.22', '3.32', '5.83',
'7.21', '15.27', '1.63', '4.04', '9.51', '0.54', '0.75', '1.06', '2.93',
'4.59', '6.85', '13.62', '9.87', '14.74', '20.46', '22.18', '24.33', '31.02',
'34.99', '35.22', '33.22', '32.46', '37.02', '29.10', '36.34', '42.11', '39.33',
'26.16', '28.25', '35.42', '31.04', '29.26', '23.91', '22.52', '23.49', '20.00',
'24.39', '17.22', '16.50', '10.69', '9.15', '9.57', '4.77', '6.67', '3.87', '2.57',
'2.84', '2.60', '3.32', '2.08', '2.93', '4.47', '4.41', '4.83', '4.86', '6.40',
'5.98', '6.10', '6.83', '5.83', '6.22', '5.71', '8.52', '9.25', '5.98', '7.52',
'7.76', '8.76', '8.55', '8.64', '7.19', '5.83', '4.59', '4.77', '4.26', '4.98',
'5.29', '5.41', '4.92', '5.59']

bench: 5229106
2013-01-26 10:18:36 +01:00
Gary Linscott e83b9075ff Simplify evaluation 2013-01-24 08:54:13 -05:00
Marco Costalba 6950d07bf4 Small reformat of split()
No functional chhange.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-01-21 23:31:33 +01:00
Marco Costalba 054d117d25 Fix an idiotic icc warning
Intel Compiler has 'invented' this pearl:

warning #1476: field uses tail padding of a base class

Just becuase we have subclassed MainThread and added
the field 'bool thinking'.

Pure nosense. Silence the warning.

No functional change.
2013-01-20 17:36:24 +01:00
Marco Costalba 62b32a4737 Futher renaming in thread.cpp
No functional change.
2013-01-20 17:35:55 +01:00
Marco Costalba 588670e8d2 Big renaming in thread stuff
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-01-16 20:00:05 +01:00
Marco Costalba c465f4c4df Fix race while exiting
Fix again TimerThread::idle_loop() to prevent a
theoretical race with 'exit' flag in ~Thread().

Indeed in Thread d'tor we raise 'exit' and then
call notify() that is lock protected, so we
have to check again for 'exit' before going to
sleep in idle_loop().

Also same change in Thread::idle_loop() where we
now check for 'exit' before to go to sleep.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2013-01-16 19:58:55 +01:00
Lucas Braesch 8737b26a23 Remove Threat Extension
Great code simplification: - instead do not futility
prune threat refutations. allows_move() is therefore removed.

4000 games at 50,000 nodes/move:
1085-989-1926 [51.2%] LOS=98.3%

4000 games in 10"+0.1"
756-751-2493 [50.1%] LOS=55.1%

EDIT: I have retested the patch of Lucas in a slightly different form
(without pruning in PvNode) and test mre or less confirms that
60 lines of code are totally unuseful:

After 6195 games at 15"+0.05"
1333 - 1325 - 3537 ELO 0

bench 5140990
2013-01-15 17:03:56 +01:00
Marco Costalba 78a9531773 Fix a bug in timer loop
Silly logic bug introduced in dda7de17e7

Timer thread, when msec = 0, instead of going
to sleep, calls check_time() in an endless loop.

Spotted and reported by snino64 due to abnormally
high CPU usage.

No functional change.
2013-01-14 19:32:30 +01:00
Marco Costalba d1143794a0 Polymorphic Thread hierarchy
Subclass MainThread and TimerThread and declare
idle_loop() virtual. This allow us to cleanly
remove a good bunch of hacks, relying on C++
polymorphism to do the job.

No functional change.
2013-01-14 02:01:37 +01:00
Marco Costalba e70eae2c91 Don't use do_sleep flag
Rename it is_finished and use it only in main
thread to signal search is finished. This allows
us to simplify the complex SMP logic.

Ultra tricky patch: deep test is required under
wide conditions like pondering on and option
"Use Sleeping Threads" set to false.

No functional change.
2013-01-14 00:02:32 +01:00
Marco Costalba 99ae47716a Re-add the hack
This reverts commit 869c924410

I misunderstood here. Actually it can happen that
thread is created but still not entered idle_loop
and at the same time start_searching() is called.

Becuase 'do_sleep' is set start_searching() will
set it to false and start the search, but when,
at last, the thread enters idle_loop(), resets
the flag and goes to sleep: not what we want.

Revert the hack waiting for a better solution
in the next patches.

No functional change.
2013-01-13 23:56:04 +01:00
Marco Costalba dda7de17e7 Retire set_timer()
Also assure in Thread::timer_loop() that when
timer interval is 0 (timer is disabled) we
never call check_time()

No functional change.
2013-01-13 18:24:43 +01:00
Marco Costalba 869c924410 Retire obsolete race hack
This hack was introduced in d282cf6964
to workaround a race with start_searching(),
but these days is no more needed.

No functional change.
2013-01-13 17:05:30 +01:00
Marco Costalba ea6c1f7a17 Retire Threads wake_up() and sleep()
These functions are used in just one place.
And generalize wait_for_stop()

No functional change.
2013-01-13 16:57:40 +01:00
Marco Costalba 81cd7d787e Rename wake_up() to notify_one()
To align to C++ std::thread conventions.

No functional change.
2013-01-13 16:43:26 +01:00
Marco Costalba a523cea772 Unify 'ponderhit' handling
Finally we can now merge the 'ponderhit' case with
'stop' and 'quit'.

The patches have been done step by step to help debugging
becuase this is really tricky code.

No functional change.
2013-01-13 14:39:41 +01:00
Marco Costalba a66a7c3870 Small change to "ponderhit" handling
Reset Limits.ponder only if search continue, but if
we are going to stop the search there is no need
(and is also confusing) to clear the 'ponder' flag.

This mimics the behaviour upon rceiving 'stop' when
pondering.

No functional change.
2013-01-13 14:32:30 +01:00
Marco Costalba 89a89eb605 Simplify and rename wait_for_stop_or_ponderhit()
Setting stopOnPonderhit is now done by the caller.

No functional change.
2013-01-13 14:15:19 +01:00
Marco Costalba 3b14b17664 Simplify a condition in search()
And rearrange best value update in case of SpNode.

No functional change.
2013-01-13 13:17:16 +01:00
Marco Costalba 3a836dab59 Clarify SAN disambiguation in case of a pinned piece
In SAN notation when two pieces of the same type can
move to a given destination square, a disambiguation
additional info (like starting file) shall be added
to the SAN move.

If one of the two pieces is pinned, the corresponding
move _could_ be illegal and in this case disambiguation
is not needed. But to be pinned alone it is not enough
to deduce that the move is illegal, for instance in this
position:

R3rk2/2r6/8/8/8/8/8/K7 b - - 0 1

The move Rc8 is ambiguous although the rook in e8 is pinned
and the correct SAN notation should be Rcc8.

No functional change.
2013-01-12 13:21:14 +01:00
Marco Costalba e1191b35e8 Async 'stop' command
Don't wait for the search to finish after a 'stop'
command, but keep processing the GUI input if any.

Also explicitly wake up the main thread (that could be
sleeping) after a 'stop' or 'quit' command and do not
rely on wait_for_search_finished() doing it for us.

This patch cleans up the code and functions's definitions,
but it is risky and needs a good test under different
conditions to be sure it does not introduces hungs up.

No functional change.
2013-01-12 12:06:55 +01:00
Marco Costalba edce2a8448 Revert so called "fromNull patch"
Revert patch c581b7ea36

Seems a regression after testing from Gary:
ELO: 7.24 +- 99%: 17.03 95%: 12.93
LOS: 97.86%
Wins: 439 Losses: 381 Draws: 1962

And mine:
After 5410 games at 15"+0.05
Wins: 936 Losses: 1141 Draws: 3333  ELO -13

Moreover we know that there is a regression in the range
of patches which include the fromNull patch.

Probably this is not the only regression since 2.3.1 and
perhaps the idea under fromNull is good, but at the moment,
while in deep regression hunting, better to be on the safe
side and revert it entirely.

My guess on why this is a regression is that using the
negated evaluation of previous ply in case of null search
fails to take in account the king safety asymmetry between
the two colors. This is of course just a guess.

bench 5503830
2013-01-06 23:06:20 +01:00
Marco Costalba 9b1cf3cf43 Have fun with union in book.cpp
Fancy way to use an union to map polyglot
zobrist keys in one go.

Also some renaming while there.

No functional change.
2013-01-06 12:06:19 +01:00
Marco Costalba bff65a211f Retire 'Cowardice' and 'Aggressiveness' UCI options
They are not self-describing and create a lot of user
requests about them.

Given that the values are already well tuned there
is no need to expose them as UCI options.

No functional change.
2013-01-04 17:11:24 +01:00
Marco Costalba 2d60995c00 Retire 'mate in x' hack
Sometimes is faster, but not always and on very long mates
produces strange scores probably due to truncation of PV
artifacts.

So simply perform normal search also in case of UCI 'mate x'
command, with the only difference that when a mate in x is
found search returns immediately.

No functional change.
2013-01-04 16:30:46 +01:00
Marco Costalba 0454bbc54f Don't exit if unable to find bench file
Now that we can call 'bench' command also from
interactive terminal it makes no more sense to
exit the application if the user types a wrong
file name.

No functional change.
2013-01-04 14:52:21 +01:00
Marco Costalba 900e2d4e1e Teach file_to_char() about upper/lower case
This allows to further simplify Position::fen()

No functional change.
2013-01-04 14:45:04 +01:00
Marco Costalba 9d1151575d Reformat FEN construction
Simplify and shrink code.

No functional change.
2013-01-04 13:38:14 +01:00
Marco Costalba 193741218c Remove some obsolete asserts on TT values
Now that insert_pv_in_tt() stores VALUE_NONE in
TT's position evaluation those 2 asserts are
obselete.

No functional change.
2012-12-31 17:20:39 +01:00
Marco Costalba 896420b166 Allow to pass a 'seed' to RKISS
This somewhat simplifies the code.

Suggested by Lucas Braesch.

No functional change.
2012-12-31 11:59:53 +01:00
Marco Costalba 009a0f88e0 Micro-optimization in evaluate_space()
Since &-ing with SpaceMask restricts the set to the home
half of the board, it is possible to use just one popcount
instead of 2 by shifting "safe" to the other half of the
board. This gives a small speedup especially on systems
where hardware popcount is not available.

Patch kindly sent by Richard Vida.

No functional change.
2012-12-31 11:34:18 +01:00
Marco Costalba e1d681458e Add 'mate' limit to 'bench' command
It is now possible to run SF on a 'mate in x' testsuite.

For instance in case of a file with fen strings of
positions with mate in 10 we can now 'bench' on it:

stockfish bench 128 1 10 mate_in_10.epd mate

No functional change.
2012-12-30 15:49:02 +01:00
Marco Costalba ce063f59cd Handle UCI command "mate in x moves"
Following a user request I added the handling of UCI:

go mate x

Currently we just return from a PV node if x moves have been
done. Probably not the best approach. I have looked at Fruit/Toga
sources and there is even simpler: engine falls back on a fixed
depth search.

No functional change.
2012-12-30 14:43:23 +01:00
Marco Costalba 3cf6471738 Revert evaluation cache
And return on using TT as backing store for position
evaluations.

Tests (even on single thread) show eval cache was a regression.
In multi thread result should be even worst because eval cache
is a per-thread struct, while TT is shared.

After 4957 games at 15"+0.05 (single thread)
eval cache vs master 969 - 1093 - 2895  -9 ELO

So previous reported result of +18 ELO was probably due to an
issue in the testing framework (a bug in cutechess-cli) that
has been fixed in the meanwhile.

bench: 5386711
2012-12-27 13:57:17 +01:00
Marco Costalba f78b68b7ff Add list of legal moves to Position::pretty()
Along the same lines of previous patch now we add
the list of the legal moves in the given position.

No functional change.
2012-12-27 11:34:48 +01:00
Marco Costalba e9ab7353de Add checkers info to Position::pretty()
In case current position is under check, list the
squares of the checker(s) pieces.

This should satisfy a specific user request.

No functional change.
2012-12-26 18:28:45 +01:00
Marco Costalba 9d1978e217 Remove two obsolete asserts in prevents_move
Now that this function is called also to calculate
move's extensions the asserts are no more valid.

No functional change.
2012-12-26 12:27:41 +01:00
Marco Costalba db097921bc Rename yields_to_threat and prevents_threat
Follow Lucas suggestions and better name these
two functions.

No functional change.
2012-12-26 12:21:59 +01:00
Marco Costalba 894c43a1d6 Introduce Null Threat extension
In case of null search at low depths returns a fail low
due to a threat then, rather than return beta-1 (to cause
a re-search at full depth in the parent node), we set a flag
threatExtension = true (false by default) that will cause
moves that prevent the threat to be extended of one ply
in the following search.

Idea and patch is by Lucas Braesch.

Lucas also did the tests:
1500 games in 5"+0.05":
SF_threatExtension vs SF_20121222: 366 - 331 - 803 [51.2%] LOS=90.8%

3000 games in 10"+0.1":
SF_threatExtension vs SF_20121222: 610 - 559 - 1831 [50.8%] LOS=93.2%

Tests confirmed by Gary after 10570 games,
ELO: 2.79 +- 99%: 8.72 95%: 6.63
LOS: 94.08%
Wins: 1523 Losses: 1438 Draws: 7607

And finally by me at 15"+0.05, single thread, 3824 games
threatExtension vs master 768 - 692 - 2364  +7 ELO

bench 4918443
2012-12-25 19:17:27 +01:00
Marco Costalba b5b799b5ab Fix a couple of extra spaces
No functional change.
2012-12-25 18:48:41 +01:00
Marco Costalba e82382703c Retire Position::in_check()
It is redundant with Position::checkers()

No functional change.
2012-12-25 17:59:35 +01:00
Marco Costalba 3b49aeb4f2 Retire Position::move_is_legal()
Use the new contains() method of struct MoveList

No functional change.
2012-12-25 11:51:08 +01:00
Marco Costalba 423c6d8a8a Small tweak in is_pseudo_legal()
This is difficult code becuase a bug here could lead
to very subtle crashes in case of SMP games where we
have TT move corruption due to concurrent access.

Anyhow I have fully verified te code throwing at it
random moves. It shoudl work.

No functional change.
2012-12-25 11:31:32 +01:00
Marco Costalba 158014b39d Introduce namespace Pawns
And retire old struct PawnTable along the same lines
of previous patch.

No functional change.
2012-12-22 11:38:36 +01:00
Marco Costalba 231f62baf7 Introduce namespace Material
And retire old struct MaterialTable simplifying the code.

No functional change.
2012-12-16 12:58:39 +01:00
Marco Costalba 52bbf372bb Don't need to check for bestValue < beta to split
With rearrangement of fail high code this condition
is no more necessary.

Found by Jörg Oster.

No fuctional change.
2012-12-15 13:11:10 +01:00
Marco Costalba 3ddf91d9d1 Remove an extra semicolon
No functional change.
2012-12-15 11:20:04 +01:00
Marco Costalba a2f46446cf Revert store of distinct upper and lower bounds
Test by Joona prooves the new feature don't value 70 added lines of code.

Grand totals after 10040 games (crashes: 0) for tt_both

master_9edc7 - 6a93488_6a934: 1756 - 1688 - 6596 ELO +2 (+- 2.7)

Confirmed by test of Gary:

After 8680 games:
ELO: 0.80 +- 99%: 9.62 95%: 7.31
LOS: 65.38%
Wins: 1288 Losses: 1268 Draws: 6130

Thanks a lot to both for testing it !!!

bench 5149248
2012-12-15 11:18:52 +01:00
Marco Costalba 9edc7d6958 Merge branch 'eval_cache'
Unusually good result. Defenitly needs further verifications.

After 2160 games at 15"+0.05
Mod vs Orig 486 - 367 - 1307 ELO +19

bench: 6261882
2012-12-10 09:26:02 +01:00
Marco Costalba 55db871472 Fix comparison with alpha, not beta
This silly bug seems the reason of the unsual bench
value.

bench: 6261882
2012-12-09 14:19:21 +01:00
Marco Costalba da98a45bcb Ensure valueLower <= valueUpper
In case a TTEntry stores both an upper and a lower bound
ensure that upper bound is not smaller than lower bound.

bench 1813815
2012-12-09 14:14:44 +01:00
Marco Costalba feeafb0a50 Store distinct upper and lower bound scores
This is more complex than what I'd like but I
was unable to split in small chunks.

Here we add 2 slots to TTEntry (valueUpper and depthUpper)
so that sizeof(TTEntry) returns to the original 16 bytes
and we can pack exactly 4 entries in a 64 bytes cache line.

Now we save an upper bound score alongside a lower (exact)
score. The idea is to increase TT cut-offs rates becuase
there is now an higher probability for a node to use TT info.

This patch is highly experimental and probably needs further
steps as is hinted by an unrealistic bench number:

bench: 2022385
2012-12-09 13:15:50 +01:00
Marco Costalba 22c557ca7c Micro-optimize color_of()
In almost all cases we already know in advance that
color_of() argument is different from NO_PIECE.

So avoid the check for NO_PIECE in color_of() and
test at caller site in the very few places where
this case could occur.

As a nice side effect, this patch fixes a (bogus)
warning under some versions of gcc.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-12-08 12:32:48 +01:00
thaspel 6a93488291 Update Readme.txt now that we support 64 threads
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-12-04 19:57:16 +01:00
Marco Costalba 6fa83f5188 Merge branch 'eval_cache'
Use an eval cache instead of TT to store node
position evaluations.

It is already an improvment and, because it frees
two TT entry slots, paves the way to extend TT to
store both upper and lower bounds.

After 4855 games, single thread, 15"+0.05
Mod vs Orig 1165 -920 - 2770 ELO +18

bench: 5149248
2012-12-04 08:05:15 +01:00
Marco Costalba ce248e7920 Increase MAX_THREADS to 64
And document why this is an hard limit. It
seems for some (lucky) people 32 threads
are not enough.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-12-04 07:57:15 +01:00
Marco Costalba 23bdd06442 Reintroduce eval optimizaion from null search
Now that conversion to eval cache is finished
we can reintroduce this optimization.

bench: 5149248
2012-12-01 18:11:01 +01:00
Marco Costalba 98cd8239cc Don't save eval score in TT
This patch completes the removal of eval info
in TT table.

No functional change.
2012-12-01 15:19:50 +01:00
Marco Costalba 2a585b63b8 Don't use TT just to save a node evaluation
In search(), after we evalute the position, in case there
isn't any TT entry we create one with just the evaluation
score.

This patches removes that code. The reason becuase the patch
deserves a single commit it is becuase introduces a (very small)
functional change due to the fact that the total number of
TT stores is less now and this slightly alters the TT hits
of our benchmark.

bench: 4983262
2012-12-01 15:13:08 +01:00
Marco Costalba a5ea3a202e Don't read eval from TT anymore
Rely fully on eval cache. Note that we still save eval
info to TT, this is not needed at this moment and will be
removed in future patches. We keep it so to have a "non
functional change" patch.

No functional change.
2012-12-01 15:01:00 +01:00
Marco Costalba 69de670353 Retire eval optimization from null search
Simplify things preparing for further changes.

bench: 4983282
2012-12-01 14:48:13 +01:00
Marco Costalba 4e5d834e8e Add eval cache infrastructure
With this patch series we want to introduce a per-thread
evaluation cache to store node evaluation and do not
rely anymore on the TT table for this.

This patch just introduces the infrastructure.

No functional change.
2012-12-01 14:01:20 +01:00
Marco Costalba 4502917665 Don't double check for move legality
In case of a RootNode or a SpNode move has
been already checked for legality so we can
skip a redundant check.

Spotted by Frank Genot.

No functional change.
2012-11-27 20:05:03 +01:00
Marco Costalba 5af8179647 Update bestValue when futility pruning
In qsearch we should update the bestValue as we do
in case of futilityValue < beta, also when pruning
moves with non-positive see.

Spotted by Lucas Braesch

Bench: 5695710
2012-11-26 16:13:36 +01:00
Marco Costalba 55df3fa2d7 Avoid spamming the GUI in multipv search
Send the PV lines to GUI only once at the end of the
PV search loop or just in case of long searches.

We need to sync also sending of "currmove" info to
avoid sending info on current move without first
informing the GUI on the PV line we are searching on.

No functional change.
2012-11-18 11:46:37 +01:00
Marco Costalba 8367cf15da Triviality in move_gives_check()
It seems even a bit faster, although handling of the special
cases is not the hot path.

No functional change.
2012-11-17 13:08:04 +01:00
Marco Costalba 942989939a Better document fail-high condition
At this point we have already verified (value > alpha)
and this implies, in case of a non-PV node, where search
window size is zero, that value >= beta.

This is not so self-evident, so document the code with
an assert condition.

No functional change.
2012-11-17 12:50:10 +01:00
Marco Costalba 239d7b3fd1 Don't access std::cout from Position::print()
Let the caller to decide where to redirect (cout or cerr) the
ASCII representation of the position. Rename the function to
reflect this.

Renamed also from_fen() and to_fen() to set() and fen() respectively.

No functional change.
2012-11-11 13:30:20 +01:00
Marco Costalba 116234d6c6 Restore old BOUND_EXACT logic in qsearch
In case a PvNode node has a static evaluation above alpha but
no available moves we want to flag the node as BOUND_EXACT,
not as BOUND_UPPER as is currently.

The behaviour was recently introduced with patch d471c49700
of 3/10/2012

Spotted by Hongzhi Cheng.

bench: 5558464
2012-11-11 11:52:11 +01:00
Marco Costalba c45a4e0b48 Revert "Don't prune discovered checks"
Both Lucas re-test and Jean-Francois confirrm it
is a regression.

Here Jean-Francois's results after 3600 games :
Score of 96d3b1c92b vs 3b87314: 690 - 729 - 2181 [0.495] 3600
ELO: -3.86 +- 99%: 14.94 95%: 11.35
LOS: 15.03%
Wins: 690 Losses: 729 Draws: 2181 Total: 3600

Bench: 5404066
2012-11-11 11:20:39 +01:00
Marco Costalba 96d3b1c92b Don't prune discovered checks
Don't prune and eventually extend check moves of type
DISCO_CHECK (pun intended, Lucas will understand :-) ).

Patch from Lucas Braesch that has also tested it:

Result: 879-661-2137, score=52.96%, LOS=100.00% (time 10"+0.1")

I have started a verification test right now.

bench: 6004966
2012-11-07 18:29:56 +01:00
Marco Costalba 3b87314331 Previous patch test results
From Jean-Francois's

Final result after 5000 games :
Score of c581b7e vs a878312: 1163 - 970 - 2867 [0.519] 5000
ELO: 13.35 +- 99%: 12.71 95%: 9.65
LOS: 100.00%
Wins: 1163 Losses: 970 Draws: 2867 Total: 5000

From me

After 3266 games at 20"+0,05
Score of c581b7e vs a878312: 612 - 607 - 2047

So no regression at longer TC and perhaps a little gain at
fast TC.
2012-11-07 18:07:03 +01:00
Marco Costalba c581b7ea36 Another attempt at evaluation shortcut
In this case we try a rather drastic
approach: we simply don't futility prune
in qsearch when arriving from a null move.

So we save evaluating and also save to mess
with eval margins at all because margin is used
only in futility.

Also accuracy should not be affected, actually it
improves because we don't prune anything anymore.

bench: 5404066
2012-11-05 16:12:28 +01:00
Marco Costalba a5b1f4774f Temporary revert previous patch
Performs well at very short TC of 40/4+0.05 (courtesy of Jean-Francois):
Wins: 2503 Losses: 2146 Draws: 5581 Total: 10230 +12 ELO

But is poor at longer TC of 20"+0.05
Wins: 321 Losses: 373 Draws: 1141 Total: 1808 -10 ELO

The patch was clearly a tradoff between speed and accuracy and
the most interesting part of it are test results that can be
commented as follows:

- A short TC is very sensible to any speed increase
- A longer TC is more sensible to accuracy and less to speed

So a patch that does not change speed is suitable to be tested at
short TC, while a speed/accuracy compromise patch is IMO better to
be tested at longer TC to verify loss of accuracy can be tolerated.

In this case the revert is only temporary. We will come back again
once we will be able to preserve the evaluation margin.

bench: 5809010
2012-11-05 07:41:28 +01:00
Marco Costalba 37e9802411 Skip evaluate() call after a null move
Reuse the evaluation of the parent with inverted sign and
set margin to zero (this is an hack!).

This is done only in qsearch where almost 15% of calls are
from a null move. In normal search the number of nodes where

(ss-1)->currentMove == MOVE_NULL

is almost zero and so there is no need of using this trick.

The big advantage of this patch is a speed-up due to skipped
evaluate() calls, that are very costly.

Functionality is of course affected and we will need to proper
test it later. For now we just register a 3-4% speed up.

Suggested by Hongzhi Cheng.

bench: 5051328
2012-11-04 10:48:14 +01:00
Marco Costalba a87831230d Correctly fix "break from split" patch
In case split cut-offs we return with still
some moves to go but we really want to break
out from the loop now.

No functional change.
2012-11-03 17:09:18 +01:00
Marco Costalba dd5b3086f5 Relax constrain in prevents_threat()
When testing if a move blocks the threat path there is no
reason to require the threat to be a slider. Indeed threat
can be a double pawn push like in this example:

r1bq1rk1/ppp1np1p/4n1p1/3p4/3P2Q1/2P1B3/PPBN2PP/R4RK1 w - - 0 16

Where white's move Rf6 blocks the threat f5.

As a nice side effect we can retire the now useless helper
piece_is_slider().

This patch kicks in only very rare cases, indeed the bench is
still the same!

bench: 5809010
2012-11-03 15:57:20 +01:00
Marco Costalba 47f988f05f Sync connected_threat() and yields_to_threat()
Rename stuff so to sync as much as posisble the
two related functions.

No functional change.
2012-11-03 15:36:12 +01:00
Marco Costalba 07989712af Don't 'break' upon returning from split()
There is no guarantee that split() consumes all the node's
moves. Indeed split() can return without performing any job
for instance because MAX_SPLITPOINTS_PER_THREAD is reached
or becuase no available threads are found (this latter case
is much more common).

So search must continue in those cases and we cannot force
exiting from move's loop.

Bug introduced by 1ac417edb8 of 5/10/2012

Spotted by Frank Genot.

No functional change.
2012-11-03 14:54:38 +01:00
Marco Costalba d0d69a5358 Remove a redundant condition in connected_moves()
If a previous move attacks the king (with the piece
of the threat move removed) then must be a discovered
check, otherwise it means that first move gave check
and we were not able to do a null move.

Also renamed stuff to better document the function's
context.

No functional change.
2012-11-03 14:24:29 +01:00
Marco Costalba 972de506a0 Relax constrain in connected_moves()
When testing if a piece is moving through the squares
vacated by a previous move there is no reason to require
the piece to be a slider, indeed we can have a double
pawn push like in this example:

r1q2rk1/2p1bppp/2Pp4/pN5b/Q1P1p3/4B2P/PP1R1PP1/1K5R w - - 3 18

Where black's move f5 is connected to previous move Be7 that
frees the path.

Or we can have a castle move:

r1bqkb1r/pppp1ppp/2n1pn2/1B6/4P3/2N2N2/PPPP1PPP/R1BQK2R b KQkq - 5 1

Where a previous move Bb5 allows the white to castle king side.

This time patch is mine ;-)

new bench: 5809010
2012-11-03 13:34:04 +01:00
Marco Costalba ed1574e46c Reformat connected_moves()
Prepare code for the next patch that
will affect functionality.

No functional change.
2012-11-03 13:27:44 +01:00
Marco Costalba 52f55179a8 Fix an off-by-one bug in multi pv print
We send to GUI multi-pv info after each cycle,
not just once at the end of the PV loop. This is
because at high depths a single root search can
be very slow and we want to update the gui as
soon as we have a new PV score.

Idea is good but implementation is broken because
sort() takes as arguments a pointer to the first
element and one past the last element.

So fix the bug and rename sort arguments to better
reflect their meaning.

Another hit by Hongzhi Cheng.  Impressive!

No functional change.
2012-11-03 00:36:40 +01:00
Marco Costalba bbdf9e4737 Fix a condition in connected_moves()
When checking if the moving piece p1 in a previous
move m1 defends the destination square of a move m2
we have to use the occupancy with the from square of
m2 removed so to take in account the case in which
f2 will block an x-ray attack from p1.

For instance in this position:
r2k3r/p1pp1pb1/qn3np1/1N2P3/1p3P2/2B5/PPP3QP/R3K2R b KQ - 1 9

The move eXf6 is connected to the previous move Bc3 that
defends the destination square f6.

With this patch we have about 10% more moves detected as
'connected'. Anyhow the absolute number is very low, about
4000 more moves out of 6M nodes searched.

Another issue spotted by Hongzhi "Hawk Eye" Cheng ;-)

new bench: 5757373
2012-11-02 17:18:38 +01:00
Marco Costalba 94ecdef8ac Micro-optimize pop_lsb() for 64bit case
On Intel, perhaps due to 'lea' instruction this way of
zeroing the lsb of *b seems faster than a shift+negate.

On perft (where any speed difference is magnified) I
got a 6% speed up on my Intel i5 64bit.

Suggested by Hongzhi Cheng.

No functional change.
2012-11-02 12:11:49 +01:00
Marco Costalba e3b0327812 Fix a warning under MSVC
Compiler complies that 'cnt' is initialized but
unused (in !CheckThreeFold case). Moving the
definition of 'cnt'out of the loop  seems to do
the trick.

No functional change.
2012-11-02 11:43:23 +01:00
Marco Costalba c039103b31 Pass InCheck as template parameter of qsearch()
Instead of use a variable so to resolve many conditions
already at compile time. In quiesce is also where we
have most of the InCheck nodes and is one of the most
performance critical code paths.

Speed up of 1.5% with Clang and 1% with gcc

Suggested by Hongzhi Cheng.

No functional change.
2012-11-01 18:45:38 +01:00
Marco Costalba fe1cbe2638 Use correct occupancy in connected_threat()
When checking if a move defends the threatened piece
we correctly remove from the occupancy bitboard the
moved piece. This patch removes from the occupancy also
the threatening piece so to consider the cases of moves
that defend the threatened piece x-raying through the
threat move.

As example in this position:
r3k2r/p1ppqp2/Bn4p1/3p1n2/4P1N1/5Q1P/PPP2P1P/R3K2R w KQkq - 1 10

The threat black move is dxe4. With this patch we include
(and so don't prune) white's Bb7 that would be pruned otherwise.

The number of affected position is very low, around 1% of cases,
so we don't expect ELO changes, neverthless this is the logical
and natural thing to do.

Patch suggested by Hongzhicheng.

new bench: 5323798
2012-10-30 20:27:07 +01:00
Marco Costalba 4e31c39a64 Retire move_attacks_square()
There is only one call site. This patch is a
preparation for the next one that will affect
functionality.

No functional change.
2012-10-30 20:03:35 +01:00
Marco Costalba 13f90a30ef Get rid of ReducedStateInfo struct
ReducedStateInfo is a redundant struct that is also
prone to errors, indeed must be updated any time is
updated StateInfo. It is a trick to partial copy a
StateInfo object in do_move().

This patch takes advantage of builtin macro offsetof()
to directly calculate the number of quad words to copy.
Note that we still use memcpy to do the actual job of
copying the (48 bytes) of data.

Idea by Richard Vida.

No functional and no performance change.
2012-10-29 08:07:17 +01:00
Marco Costalba ad1941fd00 Creative formatting in uci.cpp
Have some fun breaking the indentation rules :-)

No functional change.
2012-10-28 11:28:14 +01:00
Gary Linscott cee6336515 Detect drawish KQKP endings
Based off of the rules from the wikipedia page,
here: http://en.wikipedia.org/wiki/Queen_versus_pawn_endgame.

bench does not change: 5312693 but patch is real, has been
tested on specific positions.
2012-10-28 10:17:57 +01:00
Marco Costalba b3d030294b Reformat check_is_dangerous()
And shuffle some code at search.cpp tail.

No functional change.
2012-10-27 15:07:20 +02:00
Marco Costalba cd80762c13 Use std::stack instead of fixed size array
Only in not performance critical code like pretty_pv(),
otherwise continue to use the good old C-style arrays
like in extract/insert PV where I have done some code
refactoring anyhow.

No functional change.
2012-10-27 13:31:23 +02:00
Marco Costalba 00a853950f Fix broken uci notation for promotions
Silly typo (introduced in e304db9d1e) completely
messed up move notation in case of promotions causing
"Illegal move" warning in cutechess-cli.

Reported by Jörg Oster.

No functional change.
2012-10-26 16:06:47 +02:00
Marco Costalba 4c7a71a44b Fix asserts due to TT access races
In multi-threads runs with debug on we experience some
asserts due to the fact that TT access is intrinsecally
racy and its contents cannot be always trusted so must
be validated before to be used and this is what the
patch does.

No functional case.
2012-10-26 12:41:12 +02:00
Marco Costalba c594b989c0 Extend full 3 fold detection to PvNodes
And restore old behaviour of not returning from a RootNode
without updating RootMoves[].

Also renamed is_draw() template parameters to reflect a
'positive' logic (Check instead of Skip) that is easier
to follow.

New bench: 5312693
2012-10-26 11:56:33 +02:00
Marco Costalba 71f37ac1aa Merge pull request #34 from jromang/repetition
Improve 3 fold repetition detection
2012-10-26 02:01:10 -07:00
Jean-Francois Romang 77c91ac1ba Full three fold repetition detection only at root node 2012-10-25 15:57:44 +02:00
Jean-Francois Romang 5436d98fc5 Enable true 3 fold detection in search 2012-10-25 06:54:05 +00:00
Jean-Francois Romang 0587c5b605 Allow full repetition detection
Based on sshivaji 6ee19aa5389ce60181907ba53bbb50642f2d5657 commit
2012-10-25 06:28:55 +00:00
Marco Costalba c1f4000426 Fix an assert when we stop the search
When signal 'stop' is raised we return bestValue
that could be still set at -VALUE_INFINITE and
this triggers an assert. Fix it by returning
a value we know for sure is not +-VALUE_INFINITE.

Reported by 平岡拓也 Hiraoka.

No functional change.
2012-10-25 00:07:59 +02:00
Marco Costalba 22715259a0 Rename RootPosition and shuffle think()
Just slightly code reshuffles, noting interesting here...

No functional change.
2012-10-24 15:01:39 +02:00
Marco Costalba 4eda653a56 Drop Chess960 and UCIMultiPV globals and rename MultiPV
No functional change.
2012-10-24 14:31:33 +02:00
Marco Costalba e8e5b9f537 Wrap in a class Skill Level code
Note that the actual pickup is done in the class
d'tor so to be sure it is always triggered, even
in case of a sudden exit due to a 'stop' signal.

No functional change.
2012-10-24 12:52:29 +02:00
Marco Costalba 9c2b3faec4 Shuffle aspiration window loop
No functional change.
2012-10-24 11:35:53 +02:00
Marco Costalba 70b1b79264 Retire refine_eval()
Inline its content and better comment what it does through
some renaming.

No functional change.
2012-10-22 10:03:53 +02:00
Marco Costalba 5fc8b27db9 Don't copy a full Position object in print()
Function move_to_san() requires the Position to be
passed by referenced because a do/undo move is done
inside the function to detect a possible mate and to
add to the san string the corresponding '#' suffix.

Instead of passing a copy of current position pass
directly the original position object after const
casting it. This has the advantage to avoid a costly
Position copy, on the down side a bench test could
report different searched nodes if print(move) is
used, due to the additionals do_move() calls.

No functional change.
2012-10-22 00:55:16 +02:00
Marco Costalba dbbbd3880c Don't need to init board[] with NO_PIECE
Now that NO_PIECE == 0 the common memset() will
do the work.

No functional change.
2012-10-22 00:38:12 +02:00
Marco Costalba e40b06a050 Change NO_PIECE value and shrink PieceValue[]
This requires changing color_of() definition.

No functional change.
2012-10-21 11:50:56 +02:00
Marco Costalba e304db9d1e Use self-describing constants instead of numbers
And remove now useless comments.

No functional change.
2012-10-21 11:16:21 +02:00
Marco Costalba 6b909b2343 Move RootColor from Eval to Search
No functional change.
2012-10-21 09:12:02 +02:00
Marco Costalba 55bd27b8f0 Contempt factor: use DrawValue also in case of stealmates
Spotted by Jörg Oster.

No functional change (when contempt factor is not used).
2012-10-20 11:02:37 +02:00
Marco Costalba 1018474853 Fix compatibility with old Windows 95 and 98
Report and patches by bnemias.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-10-20 10:48:26 +02:00
Marco Costalba c11edd3b0b Merge pull request #33 from daylen/master
Further improve OS X compatibility

Change the minimum supported version from 10.6 to 10.0

No functional change.
2012-10-20 01:07:27 -07:00
Daylen Yang f0298862f5 Further improve compatibility when compiling on OS X
Change the minimum supported version from 10.6 to 10.0
2012-10-16 00:13:16 -07:00
Marco Costalba 739d23f2a3 Fix Makefile for PowerPC with prefetch enabled
Existing Makefile is buggy for PowerPC, it has no
SSE, yet it is given it if Prefetch is enabled,
because it isn't ARMv7.

Patch from Matthew Brades.

No functional change.
2012-10-15 01:13:41 +02:00
Marco Costalba 3aa2d6db18 Document why is safe ttValue == VALUE_NONE
We can have ttValue == VALUE_NONE when we use a TT
slot to just save a position static evaluation, but
in this case we also save DEPTH_NONE so to avoid
using the ttValue in search. This happens to work,
but due to a number of lucky and tricky cases that
we now documnet through a bunch of asserts and a
little change to value_from_tt() that has no real
effect but clarifing the code.

No functional change.
2012-10-14 12:47:16 +02:00
Marco Costalba 6a75291ab1 Set TT size to 32 MB during 'bench'
On some platforms 128 MB of RAM for TT is too much,
so run 'bench' with the default 32 MB size.

No functional change although of course now 'bench'
reports a different number: 5545018
2012-10-14 09:02:17 +02:00
Marco Costalba eb1a4f11fa Move all Contempt Factor code to search.cpp
Where it is used.

No functional change.
2012-10-13 14:49:01 +02:00
Marco Costalba d023be5a21 Retire BitCount8Bit[] table
Use popcount() instead in the only calling place.
It is used only at initialization so there is no
speed regression and anyhow even initialization
itself is not slowed down: magic bitboard setup
stays around 175 msec on my slow 32bit Core Duo.

No functional change.
2012-10-13 14:24:01 +02:00
Marco Costalba c2cd75843e Use new 64 bit De Bruijn BitScan
Allows to sync 32 and 64 bits implementations.

Idea by Kim Walisch, reported by Gerd Isenberg:
http://talkchess.com/forum/viewtopic.php?t=45554

No functional change.
2012-10-13 13:45:45 +02:00
Marco Costalba 189b6fc270 Retire can_return_tt() and rewirte TT-hit code
Simplify the code and doing this introduce a couple
of (very small) functional changes:

- Always compare to depth even in "mate value" condition
- TT cut-off in qsearch also in case of PvNode, as in search

Verified against regression with 2500 games at 30"+0.05
on 2 threads: 451 - 444 - 1602

Functional changed: new bench is 5544977
2012-10-13 11:45:14 +02:00
Marco Costalba 6c8663341e Scale contempt factor to zero at endgame
Contempt Factor is more effective at opening/middle game
and seems harmful at endgame phase. See:

http://chessprogramming.wikispaces.com/Contempt+Factor

http://web.archive.org/web/20070707023203/www.brucemo.com/compchess/programming/contempt.htm

Therefore we scale down the contempt factor while going
on with the game so to reach zero at endgame phase.

No functional change.
2012-10-13 10:40:38 +02:00
Marco Costalba aba152ea3a Fix a minor bug in search
As Joona says: "The problem is that when doing full
window search (-VALUE_INFINITE, VALUE_INFINITE), and
pruning all the moves will return fail low which is
mate score, because only clause touching alpha is
"mate distance pruning". So we are returning mate score
although we are just pruning all the moves. In reality
there probably is no mate in sight.

Bug spotted and fixed by Joona.
2012-10-11 21:12:43 +02:00
Jean-Francois Romang 7f9ebf8e86 ARM lsb/msb assembly
Implement lsb/msb using armv7 assembly instructions.
msb is the easiest one, using a gcc intrinsic that generates
code using the ARM's clz instruction. lsb is also using this
clz instruction, but with the help of ARM's 'rbit' (bit
reversing) instruction. This leads to a >2% speed gain.

I also renamed 'arm-32' to the more meaningfull 'armv7' in the Makefile

No functional change.
2012-10-11 21:01:52 +02:00
Jean-Francois Romang 4e7da9be3d Introduce arm-32 ARCH in Makefile
No functional change.
2012-10-11 07:34:48 +02:00
Marco Costalba 46c01b5083 Retire is_dangerous() and inline its content
No functional change.
2012-10-10 08:13:36 +02:00
Marco Costalba dae843d4d6 Rearrange prefetch code
No functional change.
2012-10-08 11:43:47 +02:00
Marco Costalba 78fe0cfb8d Merge pull request #29 from RyanTaker/patch-3
Add Contempt Factor to Polyglot.ini
2012-10-07 15:36:08 -07:00
RyanTaker d77d555c72 Add Contempt Factor in Polyglot
The contempt factor was previously not included in polyglot.ini
2012-10-07 09:49:55 -07:00
Marco Costalba bbaef048cd Sync qsearch with search
Port to qsearch() the same changes we recently
added to search().

Overall this search refactoring series shows
almost 2% speed up on gcc compile.

No functional change.
2012-10-07 13:15:41 +02:00
Daylen Yang 7e2d49368d Improve compatibility with older versions of Mac OS X
Use the -mmacosx-version-min flag to support older versions of Mac OS X
(like version 10.6 and 10.7) when compiled on a machine running version
10.8.
2012-10-06 17:56:12 -07:00
Marco Costalba 954fc950d9 Fix POPCNT support on mingw 64
When using asm 'popcnt' instruction the given
operand registers must be of the same type.

No functional change.
2012-10-06 13:01:44 +02:00
Marco Costalba d777c4d789 Fix mingw compile with ARCH=x86-64
When using the Makefile (as for the mingw case),
IS_64BIT and USE_BSFQ are already set with
ARCH=x86-64 and do not need to be redefined.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-10-06 12:34:09 +02:00
Marco Costalba cedbd3332a Fix Contempt Factor implementation
First disable Contempt Factor during analysis, then
calculate the modified draw score from the point of
view of the player, so from the point of view of
RootPosition color.

Thanks to Ryan Taker for suggesting the fixes.

No functional change.
2012-10-06 10:12:34 +02:00
Marco Costalba ff9ca3e76e Fix fallouts from previous patch
These kind of arch specific code is really nasty
to make it right becuase you need to verify on
all the platforms.

Now should compile properly also on ARM

Reported by Jean-Francois.

No functional change.
2012-10-06 09:09:27 +02:00
Jean-Francois Romang b8948e84b8 Allow prefetching on non-x86 platforms with gcc
In particular on ARM processors. Original patch by
Jean-Francois, sligtly modified by me to preserve
the meaning of NO_PREFETCH flag.

Verified with gcc, clang and icc that prefetch instruction
is correctly created.

No functional change.
2012-10-06 00:43:16 +02:00
Marco Costalba 1ac417edb8 Retire futility_move_count()
And remove (bestValue < beta) condition from
moves loop.

No functional change.
2012-10-05 18:24:38 +02:00
Marco Costalba d471c49700 Rewrite search best value update
A simplification and also a small speed-up of
about 1% mainly due to reducing calls to
thisThread->cutoff_occurred().

Worst case split point recovering time after a
cut-off occurred is limited to 3 msec on my slow
PC, and usually is below 1 msec, so it seems safe
to remove the cutoff_occurred() check.

No functional change.
2012-10-05 13:53:05 +02:00
Marco Costalba c9f9262a49 Add experimental contempt factor
This is very crude and very basic: simply in case
of a draw for repetition or 50 moves rule return
a negative score instead of zero according to the
contempt factor (in centipawns). If contempt is
positive engine will try to avoid draws (to use
with weaker opponents), if negative engine will
try to draw. If zero (default) there are no changes.

No functional change.
2012-10-05 08:28:23 +02:00
Marco Costalba bd7a0d4ce4 Retire EasyMoveMargin
Use a value related to PawnValue instead.

This is a different patch from previous one because
could affect game play and skill levels, although
in a mostly unmeasurable way. Indeed thresold has
been raised so easy move is a bit harder to trigger
and skill level is a bit more prone to blunders.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-10-03 18:59:24 +02:00
Marco Costalba 561eb34aea Don't hide thresolds values
Show the real value in the code, not hide it
behind a variable name, especially when there
is only once occurence.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-10-03 18:56:36 +02:00
Marco Costalba 4c91dbc28e Further push singular extension
Extend for an extra half-ply in case the node is (probably)
going to fail high. In this case the added overhead is limited.

A novelity is the way this patch has been tested: Always in
self-play but with a much longer TC to allow the singular
extension to fully kick in and also (my impression) to have
less noisy results.

Ater 1015 games on my QUAD at 60"+0.05
Mod vs Orig 173 - 150 - 692 ELO +8
2012-10-02 06:39:27 +02:00
Marco Costalba 1b6b711c44 Further rearrange search()
No functional change.
2012-10-01 10:44:04 +02:00
Marco Costalba faaa1f1116 Don't push on the stack 200KB in one go
This could be a limit on some platforms (as it seems
to be in Native Client). Patch from a SF fork on github:

https://github.com/ccherng/Stockfish/commit/47374afd6fdfabd9de183a7a67d645daad45fb21

No functional change.
2012-09-30 11:35:14 +02:00
Marco Costalba ed0fb0b05f Add support for node limited search
Handle also the SMP case. This has been quite tricky, not
trivial to enforce the node limit in SMP case becuase
with "helpful master" concept we can have recursive split
points and we cannot lock them all at once so there is the
risk of counting the same nodes more than once.

Anyhow this patch should be race free and counted nodes are
correct.

No functional change.
2012-09-30 10:19:22 +02:00
Marco Costalba e5463eb3ae Skip some useless initializations in search()
And rearrange a bit the initialization code. Still
some polishing to do, but it is a first step.

No functional change.
2012-09-29 23:12:39 +02:00
Marco Costalba d53c928261 Don't need to early check PV moves for legality
As long as isPvMove (renamed to pvMove) is set after
legality check, we can postpone legality even in PV case.

Patch aligns the PV case with the common non-pv one.

No functional change.
2012-09-29 18:05:49 +02:00
Marco Costalba bc8f5fe0bf Drop a magic in book.cpp
Mask out move's spacial flags without relying
on internal Move representation.

No functional change.
2012-09-22 11:19:10 +02:00
Marco Costalba 9204a60dbb Tweaks to bitcount functions
Seems even a bit faster now (almost 1% in 32bit case).

No functional change.
2012-09-22 10:59:33 +02:00
Marco Costalba 1cb2722c95 Restore development version 2012-09-22 00:37:18 +02:00
Marco Costalba 3caeabf73b Stockfish 2.3.1
Stockfish bench signature is: 5423738
2012-09-22 00:20:44 +02:00
Gary Linscott fdbe8006e0 Bonus for rook/queen attacking pawns on same rank
Patch and tuning by Gary Linscott from an idea of Ryan Taker.

Double tested by Gary:

Wins: 3390 Losses: 2972 Draws: 11323
LOS: 99.999992%
ELO: 8.213465 +- 99%: 6.746506 95%: 5.124415
Win%: 51.181792 +- 99%: 0.969791 95%: 0.736740

And by me:

After 5612 games 1255  1085  3271 +11 ELO
2012-09-21 23:25:13 +02:00
Marco Costalba 09acdac56b Fix compile on 64 bits
Reported by Quocvuong82.

No functional change.
2012-09-20 19:25:27 +02:00
Marco Costalba e4a0482e43 Simplify BSFTable initialization
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-09-19 20:47:53 +02:00
Marco Costalba 22a5f91aa7 Fix crash under Chess 960
We have a crash with this position:

rkqbnnbr/pppppppp/8/8/8/8/PPPPPPPP/RKQBNNBR w HAha -

What happens is that even if we are castling QUEEN_SIDE,
in this case we have kfrom (B8) < kto (C8) so the loop
that checks for attackers runs forever leading to a crash.

The fix is to check for (kto > kfrom) instead of
Side == KING_SIDE, but this is slower in the normal case of
ortodhox chess, so rewrite generate_castle() to handle the
chess960 case as a template parameter and allow the compiler
to optimize out the comparison in case of normal chess.

Reported by Ray Banks.
2012-09-16 14:14:55 +02:00
Marco Costalba 9ce7469846 Rename class Book to PolyglotBook
And move struct BookEntry out of the header where it is
not needed.

No functional change.
2012-09-16 10:32:57 +02:00
Marco Costalba 4b7dbb3922 Fix KpsK endgame
Broken by commit a44c5cf4f7 of 3 /12 / 2011 that
was labeled "No functional change" because our 'bench'
test didn't triggered that particular endgame. Indeed
we need to run a specific bench on a set of endgames
position when touching endgame.cpp because normal bench
does not cover endgames properly.

Found by MSVC 2012 code analyzer.
2012-09-16 08:57:26 +02:00
Marco Costalba e0035e9ca9 Restore development version
No functional change.
2012-09-15 11:02:08 +02:00
Marco Costalba 0a18adb02a Stockfish 2.3
Stockfish bench signature is: 5416292
2012-09-15 10:56:17 +02:00
Marco Costalba 630b3b2482 Fix compile with Intel 13.0
It seems Intel is unable to properly workout templates with 'static'
storage specifier.

Workaround using an anonymous namespace instead.

No functional change.
2012-09-15 10:55:39 +02:00
Marco Costalba 6008f6538e Don't exit earlier from aspiration window loop
Currently we exit the loop when

abs(bestValue) >= VALUE_KNOWN_WIN

but there is no logical reason for this. It seems more
natural to re-search again with full open window.

This has practically no impact in most cases, we have a
'no functional change' running 'bench' command.
2012-09-14 10:05:46 +02:00
Marco Costalba afcee1e8a4 Fix MSVC 2012 64bits warnings
Reported by Rein.

No functional change.
2012-09-14 09:57:13 +02:00
Marco Costalba e0bd0f250b Speed-up generate<LEGAL>
The trick here is to check for legality only in the
(rare) cases we have pinned pieces or a king move
or an en-passant.

This trick is able to increase the speed of perft
of more then 20%!

No functional change.
2012-09-11 20:24:31 +02:00
Marco Costalba 1598a3edf8 Remove redundancy in move generation
Introduce generate_all_moves() and remove a good
bunch of redundant code.

No functional change.
2012-09-09 17:05:02 +02:00
Marco Costalba 0dacab65eb Simplify generate_castle()
Skipping the calls to std::min(), std::man() we get
even a nice speed-up on perft.

No functional change.
2012-09-09 11:50:28 +02:00
Marco Costalba 834bd9edd7 Rename *last to *end
It is a more correct name because it points past the
last move of the list.

No functional change.
2012-09-09 10:24:40 +02:00
Marco Costalba 6e840f8033 Enable link time optimization only when optimizing
Because it is quite slow, skip it when 'optimize' flag is 'no'

No functional change.
2012-09-09 10:02:11 +02:00
Marco Costalba 2379312028 Revert "Simplify Option c'tor"
std::to_string() is C++11 material, not c++03.

So revert the patch.
2012-09-07 15:21:50 +02:00
Marco Costalba 37db62b2ea Simplify Option c'tor
No functional change.
2012-09-06 18:18:13 +02:00
Marco Costalba b50ce5ebfb Get rid of struct Time
We just need the milliseconds of current system
time for our needs. This allows to simplify the
API.

No functional change.
2012-09-04 09:38:51 +02:00
Marco Costalba 5900ab76a0 Rename current_time() to now()
Follow C++11 naming conventions.

No functional change.
2012-09-02 17:04:22 +02:00
Marco Costalba e6d8e74152 Greatly speed up SEE
Simply reshuffling the code inverting the condition in next_attacker()
yields a miraculous speed up of more than 3% under gcc!

On my laptop a bench run goes from 320Knps to 330Knps

No functional change.
2012-09-02 00:27:53 +02:00
Marco Costalba 9cdca7516c Unroll least valuable attacker loop in SEE
This allows to reduce the scanning for new X-ray attacks
according to the capturing piece type.

It seems to be just a very small speed increase in MSVC 64
bit and gcc 32 bit, I guess cache issues value more than some
instruction less to execute (as usual).

No functional change.
2012-09-01 18:57:38 +02:00
Marco Costalba 6436e75858 Slightly simplify SEE
Some renaming and small code reshuffle.

No fuctional change.
2012-09-01 15:42:04 +02:00
Marco Costalba 831f91b859 Retire Time::restart()
Simplify API.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-31 19:47:07 +02:00
Marco Costalba 1258c7aabe Don't need to memset HashTable
Default c'tor Entry() already initializes
to zero all its POD members.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-31 19:47:00 +02:00
Marco Costalba 8dcb4bc3cc Terminate threads before to exit main()
It is very difficult and risky to assure
that a running thread doesn't access a global
variable. This is currently true, but could
change in the future and we don't want to rely
on code that works 'by accident'. The threads
are still running when ThreadPool destructor is
called (after main() returns) and this could
lead to crashes if a thread accesses a global
that has been already freed. The solution is to
use an exit() function and call it while we are
still in main(), ensuring global variables are
still alive at threads termination time.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-29 19:11:44 +02:00
Marco Costalba 0a003d3ba1 Convert to sync_cout and sync_endl
Serialize access to std::cout all over the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-29 19:11:44 +02:00
Marco Costalba 92e759a676 Introduce serialization of accesses to std::cout
When many threds concurrently print you need to serialize
the access to std::cout to avoid output lines are intermixed
with the contents of each thread.

This is not strictly needed at the moment because
only main thread prints out, although some ad-hoc
test could trigger UCI::loop() printing while searching.

Anyhow we want to lift this pretty avoidable constrain
also as a prerequisite for future work.

This patch just introduces the support, next one will enable
the serialization.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-29 19:11:31 +02:00
Marco Costalba 3df2c01b57 Correctly handle handover of setup states
Before the search we setup the starting position doing all the
moves (sent by GUI) from start position to the position just
before to start searching.

To do this we use a set of StateInfo records used by each
do_move() call. These records shall be kept valid during all
the search because repetition draw detection uses them to back
track all the earlier positions keys. The problem is that, while
searching, the GUI could send another 'position' command, this
calls set_position() that clears the states! Of course a crash
follows shortly.

Before searching all the relevant parameters are copied in
start_searching() just for this reason: to fully detach data
accessed during the search from the UCI protocol handling.
So the natural solution would be to copy also the setup states.
Unfortunatly this approach does not work because StateInfo
contains a pointer to the previous record, so naively copying and
then freeing the original memory leads to a crash.

That's why we use two std::auto_ptr (one belonging to UCI and another
to Search) to safely transfer ownership of the StateInfo records to
the search, after we have setup the root position.

As a nice side-effect all the possible memory leaks are magically
sorted out for us by std::auto_ptr semantic.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-27 19:17:02 +02:00
Marco Costalba 8991a6f005 Use std::deque to store setup states 2012-08-26 18:07:54 +02:00
Marco Costalba 16f380e5c1 Document PolyGlotRandoms[] offsets
Should be more clear from where the 'magic' numbers
come from.

Also bit of reformat while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-25 20:00:07 +01:00
Marco Costalba cbd7ce468c Explicitly use threads.size()
Instead of just size(). Although code is longer,
should be more immediate to understand when reading.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-24 13:41:29 +01:00
Marco Costalba b6883c872d Introduce struct Mutex and ConditionVariable
To mimics C++11 std::mutex and std::condition_variable,
also rename locks and condition variables to be more
uniform across the classes.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-24 12:30:36 +01:00
Marco Costalba fde0b9e701 Slightly microptimize SEE
Reduce of one instruction. It seems a tad faster on
the profiler now. Very slightly but anyhow it is a
code semplification.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-24 09:30:50 +01:00
Marco Costalba 7a2825053e Use size_t as operator[] argument type
This better mimics std::vector::operator[] and
fixes a warning with MSVC 64bit.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-22 11:44:43 +01:00
Marco Costalba 4e619a13d6 Merge generate_direct_checks() in generate_moves()
Further reduce redundancy in move generation.
Veirifed no speed regression on MSVC, Clang and gcc.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-22 11:34:52 +01:00
Marco Costalba 0de9257610 Streamline generate_moves()
Greatly simplify these very performace critical functions.
Amazingly we don't have any speed regression actually under
MSVC we have the same assembly for generate_moves() !

In generate_direct_checks() 'target' is calculated only
once being a loop invariant.

On Clang there is even a slight speed up.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-22 09:23:03 +01:00
Marco Costalba b84af67f4c Reformat piece values arrays
And rename stuff while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-20 19:17:58 +01:00
Marco Costalba 7c8b7222f5 Move zobrist keys out of Position
Are used by Position but do not belong to that class,
there is only one instance of them (that's why were
defined as static), so move to a proper namespace instead.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-20 18:24:06 +01:00
Marco Costalba ec9038b7b4 Retire copy c'tor from class Position
Not needed.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-20 14:44:30 +01:00
Marco Costalba ab65d3fd0e Prefer a reference to a pointer
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-20 07:54:38 +01:00
Marco Costalba e8b7109eff Use enums instead of constants for piece values
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-20 00:08:34 +01:00
Marco Costalba 0c6ed5929c Document De Bruijn sequences
Insted of raw magic numbers.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-19 23:01:48 +01:00
Marco Costalba 2f4a9a140a Avoid wake up master thread when useless
Check we are the last slave of the split point
before to wake up the master. This should avoid
spurious wakes up.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-19 23:01:47 +01:00
Marco Costalba dba1bc354a Simplify idle_loop() signature
We can detect the split point master also from within idle_loop,
so we can call the function without parameters and remove an
overloaded member hack in Thread class.

Note that we don't need to take a lock around curSplitPoint
when entering idle_loop() because if we are the master then
curSplitPoint cannot change under our feet (because is_searching
is set and so we cannot be reallocated), if we are a slave
we enter idle_loop() only upon Thread creation and in that case
is always splitPointsCnt == 0. This is true even in the very rare
case that curSplitPoint != NULL, if we have been already allocated
even before entering idle_loop().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-19 23:01:28 +01:00
Marco Costalba 4b19430103 Prefer size_t over int for array sizes
Align to standard library conventions.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-19 11:01:46 +01:00
Marco Costalba 7c1f8dbde9 Introduce namespace Bitbases
Let's continue this namespace galore...

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-18 13:08:12 +01:00
Marco Costalba 2c1ba2ab0d Introduce namespace UCI
Ater previous patch it comes naturally to take this
extra step.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-18 12:26:22 +01:00
Marco Costalba b011818917 Retire struct OptionsMap
Directly use the underlying std::map instead and avoid
a useless inheritance.

As a nice side-effect Options global object has now a
default c'tor avoiding possible issues with globals
initializations.

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-18 11:30:27 +01:00
Marco Costalba bc4de9edae Explicitly qualify STL functions
Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-18 11:00:56 +01:00
Marco Costalba 9de4ee6d32 Retire MovePickerExt struct
Templetize MovePicker::next_move() member function instead. It
is easier and we also avoid the forwarding of MovePicker() c'tor
arguments in the common case.

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-08-18 10:46:24 +01:00
Marco Costalba 90ec4a403a Guard against 'divide by zero' in bench
Also remove an useless cast.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-30 18:05:15 +01:00
Joseph R. Prostko b61ec33f22 Added Haiku-specific changes to Makefile
First change: If Haiku is host platform, change
installation prefix to /boot/common/bin

Second change: Only link in pthreads if Haiku isn't
host platform.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-30 06:29:46 +01:00
Marco Costalba 1f7b5d9a79 Fix UCI promotion move notation
Regression introduced by revision
f0db6a6c0b

Spotted by Joona.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-26 18:16:24 +01:00
Linus Arver 5c6ba81cf2 Readme.txt: more grammar/style fixes 2012-07-19 12:09:39 -07:00
Linus Arver 591adf564a Readme.txt: grammar/stylistic fixes 2012-07-18 16:46:51 -07:00
Marco Costalba f0db6a6c0b Fix regression in move_to_san()
Broken since commit 628808a113

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-15 09:22:00 +01:00
Marco Costalba 520e680278 Introduce notation.h
And group there all the formatting functions but
uci_pv() that requires access to search.cpp variables.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-15 08:29:07 +01:00
Marco Costalba abc6a0be2f Rewrite pv_info_xxx() signatures
Use the helpers to format the PV info but without
writing to output stream (file or cout). Message
formatting and sending are two logically different
task.

Incidentaly reintroduce the pretty_pv() name,
from Glaurung memories :-)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-15 07:48:33 +01:00
Marco Costalba 9dbda6652e Include castle moves in 'dangerous' flag
Simplifies the code and seems more natural.

We have a very small fucntional change becuase now
at PV nodes castles are extended one ply anyhow.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-14 12:29:11 +01:00
Marco Costalba 5dc0df8435 Merge exclusion search conditions
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-14 12:18:14 +01:00
Marco Costalba 6e5a334c95 Remove redundant condition in is_dangerous()
A pawn on 7th is always passed so retire
this redundant condition.

No funtional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-13 17:57:26 +01:00
Marco Costalba 6becc81446 Silence a MSVC warning in class Tie
With warning level 4 MSVC complains that a default
assignment operator could not be generated due to
member 'file' is a reference (warning C4512).

Use a pointer instead of a reference and move
struct Tie outisde class Logger while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-08 10:50:23 +01:00
Marco Costalba 6b5322ce00 Rename first_1 / last_1 in lsb / msb
It seems more accurate: lsb is clear while 'first
bit' depends from where you look at the bitboard.

And fix compile in case of 64 bits platforms that
do not use BSFQ intrinsics.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-08 09:36:40 +01:00
Marco Costalba 67d91dfd50 Use last_1() to compute new TT size
Transposition table consists of a power of 2
number of TTCluster entries.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-07 19:46:13 +01:00
Marco Costalba 089e54c7fd Revert to -O3 with Clang
Instead of -O4 option that does not work with both mingw and
Linux gcc (tested with Clang 3.1).

As reported by Reed Kotler:
Turns out that -O4 is not a valid option for clang unless you have
the proper gold linker and plugins built. That's because -O4 enables
LTO, which writes out bitcode files during the compile, and then loads
those and optimizes them during the link phase.

It requires a linker that supports LLVM's LTO. There is a plugin for
Gold available as part of LLVM.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-07 19:05:38 +01:00
Marco Costalba 0b3ffb54b7 Fix signedness warning in time_to_msec()
We have a signed integer here so let the return type
take in account that.

Found by Clang with -Weverything option.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-07 16:28:39 +01:00
Marco Costalba 775488340e More idiomatic signature for operator=()
Return a reference instead of void so to enable
chained assignments like

"p = q = Position(...);"

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-05 11:55:35 +01:00
Marco Costalba 7d2530873e Streamline null search reduction formula
Although a (little) functional change, we have no ELO change
but formula it is now more clear.

After 13019 games at 30"+0.05
Mod vs Orig 2075 - 2088 - 8856 ELO 0 (+- 3.4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-07-05 11:48:17 +01:00
Marco Costalba 18505f1fc4 Clear transposition table on "ucinewgame"
It seems the standard behaviour as implemented
in most engines although UCI protocol does not
specify what to do upon "ucinewgame" command.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-30 08:00:48 +01:00
Marco Costalba dc88cd691f Templetize make_move() helpers
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-25 08:09:55 +01:00
Marco Costalba ebe8009aff Reduce indentation in UCIOption::operator=()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-25 00:14:42 +01:00
Marco Costalba 628808a113 Micro-optimize move_to_san()
Calculate the attacks only for the piece to disambiguate,
not for all.

Also some reformatting while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-24 12:38:29 +01:00
Marco Costalba dc7fd868f4 Use type_of() to categorize the moves
Needed to rename old MoveType (used in move generation)
to GenType.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-24 11:07:18 +01:00
Marco Costalba 7b4aa10708 Rename move.cpp to notation.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-24 09:46:31 +01:00
Marco Costalba 960a689769 Rename ThreadsManager to ThreadPool
It is a more standard naming convention.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-24 09:45:37 +01:00
Marco Costalba 5f5d056c8f Replace make_square() with operator|(File, Rank)
Be fancy :-)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-23 09:17:54 +01:00
Marco Costalba 9c7e2c8f9d Coding style in move.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-23 08:34:07 +01:00
Marco Costalba 4f5680950a Add min pawn-king distance to endgame evaluation
At endgame time push the king near his pawns (actually
one of them).

Original idea is from Critter (although slightly different),
implementation is mine and is completely different from the
original, in particular it is different the algorithm to
compute the minimum distance from pawns.

After 19895 games at 15"+0.05
Mod vs Orig 3638 - 3248 - 13009 ELO +7

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-22 06:31:18 +01:00
Marco Costalba 9793fa1906 Calculate min distance between king and his pawns
Just added infrastructure.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-17 09:41:18 +01:00
Marco Costalba 0446fc85de Reformat pick_random() in magics calculation
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-10 11:46:52 +01:00
Marco Costalba 764d3f44b6 Fix a compile error in opposite_colors()
Error is due to ambiguous overloading of operator^
because we have both the built-in operator^(int, int)
and the user defined operator^(Bitboard, Square).

This error does not trigger when using Makefile becuase,
due to luck, the user defined operator^(Bitboard, Square)
happens to be always defined _after_ opposite_colors() so
that compiler does not claim. But in case of Microsoft MSVC
we don't have a Makefile and the order of files compilation
is chosen by the compiler (in an unpredictable way). So it
could still happen that error is not detected (as in my case),
but in another case the order of compilation of the files could
be so that at some point both operator^ were defined before
opposite_colors() and this triggers the error.

The fix is much simpler than the explanation :-)

Reported by Quocvuong82.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-06-02 09:04:11 +01:00
Marco Costalba 0412f4a1ee Fix an issue when adding a book during the game
Currently when we fail to open a book file, for instance
if it doesn't exsist, we leave Book::open() with ifstream
failbit set. If then the book file is added, we correctly
open it at next attempt, but failbit is still set so that
after opening we exit because ifstream::good() returns false.

The fix is to reset failbit upon exiting.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-27 14:13:14 +01:00
Marco Costalba 6828325881 Retire PieceOffset[] in book.cpp
And calculate piece offset on the fly. Also
improve comments.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-27 13:46:55 +01:00
Marco Costalba 61a054b170 Fix a possible 'Division by zero'
In case a book entry has 'count' field set to 0
we crash. Spotted by Clang's static analyzer.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-26 23:17:22 +01:00
Marco Costalba 3361ad4242 Rename psq_score in ReducedStateInfo
So to be fully in sync with StateInfo, and move struct
to position.h, just below StateInfo.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-26 22:43:23 +01:00
Marco Costalba 6c9c6dd989 Fix book file regression
Revision 2aac860db3 of 27 / 4 / 2012
changed can_castle() signatures from bool to int and
this broke the code that calculates polyglot hash key
in book.cpp

Instead of directly fixing the code we prefer to change
castling rights definitions to align to the polyglot ones
(as we did in previous patch). After this step we can simply
take internal castle rights as they are and use them
directly to calculate polyglot book hash key, as we do
in this patch that fixes the regression.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-26 10:14:08 +01:00
Marco Costalba a358dfe934 Redefine enum CastleRight
To be aligned with PolyGlot book castle right definitions.

This will be used by next patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-26 10:03:35 +01:00
Marco Costalba 0ecc920a09 Add a known draw case in kpk bitbase generation
Early classify as known draws the positions
where white king is trapped on the rook file.

Suggested by Dan Honeycutt.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-19 08:45:16 +01:00
Marco Costalba f86182e791 Simplify Position::print()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-12 12:50:11 +01:00
Marco Costalba c19a6ef82d Revert "Don't split if reduced below min_split_depth"
After extensive testing (I was off line this week and let
the test go on) it seems this change is useless:

After 33968 games on a QUAD (4 threads) at 15"+0.05
Mod vs Orig 5425 - 5550 - 22993 ELO -1 (+-2.1)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-12 10:00:26 +01:00
Marco Costalba ecb84464f9 Improve previous patch
Only in case of promotion we care about an upper case
promotion piece char, so std::transform() is overkill
for the task.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-12 09:21:52 +01:00
Balint Pfliegel 1b2af05ea6 Junior promotion patch
Assumption: Junior sends promotions according to the side to move (ucase/lcase).
Fact: Stockfish generally handles promotion lcase.
Patch: Handling position fen input moves always with lcase promotions.

Ported back by Portfish. No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-12 00:04:17 +01:00
Marco Costalba caef319219 Fix compilation with Android NDK
It seems ADL lookup is broken with the STLPort library. Peter says:

The compiler is gcc 4.4.3, but I don't know how many patches they
have applied to it. I think gcc has had support for Koenig lookup
a long time. I think the problem is the type of the vector iterator.
For example, line 272 in search.cpp:

 if (bookMove && count(RootMoves.begin(), RootMoves.end(), bookMove))

gives the error:

jni/stockfish/search.cpp:272: error: 'count' was not declared in this scope

Here RootMoves is:

 std::vector<RootMove> RootMoves;

If std::vector<T>::iterator is implemented as T*, then Koenig lookup
would fail because RootMove* is not in namespace std.

I compile with the stlport implementation of STL, which in its vector
class has:

 typedef value_type* iterator;

I'm not sure if this is allowed by the C++ standard. I did not find
anything that says the iterator type must belong to namespace std.
The consensus in this thread

http://compgroups.net/comp.lang.c++.moderated/argument-dependent-lookup/433395

is that the stlport iterator type is allowed.

Report and patch by Peter Osterlund.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-11 17:16:51 +01:00
Marco Costalba 2f47844c7c Simplify attacks_bb()
And some formatting while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-01 13:16:20 +01:00
Marco Costalba b9bc6e823f Change pos.pieces() argument order
Let first argument to be the 'color'. This allows to align
pos.pieces() signatures with piece_list(), piece_count() and
all the other functions where 'color' argument is passed as
first one.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-01 12:09:20 +01:00
Marco Costalba 5e90580088 Convert constants to decimal representation
Hex representation doesn't add any value in those cases.
Preserve hex representation where more self-documenting
for instance for binary masks values.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-05-01 11:37:12 +01:00
Marco Costalba ef0496ff40 Don't split if reduced depth is below min_split_depth
It seems to increase SMP performances. To note that
this patch goes in the opposite direction of "Active
reparenting" where we try to reparent an idle slave
as soon as possible. Instead here we prefer to keep
it idle instead of splitting on a shallow / near the
leaves node.

After 11550 games on a QUAD (4 threads) at 15"+0.05
Mod vs Orig 1972 - 1752 - 7826 ELO +6 (+-3.6)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-29 23:28:07 +01:00
Marco Costalba 96f4ab48d8 Increase optimization level of Clang
Set optimization level to 4 and get a 2.564% faster binary:

Stockfish (Clang, Level 4) bench:
$ make build ARCH=osx-x86-64 COMP=clang
(Clang does not support PGO)
Average of 4 trials:
Total time (ms): 5137.5
Nodes searched: 5631135
Nodes/second: 1096084.5

Stockfish (Clang, Level 3) bench:
$ make build ARCH=osx-x86-64 COMP=clang
(Clang does not support PGO)
Average of 4 trials:
Total time (ms): 5269.25
Nodes searched: 5631135
Nodes/second: 1068679.25

Stockfish (GCC, PGO) bench:
$ make profile-build ARCH=osx-x86-64
Average of 4 trials:
Total time (ms): 5286
Nodes searched: 5631135
Nodes/second: 1065292.25

Suggestion and performance tests by Daylen Yang.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-29 23:11:23 +01:00
Marco Costalba 39c08c17c5 Remove unreachable extension condition
A PvNode that givesCheck has been already granted
an extension of ONE_PLY in previous condition.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-29 18:58:48 +01:00
Andy Duplain b93693b831 Add support for Mac clang compiler
Makefile modified to support the clang compiler under Mac.
This was tested using clang 4 under Mountain Lion, but should
also work fine under Lion and possibly under Snow Leopard.

It requires the 'Xcode 4.x Command Line Tools' to be installed.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-29 11:47:02 +01:00
Marco Costalba cdfe43eb8f Proper indenting of multiple conditions
Triviality due to a boring saturday morning.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-28 11:10:51 +01:00
Marco Costalba 456f37b8ab Rename square_empty() to is_empty()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-28 10:54:33 +01:00
Marco Costalba e8d89ca5b0 Micro-optimize king evaluation
Reuse already calculated value, instead of calling
king_safety() again.

Patch suggested by Balint Pfliegel.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-28 10:37:55 +01:00
Marco Costalba 2aac860db3 Fix wrong condition in PawnEntry::king_safety()
Since revision 374c9e6b63
we use also castling information to calculate king safety.
So before to reuse the cached king safety score we have to
veify not only king position, but also castling rights are
the same of the pre-calculated ones.

This is a very subtle bug, found only becuase even after
previous patch, consecutives runs of 'bench' _still_ showed
different numbers. Pawn tables are not cleared between 'bench'
runs and in the second run a king safety score, previously
evaluated under some castling rights, was reused by another
position with different castling rights instead of being
recalculated from scratch.

Bug spotted and tracked down by Balint Pfliegel

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-27 10:59:18 +01:00
Marco Costalba 8b00e50cb7 Clear TT before running 'bench'
Now that we can call bench multiple times
from command prompt we need to ensure searched
nodes remain constant across different runs.

Spotted by Blint Pfliegel.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-26 19:37:30 +01:00
Marco Costalba be3b8f3ae9 Retire "Active reparenting"
After 6K games at 60" + 0.1 on QUAD with 4 threads
this implementation fails to show a measurable increase,
result is well within error bar.

Perhaps with 8 or more threads resut is better but we
don't have the hardware to test. So retire for now and
in case re-add in the future if it proves good on big
machines.

The only good news is that we don't have a regression and
implementation is stable and bug-free, so could be reused
somewhere in the future.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-22 17:52:31 +01:00
Marco Costalba ce159b16b9 Fix endless reaparenting loop
The check for detecting when a split point has all the
slaves still running is done with:

   slavesMask == allSlavesMask

When a thread reparents, slavesMask is increased, then, if
the same thread finishes, because there are no more moves,
slavesMask returns to original value and the above condition
returns to be true. So that the just finished thread immediately
reparents again with the same split point, then starts and
then immediately exits in a tight loop that ends only when a
second slave finishes, so that slavesMask decrements and the
condition becomes false. This gives a spurious and anomaly
high number of faked reparents.

With this patch, that rewrites the logic to avoid this pitfall,
the reparenting success rate drops to a more realistical 5-10%
for 4 threads case.

As a side effect note that now there is no more the limit of
maxThreadsPerSplitPoint when reparenting.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-17 18:51:49 +01:00
Marco Costalba 5392007a24 Improved cutoff check when reparenting
Check for a cutoff occurred also high in
the tree and not only at current split
point.

This avoids some more wasted reparenting.

No functional chnage.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-16 18:28:08 +01:00
Marco Costalba f59323b56a Use more_than_one() instead of single_bit()
It is more correct given what the function does. In
particular single_bit() returns true also in case of
empty bitboards.

Of course also the usual renaming while there :-)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-14 09:51:59 +01:00
Marco Costalba 25a9b601b2 Reparent to latest
Instead of reparenting to oldest split point, try to reparent
to latest. The nice thing here is that we can use the YBWC
helpful master condition to allow the reparenting of a split
point master as long as is reparented to one of its slaves.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-14 07:20:31 +01:00
Marco Costalba c645aca199 Don't reparent if a cutoff is pending
And update master->splitPointsCnt under lock
protection. Not stricly necessary because
single_bit() condition takes care of false
positives anyhow, but it is a bit tricky and
moving under lock is the most natural thing
to do to avoid races with "reparenting".

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-12 21:17:52 +01:00
Marco Costalba 44432f67d7 Active Reparenting
In Young Brothers Wait Concept (YBWC) available slaves are
booked by the split point master, then start to search below
the assigned split point and, once finished, return in idle
state waiting to be booked by another master.

This patch introduces "Active Reparenting" so that when a
slave finishes its job on the assigned split point, instead
of passively waiting to be booked, searches a suitable active
split point and reprents itselfs to that split point. Then
immediately starts to search below the split point in exactly
the same way of the others split point's slaves. This reduces
to zero the time waiting in idle loop and should increase
scalability especially whit many (8 or more) cores.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-10 18:22:58 +01:00
Marco Costalba d66b765eb6 Sync compute_xxx implementations
Also refactored Position::pos_is_ok() while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-09 13:11:38 +01:00
Marco Costalba e72b93e44f Move Tempo to evaluation
Apart from the semplification it is now more clear that
the actual Tempo added was half of the indicated score.
This is becuase at start compute_psq_score() added half
Tempo unit and in do_move() white/black were respectively
adding/subtracting one Tempo unit.

Now we have directly halved Tempo constant and everything
is more clear.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-09 11:14:00 +01:00
Marco Costalba 68885f78f3 Micro-optimize do_castle_move()
Use the same tables update trick used in do_move().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-09 09:41:41 +01:00
Marco Costalba 4a310baae2 Disable book during analysis
It is still enabled during fixed limit search so to
use it during fixed depth/nodes/time matches.

Bug reported by Daylen.

No functional changes.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-09 08:09:02 +01:00
Marco Costalba d549497144 Introduce make_castle_right() helper
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-08 17:19:49 +01:00
Marco Costalba e56342ed00 Shrink castlePath[] and castleRookSquare[] sizes
Shrinking from [16] to [2][2] is able to speedup
perft of start position of almost 5% !

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-08 16:55:01 +01:00
Marco Costalba 0049d3f337 Reduce stack usage
Shrink dimensions of the biggest stack consumers arrays.
In particular movesSearched[] can be safely shrinked
without any impact on strenght or risk of crashing.
Also MAX_PLY can be reverted to 100 with almost no impact
so to limit search recursion and hence stack allocation.

A different case is for MAX_MOVES (used by Movepicker's
moves[]), because we know that do exsist some artificial
position with about 220 legal moves, so in those cases SF
will crash. Anyhow these cases are never found in games.
An open risk remains perft, especially run above handcrafted
positions.

This patch originates from a report by Daylen that found
SF crashing on his Mac OS X 10.7.3 while in deep analysys
on the following position:

8/3Q1pk1/5p2/4r3/5K2/8/8/8 w - - 0 1

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-08 08:46:04 +01:00
Marco Costalba e0cae4bef8 Fix 'bench' for Chess960 case
Now a fen file with Chess960 positions is
correctly parsed. But it is mandatory to set
"UCI_Chess960" option _before_ to call bench.

Note that this was not needed/possible before
adding the possibility to call 'bench' from
command prompt.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-07 14:07:30 +01:00
Marco Costalba 9546b79e20 Use bench to implement UI 'perft' command
Now that we can call bench on current position
we can directly use it to perform our perft.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-07 13:45:53 +01:00
Marco Costalba cc04a745e2 Teach 'bench' to run current position
Now that we can call bench from command prompt
has a sense to teach bench to run the current
set position. To do this is enough to call bench
with 'current' as fen source parameter.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-07 12:08:34 +01:00
Marco Costalba ce5b972736 Don't need to wait after a "ponderhit"
It is enough to wake up main thread. This is
a better fix than d033d5e06a.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-07 11:47:37 +01:00
Marco Costalba f30f384757 Revert "Penalty for undefended rook"
After extensive test Gary says:

"So, after 16k games at 10"+1" on an i7, the undefended rook test
looks to be not good (albeit by a very small margin).
3063 - 3093 - 9844 (-1).

I doubt that is causing the regression, but even so, it looks like
it's not worth keeping, and we can go back to the simpler undefended
minors check."

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-07 11:34:22 +01:00
Marco Costalba 676b2c8435 Replace Position::copy()
With assignment operator. And fix Position::flip().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 19:42:45 +01:00
Marco Costalba c2fc80e5d1 Revert thread_local stuff
Unfortunatly accessing thread local variable
is much slower than object data (see previous
patch log msg), so we have to revert to old code
to avoid speed regression.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 18:47:55 +01:00
Marco Costalba b1f57e92ce Use thread_local compiler specifics
Much faster then pthread_getspecific() but still a
speed regression against the original code.

Following are the nps on a bench:

Position
454165
454838
455433

tls
441046
442767
442767

ms (Win)
450521
447510
451105

ms (pthread)
422115
422115
424276

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 18:03:15 +01:00
Marco Costalba bed4075580 Fix a (theoretical) race leading to a crash
After we release the SplitPoint lock the master, suppose
is main thread, can safely return and if a "quit" command
is pending, main thread exits and associated Thread object
is freed. So when we access master->is_searching a crash
occurs.

I have never found such a race that is of course very rare
becuase assumes that from lock releasing we go to sleep for
a time long enough for the main thread to end the search and
return. But you can never know, and anyhow a race is a race.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 15:30:41 +01:00
Marco Costalba 5a2d525048 Teach UI thread to use main thread resources
So to avoid a crash when setting the moves in
UCI "position startpos moves ...." command.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 15:05:52 +01:00
Marco Costalba e1919384a2 Don't store Thread info in Position
But use the newly introduced local storage
for this. A good code semplification and also
the correct way to go.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 14:36:45 +01:00
Marco Costalba 699f700162 Introduce thread local storage
Use thread local storage to store a pointer to the thread we
are running on. This will allow to remove thread info from
Position class.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 14:36:39 +01:00
Marco Costalba 797c960d20 Rewrite pop_1st_bit() to be endian independent
With this change sources are fully endianess
independent, so we can simplify the Makefile.

Somewhat surprisingly we don't have any speed
regression !

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-06 12:38:27 +01:00
Marco Costalba 673bc5526f Use a Thread instead of an array index
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-04 12:12:08 +01:00
Marco Costalba 0439a79566 Big Position renaming
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-04 07:35:49 +01:00
Marco Costalba 37fa8adc2b Micro-optimize last_1() for 32bits
Verified assembly it is a bit simpler.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-04 06:54:48 +01:00
Marco Costalba 2f99de0c6c Fix bench with fen files regression
Erroneusly adds default positions to the fens
loaded from external file.

Bug introduced in adb71b8096

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-02 18:40:12 +01:00
Marco Costalba 7a8429d9f1 Simplify Endgames::probe()
With this API change we simplify both function and caller site.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-02 09:58:29 +01:00
Marco Costalba dda0fa1a43 Use polymorphism to resolve map() overloading
The 2 overload functions map() accept a pointer to
EndgameBase<Value> or a pointer to EndgameBase<ScaleFactor>.

Because Endgame<E> is derived from one of them we can
directly use a pointer to this class to resolve the
overload as is needed in Endgames::add().

Also made class Endgames fully parametrized and no more
hardcoded to the types (Value or ScaleFactor) of endgames
stored.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 16:16:51 +01:00
Marco Costalba 7eb6a488ad Use a std::vector to store searchMoves
A std::set (that is a rb_tree) seems really
overkill to store at most a handful of moves
and nothing in the common case.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 12:45:43 +01:00
Marco Costalba 72641dcaae Retire platform specifics include in misc.cpp
Now that platform.h is included in types.h we
don't need this stuff anymore.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 12:00:15 +01:00
Marco Costalba 6e00aa6bae Better document square flipping helpers
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 11:30:54 +01:00
Marco Costalba 9bbd27a80f Introduce Bitboards namespace
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 11:01:13 +01:00
Marco Costalba adb71b8096 Process 'bench' also from SF prompt
It is possible to start with 'stockfish', then from
command prompt type 'bench' and SF will do what you expect.
Old behaviour is anyhow preserved. As a bonus we can now
start from command line any UCI command understood by
Stockfish. The difference is that after execution of a
command from arguments SF quits, while at the end of the
same command from prompt SF stays in UCI loop.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-04-01 10:06:30 +01:00
Marco Costalba 32c504076f Use std::vector to implement HashTable
Allows some code semplification and avoids directly
allocation and managing heap memory.

Also the usual renaming while there.

No functional change and no speed regression.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-31 19:07:11 +01:00
Marco Costalba 304deb5e83 Rename Materials and Pawns hash stuff
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-31 11:59:23 +01:00
Marco Costalba d84865eac3 Complete the renaming in Search::LimitsType
This completes the job started with revision
4124c94583.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-31 09:16:09 +01:00
Marco Costalba cc6c745b54 Reset search time as early as possible
In particualr before to wake up main thread that
could take some random time. Until we don't reset
search time we are not able to correctly track
the elapsed search time and this can be dangerous
under extreme time pressure.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-31 10:01:31 +01:00
Marco Costalba d033d5e06a Revert "Call wait_for_search_finished() only when quitting"
We need to wake up main thread if it is sleeping
waiting for stop or ponderhit, so we cannot skip
calling wait_for_search_finished().

Found by Othello1984.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-30 18:58:58 +01:00
Marco Costalba b0b9bb3462 Last touches to pawns shelter code
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-29 21:14:15 +01:00
Marco Costalba 5e18b81e87 Fix an hang when max depth is reached
In this case SF stop searching and goes sleeping
waiting for a stop / ponderhit before to return
best move. So when a "stop" arrives we need to wake
up the main thread again.

Another regression introduced by 3aa471f2a9,
hopefully the last one.

Thanks to Otello1984 to reporting this.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-29 20:04:27 +01:00
Marco Costalba 10e64e0509 Refactor pawns shelter and storm
Renamed stuff and added comments. The aim is to make more
readable, at least by me ;-) , this newly added part of code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-29 00:09:27 +01:00
Marco Costalba 76622342ec Restore MS1BTable[]
Incredible typo from my side!

The 2 tables are completely different, one counts 1s the
other returns the msb position. Even more incredible
the 'stockfish bench' command returns the same number
of nodes!!!

Spotted by Justin Blanchard.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-28 19:11:37 +01:00
Marco Costalba d6e3a40c81 Silently handle "ucinewgame" command
Avoid returning "Unknown command", it seems some
GUI are misguided by this.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-28 18:24:02 +01:00
Marco Costalba 46a50cbf38 Replace MS1BTable[] with BitCount8Bit[]
We already have the necessary infrastructure
in place.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-28 14:08:19 +01:00
Marco Costalba cc2b3ece5c Merge pull request #11 from glinscott/squash
Add more detailed pawn shelter/storm evaluation

After 10670 games at 10"+0.05
Mod vs Orig 2277 - 1941 - 6452 ELO +11 !!!

The first real increase since 2.2.2, congratulations Gary !!!

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-28 07:48:30 +01:00
Marco Costalba 22e294e044 Set do_sleep out of lock protection
Fixes a not so rare crash (once every 100 games)
newly introduced. Unfortunatly I am still not
able to figure out why :-(

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-27 19:27:20 +01:00
Marco Costalba 4124c94583 Use UCI names in Search::LimitsType
There is no need to "invent" different names
from the original UCI parameters.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-27 14:26:58 +01:00
Marco Costalba a56322fde8 Merge pull request #9 from glinscott/master
Penalty for undefended rook

Almost no change at longer TC, but perhaps there
is a tiny increase....

After 17522 games at 10"+0.05
Mod vs Orig 3064 - 2967 - 11491 ELO +2

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-26 19:57:27 +01:00
Marco Costalba 3d0d0237c5 Simplify start_searching() signature
Retire the "sync" behaviour that now is up to
the caller to honour.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-26 18:59:01 +01:00
Marco Costalba d11a529904 Call wait_for_search_finished() only when quitting
When quitting we should avoid RootPosition to be
destroyed while threads are still running, leading
to a crash. In case of a "stop" or "ponderhit"
command there is no need for the UI thread to wait.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-26 18:44:30 +01:00
Marco Costalba 3aa471f2a9 Introduce and use wait_for_search_finished()
Helper function that allows us to simplify
the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-26 18:22:41 +01:00
Gary Linscott 374c9e6b63 Add more detailed pawn shelter/storm evaluation 2012-03-26 07:52:10 -04:00
Marco Costalba 32d3a07c67 Move ThreadsManager::exit() to d'tor
And add final touches to this long patch series.

All the series has been verified against regression with
20K games at fast TC.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-26 08:18:17 +01:00
Marco Costalba b978eb05dc Fix compile error with gcc
We have a clash with start_fn defined both as a
Thread memeber and as a function pointer type in
pthread_create().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 23:08:35 +01:00
Marco Costalba e4efc8b741 Reset Thread::maxPly before a new search
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 16:44:55 +01:00
Marco Costalba 58c2fe391d Fix race in ThreadsManager::sleep()
We cannot set do_sleep flag of main thread before
"bestmove" is sent to GUI, otherwise GUI could send
immediately the next "go" command that triggers
start_thinking() and because do_sleep is set UI
thread resets the flag to launch a new search. But
when shortly after main thread returns to main_loop()
flag is incorrectly reset and main thread goes to sleep
hanging the engine.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 16:37:48 +01:00
Gary Linscott dbe5e28eaa Merge remote-tracking branch 'upstream/master' 2012-03-25 10:18:29 -04:00
Marco Costalba c483ffc773 Try to mimic std::thread API
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 12:43:19 +01:00
Marco Costalba 41561c9bb8 Use std::vector<Thread*> to store threads
We store pointers instead of Thread objects because
Thread is not copy-constructible nor copy-assignable
and default ones are not suitable. So we cannot store
directly in a std::vector.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 10:23:52 +01:00
Marco Costalba 553655eb07 Refactor Thread class
Associate platform OS thread to the Thread class instead of
creating it from ThreadsManager.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 10:23:51 +01:00
Marco Costalba f01b53c374 Refactor ThreadsManager::set_size() functionality
Split the data allocation, now done (mostly once)
in read_uci_options(), from the wake up and sleeping
of the slave threads upon entering/exiting the search.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 10:23:49 +01:00
Marco Costalba 8ec421fa14 Revert "Don't sync with C library I/O buffers"
It seems is the cause of strange and rare hangs
reported by some users where Stockfish stops
responding to GUI. It is not clear why but for
the moment revert the patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 10:23:16 +01:00
Marco Costalba 11b0c7b44a Don't ceil cpu_count()
It is already done at calling site where it is
more appropiate.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-25 09:57:33 +01:00
Marco Costalba f8224fc7d3 Fix a MSVC warning
Not correct warning about use of an uninitialized
variable. The warning is not correct becuase we can
never reach the warned code when in SpNode, anyhow
the fix is simple.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-24 10:14:21 +01:00
Marco Costalba b356e0fae3 Rename lock.h to platform.h
And move some more platform specific code there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-24 10:05:17 +01:00
Marco Costalba 06f33ff1ee Remove last platform specific code form thread.cpp
A somewhat tricky function pointer cast allows us
to move the platform specifics to lock.h, the cast
is tricky because return type is not the same of the
casted function in Linux (for Windows return type is
a DWORD that is a long) but this should not be a
problem as long as the size is the same;

From: http://stackoverflow.com/questions/188839/function-pointer-cast-to-different-signature

"OpenSSL was only casting functions pointers to
other function types taking and returning the same
number of values of the same exact sizes, and this
(assuming you're not dealing with floating-point)
happens to be safe across all the platforms and
calling conventions I know of. However, anything
else is potentially unsafe."

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-23 19:28:20 +01:00
Marco Costalba c47a74ec62 Merge two loops in ThreadsManager::init()
In analogy with ThreadsManager::exit()

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-22 22:40:04 +01:00
Marco Costalba e26d13bb31 Use a local copy of tte->value()
This should avoid some aliasing issues
with TT table access.

After 3913 games at 10"+0.05
Mod vs Orig 662 - 651 - 2600  ELO +0 (+- 6.4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-22 19:32:30 +01:00
Gary Linscott f41a21fefd Penalty for undefended rook as well 2012-03-22 07:33:47 -04:00
mcostalba d4c9abb967 Merge pull request #8 from glinscott/master
Optimize undefended minor check. Little editing by
me, no change even at assembly level.

No regression after 8K games at fast TC on a 64bit CPU.
2012-03-22 07:49:11 +01:00
Gary Linscott d1e18fc7dd Optimize undefended minor check. 2012-03-21 08:19:21 -04:00
Gary Linscott 3c6a4bfbed Penalize undefended minors
Even if not under attack. This seems to be good
especially on openings.

After 12112 games at 10"+0.05
Mod vs Orig 2175 - 1997 - 7940 ELO +5 (+- 3.7)

[Patch series from Gary, little edited by me]

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-21 08:00:57 +01:00
Marco Costalba 3b7dbc4f6d Fix Logger under MSVC iostream libraries
We need splitted Tie classes because MSVC stream library
takes a lock on buffer both on reading and on writing and
this causes an hang because, while searching, the I/O
thread is locked on getline() and when main thread is
trying to std::cout() something it blocks on the same
lock waiting for I/O thread getting some input and
releasing the lock.

The solution is to use separated streambuf objects for
cin and cout.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-20 21:46:08 +01:00
Marco Costalba 17d1940278 Remove cruft from Logger class
A big code simplification and cruft removing, make
Logger class a singleton and fully self conteined.
Also add direction indicators (">>" and "<<") to
better differentiate input and output lines in the
log file.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-20 19:24:52 +01:00
Marco Costalba 258da28e79 Better on_change() argument name
Using "o" as a parameter with the on_xxx(const UICOption& o)
functions is a bit dangerous because of confusion with "0".

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-20 07:09:58 +01:00
Marco Costalba df80232495 Add also logging of std::cin
Some trial was needed to find the correct recipe but now
we log both stdin and stdout to file "io_log.txt".

Link http://spec.winprog.org/streams/ was very useful
to understand the details of iostreams implementation.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-19 21:59:25 +01:00
Marco Costalba eb28a683bd Add (smart) logging facility
By means of "Use Debug Log" UCI option it is possible to toggle
the logging of std::cout to file "out.txt" while preserving
the usual output to stdout. There is zero overhead when logging
is disabled and we achieved this without changing a single line
of exsisting code, in particular we still use std::cout as usual.

The idea and part of the code comes from this article:
http://groups.google.com/group/comp.lang.c++/msg/1d941c0f26ea0d81

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-18 23:10:24 +01:00
Marco Costalba 2dfc94e0b6 Show startup messages immediately
In particular before initialization. So that SF
seems more snappy at startup.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-18 23:10:18 +01:00
Marco Costalba 7bc3688714 Revert to byTypeBB[0] storing occupied squares
As it was in Glaurung times. Also rearranged order
so that byTypeBB[0] is accessed before byTypeBB[x]
to be more cache friendly. It seems there is even
a small speedup.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-18 12:02:23 +01:00
Marco Costalba fc3ea7365a Rename occupied_squares() to pieces()
Also some microoptimizations, were there from ages
but hidden: the renaming suddendly made them visible!

This is a good example of how better naming lets you write
better code. Naming is really a kind of black art!

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-18 11:18:38 +01:00
Marco Costalba 55376219b7 UCI buttons don't need a value
Take advantage of this to further simplify the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-17 21:44:50 +01:00
Marco Costalba 9b26356347 Don't use "OwnBook" by default
Stick to UCI protocol that says:

* by default all the opening book handling is done by the GUI,
  but there is an option for the engine to use its own book
  ("OwnBook" option, see below)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-17 13:16:51 +01:00
Marco Costalba 7c8a8e038f Retire "ucinewgame" UCI option
UCI protocol it is not clear about what the engine
should be supposed to do when "ucinewgame" is
received. Stockfish simply sets the position to
start FEN, but it is redundant becuase the GUI always
resends the position after "ucinewgame" command, so
it seems we can safely ignore that command.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-17 11:24:19 +01:00
Marco Costalba ee838f56f7 Fix UCI 'button' options
When a button fires UCIOption::operator=() is called and from
there the on_change() function. Now it happens that in case of
a button the on_change() function resets option's value to
"false" triggering again UCIOption::operator=() that calls again
on_change() and so on in an endless loop that is experienced
by the user as an application hang.

Rework the button logic to fix the issue and also be more clear
about how button works.

Reported by several people working with Scid and tracked down
to the "Clear Hash" UCI button by Steven Atkinson.

Bug recently introduced by 2ef5b4066e.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-17 10:35:01 +01:00
Marco Costalba 9934b8ec31 Don't sync with C library I/O buffers
Now we are forced to just use C++ iostream becuase
buffers are independent and using C library functions
like printf() or scanf() could yield to issues.

Speed up of about 1%.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-12 19:34:19 +01:00
Marco Costalba 3dccdf5b83 Fix time_to_msec() precision
Result of t.time * 1000 should be a 64 bit value, not an int.

Bug reported by several users.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-10 19:38:56 +01:00
Marco Costalba 4220f191d8 Introduce Eval namespace
Wrap evaluation related stuff and reshuffle
a bit the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-07 07:39:01 +01:00
Marco Costalba 843a5961e1 Double pinner bonus
Fine tune newly introduced pinner bonus score:

After 34696 games at 2"+0.05
Mod vs Orig 7474 - 7087 - 20135 ELO +3 (+- 2.4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-07 07:31:43 +01:00
Marco Costalba d8e56cbe54 Convert init of eval to async option
So to be done only once at startup and in the (unlikely)
cases that a relevant UCI parameter is changed, instead
of doing it at the beginning of each search.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-06 19:21:00 +01:00
Marco Costalba 2ef5b4066e Async UCI options actions
Introduce 'on change' actions that are triggered as soon as
an UCI option is changed by the GUI. This allows to set hash
size before to start the game, helpful especially on very fast
TC and big TT size.

As a side effect remove the 'button' type option, that now
is managed as a 'check' type.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-05 19:20:07 +01:00
Marco Costalba 482b5b7ece Use new Time class in timed_wait()
And simplify the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-05 19:18:46 +01:00
Marco Costalba 19540c9ee8 Introduce single_bit() helper
Self-documenting code instead of a tricky
bitwise tweak, not known by everybody.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-04 23:37:44 +01:00
Marco Costalba cf0561d31a Introduce pinning bonus
Add a bonus if a slider is pinning an enemy piece.
Idea from Critter.

After 27443 games at 2"+0.05
Mod vs Orig 5900 - 5518 - 16025 ELO +4 (+- 2.7)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-04 23:01:31 +01:00
Marco Costalba 161c6b025e Rewrite time measurement code
Introduce and use a new Time class designed after
QTime, from Qt framework. Should be a more clear and
self documented code.

As an added benefit we now use 64 bits internally to get
millisecs from system time. This avoids to wrap around
to 0 every 2^32 milliseconds, which is 49.71 days.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-03-03 18:01:39 +01:00
Marco Costalba b966adf103 Halve rook on open file bonus for endgame
After 42206 fast games TC 2"+0.05
Mod vs Orig 12871 - 16849 - 12486 ELO +3 (+- 2.6)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-29 06:20:48 +01:00
Marco Costalba c2a68708ef Fix a shift overflow warning
Visual Studio 11 is worried that shift result could
overflow an integer, this is impossible becuase max
value of the shift is 4, but compiler cannot know it.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-27 20:32:41 +01:00
Marco Costalba 34178205fc Micro-optmize castling moves
Pre compute castle path so to quickly test
for impeded rule.

This speeds up perft on starting position
of more than 2%.

No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-27 19:44:04 +01:00
Marco Costalba 5bb766e826 Rename promotion_piece_type() to promotion_type()
Shorter and equally clear to understand.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-26 18:39:53 +01:00
Marco Costalba 96eefc4af6 Introduce another two (bitboard,square) operators
And simplify the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-26 18:39:40 +01:00
Marco Costalba 8751b18cf0 Fix MSVC warning on streampos to size_t conversion
Fix this warning with MSVC 64 bits:

warning C4244: '=' : conversion from 'std::streampos' to 'size_t',
possible loss of data

Point is that std::streampos could be negative, while size_t
is always non-negative.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-26 12:05:37 +01:00
Marco Costalba 2608b9249d Retire ss->bestMove
And introduce SPlitPoint bestMove to pass back the
best move after a split point.

This allow to define as const the search stack passed
to split.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-21 20:31:22 +01:00
Marco Costalba 43f84efa15 Don't update bestValue in check_is_dangerous()
It is a prerequisite for next patch and simplifies
the function. testing at ultra fast TC shows no
regression.

After 24302 games at 2"+0.05
Mod vs Orig 5122 - 5038 - 13872 ELO +1 (+- 2.9)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-21 20:29:05 +01:00
Marco Costalba ea5616785e Fix a wrong check in pos_is_ok()
Bug introduced by revision a44c5cf4f7
of 3/12/2011.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-20 19:34:43 +01:00
Marco Costalba 6a48325c49 Further simplify castling rights
Reverse the meaning of castleRightsMask[sq] so that now
is stored the castling right that will be removed in
case a move starts from or arrives to sq square. This
allows to simplify the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-20 19:20:13 +01:00
Marco Costalba 50edb7cd73 Spread usage of pos.piece_moved()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-19 12:20:29 +01:00
Marco Costalba 4aadd1e401 Retire empty_squares()
Use ~pos.occupied_squares() instead and avoid to
hide the ~ computation.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-19 11:50:50 +01:00
Marco Costalba ec5b9994b5 Index en-passant zobrist keys by file
Instead of by square. This is a more conventional
approach, as reported also in:

http://chessprogramming.wikispaces.com/Zobrist+Hashing

We shrink zobEp[] from 64 to 8 keys at the cost of an extra
'and 7' at runtime to get the file out of the ep square.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-19 11:31:36 +01:00
Marco Costalba 3441e0075d Move some stuff out of lock protection in split()
We shouldn't need lock protection to increment
splitPointsCnt and set curSplitPoint of masterThread.

Anyhow because this code is very tricky and prone to
races bound the change in a single patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-19 10:32:38 +01:00
Marco Costalba 821e1c7233 Micro-optimize castleRights update
When updating castleRights in do_move() perform only one
64bit xor with zobCastle[] instead of two.

The trick here is to define zobCastle[] keys of composite
castling rights as a xor combination of the keys of the
single castling rights, instead of 16 independent keys.

Idea from Critter although implementation is different.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-19 10:04:49 +01:00
Marco Costalba 6088ac2108 Small renaming in Thread struct
Should be a bit more clear the meaning of the
single variables.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-18 10:57:00 +01:00
Marco Costalba d8349f9d0f Fix a race when extracting PV from TT
Because TT table is shared tte->move() could change
under our feet, in particular we could validate
tte->move() then the move is changed by another
thread and we call pos.do_move() with a move different
from the original validated one !

This leads to a very rare but reproducible crash once
every about 20K games.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-18 10:20:07 +01:00
Marco Costalba fd35d92c1e Increase MAX_PLY from 100 to 256
There is no need to limit the maximum ply searched to
100, with deep exclusion search extensions we could
reach it even with much smaller search depths.

The only drawback is an increase in stack usage, but
is limited mainly to id_loop(), in particular the
recursive search() functions are not affected.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-18 10:16:01 +01:00
Marco Costalba 3b906ffc27 Micro-optimize pop_1st_bit() for 32 bits
Small perft speed-up of 2% and also a code
simplification.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-13 10:07:09 +01:00
Marco Costalba 7b4b65d7a9 Templetize sliding attacks
No functional change and no speed regression, it seems
to be even a bit faster on MSVC and gcc.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-13 09:15:55 +01:00
Marco Costalba 099b5e45e6 Speedup sliders attacks for 32bit CPU
Replace a 64 bit 'and' by two 32 bits ones and
use unsigned instead of int.

This simple patch increases perft speed of 6% on
my Intel Core 2 Duo !

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-12 15:26:18 +01:00
Marco Costalba b1768c115c Don't wake up threads at the beginning of the search
But only when needed, after a split point. This behaviour
does not apply when useSleepingThreads is false, becuase
in this case threads are not woken up at split points so
must be already running.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-12 14:17:12 +01:00
Marco Costalba 86e159997c Don't reset 50-move counter after castling
Rule says should be reset only after a capture and/or
a pawn move.

This incredible bug was here since Glaurung times !

Spotted by Kiriakos.

No functional change in the test bench because we
don't reach the 50 moves limits.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-12 10:07:10 +01:00
mcostalba 576f0f6985 Merge pull request #3 from glinscott/b46bf29
Detect stalemate in evaluation
2012-02-05 21:34:32 -08:00
Gary Linscott b46bf2950f Simpler stalemate check. 2012-02-05 14:52:01 -05:00
Gary Linscott 0347339970 Detect stalemate in KXK endgames
Also, handle cases where there are 2 bishops of the same color.
2012-02-05 10:24:53 -05:00
Marco Costalba 40e939421f Add "Slow Mover" UCI parameter to adjust time management
With default value of 100 no change in regard of current
behaviour. Increasing the value makes SF to think a
longer time for each move. Decreasing the value makes SF
to move faster.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-04 10:41:09 +01:00
Marco Costalba b1cf1acb93 Move wait_for_stop_or_ponderhit() under Thread
This method belongs to Thread, not to ThreadsManager.

Reshuffle stuff in thread.cpp while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-03 17:33:09 +01:00
Marco Costalba c94cfebb7e Reduce lock contention in idle_loop
Release split point lock before to wake up
master thread. This seems to increase speed
in case "sleeping threads" are used:

After 7792 games with 4 threads at very fast TC (2"+0.05)
Mod vs Orig 1722 - 1627 - 4443 ELO +4 (+- 5.1)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-03 13:43:58 +01:00
Marco Costalba 57e942145c Fix an alignment warning with MSVC
The declared alignment is different from the one
in the definition.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-02-03 13:14:57 +01:00
Marco Costalba 51e8efdab5 Fix subtle race with slave allocation
When allocating a slave we set both is_searching
and splitPoint under lock protection.

Unfortunatly the order in which the variables are
set is not defined. This article was very clarifying:

http://software.intel.com/en-us/blogs/2007/11/30/volatile-almost-useless-for-multi-threaded-programming/

So when in idle loop we test for is_searching and then
access splitPoint, it could happen that splitPoint is still
not updated leading to a possible crash.

Fix the race lock protecting splitPoint access.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-31 20:19:25 +01:00
Marco Costalba df31398bb9 Fix bug in useless checks prune
With current code we could raise bestValue above beta,
not what is intended for.

Spotted by Richard Vida.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-31 20:17:37 +01:00
Marco Costalba a94fd3bbec Reformat kpk bitbase
Simplify and streamline the code. Verified all the
resulting bitbases are not changed.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-30 19:19:47 +01:00
Marco Costalba 1a1742ac4f Don't log search info after a stop
Fix an issue where the log file stores an incorrect +0.00
eval after a search has been stopped.

Bug reported by Ajedrecista.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-29 11:25:02 +01:00
Marco Costalba a492a9dd07 Bitwise operator overloads between Bitboard and Square
Yes, we try to be fancy here ;-)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-29 10:59:50 +01:00
Marco Costalba 875a8079bc Replace clear_bit() with xor_bit()
This allows to retire ClearMaskBB[] and use just
one SquareBB[] array to set and clear a bit.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-29 09:38:40 +01:00
Marco Costalba b76c04c097 Rename ValueType to Bound
It is a more conventional and common naming.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-27 19:49:38 +01:00
Marco Costalba f7b4983137 Do not require -lpthread when linking in mingw
With this we should compeltely remove the need
of installing third party POSIX threads library
when compiling with mingw-gcc under Windows.

Spotted by Trung Tu.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-27 19:41:24 +01:00
Marco Costalba eeef654cd7 Restore LMR depth limit
It is not clear the advantage and we don't want
to risk of introducing regressions on this
very critical parameter. So revert to old limit.

After 16003 games
Mod vs Orig 2496 - 2421 - 11086 ELO +1 (+-3)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-27 19:31:41 +01:00
Marco Costalba 7fb6fd2f55 Reformat threads code
Apart from some renaming the biggest change
is the retire of split_point_finished()
replaced by slavesMask flags. As a side
effect we now take also split point lock
when allocation available threads.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-27 19:24:22 +01:00
Marco Costalba a189a5f0c5 Use Windows threads library with mingw
Instead of Posix threads. This seems to fix time
losses of the gcc compiled version for Windows.
The patch replaces the MSVC specific _MSC_VER flag
with _WIN32 and _WIN64 that are defined both by
MSVC and mingw-gcc.

Workaround found by Jim Ablett.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-25 07:32:31 +01:00
Marco Costalba 24b25b4827 Order bad captures by MVV/LVA
Instead of by SEE. Almost no ELO change but it is
a bit easier and is a more natural choice given
that good captures are ordered in the same way.

After 10424 games
Mod vs Orig 1639 - 1604 - 7181 ELO +1 (+-3.8)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-24 07:12:19 +01:00
Marco Costalba 830ff985db Revert "Fix link time optimization gcc option"
It seems we need to pass the full optimization
flags to the linker otherwise we end up in a
slow compile:

http://lists.debian.org/debian-devel/2011/06/msg00181.html

Regression reported by Benigno Hernandez.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-23 20:52:21 +01:00
Marco Costalba 3d937e1e90 Simplify locking usage
pass references (Windows style) instead of
pointers (Posix style) as function arguments.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-23 20:30:19 +01:00
Marco Costalba 04ff9c2548 Simplify our insertion sort implementation
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-23 20:24:15 +01:00
Auguste Pop 28bf56e725 Fix link time optimization gcc option
The previous line, LDFLAGS += $(CXXFLAGS), does not make sense, and
breaks profile-build, thus changing it into: LDFLAGS += -flto.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 21:58:42 +01:00
Marco Costalba 662d1859bd Shrink sequencer table
Integrate TT_MOVE step into the first state. This allows to
avoid the first call to next_phase() in case of a TT move.

And use overflow detection instead of the bunch of STOP_XX
states to detect end of moves.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 16:06:29 +01:00
Marco Costalba 97e0b0a01e Assorted code style in movepicker.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 13:57:24 +01:00
Marco Costalba 04ac1bcabe Fix incorrect assert(PvNode == (alpha != beta - 1))
In case of a PvNode could happen that alpha == beta - 1,
for instance in case the same previous node was visited
with same beta during a non-pv search, the node failed low
and stored beta-1 in TT. Then the node is searched again
in PV mode, TT value beta-1 is retrieved and updates alpha
that now happens to be beta-1.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 11:41:34 +01:00
Marco Costalba ec9b037e5f Order the recaptures by MVV/LVA
Almost no functional change because multiple recaptures
to same square are very rare, but neverthless it seems
the correct thing to do.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 11:19:44 +01:00
Marco Costalba 6f6be95bad Rename NON_CAPTURE to QUIET
It is a more conventional naming and is nicer.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 00:42:59 +01:00
Marco Costalba b96db269a8 Reshuffle stuff in MovePicker
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-22 00:13:47 +01:00
Marco Costalba dcbb05ef39 Fix ss->currentMove when probcutting
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-21 23:30:56 +01:00
Marco Costalba e25de55fac Use an enum instead of a table as MovePicker sequencer
No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-21 16:05:54 +01:00
Marco Costalba 28892666bd Sync generate_direct_checks() with generate_piece_moves()
They are almost the same, if the function arguments would
have been the same they could even be integrated.

Also a bit of renaming while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-21 12:49:37 +01:00
Marco Costalba 6058e7cf60 Triviality in SERIALIZE_PAWNS() macro usage
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-18 17:52:55 +01:00
Justin Blanchard 2a21543c88 Remove unused #include lines 2012-01-19 00:48:53 +08:00
Justin Blanchard 007613cb5e Fix "go nodes", at least when Threads=1 2012-01-19 00:48:53 +08:00
Marco Costalba 1b69ef8e6c Don't allow LMR to fall in qsearch
And increase LMR limit. Tests show no change ELO wise,
but we prefer to take the risk to commit anyhow becuase
is a 'prune reducing' patch.

After 10749 games
Mod vs Orig: 1670 - 1676 - 7403 ELO 0 (+-3.7)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-18 08:10:40 +01:00
Marco Costalba 972dec454c Microptimize generation of pawn evasions
Skip calling promotion generation functions in
the very common case of no possible promotion
evasion. Also retire generate_pawn_captures()

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-18 07:47:42 +01:00
Marco Costalba b6b8c62ba5 Simplify pawn captures generation
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-16 19:29:15 +01:00
Marco Costalba db57b5f8f4 Fix a (bogus) warning with gcc 4.4
Fix an incorrect warning: 'bm may be used uninitialized'
with the old but still commonly used gcc 4.4

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-16 19:26:42 +01:00
Marco Costalba 4a4513d126 Retire double-profile-build Makefile target
Now that we don't support anymore popcount detection
at runtime this target is obsolete.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-15 17:02:06 +01:00
Marco Costalba 2b1324eddb Fix gcc name used in Link Time Optimization
Use $(CXX) instead of assuming compiler name is 'gcc'

Spotted by Louis Zulli.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-15 12:50:09 +01:00
Marco Costalba dfd030b67a Make init_magic() piece agnostic
All the piece dependant data is passed now as
function arguments so that the code is exactly
the same for bishop and rook.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-15 09:21:00 +01:00
Marco Costalba 20621ed3e3 Unify some template specializations
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 23:11:59 +01:00
Marco Costalba 2aec8fb956 Retire queen_attacks_bb()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 23:00:10 +01:00
Marco Costalba 3ec94abcdb Use 'adjacent' instead of 'neighboring'
It is more correct and specific. Another naming
improvement while reading Critter sources.

No functional changes.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 18:04:13 +01:00
Marco Costalba e1907e349f Restore development version
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 15:27:24 +01:00
Marco Costalba f237e8b8ea Stockfish 2.2.2
Dedicated to the new Jim's super fast builds ! :-)

stockfish bench signature is: 5447426

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 15:01:21 +01:00
Marco Costalba 103b368ab7 Move struct RootMove to Search namespace
And directly pass RootMoves instead of SearchMoves
to main thread. A class declaration is better suited
in a header and slims a bit the fatty search.cpp

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 14:22:34 +01:00
Marco Costalba a29dd88f75 Use a set to store SearchMoves
We just need to verify if a legal move is among the
SearchMoves, so we don't need a vector for this.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-14 13:06:01 +01:00
Marco Costalba a1076cc68a Fix a gcc 4.7 warning
New gcc 4.7 complains about casting a volatile pointer
to void* so assign the variables directly.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-13 07:58:21 +01:00
Marco Costalba 78908b7aed Enable easy move detection only for recaptures
It could lead to terrible mistakes otherwise, as
it happened during a game on playchess when on
this position (after white's f4):

2q4r/4b1k1/p3rpp1/3np2p/PpNpNP1P/1P1P2PQ/2P1R3/4R1K1 b - - 0 1

SF moves immediately e5xf4 instead of the correct f5.
In general during engine matches it is impossible the
opponent leaves a piece hanging or anyhow starts a
clear losing sequence. So avoid to fall in subtle traps.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-13 07:27:01 +01:00
Marco Costalba d98150dffc Use operator~ to flip colors and squares
More natural and nicer. Idea from Critter.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-12 19:37:44 +01:00
Marco Costalba c549f71f64 Small touches in FEN decoding
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-12 18:40:58 +01:00
Marco Costalba c19ea4b000 Retire is_mate()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-10 21:53:07 +01:00
Marco Costalba bede30e7a6 Introduce piece_moved() to simplify common code
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-09 22:36:45 +01:00
Marco Costalba b05fbb3733 Unify PseudoAttacks arrays
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-09 22:08:37 +01:00
Marco Costalba 9b43fd7937 Use CheckInfo to generate checks
It should help to avoid recalculating check squares
of sliding attackers for queen when already done for
bishops and rooks. Of course this helps when there are
bishop, rook and queen on the board !

Fixed also a subtle bug (use of same variable b in while
condition and in condition body) introduced recently by
revision d655147e8c that triggers
in case we have at least 2 non-pawn discovered check pieces.
This is very rare that's why didn't show in the node count
verification where we actually have a case of 2 dc pieces
in position 14, but one is a pawn.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-08 16:09:18 +01:00
Marco Costalba 87fc9dcaa3 Add castling to generation of checking moves
During generation of non-captures checks (in qsearch)
we don't consider castling moves that give check,
this patch includes also this rare case. Verified with
perft that all the non-capture checks are now generated.

There should be a very little slowdown due to the extra
work, but actually I failed to measure it. I don't expect
any ELO improvment, there is even no functional change on
the standard depth 12 search, it is just to have a correct
move generator.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-08 12:28:51 +01:00
Marco Costalba 83f3ea7ab4 Last touches to movegen.cpp
The full movegen patch series shows a speed up of almost
6% (!) on perft, and code is much more readable too.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 13:13:28 +01:00
Marco Costalba 9c8c4ff46f Retire the redundant MV_CHECK
MV_CHECK is an alias of the more appropiate named
MV_NON_CAPTURE_CHECK so use only the latter.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 12:03:19 +01:00
Marco Costalba d655147e8c Retire generate_discovered_checks
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 12:03:17 +01:00
Marco Costalba 3c675db3d0 Rearrange pawn moves generation
Functional change due only to rearrangement

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 12:03:16 +01:00
Marco Costalba cf247e7e30 Reshuffle stuff in movegen.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 12:03:15 +01:00
Marco Costalba 0026c88b3a Retire OLD_LOCKS option
And make CRITICAL_SECTION locks the only option for Windows.
This guarantees backward compatibility with all the Windows
versions (even XP and older) and an hassle free experience
when compiling for Windows. Tests performed by Ingo and
reported on talkchess confirm there is no speed penalty
against the most modern SRW locks:

http://www.talkchess.com/forum/viewtopic.php?t=41835&start=20

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 11:18:09 +01:00
Marco Costalba 6482ce2bb2 Fix compile on HP-UX 11's HP's C++
On that platform non-bracketed casting are not supported.

Reported by Richard Lloyd.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-07 00:04:08 +01:00
Marco Costalba 867a5a8cd2 Restore development version
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-06 18:09:25 +01:00
Marco Costalba 78e6b361c5 Stockfish 2.2.1
Hopefully fixed the "lose on time" issue.

stockfish bench signature is: 5457475

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-06 17:01:40 +01:00
Marco Costalba aa392c366e Extra time management safety
Further increase safety against time losses. After this
change (tested on LittleBlitzer and cutechess) I had no
more time losses at 2" and 1"+0.02 TC both on Windows
and Linux on more than 10000 games.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-06 13:24:37 +01:00
Marco Costalba f80c50bcdd Try hard not to lose on time
We try hard not to lose on time even under extreme
time pressure. We achieve this through 3 different but
coordinated steps:

    1) Increase max frequency of timer events

    2) Quickly return after a stop signal

    3) Take in account timer resolution

With these SF played under LittleBlitzer at 1"+0.02 and 3"+0
without losing on time even one game.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-06 01:41:45 +01:00
Marco Costalba d282cf6964 Avoid a race at thread creation
Before creating main thread we set its do_sleep flag to true,
then thread is created and it will go to sleep in main_loop()
after resetting do_sleep.

But if after the setting of do_sleep and before its resetting
the UI thread calls start_thinking() it will not wait on:

  if (!asyncMode)
      while (!main.do_sleep)
          cond_wait(&sleepCond, &main.sleepLock);

as it should but will immediately return before the main thread has
started the search. This very rare race show itself during bench,
when the first position is erroneusly skipped so that bench node count
results of 5309038 instead of the correct 5457475.

The patch is somewhat tricky, but is simple and it works!

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-03 21:31:50 +01:00
Marco Costalba b1fcfe4c5d Streamline generation of MV_NON_EVASION
Small speed-up of 3% in perft.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-03 19:18:38 +01:00
Marco Costalba 30418a3cfc Fix a warning under gcc
Locals left and right shadow two same named
variables in the std::ifstream base class.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-03 19:01:47 +01:00
Marco Costalba cb1709ef5e Revert cond_signal() fix
It seems it yields to missing wake-up events with the
result of SF loosing on time as reported by many people.

So revert the patch and use a more robust approach: assume
there can be spurious wake ups events and make the code to
work also in those cases.

While debugging I found that WaitForSingleObject() had wrong
parameter 0 instead of INFINITE yielding to a crash while
exiting under Windows, strangely unnoticed til now.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-02 16:44:04 +01:00
Marco Costalba 67338e6f32 Big renaming of move's helpers
The aim is to have shorter names without losing
readibility but, if possible, increasing it.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-02 12:03:54 +01:00
Marco Costalba 8300ab149c Simplify Book APIs
Retire open(), close() and name() from public visibility
and greately simplify the code. It is amazing how much
can be squeezed out of an already mature code !

No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2012-01-01 14:46:18 +01:00
Marco Costalba c00443b19e Restore development version
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-31 15:44:00 +01:00
Marco Costalba 9db9e4f7d3 Stockfish 2.2
stockfish bench signature is: 5457475

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-31 15:43:28 +01:00
Marco Costalba 22e40c8c10 Fix cond_signal() semantics when using OLD_LOCKS
In Windows when OLD_LOCKS is defined we use SetEvent() to mimic
the semantic of the POSIX pthread_cond_signal().

Unfortunatly there is not a direct mapping because with SetEvent()
the state of an event object remains signaled until it is set
explicitly to the nonsignaled state or until a single waiting thread
has been released. Instead in case of pthread_cond_signal(), if there
are no waiting threads it has no effect. What we may want is something
like PulseEvent() instead of SetEvent(). Unfortunatly it is documented
by Mcrosoft as 'unreliable' due to spurious wakes up that could
filter out the signal resetting. So we opt to reset manually any
pending signaled state before to go to sleep.

This fixes the strange misbehaves during 'stockfish bench'
when using OLD_LOCKS under Windows.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-31 15:40:12 +01:00
Marco Costalba e9296d694c Unify BitCountType selection
Now that HasPopCnt is a compile time constant we can
centralize and unify the BitCountType selection.

Also rename count_1s() in the more standard popcount()

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-31 10:46:14 +01:00
Marco Costalba fb4b4f772e Fix Windows 64 build
Broken by previous patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-31 08:48:16 +01:00
Marco Costalba f4dadee5e2 Reformat types.h
Retire obsolete code and reshuffle stuff.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 19:13:44 +01:00
Marco Costalba 808a312e1c Simplify debug functions
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 16:23:11 +01:00
Marco Costalba 93e539d909 Assorted cleanups in benchmark.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 15:23:33 +01:00
Marco Costalba 9d7a36121a Retire RootMove::nodes
Was used for time management but is no more used
today.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 14:16:15 +01:00
Marco Costalba 8307da0de7 Update copyright year to 2012
And refresh Readme.txt while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 13:52:16 +01:00
Marco Costalba ad43ce1436 Simplify printing of engine info
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 13:19:20 +01:00
Marco Costalba 0a6532a39d Retire run-time detection of hardware POPCNT
It was meant to build a single binary optimized
for any kind of CPU: with and without hardware POPCNT.

This is a nice idea but in practice was never used, or
people builds binary with popcnt enabled or not, mainly
according to their type of CPU. And it was also never
used in the official Jim's builds where, in case, would
be easier for a number of reasons, do build two different
versions: with and without SEE42 support.

So retire this feature and simplify the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 12:17:03 +01:00
Marco Costalba 20a6f99cdb Fix an off-by-one bug in ucioption.cpp
Harmless but anyhow wrong.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 11:58:55 +01:00
Marco Costalba 4554d8b2ac Better use STL algorithms in Endgame functions
This leads to a further and unexpected simplification
of this already very size optimized code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-30 11:58:54 +01:00
Marco Costalba 9cb187762a Wait for main thread to finish before to exit
Currently after a 'quit' command UI thread raises stop
signal, exits from uci_loop() and calls Threads.exit()
while the search threads are still active.

In Threads.exit() main thread is asked to terminate, but
if it is parked in idle_loop() it will exit and free its
resources (in particular the shared Movepicker object) while
sibling slaves are still active and this leads to a crash.

The fix is to let the UI thread always wait for main thread
to finish the search before to return from uci_loop().

Found by Valgrind when running with 8 threads.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-29 10:33:06 +01:00
Marco Costalba 4a8c1b2470 Use for_each() in Endgames d'tor
And fix some comments while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-29 10:25:11 +01:00
Marco Costalba 0759d8f430 Add user-defined conversions to UCIOption
Greatly improves the usage. User defined conversions
are a novelity for SF, another amazing C++ facility
at work !

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-28 19:42:50 +01:00
Marco Costalba ae65ab25d5 Fix score_to_uci()
The condition for a mate score was wrong:

abs(v) < VALUE_MATE - PLY_MAX * ONE_PLY

instead of

abs(v) < VALUE_MATE_IN_PLY_MAX

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-28 14:20:24 +01:00
Marco Costalba 24417a6cd9 Better document how mate scores are stored in TT
During the search we score a mate as "plies to mate
from the root" to compare in an homogeneous way the
values returned by different sub-trees. However we
store in TT a mate score as "plies to mate from the
current position" the let the TT value remain valid
across the game.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-28 13:33:31 +01:00
Marco Costalba ad4739a6d4 Retire SquaresByColorBB[] and enum SquareColor
Use same_color_squares() instead.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-28 10:57:08 +01:00
Marco Costalba a695ed65a8 Rename Pieces
Align with PieceType naming convention and
make them more readable.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-28 10:38:48 +01:00
Marco Costalba 750ac9ac50 Document mate distance pruning
It is simple but somewhat tricky code that deserves
a bit of documentation. A bit of renaming while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-27 18:53:00 +01:00
Marco Costalba 07f3f0384a Assert enhancements in search
Add the check that alpha < beta - 1 if and only if PvNode is true.
The current code would not flag PvNode and alpha == beta - 1. In
other words, the || is not an exclusive OR!.

Also sync assert conditions of search() and qsearch()

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-27 01:04:00 +01:00
Marco Costalba 87b483f999 Reformat UCI option code
Make a better use of C++ operators overloading to
streamline the APIs.

Also sync polyglot.ini file while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-27 00:56:11 +01:00
Marco Costalba c2d42ea833 Rename getters functions removing 'get_' prefix
Follow the suggested Qt style:

http://doc.qt.nokia.com/qq/qq13-apis.html

It seems to me simpler and easier to read.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-25 11:50:59 +01:00
Marco Costalba b5f6c2241b Restore std::cout instead of printf()
I am not able to reproduce the speed regression anymore,
and also we were using cout even before speed regression
so probably the reason is not there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-25 09:29:51 +01:00
Marco Costalba e89a8e0913 Correctly define operators in types.h
Be consistent with the way these operators are defined
in plain C (and in C++).

Spotted by Lucas Braesch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-24 09:58:46 +01:00
Marco Costalba d543a64cc7 Don't update killers for evasions
We don't use killers to order evasions, so it
seems natural do not consider an evasion cut-off
move as a possible killer. Test shows almost no
change, as it should be becuase this is a really
tiny change, but neverthless seems the correct
thing to do.

After 11893 games
Mod vs Orig 1773 - 1696 - 8424 ELO +2 (+-3.4)

Idea from Critter.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-24 09:25:20 +01:00
Marco Costalba 939b621e5c Use ADL to skip std:: qualifier
Take advantage of argument-dependent lookup (ADL) to
avoid specifying std:: qualifier in some STL functions.
When a function argument refers to a namespace (in this
case std) then the compiler will search the unqualified
function in that namespace too.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-18 21:18:51 +01:00
Marco Costalba a77a3b723f Disable again buffering at startup
Partially revert efd2167998
Without this patch SF does not send "bestmove" to GUI.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-17 16:59:32 +01:00
Marco Costalba 976270916b Headers cleanup in ucioption.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-17 15:58:19 +01:00
Marco Costalba 87f7e99bc4 Use printf() instead of std::cout()
Seems sensibly faster: On a

./stockfish bench > /dev/null

We have +2% on mingw and even +5% on MSVC !

Also removed the nice but complex enum set960 machinery,
use directly the underlying move_to_uci() function.

Speed regression reported by Heinz van Saanen.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-17 15:50:06 +01:00
Marco Costalba 72d8d27234 Retire update_history() Inline the only caller site
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-16 19:17:47 +01:00
Marco Costalba 1ae6ae9b60 Fix book move with searchmoves compatibility
Do not return the book move if is not among the
RootMoves,in particular if we have been asked to
search on a move subset with "searchmoves" then
return book move only if it is among this subset.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-16 07:23:24 +01:00
Marco Costalba af4fadebda Simplify id_loop() signature
And related assorted cleanup of this
very important function.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-14 21:56:37 +01:00
Marco Costalba 0af3af5d25 Reformat sending of PV information
Introduce pv_info_to_log() and pv_info_to_uci() and
greatly cleanup this stuff.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-14 19:22:22 +01:00
Marco Costalba 852d45a600 Further simplify aspiration code
Actually after last patch it happens that delta
starts always with the fixed value of 16.

So further remove useless code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-14 06:18:33 +01:00
Marco Costalba 96213689e4 Simplify aspiration window calculation
It seems that we just need to look at previous score to
compute aspiration window size.

After 5350 games:
Mod vs Orig 800 - 803  - 3647 ELO +0 (+- 5.2)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-14 05:56:53 +01:00
Marco Costalba 4e59c5c274 Retire RootMoveList
Diretcly use the underlying std::vector<Move> and the
STL algorithms. Also a bit of cleanup while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-12 20:13:24 +01:00
Marco Costalba 7d97ebfe7f Fix another crash triggered by previous patch
It is ok to redirect st pointer to startState, but the latter
should be updated with the content pointed by the st of the
original position. The bug is hidden when startState and *st
are the same as is the case of searching from start position,
but as soon as moves are made (as is the case when splitting)
the bug leads to a crash.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-11 11:44:09 +01:00
Marco Costalba 377c406c74 Fix a crash when quitting while searching
The Position object used by UI thread is a local variable in
uci_loop(), so after receiving "quit" command the function
returns and the position is freed from the stack.

This should not be a problem becuase in start_thinking() we copy
the position to RootPosition that is the one used by main search
thread. Unfortunatly we blindly copy also StateInfo pointer that
still points to the startState struct inside UI position. So the
pointer becomes stale as soon as UI thread leaves uci_loop() and
because this happens while main search thread is still recovering
after the 'stop' signal we have a crash.

The fix is to update the pointer to the correct startState after
the copy.

Found with Valgrind.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-11 10:31:46 +01:00
Marco Costalba 91601d7f95 Prune silly comments in search()
Comments should be informative but not pedantic / obvious.
The only exception is the function description where we
indulge a bit on the "chatty" side, but has always been like
this since Glaurung times, so we continue with this tradition.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-10 19:14:49 +01:00
Marco Costalba f3728e66f8 Allow to prune also first move
Tested togheter with previous patch; shows no regression and
is a semplification.

After 5817 games:
Mod vs Orig 939 - 892 - 3986 ELO +2 (+- 5.1)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-10 17:35:37 +01:00
Marco Costalba 3f14ed6602 Don't update bestValue when pruning
Simply return a fail-low score

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-10 17:35:36 +01:00
Marco Costalba 14df99130f Fix description of endgame scaling functions
Triggered by a comment of Eelco on talkchess. Also
a bit of cleanup while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-10 12:40:05 +01:00
Marco Costalba 47bcb892af Fix compile for 64 bits
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-08 13:55:27 +01:00
Marco Costalba 6e05055f06 Set captureThreshold according to static evaluation
Consider negative captures as good if
still enough to reach beta.

After 7502 games:
Mod vs Orig 1225 - 1158 - 5119 ELO +3 (+- 4.5)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-08 13:48:18 +01:00
Marco Costalba da6e53a436 Remove some (int) casts
A cast rarely is the right solution. In this case was enough
to redifine 3 variables with type size_t instead of int

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-07 17:30:28 +01:00
Marco Costalba 56774fff20 Retire all extensions (but checks) for non-PV nodes
It seems we don't have any added value. Note that the
moves that were used to be extended are still flagged
as dangerous so to avoid at least pruning them.

After 9555 games
Mod vs Orig 1562 - 1540 - 6453 ELO +0 (+- 4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-07 07:49:24 +01:00
Marco Costalba 3527a42063 Prefer empty() to size()
As Heinz says:

"Function empty() should have a constant run-time even
 on lousy compilers and you spare the not.

The change is even measurable: + 100-150 nodes/sec. Wow:-)"

Patch from Heinz van Saanen

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-06 19:41:48 +01:00
Marco Costalba 98352a5e84 Use operator() instead of apply() in endgames
It is more idiomatic for a functor (a function object) as are
the endgames.

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-05 21:04:32 +01:00
Marco Costalba 11a7980976 Fix disambiguation bug in move_to_san()
A pinned piece cannot move and so does not play any role
in SAN disambiguation.

Reported by Steven Edwards.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-04 12:24:15 +01:00
Marco Costalba 5b8ca1eee7 Move SearchStack under Search namespace
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-04 12:03:59 +01:00
Marco Costalba 81cd417b45 Retire move.h
Also some assorted comments fixes and other trivia.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-04 11:36:03 +01:00
Marco Costalba a44c5cf4f7 Prefer 0 to EmptyBoardBB
Easier and even faster or at least easier to optimize.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-03 12:02:13 +01:00
Marco Costalba 5c5af4fa65 Retire neighboring_files_bb() overload
Rarely used and we prefer to not hide the complexity.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-03 11:58:55 +01:00
Marco Costalba efd2167998 Don't disable IO buffering at startup
It was never clear to me why we needed this trick, and now
that we rely only on C++ std::getline() and std::cout for
input / output it is even more a mistery what this code does.

So disable it and wait to see if someone screams ;-)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-03 10:54:44 +01:00
Marco Costalba 348f824104 Tidy up comments in uci.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-03 10:48:31 +01:00
Marco Costalba 0f7cbaca75 Tidy up comments in thread.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-12-03 09:58:46 +01:00
Marco Costalba e870afa5d5 Include <cstring> in search.h
Now we use memset() directly there.

Spotted by Justin Blanchard.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-28 07:30:31 +01:00
Marco Costalba 9bacd921fa Little reformat of elapsed_search_time()
Change name and argument type.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-27 22:36:13 +01:00
Marco Costalba bb3427ca85 Detach search arguments from UI thread
Detach from the UI thread the input arguments used by
the search threads so that the UI thread is able to receive
and process any command sent by the GUI while other threads
keep searching.

With this patch there is no more need to block the UI
thread after a "stop", so it is a more reliable and
robust solution than the previous patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-27 17:46:18 +01:00
Marco Costalba 6809b57cfc After a "stop" do not read new input until search finishes
Unfortunatly xboard sends immediately the new position to
search after sending "stop" when we have a ponder miss.

Becuase main thread position is not copied but is referenced
directly from root position and the latter is modified by
the "position.." UCI command we end up with the working position
that changes under our feet while the search is still recovering
after the "stop" and this causes a crash.

This happens only with the (broken) xboard, native UCI does not
have this problem.

Reported by otello1984

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-27 12:19:33 +01:00
Marco Costalba ffa75215cc Fix a race in pondering mode
Fixes an hang when playing with ponder ON. Perhaps there is still
a very small race but now it seems engine does not hang anymore.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-26 18:29:40 +01:00
Marco Costalba c4517c013c Introduce Search namespace
Move global search-related variables under "Search" namespace.

As a side effect we can move uci_async_command() and
wait_for_stop_or_ponderhit() away from search.cpp

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-26 13:43:22 +01:00
Marco Costalba ed04c010eb Rewrite async I/O
Use the starting thread to wait for GUI input and instead use
the other threads to search. The consequence is that now think()
is alwasy started on a differnt thread than the caller that
returns immediately waiting for input. This reformat greatly
simplifies the code and is more in line with the common way
to implement this feature.

As a side effect now we don't need anymore Makefile tricks
with sleep() to allow profile builds.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-26 11:51:12 +01:00
Marco Costalba e9dc2e9e1e Reformat search dispatch code
Reduce indentation level and lines of code and tidy up
some comment.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-20 00:02:13 +01:00
Marco Costalba 9c7d72739c Fix regression with printing of debug info
Output of debug info each second was disabled
due to recent patches.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-19 14:37:57 +01:00
Marco Costalba 3fc08f8ab6 Don't check for early stop when StopOnPonderhit is set
If we are pondering we will stop the search only when
GUI sends "ponderhit" or "stop" commands or when we reach
maximum depth. In all the other cases we continue to search
so there is no need to verify for available time.

Also better clarify why wait_for_stop_or_ponderhit() before
to exit in some cases.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-19 12:04:42 +01:00
Marco Costalba c56a7ee803 Early stop: retire redundant Rml.size() == 1 case
In case there is only 1 legal move we will stop the
search at depth 10 anyway because the exclusion search
probe will fail low.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-19 11:19:12 +01:00
Marco Costalba 35fa4fc71f Rewrite early stop logic
In the "easy move" paradigm we verify if the best move has
ever changed over a good number of iterations and if the
biggest part of the searched nodes are consumed on it.
This is a kind of hacky ad indirect heuristic to deduce
if one move is much better than others.

Rewrite the early stop condition to verify directly if one
move is much better than others performing an exclusion
search.

Idea to use exclusion search for time management if form Critter.

After 12245 games at 30"+0.1
Mod vs Orig 1776 - 1775 - 8694 ELO +0 (+-3.4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-18 21:01:34 +01:00
Marco Costalba b56a098cfb CLOP: Passed pawns weights tuning
Tuned with CLOP against a pool of 3 engines. Result
verified with a direct match:

After 11720 games at 10"+0.1
Mod vs Orig 1922 - 1832 - 7966 ELO +2 (+-3.6)

So no change in self match but if CLOP is right it should
be a bit better against an engine pool.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-14 09:05:20 +01:00
Marco Costalba 4cc272cb94 Rename value in bestValue in id_loop()
The value returned by root search it is actually
our best value, so rename the variable to reflect this.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-13 10:54:30 +01:00
Marco Costalba c4fc82c6b7 Rewrite link time optimization in Makefile
Instead of binding link time optimization to the choice of
popcount support, do the right thing and add -flto option
when gcc 4.5 or later is detected.

Although it should be supported also under mingw, it happens
that it doesn't, at least on my 4.6.1 due to some known bugs.

Thanks to Mike for helping me with this patch.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-12 11:55:45 +01:00
Marco Costalba a40ded2884 Simplify passed pawns logic
Remove the bonus for no *friendly* pieces in the pawn's path and
reduce a bit the bonus based on kings proximity.

This patch is part of to the ongoing effort to remove form evaluation
all the terms that do not add value.

After 16284 games:

Mod vs Orig 2728 - 2651 - 10911 ELO +1 (+- 3.1)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-12 09:09:41 +01:00
Marco Costalba 44c78fdb7a Fix regression: engine hangs while pondering
After a "stop" due to a ponder miss Xboard sends
immediately the new position to search, without
waiting for engine to effectively stop the search.

It is not clear if this is a GUI bug (as I suspect)
or allowed behaviour, but because it won't be fixed
anyway workaround this issue making listener thread
to switch to in-sync mode as soon as a "stop" command
is received.

Thanks to Mike Whiteley for reporting this.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-10 15:08:53 +01:00
Marco Costalba ecb98a3330 Stop is not an unknown command
If GUI sends stop while we are waiting for
a command do not reply with a silly:

Unknown command: stop

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-10 15:06:28 +01:00
Marco Costalba 9b6b510ca8 Another fix to profile-build on gcc 4.6
Oliver reports profile builds error with new gcc 4.6, he says:

"We need to add -lgov with profile-generate AND profile-use.
So it has to be added to the second stage of building too.

The problem occurred first with the introduction of gcc4.6 and
I think this is because the previous version did find the gcov
library automatically. gcc4.6 needs more precise options and
does less guesses. I have seen it in debian, Ubuntu and also with
mingw on Windows. And all use gcc4.6."

This patch fixes the issue.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-09 05:20:35 +01:00
Marco Costalba c2600a73cf Fix profile-build
After async I/O patches 'bench' changed behaviour and now waits for
input at the end of the test run. This is due to listener thread stay
blocked on std::getline() even after test run is finished, as soon as
we feed something the thread unblocks and then quickly exits.

This is not a big problem, but has the bad side effect of breaking
profile builds that hang forever at the end of the test run.

The tricky workaround is to create a pipe that connects to stockfish
input and then, when test run is finished, breaking the pipe: this
makes std::getline() immediately return.

So this patch adds a 'sleep 10' piped into 'stockfish bench' test run
command. After 10 seconds sleep ends, the pipe breaks and 'bench'
finishes as usual.

Thanks to Oliver Korff for reporting the issue, and to Mike Whiteley
for having co-authored this solution.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-08 07:39:06 +01:00
Marco Costalba 43204d9ac2 Reformat all_slaves_finished()
Rename and move under ThreadsManager class.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-06 13:45:47 +01:00
Marco Costalba 369789b426 Better document and reshuffle stuff in think()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-06 13:32:25 +01:00
Marco Costalba 8fa53a5b92 Better define wait_for_stop_or_ponderhit()
Use do_uci_async_cmd() instead of process input commands
directly and clarify that what we are waiting for is
something that is able to raise StopRequest flag.

Also fix some stale comments in do_uci_async_cmd(). Here
we need to reset Limits.ponder only upon receiving "ponderhit".
In the case of "quit" or "stop" resetting Limits.ponder has no
effect because the search is going to be stopped anyway.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-06 12:48:16 +01:00
Marco Costalba d58176bfea Use a timer to avoid polling
The timer will be fired asynchronously to handle
time management flags, while other threads are
searching.

This implementation uses a thread waiting on a
timed condition variable instead of real timers.
This approach allow to reduce platform dependant
code to a minimum and also is the most portable given
that timers libraries are very different among platforms
and also the best ones are not compatible with olds
Windows.

Also retire the now unused polling code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-05 18:19:38 +01:00
Marco Costalba 0095f423f2 Retire now unused input_available()
With our new listener thread we don't need anymore
this ugly and platform dependent code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-05 08:43:44 +01:00
Marco Costalba 2617aa415e Rewrite how commands from GUI are read
Instead of polling for input use a dedicated listener
thread to read commands from the GUI independently
from other threads.

To do this properly we have to delegate to the listener
all the reading from the GUI: while searching but also
while waiting for a command, like in std::getline().

So we have two possible behaviours: in-sync mode, in which
the thread mimics std::getline() and the caller blocks until
something is read from GUI, and async mode where the listener
continuously reads and processes GUI commands while other
threads are searching.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-05 08:35:17 +01:00
Marco Costalba 22b9307aba Further touches to magic bitboards code
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-11-01 10:38:01 +01:00
Marco Costalba ac7339877b Fix compile error in cpu_count()
The std::min() template function requires both arguments
to be of the same type.

But here we have the integer MAX_THREADS compared to a long:

long sysconf(int name);

So cast to integer and fix the compile.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-31 20:03:30 +01:00
Marco Costalba 90890844ad Document magics bitboards code
Add comments and rename stuff to better clarify what the
magic bitboard initialization code does.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-31 18:57:13 +01:00
mcostalba 8a89b12641 Merge pull request #1 from Panthee/master
Code Cleanup - Replacing macros Min() and Max() with corresponding STL algorithms std::min() and std::max()
2011-10-31 05:32:18 -07:00
Alexander Kure 4a3b162c8c Retire update_gains() 2011-10-31 03:28:59 -04:00
Alexander Kure 5c8af7ccb8 Replaced macros Min() and Max() with corresponding STL algorithms std::min() and std::max() 2011-10-31 00:38:44 -04:00
Marco Costalba 7942e6f3bf Retire update_gains()
Called from one place only.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 23:05:31 +01:00
Marco Costalba c6f497f09d Fix small bug in move_attacks_square()
We test if the piece moved in 'to' attacks the square 's' with:

bit_is_set(attacks_from(piece, to), s))

But we should instead consider the new occupancy, changed after
the piece is moved, and so test with:

bit_is_set(attacks_from(piece, to, occ), s))

Otherwise we can miss some cases, for instance a queen in b1 that
moves in c1 is not detected to attack a1 while instead she does.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 19:31:50 +01:00
Marco Costalba 29be28e1a2 Inline pinned_pieces() and discovered_check_candidates()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 18:48:23 +01:00
Marco Costalba e7939f450f Code style and 80 chars cols in Position::from_fen()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 18:02:18 +01:00
Marco Costalba 81801d395f Sync do_move() and undo_move()
It is not possible to unify due to the fact that the
sequence steps are reversed. What we can do is to try
to sync comments and code as much as we can to easy
reading and documentation.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 13:58:44 +01:00
Marco Costalba bc76c62c63 Explicitly use a dedicated bitboard for occupied squares
Instead of byTypeBB[0]. This better self-documents the code.

No functional and speed change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 13:51:20 +01:00
Marco Costalba fd5d6c5340 Retire do_capture_move()
It is called only in do_move() that now has been fully
expanded. This is the most time consuming function of
all the engine.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-30 10:59:55 +01:00
Marco Costalba 08abe8b4a3 Retire undo_null_move()
Use a templetized do_null_move() to do/undo the null move.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-29 18:01:56 +01:00
Marco Costalba e896368496 Retire undo_castle_move()
Use a templetized do_castle_move() to do/undo the castling.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-29 17:31:15 +01:00
Marco Costalba 2fe4e10b0b Retire Position::set_castling_rights()
Is called in just one place. And rename set_castle() in the
now free to use and more appropiate set_castle_right().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-29 13:43:44 +01:00
Marco Costalba f2e78d9f84 Retire PieceValueXXX[] getters
They don't add any value given that the corresponding
table has global visibility anyhow.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-22 16:06:59 +01:00
Marco Costalba b5bbc1f713 Simplify the promotion case of move_gives_check()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-22 15:01:21 +01:00
Marco Costalba 23943208ec Remove redundancy in definitions of attack helpers
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-22 14:25:53 +01:00
Marco Costalba c555e1aa96 Convert PST tables to relative values
This is a prerequisite to allow changing piece values
at runtime, needed for tuning.

Also use scores instead of separated midgame and endgame values.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-22 13:17:24 +01:00
Marco Costalba 36cc214703 Increase Mobility
First tuning with CLOP against a pool of 3 engines. Result
verified with a direct match:

After 8736 games at 10"+0.1
Mod vs Orig 1470 - 1496 - 5770 ELO -1 (+-4.3)

So no change in self match but if CLOP is right it should
be a bit better against an engine pool.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-20 07:06:05 +01:00
Marco Costalba 8f7d36a85d Better document mate and stalemate detection
In particular add that we can have an harmless false positive
in case StopRequest or thread.cutoff_occurred() are set.

Reported by David Lee.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-18 20:01:21 +01:00
Marco Costalba e7cfe42d3f Use newly added log facility instead of LogFile
As a side effect now log file is open and closed every
time it is used instead of remaining open for the whole
thinking time.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-17 21:07:14 +01:00
Marco Costalba 500fff920b Add basic log facility
Mainly used to log stuff to a file while playing, when
stdout is used for the comunication with the GUI.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-17 21:07:00 +01:00
Marco Costalba 782c3f36cc Fix compile error in debug mode
Build broken by commit 3141490374
where we renamed move_is_ok() in is_ok() and this clashes
with the same named method in Position that overrides the
move's one causing compile errors.

The fix is to rename the method in Position.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-16 23:56:25 +01:00
Marco Costalba 8272dcb6cd Link Time Optimization doesn't needs -static
Justin reports that it breaks the compilation on Fedore 15 and as Tom says:

-static is only needed to work around the gcc on ubuntu 11.10 beta bug.
If -static introduces issues on its own then it is better to remove it.
It will not be needed in most environments.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-09 08:24:56 +01:00
Marco Costalba fec623d68d Better document how MultiPV search works
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-08 12:11:53 +01:00
Marco Costalba a3bf09c5c9 Send again all the PV lines in multiPV searching
Partially revert 1036cadcec because UCI protocol
in case of multipv explicitly requires:

for the best move/pv add "multipv 1" in the string when you send the pv.
in k-best mode always send all k variants in k strings together.

Thanks to Justin Blanchard for pointing this out.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-08 10:24:54 +01:00
Marco Costalba 325dedc7da Added gcc -msse3 support
It is enabled when selecting x86-64-modern target, this gives
another nice speed up:

On a Core i5-2500 (3300 Mhz, Sandy Bridge):

64 bit download version: 1597151 n/s

-flto : 1659664 n/s

-flto -msse3: 1732344 n/s

Patch suggested by Tom Vijlbrief.

Also unify flto, popcount and msse3 optimization under "modern"
target, note that this can break the "modern" build on old gcc that
do not support -flto option: in this case update gcc ;-) or default
to the standard build.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-08 08:50:11 +01:00
Tom Vijlbrief 09b5949cd2 Added gcc lto (Link Time Optimization) option
Just by adding the -flto option to CXXFLAGS link command
we can gain a few percent in speed.

On a Core i5-2500 (3300 Mhz, Sandy Bridge):

64 bit download version:

Without -flto: 1597151 n/s
With -flto : 1659664 n/s

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-06 20:10:12 +01:00
Marco Costalba 3141490374 Shrink names of move helpers
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-03 14:18:57 +01:00
Marco Costalba 1a8e3f0b2e Small touches in position.h
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-03 14:18:56 +01:00
Marco Costalba 80dd90f972 Small touches to book.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-03 14:18:30 +01:00
Marco Costalba c2c185423b Better naming borrowed from Critter
In line with http://chessprogramming.wikispaces.com conventions.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-02 10:16:59 +01:00
Marco Costalba 25b4d0c127 Revert "Retire Rml full PV search at depth == 1"
Yet another random crash source !

Hopefully this is the last one.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-02 07:40:46 +01:00
Marco Costalba 2225e3bbe7 Rename kingZone[] and reverse the king's color
Seems easier to understand to me. From Critter.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-02 06:54:57 +01:00
Marco Costalba a90a990118 Document why Book is defined static
This was not clear to someone on talkchess and actually
is not trivial to understand.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-02 06:54:57 +01:00
Marco Costalba 83d8fe2d59 When exiting wake up all threads at once
It seems we have a very rare crash under Linux, once
every 10K games without this patch.

Is faster to wake up all the threads, especially on SMP,
where the threads can then exit in parallel while the main
thread is waiting for the next one to terminate.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-10-02 06:54:43 +01:00
Marco Costalba 63a04134d0 No need to test for MOVE_NONE before move_is_ok()
Function move_is_ok() already catches the move == MOVE_NONE case.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-25 16:30:29 +01:00
Marco Costalba 6bc16f3ff1 Update killers after a TT hit
Almost no increase but seems the logic thing to do.

After 16707 games 2771 - 2595 - 11341 ELO +3 (+- 3.2)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-24 17:20:35 +01:00
Marco Costalba 10b24af98a Correctly score capture underpromotions
Be sure a queen capture promotion is tried in front of
an underpromotion.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-24 17:13:14 +01:00
Marco Costalba be9aba2fa0 Don't lock before check for termination
Restore old locking scheme changed with
commit 1e92df6b20.

This seems to prevent a very rare crash that occurs
once every 5-10K games.

With this patch we have no crashes after 33K games.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-24 17:12:36 +01:00
Marco Costalba 35018fa307 Use the map type template parameter to access map()
It is more natural than using the family subtype and also
use two single maps instead of a std::pair.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-11 11:38:23 +01:00
Marco Costalba b706165527 Lookup square distance instead of calculate on the fly
Microptimization that gives a +0.7% speed increase.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-09-11 10:11:43 +01:00
Marco Costalba 6963c3802d Detect family type of endgame from its enum value
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-22 00:49:47 +01:00
Marco Costalba 3b67636f0e Indulge a bit on the template wizardy
Push the template pedal a bit in our "showoff" endgame code ;-)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-14 13:15:58 +01:00
Marco Costalba 48e39c5c8e Small simplification of endgame functions API
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-14 12:12:14 +01:00
Marco Costalba 13524bea9b Fix use of uninitialized variable
When initializing endgames map we build a faked FEN string
in mat_key() to get the position hash's key.

This fen string lacks full move numbers, so when parsing the
fen in Position::from_fen() we leave startPosPly un-initialized.

Spotted by Valgrind (this is a kind of bug that is almost impossible
for humans to find).

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-12 11:56:11 +01:00
Marco Costalba 500c7f44ab Fix silly icc remark #2259
Another stupid remark to quiet out:

remark #2259: non-pointer conversion from "int" to "UINT16={unsigned short}"
may lose significant bits

In this case icc always converts to an integer the result of a shift operation
if the bit size of the operand is smaller, hence the warning when assignin
back to n.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-12 11:54:54 +01:00
Marco Costalba d156e7a20b Use a boolean instead as thread's state
Now that we have just two mutually exclusive thread's states
we can repleace them by a simple boolean.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-10 12:26:52 +01:00
Marco Costalba c386ce0023 Remove Thread::WORKISWAITING
Set the state directly to Thread::SEARCHING

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-10 12:05:33 +01:00
Marco Costalba b69d9ee3f7 Don't need pthread_detach() after pthread_join()
Spotted by Joona and verified with Valgrind.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-10 08:17:15 +01:00
Marco Costalba 7d5b8fcf77 Change start_routine argument
Directly pass the thread pointer.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-09 14:19:44 +01:00
Marco Costalba bc6a6e04a0 Retire Rml full PV search at depth == 1
Now that Rml ordering is based on normal MovePicker logic,
apart for the ttMove that is given, we can avoid to score
all the root moves at depth 1. We only need it for easy move
detection logic, but in this case we just need to score the
first two best moves and not all the Rml set.

No regression after 6400 games
Mod vs Orig 1052 1012 4336 ELO +2 (+- 4.9)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-09 09:58:10 +01:00
Marco Costalba e5ffe9959c Retire ThreadsManager::init_hash_tables()
Allocation of pawn and material hash tables should
be strictly bounded to the change of the number of
activeThreads, so move the code inside set_size().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 23:08:34 +01:00
Marco Costalba 86b95f2105 Retire Thread::TERMINATED
Use proper way to detect for thread terimnation instead of
our homegrown flag.

It adds more code than it removes and adds also platform specific
code, neverthless I think is the way to go becuase Thread::TERMINATED
flag is intrinsecly racy given that when we raise it thread is still
_not_ terminated nor it can be, and also we don't want to reinvent
the (broken) wheel of thread termination detection when there is
already available the proper one.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 22:10:01 +01:00
Marco Costalba 1e92df6b20 Retire Thread::INITIALIZING
Was used to prevent issues when creating multiple threads
on Windows, but now it seems we can remove it safely.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 14:12:02 +01:00
Marco Costalba ba85c59d96 Move idle_loop() under Thread
This greatly removes clutter from the difficult idle_loop() function

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 13:23:03 +01:00
Marco Costalba dafd5b5864 Tidy up comments in thread.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 11:49:58 +01:00
Marco Costalba eabba1119f Retire broken SendSearchedNodes
Now that we can split at root it happens that SendSearchedNodes
works only once at the end of the iteration, but this is useless
becuase speed info is sent anyhow toghter with the pv line.

So retire for now, waiting to find something SMP compatible.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 10:08:41 +01:00
Marco Costalba 06e0d48794 Sync search() and qsearch() alpha update
Change qsearch() to reflect alpha update logic
of search().

To be consistent changed also moves loop condition and
futility pruning condition.

No regression after 5072 games at TC 10"+0.1

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 07:46:01 +01:00
Marco Costalba 4a71c86270 Retire Thread::BOOKED
Start a slave as soon as is allocated.

No functional change with faked split.

Regression tested the full split() series and after
2000 games no regression and no crash.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-08 07:45:24 +01:00
Marco Costalba 68583f4d48 Fix obey the "maxThreadsPerSplitPoint" setting
Spotted by Joona.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 15:15:19 +01:00
Marco Costalba 5b35c149e8 Do not modify alpha in split()
When calling split or we immediately return because unable to
find available slaves, or we start searching on _all_ the moves
of the node or until a cut-off occurs, so that when returning
from split we immediately leave the moves loop.

Because of this we don't need to change alpha inside split() and
we can use a signature similar to search() so to better clarify
that split() is actually a search on the remaining node's moves.

No functional change with faked split.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 15:10:41 +01:00
Marco Costalba 9a542d9698 Initialize a new split point out of lock
Allocate and initialize a new split point
out of lock becuase modified data is local to
master thread only.

Also better document why we need a lock at
the end of split().

No functional change with faked split.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 14:37:14 +01:00
Marco Costalba cd27ed9a74 Update comment on why we call root search with ss+1
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 10:36:14 +01:00
Marco Costalba b9f8cb7837 Fix an assert when stopping the search
When StopRequest is raised we cannot immediately exit the
move loop but first we need to update bestValue so to avoid
assert:

assert(bestValue > -VALUE_INFINITE && bestValue < VALUE_INFINITE);

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 10:36:13 +01:00
Joona Kiiski ffa150bec3 Split at root!
Another great success by Joona !

After 5876 games at 10"+0.1
Mod vs Orig: 1073 - 849 - 3954 ELO +13 (+- 5.2)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-07 10:35:56 +01:00
Joona Kiiski 13bc6ba2c6 Preparations for splitting at root
No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-06 10:31:30 +01:00
Marco Costalba 30b3edf328 Simplify MovePickerExt<>
Now that we don't special case the root moves anymore
we don't need to pass NodeType anymore as template parameter,
a simple bool to detect a SpNode will be enough.

Spotted by Joona.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-05 07:23:54 +01:00
Marco Costalba 7902d6089e Fix silly bug in uci loop
After issuing "go"-command, at the end of the search
SF shows: "Unknown command: ...".

Spotted by Joona.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-05 07:15:45 +01:00
Marco Costalba b6b0878da8 Fix a (silly) warning under icc compiler
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-04 10:10:03 +01:00
Marco Costalba 3a76163aba Use std::lexicographical_compare() in UCI options
Instead of our home grown function to perform a case
insensitive compare on option names as required by UCI
protocol.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-03 15:00:23 +01:00
Marco Costalba 1e7c6fc761 Consistently set ttMove to Rml.pv[0] in root node
No functional change, but reduce risks of subtle aliasing bugs.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 22:28:03 +01:00
Marco Costalba c50ad85c3c Fix an off-by-one error in UCI print loop
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 21:47:31 +01:00
Marco Costalba bb6a6e159a Rename ok_to_use_TT() in can_return_tt()
Seems more appropiate. From Lucas Braesch's DoubleCheck.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 18:49:06 +01:00
Marco Costalba 1036cadcec Send PV only for updated lines
It seems FritzGUI already remembers the old lines, so
we just need to update PV info only for the new lines.

Also introduced prevScore field in RootMove to avoid
a bulk copy of Rml.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 18:00:47 +01:00
Marco Costalba c71ae794df Sort PV moves always in two steps
This should fix following issue:

Suppose the search with MultiPVIteration == 0 returns an exact score

move = Nxf4, score = 100

Now search with MultiPVIteration == 1 and get two scores

move = Qg8, score = 150
move = Ra1, score = 180

If we now reorder all the moves in one step we end up with

pv[0] = Ra1, pv[1] = Qg8

Instead reordering as the current patch we end up in:

pv[0] = Ra1, pv[1] = Nxf4

preserving the first searched move.

No functional change in single PV.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 18:00:47 +01:00
Marco Costalba 409930e98c Small cleanup of previous patches
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 18:00:22 +01:00
Joona Kiiski b88f7df387 Reimplement MultiPV mode
No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 06:47:33 +01:00
Joona Kiiski adcfffceeb Reimplement support for "searchmoves" option
No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 06:47:22 +01:00
Joona Kiiski a3c1d64a5f Remove now unused RootMove.non_pv_score
No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 06:47:06 +01:00
Joona Kiiski ce24a229df Make root search to use standard MovePicker.
This patch temporarily breaks MultiPV and searchmove
features, but they will be re-implemented in future
patches.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-08-02 06:46:50 +01:00
Joona Kiiski ce619b3b6c Don't probe TT at RootNode
In that case we should also update RootMoveList to avoid
bogus output

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-29 12:42:22 +01:00
Marco Costalba 2e2a4b4ea3 Fix pretty_pv() output in Chess960
And move it to search.cpp

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-26 12:39:15 +01:00
Joona Kiiski 7f519a2463 Fix PV output in Chess960
We missed to set chess960 flag into the std::stringstream used to
setup the PV line.

Bug introduced with commit f803f33e63
of 30/12/2010 when we started to print PV line into a std::stringstream
instead of directly into cout, where the chess960 flag is correctly set.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 18:27:29 +01:00
Marco Costalba ff1ecb5d6c Tidy up benchmark.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 08:53:36 +01:00
Marco Costalba dab1cd8af9 Rename execute_uci_command() to uci_loop()
As a side effect now root position can be directly
allocated on the stack and doesn't need to be defined
static anymore.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 08:12:07 +01:00
Marco Costalba 5b0c6b9bc0 Unhide the istringstream behind UCIParser
It is misnamed because it is not a parser, perhaps a
tokenizer, anyhow better call it for what it is, an
input string stream.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 07:39:07 +01:00
Marco Costalba fc290dc30b Fix startpos_ply_counter() regression
Return the correct number of played plies at the end
of the setup moves. Currently it always returns 0 when
starting from start position, like in real games.

We fix this adding st->pliesFromNull that starts from 0
and is incremented after each setup move and is never
reset in the setup phase but only once search is started.

It is an hack because startpos_ply_counter() will return
different values during the search and is correct only
at the beginning of the search.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:59:14 +01:00
Marco Costalba 5f7eb20090 Use a circular buffer to keep track of setup states
This fixes a regression on real games due to the fact that
we have some mismatches:

    history[st->gamePly - i] != stp->key

when st->gamePly - i == 0,this is due to a nasty bug I have
introduced when using std::vector<> as StateInfo backup. The
point is that StateInfo keeps inside a pointer to the previous
StateInfo in a kind of linked list. But when std::vector<> is
resized reallocates a larger chunk of memory and moves the
data there so these pointers became stale.

This patch fixes the issue.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:15:40 +01:00
Marco Costalba 03ad183384 Don't update gamePly after each move
We just need startup value to calculate available
thinking time. So remove from state.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:15:39 +01:00
Marco Costalba 527a2ec541 Use std::vector<Move> to store UCI search moves
Avoid the ugly and anyhow incorrect hard limit on the
maximum number of moves and allow to handle an arbitrary
number of moves to search.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:15:38 +01:00
Marco Costalba 3185c36a65 Use st->gamePly to store fullMoves
This allow to retire do_setup_move() and also to simplify
draw detection logic becuase now we always have:

Min(st->rule50, st->gamePly) = st->rule50

This was already true when starting from starting position,
but now is true even when starting from a FEN string because
now we take in account fullmove number in counting gamePly so
that it is always.

st->rule50 <= st->gamePly

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:15:36 +01:00
Marco Costalba 3d8140a541 Retire history[]
Use key saved in state instead.

No functional change (in real games) and no speed regression.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:15:35 +01:00
Joona Kiiski a6fc3d6ee5 Do not exit early even when seeing mate
Fixes the reported KNNK ending problem:

http://talkchess.com/forum/viewtopic.php?t=39347

Joona says:

Now I finally had a time to take a look at on this issue.

I've reproduced the problem starting from this position:
1B6/1B2k3/P7/1P3p2/1K6/8/4b3/4b3 w - - 6 85

I made Stockfish play as white and Fruit as black.
I repeated test ten times and once SF was not able to deliver mate.

But I observed several times that SF had reported on last something like mate in 10.
However next time it played move with score mate in 15.

Easiest way to solve the problem is attached as a patch. I tested it several times and SF always
ended up playing the optimal move. Of course the downside is that now delivering mate
takes a bit longer, but IMO it's better to lose once in a while by time in sudden death
game than not being able to deliver simple mate with long time controls.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-24 06:13:29 +01:00
Marco Costalba a1f9bf19d9 Revert previous patches due to bug
We have a bug (possibly because of returning draw from
root move list), it is possible to see when looking at
games with a GUI, we can see rarely but consistently the
score return as #0 for many depths until it comes back to
normal values.

Revert patches until it is not fixed.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-20 03:34:39 +01:00
Joona Kiiski 4ad6a3496b Move the draw check also for qsearch
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-19 17:36:22 +01:00
Joona Kiiski 969ad8001c Move draw checks right after doing the move
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-19 17:36:14 +01:00
Marco Costalba 07e0dd27fb Small touches in set_option()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-17 12:55:26 +01:00
Marco Costalba b8eb699db7 Validate input UCI moves
Running following command:

  position startpos moves e1e8

Makes SF to assert in debug mode in do_move() but to accept
bad input and continue in release mode where probably it is
going to crash little later.

So validate input before to feed do_move().

Suggestion by Yakovlev Vadim.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-17 10:22:55 +01:00
Marco Costalba ad1f28bc1c Remove some useless include
Spotted by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-16 22:59:14 +01:00
Marco Costalba 67686b7684 Don't need to assert for pos.is_ok() when position is constant
It's only necessary to do the checking at the end of every non-const
member (including the constructors and from_fen()) of class Position.
Once the post-condition of every modifier guarantees the class invariant,
we don't need to verify sanity of the position as preconditions for outside
callers such as movegen, search etc. For non-class types such as Move and
Square we still need to assert of course.

Suggested by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-16 10:53:34 +01:00
Marco Costalba 69f4954df1 Change hidden checkers API
After previous patch is no more needed to pass
the color, becuase it is always the side to move.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-16 10:12:46 +01:00
Marco Costalba 4894231ff7 Simplified discovered check connected_moves()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-16 09:50:35 +01:00
Marco Costalba 36bb57a47e Simplify and micro-optimize hidden_checkers()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-16 08:53:21 +01:00
Marco Costalba a042449f5d No need to declare default Position c'tor
Pointed out by Rein Halbersma.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-15 21:53:47 +01:00
Marco Costalba f5e434c2d9 Fix a warning under MSVC
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-15 12:20:31 +01:00
Marco Costalba 37d551fb39 Fix parametrized direction in pawns generation
It worked by accident because we always called both directions,
but definition was wrong.

Functional change due to different generation order, but
perft numbers are the correct ones.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-15 12:13:55 +01:00
Marco Costalba 9c8c0de538 Cleanup handling of Delta enums
Ispired by Rein's code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-15 12:13:48 +01:00
Marco Costalba fbdabe2975 Use std library to sort moves
Functional change due to the fact that now pick_best() is
stable, but should be no change in strenght.

Example code and ideas by Rein Halbersma.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-14 12:43:13 +01:00
Marco Costalba 0cfb30e5be Fix icc's "unreachable code" warning
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-11 11:17:50 +01:00
Marco Costalba c081a81daf Teach to_fen() about Halfmove and Fullmove number
And also fix the last '/' at the end of the piece placement field.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-03 12:08:02 +01:00
Marco Costalba 155bed18f5 Use MoveList also in Position::move_is_pl_slow()
And rename it in Position::move_is_legal()

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-03 11:00:28 +01:00
Marco Costalba 95d9687d95 Retire move_is_short_castle() and move_is_long_castle()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-03 10:27:15 +01:00
Marco Costalba ff41b8df76 Restore startpos_ply_counter() instead of full_moves()
And pass correct currentPly to TimeManager::init().

This restores old behaviour, in particular now black has
a different timing than white becuase is no more:

currentPly = 2 * fullMoveNumber;

but becomes

2 * (fullMoves - 1) + int(sideToMove == BLACK)

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-03 06:54:46 +01:00
Marco Costalba 53ccba8457 Introduce and use struct MoveList
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-02 14:04:33 +01:00
Marco Costalba 7ac6e3b850 Remove MSVC debug window hack in bench
Interference with profile-build under mingw

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-02 11:52:22 +01:00
Marco Costalba d15217b953 Rearrange structs to avoid internal padding
Found with gcc -Wpadded gcc option.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-02 11:01:12 +01:00
Marco Costalba bc54a44010 Revert PHQ-2
No clear advantage again standard, possibly we will
retest before to release.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-07-02 09:47:06 +01:00
Marco Costalba 3f806df0ca Small touches to do/undo_castle_move()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-30 20:04:46 +01:00
Marco Costalba f3799aa750 Better document generate_castle_moves()
No functional change also in Chess960

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-30 19:58:44 +01:00
Marco Costalba 30ca6935a5 Small tweaks to search()
No functional change also in faked split mode

To be sure verified in real games with 4 threads TC 2"+0.1
After 11125 games 2497 - 2469 - 6159 ELO +0 (+- 4.4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-30 06:54:09 +01:00
Marco Costalba f25582d4b8 Remove duplicated enum Phase definition
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-29 17:54:12 +01:00
Marco Costalba fb2fdb21d3 Retire find_checkers()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 18:01:51 +01:00
Marco Costalba 31a0d2200c Retire square_is_weak()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:22:33 +01:00
Marco Costalba e0a00c4996 Retire one piece_list() overload
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:11:03 +01:00
Marco Costalba e5077dc11e Rename pieces_of_color() in pieces()
To be uniform with other overloads.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:10:44 +01:00
Marco Costalba 01a191936e Retire redundant square_is_occupied()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:10:27 +01:00
Marco Costalba 8094b2add8 Change Position::pst() signature
To be more clear what is the underlying table.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:10:18 +01:00
Marco Costalba ffb638995d Fix Shredder-FEN regression in from_fen()
Fix also an incredible 3% speed regression by an almost
never called function. I guess this is due to mingw very low
quality standard libraries implementation.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-28 17:10:05 +01:00
Marco Costalba c2a4856f9e Greatly simplify castling rights handling
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-27 19:26:30 +01:00
Marco Costalba c9b24c3358 Assume input FEN string is correct in from_fen()
And also tolerate a 0 value for full move number.

Revert BUG_41 patch, now we set initial King file only
if a castling is possible, so we don't need the fix
anymore in case of correct FEN.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-27 12:08:12 +01:00
Marco Costalba 9305983018 Fix a bug in Position::is_ok()
If we cannot castle castleRightsMask[] could be not valid,
for instance when king initial file is FILE_A as queen rook.

In this case castleRightsMask[] at initialQRFile is different
from the expected (ALL_CASTLES ^ WHITE_OOO).

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-26 11:44:12 +01:00
Marco Costalba ae2f5f25cd Rename type_of_piece() and color_of_piece()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-26 10:52:42 +01:00
Marco Costalba 923b14afaa Retire Position::color_of_piece_on()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-26 10:37:13 +01:00
Marco Costalba a9782b94e6 Retire Position::type_of_piece_on()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-26 10:28:54 +01:00
Marco Costalba 351ef5c85b Retire seeValues[] and move PieceValue[] out of Position
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-26 10:19:37 +01:00
Marco Costalba be2925b3c5 Restore user weights to 100
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-24 10:31:03 +01:00
Marco Costalba ace27831ee PHQ settings for King and Mobility
See:
http://talkchess.com/forum/viewtopic.php?t=39327

After 8130 games on QUAD at 20"+0.1
1342 - 1359 - 5429 ELO +0 (+- 4.4)

Tried also version with just king settings changed:
After 5932 games 962 - 1052 - 3918 ELO -5 (+- 5.2)

And with just mobility settings changed:
After 4114 games 618 - 619  - 2877 ELO +0 (+- 5.9)

Frank has tested only 1200 games, but at longer TC and
against many engines, so because PHQ results are not worst
than other combination and not worst than original let's
commit his version.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-24 09:47:30 +01:00
Marco Costalba 6a1707889c Fix move_is_capture() to detect capture promotions
We miss to account as a capture a promotion capture !

Incredible bug in this critical function that is here
since a long time (b50921fd5c of 21/10/2009 !!)

This patch fixes the bug and readds the faster
move_is_capture_or_promotion() that slightly increases
overall speed.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-23 12:09:31 +01:00
Marco Costalba 1c42d15340 Rewrite how uci info is sent to GUI
It is now much more modular than before and also we
always send the seldepth when we send the depth, this
avoids to make seldepth disappearing from GUI at the
start of a new iteration.

Print also fails high/low pv lines at high enough
search depths.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-22 17:32:22 +01:00
Marco Costalba fe26967ea0 Simplify sliding_attacks()
Easy, almost trivial simplification, I don't understand
how I missed this before !!

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-21 19:01:00 +01:00
Marco Costalba e7413417ce Retire ksq from CheckInfo
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-21 12:13:26 +01:00
Marco Costalba 018022b866 Use CheckInfo to store pinned bitboard
This trivial change gives an impressive 2,5% speedup !!!!

Also retire one unused move_gives_check() overload.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-20 12:09:11 +01:00
Marco Costalba 7ea38e980b Omit mate distance pruning at root
Restore original behaviour, before root unification and
remove a now useless ugly hack for alpha in multi-pv case.

No functional change

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-18 17:11:09 +01:00
Marco Costalba 8173d92dd2 Use an array index instead of an iterator in root list
It is not correct to use an iterator stick on a vector that
is sorted becuase iterator is invalidated in general case.

It happens to work by accident because iterators are implemented
as pointers and so they behave in the same (correct) way then
using array indices, but the latters are the correct thing to use.

Also better document the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-18 08:03:59 +01:00
Marco Costalba 808a4fe817 Remove useless bestValue = alpha assignement
It is a fossil from the root_search() era, no more
needed today.

Spotted by Onno

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-15 12:49:59 +01:00
Marco Costalba 72760c05c6 Try only recaptures in qsearch if depth is very low
This avoids search explosion in qsearch for some
patological cases like:

r1n1n1b1/1P1P1P1P/1N1N1N2/2RnQrRq/2pKp3/3BNQbQ/k7/4Bq2 w - - 0 1

After 9078 games 20"+0.1 QUAD:
Mod vs Orig 1413 - 1319 - 6346 ELO +3 (+- 4)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-15 12:17:49 +01:00
Marco Costalba 15683034a7 Speed up kpk initialization
The trick is to classify more position at first cycle,
so to reduce following work. Speed up is of about 50% !

Also some cleanup while there.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-15 12:11:35 +01:00
Marco Costalba 4c9d570e43 Use Carry-Rippler trick to speed up magics
Nice trick discovered on:

http://chessprogramming.wikispaces.com/Traversing+Subsets+of+a+Set

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-15 02:15:41 +01:00
Marco Costalba fc519ca74a Retire init_piece_square_tables()
Merge in init_zobrist() and rename the latter.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-13 17:28:09 +01:00
Marco Costalba a24c0a736c Increase LMR limit by one ply
Seems there is no regression so prefer to prune less.

After 8278 games
Mod vs Orig 1246 - 1265 - 5767 +0 ELO (+- 4.2)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-12 21:14:07 +01:00
Marco Costalba b3a0b389d2 Better self-document init_zobrist()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-12 11:50:19 +01:00
Marco Costalba 47959c56fd Fix initialization of BSFTable[]
We should start from i = 0, it works by accident because
static storage BSFTable[] is init to zero by default.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-12 11:37:02 +01:00
Marco Costalba aee75ae105 Don't update_gains() in qsearch
Is almost unuseful becuase captures are skipped.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-12 07:09:30 +01:00
Marco Costalba 4b8a7f2793 Fix score_captures() for the case of capture promotions
In case we have more than one promotion move, prefer
the one that captures the biggest piece.

Almost no functional change, anyhow I don't expect any
ELO change, it is just the correct thing to do.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-11 22:12:13 +01:00
Marco Costalba 89ec224cb9 Retire some unused functions
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-11 15:56:12 +01:00
Marco Costalba b21a5e2f06 Micro-optimize castling handling in do_move()
And better self-document the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-11 15:31:39 +01:00
Marco Costalba 735cac5d53 Retire PieceLetters struct
Use a much simpler std::string instead.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-10 17:12:10 +01:00
Marco Costalba bc4f3155ae Better document move_to_san()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-10 11:59:53 +01:00
Marco Costalba a21a110188 Revert refinedValue in ProbCut
It seems much worst in number of nodes seacrhed to reach
the depth and anyhow does not give any advantage to the
Onno's oroginal one.

So revert by now and perhaps readd when we find something
clearly better.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-09 21:59:24 +01:00
Marco Costalba b414fc0dfd Use double rotate for magic generation
Allow to choose among 4096 instances of pseudo-random
sequences instead of the previous 64 so the probability
to find a better sequence increases and actually we have
a much better 64 bit case and we can also use the 64 bit
version of pick_magic() also for 32 bits and althoug sub
optimal, because now we can have more choices results are
even slightly better also for 32 bit.

Use also a faster submask().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-09 18:47:58 +01:00
Marco Costalba e0215f3222 Use refinedValue in ProbCut condition
After 12613 games at 20"+0.1 on QUAD
Mod vs Orig 1870 - 1863 - 8880 ELO +0 (+- 3.3)

So no performance change but it is a code semplification
and also is more easy to understand.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-08 18:03:26 +01:00
Marco Costalba 3dfff5bdae Small pick_magic() touches
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-08 17:59:47 +01:00
Marco Costalba d632e77058 Find magics on the fly
Good result for 32 bit case where computation is very fast,
still not satisfying on 64 bit case where the magics seem
a bit harder to get.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-07 20:59:12 +01:00
Marco Costalba 6d665b7f78 Partially revert previous patches
Due to a -2% speed penalty. This patch takes the best
of the previous series without the regression due to
introduction of Magic struct.

Speedup against previous revision is of almost 3% !!!!

No functional change both in 32 and 64 bits.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-06 11:56:17 +01:00
Marco Costalba b1b0c64046 Skip offset calculation in slider attacks
Another small simplification and micro optimization.

No functional change in both 32 and 64 bits.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-05 15:01:22 +01:00
Marco Costalba 670cee44f0 Get rid of Shift[] tables
We can calculate them counting the masks bits.

Also small tweak to sliding_attacks()

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-05 14:13:41 +01:00
Marco Costalba 3092f0c646 Better name and document magic botboard stuff
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-05 12:28:36 +01:00
Marco Costalba 2f6142cb9b Try to keep memory access in the same cache line
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-05 12:27:53 +01:00
Marco Costalba 3b2bcee0a8 Skip draw by repetition check in qsearch
Cut in half the time spent in pos.draw() that accounts
for a whopping 1% of total time !

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-04 11:29:54 +01:00
Marco Costalba 91407f4f74 Move bitboards initializations under one function
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-04 11:05:12 +01:00
Marco Costalba 025d57855a Calculate Bit Scan tables at initialization
Instead of hard-coding them.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-04 09:59:18 +01:00
Marco Costalba 8647fbd6ed Do not sort negative non captures at low depth
Speedup of the whole 3 patch series is of 2,5% !!

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-04 06:35:24 +01:00
Marco Costalba 811037c845 Split non capture in two sets when ordering
But keep same ordering. This patch is prerequisite
for future work.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-03 22:10:23 +01:00
Marco Costalba 181cc3f93f Inline extension()
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-03 13:14:03 +01:00
Marco Costalba 91b919fd1d Use TT also in Root nodes
And other small stuff ti be tested in SMP

No functional change in single thread.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-03 12:59:50 +01:00
Marco Costalba 5ff6289afd Microoptimize generate<MV_EVASION>
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-06-02 12:49:40 +01:00
Marco Costalba fca0a2dd88 New extended probcut implementation
Here the idea is to test probcut not only after bad
captures, but after any bad move, i.e. any move that
leaves the opponent with a good capture.

Ported by a patch from Onno, the difference from
original version is that we have moved probcut after
null search.

After 7917 games 4 threads 20"+0.1
Mod vs Orig: 1261 - 1095 - 5561 ELO +7 (+- 4.2) LOS 96%

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-30 20:49:04 +01:00
Marco Costalba 462a39ec49 Fix SAN disambiguation bug
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-29 10:52:03 +01:00
Marco Costalba 70125a3be0 Rename PH_TT_MOVES in PH_TT_MOVE
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-29 09:17:03 +01:00
Marco Costalba cff8877a1a Retire mateKiller
Practically useless:
After 6456 games 1281 - 1293 - 3882 ELO +0 (+- 5.5)

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-28 22:50:22 +01:00
Marco Costalba 8f51f09de7 Unify MovePickerExt template parameters
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-28 13:23:32 +01:00
Marco Costalba 013dc43d5d Unify search() template parameters
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-28 12:44:39 +01:00
Marco Costalba 853e2a9495 Fix moveCount after legality check delay
We really want PV moves and also Split Point moves to be
legal to avoid messing the move counter and corresonding
PV move detection or shared Split Point's counter variable.

This fixes a real bug where a position with only one move
allowed returns bestValue == -VALUE_INFINITE if the move
turns out to be illegal.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-28 11:16:00 +01:00
Marco Costalba 4c3a000211 A bit of reformatting after previous series
And some documentation update.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-24 21:24:40 +01:00
Marco Costalba 9884573561 Test for legality only after futility pruning
Although there is a small functional change it seems
an improvment of about 2% in speed !

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-23 20:21:02 +01:00
Marco Costalba 4b232f5ddc Move legal check out of MovePicker
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-23 20:20:52 +01:00
Marco Costalba 45ce92b89c Rename move_is_legal() in move_is_pl()
We disjoint pseudo legal detection from full legal detection.

It will be used by future patches.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-23 20:20:42 +01:00
Marco Costalba bc86668ba4 Output debug info to cerr
So to be clearly visible when redirecting stdout to /dev/null

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-23 20:20:31 +01:00
Marco Costalba 0783211950 Fix a shadowed variable warning under gcc
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-23 20:20:19 +01:00
Marco Costalba 13fe7ee4df Bug wrong evasion detection for king moves
When we are in check and we move the king then testing with
pl_move_is_legal(m, pinned) is not enough becuase we cannot
rely on attackers_to() but we have to explicitly remove the
king form the occupied bitboard to catch as invalid moves like
b1a1 when opposite queen is on c1.

Our move generator already produces correct evasions so we
just need to add the extra verification to move_is_legal().

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-22 11:51:28 +01:00
Marco Costalba 21fc66c246 Add file distance condition in move_is_legal()
Found another missed control in move_is_legal() thanks to
brute force testing.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-22 09:49:44 +01:00
Marco Costalba ff9e49bac9 Remove useless casts in types.h
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-22 08:52:39 +01:00
Marco Costalba 90e83fa879 Promotion piece must be empty if is not a promotion
Add a new check in  move_is_legal()

Avoid useless casting in move.h while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-22 08:48:19 +01:00
Marco Costalba 3ef4fdeaa0 Introduce MovePicker::isBadCapture() and use in probcut
Small functional change due to the fact that now we skip
probcut on evasions.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-21 10:40:36 +01:00
Marco Costalba 13d8af1852 Correctly handle castle in see()
Suggested by Onno.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-20 06:58:53 +01:00
Marco Costalba 5b7a141065 Fix brekage from previous patches
It is interesting the fact that we need to test for
move_is_castle(m) anyway and not relying on testing
if destination square is attacked. Indeed the latter
condition fails if the castling rook is attacked,
castling is coded as "king captures the rook" but it
is legal in that case.

Verified no functional change with beginning of the series.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-18 07:05:36 +01:00
Marco Costalba 89a06f6651 Micro-optimize pl_move_is_legal()
Remove the check for castling moves because it is
already implicit in the check for king moves and castling
is so rare that doing the check is just a slow down.

Thanks to Marek Kwiatkowski.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 23:47:26 +01:00
Marco Costalba a2e924039b Retire move_is_capture_or_promotion()
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 23:39:14 +01:00
Marco Costalba 85d1f9c5ec Fix move_is_capture() definition
The structure of move is changed so should also the two
functions. It happens that it works by accident !

Bug spotted by Marek Kwiatkowski

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 23:31:40 +01:00
Marco Costalba e444e18d2b Retire test for king moves in see()
We already test this condition in see_sign() and
so it is almost always a redundant verification.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 12:25:49 +01:00
Marco Costalba ad0fdf0da6 Retire Position::see(Square from, Square to) overload
Alomst unuseful.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 12:17:57 +01:00
Marco Costalba fbbc7e421c Prefer an assert to a comment in position.h
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 12:17:48 +01:00
Marco Costalba 1cc42ac9e4 Let 'make' with no arguments to show compilation options
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 12:17:38 +01:00
Marco Costalba 6624105b5b Prefer ttMove to tte->move() in search()
Avoids aliasing problems due to TT overwrites.

Node changes becuase of IID.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-17 12:17:29 +01:00
Marco Costalba de58594b0f Use standard naming conventions in book.cpp
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-08 11:50:38 +01:00
Marco Costalba d5f0b91c06 Reintroduce operator>>() in Book class
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-08 09:15:18 +01:00
Marco Costalba f78488bbdb Restore development version
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-08 09:11:36 +01:00
Marco Costalba e3b23eb818 Stockfish 2.1.1
stockfish bench signature is: 6487630

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-08 08:46:33 +01:00
Marco Costalba d494725400 Spell fix in evaluate.cpp
Spotted by Eelco.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-07 10:08:53 +01:00
Marco Costalba 2535dc1340 Fix reading of book file
Bug is subtle because appears only under MSVC 32 bits in
optimized version, hence was missed before.

Bug is due to the fact that evaluation order of terms of a
sum is undefined by the standard, so in get_int() we have:

return 256 * get_int<n-1>() + bookFile.get();

And if get() is evaluated before get_int() we have a corrupted
key.

The patch rewrites the code in a more natural and predictable way.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-07 09:51:33 +01:00
Marco Costalba 339bf9b7e1 Remove redundant assignment in search()
It is already assigned few lines before.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-05 12:16:26 +01:00
Marco Costalba bb86691a0d Restore development version
No functional change.
2011-05-05 06:35:42 +01:00
Marco Costalba f7fee4c616 Fix a warning in debug mode
No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
2011-05-03 19:29:21 +01:00
59 changed files with 10368 additions and 11588 deletions
+73
View File
@@ -0,0 +1,73 @@
language: cpp
sudo: required
dist: trusty
matrix:
include:
- os: linux
compiler: gcc
addons:
apt:
sources: ['ubuntu-toolchain-r-test']
packages: ['g++-6', 'g++-6-multilib', 'g++-multilib', 'valgrind', 'expect']
env:
- COMPILER=g++-6
- COMP=gcc
- os: linux
compiler: clang
addons:
apt:
sources: ['ubuntu-toolchain-r-test']
packages: ['clang', 'g++-multilib', 'valgrind', 'expect']
env:
- COMPILER=clang++
- COMP=clang
- os: osx
compiler: gcc
env:
- COMPILER=g++
- COMP=gcc
- os: osx
compiler: clang
env:
- COMPILER=clang++ V='Apple LLVM 6.0' # Apple LLVM version 6.0 (clang-600.0.54) (based on LLVM 3.5svn)
- COMP=clang
branches:
only:
- master
before_script:
- cd src
script:
# Obtain bench reference from git log
- git log HEAD | grep "\b[Bb]ench[ :]\+[0-9]\{7\}" | head -n 1 | sed "s/[^0-9]*\([0-9][0-9]*\)/\1/g" > git_sig
- export benchref=$(cat git_sig)
- echo "Reference bench:" $benchref
#
# Verify bench number against various builds
- export CXXFLAGS=-Werror
- make clean && make -j2 ARCH=x86-64 optimize=no debug=yes build && ../tests/signature.sh $benchref
- make clean && make -j2 ARCH=x86-32 optimize=no debug=yes build && ../tests/signature.sh $benchref
- make clean && make -j2 ARCH=x86-32 build && ../tests/signature.sh $benchref
- make clean && make -j2 ARCH=x86-64 build && ../tests/signature.sh $benchref
#
# Check perft and reproducible search
- ../tests/perft.sh
- ../tests/reprosearch.sh
#
# Valgrind
#
- export CXXFLAGS=-O1
- if [ -x "$(command -v valgrind )" ]; then make clean && make -j2 ARCH=x86-64 debug=yes optimize=no build > /dev/null && ../tests/instrumented.sh --valgrind; fi
- if [ -x "$(command -v valgrind )" ]; then ../tests/instrumented.sh --valgrind-thread; fi
#
# Sanitizer
#
# Use g++-6 as a proxy for having sanitizers, might need revision as they become available for more recent versions of clang/gcc
- if [[ "$COMPILER" == "g++-6" ]]; then make clean && make -j2 ARCH=x86-64 sanitize=undefined optimize=no debug=yes build > /dev/null && ../tests/instrumented.sh --sanitizer-undefined; fi
- if [[ "$COMPILER" == "g++-6" ]]; then make clean && make -j2 ARCH=x86-64 sanitize=thread optimize=no debug=yes build > /dev/null && ../tests/instrumented.sh --sanitizer-thread; fi
+98
View File
@@ -0,0 +1,98 @@
# Generated with 'git shortlog -sn | cut -c8-', which sorts by commits, manually ordered the first four authors, merged duplicates
Tord Romstad
Marco Costalba (mcostalba)
Joona Kiiski (zamar)
Gary Linscott (glinscott)
Lucas Braesch (lucasart)
Bill Henry (VoyagerOne)
mstembera
Stéphane Nicolet (Stephane Nicolet, snicolet)
Stefan Geschwentner
Alain SAVARD (Rocky640)
Jörg Oster (Joerg Oster, joergoster)
Reuven Peleg
Chris Caino (Chris Cain, ceebo)
Jean-Francois Romang
homoSapiensSapiens
Leonid Pechenik
Stefano Cardanobile (Stefano80)
Arjun Temurnikar
Uri Blass (uriblass)
jundery
Ajith (ajithcj)
hxim
Ralph Stößer (Ralph Stoesser)
Guenther Demetz
Jonathan Calovski (Mysseno)
Tom Vijlbrief
mbootsector
Daylen Yang
ElbertoOne
Henri Wiechers
loco-loco
Joost VandeVondele (Joost Vandevondele)
Ronald de Man (syzygy)
DU-jdto
David Zar
Eelco de Groot
Jerry Donald
NicklasPersson
Ryan Schmitt
Alexander Kure
Dan Schmidt
H. Felix Wittmann
Jacques
Joseph R. Prostko
Justin Blanchard
Linus Arver
Luca Brivio
Lyudmil Antonov
Rodrigo Exterckötter Tjäder
Ron Britvich
RyanTaker
Vince Negri
erbsenzaehler
Joseph Hellis (jhellis3)
shane31
Andrew Grant
Andy Duplain
Auguste Pop
Balint Pfliegel
Dariusz Orzechowski
DiscanX
Ernesto Gatti
Gregor Cramer
Hiraoka Takuya (HiraokaTakuya)
Hongzhi Cheng
IIvec
Kelly Wilson
Ken T Takusagawa
Kojirion
Krgp
Matt Sullivan
Matthew Lai
Matthew Sullivan
Michel Van den Bergh
Niklas Fiekas
Oskar Werkelin Ahlin
Pablo Vazquez
Pascal Romaret
Raminder Singh
Richard Lloyd
Ryan Takker
Thanar2
absimaldata
atumanian
braich
fanon
gamander
gguliash
kinderchocolate
pellanda
ppigazzini
renouve
sf-x
thaspel
unknown
+123
View File
@@ -0,0 +1,123 @@
### Overview
[![Build Status](https://travis-ci.org/official-stockfish/Stockfish.svg?branch=master)](https://travis-ci.org/official-stockfish/Stockfish)
[![Build Status](https://ci.appveyor.com/api/projects/status/github/official-stockfish/Stockfish?svg=true)](https://ci.appveyor.com/project/mcostalba/stockfish)
Stockfish is a free UCI chess engine derived from Glaurung 2.1. It is
not a complete chess program and requires some UCI-compatible GUI
(e.g. XBoard with PolyGlot, eboard, Arena, Sigma Chess, Shredder, Chess
Partner or Fritz) in order to be used comfortably. Read the
documentation for your GUI of choice for information about how to use
Stockfish with it.
This version of Stockfish supports up to 512 cores. The engine defaults
to one search thread, so it is therefore recommended to inspect the value of
the *Threads* UCI parameter, and to make sure it equals the number of CPU
cores on your computer.
This version of Stockfish has support for Syzygybases.
### Files
This distribution of Stockfish consists of the following files:
* Readme.md, the file you are currently reading.
* Copying.txt, a text file containing the GNU General Public License.
* src, a subdirectory containing the full source code, including a Makefile
that can be used to compile Stockfish on Unix-like systems.
### Syzygybases
**Configuration**
Syzygybases are configured using the UCI options "SyzygyPath",
"SyzygyProbeDepth", "Syzygy50MoveRule" and "SyzygyProbeLimit".
The option "SyzygyPath" should be set to the directory or directories that
contain the .rtbw and .rtbz files. Multiple directories should be
separated by ";" on Windows and by ":" on Unix-based operating systems.
**Do not use spaces around the ";" or ":".**
Example: `C:\tablebases\wdl345;C:\tablebases\wdl6;D:\tablebases\dtz345;D:\tablebases\dtz6`
It is recommended to store .rtbw files on an SSD. There is no loss in
storing the .rtbz files on a regular HD.
Increasing the "SyzygyProbeDepth" option lets the engine probe less
aggressively. Set this option to a higher value if you experience too much
slowdown (in terms of nps) due to TB probing.
Set the "Syzygy50MoveRule" option to false if you want tablebase positions
that are drawn by the 50-move rule to count as win or loss. This may be useful
for correspondence games (because of tablebase adjudication).
The "SyzygyProbeLimit" option should normally be left at its default value.
**What to expect**
If the engine is searching a position that is not in the tablebases (e.g.
a position with 7 pieces), it will access the tablebases during the search.
If the engine reports a very large score (typically 123.xx), this means
that it has found a winning line into a tablebase position.
If the engine is given a position to search that is in the tablebases, it
will use the tablebases at the beginning of the search to preselect all
good moves, i.e. all moves that preserve the win or preserve the draw while
taking into account the 50-move rule.
It will then perform a search only on those moves. **The engine will not move
immediately**, unless there is only a single good move. **The engine likely
will not report a mate score even if the position is known to be won.**
It is therefore clear that behaviour is not identical to what one might
be used to with Nalimov tablebases. There are technical reasons for this
difference, the main technical reason being that Nalimov tablebases use the
DTM metric (distance-to-mate), while Syzygybases use a variation of the
DTZ metric (distance-to-zero, zero meaning any move that resets the 50-move
counter). This special metric is one of the reasons that Syzygybases are
more compact than Nalimov tablebases, while still storing all information
needed for optimal play and in addition being able to take into account
the 50-move rule.
### Compiling it yourself
On Unix-like systems, it should be possible to compile Stockfish
directly from the source code with the included Makefile.
Stockfish has support for 32 or 64-bit CPUs, the hardware POPCNT
instruction, big-endian machines such as Power PC, and other platforms.
In general it is recommended to run `make help` to see a list of make
targets with corresponding descriptions. When not using the Makefile to
compile (for instance with Microsoft MSVC) you need to manually
set/unset some switches in the compiler command line; see file *types.h*
for a quick reference.
### Resource For Understanding the Code Base
* [Chessprogramingwiki](https://chessprogramming.wikispaces.com) has good overall chess engines explanations
(techniques used here are well explained like hash maps etc), it was
also recommended by the [support at stockfish.](http://support.stockfishchess.org/discussions/questions/1132-how-to-understand-stockfish-sources)
* [Here](https://chessprogramming.wikispaces.com/Stockfish) you can find a set of features and techniques used by stockfish and each of them is explained at the wiki, however, it's a generic way rather than focusing on stockfish's own implementation, but it will still help you.
### Terms of use
Stockfish is free, and distributed under the **GNU General Public License**
(GPL). Essentially, this means that you are free to do almost exactly
what you want with the program, including distributing it among your
friends, making it available for download from your web site, selling
it (either by itself or as part of some bigger software package), or
using it as the starting point for a software project of your own.
The only real limitation is that whenever you distribute Stockfish in
some way, you must always include the full source code, or a pointer
to where the source code can be found. If you make any changes to the
source code, these changes must also be made available under the GPL.
For full details, read the copy of the GPL found in the file named
*Copying.txt*.
-88
View File
@@ -1,88 +0,0 @@
1. Introduction
---------------
Stockfish is a free UCI chess engine derived from Glaurung 2.1. It is
not a complete chess program, but requires some UCI compatible GUI
(like XBoard with PolyGlot, eboard, Josè, Arena, Sigma Chess, Shredder,
Chess Partner, or Fritz) in order to be used comfortably. Read the
documentation for your GUI of choice for information about how to use
Stockfish with your GUI.
This version of Stockfish supports up to 32 CPUs, but has not been
tested thoroughly with more than 4. The program tries to detect the
number of CPUs on your computer and set the number of search threads
accordingly, but please be aware that the detection is not always
correct. It is therefore recommended to inspect the value of the
"Threads" UCI parameter, and to make sure it equals the number of CPU
cores on your computer. If you are using more than four threads, it
is recommended to raise the value of "Minimum Split Depth" UCI parameter
to 6.
2. Files
--------
This distribution of Stockfish consists of the following files:
* Readme.txt, the file you are currently reading.
* Copying.txt, a text file containing the GNU General Public
License.
* src/, a subdirectory containing the full source code, including a
Makefile that can be used to compile Stockfish on Unix-like
systems. For further information about how to compile Stockfish
yourself, read section 4 below.
* polyglot.ini, for using Stockfish with Fabien Letouzey's PolyGlot
adapter.
3. Opening books
----------------
This version of Stockfish has support for PolyGlot opening books.
For information about how to create such books, consult the PolyGlot
documentation. The book file can be selected by setting the UCI
parameter "Book File".
4. Compiling it yourself
------------------------
On Unix-like systems, it should usually be possible to compile
Stockfish directly from the source code with the included Makefile.
For big-endian machines like Power PC you need to enable the proper
flag changing from -DNBIGENDIAN to -DBIGENDIAN in the Makefile.
Stockfish has POPCNT instruction runtime detection and support. This can
give an extra speed on Core i7 or similar systems. To enable this feature
compile with 'make icc-profile-popcnt'
On 64 bit systems the 'bsfq' assembly instruction will be used for bit
counting. Detection is automatic at compile time, but in case you experience
compile problems you can comment out #define USE_BSFQ line in types.h
In general is recommended to run 'make help' to see a list of make targets
with corresponding descriptions.
5. Terms of use
---------------
Stockfish is free, and distributed under the GNU General Public License
(GPL). Essentially, this means that you are free to do almost exactly
what you want with the program, including distributing it among your
friends, making it available for download from your web site, selling
it (either by itself or as part of some bigger software package), or
using it as the starting point for a software project of your own.
The only real limitation is that whenever you distribute Stockfish in
some way, you must always include the full source code, or a pointer
to where the source code can be found. If you make any changes to the
source code, these changes must also be made available under the GPL.
For full details, read the copy of the GPL found in the file named
Copying.txt.
+132
View File
@@ -0,0 +1,132 @@
Contributors with >10,000 CPU hours as of January 23, 2018
Thank you!
Username CPU Hours Games played
mibere 518300 41835669
crunchy 375564 29121434
cw 371664 28748719
fastgm 299773 20765374
JojoM 220590 15299913
glinscott 204517 13932027
bking_US 187568 12233168
ctoks 169342 13475495
spams 149531 10940322
Thanar 137015 11714855
velislav 127305 10047749
vdbergh 121741 9056874
malala 117291 8126488
vdv 117218 8289983
leszek 114825 8331897
dsmith 114010 7622414
CSU_Dynasty 113516 9582758
sqrt2 112407 8782694
marrco 111143 8222921
drabel 108168 9061580
BrunoBanani 104938 7448565
Data 94621 8433010
CoffeeOne 90394 3964243
BRAVONE 80811 5341681
psk 77195 6156031
brabos 70284 5685893
Fisherman 66650 5572406
nssy 64587 5369140
Pking_cda 64499 5704075
sterni1971 63488 5070004
mgrabiak 62385 5420812
tvijlbrief 58957 4154234
jromang 58854 4704502
dv8silencer 57421 3961325
sunu 56620 4609155
tinker 56039 4204914
biffhero 55743 4810039
teddybaer 52982 4740444
bcross 50548 5071599
renouve 50318 3544864
Freja 50296 3805120
robnjr 47504 4131742
eva42 46542 4044694
davar 46538 4030604
finfish 46244 3481661
rap 46201 3219490
ttruscott 45037 3645430
solarlight 44155 4074841
TueRens 41372 3891510
ElbertoOne 41321 3920894
Antihistamine 39218 2792761
mhunt 38991 2697512
bigpen0r 37820 3149955
homyur 35569 3009637
VoyagerOne 35137 3302650
mhoram 34770 2684128
racerschmacer 33022 3231055
speedycpu 32043 2531964
EthanOConnor 31638 2143255
oryx 29574 2767730
Pyafue 28885 1986098
jkiiski 28014 1923255
Garf 27579 2770144
slakovv 27017 2031279
Bobo1239 27000 2488707
pb00067 26817 2306694
robal 26337 2316795
hyperbolic.tom 26248 2200777
rkl 24898 2236013
SC 23988 2126825
nabildanial 23524 1586321
achambord 23495 1942546
Sharaf_DG 22975 1790697
chriswk 22876 1947731
anst 22568 2013953
Patrick_G 22435 1682293
cuistot 22201 1383031
gri 21901 1820968
Prcuvu 21182 1890546
Zirie 21171 1493227
JanErik 20596 1791991
Isidor 20560 1730290
xor12 20535 1819280
team-oh 20364 1653708
nesoneg 20264 1493435
rstoesser 19802 1335177
grandphish2 19402 1834196
sg4032 18427 1671742
dew 18263 1423326
ianh2105 18133 1668562
MazeOfGalious 18022 1644593
ville 17900 1539130
j3corre 17607 975954
eudhan 17502 1424648
iisiraider 17175 1118788
jundery 17172 1115855
SFTUser 16635 1363975
purplefishies 16621 1106850
DragonLord 16599 1252348
chris 15274 1575333
xoto 14900 1486261
dju 14861 901552
dex 14647 1228763
nordlandia 14551 1369718
ronaldjerum 14361 1210607
OssumOpossum 14149 1029265
IgorLeMasson 13844 1228391
enedene 13762 935618
ako027ako 13442 1250249
AdrianSA 13324 924980
bpfliegel 13318 886523
ncfish1 13056 932344
wei 12863 1369596
jpulman 12776 854815
horst.prack 12436 1151505
joster 12424 986622
cisco2015 12265 1205019
fatmurphy 12015 901134
modolief 11228 926456
Dark_wizzie 11214 1017910
mschmidt 10973 818594
eastorwest 10970 1117836
infinity 10762 746397
SapphireBrand 10692 1024604
Thomas A. 10553 736094
pgontarz 10294 878746
Andrew Grant 10195 922933
stocky 10083 718114
+71
View File
@@ -0,0 +1,71 @@
version: 1.0.{build}
clone_depth: 50
branches:
only:
- master
- appveyor
# Operating system (build VM template)
os: Visual Studio 2015
# Build platform, i.e. x86, x64, AnyCPU. This setting is optional.
platform:
- x86
- x64
# build Configuration, i.e. Debug, Release, etc.
configuration:
- Debug
- Release
matrix:
# The build fail immediately once one of the job fails
fast_finish: true
# Scripts that are called at very beginning, before repo cloning
init:
- cmake --version
- msbuild /version
before_build:
- ps: |
# Get sources
$src = get-childitem -Path *.cpp -Recurse | select -ExpandProperty FullName
$src = $src -join ' '
$src = $src.Replace("\", "/")
# Build CMakeLists.txt
$t = 'cmake_minimum_required(VERSION 3.8)',
'project(Stockfish)',
'set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_SOURCE_DIR}/src)',
'set(source_files', $src, ')',
'add_executable(stockfish ${source_files})'
# Write CMakeLists.txt withouth BOM
$MyPath = (Get-Item -Path "." -Verbose).FullName + '\CMakeLists.txt'
$Utf8NoBomEncoding = New-Object System.Text.UTF8Encoding $False
[System.IO.File]::WriteAllLines($MyPath, $t, $Utf8NoBomEncoding)
# Obtain bench reference from git log
$b = git log HEAD | sls "\b[Bb]ench[ :]+[0-9]{7}" | select -first 1
$bench = $b -match '\D+(\d+)' | % { $matches[1] }
Write-Host "Reference bench:" $bench
$g = "Visual Studio 14 2015"
If (${env:PLATFORM} -eq 'x64') { $g = $g + ' Win64' }
cmake -G "${g}" .
Write-Host "Generated files for: " $g
build_script:
- cmake --build . --config %CONFIGURATION% -- /verbosity:minimal
before_test:
- cd src/%CONFIGURATION%
- ps: |
# Verify bench number
./stockfish bench 2> out.txt 1> null
$s = (gc "./out.txt" | out-string)
$r = ($s -match 'Nodes searched \D+(\d+)' | % { $matches[1] })
Write-Host "Engine bench:" $r
Write-Host "Reference bench:" $bench
If ($r -ne $bench) { exit 1 }
-39
View File
@@ -1,39 +0,0 @@
[PolyGlot]
EngineDir = .
EngineCommand = ./stockfish
Book = false
BookFile = book.bin
Log = false
LogFile = stockfish.log
Resign = true
ResignScore = 600
[Engine]
Hash = 128
Threads = 1
OwnBook = false
Book File = book.bin
Best Book Move = false
Use Search Log = false
Search Log Filename = SearchLog.txt
Mobility (Middle Game) = 100
Mobility (Endgame) = 100
Passed Pawns (Middle Game) = 100
Passed Pawns (Endgame) = 100
Space = 100
Aggressiveness = 100
Cowardice = 100
Minimum Split Depth = 4
Maximum Number of Threads per Split Point = 5
Use Sleeping Threads = false
Skill Level = 20
Emergency Move Horizon = 40
Emergency Base Time = 200
Emergency Move Time = 70
Minimum Thinking Time = 20
+336 -301
View File
@@ -1,6 +1,7 @@
# Stockfish, a UCI chess playing engine derived from Glaurung 2.1 # Stockfish, a UCI chess playing engine derived from Glaurung 2.1
# Copyright (C) 2004-2008 Tord Romstad (Glaurung author) # Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
# Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad # Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
# Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
# #
# Stockfish is free software: you can redistribute it and/or modify # Stockfish is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by # it under the terms of the GNU General Public License as published by
@@ -20,6 +21,12 @@
### Section 1. General Configuration ### Section 1. General Configuration
### ========================================================================== ### ==========================================================================
### Establish the operating system name
KERNEL = $(shell uname -s)
ifeq ($(KERNEL),Linux)
OS = $(shell uname -o)
endif
### Executable name ### Executable name
EXE = stockfish EXE = stockfish
@@ -28,12 +35,12 @@ PREFIX = /usr/local
BINDIR = $(PREFIX)/bin BINDIR = $(PREFIX)/bin
### Built-in benchmark for pgo-builds ### Built-in benchmark for pgo-builds
PGOBENCH = ./$(EXE) bench 32 1 10 default depth PGOBENCH = ./$(EXE) bench
### Object files ### Object files
OBJS = benchmark.o bitbase.o bitboard.o book.o endgame.o evaluate.o main.o \ OBJS = benchmark.o bitbase.o bitboard.o endgame.o evaluate.o main.o \
material.o misc.o move.o movegen.o movepick.o pawns.o position.o \ material.o misc.o movegen.o movepick.o pawns.o position.o psqt.o \
search.o thread.o timeman.o tt.o uci.o ucioption.o search.o thread.o timeman.o tt.o uci.o ucioption.o syzygy/tbprobe.o
### ========================================================================== ### ==========================================================================
### Section 2. High-level Configuration ### Section 2. High-level Configuration
@@ -42,148 +49,89 @@ OBJS = benchmark.o bitbase.o bitboard.o book.o endgame.o evaluate.o main.o \
# flag --- Comp switch --- Description # flag --- Comp switch --- Description
# ---------------------------------------------------------------------------- # ----------------------------------------------------------------------------
# #
# debug = no/yes --- -DNDEBUG --- Enable/Disable debug mode # debug = yes/no --- -DNDEBUG --- Enable/Disable debug mode
# sanitize = undefined/thread/no (-fsanitize )
# --- ( undefined ) --- enable undefined behavior checks
# --- ( thread ) --- enable threading error checks
# optimize = yes/no --- (-O3/-fast etc.) --- Enable/Disable optimizations # optimize = yes/no --- (-O3/-fast etc.) --- Enable/Disable optimizations
# arch = (name) --- (-arch) --- Target architecture # arch = (name) --- (-arch) --- Target architecture
# os = (name) --- --- Target operating system # bits = 64/32 --- -DIS_64BIT --- 64-/32-bit operating system
# bits = 64/32 --- -DIS_64BIT --- 64-/32-bit operating system # prefetch = yes/no --- -DUSE_PREFETCH --- Use prefetch asm-instruction
# bigendian = no/yes --- -DBIGENDIAN --- big/little-endian byte order # popcnt = yes/no --- -DUSE_POPCNT --- Use popcnt asm-instruction
# prefetch = no/yes --- -DUSE_PREFETCH --- Use prefetch x86 asm-instruction # sse = yes/no --- -msse --- Use Intel Streaming SIMD Extensions
# bsfq = no/yes --- -DUSE_BSFQ --- Use bsfq x86_64 asm-instruction # pext = yes/no --- -DUSE_PEXT --- Use pext x86_64 asm-instruction
# --- (Works only with GCC and ICC 64-bit)
# popcnt = no/yes --- -DUSE_POPCNT --- Use popcnt x86_64 asm-instruction
# #
# Note that Makefile is space sensitive, so when adding new architectures # Note that Makefile is space sensitive, so when adding new architectures
# or modifying existing flags, you have to make sure there are no extra spaces # or modifying existing flags, you have to make sure there are no extra spaces
# at the end of the line for flag values. # at the end of the line for flag values.
### 2.1. General ### 2.1. General and architecture defaults
debug = no
optimize = yes optimize = yes
debug = no
sanitize = no
bits = 32
prefetch = no
popcnt = no
sse = no
pext = no
### 2.2 Architecture specific ### 2.2 Architecture specific
# General-section
ifeq ($(ARCH),general-64)
arch = any
os = any
bits = 64
bigendian = no
prefetch = no
bsfq = no
popcnt = no
endif
ifeq ($(ARCH),general-32) ifeq ($(ARCH),general-32)
arch = any arch = any
os = any
bits = 32
bigendian = no
prefetch = no
bsfq = no
popcnt = no
endif
ifeq ($(ARCH),bigendian-64)
arch = any
os = any
bits = 64
bigendian = yes
prefetch = no
bsfq = no
popcnt = no
endif
ifeq ($(ARCH),bigendian-32)
arch = any
os = any
bits = 32
bigendian = yes
prefetch = no
bsfq = no
popcnt = no
endif
# x86-section
ifeq ($(ARCH),x86-64)
arch = x86_64
os = any
bits = 64
bigendian = no
prefetch = yes
bsfq = yes
popcnt = no
endif
ifeq ($(ARCH),x86-64-modern)
arch = x86_64
os = any
bits = 64
bigendian = no
prefetch = yes
bsfq = yes
popcnt = yes
endif
ifeq ($(ARCH),x86-32)
arch = i386
os = any
bits = 32
bigendian = no
prefetch = yes
bsfq = no
popcnt = no
endif endif
ifeq ($(ARCH),x86-32-old) ifeq ($(ARCH),x86-32-old)
arch = i386 arch = i386
os = any
bits = 32
bigendian = no
prefetch = no
bsfq = no
popcnt = no
endif endif
# osx-section ifeq ($(ARCH),x86-32)
ifeq ($(ARCH),osx-ppc-64)
arch = ppc64
os = osx
bits = 64
bigendian = yes
prefetch = no
bsfq = no
popcnt = no
endif
ifeq ($(ARCH),osx-ppc-32)
arch = ppc
os = osx
bits = 32
bigendian = yes
prefetch = no
bsfq = no
popcnt = no
endif
ifeq ($(ARCH),osx-x86-64)
arch = x86_64
os = osx
bits = 64
bigendian = no
prefetch = yes
bsfq = yes
popcnt = no
endif
ifeq ($(ARCH),osx-x86-32)
arch = i386 arch = i386
os = osx
bits = 32
bigendian = no
prefetch = yes prefetch = yes
bsfq = no sse = yes
popcnt = no endif
ifeq ($(ARCH),general-64)
arch = any
bits = 64
endif
ifeq ($(ARCH),x86-64)
arch = x86_64
bits = 64
prefetch = yes
sse = yes
endif
ifeq ($(ARCH),x86-64-modern)
arch = x86_64
bits = 64
prefetch = yes
popcnt = yes
sse = yes
endif
ifeq ($(ARCH),x86-64-bmi2)
arch = x86_64
bits = 64
prefetch = yes
popcnt = yes
sse = yes
pext = yes
endif
ifeq ($(ARCH),armv7)
arch = armv7
prefetch = yes
endif
ifeq ($(ARCH),ppc-32)
arch = ppc
endif
ifeq ($(ARCH),ppc-64)
arch = ppc64
bits = 64
endif endif
@@ -192,75 +140,170 @@ endif
### ========================================================================== ### ==========================================================================
### 3.1 Selecting compiler (default = gcc) ### 3.1 Selecting compiler (default = gcc)
CXXFLAGS += -Wall -Wcast-qual -fno-exceptions -std=c++11 $(EXTRACXXFLAGS)
DEPENDFLAGS += -std=c++11
LDFLAGS += $(EXTRALDFLAGS)
ifeq ($(COMP),) ifeq ($(COMP),)
COMP=gcc COMP=gcc
endif endif
ifeq ($(COMP),mingw)
comp=mingw
CXX=g++
profile_prepare = gcc-profile-prepare
profile_make = gcc-profile-make
profile_use = gcc-profile-use
profile_clean = gcc-profile-clean
endif
ifeq ($(COMP),gcc) ifeq ($(COMP),gcc)
comp=gcc comp=gcc
CXX=g++ CXX=g++
profile_prepare = gcc-profile-prepare CXXFLAGS += -pedantic -Wextra -Wshadow
profile_make = gcc-profile-make
profile_use = gcc-profile-use ifeq ($(ARCH),armv7)
profile_clean = gcc-profile-clean ifeq ($(OS),Android)
CXXFLAGS += -m$(bits)
LDFLAGS += -m$(bits)
endif
else
CXXFLAGS += -m$(bits)
LDFLAGS += -m$(bits)
endif
ifneq ($(KERNEL),Darwin)
LDFLAGS += -Wl,--no-as-needed
endif
endif
ifeq ($(COMP),mingw)
comp=mingw
ifeq ($(KERNEL),Linux)
ifeq ($(bits),64)
ifeq ($(shell which x86_64-w64-mingw32-c++-posix),)
CXX=x86_64-w64-mingw32-c++
else
CXX=x86_64-w64-mingw32-c++-posix
endif
else
ifeq ($(shell which i686-w64-mingw32-c++-posix),)
CXX=i686-w64-mingw32-c++
else
CXX=i686-w64-mingw32-c++-posix
endif
endif
else
CXX=g++
endif
CXXFLAGS += -Wextra -Wshadow
LDFLAGS += -static
endif endif
ifeq ($(COMP),icc) ifeq ($(COMP),icc)
comp=icc comp=icc
CXX=icpc CXX=icpc
profile_prepare = icc-profile-prepare CXXFLAGS += -diag-disable 1476,10120 -Wcheck -Wabi -Wdeprecated -strict-ansi
profile_make = icc-profile-make
profile_use = icc-profile-use
profile_clean = icc-profile-clean
endif endif
### 3.2 General compiler settings ifeq ($(COMP),clang)
CXXFLAGS = -g -Wall -Wcast-qual -fno-exceptions -fno-rtti $(EXTRACXXFLAGS) comp=clang
CXX=clang++
ifeq ($(comp),gcc) CXXFLAGS += -pedantic -Wextra -Wshadow
CXXFLAGS += -ansi -pedantic -Wno-long-long -Wextra -Wshadow ifneq ($(KERNEL),Darwin)
ifneq ($(KERNEL),OpenBSD)
LDFLAGS += -latomic
endif
endif endif
ifeq ($(comp),mingw) ifeq ($(ARCH),armv7)
CXXFLAGS += -Wextra -Wshadow ifeq ($(OS),Android)
CXXFLAGS += -m$(bits)
LDFLAGS += -m$(bits)
endif
else
CXXFLAGS += -m$(bits)
LDFLAGS += -m$(bits)
endif
endif endif
ifeq ($(comp),icc) ifeq ($(comp),icc)
CXXFLAGS += -wd383,981,1418,1419,10187,10188,11505,11503 -Wcheck -Wabi -Wdeprecated -strict-ansi profile_make = icc-profile-make
profile_use = icc-profile-use
else
ifeq ($(comp),clang)
profile_make = clang-profile-make
profile_use = clang-profile-use
else
profile_make = gcc-profile-make
profile_use = gcc-profile-use
endif
endif endif
ifeq ($(os),osx) ifeq ($(KERNEL),Darwin)
CXXFLAGS += -arch $(arch) CXXFLAGS += -arch $(arch) -mmacosx-version-min=10.9
LDFLAGS += -arch $(arch) -mmacosx-version-min=10.9
endif endif
### 3.3 General linker settings ### Travis CI script uses COMPILER to overwrite CXX
LDFLAGS = -lpthread $(EXTRALDFLAGS) ifdef COMPILER
COMPCXX=$(COMPILER)
ifeq ($(os),osx)
LDFLAGS += -arch $(arch)
endif endif
### 3.4 Debugging ### Allow overwriting CXX from command line
ifdef COMPCXX
CXX=$(COMPCXX)
endif
### On mingw use Windows threads, otherwise POSIX
ifneq ($(comp),mingw)
# On Android Bionic's C library comes with its own pthread implementation bundled in
ifneq ($(OS),Android)
# Haiku has pthreads in its libroot, so only link it in on other platforms
ifneq ($(KERNEL),Haiku)
LDFLAGS += -lpthread
endif
endif
endif
### 3.2.1 Debugging
ifeq ($(debug),no) ifeq ($(debug),no)
CXXFLAGS += -DNDEBUG CXXFLAGS += -DNDEBUG
else
CXXFLAGS += -g
endif endif
### 3.5 Optimization ### 3.2.2 Debugging with undefined behavior sanitizers
ifneq ($(sanitize),no)
CXXFLAGS += -g3 -fsanitize=$(sanitize) -fuse-ld=gold
LDFLAGS += -fsanitize=$(sanitize) -fuse-ld=gold
endif
### 3.3 Optimization
ifeq ($(optimize),yes) ifeq ($(optimize),yes)
ifeq ($(comp),gcc) CXXFLAGS += -O3
CXXFLAGS += -O3
ifeq ($(os),osx) ifeq ($(comp),gcc)
ifeq ($(KERNEL),Darwin)
ifeq ($(arch),i386)
CXXFLAGS += -mdynamic-no-pic
endif
ifeq ($(arch),x86_64)
CXXFLAGS += -mdynamic-no-pic
endif
endif
ifeq ($(OS), Android)
CXXFLAGS += -fno-gcse -mthumb -march=armv7-a -mfloat-abi=softfp
endif
endif
ifeq ($(comp),icc)
ifeq ($(KERNEL),Darwin)
CXXFLAGS += -mdynamic-no-pic
endif
endif
ifeq ($(comp),clang)
ifeq ($(KERNEL),Darwin)
CXXFLAGS += -flto
LDFLAGS += $(CXXFLAGS)
ifeq ($(arch),i386) ifeq ($(arch),i386)
CXXFLAGS += -mdynamic-no-pic CXXFLAGS += -mdynamic-no-pic
endif endif
@@ -269,153 +312,142 @@ ifeq ($(optimize),yes)
endif endif
endif endif
endif endif
ifeq ($(comp),mingw)
CXXFLAGS += -O3
endif
ifeq ($(comp),icc)
ifeq ($(os),osx)
CXXFLAGS += -fast -mdynamic-no-pic
else
CXXFLAGS += -O3
endif
endif
endif endif
### 3.6. Bits ### 3.4 Bits
ifeq ($(bits),64) ifeq ($(bits),64)
CXXFLAGS += -DIS_64BIT CXXFLAGS += -DIS_64BIT
endif endif
### 3.7 Endianess ### 3.5 prefetch
ifeq ($(bigendian),yes)
CXXFLAGS += -DBIGENDIAN
endif
### 3.8 prefetch
ifeq ($(prefetch),yes) ifeq ($(prefetch),yes)
CXXFLAGS += -msse ifeq ($(sse),yes)
DEPENDFLAGS += -msse CXXFLAGS += -msse
DEPENDFLAGS += -msse
endif
else else
CXXFLAGS += -DNO_PREFETCH CXXFLAGS += -DNO_PREFETCH
endif endif
### 3.9 bsfq ### 3.6 popcnt
ifeq ($(bsfq),yes) ifeq ($(popcnt),yes)
CXXFLAGS += -DUSE_BSFQ ifeq ($(comp),icc)
CXXFLAGS += -msse3 -DUSE_POPCNT
else
CXXFLAGS += -msse3 -mpopcnt -DUSE_POPCNT
endif
endif endif
### 3.10 popcnt ### 3.7 pext
ifeq ($(popcnt),yes) ifeq ($(pext),yes)
CXXFLAGS += -DUSE_POPCNT CXXFLAGS += -DUSE_PEXT
ifeq ($(comp),$(filter $(comp),gcc clang mingw))
CXXFLAGS += -mbmi2
endif
endif endif
### 3.8 Link Time Optimization, it works since gcc 4.5 but not on mingw under Windows.
### This is a mix of compile and link time options because the lto link phase
### needs access to the optimization flags.
ifeq ($(comp),gcc)
ifeq ($(optimize),yes)
ifeq ($(debug),no)
CXXFLAGS += -flto
LDFLAGS += $(CXXFLAGS)
endif
endif
endif
ifeq ($(comp),mingw)
ifeq ($(KERNEL),Linux)
ifeq ($(optimize),yes)
ifeq ($(debug),no)
CXXFLAGS += -flto
LDFLAGS += $(CXXFLAGS)
endif
endif
endif
endif
### 3.9 Android 5 can only run position independent executables. Note that this
### breaks Android 4.0 and earlier.
ifeq ($(OS), Android)
CXXFLAGS += -fPIE
LDFLAGS += -fPIE -pie
endif
### ========================================================================== ### ==========================================================================
### Section 4. Public targets ### Section 4. Public targets
### ========================================================================== ### ==========================================================================
default:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) build
help: help:
@echo "" @echo ""
@echo "To compile stockfish, type: " @echo "To compile stockfish, type: "
@echo "" @echo ""
@echo "make target ARCH=arch [COMP=comp]" @echo "make target ARCH=arch [COMP=compiler] [COMPCXX=cxx]"
@echo "" @echo ""
@echo "Supported targets:" @echo "Supported targets:"
@echo "" @echo ""
@echo "build > Build unoptimized version" @echo "build > Standard build"
@echo "profile-build > Build PGO-optimized version" @echo "profile-build > PGO build"
@echo "popcnt-profile-build > Build PGO-optimized version with optional popcnt-support" @echo "strip > Strip executable"
@echo "strip > Strip executable" @echo "install > Install executable"
@echo "install > Install executable" @echo "clean > Clean up"
@echo "clean > Clean up"
@echo "testrun > Make sample run"
@echo "" @echo ""
@echo "Supported archs:" @echo "Supported archs:"
@echo "" @echo ""
@echo "x86-64 > x86 64-bit" @echo "x86-64 > x86 64-bit"
@echo "x86-64-modern > x86 64-bit with runtime support for popcnt-instruction" @echo "x86-64-modern > x86 64-bit with popcnt support"
@echo "x86-32 > x86 32-bit excluding very old hardware without SSE-support" @echo "x86-64-bmi2 > x86 64-bit with pext support"
@echo "x86-32-old > x86 32-bit including also very old hardware" @echo "x86-32 > x86 32-bit with SSE support"
@echo "osx-ppc-64 > PPC-Mac OS X 64 bit" @echo "x86-32-old > x86 32-bit fall back for old hardware"
@echo "osx-ppc-32 > PPC-Mac OS X 32 bit" @echo "ppc-64 > PPC 64-bit"
@echo "osx-x86-64 > x86-Mac OS X 64 bit" @echo "ppc-32 > PPC 32-bit"
@echo "osx-x86-32 > x86-Mac OS X 32 bit" @echo "armv7 > ARMv7 32-bit"
@echo "general-64 > unspecified 64-bit" @echo "general-64 > unspecified 64-bit"
@echo "general-32 > unspecified 32-bit" @echo "general-32 > unspecified 32-bit"
@echo "bigendian-64 > unspecified 64-bit with bigendian byte order"
@echo "bigendian-32 > unspecified 32-bit with bigendian byte order"
@echo "" @echo ""
@echo "Supported comps:" @echo "Supported compilers:"
@echo "" @echo ""
@echo "gcc > Gnu compiler (default)" @echo "gcc > Gnu compiler (default)"
@echo "icc > Intel compiler" @echo "mingw > Gnu compiler with MinGW under Windows"
@echo "mingw > Gnu compiler with MinGW under Windows" @echo "clang > LLVM Clang compiler"
@echo "icc > Intel compiler"
@echo "" @echo ""
@echo "Non-standard targets:" @echo "Simple examples. If you don't know what to do, you likely want to run: "
@echo "" @echo ""
@echo "make hpux > Compile for HP-UX. Compiler = aCC" @echo "make build ARCH=x86-64 (This is for 64-bit systems)"
@echo "make build ARCH=x86-32 (This is for 32-bit systems)"
@echo "" @echo ""
@echo "Examples. If you don't know what to do, you likely want to run: " @echo "Advanced examples, for experienced users: "
@echo "" @echo ""
@echo "make profile-build ARCH=x86-64 (This is for 64-bit systems)" @echo "make build ARCH=x86-64 COMP=clang"
@echo "make profile-build ARCH=x86-32 (This is for 32-bit systems)" @echo "make profile-build ARCH=x86-64-modern COMP=gcc COMPCXX=g++-4.8"
@echo "" @echo ""
build:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) config-sanity .PHONY: help build profile-build strip install clean objclean profileclean help \
config-sanity icc-profile-use icc-profile-make gcc-profile-use gcc-profile-make \
clang-profile-use clang-profile-make
build: config-sanity
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) all $(MAKE) ARCH=$(ARCH) COMP=$(COMP) all
profile-build: profile-build: config-sanity objclean profileclean
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) config-sanity
@echo "" @echo ""
@echo "Step 0/4. Preparing for profile build." @echo "Step 1/4. Building instrumented executable ..."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_prepare)
@echo ""
@echo "Step 1/4. Building executable for benchmark ..."
@touch *.cpp *.h
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_make) $(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_make)
@echo "" @echo ""
@echo "Step 2/4. Running benchmark for pgo-build ..." @echo "Step 2/4. Running benchmark for pgo-build ..."
@$(PGOBENCH) > /dev/null $(PGOBENCH) > /dev/null
@echo "" @echo ""
@echo "Step 3/4. Building final executable ..." @echo "Step 3/4. Building optimized executable ..."
@touch *.cpp $(MAKE) ARCH=$(ARCH) COMP=$(COMP) objclean
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_use) $(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_use)
@echo "" @echo ""
@echo "Step 4/4. Deleting profile data ..." @echo "Step 4/4. Deleting profile data ..."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_clean) $(MAKE) ARCH=$(ARCH) COMP=$(COMP) profileclean
popcnt-profile-build:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) config-sanity
@echo ""
@echo "Step 0/6. Preparing for profile build."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_prepare)
@echo ""
@echo "Step 1/6. Building executable for benchmark (popcnt disabled)..."
@touch *.cpp *.h
$(MAKE) ARCH=x86-64 COMP=$(COMP) $(profile_make)
@echo ""
@echo "Step 2/6. Running benchmark for pgo-build (popcnt disabled)..."
@$(PGOBENCH) > /dev/null
@echo ""
@echo "Step 3/6. Building executable for benchmark (popcnt enabled)..."
@touch *.cpp *.h
$(MAKE) ARCH=x86-64-modern COMP=$(COMP) $(profile_make)
@echo ""
@echo "Step 4/6. Running benchmark for pgo-build (popcnt enabled)..."
@$(PGOBENCH) > /dev/null
@echo ""
@echo "Step 5/6. Building final executable ..."
@touch *.cpp *.h
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_use)
@echo ""
@echo "Step 6/6. Deleting profile data ..."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_clean)
@echo ""
strip: strip:
strip $(EXE) strip $(EXE)
@@ -425,11 +457,22 @@ install:
-cp $(EXE) $(BINDIR) -cp $(EXE) $(BINDIR)
-strip $(BINDIR)/$(EXE) -strip $(BINDIR)/$(EXE)
clean: #clean all
$(RM) $(EXE) $(EXE).exe *.o .depend *~ core bench.txt *.gcda clean: objclean profileclean
@rm -f .depend *~ core
testrun: # clean binaries and objects
@$(PGOBENCH) objclean:
@rm -f $(EXE) $(EXE).exe *.o ./syzygy/*.o
# clean auxiliary profiling files
profileclean:
@rm -rf profdir
@rm -f bench.txt *.gcda ./syzygy/*.gcda *.gcno ./syzygy/*.gcno
@rm -f stockfish.profdata *.profraw
default:
help
### ========================================================================== ### ==========================================================================
### Section 5. Private targets ### Section 5. Private targets
@@ -441,14 +484,16 @@ config-sanity:
@echo "" @echo ""
@echo "Config:" @echo "Config:"
@echo "debug: '$(debug)'" @echo "debug: '$(debug)'"
@echo "sanitize: '$(sanitize)'"
@echo "optimize: '$(optimize)'" @echo "optimize: '$(optimize)'"
@echo "arch: '$(arch)'" @echo "arch: '$(arch)'"
@echo "os: '$(os)'"
@echo "bits: '$(bits)'" @echo "bits: '$(bits)'"
@echo "bigendian: '$(bigendian)'" @echo "kernel: '$(KERNEL)'"
@echo "os: '$(OS)'"
@echo "prefetch: '$(prefetch)'" @echo "prefetch: '$(prefetch)'"
@echo "bsfq: '$(bsfq)'"
@echo "popcnt: '$(popcnt)'" @echo "popcnt: '$(popcnt)'"
@echo "sse: '$(sse)'"
@echo "pext: '$(pext)'"
@echo "" @echo ""
@echo "Flags:" @echo "Flags:"
@echo "CXX: $(CXX)" @echo "CXX: $(CXX)"
@@ -458,22 +503,32 @@ config-sanity:
@echo "Testing config sanity. If this fails, try 'make help' ..." @echo "Testing config sanity. If this fails, try 'make help' ..."
@echo "" @echo ""
@test "$(debug)" = "yes" || test "$(debug)" = "no" @test "$(debug)" = "yes" || test "$(debug)" = "no"
@test "$(sanitize)" = "undefined" || test "$(sanitize)" = "thread" || test "$(sanitize)" = "no"
@test "$(optimize)" = "yes" || test "$(optimize)" = "no" @test "$(optimize)" = "yes" || test "$(optimize)" = "no"
@test "$(arch)" = "any" || test "$(arch)" = "x86_64" || test "$(arch)" = "i386" || \ @test "$(arch)" = "any" || test "$(arch)" = "x86_64" || test "$(arch)" = "i386" || \
test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || test "$(arch)" = "armv7"
@test "$(os)" = "any" || test "$(os)" = "osx"
@test "$(bits)" = "32" || test "$(bits)" = "64" @test "$(bits)" = "32" || test "$(bits)" = "64"
@test "$(bigendian)" = "yes" || test "$(bigendian)" = "no"
@test "$(prefetch)" = "yes" || test "$(prefetch)" = "no" @test "$(prefetch)" = "yes" || test "$(prefetch)" = "no"
@test "$(bsfq)" = "yes" || test "$(bsfq)" = "no"
@test "$(popcnt)" = "yes" || test "$(popcnt)" = "no" @test "$(popcnt)" = "yes" || test "$(popcnt)" = "no"
@test "$(comp)" = "gcc" || test "$(comp)" = "icc" || test "$(comp)" = "mingw" @test "$(sse)" = "yes" || test "$(sse)" = "no"
@test "$(pext)" = "yes" || test "$(pext)" = "no"
@test "$(comp)" = "gcc" || test "$(comp)" = "icc" || test "$(comp)" = "mingw" || test "$(comp)" = "clang"
$(EXE): $(OBJS) $(EXE): $(OBJS)
$(CXX) -o $@ $(OBJS) $(LDFLAGS) $(CXX) -o $@ $(OBJS) $(LDFLAGS)
gcc-profile-prepare: clang-profile-make:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) gcc-profile-clean $(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
EXTRACXXFLAGS='-fprofile-instr-generate ' \
EXTRALDFLAGS=' -fprofile-instr-generate' \
all
clang-profile-use:
llvm-profdata merge -output=stockfish.profdata *.profraw
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
EXTRACXXFLAGS='-fprofile-instr-use=stockfish.profdata' \
EXTRALDFLAGS='-fprofile-use ' \
all
gcc-profile-make: gcc-profile-make:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \ $(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
@@ -483,17 +538,12 @@ gcc-profile-make:
gcc-profile-use: gcc-profile-use:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \ $(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
EXTRACXXFLAGS='-fprofile-use' \ EXTRACXXFLAGS='-fprofile-use -fno-peel-loops -fno-tracer' \
EXTRALDFLAGS='-lgcov' \
all all
gcc-profile-clean:
@rm -rf *.gcda *.gcno bench.txt
icc-profile-prepare:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) icc-profile-clean
@mkdir profdir
icc-profile-make: icc-profile-make:
@mkdir -p profdir
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \ $(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
EXTRACXXFLAGS='-prof-gen=srcpos -prof_dir ./profdir' \ EXTRACXXFLAGS='-prof-gen=srcpos -prof_dir ./profdir' \
all all
@@ -503,23 +553,8 @@ icc-profile-use:
EXTRACXXFLAGS='-prof_use -prof_dir ./profdir' \ EXTRACXXFLAGS='-prof_use -prof_dir ./profdir' \
all all
icc-profile-clean:
@rm -rf profdir bench.txt
.depend: .depend:
-@$(CXX) $(DEPENDFLAGS) -MM $(OBJS:.o=.cpp) > $@ 2> /dev/null -@$(CXX) $(DEPENDFLAGS) -MM $(OBJS:.o=.cpp) > $@ 2> /dev/null
-include .depend -include .depend
### ==========================================================================
### Section 6. Non-standard targets
### ==========================================================================
hpux:
$(MAKE) \
CXX='/opt/aCC/bin/aCC -AA +hpxstd98 -DBIGENDIAN -mt +O3 -DNDEBUG -DNO_PREFETCH' \
CXXFLAGS="" \
LDFLAGS="" \
all
+94 -91
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -19,21 +20,23 @@
#include <fstream> #include <fstream>
#include <iostream> #include <iostream>
#include <istream>
#include <vector> #include <vector>
#include "position.h" #include "position.h"
#include "search.h"
#include "ucioption.h"
using namespace std; using namespace std;
static const string Defaults[] = { namespace {
const vector<string> Defaults = {
"setoption name UCI_Chess960 value false",
"rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1", "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1",
"r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -", "r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq - 0 10",
"8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - -", "8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - - 0 11",
"4rrk1/pp1n3p/3q2pQ/2p1pb2/2PP4/2P3N1/P2B2PP/4RRK1 b - - 7 19", "4rrk1/pp1n3p/3q2pQ/2p1pb2/2PP4/2P3N1/P2B2PP/4RRK1 b - - 7 19",
"rq3rk1/ppp2ppp/1bnpb3/3N2B1/3NP3/7P/PPPQ1PP1/2KR3R w - - 7 14", "rq3rk1/ppp2ppp/1bnpb3/3N2B1/3NP3/7P/PPPQ1PP1/2KR3R w - - 7 14 moves d4e6",
"r1bq1r1k/1pp1n1pp/1p1p4/4p2Q/4Pp2/1BNP4/PPP2PPP/3R1RK1 w - - 2 14", "r1bq1r1k/1pp1n1pp/1p1p4/4p2Q/4Pp2/1BNP4/PPP2PPP/3R1RK1 w - - 2 14 moves g2g4",
"r3r1k1/2p2ppp/p1p1bn2/8/1q2P3/2NPQN2/PPP3PP/R4RK1 b - - 2 15", "r3r1k1/2p2ppp/p1p1bn2/8/1q2P3/2NPQN2/PPP3PP/R4RK1 b - - 2 15",
"r1bbk1nr/pp3p1p/2n5/1N4p1/2Np1B2/8/PPP2PPP/2KR1B1R w kq - 0 13", "r1bbk1nr/pp3p1p/2n5/1N4p1/2Np1B2/8/PPP2PPP/2KR1B1R w kq - 0 13",
"r1bq1rk1/ppp1nppp/4n3/3p3Q/3P4/1BP1B3/PP1N2PP/R4RK1 w - - 1 16", "r1bq1rk1/ppp1nppp/4n3/3p3Q/3P4/1BP1B3/PP1N2PP/R4RK1 w - - 1 16",
@@ -44,110 +47,110 @@ static const string Defaults[] = {
"r1q2rk1/2p1bppp/2Pp4/p6b/Q1PNp3/4B3/PP1R1PPP/2K4R w - - 2 18", "r1q2rk1/2p1bppp/2Pp4/p6b/Q1PNp3/4B3/PP1R1PPP/2K4R w - - 2 18",
"4k2r/1pb2ppp/1p2p3/1R1p4/3P4/2r1PN2/P4PPP/1R4K1 b - - 3 22", "4k2r/1pb2ppp/1p2p3/1R1p4/3P4/2r1PN2/P4PPP/1R4K1 b - - 3 22",
"3q2k1/pb3p1p/4pbp1/2r5/PpN2N2/1P2P2P/5PP1/Q2R2K1 b - - 4 26", "3q2k1/pb3p1p/4pbp1/2r5/PpN2N2/1P2P2P/5PP1/Q2R2K1 b - - 4 26",
"" "6k1/6p1/6Pp/ppp5/3pn2P/1P3K2/1PP2P2/3N4 b - - 0 1",
"3b4/5kp1/1p1p1p1p/pP1PpP1P/P1P1P3/3KN3/8/8 w - - 0 1",
"2K5/p7/7P/5pR1/8/5k2/r7/8 w - - 0 1 moves g5g6 f3e3 g6g5 e3f3",
"8/6pk/1p6/8/PP3p1p/5P2/4KP1q/3Q4 w - - 0 1",
"7k/3p2pp/4q3/8/4Q3/5Kp1/P6b/8 w - - 0 1",
"8/2p5/8/2kPKp1p/2p4P/2P5/3P4/8 w - - 0 1",
"8/1p3pp1/7p/5P1P/2k3P1/8/2K2P2/8 w - - 0 1",
"8/pp2r1k1/2p1p3/3pP2p/1P1P1P1P/P5KR/8/8 w - - 0 1",
"8/3p4/p1bk3p/Pp6/1Kp1PpPp/2P2P1P/2P5/5B2 b - - 0 1",
"5k2/7R/4P2p/5K2/p1r2P1p/8/8/8 b - - 0 1",
"6k1/6p1/P6p/r1N5/5p2/7P/1b3PP1/4R1K1 w - - 0 1",
"1r3k2/4q3/2Pp3b/3Bp3/2Q2p2/1p1P2P1/1P2KP2/3N4 w - - 0 1",
"6k1/4pp1p/3p2p1/P1pPb3/R7/1r2P1PP/3B1P2/6K1 w - - 0 1",
"8/3p3B/5p2/5P2/p7/PP5b/k7/6K1 w - - 0 1",
// 5-man positions
"8/8/8/8/5kp1/P7/8/1K1N4 w - - 0 1", // Kc2 - mate
"8/8/8/5N2/8/p7/8/2NK3k w - - 0 1", // Na2 - mate
"8/3k4/8/8/8/4B3/4KB2/2B5 w - - 0 1", // draw
// 6-man positions
"8/8/1P6/5pr1/8/4R3/7k/2K5 w - - 0 1", // Re5 - mate
"8/2p4P/8/kr6/6R1/8/8/1K6 w - - 0 1", // Ka2 - mate
"8/8/3P3k/8/1p6/8/1P6/1K3n2 b - - 0 1", // Nd2 - draw
// 7-man positions
"8/R7/2q5/8/6k1/8/1P5p/K6R w - - 0 124", // Draw
// Mate and stalemate positions
"6k1/3b3r/1p1p4/p1n2p2/1PPNpP1q/P3Q1p1/1R1RB1P1/5K2 b - - 0 1",
"r2r1n2/pp2bk2/2p1p2p/3q4/3PN1QP/2P3R1/P4PP1/5RK1 w - - 0 1",
"8/8/8/8/8/6k1/6p1/6K1 w - -",
"7k/7P/6K1/8/3B4/8/8/8 b - -",
// Chess 960
"setoption name UCI_Chess960 value true",
"bbqnnrkr/pppppppp/8/8/8/8/PPPPPPPP/BBQNNRKR w KQkq - 0 1 moves g2g3 d7d5 d2d4 c8h3 c1g5 e8d6 g5e7 f7f6",
"setoption name UCI_Chess960 value false"
}; };
} // namespace
/// benchmark() runs a simple benchmark by letting Stockfish analyze a set /// setup_bench() builds a list of UCI commands to be run by bench. There
/// of positions for a given limit each. There are five parameters; the /// are five parameters: TT size in MB, number of search threads that
/// transposition table size, the number of search threads that should /// should be used, the limit value spent for each position, a file name
/// be used, the limit value spent for each position (optional, default /// where to look for positions in FEN format and the type of the limit:
/// is ply 12), an optional file name where to look for positions in fen /// depth, perft, nodes and movetime (in millisecs).
/// format (default are the BenchmarkPositions defined above) and the type ///
/// of the limit value: depth (default), time in secs or number of nodes. /// bench -> search default positions up to depth 13
/// The analysis is written to a file named bench.txt. /// bench 64 1 15 -> search default positions up to depth 15 (TT = 64MB)
/// bench 64 4 5000 current movetime -> search current position with 4 threads for 5 sec
/// bench 64 1 100000 default nodes -> search default positions for 100K nodes each
/// bench 16 1 5 default perft -> run a perft 5 on default positions
void benchmark(int argc, char* argv[]) { vector<string> setup_bench(const Position& current, istream& is) {
vector<string> fenList; vector<string> fens, list;
SearchLimits limits; string go, token;
int64_t totalNodes;
int time;
// Load default positions
for (int i = 0; !Defaults[i].empty(); i++)
fenList.push_back(Defaults[i]);
// Assign default values to missing arguments // Assign default values to missing arguments
string ttSize = argc > 2 ? argv[2] : "128"; string ttSize = (is >> token) ? token : "16";
string threads = argc > 3 ? argv[3] : "1"; string threads = (is >> token) ? token : "1";
string valStr = argc > 4 ? argv[4] : "12"; string limit = (is >> token) ? token : "13";
string fenFile = argc > 5 ? argv[5] : "default"; string fenFile = (is >> token) ? token : "default";
string valType = argc > 6 ? argv[6] : "depth"; string limitType = (is >> token) ? token : "depth";
Options["Hash"].set_value(ttSize); go = "go " + limitType + " " + limit;
Options["Threads"].set_value(threads);
Options["OwnBook"].set_value("false"); if (fenFile == "default")
fens = Defaults;
else if (fenFile == "current")
fens.push_back(current.fen());
// Search should be limited by nodes, time or depth ?
if (valType == "nodes")
limits.maxNodes = atoi(valStr.c_str());
else if (valType == "time")
limits.maxTime = 1000 * atoi(valStr.c_str()); // maxTime is in ms
else else
limits.maxDepth = atoi(valStr.c_str());
// Do we need to load positions from a given FEN file ?
if (fenFile != "default")
{ {
string fen; string fen;
ifstream f(fenFile.c_str()); ifstream file(fenFile);
if (f.is_open()) if (!file.is_open())
{ {
fenList.clear(); cerr << "Unable to open file " << fenFile << endl;
while (getline(f, fen))
if (!fen.empty())
fenList.push_back(fen);
f.close();
}
else
{
cerr << "Unable to open FEN file " << fenFile << endl;
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
while (getline(file, fen))
if (!fen.empty())
fens.push_back(fen);
file.close();
} }
// Ok, let's start the benchmark ! list.emplace_back("ucinewgame");
totalNodes = 0; list.emplace_back("setoption name Threads value " + threads);
time = get_system_time(); list.emplace_back("setoption name Hash value " + ttSize);
for (size_t i = 0; i < fenList.size(); i++) for (const string& fen : fens)
{ if (fen.find("setoption") != string::npos)
Move moves[] = { MOVE_NONE }; list.emplace_back(fen);
Position pos(fenList[i], false, 0);
cerr << "\nBench position: " << i + 1 << '/' << fenList.size() << endl;
if (valType == "perft")
{
int64_t cnt = perft(pos, limits.maxDepth * ONE_PLY);
totalNodes += cnt;
cerr << "\nPerft " << limits.maxDepth << " nodes counted: " << cnt << endl;
}
else else
{ {
if (!think(pos, limits, moves)) list.emplace_back("position fen " + fen);
break; list.emplace_back(go);
totalNodes += pos.nodes_searched();
} }
}
time = get_system_time() - time; return list;
cerr << "\n==============================="
<< "\nTotal time (ms) : " << time
<< "\nNodes searched : " << totalNodes
<< "\nNodes/second : " << (int)(totalNodes / (time / 1000.0)) << endl << endl;
// MS Visual C++ debug window always unconditionally closes when program
// exits, this is bad because we want to read results before.
#if (defined(WINDOWS) || defined(WIN32) || defined(WIN64))
cerr << "Press any key to exit" << endl;
cin >> time;
#endif
} }
+122 -226
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,268 +18,163 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <algorithm>
#include <cassert> #include <cassert>
#include <numeric>
#include <vector>
#include "bitboard.h" #include "bitboard.h"
#include "types.h" #include "types.h"
namespace { namespace {
enum Result { // There are 24 possible pawn squares: the first 4 files and ranks from 2 to 7
RESULT_UNKNOWN, const unsigned MAX_INDEX = 2*24*64*64; // stm * psq * wksq * bksq = 196608
RESULT_INVALID,
RESULT_WIN,
RESULT_LOSS,
RESULT_DRAW
};
struct KPKPosition {
void from_index(int index);
bool is_legal() const;
bool is_immediate_draw() const;
bool is_immediate_win() const;
Bitboard wk_attacks() const { return StepAttacksBB[WK][whiteKingSquare]; }
Bitboard bk_attacks() const { return StepAttacksBB[BK][blackKingSquare]; }
Bitboard pawn_attacks() const { return StepAttacksBB[WP][pawnSquare]; }
Square whiteKingSquare, blackKingSquare, pawnSquare;
Color sideToMove;
};
// The possible pawns squares are 24, the first 4 files and ranks from 2 to 7
const int IndexMax = 2 * 24 * 64 * 64; // color * wp_sq * wk_sq * bk_sq
// Each uint32_t stores results of 32 positions, one per bit // Each uint32_t stores results of 32 positions, one per bit
uint32_t KPKBitbase[IndexMax / 32]; uint32_t KPKBitbase[MAX_INDEX / 32];
Result classify_wtm(const KPKPosition& pos, const Result bb[]); // A KPK bitbase index is an integer in [0, IndexMax] range
Result classify_btm(const KPKPosition& pos, const Result bb[]); //
int compute_index(Square wksq, Square bksq, Square wpsq, Color stm); // Information is mapped in a way that minimizes the number of iterations:
} //
// bit 0- 5: white king square (from SQ_A1 to SQ_H8)
// bit 6-11: black king square (from SQ_A1 to SQ_H8)
uint32_t probe_kpk_bitbase(Square wksq, Square wpsq, Square bksq, Color stm) { // bit 12: side to move (WHITE or BLACK)
// bit 13-14: white pawn file (from FILE_A to FILE_D)
int index = compute_index(wksq, bksq, wpsq, stm); // bit 15-17: white pawn RANK_7 - rank (from RANK_7 - RANK_7 to RANK_7 - RANK_2)
unsigned index(Color us, Square bksq, Square wksq, Square psq) {
return KPKBitbase[index / 32] & (1 << (index & 31)); return wksq | (bksq << 6) | (us << 12) | (file_of(psq) << 13) | ((RANK_7 - rank_of(psq)) << 15);
}
void init_kpk_bitbase() {
Result bb[IndexMax];
KPKPosition pos;
bool repeat;
// Initialize table
for (int i = 0; i < IndexMax; i++)
{
pos.from_index(i);
bb[i] = !pos.is_legal() ? RESULT_INVALID
: pos.is_immediate_draw() ? RESULT_DRAW
: pos.is_immediate_win() ? RESULT_WIN : RESULT_UNKNOWN;
} }
// Iterate until all positions are classified (30 cycles needed) enum Result {
do { INVALID = 0,
repeat = false; UNKNOWN = 1,
DRAW = 2,
WIN = 4
};
for (int i = 0; i < IndexMax; i++) Result& operator|=(Result& r, Result v) { return r = Result(r | v); }
if (bb[i] == RESULT_UNKNOWN)
{
pos.from_index(i);
bb[i] = (pos.sideToMove == WHITE) ? classify_wtm(pos, bb) struct KPKPosition {
: classify_btm(pos, bb); KPKPosition() = default;
if (bb[i] != RESULT_UNKNOWN) explicit KPKPosition(unsigned idx);
repeat = true; operator Result() const { return result; }
} Result classify(const std::vector<KPKPosition>& db)
{ return us == WHITE ? classify<WHITE>(db) : classify<BLACK>(db); }
} while (repeat); template<Color Us> Result classify(const std::vector<KPKPosition>& db);
// Map 32 position results into one KPKBitbase[] entry Color us;
for (int i = 0; i < IndexMax / 32; i++) Square ksq[COLOR_NB], psq;
for (int j = 0; j < 32; j++) Result result;
if (bb[32 * i + j] == RESULT_WIN || bb[32 * i + j] == RESULT_LOSS) };
KPKBitbase[i] |= (1 << j);
} // namespace
bool Bitbases::probe(Square wksq, Square wpsq, Square bksq, Color us) {
assert(file_of(wpsq) <= FILE_D);
unsigned idx = index(us, bksq, wksq, wpsq);
return KPKBitbase[idx / 32] & (1 << (idx & 0x1F));
}
void Bitbases::init() {
std::vector<KPKPosition> db(MAX_INDEX);
unsigned idx, repeat = 1;
// Initialize db with known win / draw positions
for (idx = 0; idx < MAX_INDEX; ++idx)
db[idx] = KPKPosition(idx);
// Iterate through the positions until none of the unknown positions can be
// changed to either wins or draws (15 cycles needed).
while (repeat)
for (repeat = idx = 0; idx < MAX_INDEX; ++idx)
repeat |= (db[idx] == UNKNOWN && db[idx].classify(db) != UNKNOWN);
// Map 32 results into one KPKBitbase[] entry
for (idx = 0; idx < MAX_INDEX; ++idx)
if (db[idx] == WIN)
KPKBitbase[idx / 32] |= 1 << (idx & 0x1F);
} }
namespace { namespace {
// A KPK bitbase index is an integer in [0, IndexMax] range KPKPosition::KPKPosition(unsigned idx) {
//
// Information is mapped in this way
//
// bit 0: side to move (WHITE or BLACK)
// bit 1- 6: black king square (from SQ_A1 to SQ_H8)
// bit 7-12: white king square (from SQ_A1 to SQ_H8)
// bit 13-14: white pawn file (from FILE_A to FILE_D)
// bit 15-17: white pawn rank - 1 (from RANK_2 - 1 to RANK_7 - 1)
int compute_index(Square wksq, Square bksq, Square wpsq, Color stm) { ksq[WHITE] = Square((idx >> 0) & 0x3F);
ksq[BLACK] = Square((idx >> 6) & 0x3F);
us = Color ((idx >> 12) & 0x01);
psq = make_square(File((idx >> 13) & 0x3), Rank(RANK_7 - ((idx >> 15) & 0x7)));
assert(square_file(wpsq) <= FILE_D); // Check if two pieces are on the same square or if a king can be captured
if ( distance(ksq[WHITE], ksq[BLACK]) <= 1
|| ksq[WHITE] == psq
|| ksq[BLACK] == psq
|| (us == WHITE && (PawnAttacks[WHITE][psq] & ksq[BLACK])))
result = INVALID;
int p = int(square_file(wpsq)) + 4 * int(square_rank(wpsq) - 1); // Immediate win if a pawn can be promoted without getting captured
int r = int(stm) + 2 * int(bksq) + 128 * int(wksq) + 8192 * p; else if ( us == WHITE
&& rank_of(psq) == RANK_7
&& ksq[us] != psq + NORTH
&& ( distance(ksq[~us], psq + NORTH) > 1
|| (PseudoAttacks[KING][ksq[us]] & (psq + NORTH))))
result = WIN;
assert(r >= 0 && r < IndexMax); // Immediate draw if it is a stalemate or a king captures undefended pawn
else if ( us == BLACK
&& ( !(PseudoAttacks[KING][ksq[us]] & ~(PseudoAttacks[KING][ksq[~us]] | PawnAttacks[~us][psq]))
|| (PseudoAttacks[KING][ksq[us]] & psq & ~PseudoAttacks[KING][ksq[~us]])))
result = DRAW;
return r; // Position will be classified later
}
void KPKPosition::from_index(int index) {
int s = (index / 8192) % 24;
sideToMove = Color(index % 2);
blackKingSquare = Square((index / 2) % 64);
whiteKingSquare = Square((index / 128) % 64);
pawnSquare = make_square(File(s % 4), Rank(s / 4 + 1));
}
bool KPKPosition::is_legal() const {
if ( whiteKingSquare == pawnSquare
|| whiteKingSquare == blackKingSquare
|| blackKingSquare == pawnSquare)
return false;
if (sideToMove == WHITE)
{
if ( bit_is_set(wk_attacks(), blackKingSquare)
|| bit_is_set(pawn_attacks(), blackKingSquare))
return false;
}
else if (bit_is_set(bk_attacks(), whiteKingSquare))
return false;
return true;
}
bool KPKPosition::is_immediate_draw() const {
if (sideToMove == BLACK)
{
Bitboard wka = wk_attacks();
Bitboard bka = bk_attacks();
// Case 1: Stalemate
if ((bka & ~(wka | pawn_attacks())) == EmptyBoardBB)
return true;
// Case 2: King can capture pawn
if (bit_is_set(bka, pawnSquare) && !bit_is_set(wka, pawnSquare))
return true;
}
else else
{ result = UNKNOWN;
// Case 1: Stalemate (possible pawn files are only from A to D)
if ( whiteKingSquare == SQ_A8
&& pawnSquare == SQ_A7
&& (blackKingSquare == SQ_C7 || blackKingSquare == SQ_C8))
return true;
}
return false;
} }
bool KPKPosition::is_immediate_win() const { template<Color Us>
Result KPKPosition::classify(const std::vector<KPKPosition>& db) {
// The position is an immediate win if it is white to move and the // White to move: If one move leads to a position classified as WIN, the result
// white pawn can be promoted without getting captured. // of the current position is WIN. If all moves lead to positions classified
return sideToMove == WHITE // as DRAW, the current position is classified as DRAW, otherwise the current
&& square_rank(pawnSquare) == RANK_7 // position is classified as UNKNOWN.
&& whiteKingSquare != pawnSquare + DELTA_N //
&& ( square_distance(blackKingSquare, pawnSquare + DELTA_N) > 1 // Black to move: If one move leads to a position classified as DRAW, the result
|| bit_is_set(wk_attacks(), pawnSquare + DELTA_N)); // of the current position is DRAW. If all moves lead to positions classified
} // as WIN, the position is classified as WIN, otherwise the current position is
// classified as UNKNOWN.
Result classify_wtm(const KPKPosition& pos, const Result bb[]) { const Color Them = (Us == WHITE ? BLACK : WHITE);
const Result Good = (Us == WHITE ? WIN : DRAW);
const Result Bad = (Us == WHITE ? DRAW : WIN);
// If one move leads to a position classified as RESULT_LOSS, the result Result r = INVALID;
// of the current position is RESULT_WIN. If all moves lead to positions Bitboard b = PseudoAttacks[KING][ksq[Us]];
// classified as RESULT_DRAW, the current position is classified RESULT_DRAW
// otherwise the current position is classified as RESULT_UNKNOWN.
bool unknownFound = false;
Bitboard b;
Square s;
Result r;
// King moves
b = pos.wk_attacks();
while (b) while (b)
r |= Us == WHITE ? db[index(Them, ksq[Them] , pop_lsb(&b), psq)]
: db[index(Them, pop_lsb(&b), ksq[Them] , psq)];
if (Us == WHITE)
{ {
s = pop_1st_bit(&b); if (rank_of(psq) < RANK_7) // Single push
r = bb[compute_index(s, pos.blackKingSquare, pos.pawnSquare, BLACK)]; r |= db[index(Them, ksq[Them], ksq[Us], psq + NORTH)];
if (r == RESULT_LOSS) if ( rank_of(psq) == RANK_2 // Double push
return RESULT_WIN; && psq + NORTH != ksq[Us]
&& psq + NORTH != ksq[Them])
if (r == RESULT_UNKNOWN) r |= db[index(Them, ksq[Them], ksq[Us], psq + NORTH + NORTH)];
unknownFound = true;
} }
// Pawn moves return result = r & Good ? Good : r & UNKNOWN ? UNKNOWN : Bad;
if (square_rank(pos.pawnSquare) < RANK_7)
{
s = pos.pawnSquare + DELTA_N;
r = bb[compute_index(pos.whiteKingSquare, pos.blackKingSquare, s, BLACK)];
if (r == RESULT_LOSS)
return RESULT_WIN;
if (r == RESULT_UNKNOWN)
unknownFound = true;
// Double pawn push
if ( square_rank(s) == RANK_3
&& s != pos.whiteKingSquare
&& s != pos.blackKingSquare)
{
s += DELTA_N;
r = bb[compute_index(pos.whiteKingSquare, pos.blackKingSquare, s, BLACK)];
if (r == RESULT_LOSS)
return RESULT_WIN;
if (r == RESULT_UNKNOWN)
unknownFound = true;
}
}
return unknownFound ? RESULT_UNKNOWN : RESULT_DRAW;
} }
} // namespace
Result classify_btm(const KPKPosition& pos, const Result bb[]) {
// If one move leads to a position classified as RESULT_DRAW, the result
// of the current position is RESULT_DRAW. If all moves lead to positions
// classified as RESULT_WIN, the current position is classified as
// RESULT_LOSS. Otherwise, the current position is classified as
// RESULT_UNKNOWN.
bool unknownFound = false;
Bitboard b;
Square s;
Result r;
// King moves
b = pos.bk_attacks();
while (b)
{
s = pop_1st_bit(&b);
r = bb[compute_index(pos.whiteKingSquare, s, pos.pawnSquare, WHITE)];
if (r == RESULT_DRAW)
return RESULT_DRAW;
if (r == RESULT_UNKNOWN)
unknownFound = true;
}
return unknownFound ? RESULT_UNKNOWN : RESULT_LOSS;
}
}
+260 -420
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,470 +18,309 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <iostream> #include <algorithm>
#include "bitboard.h" #include "bitboard.h"
#include "bitcount.h" #include "misc.h"
#if defined(IS_64BIT) uint8_t PopCnt16[1 << 16];
int SquareDistance[SQUARE_NB][SQUARE_NB];
const uint64_t BMult[64] = { Bitboard SquareBB[SQUARE_NB];
0x0440049104032280ULL, 0x1021023C82008040ULL, 0x0404040082000048ULL, Bitboard FileBB[FILE_NB];
0x48C4440084048090ULL, 0x2801104026490000ULL, 0x4100880442040800ULL, Bitboard RankBB[RANK_NB];
0x0181011002E06040ULL, 0x9101004104200E00ULL, 0x1240848848310401ULL, Bitboard AdjacentFilesBB[FILE_NB];
0x2000142828050024ULL, 0x00001004024D5000ULL, 0x0102044400800200ULL, Bitboard ForwardRanksBB[COLOR_NB][RANK_NB];
0x8108108820112000ULL, 0xA880818210C00046ULL, 0x4008008801082000ULL, Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
0x0060882404049400ULL, 0x0104402004240810ULL, 0x000A002084250200ULL, Bitboard LineBB[SQUARE_NB][SQUARE_NB];
0x00100B0880801100ULL, 0x0004080201220101ULL, 0x0044008080A00000ULL, Bitboard DistanceRingBB[SQUARE_NB][8];
0x0000202200842000ULL, 0x5006004882D00808ULL, 0x0000200045080802ULL, Bitboard ForwardFileBB[COLOR_NB][SQUARE_NB];
0x0086100020200601ULL, 0xA802080A20112C02ULL, 0x0080411218080900ULL, Bitboard PassedPawnMask[COLOR_NB][SQUARE_NB];
0x000200A0880080A0ULL, 0x9A01010000104000ULL, 0x0028008003100080ULL, Bitboard PawnAttackSpan[COLOR_NB][SQUARE_NB];
0x0211021004480417ULL, 0x0401004188220806ULL, 0x00825051400C2006ULL, Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
0x00140C0210943000ULL, 0x0000242800300080ULL, 0x00C2208120080200ULL, Bitboard PawnAttacks[COLOR_NB][SQUARE_NB];
0x2430008200002200ULL, 0x1010100112008040ULL, 0x8141050100020842ULL,
0x0000822081014405ULL, 0x800C049E40400804ULL, 0x4A0404028A000820ULL,
0x0022060201041200ULL, 0x0360904200840801ULL, 0x0881A08208800400ULL,
0x0060202C00400420ULL, 0x1204440086061400ULL, 0x0008184042804040ULL,
0x0064040315300400ULL, 0x0C01008801090A00ULL, 0x0808010401140C00ULL,
0x04004830C2020040ULL, 0x0080005002020054ULL, 0x40000C14481A0490ULL,
0x0010500101042048ULL, 0x1010100200424000ULL, 0x0000640901901040ULL,
0x00000A0201014840ULL, 0x00840082AA011002ULL, 0x010010840084240AULL,
0x0420400810420608ULL, 0x8D40230408102100ULL, 0x4A00200612222409ULL,
0x0A08520292120600ULL
};
const uint64_t RMult[64] = {
0x0A8002C000108020ULL, 0x4440200140003000ULL, 0x8080200010011880ULL,
0x0380180080141000ULL, 0x1A00060008211044ULL, 0x410001000A0C0008ULL,
0x9500060004008100ULL, 0x0100024284A20700ULL, 0x0000802140008000ULL,
0x0080C01002A00840ULL, 0x0402004282011020ULL, 0x9862000820420050ULL,
0x0001001448011100ULL, 0x6432800200800400ULL, 0x040100010002000CULL,
0x0002800D0010C080ULL, 0x90C0008000803042ULL, 0x4010004000200041ULL,
0x0003010010200040ULL, 0x0A40828028001000ULL, 0x0123010008000430ULL,
0x0024008004020080ULL, 0x0060040001104802ULL, 0x00582200028400D1ULL,
0x4000802080044000ULL, 0x0408208200420308ULL, 0x0610038080102000ULL,
0x3601000900100020ULL, 0x0000080080040180ULL, 0x00C2020080040080ULL,
0x0080084400100102ULL, 0x4022408200014401ULL, 0x0040052040800082ULL,
0x0B08200280804000ULL, 0x008A80A008801000ULL, 0x4000480080801000ULL,
0x0911808800801401ULL, 0x822A003002001894ULL, 0x401068091400108AULL,
0x000004A10A00004CULL, 0x2000800640008024ULL, 0x1486408102020020ULL,
0x000100A000D50041ULL, 0x00810050020B0020ULL, 0x0204000800808004ULL,
0x00020048100A000CULL, 0x0112000831020004ULL, 0x0009000040810002ULL,
0x0440490200208200ULL, 0x8910401000200040ULL, 0x6404200050008480ULL,
0x4B824A2010010100ULL, 0x04080801810C0080ULL, 0x00000400802A0080ULL,
0x8224080110026400ULL, 0x40002C4104088200ULL, 0x01002100104A0282ULL,
0x1208400811048021ULL, 0x3201014A40D02001ULL, 0x0005100019200501ULL,
0x0101000208001005ULL, 0x0002008450080702ULL, 0x001002080301D00CULL,
0x410201CE5C030092ULL
};
const int BShift[64] = {
58, 59, 59, 59, 59, 59, 59, 58, 59, 59, 59, 59, 59, 59, 59, 59,
59, 59, 57, 57, 57, 57, 59, 59, 59, 59, 57, 55, 55, 57, 59, 59,
59, 59, 57, 55, 55, 57, 59, 59, 59, 59, 57, 57, 57, 57, 59, 59,
59, 59, 59, 59, 59, 59, 59, 59, 58, 59, 59, 59, 59, 59, 59, 58
};
const int RShift[64] = {
52, 53, 53, 53, 53, 53, 53, 52, 53, 54, 54, 54, 54, 54, 54, 53,
53, 54, 54, 54, 54, 54, 54, 53, 53, 54, 54, 54, 54, 54, 54, 53,
53, 54, 54, 54, 54, 54, 54, 53, 53, 54, 54, 54, 54, 54, 54, 53,
53, 54, 54, 54, 54, 54, 54, 53, 52, 53, 53, 53, 53, 53, 53, 52
};
#else // if !defined(IS_64BIT)
const uint64_t BMult[64] = {
0x54142844C6A22981ULL, 0x710358A6EA25C19EULL, 0x704F746D63A4A8DCULL,
0xBFED1A0B80F838C5ULL, 0x90561D5631E62110ULL, 0x2804260376E60944ULL,
0x84A656409AA76871ULL, 0xF0267F64C28B6197ULL, 0x70764EBB762F0585ULL,
0x92AA09E0CFE161DEULL, 0x41EE1F6BB266F60EULL, 0xDDCBF04F6039C444ULL,
0x5A3FAB7BAC0D988AULL, 0xD3727877FA4EAA03ULL, 0xD988402D868DDAAEULL,
0x812B291AFA075C7CULL, 0x94FAF987B685A932ULL, 0x3ED867D8470D08DBULL,
0x92517660B8901DE8ULL, 0x2D97E43E058814B4ULL, 0x880A10C220B25582ULL,
0xC7C6520D1F1A0477ULL, 0xDBFC7FBCD7656AA6ULL, 0x78B1B9BFB1A2B84FULL,
0x2F20037F112A0BC1ULL, 0x657171EA2269A916ULL, 0xC08302B07142210EULL,
0x0880A4403064080BULL, 0x3602420842208C00ULL, 0x852800DC7E0B6602ULL,
0x595A3FBBAA0F03B2ULL, 0x9F01411558159D5EULL, 0x2B4A4A5F88B394F2ULL,
0x4AFCBFFC292DD03AULL, 0x4A4094A3B3F10522ULL, 0xB06F00B491F30048ULL,
0xD5B3820280D77004ULL, 0x8B2E01E7C8E57A75ULL, 0x2D342794E886C2E6ULL,
0xC302C410CDE21461ULL, 0x111F426F1379C274ULL, 0xE0569220ABB31588ULL,
0x5026D3064D453324ULL, 0xE2076040C343CD8AULL, 0x93EFD1E1738021EEULL,
0xB680804BED143132ULL, 0x44E361B21986944CULL, 0x44C60170EF5C598CULL,
0xF4DA475C195C9C94ULL, 0xA3AFBB5F72060B1DULL, 0xBC75F410E41C4FFCULL,
0xB51C099390520922ULL, 0x902C011F8F8EC368ULL, 0x950B56B3D6F5490AULL,
0x3909E0635BF202D0ULL, 0x5744F90206EC10CCULL, 0xDC59FD76317ABBC1ULL,
0x881C7C67FCBFC4F6ULL, 0x47CA41E7E440D423ULL, 0xEB0C88112048D004ULL,
0x51C60E04359AEF1AULL, 0x1AA1FE0E957A5554ULL, 0xDD9448DB4F5E3104ULL,
0xDC01F6DCA4BEBBDCULL,
};
const uint64_t RMult[64] = {
0xD7445CDEC88002C0ULL, 0xD0A505C1F2001722ULL, 0xE065D1C896002182ULL,
0x9A8C41E75A000892ULL, 0x8900B10C89002AA8ULL, 0x9B28D1C1D60005A2ULL,
0x015D6C88DE002D9AULL, 0xB1DBFC802E8016A9ULL, 0x149A1042D9D60029ULL,
0xB9C08050599E002FULL, 0x132208C3AF300403ULL, 0xC1000CE2E9C50070ULL,
0x9D9AA13C99020012ULL, 0xB6B078DAF71E0046ULL, 0x9D880182FB6E002EULL,
0x52889F467E850037ULL, 0xDA6DC008D19A8480ULL, 0x468286034F902420ULL,
0x7140AC09DC54C020ULL, 0xD76FFFFA39548808ULL, 0xEA901C4141500808ULL,
0xC91004093F953A02ULL, 0x02882AFA8F6BB402ULL, 0xAEBE335692442C01ULL,
0x0E904A22079FB91EULL, 0x13A514851055F606ULL, 0x76C782018C8FE632ULL,
0x1DC012A9D116DA06ULL, 0x3C9E0037264FFFA6ULL, 0x2036002853C6E4A2ULL,
0xE3FE08500AFB47D4ULL, 0xF38AF25C86B025C2ULL, 0xC0800E2182CF9A40ULL,
0x72002480D1F60673ULL, 0x2500200BAE6E9B53ULL, 0xC60018C1EEFCA252ULL,
0x0600590473E3608AULL, 0x46002C4AB3FE51B2ULL, 0xA200011486BCC8D2ULL,
0xB680078095784C63ULL, 0x2742002639BF11AEULL, 0xC7D60021A5BDB142ULL,
0xC8C04016BB83D820ULL, 0xBD520028123B4842ULL, 0x9D1600344AC2A832ULL,
0x6A808005631C8A05ULL, 0x604600A148D5389AULL, 0xE2E40103D40DEA65ULL,
0x945B5A0087C62A81ULL, 0x012DC200CD82D28EULL, 0x2431C600B5F9EF76ULL,
0xFB142A006A9B314AULL, 0x06870E00A1C97D62ULL, 0x2A9DB2004A2689A2ULL,
0xD3594600CAF5D1A2ULL, 0xEE0E4900439344A7ULL, 0x89C4D266CA25007AULL,
0x3E0013A2743F97E3ULL, 0x0180E31A0431378AULL, 0x3A9E465A4D42A512ULL,
0x98D0A11A0C0D9CC2ULL, 0x8E711C1ABA19B01EULL, 0x8DCDC836DD201142ULL,
0x5AC08A4735370479ULL,
};
const int BShift[64] = {
26, 27, 27, 27, 27, 27, 27, 26, 27, 27, 27, 27, 27, 27, 27, 27,
27, 27, 25, 25, 25, 25, 27, 27, 27, 27, 25, 23, 23, 25, 27, 27,
27, 27, 25, 23, 23, 25, 27, 27, 27, 27, 25, 25, 25, 25, 27, 27,
27, 27, 27, 27, 27, 27, 27, 27, 26, 27, 27, 27, 27, 27, 27, 26
};
const int RShift[64] = {
20, 21, 21, 21, 21, 21, 21, 20, 21, 22, 22, 22, 22, 22, 22, 21,
21, 22, 22, 22, 22, 22, 22, 21, 21, 22, 22, 22, 22, 22, 22, 21,
21, 22, 22, 22, 22, 22, 22, 21, 21, 22, 22, 22, 22, 22, 22, 21,
21, 22, 22, 22, 22, 22, 22, 21, 20, 21, 21, 21, 21, 21, 21, 20
};
#endif // defined(IS_64BIT)
// Global bitboards definitions with static storage duration are
// automatically set to zero before enter main().
Bitboard RMask[64];
int RAttackIndex[64];
Bitboard RAttacks[0x19000];
Bitboard BMask[64];
int BAttackIndex[64];
Bitboard BAttacks[0x1480];
Bitboard SetMaskBB[65];
Bitboard ClearMaskBB[65];
Bitboard SquaresByColorBB[2];
Bitboard FileBB[8];
Bitboard RankBB[8];
Bitboard NeighboringFilesBB[8];
Bitboard ThisAndNeighboringFilesBB[8];
Bitboard InFrontBB[2][8];
Bitboard StepAttacksBB[16][64];
Bitboard BetweenBB[64][64];
Bitboard SquaresInFrontMask[2][64];
Bitboard PassedPawnMask[2][64];
Bitboard AttackSpanMask[2][64];
Bitboard BishopPseudoAttacks[64];
Bitboard RookPseudoAttacks[64];
Bitboard QueenPseudoAttacks[64];
uint8_t BitCount8Bit[256];
Magic RookMagics[SQUARE_NB];
Magic BishopMagics[SQUARE_NB];
namespace { namespace {
void init_masks(); // De Bruijn sequences. See chessprogramming.wikispaces.com/BitScan
void init_step_attacks(); const uint64_t DeBruijn64 = 0x3F79D71B4CB0A89ULL;
void init_pseudo_attacks(); const uint32_t DeBruijn32 = 0x783A9B23;
void init_between_bitboards();
Bitboard index_to_bitboard(int index, Bitboard mask); int MSBTable[256]; // To implement software msb()
Bitboard sliding_attacks(int sq, Bitboard occupied, int deltas[][2], Square BSFTable[SQUARE_NB]; // To implement software bitscan
int fmin, int fmax, int rmin, int rmax); Bitboard RookTable[0x19000]; // To store rook attacks
void init_sliding_attacks(Bitboard attacks[], int attackIndex[], Bitboard mask[], Bitboard BishopTable[0x1480]; // To store bishop attacks
const int shift[], const Bitboard mult[], int deltas[][2]);
void init_magics(Bitboard table[], Magic magics[], Direction directions[]);
// bsf_index() returns the index into BSFTable[] to look up the bitscan. Uses
// Matt Taylor's folding for 32 bit case, extended to 64 bit by Kim Walisch.
unsigned bsf_index(Bitboard b) {
b ^= b - 1;
return Is64Bit ? (b * DeBruijn64) >> 58
: ((unsigned(b) ^ unsigned(b >> 32)) * DeBruijn32) >> 26;
}
// popcount16() counts the non-zero bits using SWAR-Popcount algorithm
unsigned popcount16(unsigned u) {
u -= (u >> 1) & 0x5555U;
u = ((u >> 2) & 0x3333U) + (u & 0x3333U);
u = ((u >> 4) + u) & 0x0F0FU;
return (u * 0x0101U) >> 8;
}
} }
#ifdef NO_BSF
/// print_bitboard() prints a bitboard in an easily readable format to the /// Software fall-back of lsb() and msb() for CPU lacking hardware support
/// standard output. This is sometimes useful for debugging.
void print_bitboard(Bitboard b) { Square lsb(Bitboard b) {
assert(b);
return BSFTable[bsf_index(b)];
}
for (Rank r = RANK_8; r >= RANK_1; r--) Square msb(Bitboard b) {
assert(b);
unsigned b32;
int result = 0;
if (b > 0xFFFFFFFF)
{ {
std::cout << "+---+---+---+---+---+---+---+---+" << '\n'; b >>= 32;
for (File f = FILE_A; f <= FILE_H; f++) result = 32;
std::cout << "| " << (bit_is_set(b, make_square(f, r)) ? 'X' : ' ') << ' ';
std::cout << "|\n";
} }
std::cout << "+---+---+---+---+---+---+---+---+" << std::endl;
b32 = unsigned(b);
if (b32 > 0xFFFF)
{
b32 >>= 16;
result += 16;
}
if (b32 > 0xFF)
{
b32 >>= 8;
result += 8;
}
return Square(result + MSBTable[b32]);
}
#endif // ifdef NO_BSF
/// Bitboards::pretty() returns an ASCII representation of a bitboard suitable
/// to be printed to standard output. Useful for debugging.
const std::string Bitboards::pretty(Bitboard b) {
std::string s = "+---+---+---+---+---+---+---+---+\n";
for (Rank r = RANK_8; r >= RANK_1; --r)
{
for (File f = FILE_A; f <= FILE_H; ++f)
s += b & make_square(f, r) ? "| X " : "| ";
s += "|\n+---+---+---+---+---+---+---+---+\n";
}
return s;
} }
/// first_1() finds the least significant nonzero bit in a nonzero bitboard. /// Bitboards::init() initializes various bitboard tables. It is called at
/// pop_1st_bit() finds and clears the least significant nonzero bit in a /// startup and relies on global objects to be already zero-initialized.
/// nonzero bitboard.
#if defined(IS_64BIT) && !defined(USE_BSFQ) void Bitboards::init() {
static CACHE_LINE_ALIGNMENT for (unsigned i = 0; i < (1 << 16); ++i)
const int BitTable[64] = { PopCnt16[i] = (uint8_t) popcount16(i);
0, 1, 2, 7, 3, 13, 8, 19, 4, 25, 14, 28, 9, 34, 20, 40, 5, 17, 26,
38, 15, 46, 29, 48, 10, 31, 35, 54, 21, 50, 41, 57, 63, 6, 12, 18, 24, 27,
33, 39, 16, 37, 45, 47, 30, 53, 49, 56, 62, 11, 23, 32, 36, 44, 52, 55, 61,
22, 43, 51, 60, 42, 59, 58
};
Square first_1(Bitboard b) { for (Square s = SQ_A1; s <= SQ_H8; ++s)
return Square(BitTable[((b & -b) * 0x218a392cd3d5dbfULL) >> 58]); {
SquareBB[s] = 1ULL << s;
BSFTable[bsf_index(SquareBB[s])] = s;
}
for (Bitboard b = 2; b < 256; ++b)
MSBTable[b] = MSBTable[b - 1] + !more_than_one(b);
for (File f = FILE_A; f <= FILE_H; ++f)
FileBB[f] = f > FILE_A ? FileBB[f - 1] << 1 : FileABB;
for (Rank r = RANK_1; r <= RANK_8; ++r)
RankBB[r] = r > RANK_1 ? RankBB[r - 1] << 8 : Rank1BB;
for (File f = FILE_A; f <= FILE_H; ++f)
AdjacentFilesBB[f] = (f > FILE_A ? FileBB[f - 1] : 0) | (f < FILE_H ? FileBB[f + 1] : 0);
for (Rank r = RANK_1; r < RANK_8; ++r)
ForwardRanksBB[WHITE][r] = ~(ForwardRanksBB[BLACK][r + 1] = ForwardRanksBB[BLACK][r] | RankBB[r]);
for (Color c = WHITE; c <= BLACK; ++c)
for (Square s = SQ_A1; s <= SQ_H8; ++s)
{
ForwardFileBB [c][s] = ForwardRanksBB[c][rank_of(s)] & FileBB[file_of(s)];
PawnAttackSpan[c][s] = ForwardRanksBB[c][rank_of(s)] & AdjacentFilesBB[file_of(s)];
PassedPawnMask[c][s] = ForwardFileBB [c][s] | PawnAttackSpan[c][s];
}
for (Square s1 = SQ_A1; s1 <= SQ_H8; ++s1)
for (Square s2 = SQ_A1; s2 <= SQ_H8; ++s2)
if (s1 != s2)
{
SquareDistance[s1][s2] = std::max(distance<File>(s1, s2), distance<Rank>(s1, s2));
DistanceRingBB[s1][SquareDistance[s1][s2] - 1] |= s2;
}
int steps[][5] = { {}, { 7, 9 }, { 6, 10, 15, 17 }, {}, {}, {}, { 1, 7, 8, 9 } };
for (Color c = WHITE; c <= BLACK; ++c)
for (PieceType pt : { PAWN, KNIGHT, KING })
for (Square s = SQ_A1; s <= SQ_H8; ++s)
for (int i = 0; steps[pt][i]; ++i)
{
Square to = s + Direction(c == WHITE ? steps[pt][i] : -steps[pt][i]);
if (is_ok(to) && distance(s, to) < 3)
{
if (pt == PAWN)
PawnAttacks[c][s] |= to;
else
PseudoAttacks[pt][s] |= to;
}
}
Direction RookDirections[] = { NORTH, EAST, SOUTH, WEST };
Direction BishopDirections[] = { NORTH_EAST, SOUTH_EAST, SOUTH_WEST, NORTH_WEST };
init_magics(RookTable, RookMagics, RookDirections);
init_magics(BishopTable, BishopMagics, BishopDirections);
for (Square s1 = SQ_A1; s1 <= SQ_H8; ++s1)
{
PseudoAttacks[QUEEN][s1] = PseudoAttacks[BISHOP][s1] = attacks_bb<BISHOP>(s1, 0);
PseudoAttacks[QUEEN][s1] |= PseudoAttacks[ ROOK][s1] = attacks_bb< ROOK>(s1, 0);
for (PieceType pt : { BISHOP, ROOK })
for (Square s2 = SQ_A1; s2 <= SQ_H8; ++s2)
{
if (!(PseudoAttacks[pt][s1] & s2))
continue;
LineBB[s1][s2] = (attacks_bb(pt, s1, 0) & attacks_bb(pt, s2, 0)) | s1 | s2;
BetweenBB[s1][s2] = attacks_bb(pt, s1, SquareBB[s2]) & attacks_bb(pt, s2, SquareBB[s1]);
}
}
} }
Square pop_1st_bit(Bitboard* b) {
Bitboard bb = *b;
*b &= (*b - 1);
return Square(BitTable[((bb & -bb) * 0x218a392cd3d5dbfULL) >> 58]);
}
#elif !defined(USE_BSFQ)
static CACHE_LINE_ALIGNMENT
const int BitTable[64] = {
63, 30, 3, 32, 25, 41, 22, 33, 15, 50, 42, 13, 11, 53, 19, 34, 61, 29, 2,
51, 21, 43, 45, 10, 18, 47, 1, 54, 9, 57, 0, 35, 62, 31, 40, 4, 49, 5,
52, 26, 60, 6, 23, 44, 46, 27, 56, 16, 7, 39, 48, 24, 59, 14, 12, 55, 38,
28, 58, 20, 37, 17, 36, 8
};
Square first_1(Bitboard b) {
b ^= (b - 1);
uint32_t fold = int(b) ^ int(b >> 32);
return Square(BitTable[(fold * 0x783a9b23) >> 26]);
}
// Use type-punning
union b_union {
Bitboard b;
struct {
#if defined (BIGENDIAN)
uint32_t h;
uint32_t l;
#else
uint32_t l;
uint32_t h;
#endif
} dw;
};
Square pop_1st_bit(Bitboard* bb) {
b_union u;
Square ret;
u.b = *bb;
if (u.dw.l)
{
ret = Square(BitTable[((u.dw.l ^ (u.dw.l - 1)) * 0x783a9b23) >> 26]);
u.dw.l &= (u.dw.l - 1);
*bb = u.b;
return ret;
}
ret = Square(BitTable[((~(u.dw.h ^ (u.dw.h - 1))) * 0x783a9b23) >> 26]);
u.dw.h &= (u.dw.h - 1);
*bb = u.b;
return ret;
}
#endif // !defined(USE_BSFQ)
/// init_bitboards() initializes various bitboard arrays. It is called during
/// program initialization.
void init_bitboards() {
int rookDeltas[4][2] = {{0,1},{0,-1},{1,0},{-1,0}};
int bishopDeltas[4][2] = {{1,1},{-1,1},{1,-1},{-1,-1}};
init_masks();
init_step_attacks();
init_sliding_attacks(RAttacks, RAttackIndex, RMask, RShift, RMult, rookDeltas);
init_sliding_attacks(BAttacks, BAttackIndex, BMask, BShift, BMult, bishopDeltas);
init_pseudo_attacks();
init_between_bitboards();
}
namespace { namespace {
// All functions below are used to precompute various bitboards during Bitboard sliding_attack(Direction directions[], Square sq, Bitboard occupied) {
// program initialization. Some of the functions may be difficult to
// understand, but they all seem to work correctly, and it should never
// be necessary to touch any of them.
void init_masks() { Bitboard attack = 0;
SquaresByColorBB[DARK] = 0xAA55AA55AA55AA55ULL; for (int i = 0; i < 4; ++i)
SquaresByColorBB[LIGHT] = ~SquaresByColorBB[DARK]; for (Square s = sq + directions[i];
is_ok(s) && distance(s, s - directions[i]) == 1;
FileBB[FILE_A] = FileABB; s += directions[i])
RankBB[RANK_1] = Rank1BB;
for (int f = FILE_B; f <= FILE_H; f++)
{
FileBB[f] = FileBB[f - 1] << 1;
RankBB[f] = RankBB[f - 1] << 8;
}
for (int f = FILE_A; f <= FILE_H; f++)
{
NeighboringFilesBB[f] = (f > FILE_A ? FileBB[f - 1] : 0) | (f < FILE_H ? FileBB[f + 1] : 0);
ThisAndNeighboringFilesBB[f] = FileBB[f] | NeighboringFilesBB[f];
}
for (int rw = RANK_7, rb = RANK_2; rw >= RANK_1; rw--, rb++)
{
InFrontBB[WHITE][rw] = InFrontBB[WHITE][rw + 1] | RankBB[rw + 1];
InFrontBB[BLACK][rb] = InFrontBB[BLACK][rb - 1] | RankBB[rb - 1];
}
SetMaskBB[SQ_NONE] = EmptyBoardBB;
ClearMaskBB[SQ_NONE] = ~SetMaskBB[SQ_NONE];
for (Square s = SQ_A1; s <= SQ_H8; s++)
{
SetMaskBB[s] = (1ULL << s);
ClearMaskBB[s] = ~SetMaskBB[s];
}
for (Color c = WHITE; c <= BLACK; c++)
for (Square s = SQ_A1; s <= SQ_H8; s++)
{ {
SquaresInFrontMask[c][s] = in_front_bb(c, s) & file_bb(s); attack |= s;
PassedPawnMask[c][s] = in_front_bb(c, s) & this_and_neighboring_files_bb(s);
AttackSpanMask[c][s] = in_front_bb(c, s) & neighboring_files_bb(s);
}
for (Bitboard b = 0; b < 256; b++) if (occupied & s)
BitCount8Bit[b] = (uint8_t)count_1s<CNT32>(b);
}
void init_step_attacks() {
const int step[][9] = {
{0},
{7,9,0}, {17,15,10,6,-6,-10,-15,-17,0}, {0}, {0}, {0},
{9,7,-7,-9,8,1,-1,-8,0}, {0}, {0},
{-7,-9,0}, {17,15,10,6,-6,-10,-15,-17,0}, {0}, {0}, {0},
{9,7,-7,-9,8,1,-1,-8,0}
};
for (Square s = SQ_A1; s <= SQ_H8; s++)
for (Piece pc = WP; pc <= BK; pc++)
for (int k = 0; step[pc][k] != 0; k++)
{
Square to = s + Square(step[pc][k]);
if (square_is_ok(to) && square_distance(s, to) < 3)
set_bit(&StepAttacksBB[pc][s], to);
}
}
Bitboard sliding_attacks(int sq, Bitboard occupied, int deltas[][2],
int fmin, int fmax, int rmin, int rmax) {
int dx, dy, f, r;
int rk = sq / 8;
int fl = sq % 8;
Bitboard attacks = EmptyBoardBB;
for (int i = 0; i < 4; i++)
{
dx = deltas[i][0];
dy = deltas[i][1];
f = fl + dx;
r = rk + dy;
while ( (dx == 0 || (f >= fmin && f <= fmax))
&& (dy == 0 || (r >= rmin && r <= rmax)))
{
attacks |= (1ULL << (f + r * 8));
if (occupied & (1ULL << (f + r * 8)))
break; break;
f += dx;
r += dy;
} }
}
return attacks; return attack;
} }
Bitboard index_to_bitboard(int index, Bitboard mask) {
Bitboard result = EmptyBoardBB; // init_magics() computes all rook and bishop attacks at startup. Magic
int sq, cnt = 0; // bitboards are used to look up attacks of sliding pieces. As a reference see
// chessprogramming.wikispaces.com/Magic+Bitboards. In particular, here we
// use the so called "fancy" approach.
while (mask) void init_magics(Bitboard table[], Magic magics[], Direction directions[]) {
// Optimal PRNG seeds to pick the correct magics in the shortest time
int seeds[][RANK_NB] = { { 8977, 44560, 54343, 38998, 5731, 95205, 104912, 17020 },
{ 728, 10316, 55013, 32803, 12281, 15100, 16645, 255 } };
Bitboard occupancy[4096], reference[4096], edges, b;
int epoch[4096] = {}, cnt = 0, size = 0;
for (Square s = SQ_A1; s <= SQ_H8; ++s)
{ {
sq = pop_1st_bit(&mask); // Board edges are not considered in the relevant occupancies
edges = ((Rank1BB | Rank8BB) & ~rank_bb(s)) | ((FileABB | FileHBB) & ~file_bb(s));
if (index & (1 << cnt++)) // Given a square 's', the mask is the bitboard of sliding attacks from
result |= (1ULL << sq); // 's' computed on an empty board. The index must be big enough to contain
} // all the attacks for each possible subset of the mask and so is 2 power
return result; // the number of 1s of the mask. Hence we deduce the size of the shift to
} // apply to the 64 or 32 bits word to get the index.
Magic& m = magics[s];
m.mask = sliding_attack(directions, s, 0) & ~edges;
m.shift = (Is64Bit ? 64 : 32) - popcount(m.mask);
void init_sliding_attacks(Bitboard attacks[], int attackIndex[], Bitboard mask[], // Set the offset for the attacks table of the square. We have individual
const int shift[], const Bitboard mult[], int deltas[][2]) { // table sizes for each square with "Fancy Magic Bitboards".
Bitboard b, v; m.attacks = s == SQ_A1 ? table : magics[s - 1].attacks + size;
int i, j, index;
for (i = index = 0; i < 64; i++) // Use Carry-Rippler trick to enumerate all subsets of masks[s] and
{ // store the corresponding sliding attack bitboard in reference[].
attackIndex[i] = index; b = size = 0;
mask[i] = sliding_attacks(i, 0, deltas, 1, 6, 1, 6); do {
j = 1 << ((CpuIs64Bit ? 64 : 32) - shift[i]); occupancy[size] = b;
reference[size] = sliding_attack(directions, s, b);
for (int k = 0; k < j; k++) if (HasPext)
m.attacks[pext(b, m.mask)] = reference[size];
size++;
b = (b - m.mask) & m.mask;
} while (b);
if (HasPext)
continue;
PRNG rng(seeds[Is64Bit][rank_of(s)]);
// Find a magic for square 's' picking up an (almost) random number
// until we find the one that passes the verification test.
for (int i = 0; i < size; )
{ {
b = index_to_bitboard(k, mask[i]); for (m.magic = 0; popcount((m.magic * m.mask) >> 56) < 6; )
v = CpuIs64Bit ? b * mult[i] : unsigned(b * mult[i] ^ (b >> 32) * (mult[i] >> 32)); m.magic = rng.sparse_rand<Bitboard>();
attacks[index + (v >> shift[i])] = sliding_attacks(i, b, deltas, 0, 7, 0, 7);
}
index += j;
}
}
void init_pseudo_attacks() { // A good magic must map every possible occupancy to an index that
// looks up the correct sliding attack in the attacks[s] database.
for (Square s = SQ_A1; s <= SQ_H8; s++) // Note that we build up the database for square 's' as a side
{ // effect of verifying the magic. Keep track of the attempt count
BishopPseudoAttacks[s] = bishop_attacks_bb(s, EmptyBoardBB); // and save it in epoch[], little speed-up trick to avoid resetting
RookPseudoAttacks[s] = rook_attacks_bb(s, EmptyBoardBB); // m.attacks[] after every failed attempt.
QueenPseudoAttacks[s] = queen_attacks_bb(s, EmptyBoardBB); for (++cnt, i = 0; i < size; ++i)
}
}
void init_between_bitboards() {
Square s1, s2, s3, d;
int f, r;
for (s1 = SQ_A1; s1 <= SQ_H8; s1++)
for (s2 = SQ_A1; s2 <= SQ_H8; s2++)
if (bit_is_set(QueenPseudoAttacks[s1], s2))
{ {
f = file_distance(s1, s2); unsigned idx = m.index(occupancy[i]);
r = rank_distance(s1, s2);
d = (s2 - s1) / Max(f, r); if (epoch[idx] < cnt)
{
for (s3 = s1 + d; s3 != s2; s3 += d) epoch[idx] = cnt;
set_bit(&(BetweenBB[s1][s2]), s3); m.attacks[idx] = reference[i];
}
else if (m.attacks[idx] != reference[i])
break;
} }
}
}
} }
} }
+222 -178
View File
@@ -1,14 +1,14 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or the Free Software Foundation, either version 3 of the License, or
(at your option) any later version. (at your option) any later version.
Stockfish is distributed in the hope that it will be useful, Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
@@ -18,12 +18,29 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(BITBOARD_H_INCLUDED) #ifndef BITBOARD_H_INCLUDED
#define BITBOARD_H_INCLUDED #define BITBOARD_H_INCLUDED
#include <string>
#include "types.h" #include "types.h"
const Bitboard EmptyBoardBB = 0; namespace Bitbases {
void init();
bool probe(Square wksq, Square wpsq, Square bksq, Color us);
}
namespace Bitboards {
void init();
const std::string pretty(Bitboard b);
}
const Bitboard AllSquares = ~Bitboard(0);
const Bitboard DarkSquares = 0xAA55AA55AA55AA55ULL;
const Bitboard FileABB = 0x0101010101010101ULL; const Bitboard FileABB = 0x0101010101010101ULL;
const Bitboard FileBBB = FileABB << 1; const Bitboard FileBBB = FileABB << 1;
@@ -43,79 +60,85 @@ const Bitboard Rank6BB = Rank1BB << (8 * 5);
const Bitboard Rank7BB = Rank1BB << (8 * 6); const Bitboard Rank7BB = Rank1BB << (8 * 6);
const Bitboard Rank8BB = Rank1BB << (8 * 7); const Bitboard Rank8BB = Rank1BB << (8 * 7);
extern Bitboard SquaresByColorBB[2]; extern int SquareDistance[SQUARE_NB][SQUARE_NB];
extern Bitboard FileBB[8];
extern Bitboard NeighboringFilesBB[8];
extern Bitboard ThisAndNeighboringFilesBB[8];
extern Bitboard RankBB[8];
extern Bitboard InFrontBB[2][8];
extern Bitboard SetMaskBB[65]; extern Bitboard SquareBB[SQUARE_NB];
extern Bitboard ClearMaskBB[65]; extern Bitboard FileBB[FILE_NB];
extern Bitboard RankBB[RANK_NB];
extern Bitboard StepAttacksBB[16][64]; extern Bitboard AdjacentFilesBB[FILE_NB];
extern Bitboard BetweenBB[64][64]; extern Bitboard ForwardRanksBB[COLOR_NB][RANK_NB];
extern Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
extern Bitboard SquaresInFrontMask[2][64]; extern Bitboard LineBB[SQUARE_NB][SQUARE_NB];
extern Bitboard PassedPawnMask[2][64]; extern Bitboard DistanceRingBB[SQUARE_NB][8];
extern Bitboard AttackSpanMask[2][64]; extern Bitboard ForwardFileBB[COLOR_NB][SQUARE_NB];
extern Bitboard PassedPawnMask[COLOR_NB][SQUARE_NB];
extern const uint64_t RMult[64]; extern Bitboard PawnAttackSpan[COLOR_NB][SQUARE_NB];
extern const int RShift[64]; extern Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
extern Bitboard RMask[64]; extern Bitboard PawnAttacks[COLOR_NB][SQUARE_NB];
extern int RAttackIndex[64];
extern Bitboard RAttacks[0x19000];
extern const uint64_t BMult[64];
extern const int BShift[64];
extern Bitboard BMask[64];
extern int BAttackIndex[64];
extern Bitboard BAttacks[0x1480];
extern Bitboard BishopPseudoAttacks[64];
extern Bitboard RookPseudoAttacks[64];
extern Bitboard QueenPseudoAttacks[64];
extern uint8_t BitCount8Bit[256];
/// Functions for testing whether a given bit is set in a bitboard, and for /// Magic holds all magic bitboards relevant data for a single square
/// setting and clearing bits. struct Magic {
Bitboard mask;
Bitboard magic;
Bitboard* attacks;
unsigned shift;
inline Bitboard bit_is_set(Bitboard b, Square s) { // Compute the attack's index using the 'magic bitboards' approach
return b & SetMaskBB[s]; unsigned index(Bitboard occupied) const {
if (HasPext)
return unsigned(pext(occupied, mask));
if (Is64Bit)
return unsigned(((occupied & mask) * magic) >> shift);
unsigned lo = unsigned(occupied) & unsigned(mask);
unsigned hi = unsigned(occupied >> 32) & unsigned(mask >> 32);
return (lo * unsigned(magic) ^ hi * unsigned(magic >> 32)) >> shift;
}
};
extern Magic RookMagics[SQUARE_NB];
extern Magic BishopMagics[SQUARE_NB];
/// Overloads of bitwise operators between a Bitboard and a Square for testing
/// whether a given bit is set in a bitboard, and for setting and clearing bits.
inline Bitboard operator&(Bitboard b, Square s) {
return b & SquareBB[s];
} }
inline void set_bit(Bitboard *b, Square s) { inline Bitboard operator|(Bitboard b, Square s) {
*b |= SetMaskBB[s]; return b | SquareBB[s];
} }
inline void clear_bit(Bitboard *b, Square s) { inline Bitboard operator^(Bitboard b, Square s) {
*b &= ClearMaskBB[s]; return b ^ SquareBB[s];
} }
inline Bitboard& operator|=(Bitboard& b, Square s) {
/// Functions used to update a bitboard after a move. This is faster return b |= SquareBB[s];
/// then calling a sequence of clear_bit() + set_bit()
inline Bitboard make_move_bb(Square from, Square to) {
return SetMaskBB[from] | SetMaskBB[to];
} }
inline void do_move_bb(Bitboard *b, Bitboard move_bb) { inline Bitboard& operator^=(Bitboard& b, Square s) {
*b ^= move_bb; return b ^= SquareBB[s];
} }
constexpr bool more_than_one(Bitboard b) {
return b & (b - 1);
}
/// rank_bb() and file_bb() take a file or a square as input and return /// rank_bb() and file_bb() return a bitboard representing all the squares on
/// a bitboard representing all squares on the given file or rank. /// the given file or rank.
inline Bitboard rank_bb(Rank r) { inline Bitboard rank_bb(Rank r) {
return RankBB[r]; return RankBB[r];
} }
inline Bitboard rank_bb(Square s) { inline Bitboard rank_bb(Square s) {
return RankBB[square_rank(s)]; return RankBB[rank_of(s)];
} }
inline Bitboard file_bb(File f) { inline Bitboard file_bb(File f) {
@@ -123,173 +146,194 @@ inline Bitboard file_bb(File f) {
} }
inline Bitboard file_bb(Square s) { inline Bitboard file_bb(Square s) {
return FileBB[square_file(s)]; return FileBB[file_of(s)];
} }
/// neighboring_files_bb takes a file or a square as input and returns a /// shift() moves a bitboard one step along direction D. Mainly for pawns
/// bitboard representing all squares on the neighboring files.
inline Bitboard neighboring_files_bb(File f) { template<Direction D>
return NeighboringFilesBB[f]; constexpr Bitboard shift(Bitboard b) {
} return D == NORTH ? b << 8 : D == SOUTH ? b >> 8
: D == NORTH_EAST ? (b & ~FileHBB) << 9 : D == SOUTH_EAST ? (b & ~FileHBB) >> 7
inline Bitboard neighboring_files_bb(Square s) { : D == NORTH_WEST ? (b & ~FileABB) << 7 : D == SOUTH_WEST ? (b & ~FileABB) >> 9
return NeighboringFilesBB[square_file(s)]; : 0;
} }
/// this_and_neighboring_files_bb takes a file or a square as input and returns /// adjacent_files_bb() returns a bitboard representing all the squares on the
/// a bitboard representing all squares on the given and neighboring files. /// adjacent files of the given one.
inline Bitboard this_and_neighboring_files_bb(File f) { inline Bitboard adjacent_files_bb(File f) {
return ThisAndNeighboringFilesBB[f]; return AdjacentFilesBB[f];
}
inline Bitboard this_and_neighboring_files_bb(Square s) {
return ThisAndNeighboringFilesBB[square_file(s)];
} }
/// in_front_bb() takes a color and a rank or square as input, and returns a /// between_bb() returns a bitboard representing all the squares between the two
/// bitboard representing all the squares on all ranks in front of the rank /// given ones. For instance, between_bb(SQ_C4, SQ_F7) returns a bitboard with
/// (or square), from the given color's point of view. For instance, /// the bits for square d5 and e6 set. If s1 and s2 are not on the same rank, file
/// in_front_bb(WHITE, RANK_5) will give all squares on ranks 6, 7 and 8, while /// or diagonal, 0 is returned.
/// in_front_bb(BLACK, SQ_D3) will give all squares on ranks 1 and 2.
inline Bitboard in_front_bb(Color c, Rank r) { inline Bitboard between_bb(Square s1, Square s2) {
return InFrontBB[c][r];
}
inline Bitboard in_front_bb(Color c, Square s) {
return InFrontBB[c][square_rank(s)];
}
/// Functions for computing sliding attack bitboards. rook_attacks_bb(),
/// bishop_attacks_bb() and queen_attacks_bb() all take a square and a
/// bitboard of occupied squares as input, and return a bitboard representing
/// all squares attacked by a rook, bishop or queen on the given square.
#if defined(IS_64BIT)
inline Bitboard rook_attacks_bb(Square s, Bitboard blockers) {
Bitboard b = blockers & RMask[s];
return RAttacks[RAttackIndex[s] + ((b * RMult[s]) >> RShift[s])];
}
inline Bitboard bishop_attacks_bb(Square s, Bitboard blockers) {
Bitboard b = blockers & BMask[s];
return BAttacks[BAttackIndex[s] + ((b * BMult[s]) >> BShift[s])];
}
#else // if !defined(IS_64BIT)
inline Bitboard rook_attacks_bb(Square s, Bitboard blockers) {
Bitboard b = blockers & RMask[s];
return RAttacks[RAttackIndex[s] +
(unsigned(int(b) * int(RMult[s]) ^ int(b >> 32) * int(RMult[s] >> 32)) >> RShift[s])];
}
inline Bitboard bishop_attacks_bb(Square s, Bitboard blockers) {
Bitboard b = blockers & BMask[s];
return BAttacks[BAttackIndex[s] +
(unsigned(int(b) * int(BMult[s]) ^ int(b >> 32) * int(BMult[s] >> 32)) >> BShift[s])];
}
#endif
inline Bitboard queen_attacks_bb(Square s, Bitboard blockers) {
return rook_attacks_bb(s, blockers) | bishop_attacks_bb(s, blockers);
}
/// squares_between returns a bitboard representing all squares between
/// two squares. For instance, squares_between(SQ_C4, SQ_F7) returns a
/// bitboard with the bits for square d5 and e6 set. If s1 and s2 are not
/// on the same line, file or diagonal, EmptyBoardBB is returned.
inline Bitboard squares_between(Square s1, Square s2) {
return BetweenBB[s1][s2]; return BetweenBB[s1][s2];
} }
/// squares_in_front_of takes a color and a square as input, and returns a /// forward_ranks_bb() returns a bitboard representing all the squares on all the ranks
/// bitboard representing all squares along the line in front of the square, /// in front of the given one, from the point of view of the given color. For
/// from the point of view of the given color. Definition of the table is: /// instance, forward_ranks_bb(BLACK, SQ_D3) will return the 16 squares on ranks 1 and 2.
/// SquaresInFrontOf[c][s] = in_front_bb(c, s) & file_bb(s)
inline Bitboard squares_in_front_of(Color c, Square s) { inline Bitboard forward_ranks_bb(Color c, Square s) {
return SquaresInFrontMask[c][s]; return ForwardRanksBB[c][rank_of(s)];
} }
/// passed_pawn_mask takes a color and a square as input, and returns a /// forward_file_bb() returns a bitboard representing all the squares along the line
/// bitboard mask which can be used to test if a pawn of the given color on /// in front of the given one, from the point of view of the given color:
/// the given square is a passed pawn. Definition of the table is: /// ForwardFileBB[c][s] = forward_ranks_bb(c, s) & file_bb(s)
/// PassedPawnMask[c][s] = in_front_bb(c, s) & this_and_neighboring_files_bb(s)
inline Bitboard forward_file_bb(Color c, Square s) {
return ForwardFileBB[c][s];
}
/// pawn_attack_span() returns a bitboard representing all the squares that can be
/// attacked by a pawn of the given color when it moves along its file, starting
/// from the given square:
/// PawnAttackSpan[c][s] = forward_ranks_bb(c, s) & adjacent_files_bb(file_of(s));
inline Bitboard pawn_attack_span(Color c, Square s) {
return PawnAttackSpan[c][s];
}
/// passed_pawn_mask() returns a bitboard mask which can be used to test if a
/// pawn of the given color and on the given square is a passed pawn:
/// PassedPawnMask[c][s] = pawn_attack_span(c, s) | forward_file_bb(c, s)
inline Bitboard passed_pawn_mask(Color c, Square s) { inline Bitboard passed_pawn_mask(Color c, Square s) {
return PassedPawnMask[c][s]; return PassedPawnMask[c][s];
} }
/// attack_span_mask takes a color and a square as input, and returns a bitboard /// aligned() returns true if the squares s1, s2 and s3 are aligned either on a
/// representing all squares that can be attacked by a pawn of the given color /// straight or on a diagonal line.
/// when it moves along its file starting from the given square. Definition is:
/// AttackSpanMask[c][s] = in_front_bb(c, s) & neighboring_files_bb(s);
inline Bitboard attack_span_mask(Color c, Square s) { inline bool aligned(Square s1, Square s2, Square s3) {
return AttackSpanMask[c][s]; return LineBB[s1][s2] & s3;
} }
/// squares_aligned returns true if the squares s1, s2 and s3 are aligned /// distance() functions return the distance between x and y, defined as the
/// either on a straight or on a diagonal line. /// number of steps for a king in x to reach y. Works with squares, ranks, files.
inline bool squares_aligned(Square s1, Square s2, Square s3) { template<typename T> inline int distance(T x, T y) { return x < y ? y - x : x - y; }
return (BetweenBB[s1][s2] | BetweenBB[s1][s3] | BetweenBB[s2][s3]) template<> inline int distance<Square>(Square x, Square y) { return SquareDistance[x][y]; }
& ((1ULL << s1) | (1ULL << s2) | (1ULL << s3));
template<typename T1, typename T2> inline int distance(T2 x, T2 y);
template<> inline int distance<File>(Square x, Square y) { return distance(file_of(x), file_of(y)); }
template<> inline int distance<Rank>(Square x, Square y) { return distance(rank_of(x), rank_of(y)); }
/// attacks_bb() returns a bitboard representing all the squares attacked by a
/// piece of type Pt (bishop or rook) placed on 's'.
template<PieceType Pt>
inline Bitboard attacks_bb(Square s, Bitboard occupied) {
const Magic& m = Pt == ROOK ? RookMagics[s] : BishopMagics[s];
return m.attacks[m.index(occupied)];
}
inline Bitboard attacks_bb(PieceType pt, Square s, Bitboard occupied) {
assert(pt != PAWN);
switch (pt)
{
case BISHOP: return attacks_bb<BISHOP>(s, occupied);
case ROOK : return attacks_bb< ROOK>(s, occupied);
case QUEEN : return attacks_bb<BISHOP>(s, occupied) | attacks_bb<ROOK>(s, occupied);
default : return PseudoAttacks[pt][s];
}
} }
/// first_1() finds the least significant nonzero bit in a nonzero bitboard. /// popcount() counts the number of non-zero bits in a bitboard
/// pop_1st_bit() finds and clears the least significant nonzero bit in a
/// nonzero bitboard.
#if defined(USE_BSFQ) inline int popcount(Bitboard b) {
#if defined(_MSC_VER) && !defined(__INTEL_COMPILER) #ifndef USE_POPCNT
FORCE_INLINE Square first_1(Bitboard b) { extern uint8_t PopCnt16[1 << 16];
unsigned long index; union { Bitboard bb; uint16_t u[4]; } v = { b };
_BitScanForward64(&index, b); return PopCnt16[v.u[0]] + PopCnt16[v.u[1]] + PopCnt16[v.u[2]] + PopCnt16[v.u[3]];
return (Square) index;
#elif defined(_MSC_VER) || defined(__INTEL_COMPILER)
return (int)_mm_popcnt_u64(b);
#else // Assumed gcc or compatible compiler
return __builtin_popcountll(b);
#endif
} }
/// lsb() and msb() return the least/most significant bit in a non-zero bitboard
#if defined(__GNUC__)
inline Square lsb(Bitboard b) {
assert(b);
return Square(__builtin_ctzll(b));
}
inline Square msb(Bitboard b) {
assert(b);
return Square(63 ^ __builtin_clzll(b));
}
#elif defined(_WIN64) && defined(_MSC_VER)
inline Square lsb(Bitboard b) {
assert(b);
unsigned long idx;
_BitScanForward64(&idx, b);
return (Square) idx;
}
inline Square msb(Bitboard b) {
assert(b);
unsigned long idx;
_BitScanReverse64(&idx, b);
return (Square) idx;
}
#else #else
FORCE_INLINE Square first_1(Bitboard b) { // Assembly code by Heinz van Saanen #define NO_BSF // Fallback on software implementation for other cases
Bitboard dummy;
__asm__("bsfq %1, %0": "=r"(dummy): "rm"(b) ); Square lsb(Bitboard b);
return (Square) dummy; Square msb(Bitboard b);
}
#endif #endif
FORCE_INLINE Square pop_1st_bit(Bitboard* b) {
const Square s = first_1(*b); /// pop_lsb() finds and clears the least significant bit in a non-zero bitboard
*b &= ~(1ULL<<s);
inline Square pop_lsb(Bitboard* b) {
const Square s = lsb(*b);
*b &= *b - 1;
return s; return s;
} }
#else // if !defined(USE_BSFQ)
extern Square first_1(Bitboard b); /// frontmost_sq() and backmost_sq() return the square corresponding to the
extern Square pop_1st_bit(Bitboard* b); /// most/least advanced bit relative to the given color.
#endif inline Square frontmost_sq(Color c, Bitboard b) { return c == WHITE ? msb(b) : lsb(b); }
inline Square backmost_sq(Color c, Bitboard b) { return c == WHITE ? lsb(b) : msb(b); }
#endif // #ifndef BITBOARD_H_INCLUDED
extern void print_bitboard(Bitboard b);
extern void init_bitboards();
#endif // !defined(BITBOARD_H_INCLUDED)
-98
View File
@@ -1,98 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(BITCOUNT_H_INCLUDED)
#define BITCOUNT_H_INCLUDED
#include "types.h"
enum BitCountType {
CNT64,
CNT64_MAX15,
CNT32,
CNT32_MAX15,
CNT_POPCNT
};
/// count_1s() counts the number of nonzero bits in a bitboard.
/// We have different optimized versions according if platform
/// is 32 or 64 bits, and to the maximum number of nonzero bits.
/// We also support hardware popcnt instruction. See Readme.txt
/// on how to pgo compile with popcnt support.
template<BitCountType> inline int count_1s(Bitboard);
template<>
inline int count_1s<CNT64>(Bitboard b) {
b -= ((b>>1) & 0x5555555555555555ULL);
b = ((b>>2) & 0x3333333333333333ULL) + (b & 0x3333333333333333ULL);
b = ((b>>4) + b) & 0x0F0F0F0F0F0F0F0FULL;
b *= 0x0101010101010101ULL;
return int(b >> 56);
}
template<>
inline int count_1s<CNT64_MAX15>(Bitboard b) {
b -= (b>>1) & 0x5555555555555555ULL;
b = ((b>>2) & 0x3333333333333333ULL) + (b & 0x3333333333333333ULL);
b *= 0x1111111111111111ULL;
return int(b >> 60);
}
template<>
inline int count_1s<CNT32>(Bitboard b) {
unsigned w = unsigned(b >> 32), v = unsigned(b);
v -= (v >> 1) & 0x55555555; // 0-2 in 2 bits
w -= (w >> 1) & 0x55555555;
v = ((v >> 2) & 0x33333333) + (v & 0x33333333); // 0-4 in 4 bits
w = ((w >> 2) & 0x33333333) + (w & 0x33333333);
v = ((v >> 4) + v) & 0x0F0F0F0F; // 0-8 in 8 bits
v += (((w >> 4) + w) & 0x0F0F0F0F); // 0-16 in 8 bits
v *= 0x01010101; // mul is fast on amd procs
return int(v >> 24);
}
template<>
inline int count_1s<CNT32_MAX15>(Bitboard b) {
unsigned w = unsigned(b >> 32), v = unsigned(b);
v -= (v >> 1) & 0x55555555; // 0-2 in 2 bits
w -= (w >> 1) & 0x55555555;
v = ((v >> 2) & 0x33333333) + (v & 0x33333333); // 0-4 in 4 bits
w = ((w >> 2) & 0x33333333) + (w & 0x33333333);
v += w; // 0-8 in 4 bits
v *= 0x11111111;
return int(v >> 28);
}
template<>
inline int count_1s<CNT_POPCNT>(Bitboard b) {
#if !defined(USE_POPCNT)
return int(b != 0); // Avoid 'b not used' warning
#elif defined(_MSC_VER) && defined(__INTEL_COMPILER)
return _mm_popcnt_u64(b);
#elif defined(_MSC_VER)
return (int)__popcnt64(b);
#elif defined(__GNUC__)
unsigned long ret;
__asm__("popcnt %1, %0" : "=r" (ret) : "r" (b));
return ret;
#endif
}
#endif // !defined(BITCOUNT_H_INCLUDED)
-525
View File
@@ -1,525 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
/*
The code in this file is based on the opening book code in PolyGlot
by Fabien Letouzey. PolyGlot is available under the GNU General
Public License, and can be downloaded from http://wbec-ridderkerk.nl
*/
#include <cassert>
#include <iostream>
#include "book.h"
#include "movegen.h"
using namespace std;
namespace {
// Random numbers from PolyGlot, used to compute book hash keys
const uint64_t Random64[781] = {
0x9D39247E33776D41ULL, 0x2AF7398005AAA5C7ULL, 0x44DB015024623547ULL,
0x9C15F73E62A76AE2ULL, 0x75834465489C0C89ULL, 0x3290AC3A203001BFULL,
0x0FBBAD1F61042279ULL, 0xE83A908FF2FB60CAULL, 0x0D7E765D58755C10ULL,
0x1A083822CEAFE02DULL, 0x9605D5F0E25EC3B0ULL, 0xD021FF5CD13A2ED5ULL,
0x40BDF15D4A672E32ULL, 0x011355146FD56395ULL, 0x5DB4832046F3D9E5ULL,
0x239F8B2D7FF719CCULL, 0x05D1A1AE85B49AA1ULL, 0x679F848F6E8FC971ULL,
0x7449BBFF801FED0BULL, 0x7D11CDB1C3B7ADF0ULL, 0x82C7709E781EB7CCULL,
0xF3218F1C9510786CULL, 0x331478F3AF51BBE6ULL, 0x4BB38DE5E7219443ULL,
0xAA649C6EBCFD50FCULL, 0x8DBD98A352AFD40BULL, 0x87D2074B81D79217ULL,
0x19F3C751D3E92AE1ULL, 0xB4AB30F062B19ABFULL, 0x7B0500AC42047AC4ULL,
0xC9452CA81A09D85DULL, 0x24AA6C514DA27500ULL, 0x4C9F34427501B447ULL,
0x14A68FD73C910841ULL, 0xA71B9B83461CBD93ULL, 0x03488B95B0F1850FULL,
0x637B2B34FF93C040ULL, 0x09D1BC9A3DD90A94ULL, 0x3575668334A1DD3BULL,
0x735E2B97A4C45A23ULL, 0x18727070F1BD400BULL, 0x1FCBACD259BF02E7ULL,
0xD310A7C2CE9B6555ULL, 0xBF983FE0FE5D8244ULL, 0x9F74D14F7454A824ULL,
0x51EBDC4AB9BA3035ULL, 0x5C82C505DB9AB0FAULL, 0xFCF7FE8A3430B241ULL,
0x3253A729B9BA3DDEULL, 0x8C74C368081B3075ULL, 0xB9BC6C87167C33E7ULL,
0x7EF48F2B83024E20ULL, 0x11D505D4C351BD7FULL, 0x6568FCA92C76A243ULL,
0x4DE0B0F40F32A7B8ULL, 0x96D693460CC37E5DULL, 0x42E240CB63689F2FULL,
0x6D2BDCDAE2919661ULL, 0x42880B0236E4D951ULL, 0x5F0F4A5898171BB6ULL,
0x39F890F579F92F88ULL, 0x93C5B5F47356388BULL, 0x63DC359D8D231B78ULL,
0xEC16CA8AEA98AD76ULL, 0x5355F900C2A82DC7ULL, 0x07FB9F855A997142ULL,
0x5093417AA8A7ED5EULL, 0x7BCBC38DA25A7F3CULL, 0x19FC8A768CF4B6D4ULL,
0x637A7780DECFC0D9ULL, 0x8249A47AEE0E41F7ULL, 0x79AD695501E7D1E8ULL,
0x14ACBAF4777D5776ULL, 0xF145B6BECCDEA195ULL, 0xDABF2AC8201752FCULL,
0x24C3C94DF9C8D3F6ULL, 0xBB6E2924F03912EAULL, 0x0CE26C0B95C980D9ULL,
0xA49CD132BFBF7CC4ULL, 0xE99D662AF4243939ULL, 0x27E6AD7891165C3FULL,
0x8535F040B9744FF1ULL, 0x54B3F4FA5F40D873ULL, 0x72B12C32127FED2BULL,
0xEE954D3C7B411F47ULL, 0x9A85AC909A24EAA1ULL, 0x70AC4CD9F04F21F5ULL,
0xF9B89D3E99A075C2ULL, 0x87B3E2B2B5C907B1ULL, 0xA366E5B8C54F48B8ULL,
0xAE4A9346CC3F7CF2ULL, 0x1920C04D47267BBDULL, 0x87BF02C6B49E2AE9ULL,
0x092237AC237F3859ULL, 0xFF07F64EF8ED14D0ULL, 0x8DE8DCA9F03CC54EULL,
0x9C1633264DB49C89ULL, 0xB3F22C3D0B0B38EDULL, 0x390E5FB44D01144BULL,
0x5BFEA5B4712768E9ULL, 0x1E1032911FA78984ULL, 0x9A74ACB964E78CB3ULL,
0x4F80F7A035DAFB04ULL, 0x6304D09A0B3738C4ULL, 0x2171E64683023A08ULL,
0x5B9B63EB9CEFF80CULL, 0x506AACF489889342ULL, 0x1881AFC9A3A701D6ULL,
0x6503080440750644ULL, 0xDFD395339CDBF4A7ULL, 0xEF927DBCF00C20F2ULL,
0x7B32F7D1E03680ECULL, 0xB9FD7620E7316243ULL, 0x05A7E8A57DB91B77ULL,
0xB5889C6E15630A75ULL, 0x4A750A09CE9573F7ULL, 0xCF464CEC899A2F8AULL,
0xF538639CE705B824ULL, 0x3C79A0FF5580EF7FULL, 0xEDE6C87F8477609DULL,
0x799E81F05BC93F31ULL, 0x86536B8CF3428A8CULL, 0x97D7374C60087B73ULL,
0xA246637CFF328532ULL, 0x043FCAE60CC0EBA0ULL, 0x920E449535DD359EULL,
0x70EB093B15B290CCULL, 0x73A1921916591CBDULL, 0x56436C9FE1A1AA8DULL,
0xEFAC4B70633B8F81ULL, 0xBB215798D45DF7AFULL, 0x45F20042F24F1768ULL,
0x930F80F4E8EB7462ULL, 0xFF6712FFCFD75EA1ULL, 0xAE623FD67468AA70ULL,
0xDD2C5BC84BC8D8FCULL, 0x7EED120D54CF2DD9ULL, 0x22FE545401165F1CULL,
0xC91800E98FB99929ULL, 0x808BD68E6AC10365ULL, 0xDEC468145B7605F6ULL,
0x1BEDE3A3AEF53302ULL, 0x43539603D6C55602ULL, 0xAA969B5C691CCB7AULL,
0xA87832D392EFEE56ULL, 0x65942C7B3C7E11AEULL, 0xDED2D633CAD004F6ULL,
0x21F08570F420E565ULL, 0xB415938D7DA94E3CULL, 0x91B859E59ECB6350ULL,
0x10CFF333E0ED804AULL, 0x28AED140BE0BB7DDULL, 0xC5CC1D89724FA456ULL,
0x5648F680F11A2741ULL, 0x2D255069F0B7DAB3ULL, 0x9BC5A38EF729ABD4ULL,
0xEF2F054308F6A2BCULL, 0xAF2042F5CC5C2858ULL, 0x480412BAB7F5BE2AULL,
0xAEF3AF4A563DFE43ULL, 0x19AFE59AE451497FULL, 0x52593803DFF1E840ULL,
0xF4F076E65F2CE6F0ULL, 0x11379625747D5AF3ULL, 0xBCE5D2248682C115ULL,
0x9DA4243DE836994FULL, 0x066F70B33FE09017ULL, 0x4DC4DE189B671A1CULL,
0x51039AB7712457C3ULL, 0xC07A3F80C31FB4B4ULL, 0xB46EE9C5E64A6E7CULL,
0xB3819A42ABE61C87ULL, 0x21A007933A522A20ULL, 0x2DF16F761598AA4FULL,
0x763C4A1371B368FDULL, 0xF793C46702E086A0ULL, 0xD7288E012AEB8D31ULL,
0xDE336A2A4BC1C44BULL, 0x0BF692B38D079F23ULL, 0x2C604A7A177326B3ULL,
0x4850E73E03EB6064ULL, 0xCFC447F1E53C8E1BULL, 0xB05CA3F564268D99ULL,
0x9AE182C8BC9474E8ULL, 0xA4FC4BD4FC5558CAULL, 0xE755178D58FC4E76ULL,
0x69B97DB1A4C03DFEULL, 0xF9B5B7C4ACC67C96ULL, 0xFC6A82D64B8655FBULL,
0x9C684CB6C4D24417ULL, 0x8EC97D2917456ED0ULL, 0x6703DF9D2924E97EULL,
0xC547F57E42A7444EULL, 0x78E37644E7CAD29EULL, 0xFE9A44E9362F05FAULL,
0x08BD35CC38336615ULL, 0x9315E5EB3A129ACEULL, 0x94061B871E04DF75ULL,
0xDF1D9F9D784BA010ULL, 0x3BBA57B68871B59DULL, 0xD2B7ADEEDED1F73FULL,
0xF7A255D83BC373F8ULL, 0xD7F4F2448C0CEB81ULL, 0xD95BE88CD210FFA7ULL,
0x336F52F8FF4728E7ULL, 0xA74049DAC312AC71ULL, 0xA2F61BB6E437FDB5ULL,
0x4F2A5CB07F6A35B3ULL, 0x87D380BDA5BF7859ULL, 0x16B9F7E06C453A21ULL,
0x7BA2484C8A0FD54EULL, 0xF3A678CAD9A2E38CULL, 0x39B0BF7DDE437BA2ULL,
0xFCAF55C1BF8A4424ULL, 0x18FCF680573FA594ULL, 0x4C0563B89F495AC3ULL,
0x40E087931A00930DULL, 0x8CFFA9412EB642C1ULL, 0x68CA39053261169FULL,
0x7A1EE967D27579E2ULL, 0x9D1D60E5076F5B6FULL, 0x3810E399B6F65BA2ULL,
0x32095B6D4AB5F9B1ULL, 0x35CAB62109DD038AULL, 0xA90B24499FCFAFB1ULL,
0x77A225A07CC2C6BDULL, 0x513E5E634C70E331ULL, 0x4361C0CA3F692F12ULL,
0xD941ACA44B20A45BULL, 0x528F7C8602C5807BULL, 0x52AB92BEB9613989ULL,
0x9D1DFA2EFC557F73ULL, 0x722FF175F572C348ULL, 0x1D1260A51107FE97ULL,
0x7A249A57EC0C9BA2ULL, 0x04208FE9E8F7F2D6ULL, 0x5A110C6058B920A0ULL,
0x0CD9A497658A5698ULL, 0x56FD23C8F9715A4CULL, 0x284C847B9D887AAEULL,
0x04FEABFBBDB619CBULL, 0x742E1E651C60BA83ULL, 0x9A9632E65904AD3CULL,
0x881B82A13B51B9E2ULL, 0x506E6744CD974924ULL, 0xB0183DB56FFC6A79ULL,
0x0ED9B915C66ED37EULL, 0x5E11E86D5873D484ULL, 0xF678647E3519AC6EULL,
0x1B85D488D0F20CC5ULL, 0xDAB9FE6525D89021ULL, 0x0D151D86ADB73615ULL,
0xA865A54EDCC0F019ULL, 0x93C42566AEF98FFBULL, 0x99E7AFEABE000731ULL,
0x48CBFF086DDF285AULL, 0x7F9B6AF1EBF78BAFULL, 0x58627E1A149BBA21ULL,
0x2CD16E2ABD791E33ULL, 0xD363EFF5F0977996ULL, 0x0CE2A38C344A6EEDULL,
0x1A804AADB9CFA741ULL, 0x907F30421D78C5DEULL, 0x501F65EDB3034D07ULL,
0x37624AE5A48FA6E9ULL, 0x957BAF61700CFF4EULL, 0x3A6C27934E31188AULL,
0xD49503536ABCA345ULL, 0x088E049589C432E0ULL, 0xF943AEE7FEBF21B8ULL,
0x6C3B8E3E336139D3ULL, 0x364F6FFA464EE52EULL, 0xD60F6DCEDC314222ULL,
0x56963B0DCA418FC0ULL, 0x16F50EDF91E513AFULL, 0xEF1955914B609F93ULL,
0x565601C0364E3228ULL, 0xECB53939887E8175ULL, 0xBAC7A9A18531294BULL,
0xB344C470397BBA52ULL, 0x65D34954DAF3CEBDULL, 0xB4B81B3FA97511E2ULL,
0xB422061193D6F6A7ULL, 0x071582401C38434DULL, 0x7A13F18BBEDC4FF5ULL,
0xBC4097B116C524D2ULL, 0x59B97885E2F2EA28ULL, 0x99170A5DC3115544ULL,
0x6F423357E7C6A9F9ULL, 0x325928EE6E6F8794ULL, 0xD0E4366228B03343ULL,
0x565C31F7DE89EA27ULL, 0x30F5611484119414ULL, 0xD873DB391292ED4FULL,
0x7BD94E1D8E17DEBCULL, 0xC7D9F16864A76E94ULL, 0x947AE053EE56E63CULL,
0xC8C93882F9475F5FULL, 0x3A9BF55BA91F81CAULL, 0xD9A11FBB3D9808E4ULL,
0x0FD22063EDC29FCAULL, 0xB3F256D8ACA0B0B9ULL, 0xB03031A8B4516E84ULL,
0x35DD37D5871448AFULL, 0xE9F6082B05542E4EULL, 0xEBFAFA33D7254B59ULL,
0x9255ABB50D532280ULL, 0xB9AB4CE57F2D34F3ULL, 0x693501D628297551ULL,
0xC62C58F97DD949BFULL, 0xCD454F8F19C5126AULL, 0xBBE83F4ECC2BDECBULL,
0xDC842B7E2819E230ULL, 0xBA89142E007503B8ULL, 0xA3BC941D0A5061CBULL,
0xE9F6760E32CD8021ULL, 0x09C7E552BC76492FULL, 0x852F54934DA55CC9ULL,
0x8107FCCF064FCF56ULL, 0x098954D51FFF6580ULL, 0x23B70EDB1955C4BFULL,
0xC330DE426430F69DULL, 0x4715ED43E8A45C0AULL, 0xA8D7E4DAB780A08DULL,
0x0572B974F03CE0BBULL, 0xB57D2E985E1419C7ULL, 0xE8D9ECBE2CF3D73FULL,
0x2FE4B17170E59750ULL, 0x11317BA87905E790ULL, 0x7FBF21EC8A1F45ECULL,
0x1725CABFCB045B00ULL, 0x964E915CD5E2B207ULL, 0x3E2B8BCBF016D66DULL,
0xBE7444E39328A0ACULL, 0xF85B2B4FBCDE44B7ULL, 0x49353FEA39BA63B1ULL,
0x1DD01AAFCD53486AULL, 0x1FCA8A92FD719F85ULL, 0xFC7C95D827357AFAULL,
0x18A6A990C8B35EBDULL, 0xCCCB7005C6B9C28DULL, 0x3BDBB92C43B17F26ULL,
0xAA70B5B4F89695A2ULL, 0xE94C39A54A98307FULL, 0xB7A0B174CFF6F36EULL,
0xD4DBA84729AF48ADULL, 0x2E18BC1AD9704A68ULL, 0x2DE0966DAF2F8B1CULL,
0xB9C11D5B1E43A07EULL, 0x64972D68DEE33360ULL, 0x94628D38D0C20584ULL,
0xDBC0D2B6AB90A559ULL, 0xD2733C4335C6A72FULL, 0x7E75D99D94A70F4DULL,
0x6CED1983376FA72BULL, 0x97FCAACBF030BC24ULL, 0x7B77497B32503B12ULL,
0x8547EDDFB81CCB94ULL, 0x79999CDFF70902CBULL, 0xCFFE1939438E9B24ULL,
0x829626E3892D95D7ULL, 0x92FAE24291F2B3F1ULL, 0x63E22C147B9C3403ULL,
0xC678B6D860284A1CULL, 0x5873888850659AE7ULL, 0x0981DCD296A8736DULL,
0x9F65789A6509A440ULL, 0x9FF38FED72E9052FULL, 0xE479EE5B9930578CULL,
0xE7F28ECD2D49EECDULL, 0x56C074A581EA17FEULL, 0x5544F7D774B14AEFULL,
0x7B3F0195FC6F290FULL, 0x12153635B2C0CF57ULL, 0x7F5126DBBA5E0CA7ULL,
0x7A76956C3EAFB413ULL, 0x3D5774A11D31AB39ULL, 0x8A1B083821F40CB4ULL,
0x7B4A38E32537DF62ULL, 0x950113646D1D6E03ULL, 0x4DA8979A0041E8A9ULL,
0x3BC36E078F7515D7ULL, 0x5D0A12F27AD310D1ULL, 0x7F9D1A2E1EBE1327ULL,
0xDA3A361B1C5157B1ULL, 0xDCDD7D20903D0C25ULL, 0x36833336D068F707ULL,
0xCE68341F79893389ULL, 0xAB9090168DD05F34ULL, 0x43954B3252DC25E5ULL,
0xB438C2B67F98E5E9ULL, 0x10DCD78E3851A492ULL, 0xDBC27AB5447822BFULL,
0x9B3CDB65F82CA382ULL, 0xB67B7896167B4C84ULL, 0xBFCED1B0048EAC50ULL,
0xA9119B60369FFEBDULL, 0x1FFF7AC80904BF45ULL, 0xAC12FB171817EEE7ULL,
0xAF08DA9177DDA93DULL, 0x1B0CAB936E65C744ULL, 0xB559EB1D04E5E932ULL,
0xC37B45B3F8D6F2BAULL, 0xC3A9DC228CAAC9E9ULL, 0xF3B8B6675A6507FFULL,
0x9FC477DE4ED681DAULL, 0x67378D8ECCEF96CBULL, 0x6DD856D94D259236ULL,
0xA319CE15B0B4DB31ULL, 0x073973751F12DD5EULL, 0x8A8E849EB32781A5ULL,
0xE1925C71285279F5ULL, 0x74C04BF1790C0EFEULL, 0x4DDA48153C94938AULL,
0x9D266D6A1CC0542CULL, 0x7440FB816508C4FEULL, 0x13328503DF48229FULL,
0xD6BF7BAEE43CAC40ULL, 0x4838D65F6EF6748FULL, 0x1E152328F3318DEAULL,
0x8F8419A348F296BFULL, 0x72C8834A5957B511ULL, 0xD7A023A73260B45CULL,
0x94EBC8ABCFB56DAEULL, 0x9FC10D0F989993E0ULL, 0xDE68A2355B93CAE6ULL,
0xA44CFE79AE538BBEULL, 0x9D1D84FCCE371425ULL, 0x51D2B1AB2DDFB636ULL,
0x2FD7E4B9E72CD38CULL, 0x65CA5B96B7552210ULL, 0xDD69A0D8AB3B546DULL,
0x604D51B25FBF70E2ULL, 0x73AA8A564FB7AC9EULL, 0x1A8C1E992B941148ULL,
0xAAC40A2703D9BEA0ULL, 0x764DBEAE7FA4F3A6ULL, 0x1E99B96E70A9BE8BULL,
0x2C5E9DEB57EF4743ULL, 0x3A938FEE32D29981ULL, 0x26E6DB8FFDF5ADFEULL,
0x469356C504EC9F9DULL, 0xC8763C5B08D1908CULL, 0x3F6C6AF859D80055ULL,
0x7F7CC39420A3A545ULL, 0x9BFB227EBDF4C5CEULL, 0x89039D79D6FC5C5CULL,
0x8FE88B57305E2AB6ULL, 0xA09E8C8C35AB96DEULL, 0xFA7E393983325753ULL,
0xD6B6D0ECC617C699ULL, 0xDFEA21EA9E7557E3ULL, 0xB67C1FA481680AF8ULL,
0xCA1E3785A9E724E5ULL, 0x1CFC8BED0D681639ULL, 0xD18D8549D140CAEAULL,
0x4ED0FE7E9DC91335ULL, 0xE4DBF0634473F5D2ULL, 0x1761F93A44D5AEFEULL,
0x53898E4C3910DA55ULL, 0x734DE8181F6EC39AULL, 0x2680B122BAA28D97ULL,
0x298AF231C85BAFABULL, 0x7983EED3740847D5ULL, 0x66C1A2A1A60CD889ULL,
0x9E17E49642A3E4C1ULL, 0xEDB454E7BADC0805ULL, 0x50B704CAB602C329ULL,
0x4CC317FB9CDDD023ULL, 0x66B4835D9EAFEA22ULL, 0x219B97E26FFC81BDULL,
0x261E4E4C0A333A9DULL, 0x1FE2CCA76517DB90ULL, 0xD7504DFA8816EDBBULL,
0xB9571FA04DC089C8ULL, 0x1DDC0325259B27DEULL, 0xCF3F4688801EB9AAULL,
0xF4F5D05C10CAB243ULL, 0x38B6525C21A42B0EULL, 0x36F60E2BA4FA6800ULL,
0xEB3593803173E0CEULL, 0x9C4CD6257C5A3603ULL, 0xAF0C317D32ADAA8AULL,
0x258E5A80C7204C4BULL, 0x8B889D624D44885DULL, 0xF4D14597E660F855ULL,
0xD4347F66EC8941C3ULL, 0xE699ED85B0DFB40DULL, 0x2472F6207C2D0484ULL,
0xC2A1E7B5B459AEB5ULL, 0xAB4F6451CC1D45ECULL, 0x63767572AE3D6174ULL,
0xA59E0BD101731A28ULL, 0x116D0016CB948F09ULL, 0x2CF9C8CA052F6E9FULL,
0x0B090A7560A968E3ULL, 0xABEEDDB2DDE06FF1ULL, 0x58EFC10B06A2068DULL,
0xC6E57A78FBD986E0ULL, 0x2EAB8CA63CE802D7ULL, 0x14A195640116F336ULL,
0x7C0828DD624EC390ULL, 0xD74BBE77E6116AC7ULL, 0x804456AF10F5FB53ULL,
0xEBE9EA2ADF4321C7ULL, 0x03219A39EE587A30ULL, 0x49787FEF17AF9924ULL,
0xA1E9300CD8520548ULL, 0x5B45E522E4B1B4EFULL, 0xB49C3B3995091A36ULL,
0xD4490AD526F14431ULL, 0x12A8F216AF9418C2ULL, 0x001F837CC7350524ULL,
0x1877B51E57A764D5ULL, 0xA2853B80F17F58EEULL, 0x993E1DE72D36D310ULL,
0xB3598080CE64A656ULL, 0x252F59CF0D9F04BBULL, 0xD23C8E176D113600ULL,
0x1BDA0492E7E4586EULL, 0x21E0BD5026C619BFULL, 0x3B097ADAF088F94EULL,
0x8D14DEDB30BE846EULL, 0xF95CFFA23AF5F6F4ULL, 0x3871700761B3F743ULL,
0xCA672B91E9E4FA16ULL, 0x64C8E531BFF53B55ULL, 0x241260ED4AD1E87DULL,
0x106C09B972D2E822ULL, 0x7FBA195410E5CA30ULL, 0x7884D9BC6CB569D8ULL,
0x0647DFEDCD894A29ULL, 0x63573FF03E224774ULL, 0x4FC8E9560F91B123ULL,
0x1DB956E450275779ULL, 0xB8D91274B9E9D4FBULL, 0xA2EBEE47E2FBFCE1ULL,
0xD9F1F30CCD97FB09ULL, 0xEFED53D75FD64E6BULL, 0x2E6D02C36017F67FULL,
0xA9AA4D20DB084E9BULL, 0xB64BE8D8B25396C1ULL, 0x70CB6AF7C2D5BCF0ULL,
0x98F076A4F7A2322EULL, 0xBF84470805E69B5FULL, 0x94C3251F06F90CF3ULL,
0x3E003E616A6591E9ULL, 0xB925A6CD0421AFF3ULL, 0x61BDD1307C66E300ULL,
0xBF8D5108E27E0D48ULL, 0x240AB57A8B888B20ULL, 0xFC87614BAF287E07ULL,
0xEF02CDD06FFDB432ULL, 0xA1082C0466DF6C0AULL, 0x8215E577001332C8ULL,
0xD39BB9C3A48DB6CFULL, 0x2738259634305C14ULL, 0x61CF4F94C97DF93DULL,
0x1B6BACA2AE4E125BULL, 0x758F450C88572E0BULL, 0x959F587D507A8359ULL,
0xB063E962E045F54DULL, 0x60E8ED72C0DFF5D1ULL, 0x7B64978555326F9FULL,
0xFD080D236DA814BAULL, 0x8C90FD9B083F4558ULL, 0x106F72FE81E2C590ULL,
0x7976033A39F7D952ULL, 0xA4EC0132764CA04BULL, 0x733EA705FAE4FA77ULL,
0xB4D8F77BC3E56167ULL, 0x9E21F4F903B33FD9ULL, 0x9D765E419FB69F6DULL,
0xD30C088BA61EA5EFULL, 0x5D94337FBFAF7F5BULL, 0x1A4E4822EB4D7A59ULL,
0x6FFE73E81B637FB3ULL, 0xDDF957BC36D8B9CAULL, 0x64D0E29EEA8838B3ULL,
0x08DD9BDFD96B9F63ULL, 0x087E79E5A57D1D13ULL, 0xE328E230E3E2B3FBULL,
0x1C2559E30F0946BEULL, 0x720BF5F26F4D2EAAULL, 0xB0774D261CC609DBULL,
0x443F64EC5A371195ULL, 0x4112CF68649A260EULL, 0xD813F2FAB7F5C5CAULL,
0x660D3257380841EEULL, 0x59AC2C7873F910A3ULL, 0xE846963877671A17ULL,
0x93B633ABFA3469F8ULL, 0xC0C0F5A60EF4CDCFULL, 0xCAF21ECD4377B28CULL,
0x57277707199B8175ULL, 0x506C11B9D90E8B1DULL, 0xD83CC2687A19255FULL,
0x4A29C6465A314CD1ULL, 0xED2DF21216235097ULL, 0xB5635C95FF7296E2ULL,
0x22AF003AB672E811ULL, 0x52E762596BF68235ULL, 0x9AEBA33AC6ECC6B0ULL,
0x944F6DE09134DFB6ULL, 0x6C47BEC883A7DE39ULL, 0x6AD047C430A12104ULL,
0xA5B1CFDBA0AB4067ULL, 0x7C45D833AFF07862ULL, 0x5092EF950A16DA0BULL,
0x9338E69C052B8E7BULL, 0x455A4B4CFE30E3F5ULL, 0x6B02E63195AD0CF8ULL,
0x6B17B224BAD6BF27ULL, 0xD1E0CCD25BB9C169ULL, 0xDE0C89A556B9AE70ULL,
0x50065E535A213CF6ULL, 0x9C1169FA2777B874ULL, 0x78EDEFD694AF1EEDULL,
0x6DC93D9526A50E68ULL, 0xEE97F453F06791EDULL, 0x32AB0EDB696703D3ULL,
0x3A6853C7E70757A7ULL, 0x31865CED6120F37DULL, 0x67FEF95D92607890ULL,
0x1F2B1D1F15F6DC9CULL, 0xB69E38A8965C6B65ULL, 0xAA9119FF184CCCF4ULL,
0xF43C732873F24C13ULL, 0xFB4A3D794A9A80D2ULL, 0x3550C2321FD6109CULL,
0x371F77E76BB8417EULL, 0x6BFA9AAE5EC05779ULL, 0xCD04F3FF001A4778ULL,
0xE3273522064480CAULL, 0x9F91508BFFCFC14AULL, 0x049A7F41061A9E60ULL,
0xFCB6BE43A9F2FE9BULL, 0x08DE8A1C7797DA9BULL, 0x8F9887E6078735A1ULL,
0xB5B4071DBFC73A66ULL, 0x230E343DFBA08D33ULL, 0x43ED7F5A0FAE657DULL,
0x3A88A0FBBCB05C63ULL, 0x21874B8B4D2DBC4FULL, 0x1BDEA12E35F6A8C9ULL,
0x53C065C6C8E63528ULL, 0xE34A1D250E7A8D6BULL, 0xD6B04D3B7651DD7EULL,
0x5E90277E7CB39E2DULL, 0x2C046F22062DC67DULL, 0xB10BB459132D0A26ULL,
0x3FA9DDFB67E2F199ULL, 0x0E09B88E1914F7AFULL, 0x10E8B35AF3EEAB37ULL,
0x9EEDECA8E272B933ULL, 0xD4C718BC4AE8AE5FULL, 0x81536D601170FC20ULL,
0x91B534F885818A06ULL, 0xEC8177F83F900978ULL, 0x190E714FADA5156EULL,
0xB592BF39B0364963ULL, 0x89C350C893AE7DC1ULL, 0xAC042E70F8B383F2ULL,
0xB49B52E587A1EE60ULL, 0xFB152FE3FF26DA89ULL, 0x3E666E6F69AE2C15ULL,
0x3B544EBE544C19F9ULL, 0xE805A1E290CF2456ULL, 0x24B33C9D7ED25117ULL,
0xE74733427B72F0C1ULL, 0x0A804D18B7097475ULL, 0x57E3306D881EDB4FULL,
0x4AE7D6A36EB5DBCBULL, 0x2D8D5432157064C8ULL, 0xD1E649DE1E7F268BULL,
0x8A328A1CEDFE552CULL, 0x07A3AEC79624C7DAULL, 0x84547DDC3E203C94ULL,
0x990A98FD5071D263ULL, 0x1A4FF12616EEFC89ULL, 0xF6F7FD1431714200ULL,
0x30C05B1BA332F41CULL, 0x8D2636B81555A786ULL, 0x46C9FEB55D120902ULL,
0xCCEC0A73B49C9921ULL, 0x4E9D2827355FC492ULL, 0x19EBB029435DCB0FULL,
0x4659D2B743848A2CULL, 0x963EF2C96B33BE31ULL, 0x74F85198B05A2E7DULL,
0x5A0F544DD2B1FB18ULL, 0x03727073C2E134B1ULL, 0xC7F6AA2DE59AEA61ULL,
0x352787BAA0D7C22FULL, 0x9853EAB63B5E0B35ULL, 0xABBDCDD7ED5C0860ULL,
0xCF05DAF5AC8D77B0ULL, 0x49CAD48CEBF4A71EULL, 0x7A4C10EC2158C4A6ULL,
0xD9E92AA246BF719EULL, 0x13AE978D09FE5557ULL, 0x730499AF921549FFULL,
0x4E4B705B92903BA4ULL, 0xFF577222C14F0A3AULL, 0x55B6344CF97AAFAEULL,
0xB862225B055B6960ULL, 0xCAC09AFBDDD2CDB4ULL, 0xDAF8E9829FE96B5FULL,
0xB5FDFC5D3132C498ULL, 0x310CB380DB6F7503ULL, 0xE87FBB46217A360EULL,
0x2102AE466EBB1148ULL, 0xF8549E1A3AA5E00DULL, 0x07A69AFDCC42261AULL,
0xC4C118BFE78FEAAEULL, 0xF9F4892ED96BD438ULL, 0x1AF3DBE25D8F45DAULL,
0xF5B4B0B0D2DEEEB4ULL, 0x962ACEEFA82E1C84ULL, 0x046E3ECAAF453CE9ULL,
0xF05D129681949A4CULL, 0x964781CE734B3C84ULL, 0x9C2ED44081CE5FBDULL,
0x522E23F3925E319EULL, 0x177E00F9FC32F791ULL, 0x2BC60A63A6F3B3F2ULL,
0x222BBFAE61725606ULL, 0x486289DDCC3D6780ULL, 0x7DC7785B8EFDFC80ULL,
0x8AF38731C02BA980ULL, 0x1FAB64EA29A2DDF7ULL, 0xE4D9429322CD065AULL,
0x9DA058C67844F20CULL, 0x24C0E332B70019B0ULL, 0x233003B5A6CFE6ADULL,
0xD586BD01C5C217F6ULL, 0x5E5637885F29BC2BULL, 0x7EBA726D8C94094BULL,
0x0A56A5F0BFE39272ULL, 0xD79476A84EE20D06ULL, 0x9E4C1269BAA4BF37ULL,
0x17EFEE45B0DEE640ULL, 0x1D95B0A5FCF90BC6ULL, 0x93CBE0B699C2585DULL,
0x65FA4F227A2B6D79ULL, 0xD5F9E858292504D5ULL, 0xC2B5A03F71471A6FULL,
0x59300222B4561E00ULL, 0xCE2F8642CA0712DCULL, 0x7CA9723FBB2E8988ULL,
0x2785338347F2BA08ULL, 0xC61BB3A141E50E8CULL, 0x150F361DAB9DEC26ULL,
0x9F6A419D382595F4ULL, 0x64A53DC924FE7AC9ULL, 0x142DE49FFF7A7C3DULL,
0x0C335248857FA9E7ULL, 0x0A9C32D5EAE45305ULL, 0xE6C42178C4BBB92EULL,
0x71F1CE2490D20B07ULL, 0xF1BCC3D275AFE51AULL, 0xE728E8C83C334074ULL,
0x96FBF83A12884624ULL, 0x81A1549FD6573DA5ULL, 0x5FA7867CAF35E149ULL,
0x56986E2EF3ED091BULL, 0x917F1DD5F8886C61ULL, 0xD20D8C88C8FFE65FULL,
0x31D71DCE64B2C310ULL, 0xF165B587DF898190ULL, 0xA57E6339DD2CF3A0ULL,
0x1EF6E6DBB1961EC9ULL, 0x70CC73D90BC26E24ULL, 0xE21A6B35DF0C3AD7ULL,
0x003A93D8B2806962ULL, 0x1C99DED33CB890A1ULL, 0xCF3145DE0ADD4289ULL,
0xD0E4427A5514FB72ULL, 0x77C621CC9FB3A483ULL, 0x67A34DAC4356550BULL,
0xF8D626AAAF278509ULL
};
// Indices to the Random64[] array
const int PieceIdx = 0;
const int CastleIdx = 768;
const int EnPassantIdx = 772;
const int TurnIdx = 780;
// book_key() builds up a PolyGlot hash key out of a position
uint64_t book_key(const Position& pos) {
// Piece offset is calculated as (64 * PolyPieceType + square), where
// PolyPieceType is: BP = 0, WP = 1, BN = 2, WN = 3 .... BK = 10, WK = 11
static const int PieceToPoly[] = { 0, 1, 3, 5, 7, 9, 11, 0, 0, 0, 2, 4, 6, 8, 10 };
uint64_t result = 0;
Bitboard b = pos.occupied_squares();
while (b)
{
Square s = pop_1st_bit(&b);
int p = PieceToPoly[int(pos.piece_on(s))];
result ^= Random64[PieceIdx + 64 * p + int(s)];
}
if (pos.can_castle_kingside(WHITE))
result ^= Random64[CastleIdx + 0];
if (pos.can_castle_queenside(WHITE))
result ^= Random64[CastleIdx + 1];
if (pos.can_castle_kingside(BLACK))
result ^= Random64[CastleIdx + 2];
if (pos.can_castle_queenside(BLACK))
result ^= Random64[CastleIdx + 3];
if (pos.ep_square() != SQ_NONE)
result ^= Random64[EnPassantIdx + square_file(pos.ep_square())];
if (pos.side_to_move() == WHITE)
result ^= Random64[TurnIdx];
return result;
}
}
/// Book c'tor. Make random number generation less deterministic, for book moves
Book::Book() {
for (int i = abs(get_system_time() % 10000); i > 0; i--)
RKiss.rand<unsigned>();
}
/// Book destructor. Be sure file is closed before we leave.
Book::~Book() {
close();
}
/// Book::close() closes the file only if it is open, otherwise
/// we can end up in a little mess due to how std::ifstream works.
void Book::close() {
if (bookFile.is_open())
bookFile.close();
bookName = "";
}
/// Book::open() opens a book file with a given file name
void Book::open(const string& fileName) {
// Close old file before opening the new
close();
bookFile.open(fileName.c_str(), ifstream::in | ifstream::binary);
// Silently return when asked to open a non-exsistent file
if (!bookFile.is_open())
return;
// Get the book size in number of entries
bookFile.seekg(0, ios::end);
bookSize = long(bookFile.tellg()) / sizeof(BookEntry);
if (!bookFile.good())
{
cerr << "Failed to open book file " << fileName << endl;
exit(EXIT_FAILURE);
}
// Set only if successful
bookName = fileName;
}
/// Book::get_move() gets a book move for a given position. Returns
/// MOVE_NONE if no book move is found. If findBestMove is true then
/// return always the highest rated book move.
Move Book::get_move(const Position& pos, bool findBestMove) {
if (!bookFile.is_open() || bookSize == 0)
return MOVE_NONE;
BookEntry entry;
int bookMove = MOVE_NONE;
unsigned score, scoresSum = 0, bestScore = 0;
uint64_t key = book_key(pos);
// Choose a book move among the possible moves for the given position
for (int idx = find_entry(key); idx < bookSize; idx++)
{
entry = read_entry(idx);
if (entry.key != key)
break;
score = entry.count;
if (!findBestMove)
{
// Choose book move according to its score. If a move has a very
// high score it has more probability to be choosen then a one with
// lower score. Note that first entry is always chosen.
scoresSum += score;
if (RKiss.rand<unsigned>() % scoresSum < score)
bookMove = entry.move;
}
else if (score > bestScore)
{
bestScore = score;
bookMove = entry.move;
}
}
// A PolyGlot book move is encoded as follows:
//
// bit 0- 5: destination square (from 0 to 63)
// bit 6-11: origin square (from 0 to 63)
// bit 12-13-14: promotion piece (from KNIGHT == 1 to QUEEN == 4)
//
// Castling moves follow "king captures rook" representation. So in case
// book move is a promotion we have to convert to our representation, in
// all other cases we can directly compare with a Move after having
// masked out special Move's flags that are not supported by PolyGlot.
int p = (bookMove >> 12) & 7;
if (p)
bookMove = int(make_promotion_move(move_from(Move(bookMove)),
move_to(Move(bookMove)), PieceType(p + 1)));
// Verify the book move (if any) is legal
MoveStack mlist[MAX_MOVES];
MoveStack* last = generate<MV_LEGAL>(pos, mlist);
for (MoveStack* cur = mlist; cur != last; cur++)
if ((int(cur->move) & ~(3 << 14)) == bookMove) // Mask out special flags
return cur->move;
return MOVE_NONE;
}
/// Book::find_entry() takes a book key as input, and does a binary search
/// through the book file for the given key. The index to the first book
/// entry with the same key as the input is returned. When the key is not
/// found in the book file, bookSize is returned.
int Book::find_entry(uint64_t key) {
int left, right, mid;
// Binary search (finds the leftmost entry)
left = 0;
right = bookSize - 1;
assert(left <= right);
while (left < right)
{
mid = (left + right) / 2;
assert(mid >= left && mid < right);
if (key <= read_entry(mid).key)
right = mid;
else
left = mid + 1;
}
assert(left == right);
return read_entry(left).key == key ? left : bookSize;
}
/// Book::read_entry() takes an integer index, and returns the BookEntry
/// at the given index in the book file.
BookEntry Book::read_entry(int idx) {
assert(idx >= 0 && idx < bookSize);
assert(bookFile.is_open());
BookEntry e;
bookFile.seekg(idx * sizeof(BookEntry), ios_base::beg);
*this >> e.key >> e.move >> e.count >> e.learn;
if (!bookFile.good())
{
cerr << "Failed to read book entry at index " << idx << endl;
exit(EXIT_FAILURE);
}
return e;
}
-71
View File
@@ -1,71 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(BOOK_H_INCLUDED)
#define BOOK_H_INCLUDED
#include <fstream>
#include <string>
#include "move.h"
#include "position.h"
#include "rkiss.h"
// A Polyglot book is a series of "entries" of 16 bytes. All integers are
// stored highest byte first (regardless of size). The entries are ordered
// according to key. Lowest key first.
struct BookEntry {
uint64_t key;
uint16_t move;
uint16_t count;
uint32_t learn;
};
class Book {
public:
Book();
~Book();
void open(const std::string& fileName);
void close();
Move get_move(const Position& pos, bool findBestMove);
const std::string name() const { return bookName; }
private:
// read n chars from the file stream and converts them in an
// integer number. Integers are stored with highest byte first.
template<int n> uint64_t get_int();
template<typename T>
Book& operator>>(T& n) { n = (T)get_int<sizeof(T)>(); return *this; }
BookEntry read_entry(int idx);
int find_entry(uint64_t key);
std::ifstream bookFile;
std::string bookName;
int bookSize;
RKISS RKiss;
};
// Yes, we indulge a bit here ;-)
template<int n> inline uint64_t Book::get_int() { return 256 * get_int<n-1>() + bookFile.get(); }
template<> inline uint64_t Book::get_int<1>() { return bookFile.get(); }
#endif // !defined(BOOK_H_INCLUDED)
+476 -599
View File
File diff suppressed because it is too large Load Diff
+57 -44
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,96 +18,108 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(ENDGAME_H_INCLUDED) #ifndef ENDGAME_H_INCLUDED
#define ENDGAME_H_INCLUDED #define ENDGAME_H_INCLUDED
#include <string>
#include <map> #include <map>
#include <memory>
#include <string>
#include <type_traits>
#include <utility>
#include "position.h" #include "position.h"
#include "types.h" #include "types.h"
/// EndgameType lists all supported endgames /// EndgameCode lists all supported endgame functions by corresponding codes
enum EndgameType { enum EndgameCode {
// Evaluation functions
EVALUATION_FUNCTIONS,
KNNK, // KNN vs K
KXK, // Generic "mate lone king" eval KXK, // Generic "mate lone king" eval
KBNK, // KBN vs K KBNK, // KBN vs K
KPK, // KP vs K KPK, // KP vs K
KRKP, // KR vs KP KRKP, // KR vs KP
KRKB, // KR vs KB KRKB, // KR vs KB
KRKN, // KR vs KN KRKN, // KR vs KN
KQKP, // KQ vs KP
KQKR, // KQ vs KR KQKR, // KQ vs KR
KBBKN, // KBB vs KN
KNNK, // KNN vs K
KmmKm, // K and two minors vs K and one or two minors
SCALING_FUNCTIONS,
// Scaling functions KBPsK, // KB and pawns vs K
KQKRPs, // KQ vs KR and pawns
KBPsK, // KB+pawns vs K
KQKRPs, // KQ vs KR+pawns
KRPKR, // KRP vs KR KRPKR, // KRP vs KR
KRPKB, // KRP vs KB
KRPPKRP, // KRPP vs KRP KRPPKRP, // KRPP vs KRP
KPsK, // King and pawns vs king KPsK, // K and pawns vs K
KBPKB, // KBP vs KB KBPKB, // KBP vs KB
KBPPKB, // KBPP vs KB KBPPKB, // KBPP vs KB
KBPKN, // KBP vs KN KBPKN, // KBP vs KN
KNPK, // KNP vs K KNPK, // KNP vs K
KNPKB, // KNP vs KB
KPKP // KP vs KP KPKP // KP vs KP
}; };
/// Base and derived templates for endgame evaluation and scaling functions /// Endgame functions can be of two types depending on whether they return a
/// Value or a ScaleFactor.
template<EndgameCode E> using
eg_type = typename std::conditional<(E < SCALING_FUNCTIONS), Value, ScaleFactor>::type;
/// Base and derived functors for endgame evaluation and scaling functions
template<typename T> template<typename T>
struct EndgameBase { struct EndgameBase {
typedef EndgameBase<T> Base; explicit EndgameBase(Color c) : strongSide(c), weakSide(~c) {}
virtual ~EndgameBase() = default;
virtual T operator()(const Position&) const = 0;
virtual ~EndgameBase() {} const Color strongSide, weakSide;
virtual Color color() const = 0;
virtual T apply(const Position&) const = 0;
}; };
template<typename T, EndgameType> template<EndgameCode E, typename T = eg_type<E>>
struct Endgame : public EndgameBase<T> { struct Endgame : public EndgameBase<T> {
explicit Endgame(Color c) : strongerSide(c), weakerSide(opposite_color(c)) {} explicit Endgame(Color c) : EndgameBase<T>(c) {}
Color color() const { return strongerSide; } T operator()(const Position&) const override;
T apply(const Position&) const;
private:
Color strongerSide, weakerSide;
}; };
/// Endgames class stores in two std::map the pointers to endgame evaluation /// The Endgames class stores the pointers to endgame evaluation and scaling
/// and scaling base objects. Then we use polymorphism to invoke the actual /// base objects in two std::map. We use polymorphism to invoke the actual
/// endgame function calling its apply() method that is virtual. /// endgame function by calling its virtual operator().
class Endgames { class Endgames {
typedef std::map<Key, EndgameBase<Value>* > EFMap; template<typename T> using Ptr = std::unique_ptr<EndgameBase<T>>;
typedef std::map<Key, EndgameBase<ScaleFactor>* > SFMap; template<typename T> using Map = std::map<Key, Ptr<T>>;
template<typename T>
Map<T>& map() {
return std::get<std::is_same<T, ScaleFactor>::value>(maps);
}
template<EndgameCode E, typename T = eg_type<E>, typename P = Ptr<T>>
void add(const std::string& code) {
StateInfo st;
map<T>()[Position().set(code, WHITE, &st).material_key()] = P(new Endgame<E>(WHITE));
map<T>()[Position().set(code, BLACK, &st).material_key()] = P(new Endgame<E>(BLACK));
}
std::pair<Map<Value>, Map<ScaleFactor>> maps;
public: public:
Endgames(); Endgames();
~Endgames();
template<class T> T* get(Key key) const;
private: template<typename T>
template<class T> void add(const std::string& keyCode); EndgameBase<T>* probe(Key key) {
return map<T>().count(key) ? map<T>()[key].get() : nullptr;
// Here we store two maps, for evaluate and scaling functions... }
std::pair<EFMap, SFMap> maps;
// ...and here is the accessing template function
template<typename T> const std::map<Key, T*>& get() const;
}; };
#endif // !defined(ENDGAME_H_INCLUDED) #endif // #ifndef ENDGAME_H_INCLUDED
+742 -1011
View File
File diff suppressed because it is too large Load Diff
+16 -6
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,15 +18,24 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(EVALUATE_H_INCLUDED) #ifndef EVALUATE_H_INCLUDED
#define EVALUATE_H_INCLUDED #define EVALUATE_H_INCLUDED
#include <string>
#include "types.h" #include "types.h"
class Position; class Position;
extern Value evaluate(const Position& pos, Value& margin); namespace Eval {
extern std::string trace_evaluate(const Position& pos);
extern void read_evaluation_uci_options(Color sideToMove);
#endif // !defined(EVALUATE_H_INCLUDED) const Value Tempo = Value(20); // Must be visible to search
extern Score Contempt;
std::string trace(const Position& pos);
Value evaluate(const Position& pos);
}
#endif // #ifndef EVALUATE_H_INCLUDED
-70
View File
@@ -1,70 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(HISTORY_H_INCLUDED)
#define HISTORY_H_INCLUDED
#include <cstring>
#include "types.h"
/// The History class stores statistics about how often different moves
/// have been successful or unsuccessful during the current search. These
/// statistics are used for reduction and move ordering decisions. History
/// entries are stored according only to moving piece and destination square,
/// in particular two moves with different origin but same destination and
/// same piece will be considered identical.
class History {
public:
void clear();
Value value(Piece p, Square to) const;
void update(Piece p, Square to, Value bonus);
Value gain(Piece p, Square to) const;
void update_gain(Piece p, Square to, Value g);
static const Value MaxValue = Value(2000);
private:
Value history[16][64]; // [piece][to_square]
Value maxGains[16][64]; // [piece][to_square]
};
inline void History::clear() {
memset(history, 0, 16 * 64 * sizeof(Value));
memset(maxGains, 0, 16 * 64 * sizeof(Value));
}
inline Value History::value(Piece p, Square to) const {
return history[p][to];
}
inline void History::update(Piece p, Square to, Value bonus) {
if (abs(history[p][to] + bonus) < MaxValue) history[p][to] += bonus;
}
inline Value History::gain(Piece p, Square to) const {
return maxGains[p][to];
}
inline void History::update_gain(Piece p, Square to, Value g) {
maxGains[p][to] = Max(g, maxGains[p][to] - 1);
}
#endif // !defined(HISTORY_H_INCLUDED)
-79
View File
@@ -1,79 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(LOCK_H_INCLUDED)
#define LOCK_H_INCLUDED
#if !defined(_MSC_VER)
# include <pthread.h>
typedef pthread_mutex_t Lock;
typedef pthread_cond_t WaitCondition;
# define lock_init(x) pthread_mutex_init(x, NULL)
# define lock_grab(x) pthread_mutex_lock(x)
# define lock_release(x) pthread_mutex_unlock(x)
# define lock_destroy(x) pthread_mutex_destroy(x)
# define cond_destroy(x) pthread_cond_destroy(x)
# define cond_init(x) pthread_cond_init(x, NULL)
# define cond_signal(x) pthread_cond_signal(x)
# define cond_wait(x,y) pthread_cond_wait(x,y)
#else
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#undef WIN32_LEAN_AND_MEAN
// Default fast and race free locks and condition variables
#if !defined(OLD_LOCKS)
typedef SRWLOCK Lock;
typedef CONDITION_VARIABLE WaitCondition;
# define lock_init(x) InitializeSRWLock(x)
# define lock_grab(x) AcquireSRWLockExclusive(x)
# define lock_release(x) ReleaseSRWLockExclusive(x)
# define lock_destroy(x) (x)
# define cond_destroy(x) (x)
# define cond_init(x) InitializeConditionVariable(x)
# define cond_signal(x) WakeConditionVariable(x)
# define cond_wait(x,y) SleepConditionVariableSRW(x, y, INFINITE,0)
// Fallback solution to build for Windows XP and older versions, note that
// cond_wait() is racy between lock_release() and WaitForSingleObject().
#else
typedef CRITICAL_SECTION Lock;
typedef HANDLE WaitCondition;
# define lock_init(x) InitializeCriticalSection(x)
# define lock_grab(x) EnterCriticalSection(x)
# define lock_release(x) LeaveCriticalSection(x)
# define lock_destroy(x) DeleteCriticalSection(x)
# define cond_init(x) { *x = CreateEvent(0, FALSE, FALSE, 0); }
# define cond_destroy(x) CloseHandle(*x)
# define cond_signal(x) SetEvent(*x)
# define cond_wait(x,y) { lock_release(y); WaitForSingleObject(*x, INFINITE); lock_grab(y); }
#endif
#endif
#endif // !defined(LOCK_H_INCLUDED)
+23 -55
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,71 +18,38 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
// To profile with callgrind uncomment following line
//#define USE_CALLGRIND
#include <cstdio>
#include <iostream> #include <iostream>
#include <string>
#include "bitboard.h" #include "bitboard.h"
#include "evaluate.h"
#include "position.h" #include "position.h"
#include "thread.h"
#include "search.h" #include "search.h"
#include "ucioption.h" #include "thread.h"
#include "tt.h"
#include "uci.h"
#include "syzygy/tbprobe.h"
#ifdef USE_CALLGRIND namespace PSQT {
#include <valgrind/callgrind.h> void init();
#endif }
using namespace std;
extern bool execute_uci_command(const string& cmd);
extern void benchmark(int argc, char* argv[]);
extern void init_kpk_bitbase();
int main(int argc, char* argv[]) { int main(int argc, char* argv[]) {
// Disable IO buffering for C and C++ standard libraries std::cout << engine_info() << std::endl;
setvbuf(stdin, NULL, _IONBF, 0);
setvbuf(stdout, NULL, _IONBF, 0);
cout.rdbuf()->pubsetbuf(NULL, 0);
cin.rdbuf()->pubsetbuf(NULL, 0);
// Startup initializations UCI::init(Options);
init_bitboards(); PSQT::init();
Position::init_zobrist(); Bitboards::init();
Position::init_piece_square_tables(); Position::init();
init_kpk_bitbase(); Bitbases::init();
init_search(); Search::init();
Threads.init(); Pawns::init();
Tablebases::init(Options["SyzygyPath"]);
TT.resize(Options["Hash"]);
Threads.set(Options["Threads"]);
Search::clear(); // After threads are up
#ifdef USE_CALLGRIND UCI::loop(argc, argv);
CALLGRIND_START_INSTRUMENTATION;
#endif
if (argc < 2) Threads.set(0);
{
// Print copyright notice
cout << engine_name() << " by " << engine_authors() << endl;
if (CpuHasPOPCNT)
cout << "Good! CPU has hardware POPCNT." << endl;
// Wait for a command from the user, and passes this command to
// execute_uci_command() and also intercepts EOF from stdin to
// ensure that we exit gracefully if the GUI dies unexpectedly.
string cmd;
while (getline(cin, cmd) && execute_uci_command(cmd)) {}
}
else if (string(argv[1]) == "bench" && argc < 8)
benchmark(argc, argv);
else
cout << "Usage: stockfish bench [hash size = 128] [threads = 1] "
<< "[limit = 12] [fen positions file = default] "
<< "[limited by depth, time, nodes or perft = depth]" << endl;
Threads.exit();
return 0; return 0;
} }
+159 -216
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,269 +18,211 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <algorithm> // For std::min
#include <cassert> #include <cassert>
#include <cstring> #include <cstring> // For std::memset
#include "material.h" #include "material.h"
#include "thread.h"
using namespace std; using namespace std;
namespace { namespace {
// Values modified by Joona Kiiski // Polynomial material imbalance parameters
const Value MidgameLimit = Value(15581);
const Value EndgameLimit = Value(3998);
// Scale factors used when one side has no more pawns const int QuadraticOurs[][PIECE_TYPE_NB] = {
const int NoPawnsSF[4] = { 6, 12, 32 }; // OUR PIECES
// pair pawn knight bishop rook queen
{1667 }, // Bishop pair
{ 40, 0 }, // Pawn
{ 32, 255, -3 }, // Knight OUR PIECES
{ 0, 104, 4, 0 }, // Bishop
{ -26, -2, 47, 105, -149 }, // Rook
{-189, 24, 117, 133, -134, -10 } // Queen
};
// Polynomial material balance parameters const int QuadraticTheirs[][PIECE_TYPE_NB] = {
const Value RedundantQueenPenalty = Value(320); // THEIR PIECES
const Value RedundantRookPenalty = Value(554); // pair pawn knight bishop rook queen
{ 0 }, // Bishop pair
{ 36, 0 }, // Pawn
{ 9, 63, 0 }, // Knight OUR PIECES
{ 59, 65, 42, 0 }, // Bishop
{ 46, 39, 24, -24, 0 }, // Rook
{ 97, 100, -42, 137, 268, 0 } // Queen
};
const int LinearCoefficients[6] = { 1617, -162, -1172, -190, 105, 26 }; // Endgame evaluation and scaling functions are accessed directly and not through
// the function maps because they correspond to more than one material hash key.
Endgame<KXK> EvaluateKXK[] = { Endgame<KXK>(WHITE), Endgame<KXK>(BLACK) };
const int QuadraticCoefficientsSameColor[][8] = { Endgame<KBPsK> ScaleKBPsK[] = { Endgame<KBPsK>(WHITE), Endgame<KBPsK>(BLACK) };
{ 7, 7, 7, 7, 7, 7 }, { 39, 2, 7, 7, 7, 7 }, { 35, 271, -4, 7, 7, 7 }, Endgame<KQKRPs> ScaleKQKRPs[] = { Endgame<KQKRPs>(WHITE), Endgame<KQKRPs>(BLACK) };
{ 7, 25, 4, 7, 7, 7 }, { -27, -2, 46, 100, 56, 7 }, { 58, 29, 83, 148, -3, -25 } }; Endgame<KPsK> ScaleKPsK[] = { Endgame<KPsK>(WHITE), Endgame<KPsK>(BLACK) };
Endgame<KPKP> ScaleKPKP[] = { Endgame<KPKP>(WHITE), Endgame<KPKP>(BLACK) };
const int QuadraticCoefficientsOppositeColor[][8] = { // Helper used to detect a given material distribution
{ 41, 41, 41, 41, 41, 41 }, { 37, 41, 41, 41, 41, 41 }, { 10, 62, 41, 41, 41, 41 }, bool is_KXK(const Position& pos, Color us) {
{ 57, 64, 39, 41, 41, 41 }, { 50, 40, 23, -22, 41, 41 }, { 106, 101, 3, 151, 171, 41 } }; return !more_than_one(pos.pieces(~us))
&& pos.non_pawn_material(us) >= RookValueMg;
// Endgame evaluation and scaling functions accessed direcly and not through
// the function maps because correspond to more then one material hash key.
Endgame<Value, KmmKm> EvaluateKmmKm[] = { Endgame<Value, KmmKm>(WHITE), Endgame<Value, KmmKm>(BLACK) };
Endgame<Value, KXK> EvaluateKXK[] = { Endgame<Value, KXK>(WHITE), Endgame<Value, KXK>(BLACK) };
Endgame<ScaleFactor, KBPsK> ScaleKBPsK[] = { Endgame<ScaleFactor, KBPsK>(WHITE), Endgame<ScaleFactor, KBPsK>(BLACK) };
Endgame<ScaleFactor, KQKRPs> ScaleKQKRPs[] = { Endgame<ScaleFactor, KQKRPs>(WHITE), Endgame<ScaleFactor, KQKRPs>(BLACK) };
Endgame<ScaleFactor, KPsK> ScaleKPsK[] = { Endgame<ScaleFactor, KPsK>(WHITE), Endgame<ScaleFactor, KPsK>(BLACK) };
Endgame<ScaleFactor, KPKP> ScaleKPKP[] = { Endgame<ScaleFactor, KPKP>(WHITE), Endgame<ScaleFactor, KPKP>(BLACK) };
// Helper templates used to detect a given material distribution
template<Color Us> bool is_KXK(const Position& pos) {
const Color Them = (Us == WHITE ? BLACK : WHITE);
return pos.non_pawn_material(Them) == VALUE_ZERO
&& pos.piece_count(Them, PAWN) == 0
&& pos.non_pawn_material(Us) >= RookValueMidgame;
} }
template<Color Us> bool is_KBPsKs(const Position& pos) { bool is_KBPsKs(const Position& pos, Color us) {
return pos.non_pawn_material(Us) == BishopValueMidgame return pos.non_pawn_material(us) == BishopValueMg
&& pos.piece_count(Us, BISHOP) == 1 && pos.count<BISHOP>(us) == 1
&& pos.piece_count(Us, PAWN) >= 1; && pos.count<PAWN >(us) >= 1;
} }
template<Color Us> bool is_KQKRPs(const Position& pos) { bool is_KQKRPs(const Position& pos, Color us) {
return !pos.count<PAWN>(us)
&& pos.non_pawn_material(us) == QueenValueMg
&& pos.count<QUEEN>(us) == 1
&& pos.count<ROOK>(~us) == 1
&& pos.count<PAWN>(~us) >= 1;
}
/// imbalance() calculates the imbalance by comparing the piece count of each
/// piece type for both colors.
template<Color Us>
int imbalance(const int pieceCount[][PIECE_TYPE_NB]) {
const Color Them = (Us == WHITE ? BLACK : WHITE); const Color Them = (Us == WHITE ? BLACK : WHITE);
return pos.piece_count(Us, PAWN) == 0
&& pos.non_pawn_material(Us) == QueenValueMidgame int bonus = 0;
&& pos.piece_count(Us, QUEEN) == 1
&& pos.piece_count(Them, ROOK) == 1 // Second-degree polynomial material imbalance, by Tord Romstad
&& pos.piece_count(Them, PAWN) >= 1; for (int pt1 = NO_PIECE_TYPE; pt1 <= QUEEN; ++pt1)
{
if (!pieceCount[Us][pt1])
continue;
int v = 0;
for (int pt2 = NO_PIECE_TYPE; pt2 <= pt1; ++pt2)
v += QuadraticOurs[pt1][pt2] * pieceCount[Us][pt2]
+ QuadraticTheirs[pt1][pt2] * pieceCount[Them][pt2];
bonus += pieceCount[Us][pt1] * v;
}
return bonus;
} }
} // namespace } // namespace
namespace Material {
/// MaterialInfoTable c'tor and d'tor allocate and free the space for Endgames /// Material::probe() looks up the current position's material configuration in
/// the material hash table. It returns a pointer to the Entry if the position
/// is found. Otherwise a new Entry is computed and stored there, so we don't
/// have to recompute all when the same material configuration occurs again.
void MaterialInfoTable::init() { Base::init(); if (!funcs) funcs = new Endgames(); } Entry* probe(const Position& pos) {
MaterialInfoTable::~MaterialInfoTable() { delete funcs; }
Key key = pos.material_key();
Entry* e = pos.this_thread()->materialTable[key];
/// MaterialInfoTable::get_material_info() takes a position object as input, if (e->key == key)
/// computes or looks up a MaterialInfo object, and returns a pointer to it. return e;
/// If the material configuration is not already present in the table, it
/// is stored there, so we don't have to recompute everything when the
/// same material configuration occurs again.
MaterialInfo* MaterialInfoTable::get_material_info(const Position& pos) const { std::memset(e, 0, sizeof(Entry));
e->key = key;
Key key = pos.get_material_key(); e->factor[WHITE] = e->factor[BLACK] = (uint8_t)SCALE_FACTOR_NORMAL;
MaterialInfo* mi = probe(key);
// If mi->key matches the position's material hash key, it means that we
// have analysed this material configuration before, and we can simply
// return the information we found the last time instead of recomputing it.
if (mi->key == key)
return mi;
// Initialize MaterialInfo entry
memset(mi, 0, sizeof(MaterialInfo));
mi->key = key;
mi->factor[WHITE] = mi->factor[BLACK] = (uint8_t)SCALE_FACTOR_NORMAL;
// Store game phase
mi->gamePhase = MaterialInfoTable::game_phase(pos);
// Let's look if we have a specialized evaluation function for this
// particular material configuration. First we look for a fixed
// configuration one, then a generic one if previous search failed.
if ((mi->evaluationFunction = funcs->get<EndgameBase<Value> >(key)) != NULL)
return mi;
if (is_KXK<WHITE>(pos))
{
mi->evaluationFunction = &EvaluateKXK[WHITE];
return mi;
}
if (is_KXK<BLACK>(pos))
{
mi->evaluationFunction = &EvaluateKXK[BLACK];
return mi;
}
if (!pos.pieces(PAWN) && !pos.pieces(ROOK) && !pos.pieces(QUEEN))
{
// Minor piece endgame with at least one minor piece per side and
// no pawns. Note that the case KmmK is already handled by KXK.
assert((pos.pieces(KNIGHT, WHITE) | pos.pieces(BISHOP, WHITE)));
assert((pos.pieces(KNIGHT, BLACK) | pos.pieces(BISHOP, BLACK)));
if ( pos.piece_count(WHITE, BISHOP) + pos.piece_count(WHITE, KNIGHT) <= 2
&& pos.piece_count(BLACK, BISHOP) + pos.piece_count(BLACK, KNIGHT) <= 2)
{
mi->evaluationFunction = &EvaluateKmmKm[WHITE];
return mi;
}
}
// OK, we didn't find any special evaluation function for the current
// material configuration. Is there a suitable scaling function?
//
// We face problems when there are several conflicting applicable
// scaling functions and we need to decide which one to use.
EndgameBase<ScaleFactor>* sf;
if ((sf = funcs->get<EndgameBase<ScaleFactor> >(key)) != NULL)
{
mi->scalingFunction[sf->color()] = sf;
return mi;
}
// Generic scaling functions that refer to more then one material
// distribution. Should be probed after the specialized ones.
// Note that these ones don't return after setting the function.
if (is_KBPsKs<WHITE>(pos))
mi->scalingFunction[WHITE] = &ScaleKBPsK[WHITE];
if (is_KBPsKs<BLACK>(pos))
mi->scalingFunction[BLACK] = &ScaleKBPsK[BLACK];
if (is_KQKRPs<WHITE>(pos))
mi->scalingFunction[WHITE] = &ScaleKQKRPs[WHITE];
else if (is_KQKRPs<BLACK>(pos))
mi->scalingFunction[BLACK] = &ScaleKQKRPs[BLACK];
Value npm_w = pos.non_pawn_material(WHITE); Value npm_w = pos.non_pawn_material(WHITE);
Value npm_b = pos.non_pawn_material(BLACK); Value npm_b = pos.non_pawn_material(BLACK);
Value npm = std::max(EndgameLimit, std::min(npm_w + npm_b, MidgameLimit));
if (npm_w + npm_b == VALUE_ZERO) // Map total non-pawn material into [PHASE_ENDGAME, PHASE_MIDGAME]
e->gamePhase = Phase(((npm - EndgameLimit) * PHASE_MIDGAME) / (MidgameLimit - EndgameLimit));
// Let's look if we have a specialized evaluation function for this particular
// material configuration. Firstly we look for a fixed configuration one, then
// for a generic one if the previous search failed.
if ((e->evaluationFunction = pos.this_thread()->endgames.probe<Value>(key)) != nullptr)
return e;
for (Color c = WHITE; c <= BLACK; ++c)
if (is_KXK(pos, c))
{
e->evaluationFunction = &EvaluateKXK[c];
return e;
}
// OK, we didn't find any special evaluation function for the current material
// configuration. Is there a suitable specialized scaling function?
EndgameBase<ScaleFactor>* sf;
if ((sf = pos.this_thread()->endgames.probe<ScaleFactor>(key)) != nullptr)
{ {
if (pos.piece_count(BLACK, PAWN) == 0) e->scalingFunction[sf->strongSide] = sf; // Only strong color assigned
return e;
}
// We didn't find any specialized scaling function, so fall back on generic
// ones that refer to more than one material distribution. Note that in this
// case we don't return after setting the function.
for (Color c = WHITE; c <= BLACK; ++c)
{
if (is_KBPsKs(pos, c))
e->scalingFunction[c] = &ScaleKBPsK[c];
else if (is_KQKRPs(pos, c))
e->scalingFunction[c] = &ScaleKQKRPs[c];
}
if (npm_w + npm_b == VALUE_ZERO && pos.pieces(PAWN)) // Only pawns on the board
{
if (!pos.count<PAWN>(BLACK))
{ {
assert(pos.piece_count(WHITE, PAWN) >= 2); assert(pos.count<PAWN>(WHITE) >= 2);
mi->scalingFunction[WHITE] = &ScaleKPsK[WHITE];
e->scalingFunction[WHITE] = &ScaleKPsK[WHITE];
} }
else if (pos.piece_count(WHITE, PAWN) == 0) else if (!pos.count<PAWN>(WHITE))
{ {
assert(pos.piece_count(BLACK, PAWN) >= 2); assert(pos.count<PAWN>(BLACK) >= 2);
mi->scalingFunction[BLACK] = &ScaleKPsK[BLACK];
e->scalingFunction[BLACK] = &ScaleKPsK[BLACK];
} }
else if (pos.piece_count(WHITE, PAWN) == 1 && pos.piece_count(BLACK, PAWN) == 1) else if (pos.count<PAWN>(WHITE) == 1 && pos.count<PAWN>(BLACK) == 1)
{ {
// This is a special case because we set scaling functions // This is a special case because we set scaling functions
// for both colors instead of only one. // for both colors instead of only one.
mi->scalingFunction[WHITE] = &ScaleKPKP[WHITE]; e->scalingFunction[WHITE] = &ScaleKPKP[WHITE];
mi->scalingFunction[BLACK] = &ScaleKPKP[BLACK]; e->scalingFunction[BLACK] = &ScaleKPKP[BLACK];
} }
} }
// No pawns makes it difficult to win, even with a material advantage // Zero or just one pawn makes it difficult to win, even with a small material
if (pos.piece_count(WHITE, PAWN) == 0 && npm_w - npm_b <= BishopValueMidgame) // advantage. This catches some trivial draws like KK, KBK and KNK and gives a
{ // drawish scale factor for cases such as KRKBP and KmmKm (except for KBBKN).
mi->factor[WHITE] = uint8_t if (!pos.count<PAWN>(WHITE) && npm_w - npm_b <= BishopValueMg)
(npm_w == npm_b || npm_w < RookValueMidgame ? 0 : NoPawnsSF[Min(pos.piece_count(WHITE, BISHOP), 2)]); e->factor[WHITE] = uint8_t(npm_w < RookValueMg ? SCALE_FACTOR_DRAW :
} npm_b <= BishopValueMg ? 4 : 14);
if (pos.piece_count(BLACK, PAWN) == 0 && npm_b - npm_w <= BishopValueMidgame) if (!pos.count<PAWN>(BLACK) && npm_b - npm_w <= BishopValueMg)
{ e->factor[BLACK] = uint8_t(npm_b < RookValueMg ? SCALE_FACTOR_DRAW :
mi->factor[BLACK] = uint8_t npm_w <= BishopValueMg ? 4 : 14);
(npm_w == npm_b || npm_b < RookValueMidgame ? 0 : NoPawnsSF[Min(pos.piece_count(BLACK, BISHOP), 2)]);
}
// Compute the space weight if (pos.count<PAWN>(WHITE) == 1 && npm_w - npm_b <= BishopValueMg)
if (npm_w + npm_b >= 2 * QueenValueMidgame + 4 * RookValueMidgame + 2 * KnightValueMidgame) e->factor[WHITE] = (uint8_t) SCALE_FACTOR_ONEPAWN;
{
int minorPieceCount = pos.piece_count(WHITE, KNIGHT) + pos.piece_count(WHITE, BISHOP)
+ pos.piece_count(BLACK, KNIGHT) + pos.piece_count(BLACK, BISHOP);
mi->spaceWeight = minorPieceCount * minorPieceCount; if (pos.count<PAWN>(BLACK) == 1 && npm_b - npm_w <= BishopValueMg)
} e->factor[BLACK] = (uint8_t) SCALE_FACTOR_ONEPAWN;
// Evaluate the material imbalance. We use PIECE_TYPE_NONE as a place holder // Evaluate the material imbalance. We use PIECE_TYPE_NONE as a place holder
// for the bishop pair "extended piece", this allow us to be more flexible // for the bishop pair "extended piece", which allows us to be more flexible
// in defining bishop pair bonuses. // in defining bishop pair bonuses.
const int pieceCount[2][8] = { const int PieceCount[COLOR_NB][PIECE_TYPE_NB] = {
{ pos.piece_count(WHITE, BISHOP) > 1, pos.piece_count(WHITE, PAWN), pos.piece_count(WHITE, KNIGHT), { pos.count<BISHOP>(WHITE) > 1, pos.count<PAWN>(WHITE), pos.count<KNIGHT>(WHITE),
pos.piece_count(WHITE, BISHOP) , pos.piece_count(WHITE, ROOK), pos.piece_count(WHITE, QUEEN) }, pos.count<BISHOP>(WHITE) , pos.count<ROOK>(WHITE), pos.count<QUEEN >(WHITE) },
{ pos.piece_count(BLACK, BISHOP) > 1, pos.piece_count(BLACK, PAWN), pos.piece_count(BLACK, KNIGHT), { pos.count<BISHOP>(BLACK) > 1, pos.count<PAWN>(BLACK), pos.count<KNIGHT>(BLACK),
pos.piece_count(BLACK, BISHOP) , pos.piece_count(BLACK, ROOK), pos.piece_count(BLACK, QUEEN) } }; pos.count<BISHOP>(BLACK) , pos.count<ROOK>(BLACK), pos.count<QUEEN >(BLACK) } };
mi->value = int16_t((imbalance<WHITE>(pieceCount) - imbalance<BLACK>(pieceCount)) / 16); e->value = int16_t((imbalance<WHITE>(PieceCount) - imbalance<BLACK>(PieceCount)) / 16);
return mi; return e;
} }
} // namespace Material
/// MaterialInfoTable::imbalance() calculates imbalance comparing piece count of each
/// piece type for both colors.
template<Color Us>
int MaterialInfoTable::imbalance(const int pieceCount[][8]) {
const Color Them = (Us == WHITE ? BLACK : WHITE);
int pt1, pt2, pc, v;
int value = 0;
// Redundancy of major pieces, formula based on Kaufman's paper
// "The Evaluation of Material Imbalances in Chess"
if (pieceCount[Us][ROOK] > 0)
value -= RedundantRookPenalty * (pieceCount[Us][ROOK] - 1)
+ RedundantQueenPenalty * pieceCount[Us][QUEEN];
// Second-degree polynomial material imbalance by Tord Romstad
for (pt1 = PIECE_TYPE_NONE; pt1 <= QUEEN; pt1++)
{
pc = pieceCount[Us][pt1];
if (!pc)
continue;
v = LinearCoefficients[pt1];
for (pt2 = PIECE_TYPE_NONE; pt2 <= pt1; pt2++)
v += QuadraticCoefficientsSameColor[pt1][pt2] * pieceCount[Us][pt2]
+ QuadraticCoefficientsOppositeColor[pt1][pt2] * pieceCount[Them][pt2];
value += pc * v;
}
return value;
}
/// MaterialInfoTable::game_phase() calculates the phase given the current
/// position. Because the phase is strictly a function of the material, it
/// is stored in MaterialInfo.
Phase MaterialInfoTable::game_phase(const Position& pos) {
Value npm = pos.non_pawn_material(WHITE) + pos.non_pawn_material(BLACK);
return npm >= MidgameLimit ? PHASE_MIDGAME
: npm <= EndgameLimit ? PHASE_ENDGAME
: Phase(((npm - EndgameLimit) * 128) / (MidgameLimit - EndgameLimit));
}
+35 -79
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,101 +18,56 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(MATERIAL_H_INCLUDED) #ifndef MATERIAL_H_INCLUDED
#define MATERIAL_H_INCLUDED #define MATERIAL_H_INCLUDED
#include "endgame.h" #include "endgame.h"
#include "misc.h"
#include "position.h" #include "position.h"
#include "tt.h"
#include "types.h" #include "types.h"
const int MaterialTableSize = 8192; namespace Material {
/// MaterialInfo is a class which contains various information about a /// Material::Entry contains various information about a material configuration.
/// material configuration. It contains a material balance evaluation, /// It contains a material imbalance evaluation, a function pointer to a special
/// a function pointer to a special endgame evaluation function (which in /// endgame evaluation function (which in most cases is NULL, meaning that the
/// most cases is NULL, meaning that the standard evaluation function will /// standard evaluation function will be used), and scale factors.
/// be used), and "scale factors" for black and white.
/// ///
/// The scale factors are used to scale the evaluation score up or down. /// The scale factors are used to scale the evaluation score up or down. For
/// For instance, in KRB vs KR endgames, the score is scaled down by a factor /// instance, in KRB vs KR endgames, the score is scaled down by a factor of 4,
/// of 4, which will result in scores of absolute value less than one pawn. /// which will result in scores of absolute value less than one pawn.
class MaterialInfo { struct Entry {
friend class MaterialInfoTable; Score imbalance() const { return make_score(value, value); }
Phase game_phase() const { return gamePhase; }
bool specialized_eval_exists() const { return evaluationFunction != nullptr; }
Value evaluate(const Position& pos) const { return (*evaluationFunction)(pos); }
public: // scale_factor takes a position and a color as input and returns a scale factor
Score material_value() const; // for the given color. We have to provide the position in addition to the color
ScaleFactor scale_factor(const Position& pos, Color c) const; // because the scale factor may also be a function which should be applied to
int space_weight() const; // the position. For instance, in KBP vs K endgames, the scaling function looks
Phase game_phase() const; // for rook pawns and wrong-colored bishops.
bool specialized_eval_exists() const; ScaleFactor scale_factor(const Position& pos, Color c) const {
Value evaluate(const Position& pos) const; ScaleFactor sf = scalingFunction[c] ? (*scalingFunction[c])(pos)
: SCALE_FACTOR_NONE;
return sf != SCALE_FACTOR_NONE ? sf : ScaleFactor(factor[c]);
}
private:
Key key; Key key;
int16_t value;
uint8_t factor[2];
EndgameBase<Value>* evaluationFunction; EndgameBase<Value>* evaluationFunction;
EndgameBase<ScaleFactor>* scalingFunction[2]; EndgameBase<ScaleFactor>* scalingFunction[COLOR_NB]; // Could be one for each
int spaceWeight; // side (e.g. KPKP, KBPsKs)
int16_t value;
uint8_t factor[COLOR_NB];
Phase gamePhase; Phase gamePhase;
}; };
typedef HashTable<Entry, 8192> Table;
/// The MaterialInfoTable class represents a pawn hash table. The most important Entry* probe(const Position& pos);
/// method is get_material_info, which returns a pointer to a MaterialInfo object.
class MaterialInfoTable : public SimpleHash<MaterialInfo, MaterialTableSize> { } // namespace Material
public:
~MaterialInfoTable();
void init();
MaterialInfo* get_material_info(const Position& pos) const;
static Phase game_phase(const Position& pos);
private: #endif // #ifndef MATERIAL_H_INCLUDED
template<Color Us>
static int imbalance(const int pieceCount[][8]);
Endgames* funcs;
};
/// MaterialInfo::scale_factor takes a position and a color as input, and
/// returns a scale factor for the given color. We have to provide the
/// position in addition to the color, because the scale factor need not
/// to be a constant: It can also be a function which should be applied to
/// the position. For instance, in KBP vs K endgames, a scaling function
/// which checks for draws with rook pawns and wrong-colored bishops.
inline ScaleFactor MaterialInfo::scale_factor(const Position& pos, Color c) const {
if (!scalingFunction[c])
return ScaleFactor(factor[c]);
ScaleFactor sf = scalingFunction[c]->apply(pos);
return sf == SCALE_FACTOR_NONE ? ScaleFactor(factor[c]) : sf;
}
inline Value MaterialInfo::evaluate(const Position& pos) const {
return evaluationFunction->apply(pos);
}
inline Score MaterialInfo::material_value() const {
return make_score(value, value);
}
inline int MaterialInfo::space_weight() const {
return spaceWeight;
}
inline Phase MaterialInfo::game_phase() const {
return gamePhase;
}
inline bool MaterialInfo::specialized_eval_exists() const {
return evaluationFunction != NULL;
}
#endif // !defined(MATERIAL_H_INCLUDED)
+232 -185
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,254 +18,300 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(_MSC_VER) #ifdef _WIN32
#if _WIN32_WINNT < 0x0601
# include <sys/time.h> #undef _WIN32_WINNT
# include <sys/types.h> #define _WIN32_WINNT 0x0601 // Force to include needed API prototypes
# include <unistd.h> #endif
# if defined(__hpux)
# include <sys/pstat.h>
# endif
#else
#define _CRT_SECURE_NO_DEPRECATE
#include <windows.h> #include <windows.h>
#include <sys/timeb.h> // The needed Windows API for processor groups could be missed from old Windows
// versions, so instead of calling them directly (forcing the linker to resolve
// the calls at compile time), try to load them at runtime. To do this we need
// first to define the corresponding function pointers.
extern "C" {
typedef bool(*fun1_t)(LOGICAL_PROCESSOR_RELATIONSHIP,
PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX, PDWORD);
typedef bool(*fun2_t)(USHORT, PGROUP_AFFINITY);
typedef bool(*fun3_t)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY);
}
#endif #endif
#if !defined(NO_PREFETCH) #include <fstream>
# include <xmmintrin.h>
#endif
#include <cassert>
#include <cstdio>
#include <iomanip> #include <iomanip>
#include <iostream> #include <iostream>
#include <sstream> #include <sstream>
#include <vector>
#include "bitcount.h"
#include "misc.h" #include "misc.h"
#include "thread.h" #include "thread.h"
using namespace std; using namespace std;
/// Version number. If EngineVersion is left empty, then AppTag plus namespace {
/// current date (in the format YYMMDD) is used as a version number.
static const string AppName = "Stockfish"; /// Version number. If Version is left empty, then compile date in the format
static const string EngineVersion = "2.1"; /// DD-MM-YY and show in engine_info.
static const string AppTag = ""; const string Version = "9";
/// Our fancy logging facility. The trick here is to replace cin.rdbuf() and
/// cout.rdbuf() with two Tie objects that tie cin and cout to a file stream. We
/// can toggle the logging of std::cout and std:cin at runtime whilst preserving
/// usual I/O functionality, all without changing a single line of code!
/// Idea from http://groups.google.com/group/comp.lang.c++/msg/1d941c0f26ea0d81
/// engine_name() returns the full name of the current Stockfish version. struct Tie: public streambuf { // MSVC requires split streambuf for cin and cout
/// This will be either "Stockfish YYMMDD" (where YYMMDD is the date when
/// the program was compiled) or "Stockfish <version number>", depending
/// on whether the constant EngineVersion is empty.
const string engine_name() { Tie(streambuf* b, streambuf* l) : buf(b), logBuf(l) {}
int sync() override { return logBuf->pubsync(), buf->pubsync(); }
int overflow(int c) override { return log(buf->sputc((char)c), "<< "); }
int underflow() override { return buf->sgetc(); }
int uflow() override { return log(buf->sbumpc(), ">> "); }
streambuf *buf, *logBuf;
int log(int c, const char* prefix) {
static int last = '\n'; // Single log file
if (last == '\n')
logBuf->sputn(prefix, 3);
return last = logBuf->sputc((char)c);
}
};
class Logger {
Logger() : in(cin.rdbuf(), file.rdbuf()), out(cout.rdbuf(), file.rdbuf()) {}
~Logger() { start(""); }
ofstream file;
Tie in, out;
public:
static void start(const std::string& fname) {
static Logger l;
if (!fname.empty() && !l.file.is_open())
{
l.file.open(fname, ifstream::out);
cin.rdbuf(&l.in);
cout.rdbuf(&l.out);
}
else if (fname.empty() && l.file.is_open())
{
cout.rdbuf(l.out.buf);
cin.rdbuf(l.in.buf);
l.file.close();
}
}
};
} // namespace
/// engine_info() returns the full name of the current Stockfish version. This
/// will be either "Stockfish <Tag> DD-MM-YY" (where DD-MM-YY is the date when
/// the program was compiled) or "Stockfish <Version>", depending on whether
/// Version is empty.
const string engine_info(bool to_uci) {
const string months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec"); const string months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec");
const string cpu64(CpuIs64Bit ? " 64bit" : "");
if (!EngineVersion.empty())
return AppName + " " + EngineVersion + cpu64;
stringstream s, date(__DATE__); // From compiler, format is "Sep 21 2008"
string month, day, year; string month, day, year;
stringstream ss, date(__DATE__); // From compiler, format is "Sep 21 2008"
date >> month >> day >> year; ss << "Stockfish " << Version << setfill('0');
s << setfill('0') << AppName + " " + AppTag + " " if (Version.empty())
<< year.substr(2, 2) << setw(2) {
<< (1 + months.find(month) / 4) << setw(2) date >> month >> day >> year;
<< day << cpu64; ss << setw(2) << day << setw(2) << (1 + months.find(month) / 4) << year.substr(2);
}
return s.str(); ss << (Is64Bit ? " 64" : "")
<< (HasPext ? " BMI2" : (HasPopCnt ? " POPCNT" : ""))
<< (to_uci ? "\nid author ": " by ")
<< "T. Romstad, M. Costalba, J. Kiiski, G. Linscott";
return ss.str();
} }
/// Our brave developers! Required by UCI /// Debug functions used mainly to collect run-time statistics
static int64_t hits[2], means[2];
const string engine_authors() { void dbg_hit_on(bool b) { ++hits[0]; if (b) ++hits[1]; }
void dbg_hit_on(bool c, bool b) { if (c) dbg_hit_on(b); }
void dbg_mean_of(int v) { ++means[0]; means[1] += v; }
return "Tord Romstad, Marco Costalba and Joona Kiiski"; void dbg_print() {
if (hits[0])
cerr << "Total " << hits[0] << " Hits " << hits[1]
<< " hit rate (%) " << 100 * hits[1] / hits[0] << endl;
if (means[0])
cerr << "Total " << means[0] << " Mean "
<< (double)means[1] / means[0] << endl;
} }
/// Debug stuff. Helper functions used mainly for debugging purposes /// Used to serialize access to std::cout to avoid multiple threads writing at
/// the same time.
static uint64_t dbg_hit_cnt0; std::ostream& operator<<(std::ostream& os, SyncCout sc) {
static uint64_t dbg_hit_cnt1;
static uint64_t dbg_mean_cnt0;
static uint64_t dbg_mean_cnt1;
void dbg_print_hit_rate() { static Mutex m;
if (dbg_hit_cnt0) if (sc == IO_LOCK)
cout << "Total " << dbg_hit_cnt0 << " Hit " << dbg_hit_cnt1 m.lock();
<< " hit rate (%) " << 100 * dbg_hit_cnt1 / dbg_hit_cnt0 << endl;
}
void dbg_print_mean() { if (sc == IO_UNLOCK)
m.unlock();
if (dbg_mean_cnt0) return os;
cout << "Total " << dbg_mean_cnt0 << " Mean "
<< (float)dbg_mean_cnt1 / dbg_mean_cnt0 << endl;
}
void dbg_mean_of(int v) {
dbg_mean_cnt0++;
dbg_mean_cnt1 += v;
}
void dbg_hit_on(bool b) {
dbg_hit_cnt0++;
if (b)
dbg_hit_cnt1++;
}
void dbg_hit_on_c(bool c, bool b) { if (c) dbg_hit_on(b); }
void dbg_before() { dbg_hit_on(false); }
void dbg_after() { dbg_hit_on(true); dbg_hit_cnt0--; }
/// get_system_time() returns the current system time, measured in milliseconds
int get_system_time() {
#if defined(_MSC_VER)
struct _timeb t;
_ftime(&t);
return int(t.time * 1000 + t.millitm);
#else
struct timeval t;
gettimeofday(&t, NULL);
return t.tv_sec * 1000 + t.tv_usec / 1000;
#endif
} }
/// cpu_count() tries to detect the number of CPU cores /// Trampoline helper to avoid moving Logger to misc.h
void start_logger(const std::string& fname) { Logger::start(fname); }
int cpu_count() {
#if defined(_MSC_VER) /// prefetch() preloads the given address in L1/L2 cache. This is a non-blocking
SYSTEM_INFO s; /// function that doesn't stall the CPU waiting for data to be loaded from memory,
GetSystemInfo(&s); /// which can be quite slow.
return Min(s.dwNumberOfProcessors, MAX_THREADS); #ifdef NO_PREFETCH
void prefetch(void*) {}
#else #else
# if defined(_SC_NPROCESSORS_ONLN) void prefetch(void* addr) {
return Min(sysconf(_SC_NPROCESSORS_ONLN), MAX_THREADS);
# elif defined(__hpux) # if defined(__INTEL_COMPILER)
struct pst_dynamic psd; // This hack prevents prefetches from being optimized away by
if (pstat_getdynamic(&psd, sizeof(psd), (size_t)1, 0) == -1) // Intel compiler. Both MSVC and gcc seem not be affected by this.
return 1; __asm__ ("");
return Min(psd.psd_proc_cnt, MAX_THREADS);
# else
return 1;
# endif # endif
#endif # if defined(__INTEL_COMPILER) || defined(_MSC_VER)
_mm_prefetch((char*)addr, _MM_HINT_T0);
# else
__builtin_prefetch(addr);
# endif
} }
#endif
/// Check for console input. Original code from Beowulf, Olithink and Greko void prefetch2(void* addr) {
prefetch(addr);
prefetch((uint8_t*)addr + 64);
}
namespace WinProcGroup {
#ifndef _WIN32 #ifndef _WIN32
int input_available() { void bindThisThread(size_t) {}
fd_set readfds;
struct timeval timeout;
FD_ZERO(&readfds);
FD_SET(fileno(stdin), &readfds);
timeout.tv_sec = 0; // Set to timeout immediately
timeout.tv_usec = 0;
select(16, &readfds, 0, 0, &timeout);
return (FD_ISSET(fileno(stdin), &readfds));
}
#else #else
int input_available() { /// get_group() retrieves logical processor information using Windows specific
/// API and returns the best group id for the thread with index idx. Original
/// code from Texel by Peter Österlund.
static HANDLE inh = NULL; int get_group(size_t idx) {
static bool usePipe = false;
INPUT_RECORD rec[256];
DWORD nchars, recCnt;
if (!inh) int threads = 0;
int nodes = 0;
int cores = 0;
DWORD returnLength = 0;
DWORD byteOffset = 0;
// Early exit if the needed API is not available at runtime
HMODULE k32 = GetModuleHandle("Kernel32.dll");
auto fun1 = (fun1_t)GetProcAddress(k32, "GetLogicalProcessorInformationEx");
if (!fun1)
return -1;
// First call to get returnLength. We expect it to fail due to null buffer
if (fun1(RelationAll, nullptr, &returnLength))
return -1;
// Once we know returnLength, allocate the buffer
SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *buffer, *ptr;
ptr = buffer = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)malloc(returnLength);
// Second call, now we expect to succeed
if (!fun1(RelationAll, buffer, &returnLength))
{ {
inh = GetStdHandle(STD_INPUT_HANDLE); free(buffer);
if (GetConsoleMode(inh, &nchars)) return -1;
{
SetConsoleMode(inh, nchars & ~(ENABLE_MOUSE_INPUT | ENABLE_WINDOW_INPUT));
FlushConsoleInputBuffer(inh);
} else
usePipe = true;
} }
// When using Standard C input functions, also check if there while (ptr->Size > 0 && byteOffset + ptr->Size <= returnLength)
// is anything in the buffer. After a call to such functions, {
// the input waiting in the pipe will be copied to the buffer, if (ptr->Relationship == RelationNumaNode)
// and the call to PeekNamedPipe can indicate no input available. nodes++;
// Setting stdin to unbuffered was not enough. [from Greko]
if (stdin->_cnt > 0)
return 1;
// When running under a GUI the input commands are sent to us else if (ptr->Relationship == RelationProcessorCore)
// directly over the internal pipe. If PeekNamedPipe() returns 0 {
// then something went wrong. Probably the parent program exited. cores++;
// Returning 1 will make the next call to the input function threads += (ptr->Processor.Flags == LTP_PC_SMT) ? 2 : 1;
// return EOF, where this should be catched then. }
if (usePipe)
return PeekNamedPipe(inh, NULL, 0, NULL, &nchars, NULL) ? nchars : 1;
// Count the number of unread input records, including keyboard, byteOffset += ptr->Size;
// mouse, and window-resizing input records. ptr = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)(((char*)ptr) + ptr->Size);
GetNumberOfConsoleInputEvents(inh, &nchars); }
// Read data from console without removing it from the buffer free(buffer);
if (nchars <= 0 || !PeekConsoleInput(inh, rec, Min(nchars, 256), &recCnt))
return 0;
// Search for at least one keyboard event std::vector<int> groups;
for (DWORD i = 0; i < recCnt; i++)
if (rec[i].EventType == KEY_EVENT)
return 1;
return 0; // Run as many threads as possible on the same node until core limit is
// reached, then move on filling the next node.
for (int n = 0; n < nodes; n++)
for (int i = 0; i < cores / nodes; i++)
groups.push_back(n);
// In case a core has more than one logical processor (we assume 2) and we
// have still threads to allocate, then spread them evenly across available
// nodes.
for (int t = 0; t < threads - cores; t++)
groups.push_back(t % nodes);
// If we still have more threads than the total number of logical processors
// then return -1 and let the OS to decide what to do.
return idx < groups.size() ? groups[idx] : -1;
}
/// bindThisThread() set the group affinity of the current thread
void bindThisThread(size_t idx) {
// Use only local variables to be thread-safe
int group = get_group(idx);
if (group == -1)
return;
// Early exit if the needed API are not available at runtime
HMODULE k32 = GetModuleHandle("Kernel32.dll");
auto fun2 = (fun2_t)GetProcAddress(k32, "GetNumaNodeProcessorMaskEx");
auto fun3 = (fun3_t)GetProcAddress(k32, "SetThreadGroupAffinity");
if (!fun2 || !fun3)
return;
GROUP_AFFINITY affinity;
if (fun2(group, &affinity))
fun3(GetCurrentThread(), &affinity, nullptr);
} }
#endif #endif
} // namespace WinProcGroup
/// prefetch() preloads the given address in L1/L2 cache. This is a non
/// blocking function and do not stalls the CPU waiting for data to be
/// loaded from memory, that can be quite slow.
#if defined(NO_PREFETCH)
void prefetch(char*) {}
#else
void prefetch(char* addr) {
#if defined(__INTEL_COMPILER) || defined(__ICL)
// This hack prevents prefetches to be optimized away by
// Intel compiler. Both MSVC and gcc seems not affected.
__asm__ ("");
#endif
_mm_prefetch(addr, _MM_HINT_T2);
_mm_prefetch(addr+64, _MM_HINT_T2); // 64 bytes ahead
}
#endif
+87 -16
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,25 +18,95 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(MISC_H_INCLUDED) #ifndef MISC_H_INCLUDED
#define MISC_H_INCLUDED #define MISC_H_INCLUDED
#include <cassert>
#include <chrono>
#include <ostream>
#include <string> #include <string>
#include <vector>
#include "types.h" #include "types.h"
extern const std::string engine_name(); const std::string engine_info(bool to_uci = false);
extern const std::string engine_authors(); void prefetch(void* addr);
extern int get_system_time(); void prefetch2(void* addr);
extern int cpu_count(); void start_logger(const std::string& fname);
extern int input_available();
extern void prefetch(char* addr);
extern void dbg_hit_on(bool b); void dbg_hit_on(bool b);
extern void dbg_hit_on_c(bool c, bool b); void dbg_hit_on(bool c, bool b);
extern void dbg_before(); void dbg_mean_of(int v);
extern void dbg_after(); void dbg_print();
extern void dbg_mean_of(int v);
extern void dbg_print_hit_rate();
extern void dbg_print_mean();
#endif // !defined(MISC_H_INCLUDED) typedef std::chrono::milliseconds::rep TimePoint; // A value in milliseconds
inline TimePoint now() {
return std::chrono::duration_cast<std::chrono::milliseconds>
(std::chrono::steady_clock::now().time_since_epoch()).count();
}
template<class Entry, int Size>
struct HashTable {
Entry* operator[](Key key) { return &table[(uint32_t)key & (Size - 1)]; }
private:
std::vector<Entry> table = std::vector<Entry>(Size);
};
enum SyncCout { IO_LOCK, IO_UNLOCK };
std::ostream& operator<<(std::ostream&, SyncCout);
#define sync_cout std::cout << IO_LOCK
#define sync_endl std::endl << IO_UNLOCK
/// xorshift64star Pseudo-Random Number Generator
/// This class is based on original code written and dedicated
/// to the public domain by Sebastiano Vigna (2014).
/// It has the following characteristics:
///
/// - Outputs 64-bit numbers
/// - Passes Dieharder and SmallCrush test batteries
/// - Does not require warm-up, no zeroland to escape
/// - Internal state is a single 64-bit integer
/// - Period is 2^64 - 1
/// - Speed: 1.60 ns/call (Core i7 @3.40GHz)
///
/// For further analysis see
/// <http://vigna.di.unimi.it/ftp/papers/xorshift.pdf>
class PRNG {
uint64_t s;
uint64_t rand64() {
s ^= s >> 12, s ^= s << 25, s ^= s >> 27;
return s * 2685821657736338717LL;
}
public:
PRNG(uint64_t seed) : s(seed) { assert(seed); }
template<typename T> T rand() { return T(rand64()); }
/// Special generator used to fast init magic numbers.
/// Output values only have 1/8th of their bits set on average.
template<typename T> T sparse_rand()
{ return T(rand64() & rand64() & rand64()); }
};
/// Under Windows it is not possible for a process to run on more than one
/// logical processor group. This usually means to be limited to use max 64
/// cores. To overcome this, some special platform specific API should be
/// called to set group affinity for each thread. Original code from Texel by
/// Peter Österlund.
namespace WinProcGroup {
void bindThisThread(size_t idx);
}
#endif // #ifndef MISC_H_INCLUDED
-258
View File
@@ -1,258 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <cassert>
#include <cstring>
#include <iomanip>
#include <string>
#include <sstream>
#include "move.h"
#include "movegen.h"
#include "search.h"
using std::string;
namespace {
const string time_string(int milliseconds);
const string score_string(Value v);
}
/// move_to_uci() converts a move to a string in coordinate notation
/// (g1f3, a7a8q, etc.). The only special case is castling moves, where we
/// print in the e1g1 notation in normal chess mode, and in e1h1 notation in
/// Chess960 mode.
const string move_to_uci(Move m, bool chess960) {
Square from = move_from(m);
Square to = move_to(m);
string promotion;
if (m == MOVE_NONE)
return "(none)";
if (m == MOVE_NULL)
return "0000";
if (move_is_short_castle(m) && !chess960)
return from == SQ_E1 ? "e1g1" : "e8g8";
if (move_is_long_castle(m) && !chess960)
return from == SQ_E1 ? "e1c1" : "e8c8";
if (move_is_promotion(m))
promotion = char(tolower(piece_type_to_char(move_promotion_piece(m))));
return square_to_string(from) + square_to_string(to) + promotion;
}
/// move_from_uci() takes a position and a string representing a move in
/// simple coordinate notation and returns an equivalent Move if any.
/// Moves are guaranteed to be legal.
Move move_from_uci(const Position& pos, const string& str) {
MoveStack mlist[MAX_MOVES];
MoveStack* last = generate<MV_LEGAL>(pos, mlist);
for (MoveStack* cur = mlist; cur != last; cur++)
if (str == move_to_uci(cur->move, pos.is_chess960()))
return cur->move;
return MOVE_NONE;
}
/// move_to_san() takes a position and a move as input, where it is assumed
/// that the move is a legal move from the position. The return value is
/// a string containing the move in short algebraic notation.
const string move_to_san(Position& pos, Move m) {
assert(pos.is_ok());
assert(move_is_ok(m));
MoveStack mlist[MAX_MOVES];
Square from = move_from(m);
Square to = move_to(m);
PieceType pt = pos.type_of_piece_on(from);
string san;
if (m == MOVE_NONE)
return "(none)";
if (m == MOVE_NULL)
return "(null)";
if (move_is_long_castle(m))
san = "O-O-O";
else if (move_is_short_castle(m))
san = "O-O";
else
{
if (pt != PAWN)
{
san = piece_type_to_char(pt);
// Collect all legal moves of piece type 'pt' with destination 'to'
MoveStack* last = generate<MV_LEGAL>(pos, mlist);
int f = 0, r = 0;
for (MoveStack* cur = mlist; cur != last; cur++)
if ( move_to(cur->move) == to
&& pos.type_of_piece_on(move_from(cur->move)) == pt)
{
if (square_file(move_from(cur->move)) == square_file(from))
f++;
if (square_rank(move_from(cur->move)) == square_rank(from))
r++;
}
assert(f > 0 && r > 0);
// Disambiguation if we have more then one piece with destination 'to'
if (f == 1 && r > 1)
san += file_to_char(square_file(from));
else if (f > 1 && r == 1)
san += rank_to_char(square_rank(from));
else if (f > 1 && r > 1)
san += square_to_string(from);
}
if (pos.move_is_capture(m))
{
if (pt == PAWN)
san += file_to_char(square_file(from));
san += 'x';
}
san += square_to_string(to);
if (move_is_promotion(m))
{
san += '=';
san += piece_type_to_char(move_promotion_piece(m));
}
}
// The move gives check? We don't use pos.move_gives_check() here
// because we need to test for a mate after the move is done.
StateInfo st;
pos.do_move(m, st);
if (pos.in_check())
san += pos.is_mate() ? "#" : "+";
pos.undo_move(m);
return san;
}
/// pretty_pv() creates a human-readable string from a position and a PV.
/// It is used to write search information to the log file (which is created
/// when the UCI parameter "Use Search Log" is "true").
const string pretty_pv(Position& pos, int depth, Value score, int time, Move pv[]) {
const int64_t K = 1000;
const int64_t M = 1000000;
const int startColumn = 28;
const size_t maxLength = 80 - startColumn;
const string lf = string("\n") + string(startColumn, ' ');
StateInfo state[PLY_MAX_PLUS_2], *st = state;
Move* m = pv;
string san;
std::stringstream s;
size_t length = 0;
// First print depth, score, time and searched nodes...
s << std::setw(2) << depth
<< std::setw(8) << score_string(score)
<< std::setw(8) << time_string(time);
if (pos.nodes_searched() < M)
s << std::setw(8) << pos.nodes_searched() / 1 << " ";
else if (pos.nodes_searched() < K * M)
s << std::setw(7) << pos.nodes_searched() / K << "K ";
else
s << std::setw(7) << pos.nodes_searched() / M << "M ";
// ...then print the full PV line in short algebraic notation
while (*m != MOVE_NONE)
{
san = move_to_san(pos, *m);
length += san.length() + 1;
if (length > maxLength)
{
length = san.length() + 1;
s << lf;
}
s << san << ' ';
pos.do_move(*m++, *st++);
}
// Restore original position before to leave
while (m != pv) pos.undo_move(*--m);
return s.str();
}
namespace {
const string time_string(int millisecs) {
const int MSecMinute = 1000 * 60;
const int MSecHour = 1000 * 60 * 60;
int hours = millisecs / MSecHour;
int minutes = (millisecs % MSecHour) / MSecMinute;
int seconds = ((millisecs % MSecHour) % MSecMinute) / 1000;
std::stringstream s;
if (hours)
s << hours << ':';
s << std::setfill('0') << std::setw(2) << minutes << ':' << std::setw(2) << seconds;
return s.str();
}
const string score_string(Value v) {
std::stringstream s;
if (v >= VALUE_MATE - 200)
s << "#" << (VALUE_MATE - v + 1) / 2;
else if (v <= -VALUE_MATE + 200)
s << "-#" << (VALUE_MATE + v) / 2;
else
s << std::setprecision(2) << std::fixed << std::showpos << float(v) / PawnValueMidgame;
return s.str();
}
}
-195
View File
@@ -1,195 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(MOVE_H_INCLUDED)
#define MOVE_H_INCLUDED
#include <string>
#include "misc.h"
#include "types.h"
// Maximum number of allowed moves per position
const int MAX_MOVES = 256;
/// A move needs 16 bits to be stored
///
/// bit 0- 5: destination square (from 0 to 63)
/// bit 6-11: origin square (from 0 to 63)
/// bit 12-13: promotion piece type - 2 (from KNIGHT-2 to QUEEN-2)
/// bit 14-15: special move flag: promotion (1), en passant (2), castle (3)
///
/// Special cases are MOVE_NONE and MOVE_NULL. We can sneak these in
/// because in any normal move destination square is always different
/// from origin square while MOVE_NONE and MOVE_NULL have the same
/// origin and destination square, 0 and 1 respectively.
enum Move {
MOVE_NONE = 0,
MOVE_NULL = 65
};
struct MoveStack {
Move move;
int score;
};
inline bool operator<(const MoveStack& f, const MoveStack& s) { return f.score < s.score; }
// An helper insertion sort implementation, works with pointers and iterators
template<typename T, typename K>
inline void insertion_sort(K firstMove, K lastMove)
{
T value;
K cur, p, d;
if (firstMove != lastMove)
for (cur = firstMove + 1; cur != lastMove; cur++)
{
p = d = cur;
value = *p--;
if (*p < value)
{
do *d = *p;
while (--d != firstMove && *--p < value);
*d = value;
}
}
}
// Our dedicated sort in range [firstMove, lastMove), first splits
// positive scores from ramining then order seaprately the two sets.
template<typename T>
inline void sort_moves(T* firstMove, T* lastMove, T** lastPositive)
{
T tmp;
T *p, *d;
d = lastMove;
p = firstMove - 1;
d->score = -1; // right guard
// Split positives vs non-positives
do {
while ((++p)->score > 0) {}
if (p != d)
{
while (--d != p && d->score <= 0) {}
tmp = *p;
*p = *d;
*d = tmp;
}
} while (p != d);
// Sort just positive scored moves, remaining only when we get there
insertion_sort<T, T*>(firstMove, p);
*lastPositive = p;
}
// Picks up the best move in range [curMove, lastMove), one per cycle.
// It is faster then sorting all the moves in advance when moves are few,
// as normally are the possible captures. Note that is not a stable alghoritm.
template<typename T>
inline T pick_best(T* curMove, T* lastMove)
{
T bestMove, tmp;
bestMove = *curMove;
while (++curMove != lastMove)
{
if (bestMove < *curMove)
{
tmp = *curMove;
*curMove = bestMove;
bestMove = tmp;
}
}
return bestMove;
}
inline Square move_from(Move m) {
return Square((int(m) >> 6) & 0x3F);
}
inline Square move_to(Move m) {
return Square(m & 0x3F);
}
inline bool move_is_special(Move m) {
return m & (3 << 14);
}
inline bool move_is_promotion(Move m) {
return (m & (3 << 14)) == (1 << 14);
}
inline int move_is_ep(Move m) {
return (m & (3 << 14)) == (2 << 14);
}
inline int move_is_castle(Move m) {
return (m & (3 << 14)) == (3 << 14);
}
inline bool move_is_short_castle(Move m) {
return move_is_castle(m) && (move_to(m) > move_from(m));
}
inline bool move_is_long_castle(Move m) {
return move_is_castle(m) && (move_to(m) < move_from(m));
}
inline PieceType move_promotion_piece(Move m) {
return move_is_promotion(m) ? PieceType(((int(m) >> 12) & 3) + 2) : PIECE_TYPE_NONE;
}
inline Move make_move(Square from, Square to) {
return Move(int(to) | (int(from) << 6));
}
inline Move make_promotion_move(Square from, Square to, PieceType promotion) {
return Move(int(to) | (int(from) << 6) | ((int(promotion) - 2) << 12) | (1 << 14));
}
inline Move make_ep_move(Square from, Square to) {
return Move(int(to) | (int(from) << 6) | (2 << 14));
}
inline Move make_castle_move(Square from, Square to) {
return Move(int(to) | (int(from) << 6) | (3 << 14));
}
inline bool move_is_ok(Move m) {
return move_from(m) != move_to(m); // Catches also MOVE_NONE
}
class Position;
extern const std::string move_to_uci(Move m, bool chess960);
extern Move move_from_uci(const Position& pos, const std::string& str);
extern const std::string move_to_san(Position& pos, Move m);
extern const std::string pretty_pv(Position& pos, int depth, Value score, int time, Move pv[]);
#endif // !defined(MOVE_H_INCLUDED)
+347 -469
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -19,524 +20,401 @@
#include <cassert> #include <cassert>
#include "bitcount.h"
#include "movegen.h" #include "movegen.h"
#include "position.h"
// Simple macro to wrap a very common while loop, no facny, no flexibility,
// hardcoded list name 'mlist' and from square 'from'.
#define SERIALIZE_MOVES(b) while (b) (*mlist++).move = make_move(from, pop_1st_bit(&b))
// Version used for pawns, where the 'from' square is given as a delta from the 'to' square
#define SERIALIZE_MOVES_D(b, d) while (b) { to = pop_1st_bit(&b); (*mlist++).move = make_move(to + (d), to); }
namespace { namespace {
enum CastlingSide { template<CastlingRight Cr, bool Checks, bool Chess960>
KING_SIDE, ExtMove* generate_castling(const Position& pos, ExtMove* moveList, Color us) {
QUEEN_SIDE
};
template<CastlingSide> static const bool KingSide = (Cr == WHITE_OO || Cr == BLACK_OO);
MoveStack* generate_castle_moves(const Position&, MoveStack*, Color us);
template<Color, MoveType> if (pos.castling_impeded(Cr) || !pos.can_castle(Cr))
MoveStack* generate_pawn_moves(const Position&, MoveStack*, Bitboard, Square); return moveList;
template<PieceType Pt> // After castling, the rook and king final positions are the same in Chess960
inline MoveStack* generate_discovered_checks(const Position& pos, MoveStack* mlist, Square from) { // as they would be in standard chess.
Square kfrom = pos.square<KING>(us);
Square rfrom = pos.castling_rook_square(Cr);
Square kto = relative_square(us, KingSide ? SQ_G1 : SQ_C1);
Bitboard enemies = pos.pieces(~us);
assert(Pt != QUEEN); assert(!pos.checkers());
Bitboard b = pos.attacks_from<Pt>(from) & pos.empty_squares(); const Direction K = Chess960 ? kto > kfrom ? WEST : EAST
if (Pt == KING) : KingSide ? WEST : EAST;
for (Square s = kto; s != kfrom; s += K)
if (pos.attackers_to(s) & enemies)
return moveList;
// Because we generate only legal castling moves we need to verify that
// when moving the castling rook we do not discover some hidden checker.
// For instance an enemy queen in SQ_A1 when castling rook is in SQ_B1.
if (Chess960 && (attacks_bb<ROOK>(kto, pos.pieces() ^ rfrom) & pos.pieces(~us, ROOK, QUEEN)))
return moveList;
Move m = make<CASTLING>(kfrom, rfrom);
if (Checks && !pos.gives_check(m))
return moveList;
*moveList++ = m;
return moveList;
}
template<GenType Type, Direction D>
ExtMove* make_promotions(ExtMove* moveList, Square to, Square ksq) {
if (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
*moveList++ = make<PROMOTION>(to - D, to, QUEEN);
if (Type == QUIETS || Type == EVASIONS || Type == NON_EVASIONS)
{ {
Square ksq = pos.king_square(opposite_color(pos.side_to_move())); *moveList++ = make<PROMOTION>(to - D, to, ROOK);
b &= ~QueenPseudoAttacks[ksq]; *moveList++ = make<PROMOTION>(to - D, to, BISHOP);
*moveList++ = make<PROMOTION>(to - D, to, KNIGHT);
} }
SERIALIZE_MOVES(b);
return mlist; // Knight promotion is the only promotion that can give a direct check
// that's not already included in the queen promotion.
if (Type == QUIET_CHECKS && (PseudoAttacks[KNIGHT][to] & ksq))
*moveList++ = make<PROMOTION>(to - D, to, KNIGHT);
else
(void)ksq; // Silence a warning under MSVC
return moveList;
} }
template<PieceType Pt>
inline MoveStack* generate_direct_checks(const Position& pos, MoveStack* mlist, Color us,
Bitboard dc, Square ksq) {
assert(Pt != KING);
Bitboard checkSqs, b; template<Color Us, GenType Type>
Square from; ExtMove* generate_pawn_moves(const Position& pos, ExtMove* moveList, Bitboard target) {
const Square* ptr = pos.piece_list_begin(us, Pt);
if ((from = *ptr++) == SQ_NONE) // Compute our parametrized parameters at compile time, named according to
return mlist; // the point of view of white side.
const Color Them = (Us == WHITE ? BLACK : WHITE);
const Bitboard TRank8BB = (Us == WHITE ? Rank8BB : Rank1BB);
const Bitboard TRank7BB = (Us == WHITE ? Rank7BB : Rank2BB);
const Bitboard TRank3BB = (Us == WHITE ? Rank3BB : Rank6BB);
const Direction Up = (Us == WHITE ? NORTH : SOUTH);
const Direction Right = (Us == WHITE ? NORTH_EAST : SOUTH_WEST);
const Direction Left = (Us == WHITE ? NORTH_WEST : SOUTH_EAST);
checkSqs = pos.attacks_from<Pt>(ksq) & pos.empty_squares(); Bitboard emptySquares;
do Bitboard pawnsOn7 = pos.pieces(Us, PAWN) & TRank7BB;
Bitboard pawnsNotOn7 = pos.pieces(Us, PAWN) & ~TRank7BB;
Bitboard enemies = (Type == EVASIONS ? pos.pieces(Them) & target:
Type == CAPTURES ? target : pos.pieces(Them));
// Single and double pawn pushes, no promotions
if (Type != CAPTURES)
{ {
if ( (Pt == QUEEN && !(QueenPseudoAttacks[from] & checkSqs)) emptySquares = (Type == QUIETS || Type == QUIET_CHECKS ? target : ~pos.pieces());
|| (Pt == ROOK && !(RookPseudoAttacks[from] & checkSqs))
|| (Pt == BISHOP && !(BishopPseudoAttacks[from] & checkSqs)))
continue;
if (dc && bit_is_set(dc, from)) Bitboard b1 = shift<Up>(pawnsNotOn7) & emptySquares;
continue; Bitboard b2 = shift<Up>(b1 & TRank3BB) & emptySquares;
b = pos.attacks_from<Pt>(from) & checkSqs; if (Type == EVASIONS) // Consider only blocking squares
SERIALIZE_MOVES(b);
} while ((from = *ptr++) != SQ_NONE);
return mlist;
}
template<>
FORCE_INLINE MoveStack* generate_direct_checks<PAWN>(const Position& p, MoveStack* m, Color us, Bitboard dc, Square ksq) {
return (us == WHITE ? generate_pawn_moves<WHITE, MV_CHECK>(p, m, dc, ksq)
: generate_pawn_moves<BLACK, MV_CHECK>(p, m, dc, ksq));
}
template<PieceType Pt, MoveType Type>
FORCE_INLINE MoveStack* generate_piece_moves(const Position& p, MoveStack* m, Color us, Bitboard t) {
assert(Pt == PAWN);
assert(Type == MV_CAPTURE || Type == MV_NON_CAPTURE || Type == MV_EVASION);
return (us == WHITE ? generate_pawn_moves<WHITE, Type>(p, m, t, SQ_NONE)
: generate_pawn_moves<BLACK, Type>(p, m, t, SQ_NONE));
}
template<PieceType Pt>
FORCE_INLINE MoveStack* generate_piece_moves(const Position& pos, MoveStack* mlist, Color us, Bitboard target) {
Bitboard b;
Square from;
const Square* ptr = pos.piece_list_begin(us, Pt);
if (*ptr != SQ_NONE)
{
do {
from = *ptr;
b = pos.attacks_from<Pt>(from) & target;
SERIALIZE_MOVES(b);
} while (*++ptr != SQ_NONE);
}
return mlist;
}
template<>
FORCE_INLINE MoveStack* generate_piece_moves<KING>(const Position& pos, MoveStack* mlist, Color us, Bitboard target) {
Bitboard b;
Square from = pos.king_square(us);
b = pos.attacks_from<KING>(from) & target;
SERIALIZE_MOVES(b);
return mlist;
}
}
/// generate<MV_CAPTURE> generates all pseudo-legal captures and queen
/// promotions. Returns a pointer to the end of the move list.
///
/// generate<MV_NON_CAPTURE> generates all pseudo-legal non-captures and
/// underpromotions. Returns a pointer to the end of the move list.
///
/// generate<MV_NON_EVASION> generates all pseudo-legal captures and
/// non-captures. Returns a pointer to the end of the move list.
template<MoveType Type>
MoveStack* generate(const Position& pos, MoveStack* mlist) {
assert(pos.is_ok());
assert(!pos.in_check());
Color us = pos.side_to_move();
Bitboard target;
if (Type == MV_CAPTURE || Type == MV_NON_EVASION)
target = pos.pieces_of_color(opposite_color(us));
else if (Type == MV_NON_CAPTURE)
target = pos.empty_squares();
else
assert(false);
if (Type == MV_NON_EVASION)
{
mlist = generate_piece_moves<PAWN, MV_CAPTURE>(pos, mlist, us, target);
mlist = generate_piece_moves<PAWN, MV_NON_CAPTURE>(pos, mlist, us, pos.empty_squares());
target |= pos.empty_squares();
}
else
mlist = generate_piece_moves<PAWN, Type>(pos, mlist, us, target);
mlist = generate_piece_moves<KNIGHT>(pos, mlist, us, target);
mlist = generate_piece_moves<BISHOP>(pos, mlist, us, target);
mlist = generate_piece_moves<ROOK>(pos, mlist, us, target);
mlist = generate_piece_moves<QUEEN>(pos, mlist, us, target);
mlist = generate_piece_moves<KING>(pos, mlist, us, target);
if (Type != MV_CAPTURE)
{
if (pos.can_castle_kingside(us))
mlist = generate_castle_moves<KING_SIDE>(pos, mlist, us);
if (pos.can_castle_queenside(us))
mlist = generate_castle_moves<QUEEN_SIDE>(pos, mlist, us);
}
return mlist;
}
// Explicit template instantiations
template MoveStack* generate<MV_CAPTURE>(const Position& pos, MoveStack* mlist);
template MoveStack* generate<MV_NON_CAPTURE>(const Position& pos, MoveStack* mlist);
template MoveStack* generate<MV_NON_EVASION>(const Position& pos, MoveStack* mlist);
/// generate_non_capture_checks() generates all pseudo-legal non-captures and knight
/// underpromotions that give check. Returns a pointer to the end of the move list.
template<>
MoveStack* generate<MV_NON_CAPTURE_CHECK>(const Position& pos, MoveStack* mlist) {
assert(pos.is_ok());
assert(!pos.in_check());
Bitboard b, dc;
Square from;
Color us = pos.side_to_move();
Square ksq = pos.king_square(opposite_color(us));
assert(pos.piece_on(ksq) == make_piece(opposite_color(us), KING));
// Discovered non-capture checks
b = dc = pos.discovered_check_candidates(us);
while (b)
{
from = pop_1st_bit(&b);
switch (pos.type_of_piece_on(from))
{
case PAWN: /* Will be generated togheter with pawns direct checks */ break;
case KNIGHT: mlist = generate_discovered_checks<KNIGHT>(pos, mlist, from); break;
case BISHOP: mlist = generate_discovered_checks<BISHOP>(pos, mlist, from); break;
case ROOK: mlist = generate_discovered_checks<ROOK>(pos, mlist, from); break;
case KING: mlist = generate_discovered_checks<KING>(pos, mlist, from); break;
default: assert(false); break;
}
}
// Direct non-capture checks
mlist = generate_direct_checks<PAWN>(pos, mlist, us, dc, ksq);
mlist = generate_direct_checks<KNIGHT>(pos, mlist, us, dc, ksq);
mlist = generate_direct_checks<BISHOP>(pos, mlist, us, dc, ksq);
mlist = generate_direct_checks<ROOK>(pos, mlist, us, dc, ksq);
return generate_direct_checks<QUEEN>(pos, mlist, us, dc, ksq);
}
/// generate_evasions() generates all pseudo-legal check evasions when
/// the side to move is in check. Returns a pointer to the end of the move list.
template<>
MoveStack* generate<MV_EVASION>(const Position& pos, MoveStack* mlist) {
assert(pos.is_ok());
assert(pos.in_check());
Bitboard b, target;
Square from, checksq;
int checkersCnt = 0;
Color us = pos.side_to_move();
Square ksq = pos.king_square(us);
Bitboard checkers = pos.checkers();
Bitboard sliderAttacks = EmptyBoardBB;
assert(pos.piece_on(ksq) == make_piece(us, KING));
assert(checkers);
// Find squares attacked by slider checkers, we will remove
// them from the king evasions set so to early skip known
// illegal moves and avoid an useless legality check later.
b = checkers;
do
{
checkersCnt++;
checksq = pop_1st_bit(&b);
assert(pos.color_of_piece_on(checksq) == opposite_color(us));
switch (pos.type_of_piece_on(checksq))
{
case BISHOP: sliderAttacks |= BishopPseudoAttacks[checksq]; break;
case ROOK: sliderAttacks |= RookPseudoAttacks[checksq]; break;
case QUEEN:
// In case of a queen remove also squares attacked in the other direction to
// avoid possible illegal moves when queen and king are on adjacent squares.
if (RookPseudoAttacks[checksq] & (1ULL << ksq))
sliderAttacks |= RookPseudoAttacks[checksq] | pos.attacks_from<BISHOP>(checksq);
else
sliderAttacks |= BishopPseudoAttacks[checksq] | pos.attacks_from<ROOK>(checksq);
default:
break;
}
} while (b);
// Generate evasions for king, capture and non capture moves
b = pos.attacks_from<KING>(ksq) & ~pos.pieces_of_color(us) & ~sliderAttacks;
from = ksq;
SERIALIZE_MOVES(b);
// Generate evasions for other pieces only if not double check
if (checkersCnt > 1)
return mlist;
// Find squares where a blocking evasion or a capture of the
// checker piece is possible.
target = squares_between(checksq, ksq) | checkers;
mlist = generate_piece_moves<PAWN, MV_EVASION>(pos, mlist, us, target);
mlist = generate_piece_moves<KNIGHT>(pos, mlist, us, target);
mlist = generate_piece_moves<BISHOP>(pos, mlist, us, target);
mlist = generate_piece_moves<ROOK>(pos, mlist, us, target);
return generate_piece_moves<QUEEN>(pos, mlist, us, target);
}
/// generate<MV_LEGAL / MV_PSEUDO_LEGAL> computes a complete list of legal
/// or pseudo-legal moves in the current position.
template<>
MoveStack* generate<MV_PSEUDO_LEGAL>(const Position& pos, MoveStack* mlist) {
assert(pos.is_ok());
return pos.in_check() ? generate<MV_EVASION>(pos, mlist)
: generate<MV_NON_EVASION>(pos, mlist);
}
template<>
MoveStack* generate<MV_LEGAL>(const Position& pos, MoveStack* mlist) {
assert(pos.is_ok());
MoveStack *last, *cur = mlist;
Bitboard pinned = pos.pinned_pieces(pos.side_to_move());
last = generate<MV_PSEUDO_LEGAL>(pos, mlist);
// Remove illegal moves from the list
while (cur != last)
if (!pos.pl_move_is_legal(cur->move, pinned))
cur->move = (--last)->move;
else
cur++;
return last;
}
namespace {
template<Square Delta>
inline Bitboard move_pawns(Bitboard p) {
return Delta == DELTA_N ? p << 8 : Delta == DELTA_S ? p >> 8 :
Delta == DELTA_NE ? p << 9 : Delta == DELTA_SE ? p >> 7 :
Delta == DELTA_NW ? p << 7 : Delta == DELTA_SW ? p >> 9 : p;
}
template<MoveType Type, Square Delta>
inline MoveStack* generate_pawn_captures(MoveStack* mlist, Bitboard pawns, Bitboard target) {
const Bitboard TFileABB = (Delta == DELTA_NE || Delta == DELTA_SE ? FileABB : FileHBB);
Bitboard b;
Square to;
// Captures in the a1-h8 (a8-h1 for black) diagonal or in the h1-a8 (h8-a1 for black)
b = move_pawns<Delta>(pawns) & target & ~TFileABB;
SERIALIZE_MOVES_D(b, -Delta);
return mlist;
}
template<Color Us, MoveType Type, Square Delta>
inline MoveStack* generate_promotions(const Position& pos, MoveStack* mlist, Bitboard pawnsOn7, Bitboard target) {
const Bitboard TFileABB = (Delta == DELTA_NE || Delta == DELTA_SE ? FileABB : FileHBB);
Bitboard b;
Square to;
// Promotions and under-promotions, both captures and non-captures
b = move_pawns<Delta>(pawnsOn7) & target;
if (Delta != DELTA_N && Delta != DELTA_S)
b &= ~TFileABB;
while (b)
{
to = pop_1st_bit(&b);
if (Type == MV_CAPTURE || Type == MV_EVASION)
(*mlist++).move = make_promotion_move(to - Delta, to, QUEEN);
if (Type == MV_NON_CAPTURE || Type == MV_EVASION)
{ {
(*mlist++).move = make_promotion_move(to - Delta, to, ROOK); b1 &= target;
(*mlist++).move = make_promotion_move(to - Delta, to, BISHOP); b2 &= target;
(*mlist++).move = make_promotion_move(to - Delta, to, KNIGHT);
} }
// This is the only possible under promotion that can give a check if (Type == QUIET_CHECKS)
// not already included in the queen-promotion.
if ( Type == MV_CHECK
&& bit_is_set(pos.attacks_from<KNIGHT>(to), pos.king_square(opposite_color(Us))))
(*mlist++).move = make_promotion_move(to - Delta, to, KNIGHT);
else (void)pos; // Silence a warning under MSVC
}
return mlist;
}
template<Color Us, MoveType Type>
MoveStack* generate_pawn_moves(const Position& pos, MoveStack* mlist, Bitboard target, Square ksq) {
// Calculate our parametrized parameters at compile time, named
// according to the point of view of white side.
const Color Them = (Us == WHITE ? BLACK : WHITE);
const Bitboard TRank7BB = (Us == WHITE ? Rank7BB : Rank2BB);
const Bitboard TRank3BB = (Us == WHITE ? Rank3BB : Rank6BB);
const Square TDELTA_N = (Us == WHITE ? DELTA_N : DELTA_S);
const Square TDELTA_NE = (Us == WHITE ? DELTA_NE : DELTA_SE);
const Square TDELTA_NW = (Us == WHITE ? DELTA_NW : DELTA_SW);
Square to;
Bitboard b1, b2, dc1, dc2, pawnPushes, emptySquares;
Bitboard pawns = pos.pieces(PAWN, Us);
Bitboard pawnsOn7 = pawns & TRank7BB;
Bitboard enemyPieces = (Type == MV_CAPTURE ? target : pos.pieces_of_color(Them));
// Pre-calculate pawn pushes before changing emptySquares definition
if (Type != MV_CAPTURE)
{
emptySquares = (Type == MV_NON_CAPTURE ? target : pos.empty_squares());
pawnPushes = move_pawns<TDELTA_N>(pawns & ~TRank7BB) & emptySquares;
}
if (Type == MV_EVASION)
{
emptySquares &= target; // Only blocking squares
enemyPieces &= target; // Capture only the checker piece
}
// Promotions and underpromotions
if (pawnsOn7)
{
if (Type == MV_CAPTURE)
emptySquares = pos.empty_squares();
pawns &= ~TRank7BB;
mlist = generate_promotions<Us, Type, TDELTA_NE>(pos, mlist, pawnsOn7, enemyPieces);
mlist = generate_promotions<Us, Type, TDELTA_NW>(pos, mlist, pawnsOn7, enemyPieces);
mlist = generate_promotions<Us, Type, TDELTA_N >(pos, mlist, pawnsOn7, emptySquares);
}
// Standard captures
if (Type == MV_CAPTURE || Type == MV_EVASION)
{
mlist = generate_pawn_captures<Type, TDELTA_NE>(mlist, pawns, enemyPieces);
mlist = generate_pawn_captures<Type, TDELTA_NW>(mlist, pawns, enemyPieces);
}
// Single and double pawn pushes
if (Type != MV_CAPTURE)
{
b1 = pawnPushes & emptySquares;
b2 = move_pawns<TDELTA_N>(pawnPushes & TRank3BB) & emptySquares;
if (Type == MV_CHECK)
{ {
// Consider only pawn moves which give direct checks Square ksq = pos.square<KING>(Them);
b1 &= pos.attacks_from<PAWN>(ksq, Them); b1 &= pos.attacks_from<PAWN>(ksq, Them);
b2 &= pos.attacks_from<PAWN>(ksq, Them); b2 &= pos.attacks_from<PAWN>(ksq, Them);
// Add pawn moves which gives discovered check. This is possible only // Add pawn pushes which give discovered check. This is possible only
// if the pawn is not on the same file as the enemy king, because we // if the pawn is not on the same file as the enemy king, because we
// don't generate captures. // don't generate captures. Note that a possible discovery check
if (pawns & target) // For CHECK type target is dc bitboard // promotion has been already generated amongst the captures.
Bitboard dcCandidates = pos.discovered_check_candidates();
if (pawnsNotOn7 & dcCandidates)
{ {
dc1 = move_pawns<TDELTA_N>(pawns & target & ~file_bb(ksq)) & emptySquares; Bitboard dc1 = shift<Up>(pawnsNotOn7 & dcCandidates) & emptySquares & ~file_bb(ksq);
dc2 = move_pawns<TDELTA_N>(dc1 & TRank3BB) & emptySquares; Bitboard dc2 = shift<Up>(dc1 & TRank3BB) & emptySquares;
b1 |= dc1; b1 |= dc1;
b2 |= dc2; b2 |= dc2;
} }
} }
SERIALIZE_MOVES_D(b1, -TDELTA_N);
SERIALIZE_MOVES_D(b2, -TDELTA_N -TDELTA_N);
}
// En passant captures
if ((Type == MV_CAPTURE || Type == MV_EVASION) && pos.ep_square() != SQ_NONE)
{
assert(Us != WHITE || square_rank(pos.ep_square()) == RANK_6);
assert(Us != BLACK || square_rank(pos.ep_square()) == RANK_3);
// An en passant capture can be an evasion only if the checking piece
// is the double pushed pawn and so is in the target. Otherwise this
// is a discovery check and we are forced to do otherwise.
if (Type == MV_EVASION && !bit_is_set(target, pos.ep_square() - TDELTA_N))
return mlist;
b1 = pawns & pos.attacks_from<PAWN>(pos.ep_square(), Them);
assert(b1 != EmptyBoardBB);
while (b1) while (b1)
{ {
to = pop_1st_bit(&b1); Square to = pop_lsb(&b1);
(*mlist++).move = make_ep_move(to, pos.ep_square()); *moveList++ = make_move(to - Up, to);
}
while (b2)
{
Square to = pop_lsb(&b2);
*moveList++ = make_move(to - Up - Up, to);
} }
} }
return mlist;
// Promotions and underpromotions
if (pawnsOn7 && (Type != EVASIONS || (target & TRank8BB)))
{
if (Type == CAPTURES)
emptySquares = ~pos.pieces();
if (Type == EVASIONS)
emptySquares &= target;
Bitboard b1 = shift<Right>(pawnsOn7) & enemies;
Bitboard b2 = shift<Left >(pawnsOn7) & enemies;
Bitboard b3 = shift<Up >(pawnsOn7) & emptySquares;
Square ksq = pos.square<KING>(Them);
while (b1)
moveList = make_promotions<Type, Right>(moveList, pop_lsb(&b1), ksq);
while (b2)
moveList = make_promotions<Type, Left >(moveList, pop_lsb(&b2), ksq);
while (b3)
moveList = make_promotions<Type, Up >(moveList, pop_lsb(&b3), ksq);
}
// Standard and en-passant captures
if (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
{
Bitboard b1 = shift<Right>(pawnsNotOn7) & enemies;
Bitboard b2 = shift<Left >(pawnsNotOn7) & enemies;
while (b1)
{
Square to = pop_lsb(&b1);
*moveList++ = make_move(to - Right, to);
}
while (b2)
{
Square to = pop_lsb(&b2);
*moveList++ = make_move(to - Left, to);
}
if (pos.ep_square() != SQ_NONE)
{
assert(rank_of(pos.ep_square()) == relative_rank(Us, RANK_6));
// An en passant capture can be an evasion only if the checking piece
// is the double pushed pawn and so is in the target. Otherwise this
// is a discovery check and we are forced to do otherwise.
if (Type == EVASIONS && !(target & (pos.ep_square() - Up)))
return moveList;
b1 = pawnsNotOn7 & pos.attacks_from<PAWN>(pos.ep_square(), Them);
assert(b1);
while (b1)
*moveList++ = make<ENPASSANT>(pop_lsb(&b1), pos.ep_square());
}
}
return moveList;
} }
template<CastlingSide Side>
MoveStack* generate_castle_moves(const Position& pos, MoveStack* mlist, Color us) {
Color them = opposite_color(us); template<PieceType Pt, bool Checks>
Square ksq = pos.king_square(us); ExtMove* generate_moves(const Position& pos, ExtMove* moveList, Color us,
Bitboard target) {
assert(pos.piece_on(ksq) == make_piece(us, KING)); assert(Pt != KING && Pt != PAWN);
Square rsq = (Side == KING_SIDE ? pos.initial_kr_square(us) : pos.initial_qr_square(us)); const Square* pl = pos.squares<Pt>(us);
Square s1 = relative_square(us, Side == KING_SIDE ? SQ_G1 : SQ_C1);
Square s2 = relative_square(us, Side == KING_SIDE ? SQ_F1 : SQ_D1);
Square s;
bool illegal = false;
assert(pos.piece_on(rsq) == make_piece(us, ROOK)); for (Square from = *pl; from != SQ_NONE; from = *++pl)
{
if (Checks)
{
if ( (Pt == BISHOP || Pt == ROOK || Pt == QUEEN)
&& !(PseudoAttacks[Pt][from] & target & pos.check_squares(Pt)))
continue;
// It is a bit complicated to correctly handle Chess960 if (pos.discovered_check_candidates() & from)
for (s = Min(ksq, s1); s <= Max(ksq, s1); s++) continue;
if ( (s != ksq && s != rsq && pos.square_is_occupied(s)) }
||(pos.attackers_to(s) & pos.pieces_of_color(them)))
illegal = true;
for (s = Min(rsq, s2); s <= Max(rsq, s2); s++) Bitboard b = pos.attacks_from<Pt>(from) & target;
if (s != ksq && s != rsq && pos.square_is_occupied(s))
illegal = true;
if ( Side == QUEEN_SIDE if (Checks)
&& square_file(rsq) == FILE_B b &= pos.check_squares(Pt);
&& ( pos.piece_on(relative_square(us, SQ_A1)) == make_piece(them, ROOK)
|| pos.piece_on(relative_square(us, SQ_A1)) == make_piece(them, QUEEN)))
illegal = true;
if (!illegal) while (b)
(*mlist++).move = make_castle_move(ksq, rsq); *moveList++ = make_move(from, pop_lsb(&b));
}
return mlist; return moveList;
}
template<Color Us, GenType Type>
ExtMove* generate_all(const Position& pos, ExtMove* moveList, Bitboard target) {
const bool Checks = Type == QUIET_CHECKS;
moveList = generate_pawn_moves<Us, Type>(pos, moveList, target);
moveList = generate_moves<KNIGHT, Checks>(pos, moveList, Us, target);
moveList = generate_moves<BISHOP, Checks>(pos, moveList, Us, target);
moveList = generate_moves< ROOK, Checks>(pos, moveList, Us, target);
moveList = generate_moves< QUEEN, Checks>(pos, moveList, Us, target);
if (Type != QUIET_CHECKS && Type != EVASIONS)
{
Square ksq = pos.square<KING>(Us);
Bitboard b = pos.attacks_from<KING>(ksq) & target;
while (b)
*moveList++ = make_move(ksq, pop_lsb(&b));
}
if (Type != CAPTURES && Type != EVASIONS && pos.can_castle(Us))
{
if (pos.is_chess960())
{
moveList = generate_castling<MakeCastling<Us, KING_SIDE>::right, Checks, true>(pos, moveList, Us);
moveList = generate_castling<MakeCastling<Us, QUEEN_SIDE>::right, Checks, true>(pos, moveList, Us);
}
else
{
moveList = generate_castling<MakeCastling<Us, KING_SIDE>::right, Checks, false>(pos, moveList, Us);
moveList = generate_castling<MakeCastling<Us, QUEEN_SIDE>::right, Checks, false>(pos, moveList, Us);
}
}
return moveList;
} }
} // namespace } // namespace
/// generate<CAPTURES> generates all pseudo-legal captures and queen
/// promotions. Returns a pointer to the end of the move list.
///
/// generate<QUIETS> generates all pseudo-legal non-captures and
/// underpromotions. Returns a pointer to the end of the move list.
///
/// generate<NON_EVASIONS> generates all pseudo-legal captures and
/// non-captures. Returns a pointer to the end of the move list.
template<GenType Type>
ExtMove* generate(const Position& pos, ExtMove* moveList) {
assert(Type == CAPTURES || Type == QUIETS || Type == NON_EVASIONS);
assert(!pos.checkers());
Color us = pos.side_to_move();
Bitboard target = Type == CAPTURES ? pos.pieces(~us)
: Type == QUIETS ? ~pos.pieces()
: Type == NON_EVASIONS ? ~pos.pieces(us) : 0;
return us == WHITE ? generate_all<WHITE, Type>(pos, moveList, target)
: generate_all<BLACK, Type>(pos, moveList, target);
}
// Explicit template instantiations
template ExtMove* generate<CAPTURES>(const Position&, ExtMove*);
template ExtMove* generate<QUIETS>(const Position&, ExtMove*);
template ExtMove* generate<NON_EVASIONS>(const Position&, ExtMove*);
/// generate<QUIET_CHECKS> generates all pseudo-legal non-captures and knight
/// underpromotions that give check. Returns a pointer to the end of the move list.
template<>
ExtMove* generate<QUIET_CHECKS>(const Position& pos, ExtMove* moveList) {
assert(!pos.checkers());
Color us = pos.side_to_move();
Bitboard dc = pos.discovered_check_candidates();
while (dc)
{
Square from = pop_lsb(&dc);
PieceType pt = type_of(pos.piece_on(from));
if (pt == PAWN)
continue; // Will be generated together with direct checks
Bitboard b = pos.attacks_from(pt, from) & ~pos.pieces();
if (pt == KING)
b &= ~PseudoAttacks[QUEEN][pos.square<KING>(~us)];
while (b)
*moveList++ = make_move(from, pop_lsb(&b));
}
return us == WHITE ? generate_all<WHITE, QUIET_CHECKS>(pos, moveList, ~pos.pieces())
: generate_all<BLACK, QUIET_CHECKS>(pos, moveList, ~pos.pieces());
}
/// generate<EVASIONS> generates all pseudo-legal check evasions when the side
/// to move is in check. Returns a pointer to the end of the move list.
template<>
ExtMove* generate<EVASIONS>(const Position& pos, ExtMove* moveList) {
assert(pos.checkers());
Color us = pos.side_to_move();
Square ksq = pos.square<KING>(us);
Bitboard sliderAttacks = 0;
Bitboard sliders = pos.checkers() & ~pos.pieces(KNIGHT, PAWN);
// Find all the squares attacked by slider checkers. We will remove them from
// the king evasions in order to skip known illegal moves, which avoids any
// useless legality checks later on.
while (sliders)
{
Square checksq = pop_lsb(&sliders);
sliderAttacks |= LineBB[checksq][ksq] ^ checksq;
}
// Generate evasions for king, capture and non capture moves
Bitboard b = pos.attacks_from<KING>(ksq) & ~pos.pieces(us) & ~sliderAttacks;
while (b)
*moveList++ = make_move(ksq, pop_lsb(&b));
if (more_than_one(pos.checkers()))
return moveList; // Double check, only a king move can save the day
// Generate blocking evasions or captures of the checking piece
Square checksq = lsb(pos.checkers());
Bitboard target = between_bb(checksq, ksq) | checksq;
return us == WHITE ? generate_all<WHITE, EVASIONS>(pos, moveList, target)
: generate_all<BLACK, EVASIONS>(pos, moveList, target);
}
/// generate<LEGAL> generates all the legal moves in the given position
template<>
ExtMove* generate<LEGAL>(const Position& pos, ExtMove* moveList) {
Bitboard pinned = pos.pinned_pieces(pos.side_to_move());
Square ksq = pos.square<KING>(pos.side_to_move());
ExtMove* cur = moveList;
moveList = pos.checkers() ? generate<EVASIONS >(pos, moveList)
: generate<NON_EVASIONS>(pos, moveList);
while (cur != moveList)
if ( (pinned || from_sq(*cur) == ksq || type_of(*cur) == ENPASSANT)
&& !pos.legal(*cur))
*cur = (--moveList)->move;
else
++cur;
return moveList;
}
+51 -16
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,24 +18,58 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(MOVEGEN_H_INCLUDED) #ifndef MOVEGEN_H_INCLUDED
#define MOVEGEN_H_INCLUDED #define MOVEGEN_H_INCLUDED
#include "move.h" #include <algorithm>
#include "position.h"
enum MoveType { #include "types.h"
MV_CAPTURE,
MV_NON_CAPTURE, class Position;
MV_CHECK,
MV_NON_CAPTURE_CHECK, enum GenType {
MV_EVASION, CAPTURES,
MV_NON_EVASION, QUIETS,
MV_LEGAL, QUIET_CHECKS,
MV_PSEUDO_LEGAL EVASIONS,
NON_EVASIONS,
LEGAL
}; };
template<MoveType> struct ExtMove {
MoveStack* generate(const Position& pos, MoveStack* mlist); Move move;
int value;
#endif // !defined(MOVEGEN_H_INCLUDED) operator Move() const { return move; }
void operator=(Move m) { move = m; }
// Inhibit unwanted implicit conversions to Move
// with an ambiguity that yields to a compile error.
operator float() const = delete;
};
inline bool operator<(const ExtMove& f, const ExtMove& s) {
return f.value < s.value;
}
template<GenType>
ExtMove* generate(const Position& pos, ExtMove* moveList);
/// The MoveList struct is a simple wrapper around generate(). It sometimes comes
/// in handy to use this class instead of the low level generate() function.
template<GenType T>
struct MoveList {
explicit MoveList(const Position& pos) : last(generate<T>(pos, moveList)) {}
const ExtMove* begin() const { return moveList; }
const ExtMove* end() const { return last; }
size_t size() const { return last - moveList; }
bool contains(Move move) const {
return std::find(begin(), end(), move) != end();
}
private:
ExtMove moveList[MAX_MOVES], *last;
};
#endif // #ifndef MOVEGEN_H_INCLUDED
+266 -282
View File
@@ -1,14 +1,14 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or the Free Software Foundation, either version 3 of the License, or
(at your option) any later version. (at your option) any later version.
Stockfish is distributed in the hope that it will be useful, Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
@@ -20,331 +20,315 @@
#include <cassert> #include <cassert>
#include "movegen.h"
#include "movepick.h" #include "movepick.h"
#include "search.h"
#include "types.h"
namespace { namespace {
enum MovegenPhase { enum Stages {
PH_TT_MOVES, // Transposition table move and mate killer MAIN_SEARCH, CAPTURES_INIT, GOOD_CAPTURES, KILLERS, COUNTERMOVE, QUIET_INIT, QUIET, BAD_CAPTURES,
PH_GOOD_CAPTURES, // Queen promotions and captures with SEE values >= 0 EVASION, EVASIONS_INIT, ALL_EVASIONS,
PH_KILLERS, // Killer moves from the current ply PROBCUT, PROBCUT_INIT, PROBCUT_CAPTURES,
PH_NONCAPTURES, // Non-captures and underpromotions QSEARCH_WITH_CHECKS, QCAPTURES_1_INIT, QCAPTURES_1, QCHECKS,
PH_BAD_CAPTURES, // Queen promotions and captures with SEE values < 0 QSEARCH_NO_CHECKS, QCAPTURES_2_INIT, QCAPTURES_2,
PH_EVASIONS, // Check evasions QSEARCH_RECAPTURES, QRECAPTURES
PH_QCAPTURES, // Captures in quiescence search
PH_QCHECKS, // Non-capture checks in quiescence search
PH_STOP
}; };
CACHE_LINE_ALIGNMENT // partial_insertion_sort() sorts moves in descending order up to and including
const uint8_t MainSearchTable[] = { PH_TT_MOVES, PH_GOOD_CAPTURES, PH_KILLERS, PH_NONCAPTURES, PH_BAD_CAPTURES, PH_STOP }; // a given limit. The order of moves smaller than the limit is left unspecified.
const uint8_t EvasionTable[] = { PH_TT_MOVES, PH_EVASIONS, PH_STOP }; void partial_insertion_sort(ExtMove* begin, ExtMove* end, int limit) {
const uint8_t QsearchWithChecksTable[] = { PH_TT_MOVES, PH_QCAPTURES, PH_QCHECKS, PH_STOP };
const uint8_t QsearchWithoutChecksTable[] = { PH_TT_MOVES, PH_QCAPTURES, PH_STOP }; for (ExtMove *sortedEnd = begin, *p = begin + 1; p < end; ++p)
} if (p->value >= limit)
{
ExtMove tmp = *p, *q;
*p = *++sortedEnd;
for (q = sortedEnd; q != begin && *(q - 1) < tmp; --q)
*q = *(q - 1);
*q = tmp;
}
}
// pick_best() finds the best move in the range (begin, end) and moves it to
// the front. It's faster than sorting all the moves in advance when there
// are few moves, e.g., the possible captures.
Move pick_best(ExtMove* begin, ExtMove* end) {
std::swap(*begin, *std::max_element(begin, end));
return *begin;
}
} // namespace
/// Constructor for the MovePicker class. Apart from the position for which /// Constructors of the MovePicker class. As arguments we pass information
/// it is asked to pick legal moves, MovePicker also wants some information /// to help it to return the (presumably) good moves first, to decide which
/// to help it to return the presumably good moves first, to decide which
/// moves to return (in the quiescence search, for instance, we only want to /// moves to return (in the quiescence search, for instance, we only want to
/// search captures, promotions and some checks) and about how important good /// search captures, promotions, and some checks) and how important good move
/// move ordering is at the current node. /// ordering is at the current node.
MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const History& h, /// MovePicker constructor for the main search
SearchStack* ss, Value beta) : pos(p), H(h) { MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHistory* mh,
int searchTT = ttm; const CapturePieceToHistory* cph, const PieceToHistory** ch, Move cm, Move* killers_p)
ttMoves[0].move = ttm; : pos(p), mainHistory(mh), captureHistory(cph), contHistory(ch), countermove(cm),
badCaptureThreshold = 0; killers{killers_p[0], killers_p[1]}, depth(d){
badCaptures = moves + MAX_MOVES;
assert(d > DEPTH_ZERO); assert(d > DEPTH_ZERO);
pinned = p.pinned_pieces(pos.side_to_move()); stage = pos.checkers() ? EVASION : MAIN_SEARCH;
ttMove = ttm && pos.pseudo_legal(ttm) ? ttm : MOVE_NONE;
if (p.in_check()) stage += (ttMove == MOVE_NONE);
{
ttMoves[1].move = killers[0].move = killers[1].move = MOVE_NONE;
phasePtr = EvasionTable;
}
else
{
ttMoves[1].move = (ss->mateKiller == ttm) ? MOVE_NONE : ss->mateKiller;
searchTT |= ttMoves[1].move;
killers[0].move = ss->killers[0];
killers[1].move = ss->killers[1];
// Consider sligtly negative captures as good if at low
// depth and far from beta.
if (ss && ss->eval < beta - PawnValueMidgame && d < 3 * ONE_PLY)
badCaptureThreshold = -PawnValueMidgame;
phasePtr = MainSearchTable;
}
phasePtr += int(!searchTT) - 1;
go_next_phase();
} }
MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const History& h) /// MovePicker constructor for quiescence search
: pos(p), H(h) { MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHistory* mh, const CapturePieceToHistory* cph, Square s)
int searchTT = ttm; : pos(p), mainHistory(mh), captureHistory(cph) {
ttMoves[0].move = ttm;
ttMoves[1].move = MOVE_NONE;
assert(d <= DEPTH_ZERO); assert(d <= DEPTH_ZERO);
pinned = p.pinned_pieces(pos.side_to_move()); if (pos.checkers())
stage = EVASION;
else if (d > DEPTH_QS_NO_CHECKS)
stage = QSEARCH_WITH_CHECKS;
else if (d > DEPTH_QS_RECAPTURES)
stage = QSEARCH_NO_CHECKS;
if (p.in_check())
phasePtr = EvasionTable;
else if (d >= DEPTH_QS_CHECKS)
phasePtr = QsearchWithChecksTable;
else else
{ {
phasePtr = QsearchWithoutChecksTable; stage = QSEARCH_RECAPTURES;
recaptureSquare = s;
// Skip TT move if is not a capture or a promotion, this avoids
// qsearch tree explosion due to a possible perpetual check or
// similar rare cases when TT table is full.
if (ttm != MOVE_NONE && !pos.move_is_capture_or_promotion(ttm))
searchTT = ttMoves[0].move = MOVE_NONE;
}
phasePtr += int(!searchTT) - 1;
go_next_phase();
}
/// MovePicker::go_next_phase() generates, scores and sorts the next bunch
/// of moves when there are no more moves to try for the current phase.
void MovePicker::go_next_phase() {
curMove = moves;
phase = *(++phasePtr);
switch (phase) {
case PH_TT_MOVES:
curMove = ttMoves;
lastMove = curMove + 2;
return;
case PH_GOOD_CAPTURES:
lastMove = generate<MV_CAPTURE>(pos, moves);
score_captures();
return;
case PH_KILLERS:
curMove = killers;
lastMove = curMove + 2;
return;
case PH_NONCAPTURES:
lastMove = generate<MV_NON_CAPTURE>(pos, moves);
score_noncaptures();
sort_moves(moves, lastMove, &lastGoodNonCapture);
return;
case PH_BAD_CAPTURES:
// Bad captures SEE value is already calculated so just pick
// them in order to get SEE move ordering.
curMove = badCaptures;
lastMove = moves + MAX_MOVES;
return;
case PH_EVASIONS:
assert(pos.in_check());
lastMove = generate<MV_EVASION>(pos, moves);
score_evasions();
return;
case PH_QCAPTURES:
lastMove = generate<MV_CAPTURE>(pos, moves);
score_captures();
return;
case PH_QCHECKS:
lastMove = generate<MV_NON_CAPTURE_CHECK>(pos, moves);
return;
case PH_STOP:
lastMove = curMove + 1; // Avoid another go_next_phase() call
return;
default:
assert(false);
return; return;
} }
ttMove = ttm && pos.pseudo_legal(ttm) ? ttm : MOVE_NONE;
stage += (ttMove == MOVE_NONE);
} }
/// MovePicker constructor for ProbCut: we generate captures with SEE higher
/// than or equal to the given threshold.
MovePicker::MovePicker(const Position& p, Move ttm, Value th, const CapturePieceToHistory* cph)
: pos(p), captureHistory(cph), threshold(th) {
/// MovePicker::score_captures(), MovePicker::score_noncaptures() and assert(!pos.checkers());
/// MovePicker::score_evasions() assign a numerical move ordering score
/// to each move in a move list. The moves with highest scores will be
/// picked first by get_next_move().
void MovePicker::score_captures() { stage = PROBCUT;
// Winning and equal captures in the main search are ordered by MVV/LVA. ttMove = ttm
// Suprisingly, this appears to perform slightly better than SEE based && pos.pseudo_legal(ttm)
// move ordering. The reason is probably that in a position with a winning && pos.capture(ttm)
// capture, capturing a more valuable (but sufficiently defended) piece && pos.see_ge(ttm, threshold) ? ttm : MOVE_NONE;
// first usually doesn't hurt. The opponent will have to recapture, and
// the hanging piece will still be hanging (except in the unusual cases
// where it is possible to recapture with the hanging piece). Exchanging
// big pieces before capturing a hanging piece probably helps to reduce
// the subtree size.
// In main search we want to push captures with negative SEE values to
// badCaptures[] array, but instead of doing it now we delay till when
// the move has been picked up in pick_move_from_list(), this way we save
// some SEE calls in case we get a cutoff (idea from Pablo Vazquez).
Move m;
// Use MVV/LVA ordering stage += (ttMove == MOVE_NONE);
for (MoveStack* cur = moves; cur != lastMove; cur++)
{
m = cur->move;
if (move_is_promotion(m))
cur->score = QueenValueMidgame;
else
cur->score = pos.midgame_value_of_piece_on(move_to(m))
- pos.type_of_piece_on(move_from(m));
}
} }
void MovePicker::score_noncaptures() { /// score() assigns a numerical value to each move in a list, used for sorting.
/// Captures are ordered by Most Valuable Victim (MVV), preferring captures
/// with a good history. Quiets are ordered using the histories.
template<GenType Type>
void MovePicker::score() {
Move m; static_assert(Type == CAPTURES || Type == QUIETS || Type == EVASIONS, "Wrong type");
Square from;
for (MoveStack* cur = moves; cur != lastMove; cur++) for (auto& m : *this)
{ if (Type == CAPTURES)
m = cur->move; m.value = PieceValue[MG][pos.piece_on(to_sq(m))]
from = move_from(m); + Value((*captureHistory)[pos.moved_piece(m)][to_sq(m)][type_of(pos.piece_on(to_sq(m)))]);
cur->score = H.value(pos.piece_on(from), move_to(m));
} else if (Type == QUIETS)
m.value = (*mainHistory)[pos.side_to_move()][from_to(m)]
+ (*contHistory[0])[pos.moved_piece(m)][to_sq(m)]
+ (*contHistory[1])[pos.moved_piece(m)][to_sq(m)]
+ (*contHistory[3])[pos.moved_piece(m)][to_sq(m)];
else // Type == EVASIONS
{
if (pos.capture(m))
m.value = PieceValue[MG][pos.piece_on(to_sq(m))]
- Value(type_of(pos.moved_piece(m)));
else
m.value = (*mainHistory)[pos.side_to_move()][from_to(m)] - (1 << 28);
}
} }
void MovePicker::score_evasions() { /// next_move() is the most important method of the MovePicker class. It returns
// Try good captures ordered by MVV/LVA, then non-captures if /// a new pseudo legal move every time it is called, until there are no more moves
// destination square is not under attack, ordered by history /// left. It picks the move with the biggest value from a list of generated moves
// value, and at the end bad-captures and non-captures with a /// taking care not to return the ttMove if it has already been searched.
// negative SEE. This last group is ordered by the SEE score.
Move m;
int seeScore;
// Skip if we don't have at least two moves to order Move MovePicker::next_move(bool skipQuiets) {
if (lastMove < moves + 2)
return;
for (MoveStack* cur = moves; cur != lastMove; cur++)
{
m = cur->move;
if ((seeScore = pos.see_sign(m)) < 0)
cur->score = seeScore - History::MaxValue; // Be sure we are at the bottom
else if (pos.move_is_capture(m))
cur->score = pos.midgame_value_of_piece_on(move_to(m))
- pos.type_of_piece_on(move_from(m)) + History::MaxValue;
else
cur->score = H.value(pos.piece_on(move_from(m)), move_to(m));
}
}
/// MovePicker::get_next_move() is the most important method of the MovePicker
/// class. It returns a new legal move every time it is called, until there
/// are no more moves left. It picks the move with the biggest score from a list
/// of generated moves taking care not to return the tt move if has already been
/// searched previously. Note that this function is not thread safe so should be
/// lock protected by caller when accessed through a shared MovePicker object.
Move MovePicker::get_next_move() {
Move move; Move move;
while (true) switch (stage) {
{
while (curMove == lastMove)
go_next_phase();
switch (phase) { case MAIN_SEARCH: case EVASION: case QSEARCH_WITH_CHECKS:
case QSEARCH_NO_CHECKS: case PROBCUT:
++stage;
return ttMove;
case PH_TT_MOVES: case CAPTURES_INIT:
move = (curMove++)->move; endBadCaptures = cur = moves;
if ( move != MOVE_NONE endMoves = generate<CAPTURES>(pos, cur);
&& pos.move_is_legal(move, pinned)) score<CAPTURES>();
return move; ++stage;
break; /* fallthrough */
case PH_GOOD_CAPTURES: case GOOD_CAPTURES:
move = pick_best(curMove++, lastMove).move; while (cur < endMoves)
if ( move != ttMoves[0].move {
&& move != ttMoves[1].move move = pick_best(cur++, endMoves);
&& pos.pl_move_is_legal(move, pinned)) if (move != ttMove)
{ {
// Check for a non negative SEE now if (pos.see_ge(move, Value(-55 * (cur-1)->value / 1024)))
int seeValue = pos.see_sign(move);
if (seeValue >= badCaptureThreshold)
return move; return move;
// Losing capture, move it to the tail of the array, note // Losing capture, move it to the beginning of the array
// that move has now been already checked for legality. *endBadCaptures++ = move;
(--badCaptures)->move = move;
badCaptures->score = seeValue;
} }
break;
case PH_KILLERS:
move = (curMove++)->move;
if ( move != MOVE_NONE
&& pos.move_is_legal(move, pinned)
&& move != ttMoves[0].move
&& move != ttMoves[1].move
&& !pos.move_is_capture(move))
return move;
break;
case PH_NONCAPTURES:
// Sort negative scored moves only when we get there
if (curMove == lastGoodNonCapture)
insertion_sort<MoveStack>(lastGoodNonCapture, lastMove);
move = (curMove++)->move;
if ( move != ttMoves[0].move
&& move != ttMoves[1].move
&& move != killers[0].move
&& move != killers[1].move
&& pos.pl_move_is_legal(move, pinned))
return move;
break;
case PH_BAD_CAPTURES:
move = pick_best(curMove++, lastMove).move;
return move;
case PH_EVASIONS:
case PH_QCAPTURES:
move = pick_best(curMove++, lastMove).move;
if ( move != ttMoves[0].move
&& pos.pl_move_is_legal(move, pinned))
return move;
break;
case PH_QCHECKS:
move = (curMove++)->move;
if ( move != ttMoves[0].move
&& pos.pl_move_is_legal(move, pinned))
return move;
break;
case PH_STOP:
return MOVE_NONE;
default:
assert(false);
break;
} }
++stage;
move = killers[0]; // First killer move
if ( move != MOVE_NONE
&& move != ttMove
&& pos.pseudo_legal(move)
&& !pos.capture(move))
return move;
/* fallthrough */
case KILLERS:
++stage;
move = killers[1]; // Second killer move
if ( move != MOVE_NONE
&& move != ttMove
&& pos.pseudo_legal(move)
&& !pos.capture(move))
return move;
/* fallthrough */
case COUNTERMOVE:
++stage;
move = countermove;
if ( move != MOVE_NONE
&& move != ttMove
&& move != killers[0]
&& move != killers[1]
&& pos.pseudo_legal(move)
&& !pos.capture(move))
return move;
/* fallthrough */
case QUIET_INIT:
cur = endBadCaptures;
endMoves = generate<QUIETS>(pos, cur);
score<QUIETS>();
partial_insertion_sort(cur, endMoves, -4000 * depth / ONE_PLY);
++stage;
/* fallthrough */
case QUIET:
while ( cur < endMoves
&& (!skipQuiets || cur->value >= VALUE_ZERO))
{
move = *cur++;
if ( move != ttMove
&& move != killers[0]
&& move != killers[1]
&& move != countermove)
return move;
}
++stage;
cur = moves; // Point to beginning of bad captures
/* fallthrough */
case BAD_CAPTURES:
if (cur < endBadCaptures)
return *cur++;
break;
case EVASIONS_INIT:
cur = moves;
endMoves = generate<EVASIONS>(pos, cur);
score<EVASIONS>();
++stage;
/* fallthrough */
case ALL_EVASIONS:
while (cur < endMoves)
{
move = pick_best(cur++, endMoves);
if (move != ttMove)
return move;
}
break;
case PROBCUT_INIT:
cur = moves;
endMoves = generate<CAPTURES>(pos, cur);
score<CAPTURES>();
++stage;
/* fallthrough */
case PROBCUT_CAPTURES:
while (cur < endMoves)
{
move = pick_best(cur++, endMoves);
if ( move != ttMove
&& pos.see_ge(move, threshold))
return move;
}
break;
case QCAPTURES_1_INIT: case QCAPTURES_2_INIT:
cur = moves;
endMoves = generate<CAPTURES>(pos, cur);
score<CAPTURES>();
++stage;
/* fallthrough */
case QCAPTURES_1: case QCAPTURES_2:
while (cur < endMoves)
{
move = pick_best(cur++, endMoves);
if (move != ttMove)
return move;
}
if (stage == QCAPTURES_2)
break;
cur = moves;
endMoves = generate<QUIET_CHECKS>(pos, cur);
++stage;
/* fallthrough */
case QCHECKS:
while (cur < endMoves)
{
move = cur++->move;
if (move != ttMove)
return move;
}
break;
case QSEARCH_RECAPTURES:
cur = moves;
endMoves = generate<CAPTURES>(pos, cur);
score<CAPTURES>();
++stage;
/* fallthrough */
case QRECAPTURES:
while (cur < endMoves)
{
move = pick_best(cur++, endMoves);
if (to_sq(move) == recaptureSquare)
return move;
}
break;
default:
assert(false);
} }
return MOVE_NONE;
} }
+121 -33
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,47 +18,134 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined MOVEPICK_H_INCLUDED #ifndef MOVEPICK_H_INCLUDED
#define MOVEPICK_H_INCLUDED #define MOVEPICK_H_INCLUDED
#include "history.h" #include <array>
#include "move.h" #include <limits>
#include "movegen.h"
#include "position.h" #include "position.h"
#include "types.h" #include "types.h"
struct SearchStack; /// StatBoards is a generic 2-dimensional array used to store various statistics
template<int Size1, int Size2, typename T = int16_t>
struct StatBoards : public std::array<std::array<T, Size2>, Size1> {
/// MovePicker is a class which is used to pick one legal move at a time from void fill(const T& v) {
/// the current position. It is initialized with a Position object and a few T* p = &(*this)[0][0];
/// moves we have reason to believe are good. The most important method is std::fill(p, p + sizeof(*this) / sizeof(*p), v);
/// MovePicker::get_next_move(), which returns a new legal move each time it }
/// is called, until there are no legal moves left, when MOVE_NONE is returned.
/// In order to improve the efficiency of the alpha beta algorithm, MovePicker
/// attempts to return the moves which are most likely to get a cut-off first.
class MovePicker { void update(T& entry, int bonus, const int D) {
MovePicker& operator=(const MovePicker&); // Silence a warning under MSVC assert(abs(bonus) <= D); // Ensure range is [-32 * D, 32 * D]
assert(abs(32 * D) < (std::numeric_limits<T>::max)()); // Ensure we don't overflow
public: entry += bonus * 32 - entry * abs(bonus) / D;
MovePicker(const Position&, Move, Depth, const History&, SearchStack*, Value);
MovePicker(const Position&, Move, Depth, const History&);
Move get_next_move();
private: assert(abs(entry) <= 32 * D);
void score_captures(); }
void score_noncaptures();
void score_evasions();
void go_next_phase();
const Position& pos;
const History& H;
Bitboard pinned;
MoveStack ttMoves[2], killers[2];
int badCaptureThreshold, phase;
const uint8_t* phasePtr;
MoveStack *curMove, *lastMove, *lastGoodNonCapture, *badCaptures;
MoveStack moves[MAX_MOVES];
}; };
#endif // !defined(MOVEPICK_H_INCLUDED) /// StatCubes is a generic 3-dimensional array used to store various statistics
template<int Size1, int Size2, int Size3, typename T = int16_t>
struct StatCubes : public std::array<std::array<std::array<T, Size3>, Size2>, Size1> {
void fill(const T& v) {
T* p = &(*this)[0][0][0];
std::fill(p, p + sizeof(*this) / sizeof(*p), v);
}
void update(T& entry, int bonus, const int D, const int W) {
assert(abs(bonus) <= D); // Ensure range is [-W * D, W * D]
assert(abs(W * D) < (std::numeric_limits<T>::max)()); // Ensure we don't overflow
entry += bonus * W - entry * abs(bonus) / D;
assert(abs(entry) <= W * D);
}
};
/// ButterflyBoards are 2 tables (one for each color) indexed by the move's from
/// and to squares, see chessprogramming.wikispaces.com/Butterfly+Boards
typedef StatBoards<COLOR_NB, int(SQUARE_NB) * int(SQUARE_NB)> ButterflyBoards;
/// PieceToBoards are addressed by a move's [piece][to] information
typedef StatBoards<PIECE_NB, SQUARE_NB> PieceToBoards;
/// CapturePieceToBoards are addressed by a move's [piece][to][captured piece type] information
typedef StatCubes<PIECE_NB, SQUARE_NB, PIECE_TYPE_NB> CapturePieceToBoards;
/// ButterflyHistory records how often quiet moves have been successful or
/// unsuccessful during the current search, and is used for reduction and move
/// ordering decisions. It uses ButterflyBoards as backing store.
struct ButterflyHistory : public ButterflyBoards {
void update(Color c, Move m, int bonus) {
StatBoards::update((*this)[c][from_to(m)], bonus, 324);
}
};
/// PieceToHistory is like ButterflyHistory, but is based on PieceToBoards
struct PieceToHistory : public PieceToBoards {
void update(Piece pc, Square to, int bonus) {
StatBoards::update((*this)[pc][to], bonus, 936);
}
};
/// CapturePieceToHistory is like PieceToHistory, but is based on CapturePieceToBoards
struct CapturePieceToHistory : public CapturePieceToBoards {
void update(Piece pc, Square to, PieceType captured, int bonus) {
StatCubes::update((*this)[pc][to][captured], bonus, 324, 2);
}
};
/// CounterMoveHistory stores counter moves indexed by [piece][to] of the previous
/// move, see chessprogramming.wikispaces.com/Countermove+Heuristic
typedef StatBoards<PIECE_NB, SQUARE_NB, Move> CounterMoveHistory;
/// ContinuationHistory is the history of a given pair of moves, usually the
/// current one given a previous one. History table is based on PieceToBoards
/// instead of ButterflyBoards.
typedef StatBoards<PIECE_NB, SQUARE_NB, PieceToHistory> ContinuationHistory;
/// MovePicker class is used to pick one pseudo legal move at a time from the
/// current position. The most important method is next_move(), which returns a
/// new pseudo legal move each time it is called, until there are no moves left,
/// when MOVE_NONE is returned. In order to improve the efficiency of the alpha
/// beta algorithm, MovePicker attempts to return the moves which are most likely
/// to get a cut-off first.
class MovePicker {
public:
MovePicker(const MovePicker&) = delete;
MovePicker& operator=(const MovePicker&) = delete;
MovePicker(const Position&, Move, Value, const CapturePieceToHistory*);
MovePicker(const Position&, Move, Depth, const ButterflyHistory*, const CapturePieceToHistory*, Square);
MovePicker(const Position&, Move, Depth, const ButterflyHistory*, const CapturePieceToHistory*, const PieceToHistory**, Move, Move*);
Move next_move(bool skipQuiets = false);
private:
template<GenType> void score();
ExtMove* begin() { return cur; }
ExtMove* end() { return endMoves; }
const Position& pos;
const ButterflyHistory* mainHistory;
const CapturePieceToHistory* captureHistory;
const PieceToHistory** contHistory;
Move ttMove, countermove, killers[2];
ExtMove *cur, *endMoves, *endBadCaptures;
int stage;
Square recaptureSquare;
Value threshold;
Depth depth;
ExtMove moves[MAX_MOVES];
};
#endif // #ifndef MOVEPICK_H_INCLUDED
+231 -175
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,227 +18,282 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <algorithm>
#include <cassert> #include <cassert>
#include "bitboard.h" #include "bitboard.h"
#include "bitcount.h"
#include "pawns.h" #include "pawns.h"
#include "position.h" #include "position.h"
#include "thread.h"
namespace { namespace {
#define V Value
#define S(mg, eg) make_score(mg, eg) #define S(mg, eg) make_score(mg, eg)
// Doubled pawn penalty by opposed flag and file // Isolated pawn penalty
const Score DoubledPawnPenalty[2][8] = { const Score Isolated = S(13, 18);
{ S(13, 43), S(20, 48), S(23, 48), S(23, 48),
S(23, 48), S(23, 48), S(20, 48), S(13, 43) },
{ S(13, 43), S(20, 48), S(23, 48), S(23, 48),
S(23, 48), S(23, 48), S(20, 48), S(13, 43) }};
// Isolated pawn penalty by opposed flag and file // Backward pawn penalty
const Score IsolatedPawnPenalty[2][8] = { const Score Backward = S(24, 12);
{ S(37, 45), S(54, 52), S(60, 52), S(60, 52),
S(60, 52), S(60, 52), S(54, 52), S(37, 45) },
{ S(25, 30), S(36, 35), S(40, 35), S(40, 35),
S(40, 35), S(40, 35), S(36, 35), S(25, 30) }};
// Backward pawn penalty by opposed flag and file // Connected pawn bonus by opposed, phalanx, #support and rank
const Score BackwardPawnPenalty[2][8] = { Score Connected[2][2][3][RANK_NB];
{ S(30, 42), S(43, 46), S(49, 46), S(49, 46),
S(49, 46), S(49, 46), S(43, 46), S(30, 42) },
{ S(20, 28), S(29, 31), S(33, 31), S(33, 31),
S(33, 31), S(33, 31), S(29, 31), S(20, 28) }};
// Pawn chain membership bonus by file // Doubled pawn penalty
const Score ChainBonus[8] = { const Score Doubled = S(18, 38);
S(11,-1), S(13,-1), S(13,-1), S(14,-1),
S(14,-1), S(13,-1), S(13,-1), S(11,-1) // Weakness of our pawn shelter in front of the king by [isKingFile][distance from edge][rank].
// RANK_1 = 0 is used for files where we have no pawns or our pawn is behind our king.
const Value ShelterWeakness[][int(FILE_NB) / 2][RANK_NB] = {
{ { V( 97), V(17), V( 9), V(44), V( 84), V( 87), V( 99) }, // Not On King file
{ V(106), V( 6), V(33), V(86), V( 87), V(104), V(112) },
{ V(101), V( 2), V(65), V(98), V( 58), V( 89), V(115) },
{ V( 73), V( 7), V(54), V(73), V( 84), V( 83), V(111) } },
{ { V(104), V(20), V( 6), V(27), V( 86), V( 93), V( 82) }, // On King file
{ V(123), V( 9), V(34), V(96), V(112), V( 88), V( 75) },
{ V(120), V(25), V(65), V(91), V( 66), V( 78), V(117) },
{ V( 81), V( 2), V(47), V(63), V( 94), V( 93), V(104) } }
}; };
// Candidate passed pawn bonus by rank // Danger of enemy pawns moving toward our king by [type][distance from edge][rank].
const Score CandidateBonus[8] = { // For the unopposed and unblocked cases, RANK_1 = 0 is used when opponent has
S( 0, 0), S( 6, 13), S(6,13), S(14,29), // no pawn on the given file, or their pawn is behind our king.
S(34,68), S(83,166), S(0, 0), S( 0, 0) const Value StormDanger[][4][RANK_NB] = {
{ { V( 0), V(-290), V(-274), V(57), V(41) }, // BlockedByKing
{ V( 0), V( 60), V( 144), V(39), V(13) },
{ V( 0), V( 65), V( 141), V(41), V(34) },
{ V( 0), V( 53), V( 127), V(56), V(14) } },
{ { V( 4), V( 73), V( 132), V(46), V(31) }, // Unopposed
{ V( 1), V( 64), V( 143), V(26), V(13) },
{ V( 1), V( 47), V( 110), V(44), V(24) },
{ V( 0), V( 72), V( 127), V(50), V(31) } },
{ { V( 0), V( 0), V( 79), V(23), V( 1) }, // BlockedByPawn
{ V( 0), V( 0), V( 148), V(27), V( 2) },
{ V( 0), V( 0), V( 161), V(16), V( 1) },
{ V( 0), V( 0), V( 171), V(22), V(15) } },
{ { V(22), V( 45), V( 104), V(62), V( 6) }, // Unblocked
{ V(31), V( 30), V( 99), V(39), V(19) },
{ V(23), V( 29), V( 96), V(41), V(15) },
{ V(21), V( 23), V( 116), V(41), V(15) } }
}; };
const Score PawnStructureWeight = S(233, 201); // Max bonus for king safety. Corresponds to start position with all the pawns
// in front of the king and no enemy pawn on the horizon.
const Value MaxSafetyBonus = V(258);
#undef S #undef S
#undef V
inline Score apply_weight(Score v, Score w) { template<Color Us>
return make_score((int(mg_value(v)) * mg_value(w)) / 0x100, Score evaluate(const Position& pos, Pawns::Entry* e) {
(int(eg_value(v)) * eg_value(w)) / 0x100);
const Color Them = (Us == WHITE ? BLACK : WHITE);
const Direction Up = (Us == WHITE ? NORTH : SOUTH);
const Direction Right = (Us == WHITE ? NORTH_EAST : SOUTH_WEST);
const Direction Left = (Us == WHITE ? NORTH_WEST : SOUTH_EAST);
Bitboard b, neighbours, stoppers, doubled, supported, phalanx;
Bitboard lever, leverPush;
Square s;
bool opposed, backward;
Score score = SCORE_ZERO;
const Square* pl = pos.squares<PAWN>(Us);
Bitboard ourPawns = pos.pieces( Us, PAWN);
Bitboard theirPawns = pos.pieces(Them, PAWN);
e->passedPawns[Us] = e->pawnAttacksSpan[Us] = e->weakUnopposed[Us] = 0;
e->semiopenFiles[Us] = 0xFF;
e->kingSquares[Us] = SQ_NONE;
e->pawnAttacks[Us] = shift<Right>(ourPawns) | shift<Left>(ourPawns);
e->pawnsOnSquares[Us][BLACK] = popcount(ourPawns & DarkSquares);
e->pawnsOnSquares[Us][WHITE] = pos.count<PAWN>(Us) - e->pawnsOnSquares[Us][BLACK];
// Loop through all pawns of the current color and score each pawn
while ((s = *pl++) != SQ_NONE)
{
assert(pos.piece_on(s) == make_piece(Us, PAWN));
File f = file_of(s);
e->semiopenFiles[Us] &= ~(1 << f);
e->pawnAttacksSpan[Us] |= pawn_attack_span(Us, s);
// Flag the pawn
opposed = theirPawns & forward_file_bb(Us, s);
stoppers = theirPawns & passed_pawn_mask(Us, s);
lever = theirPawns & PawnAttacks[Us][s];
leverPush = theirPawns & PawnAttacks[Us][s + Up];
doubled = ourPawns & (s - Up);
neighbours = ourPawns & adjacent_files_bb(f);
phalanx = neighbours & rank_bb(s);
supported = neighbours & rank_bb(s - Up);
// A pawn is backward when it is behind all pawns of the same color on the
// adjacent files and cannot be safely advanced.
if (!neighbours || lever || relative_rank(Us, s) >= RANK_5)
backward = false;
else
{
// Find the backmost rank with neighbours or stoppers
b = rank_bb(backmost_sq(Us, neighbours | stoppers));
// The pawn is backward when it cannot safely progress to that rank:
// either there is a stopper in the way on this rank, or there is a
// stopper on adjacent file which controls the way to that rank.
backward = (b | shift<Up>(b & adjacent_files_bb(f))) & stoppers;
assert(!(backward && (forward_ranks_bb(Them, s + Up) & neighbours)));
}
// Passed pawns will be properly scored in evaluation because we need
// full attack info to evaluate them. Include also not passed pawns
// which could become passed after one or two pawn pushes when are
// not attacked more times than defended.
if ( !(stoppers ^ lever ^ leverPush)
&& !(ourPawns & forward_file_bb(Us, s))
&& popcount(supported) >= popcount(lever)
&& popcount(phalanx) >= popcount(leverPush))
e->passedPawns[Us] |= s;
else if ( stoppers == SquareBB[s + Up]
&& relative_rank(Us, s) >= RANK_5)
{
b = shift<Up>(supported) & ~theirPawns;
while (b)
if (!more_than_one(theirPawns & PawnAttacks[Us][pop_lsb(&b)]))
e->passedPawns[Us] |= s;
}
// Score this pawn
if (supported | phalanx)
score += Connected[opposed][bool(phalanx)][popcount(supported)][relative_rank(Us, s)];
else if (!neighbours)
score -= Isolated, e->weakUnopposed[Us] += !opposed;
else if (backward)
score -= Backward, e->weakUnopposed[Us] += !opposed;
if (doubled && !supported)
score -= Doubled;
}
return score;
}
} // namespace
namespace Pawns {
/// Pawns::init() initializes some tables needed by evaluation. Instead of using
/// hard-coded tables, when makes sense, we prefer to calculate them with a formula
/// to reduce independent parameters and to allow easier tuning and better insight.
void init() {
static const int Seed[RANK_NB] = { 0, 13, 24, 18, 76, 100, 175, 330 };
for (int opposed = 0; opposed <= 1; ++opposed)
for (int phalanx = 0; phalanx <= 1; ++phalanx)
for (int support = 0; support <= 2; ++support)
for (Rank r = RANK_2; r < RANK_8; ++r)
{
int v = 17 * support;
v += (Seed[r] + (phalanx ? (Seed[r + 1] - Seed[r]) / 2 : 0)) >> opposed;
Connected[opposed][phalanx][support][r] = make_score(v, v * (r - 2) / 4);
} }
} }
/// PawnInfoTable::get_pawn_info() takes a position object as input, computes /// Pawns::probe() looks up the current position's pawns configuration in
/// a PawnInfo object, and returns a pointer to it. The result is also stored /// the pawns hash table. It returns a pointer to the Entry if the position
/// in an hash table, so we don't have to recompute everything when the same /// is found. Otherwise a new Entry is computed and stored there, so we don't
/// pawn structure occurs again. /// have to recompute all when the same pawns configuration occurs again.
PawnInfo* PawnInfoTable::get_pawn_info(const Position& pos) const { Entry* probe(const Position& pos) {
assert(pos.is_ok()); Key key = pos.pawn_key();
Entry* e = pos.this_thread()->pawnsTable[key];
Key key = pos.get_pawn_key(); if (e->key == key)
PawnInfo* pi = probe(key); return e;
// If pi->key matches the position's pawn hash key, it means that we e->key = key;
// have analysed this pawn structure before, and we can simply return e->score = evaluate<WHITE>(pos, e) - evaluate<BLACK>(pos, e);
// the information we found the last time instead of recomputing it. e->asymmetry = popcount(e->semiopenFiles[WHITE] ^ e->semiopenFiles[BLACK]);
if (pi->key == key) e->openFiles = popcount(e->semiopenFiles[WHITE] & e->semiopenFiles[BLACK]);
return pi; return e;
// Initialize PawnInfo entry
pi->key = key;
pi->passedPawns[WHITE] = pi->passedPawns[BLACK] = 0;
pi->kingSquares[WHITE] = pi->kingSquares[BLACK] = SQ_NONE;
pi->halfOpenFiles[WHITE] = pi->halfOpenFiles[BLACK] = 0xFF;
// Calculate pawn attacks
Bitboard wPawns = pos.pieces(PAWN, WHITE);
Bitboard bPawns = pos.pieces(PAWN, BLACK);
pi->pawnAttacks[WHITE] = ((wPawns << 9) & ~FileABB) | ((wPawns << 7) & ~FileHBB);
pi->pawnAttacks[BLACK] = ((bPawns >> 7) & ~FileABB) | ((bPawns >> 9) & ~FileHBB);
// Evaluate pawns for both colors and weight the result
pi->value = evaluate_pawns<WHITE>(pos, wPawns, bPawns, pi)
- evaluate_pawns<BLACK>(pos, bPawns, wPawns, pi);
pi->value = apply_weight(pi->value, PawnStructureWeight);
return pi;
} }
/// PawnInfoTable::evaluate_pawns() evaluates each pawn of the given color /// Entry::shelter_storm() calculates shelter and storm penalties for the file
/// the king is on, as well as the two closest files.
template<Color Us> template<Color Us>
Score PawnInfoTable::evaluate_pawns(const Position& pos, Bitboard ourPawns, Value Entry::shelter_storm(const Position& pos, Square ksq) {
Bitboard theirPawns, PawnInfo* pi) {
const BitCountType Max15 = CpuIs64Bit ? CNT64_MAX15 : CNT32_MAX15;
const Color Them = (Us == WHITE ? BLACK : WHITE); const Color Them = (Us == WHITE ? BLACK : WHITE);
Bitboard b; enum { BlockedByKing, Unopposed, BlockedByPawn, Unblocked };
Square s;
File f;
Rank r;
bool passed, isolated, doubled, opposed, chain, backward, candidate;
Score value = SCORE_ZERO;
const Square* ptr = pos.piece_list_begin(Us, PAWN);
// Loop through all pawns of the current color and score each pawn Bitboard b = pos.pieces(PAWN) & (forward_ranks_bb(Us, ksq) | rank_bb(ksq));
while ((s = *ptr++) != SQ_NONE) Bitboard ourPawns = b & pos.pieces(Us);
Bitboard theirPawns = b & pos.pieces(Them);
Value safety = MaxSafetyBonus;
File center = std::max(FILE_B, std::min(FILE_G, file_of(ksq)));
for (File f = File(center - 1); f <= File(center + 1); ++f)
{ {
assert(pos.piece_on(s) == make_piece(Us, PAWN)); b = ourPawns & file_bb(f);
Rank rkUs = b ? relative_rank(Us, backmost_sq(Us, b)) : RANK_1;
f = square_file(s); b = theirPawns & file_bb(f);
r = square_rank(s); Rank rkThem = b ? relative_rank(Us, frontmost_sq(Them, b)) : RANK_1;
// This file cannot be half open int d = std::min(f, ~f);
pi->halfOpenFiles[Us] &= ~(1 << f); safety -= ShelterWeakness[f == file_of(ksq)][d][rkUs]
+ StormDanger
// Our rank plus previous one. Used for chain detection [f == file_of(ksq) && rkThem == relative_rank(Us, ksq) + 1 ? BlockedByKing :
b = rank_bb(r) | rank_bb(Us == WHITE ? r - Rank(1) : r + Rank(1)); rkUs == RANK_1 ? Unopposed :
rkThem == rkUs + 1 ? BlockedByPawn : Unblocked]
// Flag the pawn as passed, isolated, doubled or member of a pawn [d][rkThem];
// chain (but not the backward one).
passed = !(theirPawns & passed_pawn_mask(Us, s));
doubled = ourPawns & squares_in_front_of(Us, s);
opposed = theirPawns & squares_in_front_of(Us, s);
isolated = !(ourPawns & neighboring_files_bb(f));
chain = ourPawns & neighboring_files_bb(f) & b;
// Test for backward pawn
backward = false;
// If the pawn is passed, isolated, or member of a pawn chain it cannot
// be backward. If there are friendly pawns behind on neighboring files
// or if can capture an enemy pawn it cannot be backward either.
if ( !(passed | isolated | chain)
&& !(ourPawns & attack_span_mask(Them, s))
&& !(pos.attacks_from<PAWN>(s, Us) & theirPawns))
{
// We now know that there are no friendly pawns beside or behind this
// pawn on neighboring files. We now check whether the pawn is
// backward by looking in the forward direction on the neighboring
// files, and seeing whether we meet a friendly or an enemy pawn first.
b = pos.attacks_from<PAWN>(s, Us);
// Note that we are sure to find something because pawn is not passed
// nor isolated, so loop is potentially infinite, but it isn't.
while (!(b & (ourPawns | theirPawns)))
Us == WHITE ? b <<= 8 : b >>= 8;
// The friendly pawn needs to be at least two ranks closer than the
// enemy pawn in order to help the potentially backward pawn advance.
backward = (b | (Us == WHITE ? b << 8 : b >> 8)) & theirPawns;
}
assert(opposed | passed | (attack_span_mask(Us, s) & theirPawns));
// A not passed pawn is a candidate to become passed if it is free to
// advance and if the number of friendly pawns beside or behind this
// pawn on neighboring files is higher or equal than the number of
// enemy pawns in the forward direction on the neighboring files.
candidate = !(opposed | passed | backward | isolated)
&& (b = attack_span_mask(Them, s + pawn_push(Us)) & ourPawns) != EmptyBoardBB
&& count_1s<Max15>(b) >= count_1s<Max15>(attack_span_mask(Us, s) & theirPawns);
// Passed pawns will be properly scored in evaluation because we need
// full attack info to evaluate passed pawns. Only the frontmost passed
// pawn on each file is considered a true passed pawn.
if (passed && !doubled)
set_bit(&(pi->passedPawns[Us]), s);
// Score this pawn
if (isolated)
value -= IsolatedPawnPenalty[opposed][f];
if (doubled)
value -= DoubledPawnPenalty[opposed][f];
if (backward)
value -= BackwardPawnPenalty[opposed][f];
if (chain)
value += ChainBonus[f];
if (candidate)
value += CandidateBonus[relative_rank(Us, s)];
} }
return value;
return safety;
} }
/// PawnInfo::updateShelter() calculates and caches king shelter. It is called /// Entry::do_king_safety() calculates a bonus for king safety. It is called only
/// only when king square changes, about 20% of total king_shelter() calls. /// when king square changes, which is about 20% of total king_safety() calls.
template<Color Us> template<Color Us>
Score PawnInfo::updateShelter(const Position& pos, Square ksq) { Score Entry::do_king_safety(const Position& pos, Square ksq) {
const int Shift = (Us == WHITE ? 8 : -8);
Bitboard pawns;
int r, shelter = 0;
if (relative_rank(Us, ksq) <= RANK_4)
{
pawns = pos.pieces(PAWN, Us) & this_and_neighboring_files_bb(ksq);
r = ksq & (7 << 3);
for (int i = 0; i < 3; i++)
{
r += Shift;
shelter += BitCount8Bit[(pawns >> r) & 0xFF] * (64 >> i);
}
}
kingSquares[Us] = ksq; kingSquares[Us] = ksq;
kingShelters[Us] = make_score(shelter, 0); castlingRights[Us] = pos.can_castle(Us);
return kingShelters[Us]; int minKingPawnDistance = 0;
Bitboard pawns = pos.pieces(Us, PAWN);
if (pawns)
while (!(DistanceRingBB[ksq][minKingPawnDistance++] & pawns)) {}
Value bonus = shelter_storm<Us>(pos, ksq);
// If we can castle use the bonus after the castling if it is bigger
if (pos.can_castle(MakeCastling<Us, KING_SIDE>::right))
bonus = std::max(bonus, shelter_storm<Us>(pos, relative_square(Us, SQ_G1)));
if (pos.can_castle(MakeCastling<Us, QUEEN_SIDE>::right))
bonus = std::max(bonus, shelter_storm<Us>(pos, relative_square(Us, SQ_C1)));
return make_score(bonus, -16 * minKingPawnDistance);
} }
// Explicit template instantiation // Explicit template instantiation
template Score PawnInfo::updateShelter<WHITE>(const Position& pos, Square ksq); template Score Entry::do_king_safety<WHITE>(const Position& pos, Square ksq);
template Score PawnInfo::updateShelter<BLACK>(const Position& pos, Square ksq); template Score Entry::do_king_safety<BLACK>(const Position& pos, Square ksq);
} // namespace Pawns
+52 -69
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,91 +18,73 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(PAWNS_H_INCLUDED) #ifndef PAWNS_H_INCLUDED
#define PAWNS_H_INCLUDED #define PAWNS_H_INCLUDED
#include "misc.h"
#include "position.h" #include "position.h"
#include "tt.h"
#include "types.h" #include "types.h"
const int PawnTableSize = 16384; namespace Pawns {
/// PawnInfo is a class which contains various information about a pawn /// Pawns::Entry contains various information about a pawn structure. A lookup
/// structure. Currently, it only includes a middle game and an end game /// to the pawn hash table (performed by calling the probe function) returns a
/// pawn structure evaluation, and a bitboard of passed pawns. We may want /// pointer to an Entry object.
/// to add further information in the future. A lookup to the pawn hash table
/// (performed by calling the get_pawn_info method in a PawnInfoTable object)
/// returns a pointer to a PawnInfo object.
class PawnInfo { struct Entry {
friend class PawnInfoTable; Score pawns_score() const { return score; }
Bitboard pawn_attacks(Color c) const { return pawnAttacks[c]; }
Bitboard passed_pawns(Color c) const { return passedPawns[c]; }
Bitboard pawn_attacks_span(Color c) const { return pawnAttacksSpan[c]; }
int weak_unopposed(Color c) const { return weakUnopposed[c]; }
int pawn_asymmetry() const { return asymmetry; }
int open_files() const { return openFiles; }
public: int semiopen_file(Color c, File f) const {
Score pawns_value() const; return semiopenFiles[c] & (1 << f);
Bitboard pawn_attacks(Color c) const; }
Bitboard passed_pawns(Color c) const;
int file_is_half_open(Color c, File f) const; int semiopen_side(Color c, File f, bool leftSide) const {
int has_open_file_to_left(Color c, File f) const; return semiopenFiles[c] & (leftSide ? (1 << f) - 1 : ~((1 << (f + 1)) - 1));
int has_open_file_to_right(Color c, File f) const; }
int pawns_on_same_color_squares(Color c, Square s) const {
return pawnsOnSquares[c][bool(DarkSquares & s)];
}
template<Color Us> template<Color Us>
Score king_shelter(const Position& pos, Square ksq); Score king_safety(const Position& pos, Square ksq) {
return kingSquares[Us] == ksq && castlingRights[Us] == pos.can_castle(Us)
? kingSafety[Us] : (kingSafety[Us] = do_king_safety<Us>(pos, ksq));
}
private:
template<Color Us> template<Color Us>
Score updateShelter(const Position& pos, Square ksq); Score do_king_safety(const Position& pos, Square ksq);
template<Color Us>
Value shelter_storm(const Position& pos, Square ksq);
Key key; Key key;
Bitboard passedPawns[2]; Score score;
Bitboard pawnAttacks[2]; Bitboard passedPawns[COLOR_NB];
Square kingSquares[2]; Bitboard pawnAttacks[COLOR_NB];
Score value; Bitboard pawnAttacksSpan[COLOR_NB];
int halfOpenFiles[2]; Square kingSquares[COLOR_NB];
Score kingShelters[2]; Score kingSafety[COLOR_NB];
int weakUnopposed[COLOR_NB];
int castlingRights[COLOR_NB];
int semiopenFiles[COLOR_NB];
int pawnsOnSquares[COLOR_NB][COLOR_NB]; // [color][light/dark squares]
int asymmetry;
int openFiles;
}; };
typedef HashTable<Entry, 16384> Table;
/// The PawnInfoTable class represents a pawn hash table. The most important void init();
/// method is get_pawn_info, which returns a pointer to a PawnInfo object. Entry* probe(const Position& pos);
class PawnInfoTable : public SimpleHash<PawnInfo, PawnTableSize> { } // namespace Pawns
public:
PawnInfo* get_pawn_info(const Position& pos) const;
private: #endif // #ifndef PAWNS_H_INCLUDED
template<Color Us>
static Score evaluate_pawns(const Position& pos, Bitboard ourPawns, Bitboard theirPawns, PawnInfo* pi);
};
inline Score PawnInfo::pawns_value() const {
return value;
}
inline Bitboard PawnInfo::pawn_attacks(Color c) const {
return pawnAttacks[c];
}
inline Bitboard PawnInfo::passed_pawns(Color c) const {
return passedPawns[c];
}
inline int PawnInfo::file_is_half_open(Color c, File f) const {
return halfOpenFiles[c] & (1 << int(f));
}
inline int PawnInfo::has_open_file_to_left(Color c, File f) const {
return halfOpenFiles[c] & ((1 << int(f)) - 1);
}
inline int PawnInfo::has_open_file_to_right(Color c, File f) const {
return halfOpenFiles[c] & ~((1 << int(f+1)) - 1);
}
template<Color Us>
inline Score PawnInfo::king_shelter(const Position& pos, Square ksq) {
return kingSquares[Us] == ksq ? kingShelters[Us] : updateShelter<Us>(pos, ksq);
}
#endif // !defined(PAWNS_H_INCLUDED)
+833 -1704
View File
File diff suppressed because it is too large Load Diff
+259 -386
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,538 +18,410 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(POSITION_H_INCLUDED) #ifndef POSITION_H_INCLUDED
#define POSITION_H_INCLUDED #define POSITION_H_INCLUDED
#include <cassert>
#include <deque>
#include <memory> // For std::unique_ptr
#include <string>
#include "bitboard.h" #include "bitboard.h"
#include "move.h"
#include "types.h" #include "types.h"
/// Maximum number of plies per game (220 should be enough, because the
/// maximum search depth is 100, and during position setup we reset the
/// move counter for every non-reversible move).
const int MaxGameLength = 220;
class Position; /// StateInfo struct stores information needed to restore a Position object to
/// its previous state when we retract a move. Whenever a move is made on the
/// struct checkInfo is initialized at c'tor time and keeps /// board (by calling Position::do_move), a StateInfo object must be passed.
/// info used to detect if a move gives check.
struct CheckInfo {
explicit CheckInfo(const Position&);
Bitboard dcCandidates;
Bitboard checkSq[8];
Square ksq;
};
/// Castle rights, encoded as bit fields
enum CastleRights {
CASTLES_NONE = 0,
WHITE_OO = 1,
BLACK_OO = 2,
WHITE_OOO = 4,
BLACK_OOO = 8,
ALL_CASTLES = 15
};
/// Game phase
enum Phase {
PHASE_ENDGAME = 0,
PHASE_MIDGAME = 128
};
/// The StateInfo struct stores information we need to restore a Position
/// object to its previous state when we retract a move. Whenever a move
/// is made on the board (by calling Position::do_move), an StateInfo object
/// must be passed as a parameter.
struct StateInfo { struct StateInfo {
Key pawnKey, materialKey;
int castleRights, rule50, gamePly, pliesFromNull;
Square epSquare;
Score value;
Value npMaterial[2];
PieceType capturedType; // Copied when making a move
Key key; Key pawnKey;
Bitboard checkersBB; Key materialKey;
Value nonPawnMaterial[COLOR_NB];
int castlingRights;
int rule50;
int pliesFromNull;
Score psq;
Square epSquare;
// Not copied when making a move (will be recomputed anyhow)
Key key;
Bitboard checkersBB;
Piece capturedPiece;
StateInfo* previous; StateInfo* previous;
Bitboard blockersForKing[COLOR_NB];
Bitboard pinnersForKing[COLOR_NB];
Bitboard checkSquares[PIECE_TYPE_NB];
}; };
/// A list to keep track of the position states along the setup moves (from the
/// start position to the position just before the search starts). Needed by
/// 'draw by repetition' detection. Use a std::deque because pointers to
/// elements are not invalidated upon list resizing.
typedef std::unique_ptr<std::deque<StateInfo>> StateListPtr;
/// The position data structure. A position consists of the following data:
/// /// Position class stores information regarding the board representation as
/// * For each piece type, a bitboard representing the squares occupied /// pieces, side to move, hash keys, castling info, etc. Important methods are
/// by pieces of that type. /// do_move() and undo_move(), used by the search to update node info when
/// * For each color, a bitboard representing the squares occupied by /// traversing the search tree.
/// pieces of that color. class Thread;
/// * A bitboard of all occupied squares.
/// * A bitboard of all checking pieces.
/// * A 64-entry array of pieces, indexed by the squares of the board.
/// * The current side to move.
/// * Information about the castling rights for both sides.
/// * The initial files of the kings and both pairs of rooks. This is
/// used to implement the Chess960 castling rules.
/// * The en passant square (which is SQ_NONE if no en passant capture is
/// possible).
/// * The squares of the kings for both sides.
/// * Hash keys for the position itself, the current pawn structure, and
/// the current material situation.
/// * Hash keys for all previous positions in the game for detecting
/// repetition draws.
/// * A counter for detecting 50 move rule draws.
class Position { class Position {
Position(); // No default or copy c'tor allowed
Position(const Position& pos);
public: public:
enum GamePhase { static void init();
MidGame,
EndGame
};
// Constructors Position() = default;
Position(const Position& pos, int threadID); Position(const Position&) = delete;
Position(const std::string& fen, bool isChess960, int threadID); Position& operator=(const Position&) = delete;
// Text input/output // FEN string input/output
void from_fen(const std::string& fen, bool isChess960); Position& set(const std::string& fenStr, bool isChess960, StateInfo* si, Thread* th);
const std::string to_fen() const; Position& set(const std::string& code, Color c, StateInfo* si);
void print(Move m = MOVE_NONE) const; const std::string fen() const;
// Copying // Position representation
void flip(); Bitboard pieces() const;
// The piece on a given square
Piece piece_on(Square s) const;
PieceType type_of_piece_on(Square s) const;
Color color_of_piece_on(Square s) const;
bool square_is_empty(Square s) const;
bool square_is_occupied(Square s) const;
Value midgame_value_of_piece_on(Square s) const;
Value endgame_value_of_piece_on(Square s) const;
// Side to move
Color side_to_move() const;
// Bitboard representation of the position
Bitboard empty_squares() const;
Bitboard occupied_squares() const;
Bitboard pieces_of_color(Color c) const;
Bitboard pieces(PieceType pt) const; Bitboard pieces(PieceType pt) const;
Bitboard pieces(PieceType pt, Color c) const;
Bitboard pieces(PieceType pt1, PieceType pt2) const; Bitboard pieces(PieceType pt1, PieceType pt2) const;
Bitboard pieces(PieceType pt1, PieceType pt2, Color c) const; Bitboard pieces(Color c) const;
Bitboard pieces(Color c, PieceType pt) const;
// Number of pieces of each color and type Bitboard pieces(Color c, PieceType pt1, PieceType pt2) const;
int piece_count(Color c, PieceType pt) const; Piece piece_on(Square s) const;
// The en passant square
Square ep_square() const; Square ep_square() const;
bool empty(Square s) const;
template<PieceType Pt> int count(Color c) const;
template<PieceType Pt> int count() const;
template<PieceType Pt> const Square* squares(Color c) const;
template<PieceType Pt> Square square(Color c) const;
// Current king position for each color // Castling
Square king_square(Color c) const; int can_castle(Color c) const;
int can_castle(CastlingRight cr) const;
bool castling_impeded(CastlingRight cr) const;
Square castling_rook_square(CastlingRight cr) const;
// Castling rights // Checking
bool can_castle_kingside(Color c) const;
bool can_castle_queenside(Color c) const;
bool can_castle(Color c) const;
Square initial_kr_square(Color c) const;
Square initial_qr_square(Color c) const;
// Bitboards for pinned pieces and discovered check candidates
Bitboard discovered_check_candidates(Color c) const;
Bitboard pinned_pieces(Color c) const;
// Checking pieces and under check information
Bitboard checkers() const; Bitboard checkers() const;
bool in_check() const; Bitboard discovered_check_candidates() const;
Bitboard pinned_pieces(Color c) const;
Bitboard check_squares(PieceType pt) const;
// Piece lists // Attacks to/from a given square
Square piece_list(Color c, PieceType pt, int index) const;
const Square* piece_list_begin(Color c, PieceType pt) const;
// Information about attacks to or from a given square
Bitboard attackers_to(Square s) const; Bitboard attackers_to(Square s) const;
Bitboard attacks_from(Piece p, Square s) const; Bitboard attackers_to(Square s, Bitboard occupied) const;
static Bitboard attacks_from(Piece p, Square s, Bitboard occ); Bitboard attacks_from(PieceType pt, Square s) const;
template<PieceType> Bitboard attacks_from(Square s) const; template<PieceType> Bitboard attacks_from(Square s) const;
template<PieceType> Bitboard attacks_from(Square s, Color c) const; template<PieceType> Bitboard attacks_from(Square s, Color c) const;
Bitboard slider_blockers(Bitboard sliders, Square s, Bitboard& pinners) const;
// Properties of moves // Properties of moves
bool pl_move_is_legal(Move m, Bitboard pinned) const; bool legal(Move m) const;
bool pl_move_is_evasion(Move m, Bitboard pinned) const; bool pseudo_legal(const Move m) const;
bool move_is_legal(const Move m) const; bool capture(Move m) const;
bool move_is_legal(const Move m, Bitboard pinned) const; bool capture_or_promotion(Move m) const;
bool move_gives_check(Move m) const; bool gives_check(Move m) const;
bool move_gives_check(Move m, const CheckInfo& ci) const; bool advanced_pawn_push(Move m) const;
bool move_is_capture(Move m) const; Piece moved_piece(Move m) const;
bool move_is_capture_or_promotion(Move m) const; Piece captured_piece() const;
bool move_is_passed_pawn_push(Move m) const;
bool move_attacks_square(Move m, Square s) const;
// Piece captured with previous moves // Piece specific
PieceType captured_piece_type() const; bool pawn_passed(Color c, Square s) const;
bool opposite_bishops() const;
// Information about pawns
bool pawn_is_passed(Color c, Square s) const;
// Weak squares
bool square_is_weak(Square s, Color c) const;
// Doing and undoing moves // Doing and undoing moves
void do_setup_move(Move m); void do_move(Move m, StateInfo& newSt);
void do_move(Move m, StateInfo& st); void do_move(Move m, StateInfo& newSt, bool givesCheck);
void do_move(Move m, StateInfo& st, const CheckInfo& ci, bool moveIsCheck);
void undo_move(Move m); void undo_move(Move m);
void do_null_move(StateInfo& st); void do_null_move(StateInfo& newSt);
void undo_null_move(); void undo_null_move();
// Static exchange evaluation // Static Exchange Evaluation
int see(Square from, Square to) const; bool see_ge(Move m, Value threshold = VALUE_ZERO) const;
int see(Move m) const;
int see_sign(Move m) const;
// Accessing hash keys // Accessing hash keys
Key get_key() const; Key key() const;
Key get_exclusion_key() const; Key key_after(Move m) const;
Key get_pawn_key() const; Key material_key() const;
Key get_material_key() const; Key pawn_key() const;
// Incremental evaluation
Score value() const;
Value non_pawn_material(Color c) const;
static Score pst_delta(Piece piece, Square from, Square to);
// Game termination checks
bool is_mate() const;
bool is_draw() const;
// Number of plies from starting position
int startpos_ply_counter() const;
// Other properties of the position // Other properties of the position
bool opposite_colored_bishops() const; Color side_to_move() const;
bool has_pawn_on_7th(Color c) const; int game_ply() const;
bool is_chess960() const; bool is_chess960() const;
Thread* this_thread() const;
// Current thread ID searching on the position bool is_draw(int ply) const;
int thread() const; int rule50_count() const;
Score psq_score() const;
int64_t nodes_searched() const; Value non_pawn_material(Color c) const;
void set_nodes_searched(int64_t n); Value non_pawn_material() const;
// Position consistency check, for debugging // Position consistency check, for debugging
bool is_ok(int* failedStep = NULL) const; bool pos_is_ok() const;
void flip();
// Static member functions
static void init_zobrist();
static void init_piece_square_tables();
private: private:
// Initialization helpers (used while setting up a position)
void set_castling_right(Color c, Square rfrom);
void set_state(StateInfo* si) const;
void set_check_info(StateInfo* si) const;
// Initialization helper functions (used while setting up a position) // Other helpers
void clear(); void put_piece(Piece pc, Square s);
void detach(); void remove_piece(Piece pc, Square s);
void put_piece(Piece p, Square s); void move_piece(Piece pc, Square from, Square to);
void do_allow_oo(Color c); template<bool Do>
void do_allow_ooo(Color c); void do_castling(Color us, Square from, Square& to, Square& rfrom, Square& rto);
bool set_castling_rights(char token);
// Helper functions for doing and undoing moves // Data members
void do_capture_move(Key& key, PieceType capture, Color them, Square to, bool ep); Piece board[SQUARE_NB];
void do_castle_move(Move m); Bitboard byTypeBB[PIECE_TYPE_NB];
void undo_castle_move(Move m); Bitboard byColorBB[COLOR_NB];
void find_checkers(); int pieceCount[PIECE_NB];
Square pieceList[PIECE_NB][16];
template<bool FindPinned> int index[SQUARE_NB];
Bitboard hidden_checkers(Color c) const; int castlingRightsMask[SQUARE_NB];
Square castlingRookSquare[CASTLING_RIGHT_NB];
// Computing hash keys from scratch (for initialization and debugging) Bitboard castlingPath[CASTLING_RIGHT_NB];
Key compute_key() const; int gamePly;
Key compute_pawn_key() const;
Key compute_material_key() const;
// Computing incremental evaluation scores and material counts
static Score pst(Color c, PieceType pt, Square s);
Score compute_value() const;
Value compute_non_pawn_material(Color c) const;
// Board
Piece board[64];
// Bitboards
Bitboard byTypeBB[8], byColorBB[2];
// Piece counts
int pieceCount[2][8]; // [color][pieceType]
// Piece lists
Square pieceList[2][8][16]; // [color][pieceType][index]
int index[64]; // [square]
// Other info
Color sideToMove; Color sideToMove;
Key history[MaxGameLength]; Thread* thisThread;
int castleRightsMask[64];
StateInfo startState;
File initialKFile, initialKRFile, initialQRFile;
bool chess960;
int startPosPlyCounter;
int threadID;
int64_t nodes;
StateInfo* st; StateInfo* st;
bool chess960;
// Static variables
static Key zobrist[2][8][64];
static Key zobEp[64];
static Key zobCastle[16];
static Key zobSideToMove;
static Score PieceSquareTable[16][64];
static Key zobExclusion;
static const Value seeValues[8];
static const Value PieceValueMidgame[17];
static const Value PieceValueEndgame[17];
}; };
inline int64_t Position::nodes_searched() const { extern std::ostream& operator<<(std::ostream& os, const Position& pos);
return nodes;
inline Color Position::side_to_move() const {
return sideToMove;
} }
inline void Position::set_nodes_searched(int64_t n) { inline bool Position::empty(Square s) const {
nodes = n; return board[s] == NO_PIECE;
} }
inline Piece Position::piece_on(Square s) const { inline Piece Position::piece_on(Square s) const {
return board[s]; return board[s];
} }
inline Color Position::color_of_piece_on(Square s) const { inline Piece Position::moved_piece(Move m) const {
return color_of_piece(piece_on(s)); return board[from_sq(m)];
} }
inline PieceType Position::type_of_piece_on(Square s) const { inline Bitboard Position::pieces() const {
return type_of_piece(piece_on(s)); return byTypeBB[ALL_PIECES];
}
inline bool Position::square_is_empty(Square s) const {
return piece_on(s) == PIECE_NONE;
}
inline bool Position::square_is_occupied(Square s) const {
return !square_is_empty(s);
}
inline Value Position::midgame_value_of_piece_on(Square s) const {
return PieceValueMidgame[piece_on(s)];
}
inline Value Position::endgame_value_of_piece_on(Square s) const {
return PieceValueEndgame[piece_on(s)];
}
inline Color Position::side_to_move() const {
return sideToMove;
}
inline Bitboard Position::occupied_squares() const {
return byTypeBB[0];
}
inline Bitboard Position::empty_squares() const {
return ~occupied_squares();
}
inline Bitboard Position::pieces_of_color(Color c) const {
return byColorBB[c];
} }
inline Bitboard Position::pieces(PieceType pt) const { inline Bitboard Position::pieces(PieceType pt) const {
return byTypeBB[pt]; return byTypeBB[pt];
} }
inline Bitboard Position::pieces(PieceType pt, Color c) const {
return byTypeBB[pt] & byColorBB[c];
}
inline Bitboard Position::pieces(PieceType pt1, PieceType pt2) const { inline Bitboard Position::pieces(PieceType pt1, PieceType pt2) const {
return byTypeBB[pt1] | byTypeBB[pt2]; return byTypeBB[pt1] | byTypeBB[pt2];
} }
inline Bitboard Position::pieces(PieceType pt1, PieceType pt2, Color c) const { inline Bitboard Position::pieces(Color c) const {
return (byTypeBB[pt1] | byTypeBB[pt2]) & byColorBB[c]; return byColorBB[c];
} }
inline int Position::piece_count(Color c, PieceType pt) const { inline Bitboard Position::pieces(Color c, PieceType pt) const {
return pieceCount[c][pt]; return byColorBB[c] & byTypeBB[pt];
} }
inline Square Position::piece_list(Color c, PieceType pt, int idx) const { inline Bitboard Position::pieces(Color c, PieceType pt1, PieceType pt2) const {
return pieceList[c][pt][idx]; return byColorBB[c] & (byTypeBB[pt1] | byTypeBB[pt2]);
} }
inline const Square* Position::piece_list_begin(Color c, PieceType pt) const { template<PieceType Pt> inline int Position::count(Color c) const {
return pieceList[c][pt]; return pieceCount[make_piece(c, Pt)];
}
template<PieceType Pt> inline int Position::count() const {
return pieceCount[make_piece(WHITE, Pt)] + pieceCount[make_piece(BLACK, Pt)];
}
template<PieceType Pt> inline const Square* Position::squares(Color c) const {
return pieceList[make_piece(c, Pt)];
}
template<PieceType Pt> inline Square Position::square(Color c) const {
assert(pieceCount[make_piece(c, Pt)] == 1);
return pieceList[make_piece(c, Pt)][0];
} }
inline Square Position::ep_square() const { inline Square Position::ep_square() const {
return st->epSquare; return st->epSquare;
} }
inline Square Position::king_square(Color c) const { inline int Position::can_castle(CastlingRight cr) const {
return pieceList[c][KING][0]; return st->castlingRights & cr;
} }
inline bool Position::can_castle_kingside(Color side) const { inline int Position::can_castle(Color c) const {
return st->castleRights & (1+int(side)); return st->castlingRights & ((WHITE_OO | WHITE_OOO) << (2 * c));
} }
inline bool Position::can_castle_queenside(Color side) const { inline bool Position::castling_impeded(CastlingRight cr) const {
return st->castleRights & (4+4*int(side)); return byTypeBB[ALL_PIECES] & castlingPath[cr];
} }
inline bool Position::can_castle(Color side) const { inline Square Position::castling_rook_square(CastlingRight cr) const {
return can_castle_kingside(side) || can_castle_queenside(side); return castlingRookSquare[cr];
} }
inline Square Position::initial_kr_square(Color c) const { template<PieceType Pt>
return relative_square(c, make_square(initialKRFile, RANK_1)); inline Bitboard Position::attacks_from(Square s) const {
} assert(Pt != PAWN);
return Pt == BISHOP || Pt == ROOK ? attacks_bb<Pt>(s, byTypeBB[ALL_PIECES])
inline Square Position::initial_qr_square(Color c) const { : Pt == QUEEN ? attacks_from<ROOK>(s) | attacks_from<BISHOP>(s)
return relative_square(c, make_square(initialQRFile, RANK_1)); : PseudoAttacks[Pt][s];
} }
template<> template<>
inline Bitboard Position::attacks_from<PAWN>(Square s, Color c) const { inline Bitboard Position::attacks_from<PAWN>(Square s, Color c) const {
return StepAttacksBB[make_piece(c, PAWN)][s]; return PawnAttacks[c][s];
} }
template<PieceType Piece> // Knight and King and white pawns inline Bitboard Position::attacks_from(PieceType pt, Square s) const {
inline Bitboard Position::attacks_from(Square s) const { return attacks_bb(pt, s, byTypeBB[ALL_PIECES]);
return StepAttacksBB[Piece][s];
} }
template<> inline Bitboard Position::attackers_to(Square s) const {
inline Bitboard Position::attacks_from<BISHOP>(Square s) const { return attackers_to(s, byTypeBB[ALL_PIECES]);
return bishop_attacks_bb(s, occupied_squares());
}
template<>
inline Bitboard Position::attacks_from<ROOK>(Square s) const {
return rook_attacks_bb(s, occupied_squares());
}
template<>
inline Bitboard Position::attacks_from<QUEEN>(Square s) const {
return attacks_from<ROOK>(s) | attacks_from<BISHOP>(s);
} }
inline Bitboard Position::checkers() const { inline Bitboard Position::checkers() const {
return st->checkersBB; return st->checkersBB;
} }
inline bool Position::in_check() const { inline Bitboard Position::discovered_check_candidates() const {
return st->checkersBB != EmptyBoardBB; return st->blockersForKing[~sideToMove] & pieces(sideToMove);
} }
inline bool Position::pawn_is_passed(Color c, Square s) const { inline Bitboard Position::pinned_pieces(Color c) const {
return !(pieces(PAWN, opposite_color(c)) & passed_pawn_mask(c, s)); return st->blockersForKing[c] & pieces(c);
} }
inline bool Position::square_is_weak(Square s, Color c) const { inline Bitboard Position::check_squares(PieceType pt) const {
return !(pieces(PAWN, opposite_color(c)) & attack_span_mask(c, s)); return st->checkSquares[pt];
} }
inline Key Position::get_key() const { inline bool Position::pawn_passed(Color c, Square s) const {
return !(pieces(~c, PAWN) & passed_pawn_mask(c, s));
}
inline bool Position::advanced_pawn_push(Move m) const {
return type_of(moved_piece(m)) == PAWN
&& relative_rank(sideToMove, from_sq(m)) > RANK_4;
}
inline Key Position::key() const {
return st->key; return st->key;
} }
inline Key Position::get_exclusion_key() const { inline Key Position::pawn_key() const {
return st->key ^ zobExclusion;
}
inline Key Position::get_pawn_key() const {
return st->pawnKey; return st->pawnKey;
} }
inline Key Position::get_material_key() const { inline Key Position::material_key() const {
return st->materialKey; return st->materialKey;
} }
inline Score Position::pst(Color c, PieceType pt, Square s) { inline Score Position::psq_score() const {
return PieceSquareTable[make_piece(c, pt)][s]; return st->psq;
}
inline Score Position::pst_delta(Piece piece, Square from, Square to) {
return PieceSquareTable[piece][to] - PieceSquareTable[piece][from];
}
inline Score Position::value() const {
return st->value;
} }
inline Value Position::non_pawn_material(Color c) const { inline Value Position::non_pawn_material(Color c) const {
return st->npMaterial[c]; return st->nonPawnMaterial[c];
} }
inline bool Position::move_is_passed_pawn_push(Move m) const { inline Value Position::non_pawn_material() const {
return st->nonPawnMaterial[WHITE] + st->nonPawnMaterial[BLACK];
Color c = side_to_move();
return piece_on(move_from(m)) == make_piece(c, PAWN)
&& pawn_is_passed(c, move_to(m));
} }
inline int Position::startpos_ply_counter() const { inline int Position::game_ply() const {
return startPosPlyCounter; return gamePly;
} }
inline bool Position::opposite_colored_bishops() const { inline int Position::rule50_count() const {
return st->rule50;
return piece_count(WHITE, BISHOP) == 1 && piece_count(BLACK, BISHOP) == 1
&& opposite_color_squares(piece_list(WHITE, BISHOP, 0), piece_list(BLACK, BISHOP, 0));
} }
inline bool Position::has_pawn_on_7th(Color c) const { inline bool Position::opposite_bishops() const {
return pieces(PAWN, c) & rank_bb(relative_rank(c, RANK_7)); return pieceCount[W_BISHOP] == 1
&& pieceCount[B_BISHOP] == 1
&& opposite_colors(square<BISHOP>(WHITE), square<BISHOP>(BLACK));
} }
inline bool Position::is_chess960() const { inline bool Position::is_chess960() const {
return chess960; return chess960;
} }
inline bool Position::move_is_capture(Move m) const { inline bool Position::capture_or_promotion(Move m) const {
assert(is_ok(m));
// Move must not be MOVE_NONE ! return type_of(m) != NORMAL ? type_of(m) != CASTLING : !empty(to_sq(m));
return (m & (3 << 15)) ? !move_is_castle(m) : !square_is_empty(move_to(m));
} }
inline bool Position::move_is_capture_or_promotion(Move m) const { inline bool Position::capture(Move m) const {
assert(is_ok(m));
// Move must not be MOVE_NONE ! // Castling is encoded as "king captures rook"
return (m & (0x1F << 12)) ? !move_is_castle(m) : !square_is_empty(move_to(m)); return (!empty(to_sq(m)) && type_of(m) != CASTLING) || type_of(m) == ENPASSANT;
} }
inline PieceType Position::captured_piece_type() const { inline Piece Position::captured_piece() const {
return st->capturedType; return st->capturedPiece;
} }
inline int Position::thread() const { inline Thread* Position::this_thread() const {
return threadID; return thisThread;
} }
inline void Position::do_allow_oo(Color c) { inline void Position::put_piece(Piece pc, Square s) {
st->castleRights |= (1 + int(c));
board[s] = pc;
byTypeBB[ALL_PIECES] |= s;
byTypeBB[type_of(pc)] |= s;
byColorBB[color_of(pc)] |= s;
index[s] = pieceCount[pc]++;
pieceList[pc][index[s]] = s;
pieceCount[make_piece(color_of(pc), ALL_PIECES)]++;
} }
inline void Position::do_allow_ooo(Color c) { inline void Position::remove_piece(Piece pc, Square s) {
st->castleRights |= (4 + 4*int(c));
// WARNING: This is not a reversible operation. If we remove a piece in
// do_move() and then replace it in undo_move() we will put it at the end of
// the list and not in its original place, it means index[] and pieceList[]
// are not invariant to a do_move() + undo_move() sequence.
byTypeBB[ALL_PIECES] ^= s;
byTypeBB[type_of(pc)] ^= s;
byColorBB[color_of(pc)] ^= s;
/* board[s] = NO_PIECE; Not needed, overwritten by the capturing one */
Square lastSquare = pieceList[pc][--pieceCount[pc]];
index[lastSquare] = index[s];
pieceList[pc][index[lastSquare]] = lastSquare;
pieceList[pc][pieceCount[pc]] = SQ_NONE;
pieceCount[make_piece(color_of(pc), ALL_PIECES)]--;
} }
#endif // !defined(POSITION_H_INCLUDED) inline void Position::move_piece(Piece pc, Square from, Square to) {
// index[from] is not updated and becomes stale. This works as long as index[]
// is accessed just by known occupied squares.
Bitboard from_to_bb = SquareBB[from] ^ SquareBB[to];
byTypeBB[ALL_PIECES] ^= from_to_bb;
byTypeBB[type_of(pc)] ^= from_to_bb;
byColorBB[color_of(pc)] ^= from_to_bb;
board[from] = NO_PIECE;
board[to] = pc;
index[to] = index[from];
pieceList[pc][index[to]] = to;
}
inline void Position::do_move(Move m, StateInfo& newSt) {
do_move(m, newSt, gives_check(m));
}
#endif // #ifndef POSITION_H_INCLUDED
+126
View File
@@ -0,0 +1,126 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <algorithm>
#include "types.h"
Value PieceValue[PHASE_NB][PIECE_NB] = {
{ VALUE_ZERO, PawnValueMg, KnightValueMg, BishopValueMg, RookValueMg, QueenValueMg },
{ VALUE_ZERO, PawnValueEg, KnightValueEg, BishopValueEg, RookValueEg, QueenValueEg }
};
namespace PSQT {
#define S(mg, eg) make_score(mg, eg)
// Bonus[PieceType][Square / 2] contains Piece-Square scores. For each piece
// type on a given square a (middlegame, endgame) score pair is assigned. Table
// is defined for files A..D and white side: it is symmetric for black side and
// second half of the files.
const Score Bonus[][RANK_NB][int(FILE_NB) / 2] = {
{ },
{ // Pawn
{ S( 0, 0), S( 0, 0), S( 0, 0), S( 0, 0) },
{ S(-11, 7), S( 6,-4), S( 7, 8), S( 3,-2) },
{ S(-18,-4), S( -2,-5), S( 19, 5), S(24, 4) },
{ S(-17, 3), S( -9, 3), S( 20,-8), S(35,-3) },
{ S( -6, 8), S( 5, 9), S( 3, 7), S(21,-6) },
{ S( -6, 8), S( -8,-5), S( -6, 2), S(-2, 4) },
{ S( -4, 3), S( 20,-9), S( -8, 1), S(-4,18) }
},
{ // Knight
{ S(-161,-105), S(-96,-82), S(-80,-46), S(-73,-14) },
{ S( -83, -69), S(-43,-54), S(-21,-17), S(-10, 9) },
{ S( -71, -50), S(-22,-39), S( 0, -7), S( 9, 28) },
{ S( -25, -41), S( 18,-25), S( 43, 6), S( 47, 38) },
{ S( -26, -46), S( 16,-25), S( 38, 3), S( 50, 40) },
{ S( -11, -54), S( 37,-38), S( 56, -7), S( 65, 27) },
{ S( -63, -65), S(-19,-50), S( 5,-24), S( 14, 13) },
{ S(-195,-109), S(-67,-89), S(-42,-50), S(-29,-13) }
},
{ // Bishop
{ S(-44,-58), S(-13,-31), S(-25,-37), S(-34,-19) },
{ S(-20,-34), S( 20, -9), S( 12,-14), S( 1, 4) },
{ S( -9,-23), S( 27, 0), S( 21, -3), S( 11, 16) },
{ S(-11,-26), S( 28, -3), S( 21, -5), S( 10, 16) },
{ S(-11,-26), S( 27, -4), S( 16, -7), S( 9, 14) },
{ S(-17,-24), S( 16, -2), S( 12, 0), S( 2, 13) },
{ S(-23,-34), S( 17,-10), S( 6,-12), S( -2, 6) },
{ S(-35,-55), S(-11,-32), S(-19,-36), S(-29,-17) }
},
{ // Rook
{ S(-25, 0), S(-16, 0), S(-16, 0), S(-9, 0) },
{ S(-21, 0), S( -8, 0), S( -3, 0), S( 0, 0) },
{ S(-21, 0), S( -9, 0), S( -4, 0), S( 2, 0) },
{ S(-22, 0), S( -6, 0), S( -1, 0), S( 2, 0) },
{ S(-22, 0), S( -7, 0), S( 0, 0), S( 1, 0) },
{ S(-21, 0), S( -7, 0), S( 0, 0), S( 2, 0) },
{ S(-12, 0), S( 4, 0), S( 8, 0), S(12, 0) },
{ S(-23, 0), S(-15, 0), S(-11, 0), S(-5, 0) }
},
{ // Queen
{ S( 0,-71), S(-4,-56), S(-3,-42), S(-1,-29) },
{ S(-4,-56), S( 6,-30), S( 9,-21), S( 8, -5) },
{ S(-2,-39), S( 6,-17), S( 9, -8), S( 9, 5) },
{ S(-1,-29), S( 8, -5), S(10, 9), S( 7, 19) },
{ S(-3,-27), S( 9, -5), S( 8, 10), S( 7, 21) },
{ S(-2,-40), S( 6,-16), S( 8,-10), S(10, 3) },
{ S(-2,-55), S( 7,-30), S( 7,-21), S( 6, -6) },
{ S(-1,-74), S(-4,-55), S(-1,-43), S( 0,-30) }
},
{ // King
{ S(267, 0), S(320, 48), S(270, 75), S(195, 84) },
{ S(264, 43), S(304, 92), S(238,143), S(180,132) },
{ S(200, 83), S(245,138), S(176,167), S(110,165) },
{ S(177,106), S(185,169), S(148,169), S(110,179) },
{ S(149,108), S(177,163), S(115,200), S( 66,203) },
{ S(118, 95), S(159,155), S( 84,176), S( 41,174) },
{ S( 87, 50), S(128, 99), S( 63,122), S( 20,139) },
{ S( 63, 9), S( 88, 55), S( 47, 80), S( 0, 90) }
}
};
#undef S
Score psq[PIECE_NB][SQUARE_NB];
// init() initializes piece-square tables: the white halves of the tables are
// copied from Bonus[] adding the piece value, then the black halves of the
// tables are initialized by flipping and changing the sign of the white scores.
void init() {
for (Piece pc = W_PAWN; pc <= W_KING; ++pc)
{
PieceValue[MG][~pc] = PieceValue[MG][pc];
PieceValue[EG][~pc] = PieceValue[EG][pc];
Score v = make_score(PieceValue[MG][pc], PieceValue[EG][pc]);
for (Square s = SQ_A1; s <= SQ_H8; ++s)
{
File f = std::min(file_of(s), ~file_of(s));
psq[ pc][ s] = v + Bonus[pc][rank_of(s)][f];
psq[~pc][~s] = -psq[pc][s];
}
}
}
} // namespace PSQT
-185
View File
@@ -1,185 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(PSQTAB_H_INCLUDED)
#define PSQTAB_H_INCLUDED
#include "types.h"
namespace {
////
//// Constants modified by Joona Kiiski
////
const Value MP = PawnValueMidgame;
const Value MK = KnightValueMidgame;
const Value MB = BishopValueMidgame;
const Value MR = RookValueMidgame;
const Value MQ = QueenValueMidgame;
const Value EP = PawnValueEndgame;
const Value EK = KnightValueEndgame;
const Value EB = BishopValueEndgame;
const Value ER = RookValueEndgame;
const Value EQ = QueenValueEndgame;
const int MgPST[][64] = {
{ },
{// Pawn
// A B C D E F G H
0, 0, 0, 0, 0, 0, 0, 0,
MP-28, MP-6, MP+ 4, MP+14, MP+14, MP+ 4, MP-6, MP-28,
MP-28, MP-6, MP+ 9, MP+36, MP+36, MP+ 9, MP-6, MP-28,
MP-28, MP-6, MP+17, MP+58, MP+58, MP+17, MP-6, MP-28,
MP-28, MP-6, MP+17, MP+36, MP+36, MP+17, MP-6, MP-28,
MP-28, MP-6, MP+ 9, MP+14, MP+14, MP+ 9, MP-6, MP-28,
MP-28, MP-6, MP+ 4, MP+14, MP+14, MP+ 4, MP-6, MP-28,
0, 0, 0, 0, 0, 0, 0, 0
},
{// Knight
// A B C D E F G H
MK-135, MK-107, MK-80, MK-67, MK-67, MK-80, MK-107, MK-135,
MK- 93, MK- 67, MK-39, MK-25, MK-25, MK-39, MK- 67, MK- 93,
MK- 53, MK- 25, MK+ 1, MK+13, MK+13, MK+ 1, MK- 25, MK- 53,
MK- 25, MK+ 1, MK+27, MK+41, MK+41, MK+27, MK+ 1, MK- 25,
MK- 11, MK+ 13, MK+41, MK+55, MK+55, MK+41, MK+ 13, MK- 11,
MK- 11, MK+ 13, MK+41, MK+55, MK+55, MK+41, MK+ 13, MK- 11,
MK- 53, MK- 25, MK+ 1, MK+13, MK+13, MK+ 1, MK- 25, MK- 53,
MK-193, MK- 67, MK-39, MK-25, MK-25, MK-39, MK- 67, MK-193
},
{// Bishop
// A B C D E F G H
MB-40, MB-40, MB-35, MB-30, MB-30, MB-35, MB-40, MB-40,
MB-17, MB+ 0, MB- 4, MB+ 0, MB+ 0, MB- 4, MB+ 0, MB-17,
MB-13, MB- 4, MB+ 8, MB+ 4, MB+ 4, MB+ 8, MB- 4, MB-13,
MB- 8, MB+ 0, MB+ 4, MB+17, MB+17, MB+ 4, MB+ 0, MB- 8,
MB- 8, MB+ 0, MB+ 4, MB+17, MB+17, MB+ 4, MB+ 0, MB- 8,
MB-13, MB- 4, MB+ 8, MB+ 4, MB+ 4, MB+ 8, MB- 4, MB-13,
MB-17, MB+ 0, MB- 4, MB+ 0, MB+ 0, MB- 4, MB+ 0, MB-17,
MB-17, MB-17, MB-13, MB- 8, MB- 8, MB-13, MB-17, MB-17
},
{// Rook
// A B C D E F G H
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12,
MR-12, MR-7, MR-2, MR+2, MR+2, MR-2, MR-7, MR-12
},
{// Queen
// A B C D E F G H
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8,
MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8, MQ+8
},
{// King
//A B C D E F G H
287, 311, 262, 214, 214, 262, 311, 287,
262, 287, 238, 190, 190, 238, 287, 262,
214, 238, 190, 142, 142, 190, 238, 214,
190, 214, 167, 119, 119, 167, 214, 190,
167, 190, 142, 94, 94, 142, 190, 167,
142, 167, 119, 69, 69, 119, 167, 142,
119, 142, 94, 46, 46, 94, 142, 119,
94, 119, 69, 21, 21, 69, 119, 94
}
};
const int EgPST[][64] = {
{ },
{// Pawn
// A B C D E F G H
0, 0, 0, 0, 0, 0, 0, 0,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8, EP-8,
0, 0, 0, 0, 0, 0, 0, 0
},
{// Knight
// A B C D E F G H
EK-104, EK-79, EK-55, EK-42, EK-42, EK-55, EK-79, EK-104,
EK- 79, EK-55, EK-30, EK-17, EK-17, EK-30, EK-55, EK- 79,
EK- 55, EK-30, EK- 6, EK+ 5, EK+ 5, EK- 6, EK-30, EK- 55,
EK- 42, EK-17, EK+ 5, EK+18, EK+18, EK+ 5, EK-17, EK- 42,
EK- 42, EK-17, EK+ 5, EK+18, EK+18, EK+ 5, EK-17, EK- 42,
EK- 55, EK-30, EK- 6, EK+ 5, EK+ 5, EK- 6, EK-30, EK- 55,
EK- 79, EK-55, EK-30, EK-17, EK-17, EK-30, EK-55, EK- 79,
EK-104, EK-79, EK-55, EK-42, EK-42, EK-55, EK-79, EK-104
},
{// Bishop
// A B C D E F G H
EB-59, EB-42, EB-35, EB-26, EB-26, EB-35, EB-42, EB-59,
EB-42, EB-26, EB-18, EB-11, EB-11, EB-18, EB-26, EB-42,
EB-35, EB-18, EB-11, EB- 4, EB- 4, EB-11, EB-18, EB-35,
EB-26, EB-11, EB- 4, EB+ 4, EB+ 4, EB- 4, EB-11, EB-26,
EB-26, EB-11, EB- 4, EB+ 4, EB+ 4, EB- 4, EB-11, EB-26,
EB-35, EB-18, EB-11, EB- 4, EB- 4, EB-11, EB-18, EB-35,
EB-42, EB-26, EB-18, EB-11, EB-11, EB-18, EB-26, EB-42,
EB-59, EB-42, EB-35, EB-26, EB-26, EB-35, EB-42, EB-59
},
{// Rook
// A B C D E F G H
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3,
ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3, ER+3
},
{// Queen
// A B C D E F G H
EQ-80, EQ-54, EQ-42, EQ-30, EQ-30, EQ-42, EQ-54, EQ-80,
EQ-54, EQ-30, EQ-18, EQ- 6, EQ- 6, EQ-18, EQ-30, EQ-54,
EQ-42, EQ-18, EQ- 6, EQ+ 6, EQ+ 6, EQ- 6, EQ-18, EQ-42,
EQ-30, EQ- 6, EQ+ 6, EQ+18, EQ+18, EQ+ 6, EQ- 6, EQ-30,
EQ-30, EQ- 6, EQ+ 6, EQ+18, EQ+18, EQ+ 6, EQ- 6, EQ-30,
EQ-42, EQ-18, EQ- 6, EQ+ 6, EQ+ 6, EQ- 6, EQ-18, EQ-42,
EQ-54, EQ-30, EQ-18, EQ- 6, EQ- 6, EQ-18, EQ-30, EQ-54,
EQ-80, EQ-54, EQ-42, EQ-30, EQ-30, EQ-42, EQ-54, EQ-80
},
{// King
//A B C D E F G H
18, 77, 105, 135, 135, 105, 77, 18,
77, 135, 165, 193, 193, 165, 135, 77,
105, 165, 193, 222, 222, 193, 165, 105,
135, 193, 222, 251, 251, 222, 193, 135,
135, 193, 222, 251, 251, 222, 193, 135,
105, 165, 193, 222, 222, 193, 165, 105,
77, 135, 165, 193, 193, 165, 135, 77,
18, 77, 105, 135, 135, 105, 77, 18
}
};
} // namespace
#endif // !defined(PSQTAB_H_INCLUDED)
-77
View File
@@ -1,77 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
This file is based on original code by Heinz van Saanen and is
available under the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
** George Marsaglia invented the RNG-Kiss-family in the early 90's.
** This is a specific version that Heinz van Saanen derived and
** tested from some public domain code by Bob Jenkins:
**
** Quite platform independent
** Passes ALL dieharder tests! Here *nix sys-rand() e.g. fails miserably:-)
** ~12 times faster than my *nix sys-rand()
** ~4 times faster than SSE2-version of Mersenne twister
** Average cycle length: ~2^126
** 64 bit seed
** Return doubles with a full 53 bit mantissa
** Thread safe
*/
#if !defined(RKISS_H_INCLUDED)
#define RKISS_H_INCLUDED
#include "types.h"
class RKISS {
// Keep variables always together
struct S { uint64_t a, b, c, d; } s;
uint64_t rotate(uint64_t x, uint64_t k) const {
return (x << k) | (x >> (64 - k));
}
// Return 64 bit unsigned integer in between [0, 2^64 - 1]
uint64_t rand64() {
const uint64_t
e = s.a - rotate(s.b, 7);
s.a = s.b ^ rotate(s.c, 13);
s.b = s.c + rotate(s.d, 37);
s.c = s.d + e;
return s.d = e + s.a;
}
// Init seed and scramble a few rounds
void raninit() {
s.a = 0xf1ea5eed;
s.b = s.c = s.d = 0xd4e12c77;
for (int i = 0; i < 73; i++)
rand64();
}
public:
RKISS() { raninit(); }
template<typename T> T rand() { return T(rand64()); }
};
#endif // !defined(RKISS_H_INCLUDED)
+1301 -1847
View File
File diff suppressed because it is too large Load Diff
+67 -34
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,57 +18,89 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(SEARCH_H_INCLUDED) #ifndef SEARCH_H_INCLUDED
#define SEARCH_H_INCLUDED #define SEARCH_H_INCLUDED
#include <cstring> #include <vector>
#include "move.h" #include "misc.h"
#include "movepick.h"
#include "types.h" #include "types.h"
class Position; class Position;
struct SplitPoint;
/// The SearchStack struct keeps track of the information we need to remember namespace Search {
/// from nodes shallower and deeper in the tree during the search. Each
/// search thread has its own array of SearchStack objects, indexed by the
/// current ply.
struct SearchStack { /// Threshold used for countermoves based pruning
const int CounterMovePruneThreshold = 0;
/// Stack struct keeps track of the information we need to remember from nodes
/// shallower and deeper in the tree during the search. Each search thread has
/// its own array of Stack objects, indexed by the current ply.
struct Stack {
Move* pv;
PieceToHistory* contHistory;
int ply; int ply;
Move currentMove; Move currentMove;
Move mateKiller;
Move excludedMove; Move excludedMove;
Move bestMove;
Move killers[2]; Move killers[2];
Depth reduction; Value staticEval;
Value eval; int statScore;
Value evalMargin; int moveCount;
bool skipNullMove;
SplitPoint* sp;
}; };
/// The SearchLimits struct stores information sent by GUI about available time /// RootMove struct is used for moves at the root of the tree. For each root move
/// to search the current move, maximum depth/time, if we are in analysis mode /// we store a score and a PV (really a refutation in the case of moves which
/// or if we have to ponder while is our opponent's side to move. /// fail low). Score is normally set at -VALUE_INFINITE for all non-pv moves.
struct SearchLimits { struct RootMove {
SearchLimits() { memset(this, 0, sizeof(SearchLimits)); } explicit RootMove(Move m) : pv(1, m) {}
bool extract_ponder_from_tt(Position& pos);
bool operator==(const Move& m) const { return pv[0] == m; }
bool operator<(const RootMove& m) const { // Sort in descending order
return m.score != score ? m.score < score
: m.previousScore < previousScore;
}
SearchLimits(int t, int i, int mtg, int mt, int md, int mn, bool inf, bool pon) Value score = -VALUE_INFINITE;
: time(t), increment(i), movesToGo(mtg), maxTime(mt), maxDepth(md), Value previousScore = -VALUE_INFINITE;
maxNodes(mn), infinite(inf), ponder(pon) {} int selDepth = 0;
std::vector<Move> pv;
bool useTimeManagement() const { return !(maxTime | maxDepth | maxNodes | int(infinite)); }
int time, increment, movesToGo, maxTime, maxDepth, maxNodes;
bool infinite, ponder;
}; };
extern void init_search(); typedef std::vector<RootMove> RootMoves;
extern int64_t perft(Position& pos, Depth depth);
extern bool think(Position& pos, const SearchLimits& limits, Move searchMoves[]);
#endif // !defined(SEARCH_H_INCLUDED)
/// LimitsType struct stores information sent by GUI about available time to
/// search the current move, maximum depth/time, or if we are in analysis mode.
struct LimitsType {
LimitsType() { // Init explicitly due to broken value-initialization of non POD in MSVC
nodes = time[WHITE] = time[BLACK] = inc[WHITE] = inc[BLACK] =
npmsec = movestogo = depth = movetime = mate = perft = infinite = 0;
}
bool use_time_management() const {
return !(mate | movetime | depth | nodes | perft | infinite);
}
std::vector<Move> searchmoves;
int time[COLOR_NB], inc[COLOR_NB], npmsec, movestogo, depth,
movetime, mate, perft, infinite;
int64_t nodes;
TimePoint startTime;
};
extern LimitsType Limits;
void init();
void clear();
} // namespace Search
#endif // #ifndef SEARCH_H_INCLUDED
File diff suppressed because it is too large Load Diff
+79
View File
@@ -0,0 +1,79 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (c) 2013 Ronald de Man
Copyright (C) 2016-2018 Marco Costalba, Lucas Braesch
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef TBPROBE_H
#define TBPROBE_H
#include <ostream>
#include "../search.h"
namespace Tablebases {
enum WDLScore {
WDLLoss = -2, // Loss
WDLBlessedLoss = -1, // Loss, but draw under 50-move rule
WDLDraw = 0, // Draw
WDLCursedWin = 1, // Win, but draw under 50-move rule
WDLWin = 2, // Win
WDLScoreNone = -1000
};
// Possible states after a probing operation
enum ProbeState {
FAIL = 0, // Probe failed (missing file table)
OK = 1, // Probe succesful
CHANGE_STM = -1, // DTZ should check the other side
ZEROING_BEST_MOVE = 2 // Best move zeroes DTZ (capture or pawn move)
};
extern int MaxCardinality;
void init(const std::string& paths);
WDLScore probe_wdl(Position& pos, ProbeState* result);
int probe_dtz(Position& pos, ProbeState* result);
bool root_probe(Position& pos, Search::RootMoves& rootMoves, Value& score);
bool root_probe_wdl(Position& pos, Search::RootMoves& rootMoves, Value& score);
void filter_root_moves(Position& pos, Search::RootMoves& rootMoves);
inline std::ostream& operator<<(std::ostream& os, const WDLScore v) {
os << (v == WDLLoss ? "Loss" :
v == WDLBlessedLoss ? "Blessed loss" :
v == WDLDraw ? "Draw" :
v == WDLCursedWin ? "Cursed win" :
v == WDLWin ? "Win" : "None");
return os;
}
inline std::ostream& operator<<(std::ostream& os, const ProbeState v) {
os << (v == FAIL ? "Failed" :
v == OK ? "Success" :
v == CHANGE_STM ? "Probed opponent side" :
v == ZEROING_BEST_MOVE ? "Best move zeroes DTZ" : "None");
return os;
}
}
#endif
+142 -288
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,329 +18,182 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <iostream> #include <algorithm> // For std::count
#include <cassert>
#include "movegen.h"
#include "search.h"
#include "thread.h" #include "thread.h"
#include "ucioption.h" #include "uci.h"
#include "syzygy/tbprobe.h"
ThreadsManager Threads; // Global object definition ThreadPool Threads; // Global object
namespace { extern "C" {
// start_routine() is the C function which is called when a new thread
// is launched. It simply calls idle_loop() with the supplied threadID.
// There are two versions of this function; one for POSIX threads and
// one for Windows threads.
#if defined(_MSC_VER)
DWORD WINAPI start_routine(LPVOID threadID) {
Threads.idle_loop(*(int*)threadID, NULL);
return 0;
}
#else
void* start_routine(void* threadID) {
Threads.idle_loop(*(int*)threadID, NULL);
return NULL;
}
#endif
} }
// wake_up() wakes up the thread, normally at the beginning of the search or, /// Thread constructor launches the thread and waits until it goes to sleep
// if "sleeping threads" is used, when there is some work to do. /// in idle_loop(). Note that 'searching' and 'exit' should be alredy set.
void Thread::wake_up() { Thread::Thread(size_t n) : idx(n), stdThread(&Thread::idle_loop, this) {
lock_grab(&sleepLock); wait_for_search_finished();
cond_signal(&sleepCond);
lock_release(&sleepLock);
} }
// cutoff_occurred() checks whether a beta cutoff has occurred in /// Thread destructor wakes up the thread in idle_loop() and waits
// the thread's currently active split point, or in some ancestor of /// for its termination. Thread should be already waiting.
// the current split point.
bool Thread::cutoff_occurred() const { Thread::~Thread() {
for (SplitPoint* sp = splitPoint; sp; sp = sp->parent) assert(!searching);
if (sp->is_betaCutoff)
return true; exit = true;
return false; start_searching();
stdThread.join();
} }
// is_available_to() checks whether the thread is available to help the thread with /// Thread::clear() reset histories, usually before a new game
// threadID "master" at a split point. An obvious requirement is that thread must be
// idle. With more than two threads, this is not by itself sufficient: If the thread
// is the master of some active split point, it is only available as a slave to the
// threads which are busy searching the split point at the top of "slave"'s split
// point stack (the "helpful master concept" in YBWC terminology).
bool Thread::is_available_to(int master) const { void Thread::clear() {
if (state != AVAILABLE) counterMoves.fill(MOVE_NONE);
return false; mainHistory.fill(0);
captureHistory.fill(0);
// Make a local copy to be sure doesn't become zero under our feet while for (auto& to : contHistory)
// testing next condition and so leading to an out of bound access. for (auto& h : to)
int localActiveSplitPoints = activeSplitPoints; h.fill(0);
// No active split points means that the thread is available as a slave for any contHistory[NO_PIECE][0].fill(Search::CounterMovePruneThreshold - 1);
// other thread otherwise apply the "helpful master" concept if possible. }
if ( !localActiveSplitPoints
|| splitPoints[localActiveSplitPoints - 1].is_slave[master])
return true;
return false; /// Thread::start_searching() wakes up the thread that will start the search
void Thread::start_searching() {
std::lock_guard<Mutex> lk(mutex);
searching = true;
cv.notify_one(); // Wake up the thread in idle_loop()
} }
// read_uci_options() updates number of active threads and other internal /// Thread::wait_for_search_finished() blocks on the condition variable
// parameters according to the UCI options values. It is called before /// until the thread has finished searching.
// to start a new search.
void ThreadsManager::read_uci_options() { void Thread::wait_for_search_finished() {
maxThreadsPerSplitPoint = Options["Maximum Number of Threads per Split Point"].value<int>(); std::unique_lock<Mutex> lk(mutex);
minimumSplitDepth = Options["Minimum Split Depth"].value<int>() * ONE_PLY; cv.wait(lk, [&]{ return !searching; });
useSleepingThreads = Options["Use Sleeping Threads"].value<bool>();
activeThreads = Options["Threads"].value<int>();
} }
// init() is called during startup. Initializes locks and condition variables /// Thread::idle_loop() is where the thread is parked, blocked on the
// and launches all threads sending them immediately to sleep. /// condition variable, when it has no work to do.
void ThreadsManager::init() { void Thread::idle_loop() {
int threadID[MAX_THREADS]; // If OS already scheduled us on a different group than 0 then don't overwrite
// the choice, eventually we are one of many one-threaded processes running on
// some Windows NUMA hardware, for instance in fishtest. To make it simple,
// just check if running threads are below a threshold, in this case all this
// NUMA machinery is not needed.
if (Options["Threads"] >= 8)
WinProcGroup::bindThisThread(idx);
// This flag is needed to properly end the threads when program exits while (true)
allThreadsShouldExit = false;
// Threads will sent to sleep as soon as created, only main thread is kept alive
activeThreads = 1;
threads[0].state = Thread::SEARCHING;
// Allocate pawn and material hash tables for main thread
init_hash_tables();
lock_init(&mpLock);
// Initialize thread and split point locks
for (int i = 0; i < MAX_THREADS; i++)
{ {
lock_init(&threads[i].sleepLock); std::unique_lock<Mutex> lk(mutex);
cond_init(&threads[i].sleepCond); searching = false;
cv.notify_one(); // Wake up anyone waiting for search finished
cv.wait(lk, [&]{ return searching; });
for (int j = 0; j < MAX_ACTIVE_SPLIT_POINTS; j++) if (exit)
lock_init(&(threads[i].splitPoints[j].lock)); return;
lk.unlock();
search();
}
}
/// ThreadPool::set() creates/destroys threads to match the requested number.
/// Created and launced threads wil go immediately to sleep in idle_loop.
/// Upon resizing, threads are recreated to allow for binding if necessary.
void ThreadPool::set(size_t requested) {
if (size() > 0) { // destroy any existing thread(s)
main()->wait_for_search_finished();
while (size() > 0)
delete back(), pop_back();
} }
// Create and startup all the threads but the main that is already running if (requested > 0) { // create new thread(s)
for (int i = 1; i < MAX_THREADS; i++) push_back(new MainThread(0));
while (size() < requested)
push_back(new Thread(size()));
clear();
}
}
/// ThreadPool::clear() sets threadPool data to initial values.
void ThreadPool::clear() {
for (Thread* th : *this)
th->clear();
main()->callsCnt = 0;
main()->previousScore = VALUE_INFINITE;
main()->previousTimeReduction = 1;
}
/// ThreadPool::start_thinking() wakes up main thread waiting in idle_loop() and
/// returns immediately. Main thread will wake up other threads and start the search.
void ThreadPool::start_thinking(Position& pos, StateListPtr& states,
const Search::LimitsType& limits, bool ponderMode) {
main()->wait_for_search_finished();
stopOnPonderhit = stop = false;
ponder = ponderMode;
Search::Limits = limits;
Search::RootMoves rootMoves;
for (const auto& m : MoveList<LEGAL>(pos))
if ( limits.searchmoves.empty()
|| std::count(limits.searchmoves.begin(), limits.searchmoves.end(), m))
rootMoves.emplace_back(m);
if (!rootMoves.empty())
Tablebases::filter_root_moves(pos, rootMoves);
// After ownership transfer 'states' becomes empty, so if we stop the search
// and call 'go' again without setting a new position states.get() == NULL.
assert(states.get() || setupStates.get());
if (states.get())
setupStates = std::move(states); // Ownership transfer, states is now empty
// We use Position::set() to set root position across threads. But there are
// some StateInfo fields (previous, pliesFromNull, capturedPiece) that cannot
// be deduced from a fen string, so set() clears them and to not lose the info
// we need to backup and later restore setupStates->back(). Note that setupStates
// is shared by threads but is accessed in read-only mode.
StateInfo tmp = setupStates->back();
for (Thread* th : *this)
{ {
threads[i].state = Thread::INITIALIZING; th->nodes = th->tbHits = th->nmp_ply = th->nmp_odd = 0;
threadID[i] = i; th->rootDepth = th->completedDepth = DEPTH_ZERO;
th->rootMoves = rootMoves;
#if defined(_MSC_VER) th->rootPos.set(pos.fen(), pos.is_chess960(), &setupStates->back(), th);
bool ok = (CreateThread(NULL, 0, start_routine, (LPVOID)&threadID[i], 0, NULL) != NULL);
#else
pthread_t pthreadID;
bool ok = (pthread_create(&pthreadID, NULL, start_routine, (void*)&threadID[i]) == 0);
pthread_detach(pthreadID);
#endif
if (!ok)
{
std::cout << "Failed to create thread number " << i << std::endl;
::exit(EXIT_FAILURE);
}
// Wait until the thread has finished launching and is gone to sleep
while (threads[i].state == Thread::INITIALIZING) {}
}
}
// exit() is called to cleanly exit the threads when the program finishes
void ThreadsManager::exit() {
// Force the woken up threads to exit idle_loop() and hence terminate
allThreadsShouldExit = true;
for (int i = 0; i < MAX_THREADS; i++)
{
// Wake up all the threads and waits for termination
if (i != 0)
{
threads[i].wake_up();
while (threads[i].state != Thread::TERMINATED) {}
}
// Now we can safely destroy the locks and wait conditions
lock_destroy(&threads[i].sleepLock);
cond_destroy(&threads[i].sleepCond);
for (int j = 0; j < MAX_ACTIVE_SPLIT_POINTS; j++)
lock_destroy(&(threads[i].splitPoints[j].lock));
} }
lock_destroy(&mpLock); setupStates->back() = tmp;
main()->start_searching();
} }
// init_hash_tables() dynamically allocates pawn and material hash tables
// according to the number of active threads. This avoids preallocating
// memory for all possible threads if only few are used as, for instance,
// on mobile devices where memory is scarce and allocating for MAX_THREADS
// threads could even result in a crash.
void ThreadsManager::init_hash_tables() {
for (int i = 0; i < activeThreads; i++)
{
threads[i].pawnTable.init();
threads[i].materialTable.init();
}
}
// available_slave_exists() tries to find an idle thread which is available as
// a slave for the thread with threadID "master".
bool ThreadsManager::available_slave_exists(int master) const {
assert(master >= 0 && master < activeThreads);
for (int i = 0; i < activeThreads; i++)
if (i != master && threads[i].is_available_to(master))
return true;
return false;
}
// split() does the actual work of distributing the work at a node between
// several available threads. If it does not succeed in splitting the
// node (because no idle threads are available, or because we have no unused
// split point objects), the function immediately returns. If splitting is
// possible, a SplitPoint object is initialized with all the data that must be
// copied to the helper threads and we tell our helper threads that they have
// been assigned work. This will cause them to instantly leave their idle loops and
// call search().When all threads have returned from search() then split() returns.
template <bool Fake>
void ThreadsManager::split(Position& pos, SearchStack* ss, Value* alpha, const Value beta,
Value* bestValue, Depth depth, Move threatMove,
int moveCount, MovePicker* mp, bool pvNode) {
assert(pos.is_ok());
assert(*bestValue >= -VALUE_INFINITE);
assert(*bestValue <= *alpha);
assert(*alpha < beta);
assert(beta <= VALUE_INFINITE);
assert(depth > DEPTH_ZERO);
assert(pos.thread() >= 0 && pos.thread() < activeThreads);
assert(activeThreads > 1);
int i, master = pos.thread();
Thread& masterThread = threads[master];
lock_grab(&mpLock);
// If no other thread is available to help us, or if we have too many
// active split points, don't split.
if ( !available_slave_exists(master)
|| masterThread.activeSplitPoints >= MAX_ACTIVE_SPLIT_POINTS)
{
lock_release(&mpLock);
return;
}
// Pick the next available split point object from the split point stack
SplitPoint& splitPoint = masterThread.splitPoints[masterThread.activeSplitPoints++];
// Initialize the split point object
splitPoint.parent = masterThread.splitPoint;
splitPoint.master = master;
splitPoint.is_betaCutoff = false;
splitPoint.depth = depth;
splitPoint.threatMove = threatMove;
splitPoint.alpha = *alpha;
splitPoint.beta = beta;
splitPoint.pvNode = pvNode;
splitPoint.bestValue = *bestValue;
splitPoint.mp = mp;
splitPoint.moveCount = moveCount;
splitPoint.pos = &pos;
splitPoint.nodes = 0;
splitPoint.ss = ss;
for (i = 0; i < activeThreads; i++)
splitPoint.is_slave[i] = false;
masterThread.splitPoint = &splitPoint;
// If we are here it means we are not available
assert(masterThread.state != Thread::AVAILABLE);
int workersCnt = 1; // At least the master is included
// Allocate available threads setting state to THREAD_BOOKED
for (i = 0; !Fake && i < activeThreads && workersCnt < maxThreadsPerSplitPoint; i++)
if (i != master && threads[i].is_available_to(master))
{
threads[i].state = Thread::BOOKED;
threads[i].splitPoint = &splitPoint;
splitPoint.is_slave[i] = true;
workersCnt++;
}
assert(Fake || workersCnt > 1);
// We can release the lock because slave threads are already booked and master is not available
lock_release(&mpLock);
// Tell the threads that they have work to do. This will make them leave
// their idle loop.
for (i = 0; i < activeThreads; i++)
if (i == master || splitPoint.is_slave[i])
{
assert(i == master || threads[i].state == Thread::BOOKED);
threads[i].state = Thread::WORKISWAITING; // This makes the slave to exit from idle_loop()
if (useSleepingThreads && i != master)
threads[i].wake_up();
}
// Everything is set up. The master thread enters the idle loop, from
// which it will instantly launch a search, because its state is
// THREAD_WORKISWAITING. We send the split point as a second parameter to the
// idle loop, which means that the main thread will return from the idle
// loop when all threads have finished their work at this split point.
idle_loop(master, &splitPoint);
// We have returned from the idle loop, which means that all threads are
// finished. Update alpha and bestValue, and return.
lock_grab(&mpLock);
*alpha = splitPoint.alpha;
*bestValue = splitPoint.bestValue;
masterThread.activeSplitPoints--;
masterThread.splitPoint = splitPoint.parent;
pos.set_nodes_searched(pos.nodes_searched() + splitPoint.nodes);
lock_release(&mpLock);
}
// Explicit template instantiations
template void ThreadsManager::split<false>(Position&, SearchStack*, Value*, const Value, Value*, Depth, Move, int, MovePicker*, bool);
template void ThreadsManager::split<true>(Position&, SearchStack*, Value*, const Value, Value*, Depth, Move, int, MovePicker*, bool);
+87 -96
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,116 +18,106 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(THREAD_H_INCLUDED) #ifndef THREAD_H_INCLUDED
#define THREAD_H_INCLUDED #define THREAD_H_INCLUDED
#include <cstring> #include <atomic>
#include <condition_variable>
#include <mutex>
#include <thread>
#include <vector>
#include "lock.h"
#include "material.h" #include "material.h"
#include "movepick.h" #include "movepick.h"
#include "pawns.h" #include "pawns.h"
#include "position.h" #include "position.h"
#include "search.h"
const int MAX_THREADS = 32; #include "thread_win32.h"
const int MAX_ACTIVE_SPLIT_POINTS = 8;
struct SplitPoint {
// Const data after splitPoint has been setup
SplitPoint* parent;
const Position* pos;
Depth depth;
bool pvNode;
Value beta;
int ply;
int master;
Move threatMove;
// Const pointers to shared data
MovePicker* mp;
SearchStack* ss;
// Shared data
Lock lock;
volatile int64_t nodes;
volatile Value alpha;
volatile Value bestValue;
volatile int moveCount;
volatile bool is_betaCutoff;
volatile bool is_slave[MAX_THREADS];
};
/// Thread struct is used to keep together all the thread related stuff like locks, /// Thread class keeps together all the thread-related stuff. We use
/// state and especially split points. We also use per-thread pawn and material hash /// per-thread pawn and material hash tables so that once we get a
/// tables so that once we get a pointer to an entry its life time is unlimited and /// pointer to an entry its life time is unlimited and we don't have
/// we don't have to care about someone changing the entry under our feet. /// to care about someone changing the entry under our feet.
struct Thread { class Thread {
enum ThreadState Mutex mutex;
{ ConditionVariable cv;
INITIALIZING, // Thread is initializing itself size_t idx;
SEARCHING, // Thread is performing work bool exit = false, searching = true; // Set before starting std::thread
AVAILABLE, // Thread is waiting for work std::thread stdThread;
BOOKED, // Other thread (master) has booked us as a slave
WORKISWAITING, // Master has ordered us to start
TERMINATED // We are quitting and thread is terminated
};
void wake_up();
bool cutoff_occurred() const;
bool is_available_to(int master) const;
MaterialInfoTable materialTable;
PawnInfoTable pawnTable;
int maxPly;
Lock sleepLock;
WaitCondition sleepCond;
volatile ThreadState state;
SplitPoint* volatile splitPoint;
volatile int activeSplitPoints;
SplitPoint splitPoints[MAX_ACTIVE_SPLIT_POINTS];
};
/// ThreadsManager class is used to handle all the threads related stuff like init,
/// starting, parking and, the most important, launching a slave thread at a split
/// point. All the access to shared thread data is done through this class.
class ThreadsManager {
/* As long as the single ThreadsManager object is defined as a global we don't
need to explicitly initialize to zero its data members because variables with
static storage duration are automatically set to zero before enter main()
*/
public: public:
Thread& operator[](int threadID) { return threads[threadID]; } explicit Thread(size_t);
void init(); virtual ~Thread();
void exit(); virtual void search();
void init_hash_tables(); void clear();
void idle_loop();
void start_searching();
void wait_for_search_finished();
int min_split_depth() const { return minimumSplitDepth; } Pawns::Table pawnsTable;
int size() const { return activeThreads; } Material::Table materialTable;
void set_size(int cnt) { activeThreads = cnt; } Endgames endgames;
size_t PVIdx;
int selDepth, nmp_ply, nmp_odd;
std::atomic<uint64_t> nodes, tbHits;
void read_uci_options(); Position rootPos;
bool available_slave_exists(int master) const; Search::RootMoves rootMoves;
void idle_loop(int threadID, SplitPoint* sp); Depth rootDepth, completedDepth;
CounterMoveHistory counterMoves;
template <bool Fake> ButterflyHistory mainHistory;
void split(Position& pos, SearchStack* ss, Value* alpha, const Value beta, Value* bestValue, CapturePieceToHistory captureHistory;
Depth depth, Move threatMove, int moveCount, MovePicker* mp, bool pvNode); ContinuationHistory contHistory;
private:
Lock mpLock;
Depth minimumSplitDepth;
int maxThreadsPerSplitPoint;
bool useSleepingThreads;
int activeThreads;
volatile bool allThreadsShouldExit;
Thread threads[MAX_THREADS];
}; };
extern ThreadsManager Threads;
#endif // !defined(THREAD_H_INCLUDED) /// MainThread is a derived class specific for main thread
struct MainThread : public Thread {
using Thread::Thread;
void search() override;
void check_time();
bool failedLow;
double bestMoveChanges, previousTimeReduction;
Value previousScore;
int callsCnt;
};
/// ThreadPool struct handles all the threads-related stuff like init, starting,
/// parking and, most importantly, launching a thread. All the access to threads
/// is done through this class.
struct ThreadPool : public std::vector<Thread*> {
void start_thinking(Position&, StateListPtr&, const Search::LimitsType&, bool = false);
void clear();
void set(size_t);
MainThread* main() const { return static_cast<MainThread*>(front()); }
uint64_t nodes_searched() const { return accumulate(&Thread::nodes); }
uint64_t tb_hits() const { return accumulate(&Thread::tbHits); }
std::atomic_bool stop, ponder, stopOnPonderhit;
private:
StateListPtr setupStates;
uint64_t accumulate(std::atomic<uint64_t> Thread::* member) const {
uint64_t sum = 0;
for (Thread* th : *this)
sum += (th->*member).load(std::memory_order_relaxed);
return sum;
}
};
extern ThreadPool Threads;
#endif // #ifndef THREAD_H_INCLUDED
+70
View File
@@ -0,0 +1,70 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef THREAD_WIN32_H_INCLUDED
#define THREAD_WIN32_H_INCLUDED
/// STL thread library used by mingw and gcc when cross compiling for Windows
/// relies on libwinpthread. Currently libwinpthread implements mutexes directly
/// on top of Windows semaphores. Semaphores, being kernel objects, require kernel
/// mode transition in order to lock or unlock, which is very slow compared to
/// interlocked operations (about 30% slower on bench test). To work around this
/// issue, we define our wrappers to the low level Win32 calls. We use critical
/// sections to support Windows XP and older versions. Unfortunately, cond_wait()
/// is racy between unlock() and WaitForSingleObject() but they have the same
/// speed performance as the SRW locks.
#include <condition_variable>
#include <mutex>
#if defined(_WIN32) && !defined(_MSC_VER)
#ifndef NOMINMAX
# define NOMINMAX // Disable macros min() and max()
#endif
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#undef WIN32_LEAN_AND_MEAN
#undef NOMINMAX
/// Mutex and ConditionVariable struct are wrappers of the low level locking
/// machinery and are modeled after the corresponding C++11 classes.
struct Mutex {
Mutex() { InitializeCriticalSection(&cs); }
~Mutex() { DeleteCriticalSection(&cs); }
void lock() { EnterCriticalSection(&cs); }
void unlock() { LeaveCriticalSection(&cs); }
private:
CRITICAL_SECTION cs;
};
typedef std::condition_variable_any ConditionVariable;
#else // Default case: use STL classes
typedef std::mutex Mutex;
typedef std::condition_variable ConditionVariable;
#endif
#endif // #ifndef THREAD_WIN32_H_INCLUDED
+91 -120
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,145 +18,115 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <algorithm>
#include <cfloat>
#include <cmath> #include <cmath>
#include "misc.h"
#include "search.h" #include "search.h"
#include "timeman.h" #include "timeman.h"
#include "ucioption.h" #include "uci.h"
TimeManagement Time; // Our global time management object
namespace { namespace {
/// Constants
const int MoveHorizon = 50; // Plan time management at most this many moves ahead
const float MaxRatio = 3.0f; // When in trouble, we can step over reserved time with this ratio
const float StealRatio = 0.33f; // However we must not steal time from remaining moves over this ratio
// MoveImportance[] is based on naive statistical analysis of "how many games are still undecided
// after n half-moves". Game is considered "undecided" as long as neither side has >275cp advantage.
// Data was extracted from CCRL game database with some simple filtering criteria.
const int MoveImportance[512] = {
7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780, 7780,
7780, 7780, 7780, 7780, 7778, 7778, 7776, 7776, 7776, 7773, 7770, 7768, 7766, 7763, 7757, 7751,
7743, 7735, 7724, 7713, 7696, 7689, 7670, 7656, 7627, 7605, 7571, 7549, 7522, 7493, 7462, 7425,
7385, 7350, 7308, 7272, 7230, 7180, 7139, 7094, 7055, 7010, 6959, 6902, 6841, 6778, 6705, 6651,
6569, 6508, 6435, 6378, 6323, 6253, 6152, 6085, 5995, 5931, 5859, 5794, 5717, 5646, 5544, 5462,
5364, 5282, 5172, 5078, 4988, 4901, 4831, 4764, 4688, 4609, 4536, 4443, 4365, 4293, 4225, 4155,
4085, 4005, 3927, 3844, 3765, 3693, 3634, 3560, 3479, 3404, 3331, 3268, 3207, 3146, 3077, 3011,
2947, 2894, 2828, 2776, 2727, 2676, 2626, 2589, 2538, 2490, 2442, 2394, 2345, 2302, 2243, 2192,
2156, 2115, 2078, 2043, 2004, 1967, 1922, 1893, 1845, 1809, 1772, 1736, 1702, 1674, 1640, 1605,
1566, 1536, 1509, 1479, 1452, 1423, 1388, 1362, 1332, 1304, 1289, 1266, 1250, 1228, 1206, 1180,
1160, 1134, 1118, 1100, 1080, 1068, 1051, 1034, 1012, 1001, 980, 960, 945, 934, 916, 900, 888,
878, 865, 852, 828, 807, 787, 770, 753, 744, 731, 722, 706, 700, 683, 676, 671, 664, 652, 641,
634, 627, 613, 604, 591, 582, 568, 560, 552, 540, 534, 529, 519, 509, 495, 484, 474, 467, 460,
450, 438, 427, 419, 410, 406, 399, 394, 387, 382, 377, 372, 366, 359, 353, 348, 343, 337, 333,
328, 321, 315, 309, 303, 298, 293, 287, 284, 281, 277, 273, 265, 261, 255, 251, 247, 241, 240,
235, 229, 218, 217, 213, 212, 208, 206, 197, 193, 191, 189, 185, 184, 180, 177, 172, 170, 170,
170, 166, 163, 159, 158, 156, 155, 151, 146, 141, 138, 136, 132, 130, 128, 125, 123, 122, 118,
118, 118, 117, 115, 114, 108, 107, 105, 105, 105, 102, 97, 97, 95, 94, 93, 91, 88, 86, 83, 80,
80, 79, 79, 79, 78, 76, 75, 72, 72, 71, 70, 68, 65, 63, 61, 61, 59, 59, 59, 58, 56, 55, 54, 54,
52, 49, 48, 48, 48, 48, 45, 45, 45, 44, 43, 41, 41, 41, 41, 40, 40, 38, 37, 36, 34, 34, 34, 33,
31, 29, 29, 29, 28, 28, 28, 28, 28, 28, 28, 27, 27, 27, 27, 27, 24, 24, 23, 23, 22, 21, 20, 20,
19, 19, 19, 19, 19, 18, 18, 18, 18, 17, 17, 17, 17, 17, 16, 16, 15, 15, 14, 14, 14, 12, 12, 11,
9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
8, 8, 8, 8, 7, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2,
2, 1, 1, 1, 1, 1, 1, 1 };
int move_importance(int ply) { return MoveImportance[Min(ply, 511)]; }
/// Function Prototypes
enum TimeType { OptimumTime, MaxTime }; enum TimeType { OptimumTime, MaxTime };
template<TimeType> const int MoveHorizon = 50; // Plan time management at most this many moves ahead
int remaining(int myTime, int movesToGo, int currentPly); const double MaxRatio = 7.09; // When in trouble, we can step over reserved time with this ratio
} const double StealRatio = 0.35; // However we must not steal time from remaining moves over this ratio
void TimeManager::pv_instability(int curChanges, int prevChanges) { // move_importance() is a skew-logistic function based on naive statistical
// analysis of "how many games are still undecided after n half-moves". Game
// is considered "undecided" as long as neither side has >275cp advantage.
// Data was extracted from the CCRL game database with some simple filtering criteria.
unstablePVExtraTime = curChanges * (optimumSearchTime / 2) double move_importance(int ply) {
+ prevChanges * (optimumSearchTime / 3);
}
const double XScale = 7.64;
const double XShift = 58.4;
const double Skew = 0.183;
void TimeManager::init(const SearchLimits& limits, int currentPly) return pow((1 + exp((ply - XShift) / XScale)), -Skew) + DBL_MIN; // Ensure non-zero
{
/* We support four different kind of time controls:
increment == 0 && movesToGo == 0 means: x basetime [sudden death!]
increment == 0 && movesToGo != 0 means: x moves in y minutes
increment > 0 && movesToGo == 0 means: x basetime + z increment
increment > 0 && movesToGo != 0 means: x moves in y minutes + z increment
Time management is adjusted by following UCI parameters:
emergencyMoveHorizon: Be prepared to always play at least this many moves
emergencyBaseTime : Always attempt to keep at least this much time (in ms) at clock
emergencyMoveTime : Plus attempt to keep at least this much time for each remaining emergency move
minThinkingTime : No matter what, use at least this much thinking before doing the move
*/
int hypMTG, hypMyTime, t1, t2;
// Read uci parameters
int emergencyMoveHorizon = Options["Emergency Move Horizon"].value<int>();
int emergencyBaseTime = Options["Emergency Base Time"].value<int>();
int emergencyMoveTime = Options["Emergency Move Time"].value<int>();
int minThinkingTime = Options["Minimum Thinking Time"].value<int>();
// Initialize to maximum values but unstablePVExtraTime that is reset
unstablePVExtraTime = 0;
optimumSearchTime = maximumSearchTime = limits.time;
// We calculate optimum time usage for different hypothetic "moves to go"-values and choose the
// minimum of calculated search time values. Usually the greatest hypMTG gives the minimum values.
for (hypMTG = 1; hypMTG <= (limits.movesToGo ? Min(limits.movesToGo, MoveHorizon) : MoveHorizon); hypMTG++)
{
// Calculate thinking time for hypothetic "moves to go"-value
hypMyTime = limits.time
+ limits.increment * (hypMTG - 1)
- emergencyBaseTime
- emergencyMoveTime * Min(hypMTG, emergencyMoveHorizon);
hypMyTime = Max(hypMyTime, 0);
t1 = minThinkingTime + remaining<OptimumTime>(hypMyTime, hypMTG, currentPly);
t2 = minThinkingTime + remaining<MaxTime>(hypMyTime, hypMTG, currentPly);
optimumSearchTime = Min(optimumSearchTime, t1);
maximumSearchTime = Min(maximumSearchTime, t2);
} }
if (Options["Ponder"].value<bool>())
optimumSearchTime += optimumSearchTime / 4;
// Make sure that maxSearchTime is not over absoluteMaxSearchTime
optimumSearchTime = Min(optimumSearchTime, maximumSearchTime);
}
namespace {
template<TimeType T> template<TimeType T>
int remaining(int myTime, int movesToGo, int currentPly) int remaining(int myTime, int movesToGo, int ply, int slowMover) {
{
const float TMaxRatio = (T == OptimumTime ? 1 : MaxRatio);
const float TStealRatio = (T == OptimumTime ? 0 : StealRatio);
int thisMoveImportance = move_importance(currentPly); const double TMaxRatio = (T == OptimumTime ? 1 : MaxRatio);
int otherMovesImportance = 0; const double TStealRatio = (T == OptimumTime ? 0 : StealRatio);
for (int i = 1; i < movesToGo; i++) double moveImportance = (move_importance(ply) * slowMover) / 100;
otherMovesImportance += move_importance(currentPly + 2 * i); double otherMovesImportance = 0;
float ratio1 = (TMaxRatio * thisMoveImportance) / float(TMaxRatio * thisMoveImportance + otherMovesImportance); for (int i = 1; i < movesToGo; ++i)
float ratio2 = (thisMoveImportance + TStealRatio * otherMovesImportance) / float(thisMoveImportance + otherMovesImportance); otherMovesImportance += move_importance(ply + 2 * i);
return int(floor(myTime * Min(ratio1, ratio2))); double ratio1 = (TMaxRatio * moveImportance) / (TMaxRatio * moveImportance + otherMovesImportance);
double ratio2 = (moveImportance + TStealRatio * otherMovesImportance) / (moveImportance + otherMovesImportance);
return int(myTime * std::min(ratio1, ratio2)); // Intel C++ asks for an explicit cast
} }
} // namespace
/// init() is called at the beginning of the search and calculates the allowed
/// thinking time out of the time control and current game ply. We support four
/// different kinds of time controls, passed in 'limits':
///
/// inc == 0 && movestogo == 0 means: x basetime [sudden death!]
/// inc == 0 && movestogo != 0 means: x moves in y minutes
/// inc > 0 && movestogo == 0 means: x basetime + z increment
/// inc > 0 && movestogo != 0 means: x moves in y minutes + z increment
void TimeManagement::init(Search::LimitsType& limits, Color us, int ply) {
int minThinkingTime = Options["Minimum Thinking Time"];
int moveOverhead = Options["Move Overhead"];
int slowMover = Options["Slow Mover"];
int npmsec = Options["nodestime"];
// If we have to play in 'nodes as time' mode, then convert from time
// to nodes, and use resulting values in time management formulas.
// WARNING: Given npms (nodes per millisecond) must be much lower then
// the real engine speed to avoid time losses.
if (npmsec)
{
if (!availableNodes) // Only once at game start
availableNodes = npmsec * limits.time[us]; // Time is in msec
// Convert from millisecs to nodes
limits.time[us] = (int)availableNodes;
limits.inc[us] *= npmsec;
limits.npmsec = npmsec;
}
startTime = limits.startTime;
optimumTime = maximumTime = std::max(limits.time[us], minThinkingTime);
const int MaxMTG = limits.movestogo ? std::min(limits.movestogo, MoveHorizon) : MoveHorizon;
// We calculate optimum time usage for different hypothetical "moves to go"-values
// and choose the minimum of calculated search time values. Usually the greatest
// hypMTG gives the minimum values.
for (int hypMTG = 1; hypMTG <= MaxMTG; ++hypMTG)
{
// Calculate thinking time for hypothetical "moves to go"-value
int hypMyTime = limits.time[us]
+ limits.inc[us] * (hypMTG - 1)
- moveOverhead * (2 + std::min(hypMTG, 40));
hypMyTime = std::max(hypMyTime, 0);
int t1 = minThinkingTime + remaining<OptimumTime>(hypMyTime, hypMTG, ply, slowMover);
int t2 = minThinkingTime + remaining<MaxTime >(hypMyTime, hypMTG, ply, slowMover);
optimumTime = std::min(t1, optimumTime);
maximumTime = std::min(t2, maximumTime);
}
if (Options["Ponder"])
optimumTime += optimumTime / 4;
} }
+21 -12
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,23 +18,31 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(TIMEMAN_H_INCLUDED) #ifndef TIMEMAN_H_INCLUDED
#define TIMEMAN_H_INCLUDED #define TIMEMAN_H_INCLUDED
struct SearchLimits; #include "misc.h"
#include "search.h"
#include "thread.h"
class TimeManager { /// The TimeManagement class computes the optimal time to think depending on
/// the maximum available time, the game move number and other parameters.
class TimeManagement {
public: public:
void init(Search::LimitsType& limits, Color us, int ply);
int optimum() const { return optimumTime; }
int maximum() const { return maximumTime; }
int elapsed() const { return int(Search::Limits.npmsec ? Threads.nodes_searched() : now() - startTime); }
void init(const SearchLimits& limits, int currentPly); int64_t availableNodes; // When in 'nodes as time' mode
void pv_instability(int curChanges, int prevChanges);
int available_time() const { return optimumSearchTime + unstablePVExtraTime; }
int maximum_time() const { return maximumSearchTime; }
private: private:
int optimumSearchTime; TimePoint startTime;
int maximumSearchTime; int optimumTime;
int unstablePVExtraTime; int maximumTime;
}; };
#endif // !defined(TIMEMAN_H_INCLUDED) extern TimeManagement Time;
#endif // #ifndef TIMEMAN_H_INCLUDED
+62 -88
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,128 +18,101 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <cassert> #include <cstring> // For std::memset
#include <cstring>
#include <iostream> #include <iostream>
#include "bitboard.h"
#include "tt.h" #include "tt.h"
TranspositionTable TT; // Our global transposition table TranspositionTable TT; // Our global transposition table
TranspositionTable::TranspositionTable() {
size = generation = 0; /// TranspositionTable::resize() sets the size of the transposition table,
entries = NULL; /// measured in megabytes. Transposition table consists of a power of 2 number
} /// of clusters and each cluster consists of ClusterSize number of TTEntry.
TranspositionTable::~TranspositionTable() { void TranspositionTable::resize(size_t mbSize) {
delete [] entries; size_t newClusterCount = mbSize * 1024 * 1024 / sizeof(Cluster);
}
if (newClusterCount == clusterCount)
/// TranspositionTable::set_size() sets the size of the transposition table,
/// measured in megabytes.
void TranspositionTable::set_size(size_t mbSize) {
size_t newSize = 1024;
// Transposition table consists of clusters and each cluster consists
// of ClusterSize number of TTEntries. Each non-empty entry contains
// information of exactly one position and newSize is the number of
// clusters we are going to allocate.
while (2ULL * newSize * sizeof(TTCluster) <= (mbSize << 20))
newSize *= 2;
if (newSize == size)
return; return;
size = newSize; clusterCount = newClusterCount;
delete [] entries;
entries = new (std::nothrow) TTCluster[size]; free(mem);
if (!entries) mem = malloc(clusterCount * sizeof(Cluster) + CacheLineSize - 1);
if (!mem)
{ {
std::cerr << "Failed to allocate " << mbSize std::cerr << "Failed to allocate " << mbSize
<< " MB for transposition table." << std::endl; << "MB for transposition table." << std::endl;
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
table = (Cluster*)((uintptr_t(mem) + CacheLineSize - 1) & ~(CacheLineSize - 1));
clear(); clear();
} }
/// TranspositionTable::clear() overwrites the entire transposition table /// TranspositionTable::clear() overwrites the entire transposition table
/// with zeroes. It is called whenever the table is resized, or when the /// with zeros. It is called whenever the table is resized, or when the
/// user asks the program to clear the table (from the UCI interface). /// user asks the program to clear the table (from the UCI interface).
void TranspositionTable::clear() { void TranspositionTable::clear() {
memset(entries, 0, size * sizeof(TTCluster)); std::memset(table, 0, clusterCount * sizeof(Cluster));
} }
/// TranspositionTable::store() writes a new entry containing position key and /// TranspositionTable::probe() looks up the current position in the transposition
/// valuable information of current position. The lowest order bits of position /// table. It returns true and a pointer to the TTEntry if the position is found.
/// key are used to decide on which cluster the position will be placed. /// Otherwise, it returns false and a pointer to an empty or least valuable TTEntry
/// When a new entry is written and there are no empty entries available in cluster, /// to be replaced later. The replace value of an entry is calculated as its depth
/// it replaces the least valuable of entries. A TTEntry t1 is considered to be /// minus 8 times its relative age. TTEntry t1 is considered more valuable than
/// more valuable than a TTEntry t2 if t1 is from the current search and t2 is from /// TTEntry t2 if its replace value is greater than that of t2.
/// a previous search, or if the depth of t1 is bigger than the depth of t2.
void TranspositionTable::store(const Key posKey, Value v, ValueType t, Depth d, Move m, Value statV, Value kingD) { TTEntry* TranspositionTable::probe(const Key key, bool& found) const {
int c1, c2, c3; TTEntry* const tte = first_entry(key);
TTEntry *tte, *replace; const uint16_t key16 = key >> 48; // Use the high 16 bits as key inside the cluster
uint32_t posKey32 = posKey >> 32; // Use the high 32 bits as key inside the cluster
tte = replace = first_entry(posKey); for (int i = 0; i < ClusterSize; ++i)
if (!tte[i].key16 || tte[i].key16 == key16)
for (int i = 0; i < ClusterSize; i++, tte++)
{
if (!tte->key() || tte->key() == posKey32) // Empty or overwrite old
{ {
// Preserve any existing ttMove if ((tte[i].genBound8 & 0xFC) != generation8 && tte[i].key16)
if (m == MOVE_NONE) tte[i].genBound8 = uint8_t(generation8 | tte[i].bound()); // Refresh
m = tte->move();
tte->save(posKey32, v, t, d, m, generation, statV, kingD); return found = (bool)tte[i].key16, &tte[i];
return;
} }
// Implement replace strategy // Find an entry to be replaced according to the replacement strategy
c1 = (replace->generation() == generation ? 2 : 0); TTEntry* replace = tte;
c2 = (tte->generation() == generation || tte->type() == VALUE_TYPE_EXACT ? -2 : 0); for (int i = 1; i < ClusterSize; ++i)
c3 = (tte->depth() < replace->depth() ? 1 : 0); // Due to our packed storage format for generation and its cyclic
// nature we add 259 (256 is the modulus plus 3 to keep the lowest
// two bound bits from affecting the result) to calculate the entry
// age correctly even after generation8 overflows into the next cycle.
if ( replace->depth8 - ((259 + generation8 - replace->genBound8) & 0xFC) * 2
> tte[i].depth8 - ((259 + generation8 - tte[i].genBound8) & 0xFC) * 2)
replace = &tte[i];
if (c1 + c2 + c3 > 0) return found = false, replace;
replace = tte; }
/// TranspositionTable::hashfull() returns an approximation of the hashtable
/// occupation during a search. The hash is x permill full, as per UCI protocol.
int TranspositionTable::hashfull() const {
int cnt = 0;
for (int i = 0; i < 1000 / ClusterSize; i++)
{
const TTEntry* tte = &table[i].entry[0];
for (int j = 0; j < ClusterSize; j++)
if ((tte[j].genBound8 & 0xFC) == generation8)
cnt++;
} }
replace->save(posKey32, v, t, d, m, generation, statV, kingD); return cnt;
}
/// TranspositionTable::probe() looks up the current position in the
/// transposition table. Returns a pointer to the TTEntry or NULL if
/// position is not found.
TTEntry* TranspositionTable::probe(const Key posKey) const {
uint32_t posKey32 = posKey >> 32;
TTEntry* tte = first_entry(posKey);
for (int i = 0; i < ClusterSize; i++, tte++)
if (tte->key() == posKey32)
return tte;
return NULL;
}
/// TranspositionTable::new_search() is called at the beginning of every new
/// search. It increments the "generation" variable, which is used to
/// distinguish transposition table entries from previous searches from
/// entries from the current search.
void TranspositionTable::new_search() {
generation++;
} }
+76 -128
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,157 +18,104 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(TT_H_INCLUDED) #ifndef TT_H_INCLUDED
#define TT_H_INCLUDED #define TT_H_INCLUDED
#include <iostream> #include "misc.h"
#include "move.h"
#include "types.h" #include "types.h"
/// TTEntry struct is the 10 bytes transposition table entry, defined as below:
/// The TTEntry is the class of transposition table entries
/// ///
/// A TTEntry needs 128 bits to be stored /// key 16 bit
/// /// move 16 bit
/// bit 0-31: key /// value 16 bit
/// bit 32-63: data /// eval value 16 bit
/// bit 64-79: value /// generation 6 bit
/// bit 80-95: depth /// bound type 2 bit
/// bit 96-111: static value /// depth 8 bit
/// bit 112-127: margin of static value
///
/// the 32 bits of the data field are so defined
///
/// bit 0-15: move
/// bit 16-20: not used
/// bit 21-22: value type
/// bit 23-31: generation
class TTEntry { struct TTEntry {
public: Move move() const { return (Move )move16; }
void save(uint32_t k, Value v, ValueType t, Depth d, Move m, int g, Value statV, Value statM) { Value value() const { return (Value)value16; }
Value eval() const { return (Value)eval16; }
Depth depth() const { return (Depth)(depth8 * int(ONE_PLY)); }
Bound bound() const { return (Bound)(genBound8 & 0x3); }
key32 = (uint32_t)k; void save(Key k, Value v, Bound b, Depth d, Move m, Value ev, uint8_t g) {
move16 = (uint16_t)m;
valueType = (uint8_t)t; assert(d / ONE_PLY * ONE_PLY == d);
generation8 = (uint8_t)g;
value16 = (int16_t)v; // Preserve any existing move for the same position
depth16 = (int16_t)d; if (m || (k >> 48) != key16)
staticValue = (int16_t)statV; move16 = (uint16_t)m;
staticMargin = (int16_t)statM;
// Don't overwrite more valuable entries
if ( (k >> 48) != key16
|| d / ONE_PLY > depth8 - 4
/* || g != (genBound8 & 0xFC) // Matching non-zero keys are already refreshed by probe() */
|| b == BOUND_EXACT)
{
key16 = (uint16_t)(k >> 48);
value16 = (int16_t)v;
eval16 = (int16_t)ev;
genBound8 = (uint8_t)(g | b);
depth8 = (int8_t)(d / ONE_PLY);
}
} }
void set_generation(int g) { generation8 = (uint8_t)g; }
uint32_t key() const { return key32; }
Depth depth() const { return (Depth)depth16; }
Move move() const { return (Move)move16; }
Value value() const { return (Value)value16; }
ValueType type() const { return (ValueType)valueType; }
int generation() const { return (int)generation8; }
Value static_value() const { return (Value)staticValue; }
Value static_value_margin() const { return (Value)staticMargin; }
private: private:
uint32_t key32; friend class TranspositionTable;
uint16_t key16;
uint16_t move16; uint16_t move16;
uint8_t valueType, generation8; int16_t value16;
int16_t value16, depth16, staticValue, staticMargin; int16_t eval16;
uint8_t genBound8;
int8_t depth8;
}; };
/// This is the number of TTEntry slots for each cluster /// A TranspositionTable consists of a power of 2 number of clusters and each
const int ClusterSize = 4; /// cluster consists of ClusterSize number of TTEntry. Each non-empty entry
/// contains information of exactly one position. The size of a cluster should
/// divide the size of a cache line size, to ensure that clusters never cross
/// TTCluster consists of ClusterSize number of TTEntries. Size of TTCluster /// cache lines. This ensures best cache performance, as the cacheline is
/// must not be bigger than a cache line size. In case it is less, it should /// prefetched, as soon as possible.
/// be padded to guarantee always aligned accesses.
struct TTCluster {
TTEntry data[ClusterSize];
};
/// The transposition table class. This is basically just a huge array containing
/// TTCluster objects, and a few methods for writing and reading entries.
class TranspositionTable { class TranspositionTable {
TranspositionTable(const TranspositionTable&); static const int CacheLineSize = 64;
TranspositionTable& operator=(const TranspositionTable&); static const int ClusterSize = 3;
struct Cluster {
TTEntry entry[ClusterSize];
char padding[2]; // Align to a divisor of the cache line size
};
static_assert(CacheLineSize % sizeof(Cluster) == 0, "Cluster size incorrect");
public: public:
TranspositionTable(); ~TranspositionTable() { free(mem); }
~TranspositionTable(); void new_search() { generation8 += 4; } // Lower 2 bits are used by Bound
void set_size(size_t mbSize); uint8_t generation() const { return generation8; }
TTEntry* probe(const Key key, bool& found) const;
int hashfull() const;
void resize(size_t mbSize);
void clear(); void clear();
void store(const Key posKey, Value v, ValueType type, Depth d, Move m, Value statV, Value kingD);
TTEntry* probe(const Key posKey) const; // The 32 lowest order bits of the key are used to get the index of the cluster
void new_search(); TTEntry* first_entry(const Key key) const {
TTEntry* first_entry(const Key posKey) const; return &table[(uint32_t(key) * uint64_t(clusterCount)) >> 32].entry[0];
void refresh(const TTEntry* tte) const; }
private: private:
size_t size; size_t clusterCount;
TTCluster* entries; Cluster* table;
uint8_t generation; // Size must be not bigger then TTEntry::generation8 void* mem;
uint8_t generation8; // Size must be not bigger than TTEntry::genBound8
}; };
extern TranspositionTable TT; extern TranspositionTable TT;
#endif // #ifndef TT_H_INCLUDED
/// TranspositionTable::first_entry() returns a pointer to the first entry of
/// a cluster given a position. The lowest order bits of the key are used to
/// get the index of the cluster.
inline TTEntry* TranspositionTable::first_entry(const Key posKey) const {
return entries[((uint32_t)posKey) & (size - 1)].data;
}
/// TranspositionTable::refresh() updates the 'generation' value of the TTEntry
/// to avoid aging. Normally called after a TT hit.
inline void TranspositionTable::refresh(const TTEntry* tte) const {
const_cast<TTEntry*>(tte)->set_generation(generation);
}
/// A simple fixed size hash table used to store pawns and material
/// configurations. It is basically just an array of Entry objects.
/// Without cluster concept or overwrite policy.
template<class Entry, int HashSize>
struct SimpleHash {
typedef SimpleHash<Entry, HashSize> Base;
void init() {
if (entries)
return;
entries = new (std::nothrow) Entry[HashSize];
if (!entries)
{
std::cerr << "Failed to allocate " << HashSize * sizeof(Entry)
<< " bytes for hash table." << std::endl;
exit(EXIT_FAILURE);
}
memset(entries, 0, HashSize * sizeof(Entry));
}
virtual ~SimpleHash() { delete [] entries; }
Entry* probe(Key key) const { return entries + ((uint32_t)key & (HashSize - 1)); }
void prefetch(Key key) const { ::prefetch((char*)probe(key)); }
protected:
Entry* entries;
};
#endif // !defined(TT_H_INCLUDED)
+333 -337
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,198 +18,212 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#if !defined(TYPES_H_INCLUDED) #ifndef TYPES_H_INCLUDED
#define TYPES_H_INCLUDED #define TYPES_H_INCLUDED
/// When compiling with provided Makefile (e.g. for Linux and OSX), configuration
/// is done automatically. To get started type 'make help'.
///
/// When Makefile is not used (e.g. with Microsoft Visual Studio) some switches
/// need to be set manually:
///
/// -DNDEBUG | Disable debugging mode. Always use this for release.
///
/// -DNO_PREFETCH | Disable use of prefetch asm-instruction. You may need this to
/// | run on some very old machines.
///
/// -DUSE_POPCNT | Add runtime support for use of popcnt asm-instruction. Works
/// | only in 64-bit mode and requires hardware with popcnt support.
///
/// -DUSE_PEXT | Add runtime support for use of pext asm-instruction. Works
/// | only in 64-bit mode and requires hardware with pext support.
#include <cassert>
#include <cctype>
#include <climits> #include <climits>
#include <cstdint>
#include <cstdlib> #include <cstdlib>
#if defined(_MSC_VER) #if defined(_MSC_VER)
// Disable some silly and noisy warning from MSVC compiler // Disable some silly and noisy warning from MSVC compiler
#pragma warning(disable: 4800) // Forcing value to bool 'true' or 'false'
#pragma warning(disable: 4127) // Conditional expression is constant #pragma warning(disable: 4127) // Conditional expression is constant
#pragma warning(disable: 4146) // Unary minus operator applied to unsigned type #pragma warning(disable: 4146) // Unary minus operator applied to unsigned type
#pragma warning(disable: 4800) // Forcing value to bool 'true' or 'false'
#endif
// MSVC does not support <inttypes.h> /// Predefined macros hell:
typedef signed __int8 int8_t; ///
typedef unsigned __int8 uint8_t; /// __GNUC__ Compiler is gcc, Clang or Intel on Linux
typedef signed __int16 int16_t; /// __INTEL_COMPILER Compiler is Intel
typedef unsigned __int16 uint16_t; /// _MSC_VER Compiler is MSVC or Intel on Windows
typedef signed __int32 int32_t; /// _WIN32 Building on Windows (any)
typedef unsigned __int32 uint32_t; /// _WIN64 Building on Windows 64 bit
typedef signed __int64 int64_t;
typedef unsigned __int64 uint64_t;
#if defined(_WIN64) && defined(_MSC_VER) // No Makefile used
# include <intrin.h> // Microsoft header for _BitScanForward64()
# define IS_64BIT
#endif
#if defined(USE_POPCNT) && (defined(__INTEL_COMPILER) || defined(_MSC_VER))
# include <nmmintrin.h> // Intel and Microsoft header for _mm_popcnt_u64()
#endif
#if !defined(NO_PREFETCH) && (defined(__INTEL_COMPILER) || defined(_MSC_VER))
# include <xmmintrin.h> // Intel and Microsoft header for _mm_prefetch()
#endif
#if defined(USE_PEXT)
# include <immintrin.h> // Header for _pext_u64() intrinsic
# define pext(b, m) _pext_u64(b, m)
#else #else
# define pext(b, m) 0
#include <inttypes.h>
#endif #endif
#define Min(x, y) (((x) < (y)) ? (x) : (y)) #ifdef USE_POPCNT
#define Max(x, y) (((x) < (y)) ? (y) : (x)) const bool HasPopCnt = true;
////
//// Configuration
////
//// For Linux and OSX configuration is done automatically using Makefile.
//// To get started type "make help".
////
//// For windows part of the configuration is detected automatically, but
//// some switches need to be set manually:
////
//// -DNDEBUG | Disable debugging mode. Use always.
////
//// -DNO_PREFETCH | Disable use of prefetch asm-instruction. A must if you want the
//// | executable to run on some very old machines.
////
//// -DUSE_POPCNT | Add runtime support for use of popcnt asm-instruction.
//// | Works only in 64-bit mode. For compiling requires hardware
//// | with popcnt support. Around 4% speed-up.
////
//// -DOLD_LOCKS | By default under Windows are used the fast Slim Reader/Writer (SRW)
//// | Locks and Condition Variables: these are not supported by Windows XP
//// | and older, to compile for those platforms you should enable OLD_LOCKS.
// Automatic detection for 64-bit under Windows
#if defined(_WIN64)
#define IS_64BIT
#endif
// Automatic detection for use of bsfq asm-instruction under Windows
#if defined(_WIN64)
#define USE_BSFQ
#endif
// Intel header for _mm_popcnt_u64() intrinsic
#if defined(USE_POPCNT) && defined(_MSC_VER) && defined(__INTEL_COMPILER)
#include <nmmintrin.h>
#endif
// Cache line alignment specification
#if defined(_MSC_VER) || defined(__INTEL_COMPILER)
#define CACHE_LINE_ALIGNMENT __declspec(align(64))
#else #else
#define CACHE_LINE_ALIGNMENT __attribute__ ((aligned(64))) const bool HasPopCnt = false;
#endif #endif
// Define a __cpuid() function for gcc compilers, for Intel and MSVC #ifdef USE_PEXT
// is already available as an intrinsic. const bool HasPext = true;
#if defined(_MSC_VER)
#include <intrin.h>
#elif defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
inline void __cpuid(int CPUInfo[4], int InfoType)
{
int* eax = CPUInfo + 0;
int* ebx = CPUInfo + 1;
int* ecx = CPUInfo + 2;
int* edx = CPUInfo + 3;
*eax = InfoType;
*ecx = 0;
__asm__("cpuid" : "=a" (*eax), "=b" (*ebx), "=c" (*ecx), "=d" (*edx)
: "0" (*eax), "2" (*ecx));
}
#else #else
inline void __cpuid(int CPUInfo[4], int) const bool HasPext = false;
{
CPUInfo[0] = CPUInfo[1] = CPUInfo[2] = CPUInfo[3] = 0;
}
#endif #endif
// Define FORCE_INLINE macro to force inlining overriding compiler choice #ifdef IS_64BIT
#if defined(_MSC_VER) const bool Is64Bit = true;
#define FORCE_INLINE __forceinline
#elif defined(__GNUC__)
#define FORCE_INLINE inline __attribute__((always_inline))
#else #else
#define FORCE_INLINE inline const bool Is64Bit = false;
#endif #endif
/// cpu_has_popcnt() detects support for popcnt instruction at runtime
inline bool cpu_has_popcnt() {
int CPUInfo[4] = {-1};
__cpuid(CPUInfo, 0x00000001);
return (CPUInfo[2] >> 23) & 1;
}
/// CpuHasPOPCNT is a global constant initialized at startup that
/// is set to true if CPU on which application runs supports popcnt
/// hardware instruction. Unless USE_POPCNT is not defined.
#if defined(USE_POPCNT)
const bool CpuHasPOPCNT = cpu_has_popcnt();
#else
const bool CpuHasPOPCNT = false;
#endif
/// CpuIs64Bit is a global constant initialized at compile time that
/// is set to true if CPU on which application runs is a 64 bits.
#if defined(IS_64BIT)
const bool CpuIs64Bit = true;
#else
const bool CpuIs64Bit = false;
#endif
#include <string>
typedef uint64_t Key; typedef uint64_t Key;
typedef uint64_t Bitboard; typedef uint64_t Bitboard;
const int PLY_MAX = 100; const int MAX_MOVES = 256;
const int PLY_MAX_PLUS_2 = PLY_MAX + 2; const int MAX_PLY = 128;
enum ValueType { /// A move needs 16 bits to be stored
VALUE_TYPE_NONE = 0, ///
VALUE_TYPE_UPPER = 1, /// bit 0- 5: destination square (from 0 to 63)
VALUE_TYPE_LOWER = 2, /// bit 6-11: origin square (from 0 to 63)
VALUE_TYPE_EXACT = VALUE_TYPE_UPPER | VALUE_TYPE_LOWER /// bit 12-13: promotion piece type - 2 (from KNIGHT-2 to QUEEN-2)
/// bit 14-15: special move flag: promotion (1), en passant (2), castling (3)
/// NOTE: EN-PASSANT bit is set only when a pawn can be captured
///
/// Special cases are MOVE_NONE and MOVE_NULL. We can sneak these in because in
/// any normal move destination square is always different from origin square
/// while MOVE_NONE and MOVE_NULL have the same origin and destination square.
enum Move : int {
MOVE_NONE,
MOVE_NULL = 65
}; };
enum Value { enum MoveType {
VALUE_ZERO = 0, NORMAL,
VALUE_DRAW = 0, PROMOTION = 1 << 14,
VALUE_KNOWN_WIN = 15000, ENPASSANT = 2 << 14,
VALUE_MATE = 30000, CASTLING = 3 << 14
VALUE_INFINITE = 30001,
VALUE_NONE = 30002,
VALUE_MATE_IN_PLY_MAX = VALUE_MATE - PLY_MAX,
VALUE_MATED_IN_PLY_MAX = -VALUE_MATE + PLY_MAX,
VALUE_ENSURE_INTEGER_SIZE_P = INT_MAX,
VALUE_ENSURE_INTEGER_SIZE_N = INT_MIN
};
enum PieceType {
PIECE_TYPE_NONE = 0,
PAWN = 1, KNIGHT = 2, BISHOP = 3, ROOK = 4, QUEEN = 5, KING = 6
};
enum Piece {
PIECE_NONE_DARK_SQ = 0, WP = 1, WN = 2, WB = 3, WR = 4, WQ = 5, WK = 6,
BP = 9, BN = 10, BB = 11, BR = 12, BQ = 13, BK = 14, PIECE_NONE = 16
}; };
enum Color { enum Color {
WHITE, BLACK, COLOR_NONE WHITE, BLACK, COLOR_NB = 2
}; };
enum Depth { enum CastlingSide {
KING_SIDE, QUEEN_SIDE, CASTLING_SIDE_NB = 2
ONE_PLY = 2,
DEPTH_ZERO = 0 * ONE_PLY,
DEPTH_QS_CHECKS = -1 * ONE_PLY,
DEPTH_QS_NO_CHECKS = -2 * ONE_PLY,
DEPTH_NONE = -127 * ONE_PLY
}; };
enum Square { enum CastlingRight {
NO_CASTLING,
WHITE_OO,
WHITE_OOO = WHITE_OO << 1,
BLACK_OO = WHITE_OO << 2,
BLACK_OOO = WHITE_OO << 3,
ANY_CASTLING = WHITE_OO | WHITE_OOO | BLACK_OO | BLACK_OOO,
CASTLING_RIGHT_NB = 16
};
template<Color C, CastlingSide S> struct MakeCastling {
static constexpr CastlingRight
right = C == WHITE ? S == QUEEN_SIDE ? WHITE_OOO : WHITE_OO
: S == QUEEN_SIDE ? BLACK_OOO : BLACK_OO;
};
enum Phase {
PHASE_ENDGAME,
PHASE_MIDGAME = 128,
MG = 0, EG = 1, PHASE_NB = 2
};
enum ScaleFactor {
SCALE_FACTOR_DRAW = 0,
SCALE_FACTOR_ONEPAWN = 48,
SCALE_FACTOR_NORMAL = 64,
SCALE_FACTOR_MAX = 128,
SCALE_FACTOR_NONE = 255
};
enum Bound {
BOUND_NONE,
BOUND_UPPER,
BOUND_LOWER,
BOUND_EXACT = BOUND_UPPER | BOUND_LOWER
};
enum Value : int {
VALUE_ZERO = 0,
VALUE_DRAW = 0,
VALUE_KNOWN_WIN = 10000,
VALUE_MATE = 32000,
VALUE_INFINITE = 32001,
VALUE_NONE = 32002,
VALUE_MATE_IN_MAX_PLY = VALUE_MATE - 2 * MAX_PLY,
VALUE_MATED_IN_MAX_PLY = -VALUE_MATE + 2 * MAX_PLY,
PawnValueMg = 171, PawnValueEg = 240,
KnightValueMg = 764, KnightValueEg = 848,
BishopValueMg = 826, BishopValueEg = 891,
RookValueMg = 1282, RookValueEg = 1373,
QueenValueMg = 2526, QueenValueEg = 2646,
MidgameLimit = 15258, EndgameLimit = 3915
};
enum PieceType {
NO_PIECE_TYPE, PAWN, KNIGHT, BISHOP, ROOK, QUEEN, KING,
ALL_PIECES = 0,
QUEEN_DIAGONAL = 7,
PIECE_TYPE_NB = 8
};
enum Piece {
NO_PIECE,
W_PAWN = 1, W_KNIGHT, W_BISHOP, W_ROOK, W_QUEEN, W_KING,
B_PAWN = 9, B_KNIGHT, B_BISHOP, B_ROOK, B_QUEEN, B_KING,
PIECE_NB = 16
};
extern Value PieceValue[PHASE_NB][PIECE_NB];
enum Depth : int {
ONE_PLY = 1,
DEPTH_ZERO = 0 * ONE_PLY,
DEPTH_QS_CHECKS = 0 * ONE_PLY,
DEPTH_QS_NO_CHECKS = -1 * ONE_PLY,
DEPTH_QS_RECAPTURES = -5 * ONE_PLY,
DEPTH_NONE = -6 * ONE_PLY,
DEPTH_MAX = MAX_PLY * ONE_PLY
};
static_assert(!(ONE_PLY & (ONE_PLY - 1)), "ONE_PLY is not a power of 2");
enum Square : int {
SQ_A1, SQ_B1, SQ_C1, SQ_D1, SQ_E1, SQ_F1, SQ_G1, SQ_H1, SQ_A1, SQ_B1, SQ_C1, SQ_D1, SQ_E1, SQ_F1, SQ_G1, SQ_H1,
SQ_A2, SQ_B2, SQ_C2, SQ_D2, SQ_E2, SQ_F2, SQ_G2, SQ_H2, SQ_A2, SQ_B2, SQ_C2, SQ_D2, SQ_E2, SQ_F2, SQ_G2, SQ_H2,
SQ_A3, SQ_B3, SQ_C3, SQ_D3, SQ_E3, SQ_F3, SQ_G3, SQ_H3, SQ_A3, SQ_B3, SQ_C3, SQ_D3, SQ_E3, SQ_F3, SQ_G3, SQ_H3,
@@ -219,252 +234,233 @@ enum Square {
SQ_A8, SQ_B8, SQ_C8, SQ_D8, SQ_E8, SQ_F8, SQ_G8, SQ_H8, SQ_A8, SQ_B8, SQ_C8, SQ_D8, SQ_E8, SQ_F8, SQ_G8, SQ_H8,
SQ_NONE, SQ_NONE,
DELTA_N = 8, SQUARE_NB = 64
DELTA_E = 1,
DELTA_S = -8,
DELTA_W = -1,
DELTA_NN = DELTA_N + DELTA_N,
DELTA_NE = DELTA_N + DELTA_E,
DELTA_SE = DELTA_S + DELTA_E,
DELTA_SS = DELTA_S + DELTA_S,
DELTA_SW = DELTA_S + DELTA_W,
DELTA_NW = DELTA_N + DELTA_W
}; };
enum File { enum Direction : int {
FILE_A, FILE_B, FILE_C, FILE_D, FILE_E, FILE_F, FILE_G, FILE_H NORTH = 8,
EAST = 1,
SOUTH = -NORTH,
WEST = -EAST,
NORTH_EAST = NORTH + EAST,
SOUTH_EAST = SOUTH + EAST,
SOUTH_WEST = SOUTH + WEST,
NORTH_WEST = NORTH + WEST
}; };
enum Rank { enum File : int {
RANK_1, RANK_2, RANK_3, RANK_4, RANK_5, RANK_6, RANK_7, RANK_8 FILE_A, FILE_B, FILE_C, FILE_D, FILE_E, FILE_F, FILE_G, FILE_H, FILE_NB
}; };
enum SquareColor { enum Rank : int {
DARK, LIGHT RANK_1, RANK_2, RANK_3, RANK_4, RANK_5, RANK_6, RANK_7, RANK_8, RANK_NB
};
enum ScaleFactor {
SCALE_FACTOR_ZERO = 0,
SCALE_FACTOR_NORMAL = 64,
SCALE_FACTOR_MAX = 128,
SCALE_FACTOR_NONE = 255
}; };
/// Score enum keeps a midgame and an endgame value in a single /// Score enum stores a middlegame and an endgame value in a single integer
/// integer (enum), first LSB 16 bits are used to store endgame /// (enum). The least significant 16 bits are used to store the endgame value
/// value, while upper bits are used for midgame value. Compiler /// and the upper 16 bits are used to store the middlegame value. Take some
/// is free to choose the enum type as long as can keep its data, /// care to avoid left-shifting a signed int to avoid undefined behavior.
/// so ensure Score to be an integer type. enum Score : int { SCORE_ZERO };
enum Score {
SCORE_ZERO = 0,
SCORE_ENSURE_INTEGER_SIZE_P = INT_MAX,
SCORE_ENSURE_INTEGER_SIZE_N = INT_MIN
};
#define ENABLE_OPERATORS_ON(T) \ constexpr Score make_score(int mg, int eg) {
inline T operator+ (const T d1, const T d2) { return T(int(d1) + int(d2)); } \ return Score((int)((unsigned int)eg << 16) + mg);
inline T operator- (const T d1, const T d2) { return T(int(d1) - int(d2)); } \ }
inline T operator* (int i, const T d) { return T(i * int(d)); } \
inline T operator* (const T d, int i) { return T(int(d) * i); } \
inline T operator/ (const T d, int i) { return T(int(d) / i); } \
inline T operator- (const T d) { return T(-int(d)); } \
inline T operator++ (T& d, int) {d = T(int(d) + 1); return d; } \
inline T operator-- (T& d, int) { d = T(int(d) - 1); return d; } \
inline void operator+= (T& d1, const T d2) { d1 = d1 + d2; } \
inline void operator-= (T& d1, const T d2) { d1 = d1 - d2; } \
inline void operator*= (T& d, int i) { d = T(int(d) * i); } \
inline void operator/= (T& d, int i) { d = T(int(d) / i); }
ENABLE_OPERATORS_ON(Value) /// Extracting the signed lower and upper 16 bits is not so trivial because
ENABLE_OPERATORS_ON(PieceType) /// according to the standard a simple cast to short is implementation defined
ENABLE_OPERATORS_ON(Piece) /// and so is a right shift of a signed integer.
ENABLE_OPERATORS_ON(Color) inline Value eg_value(Score s) {
ENABLE_OPERATORS_ON(Depth) union { uint16_t u; int16_t s; } eg = { uint16_t(unsigned(s + 0x8000) >> 16) };
ENABLE_OPERATORS_ON(Square) return Value(eg.s);
ENABLE_OPERATORS_ON(File) }
ENABLE_OPERATORS_ON(Rank)
#undef ENABLE_OPERATORS_ON inline Value mg_value(Score s) {
union { uint16_t u; int16_t s; } mg = { uint16_t(unsigned(s)) };
return Value(mg.s);
}
// Extra operators for adding integers to a Value #define ENABLE_BASE_OPERATORS_ON(T) \
inline Value operator+ (Value v, int i) { return Value(int(v) + i); } constexpr T operator+(T d1, T d2) { return T(int(d1) + int(d2)); } \
inline Value operator- (Value v, int i) { return Value(int(v) - i); } constexpr T operator-(T d1, T d2) { return T(int(d1) - int(d2)); } \
constexpr T operator-(T d) { return T(-int(d)); } \
inline T& operator+=(T& d1, T d2) { return d1 = d1 + d2; } \
inline T& operator-=(T& d1, T d2) { return d1 = d1 - d2; }
// Extracting the _signed_ lower and upper 16 bits it not so trivial #define ENABLE_INCR_OPERATORS_ON(T) \
// because according to the standard a simple cast to short is inline T& operator++(T& d) { return d = T(int(d) + 1); } \
// implementation defined and so is a right shift of a signed integer. inline T& operator--(T& d) { return d = T(int(d) - 1); }
inline Value mg_value(Score s) { return Value(((int(s) + 32768) & ~0xffff) / 0x10000); }
// Unfortunatly on Intel 64 bit we have a small speed regression, so use a faster code in #define ENABLE_FULL_OPERATORS_ON(T) \
// this case, although not 100% standard compliant it seems to work for Intel and MSVC. ENABLE_BASE_OPERATORS_ON(T) \
#if defined(IS_64BIT) && (!defined(__GNUC__) || defined(__INTEL_COMPILER)) ENABLE_INCR_OPERATORS_ON(T) \
inline Value eg_value(Score s) { return Value(int16_t(s & 0xffff)); } constexpr T operator*(int i, T d) { return T(i * int(d)); } \
#else constexpr T operator*(T d, int i) { return T(int(d) * i); } \
inline Value eg_value(Score s) { return Value((int)(unsigned(s) & 0x7fffu) - (int)(unsigned(s) & 0x8000u)); } constexpr T operator/(T d, int i) { return T(int(d) / i); } \
#endif constexpr int operator/(T d1, T d2) { return int(d1) / int(d2); } \
inline T& operator*=(T& d, int i) { return d = T(int(d) * i); } \
inline T& operator/=(T& d, int i) { return d = T(int(d) / i); }
inline Score make_score(int mg, int eg) { return Score((mg << 16) + eg); } ENABLE_FULL_OPERATORS_ON(Value)
ENABLE_FULL_OPERATORS_ON(Depth)
ENABLE_FULL_OPERATORS_ON(Direction)
// Division must be handled separately for each term ENABLE_INCR_OPERATORS_ON(PieceType)
inline Score operator/(Score s, int i) { return make_score(mg_value(s) / i, eg_value(s) / i); } ENABLE_INCR_OPERATORS_ON(Piece)
ENABLE_INCR_OPERATORS_ON(Color)
ENABLE_INCR_OPERATORS_ON(Square)
ENABLE_INCR_OPERATORS_ON(File)
ENABLE_INCR_OPERATORS_ON(Rank)
// Only declared but not defined. We don't want to multiply two scores due to ENABLE_BASE_OPERATORS_ON(Score)
// a very high risk of overflow. So user should explicitly convert to integer.
inline Score operator*(Score s1, Score s2);
// Remaining Score operators are standard #undef ENABLE_FULL_OPERATORS_ON
inline Score operator+ (const Score d1, const Score d2) { return Score(int(d1) + int(d2)); } #undef ENABLE_INCR_OPERATORS_ON
inline Score operator- (const Score d1, const Score d2) { return Score(int(d1) - int(d2)); } #undef ENABLE_BASE_OPERATORS_ON
inline Score operator* (int i, const Score d) { return Score(i * int(d)); }
inline Score operator* (const Score d, int i) { return Score(int(d) * i); }
inline Score operator- (const Score d) { return Score(-int(d)); }
inline void operator+= (Score& d1, const Score d2) { d1 = d1 + d2; }
inline void operator-= (Score& d1, const Score d2) { d1 = d1 - d2; }
inline void operator*= (Score& d, int i) { d = Score(int(d) * i); }
inline void operator/= (Score& d, int i) { d = Score(int(d) / i); }
const Value PawnValueMidgame = Value(0x0C6); /// Additional operators to add integers to a Value
const Value PawnValueEndgame = Value(0x102); constexpr Value operator+(Value v, int i) { return Value(int(v) + i); }
const Value KnightValueMidgame = Value(0x331); constexpr Value operator-(Value v, int i) { return Value(int(v) - i); }
const Value KnightValueEndgame = Value(0x34E); inline Value& operator+=(Value& v, int i) { return v = v + i; }
const Value BishopValueMidgame = Value(0x344); inline Value& operator-=(Value& v, int i) { return v = v - i; }
const Value BishopValueEndgame = Value(0x359);
const Value RookValueMidgame = Value(0x4F6);
const Value RookValueEndgame = Value(0x4FE);
const Value QueenValueMidgame = Value(0x9D9);
const Value QueenValueEndgame = Value(0x9FE);
inline Value value_mate_in(int ply) { /// Additional operators to add a Direction to a Square
inline Square operator+(Square s, Direction d) { return Square(int(s) + int(d)); }
inline Square operator-(Square s, Direction d) { return Square(int(s) - int(d)); }
inline Square& operator+=(Square &s, Direction d) { return s = s + d; }
inline Square& operator-=(Square &s, Direction d) { return s = s - d; }
/// Only declared but not defined. We don't want to multiply two scores due to
/// a very high risk of overflow. So user should explicitly convert to integer.
Score operator*(Score, Score) = delete;
/// Division of a Score must be handled separately for each term
inline Score operator/(Score s, int i) {
return make_score(mg_value(s) / i, eg_value(s) / i);
}
/// Multiplication of a Score by an integer. We check for overflow in debug mode.
inline Score operator*(Score s, int i) {
Score result = Score(int(s) * i);
assert(eg_value(result) == (i * eg_value(s)));
assert(mg_value(result) == (i * mg_value(s)));
assert((i == 0) || (result / i) == s );
return result;
}
constexpr Color operator~(Color c) {
return Color(c ^ BLACK); // Toggle color
}
constexpr Square operator~(Square s) {
return Square(s ^ SQ_A8); // Vertical flip SQ_A1 -> SQ_A8
}
constexpr File operator~(File f) {
return File(f ^ FILE_H); // Horizontal flip FILE_A -> FILE_H
}
constexpr Piece operator~(Piece pc) {
return Piece(pc ^ 8); // Swap color of piece B_KNIGHT -> W_KNIGHT
}
constexpr CastlingRight operator|(Color c, CastlingSide s) {
return CastlingRight(WHITE_OO << ((s == QUEEN_SIDE) + 2 * c));
}
constexpr Value mate_in(int ply) {
return VALUE_MATE - ply; return VALUE_MATE - ply;
} }
inline Value value_mated_in(int ply) { constexpr Value mated_in(int ply) {
return -VALUE_MATE + ply; return -VALUE_MATE + ply;
} }
inline Piece make_piece(Color c, PieceType pt) { constexpr Square make_square(File f, Rank r) {
return Piece((int(c) << 3) | int(pt)); return Square((r << 3) + f);
} }
inline PieceType type_of_piece(Piece p) { constexpr Piece make_piece(Color c, PieceType pt) {
return PieceType(int(p) & 7); return Piece((c << 3) + pt);
} }
inline Color color_of_piece(Piece p) { constexpr PieceType type_of(Piece pc) {
return Color(int(p) >> 3); return PieceType(pc & 7);
} }
inline Color opposite_color(Color c) { inline Color color_of(Piece pc) {
return Color(int(c) ^ 1); assert(pc != NO_PIECE);
return Color(pc >> 3);
} }
inline bool color_is_ok(Color c) { constexpr bool is_ok(Square s) {
return c == WHITE || c == BLACK; return s >= SQ_A1 && s <= SQ_H8;
} }
inline bool piece_type_is_ok(PieceType pt) { constexpr File file_of(Square s) {
return pt >= PAWN && pt <= KING; return File(s & 7);
} }
inline bool piece_is_ok(Piece p) { constexpr Rank rank_of(Square s) {
return piece_type_is_ok(type_of_piece(p)) && color_is_ok(color_of_piece(p)); return Rank(s >> 3);
} }
inline char piece_type_to_char(PieceType pt) { constexpr Square relative_square(Color c, Square s) {
static const char ch[] = " PNBRQK"; return Square(s ^ (c * 56));
return ch[pt];
} }
inline Square make_square(File f, Rank r) { constexpr Rank relative_rank(Color c, Rank r) {
return Square((int(r) << 3) | int(f)); return Rank(r ^ (c * 7));
} }
inline File square_file(Square s) { constexpr Rank relative_rank(Color c, Square s) {
return File(int(s) & 7); return relative_rank(c, rank_of(s));
} }
inline Rank square_rank(Square s) { inline bool opposite_colors(Square s1, Square s2) {
return Rank(int(s) >> 3);
}
inline Square flip_square(Square s) {
return Square(int(s) ^ 56);
}
inline Square flop_square(Square s) {
return Square(int(s) ^ 7);
}
inline Square relative_square(Color c, Square s) {
return Square(int(s) ^ (int(c) * 56));
}
inline Rank relative_rank(Color c, Rank r) {
return Rank(int(r) ^ (int(c) * 7));
}
inline Rank relative_rank(Color c, Square s) {
return relative_rank(c, square_rank(s));
}
inline SquareColor square_color(Square s) {
return SquareColor(int(square_rank(s) + s) & 1);
}
inline bool opposite_color_squares(Square s1, Square s2) {
int s = int(s1) ^ int(s2); int s = int(s1) ^ int(s2);
return ((s >> 3) ^ s) & 1; return ((s >> 3) ^ s) & 1;
} }
inline int file_distance(Square s1, Square s2) { constexpr Direction pawn_push(Color c) {
return abs(square_file(s1) - square_file(s2)); return c == WHITE ? NORTH : SOUTH;
} }
inline int rank_distance(Square s1, Square s2) { constexpr Square from_sq(Move m) {
return abs(square_rank(s1) - square_rank(s2)); return Square((m >> 6) & 0x3F);
} }
inline int square_distance(Square s1, Square s2) { constexpr Square to_sq(Move m) {
return Max(file_distance(s1, s2), rank_distance(s1, s2)); return Square(m & 0x3F);
} }
inline File file_from_char(char c) { constexpr int from_to(Move m) {
return File(c - 'a') + FILE_A; return m & 0xFFF;
} }
inline char file_to_char(File f) { constexpr MoveType type_of(Move m) {
return char(f - FILE_A + int('a')); return MoveType(m & (3 << 14));
} }
inline Rank rank_from_char(char c) { constexpr PieceType promotion_type(Move m) {
return Rank(c - '1') + RANK_1; return PieceType(((m >> 12) & 3) + KNIGHT);
} }
inline char rank_to_char(Rank r) { inline Move make_move(Square from, Square to) {
return char(r - RANK_1 + int('1')); return Move((from << 6) + to);
} }
inline const std::string square_to_string(Square s) { template<MoveType T>
char ch[] = { file_to_char(square_file(s)), rank_to_char(square_rank(s)), 0 }; constexpr Move make(Square from, Square to, PieceType pt = KNIGHT) {
return std::string(ch); return Move(T + ((pt - KNIGHT) << 12) + (from << 6) + to);
} }
inline bool file_is_ok(File f) { constexpr bool is_ok(Move m) {
return f >= FILE_A && f <= FILE_H; return from_sq(m) != to_sq(m); // Catch MOVE_NULL and MOVE_NONE
} }
inline bool rank_is_ok(Rank r) { #endif // #ifndef TYPES_H_INCLUDED
return r >= RANK_1 && r <= RANK_8;
}
inline bool square_is_ok(Square s) {
return s >= SQ_A1 && s <= SQ_H8;
}
inline Square pawn_push(Color c) {
return c == WHITE ? DELTA_N : DELTA_S;
}
#endif // !defined(TYPES_H_INCLUDED)
+268 -199
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -23,225 +24,293 @@
#include <string> #include <string>
#include "evaluate.h" #include "evaluate.h"
#include "misc.h" #include "movegen.h"
#include "move.h"
#include "position.h" #include "position.h"
#include "search.h" #include "search.h"
#include "ucioption.h" #include "thread.h"
#include "tt.h"
#include "timeman.h"
#include "uci.h"
#include "syzygy/tbprobe.h"
using namespace std; using namespace std;
namespace { extern vector<string> setup_bench(const Position&, istream&);
// FEN string for the initial position
const string StartPositionFEN = "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1";
// UCIParser is a class for parsing UCI input. The class
// is actually a string stream built on a given input string.
typedef istringstream UCIParser;
void set_option(UCIParser& up);
void set_position(Position& pos, UCIParser& up);
bool go(Position& pos, UCIParser& up);
void perft(Position& pos, UCIParser& up);
}
/// execute_uci_command() takes a string as input, uses a UCIParser
/// object to parse this text string as a UCI command, and calls
/// the appropriate functions. In addition to the UCI commands,
/// the function also supports a few debug commands.
bool execute_uci_command(const string& cmd) {
static Position pos(StartPositionFEN, false, 0); // The root position
UCIParser up(cmd);
string token;
up >> token; // operator>>() skips any whitespace
if (token == "quit")
return false;
if (token == "go")
return go(pos, up);
if (token == "ucinewgame")
pos.from_fen(StartPositionFEN, false);
else if (token == "isready")
cout << "readyok" << endl;
else if (token == "position")
set_position(pos, up);
else if (token == "setoption")
set_option(up);
else if (token == "perft")
perft(pos, up);
else if (token == "d")
pos.print();
else if (token == "flip")
pos.flip();
else if (token == "eval")
{
read_evaluation_uci_options(pos.side_to_move());
cout << trace_evaluate(pos) << endl;
}
else if (token == "key")
cout << "key: " << hex << pos.get_key()
<< "\nmaterial key: " << pos.get_material_key()
<< "\npawn key: " << pos.get_pawn_key() << endl;
else if (token == "uci")
cout << "id name " << engine_name()
<< "\nid author " << engine_authors()
<< "\n" << Options.print_all()
<< "\nuciok" << endl;
else
cout << "Unknown command: " << cmd << endl;
return true;
}
namespace { namespace {
// set_position() is called when engine receives the "position" UCI // FEN string of the initial position, normal chess
// command. The function sets up the position described in the given const char* StartFEN = "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1";
// fen string ("fen") or the starting position ("startpos") and then
// makes the moves given in the following move list ("moves").
void set_position(Position& pos, UCIParser& up) {
// position() is called when engine receives the "position" UCI command.
// The function sets up the position described in the given FEN string ("fen")
// or the starting position ("startpos") and then makes the moves given in the
// following move list ("moves").
void position(Position& pos, istringstream& is, StateListPtr& states) {
Move m;
string token, fen; string token, fen;
up >> token; // operator>>() skips any whitespace is >> token;
if (token == "startpos") if (token == "startpos")
{ {
pos.from_fen(StartPositionFEN, false); fen = StartFEN;
up >> token; // Consume "moves" token if any is >> token; // Consume "moves" token if any
} }
else if (token == "fen") else if (token == "fen")
{ while (is >> token && token != "moves")
while (up >> token && token != "moves")
fen += token + " "; fen += token + " ";
pos.from_fen(fen, Options["UCI_Chess960"].value<bool>());
}
else return;
// Parse move list (if any)
while (up >> token)
pos.do_setup_move(move_from_uci(pos, token));
}
// set_option() is called when engine receives the "setoption" UCI
// command. The function updates the corresponding UCI option ("name")
// to the given value ("value").
void set_option(UCIParser& up) {
string token, name;
string value = "true"; // UCI buttons don't have a "value" field
up >> token; // Consume "name" token
up >> name; // Read option name
// Handle names with included spaces
while (up >> token && token != "value")
name += " " + token;
up >> value; // Read option value
// Handle values with included spaces
while (up >> token)
value += " " + token;
if (Options.find(name) != Options.end())
Options[name].set_value(value);
else else
cout << "No such option: " << name << endl;
}
// go() is called when engine receives the "go" UCI command. The
// function sets the thinking time and other parameters from the input
// string, and then calls think(). Returns false if a quit command
// is received while thinking, true otherwise.
bool go(Position& pos, UCIParser& up) {
string token;
SearchLimits limits;
Move searchMoves[MAX_MOVES], *cur = searchMoves;
int time[] = { 0, 0 }, inc[] = { 0, 0 };
while (up >> token)
{
if (token == "infinite")
limits.infinite = true;
else if (token == "ponder")
limits.ponder = true;
else if (token == "wtime")
up >> time[WHITE];
else if (token == "btime")
up >> time[BLACK];
else if (token == "winc")
up >> inc[WHITE];
else if (token == "binc")
up >> inc[BLACK];
else if (token == "movestogo")
up >> limits.movesToGo;
else if (token == "depth")
up >> limits.maxDepth;
else if (token == "nodes")
up >> limits.maxNodes;
else if (token == "movetime")
up >> limits.maxTime;
else if (token == "searchmoves")
while (up >> token)
*cur++ = move_from_uci(pos, token);
}
*cur = MOVE_NONE;
limits.time = time[pos.side_to_move()];
limits.increment = inc[pos.side_to_move()];
assert(pos.is_ok());
return think(pos, limits, searchMoves);
}
// perft() is called when engine receives the "perft" command.
// The function calls perft() passing the required search depth
// then prints counted leaf nodes and elapsed time.
void perft(Position& pos, UCIParser& up) {
int depth, time;
int64_t n;
if (!(up >> depth))
return; return;
time = get_system_time(); states = StateListPtr(new std::deque<StateInfo>(1)); // Drop old and create a new one
pos.set(fen, Options["UCI_Chess960"], &states->back(), Threads.main());
n = perft(pos, depth * ONE_PLY); // Parse move list (if any)
while (is >> token && (m = UCI::to_move(pos, token)) != MOVE_NONE)
time = get_system_time() - time; {
states->emplace_back();
std::cout << "\nNodes " << n pos.do_move(m, states->back());
<< "\nTime (ms) " << time }
<< "\nNodes/second " << int(n / (time / 1000.0)) << std::endl;
} }
// setoption() is called when engine receives the "setoption" UCI command. The
// function updates the UCI option ("name") to the given value ("value").
void setoption(istringstream& is) {
string token, name, value;
is >> token; // Consume "name" token
// Read option name (can contain spaces)
while (is >> token && token != "value")
name += string(" ", name.empty() ? 0 : 1) + token;
// Read option value (can contain spaces)
while (is >> token)
value += string(" ", value.empty() ? 0 : 1) + token;
if (Options.count(name))
Options[name] = value;
else
sync_cout << "No such option: " << name << sync_endl;
}
// go() is called when engine receives the "go" UCI command. The function sets
// the thinking time and other parameters from the input string, then starts
// the search.
void go(Position& pos, istringstream& is, StateListPtr& states) {
Search::LimitsType limits;
string token;
bool ponderMode = false;
limits.startTime = now(); // As early as possible!
while (is >> token)
if (token == "searchmoves")
while (is >> token)
limits.searchmoves.push_back(UCI::to_move(pos, token));
else if (token == "wtime") is >> limits.time[WHITE];
else if (token == "btime") is >> limits.time[BLACK];
else if (token == "winc") is >> limits.inc[WHITE];
else if (token == "binc") is >> limits.inc[BLACK];
else if (token == "movestogo") is >> limits.movestogo;
else if (token == "depth") is >> limits.depth;
else if (token == "nodes") is >> limits.nodes;
else if (token == "movetime") is >> limits.movetime;
else if (token == "mate") is >> limits.mate;
else if (token == "perft") is >> limits.perft;
else if (token == "infinite") limits.infinite = 1;
else if (token == "ponder") ponderMode = true;
Threads.start_thinking(pos, states, limits, ponderMode);
}
// bench() is called when engine receives the "bench" command. Firstly
// a list of UCI commands is setup according to bench parameters, then
// it is run one by one printing a summary at the end.
void bench(Position& pos, istream& args, StateListPtr& states) {
string token;
uint64_t num, nodes = 0, cnt = 1;
vector<string> list = setup_bench(pos, args);
num = count_if(list.begin(), list.end(), [](string s) { return s.find("go ") == 0; });
TimePoint elapsed = now();
for (const auto& cmd : list)
{
istringstream is(cmd);
is >> skipws >> token;
if (token == "go")
{
cerr << "\nPosition: " << cnt++ << '/' << num << endl;
go(pos, is, states);
Threads.main()->wait_for_search_finished();
nodes += Threads.nodes_searched();
}
else if (token == "setoption") setoption(is);
else if (token == "position") position(pos, is, states);
else if (token == "ucinewgame") Search::clear();
}
elapsed = now() - elapsed + 1; // Ensure positivity to avoid a 'divide by zero'
dbg_print(); // Just before exiting
cerr << "\n==========================="
<< "\nTotal time (ms) : " << elapsed
<< "\nNodes searched : " << nodes
<< "\nNodes/second : " << 1000 * nodes / elapsed << endl;
}
} // namespace
/// UCI::loop() waits for a command from stdin, parses it and calls the appropriate
/// function. Also intercepts EOF from stdin to ensure gracefully exiting if the
/// GUI dies unexpectedly. When called with some command line arguments, e.g. to
/// run 'bench', once the command is executed the function returns immediately.
/// In addition to the UCI ones, also some additional debug commands are supported.
void UCI::loop(int argc, char* argv[]) {
Position pos;
string token, cmd;
StateListPtr states(new std::deque<StateInfo>(1));
auto uiThread = std::make_shared<Thread>(0);
pos.set(StartFEN, false, &states->back(), uiThread.get());
for (int i = 1; i < argc; ++i)
cmd += std::string(argv[i]) + " ";
do {
if (argc == 1 && !getline(cin, cmd)) // Block here waiting for input or EOF
cmd = "quit";
istringstream is(cmd);
token.clear(); // Avoid a stale if getline() returns empty or blank line
is >> skipws >> token;
// The GUI sends 'ponderhit' to tell us the user has played the expected move.
// So 'ponderhit' will be sent if we were told to ponder on the same move the
// user has played. We should continue searching but switch from pondering to
// normal search. In case Threads.stopOnPonderhit is set we are waiting for
// 'ponderhit' to stop the search, for instance if max search depth is reached.
if ( token == "quit"
|| token == "stop"
|| (token == "ponderhit" && Threads.stopOnPonderhit))
Threads.stop = true;
else if (token == "ponderhit")
Threads.ponder = false; // Switch to normal search
else if (token == "uci")
sync_cout << "id name " << engine_info(true)
<< "\n" << Options
<< "\nuciok" << sync_endl;
else if (token == "setoption") setoption(is);
else if (token == "go") go(pos, is, states);
else if (token == "position") position(pos, is, states);
else if (token == "ucinewgame") Search::clear();
else if (token == "isready") sync_cout << "readyok" << sync_endl;
// Additional custom non-UCI commands, mainly for debugging
else if (token == "flip") pos.flip();
else if (token == "bench") bench(pos, is, states);
else if (token == "d") sync_cout << pos << sync_endl;
else if (token == "eval") sync_cout << Eval::trace(pos) << sync_endl;
else
sync_cout << "Unknown command: " << cmd << sync_endl;
} while (token != "quit" && argc == 1); // Command line args are one-shot
}
/// UCI::value() converts a Value to a string suitable for use with the UCI
/// protocol specification:
///
/// cp <x> The score from the engine's point of view in centipawns.
/// mate <y> Mate in y moves, not plies. If the engine is getting mated
/// use negative values for y.
string UCI::value(Value v) {
assert(-VALUE_INFINITE < v && v < VALUE_INFINITE);
stringstream ss;
if (abs(v) < VALUE_MATE - MAX_PLY)
ss << "cp " << v * 100 / PawnValueEg;
else
ss << "mate " << (v > 0 ? VALUE_MATE - v + 1 : -VALUE_MATE - v) / 2;
return ss.str();
}
/// UCI::square() converts a Square to a string in algebraic notation (g1, a7, etc.)
std::string UCI::square(Square s) {
return std::string{ char('a' + file_of(s)), char('1' + rank_of(s)) };
}
/// UCI::move() converts a Move to a string in coordinate notation (g1f3, a7a8q).
/// The only special case is castling, where we print in the e1g1 notation in
/// normal chess mode, and in e1h1 notation in chess960 mode. Internally all
/// castling moves are always encoded as 'king captures rook'.
string UCI::move(Move m, bool chess960) {
Square from = from_sq(m);
Square to = to_sq(m);
if (m == MOVE_NONE)
return "(none)";
if (m == MOVE_NULL)
return "0000";
if (type_of(m) == CASTLING && !chess960)
to = make_square(to > from ? FILE_G : FILE_C, rank_of(from));
string move = UCI::square(from) + UCI::square(to);
if (type_of(m) == PROMOTION)
move += " pnbrqk"[promotion_type(m)];
return move;
}
/// UCI::to_move() converts a string representing a move in coordinate notation
/// (g1f3, a7a8q) to the corresponding legal Move, if any.
Move UCI::to_move(const Position& pos, string& str) {
if (str.length() == 5) // Junior could send promotion piece in uppercase
str[4] = char(tolower(str[4]));
for (const auto& m : MoveList<LEGAL>(pos))
if (str == UCI::move(m, pos.is_chess960()))
return m;
return MOVE_NONE;
} }
+80
View File
@@ -0,0 +1,80 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef UCI_H_INCLUDED
#define UCI_H_INCLUDED
#include <map>
#include <string>
#include "types.h"
class Position;
namespace UCI {
class Option;
/// Custom comparator because UCI options should be case insensitive
struct CaseInsensitiveLess {
bool operator() (const std::string&, const std::string&) const;
};
/// Our options container is actually a std::map
typedef std::map<std::string, Option, CaseInsensitiveLess> OptionsMap;
/// Option class implements an option as defined by UCI protocol
class Option {
typedef void (*OnChange)(const Option&);
public:
Option(OnChange = nullptr);
Option(bool v, OnChange = nullptr);
Option(const char* v, OnChange = nullptr);
Option(int v, int minv, int maxv, OnChange = nullptr);
Option& operator=(const std::string&);
void operator<<(const Option&);
operator int() const;
operator std::string() const;
private:
friend std::ostream& operator<<(std::ostream&, const OptionsMap&);
std::string defaultValue, currentValue, type;
int min, max;
size_t idx;
OnChange on_change;
};
void init(OptionsMap&);
void loop(int argc, char* argv[]);
std::string value(Value v);
std::string square(Square s);
std::string move(Move m, bool chess960);
std::string pv(const Position& pos, Depth depth, Value alpha, Value beta);
Move to_move(const Position& pos, std::string& str);
} // namespace UCI
extern UCI::OptionsMap Options;
#endif // #ifndef UCI_H_INCLUDED
+102 -108
View File
@@ -1,7 +1,8 @@
/* /*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1 Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author) Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
Copyright (C) 2015-2018 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by it under the terms of the GNU General Public License as published by
@@ -17,154 +18,147 @@
along with this program. If not, see <http://www.gnu.org/licenses/>. along with this program. If not, see <http://www.gnu.org/licenses/>.
*/ */
#include <cctype> #include <algorithm>
#include <iostream> #include <cassert>
#include <sstream> #include <ostream>
#include "misc.h" #include "misc.h"
#include "search.h"
#include "thread.h" #include "thread.h"
#include "ucioption.h" #include "tt.h"
#include "uci.h"
#include "syzygy/tbprobe.h"
using std::string; using std::string;
using std::cout;
using std::endl;
OptionsMap Options; // Global object UCI::OptionsMap Options; // Global object
namespace UCI {
/// 'On change' actions, triggered by an option's value change
void on_clear_hash(const Option&) { Search::clear(); }
void on_hash_size(const Option& o) { TT.resize(o); }
void on_logger(const Option& o) { start_logger(o); }
void on_threads(const Option& o) { Threads.set(o); }
void on_tb_path(const Option& o) { Tablebases::init(o); }
// Our case insensitive less() function as required by UCI protocol /// Our case insensitive less() function as required by UCI protocol
bool CaseInsensitiveLess::operator() (const string& s1, const string& s2) const { bool CaseInsensitiveLess::operator() (const string& s1, const string& s2) const {
int c1, c2; return std::lexicographical_compare(s1.begin(), s1.end(), s2.begin(), s2.end(),
size_t i = 0; [](char c1, char c2) { return tolower(c1) < tolower(c2); });
while (i < s1.size() && i < s2.size())
{
c1 = tolower(s1[i]);
c2 = tolower(s2[i++]);
if (c1 != c2)
return c1 < c2;
}
return s1.size() < s2.size();
} }
// stringify() converts a numeric value of type T to a std::string /// init() initializes the UCI options to their hard-coded default values
template<typename T>
static string stringify(const T& v) {
std::ostringstream ss; void init(OptionsMap& o) {
ss << v;
return ss.str(); // at most 2^32 clusters.
const int MaxHashMB = Is64Bit ? 131072 : 2048;
o["Debug Log File"] << Option("", on_logger);
o["Contempt"] << Option(20, -100, 100);
o["Threads"] << Option(1, 1, 512, on_threads);
o["Hash"] << Option(16, 1, MaxHashMB, on_hash_size);
o["Clear Hash"] << Option(on_clear_hash);
o["Ponder"] << Option(false);
o["MultiPV"] << Option(1, 1, 500);
o["Skill Level"] << Option(20, 0, 20);
o["Move Overhead"] << Option(30, 0, 5000);
o["Minimum Thinking Time"] << Option(20, 0, 5000);
o["Slow Mover"] << Option(89, 10, 1000);
o["nodestime"] << Option(0, 0, 10000);
o["UCI_Chess960"] << Option(false);
o["SyzygyPath"] << Option("<empty>", on_tb_path);
o["SyzygyProbeDepth"] << Option(1, 1, 100);
o["Syzygy50MoveRule"] << Option(true);
o["SyzygyProbeLimit"] << Option(6, 0, 6);
} }
/// OptionsMap c'tor initializes the UCI options to their hard coded default /// operator<<() is used to print all the options default values in chronological
/// values and initializes the default value of "Threads" and "Minimum Split Depth"
/// parameters according to the number of CPU cores.
OptionsMap::OptionsMap() {
OptionsMap& o = *this;
o["Use Search Log"] = UCIOption(false);
o["Search Log Filename"] = UCIOption("SearchLog.txt");
o["Book File"] = UCIOption("book.bin");
o["Best Book Move"] = UCIOption(false);
o["Mobility (Middle Game)"] = UCIOption(100, 0, 200);
o["Mobility (Endgame)"] = UCIOption(100, 0, 200);
o["Passed Pawns (Middle Game)"] = UCIOption(100, 0, 200);
o["Passed Pawns (Endgame)"] = UCIOption(100, 0, 200);
o["Space"] = UCIOption(100, 0, 200);
o["Aggressiveness"] = UCIOption(100, 0, 200);
o["Cowardice"] = UCIOption(100, 0, 200);
o["Minimum Split Depth"] = UCIOption(4, 4, 7);
o["Maximum Number of Threads per Split Point"] = UCIOption(5, 4, 8);
o["Threads"] = UCIOption(1, 1, MAX_THREADS);
o["Use Sleeping Threads"] = UCIOption(false);
o["Hash"] = UCIOption(32, 4, 8192);
o["Clear Hash"] = UCIOption(false, "button");
o["Ponder"] = UCIOption(true);
o["OwnBook"] = UCIOption(true);
o["MultiPV"] = UCIOption(1, 1, 500);
o["Skill Level"] = UCIOption(20, 0, 20);
o["Emergency Move Horizon"] = UCIOption(40, 0, 50);
o["Emergency Base Time"] = UCIOption(200, 0, 30000);
o["Emergency Move Time"] = UCIOption(70, 0, 5000);
o["Minimum Thinking Time"] = UCIOption(20, 0, 5000);
o["UCI_Chess960"] = UCIOption(false);
o["UCI_AnalyseMode"] = UCIOption(false);
// Set some SMP parameters accordingly to the detected CPU count
UCIOption& thr = o["Threads"];
UCIOption& msd = o["Minimum Split Depth"];
thr.defaultValue = thr.currentValue = stringify(cpu_count());
if (cpu_count() >= 8)
msd.defaultValue = msd.currentValue = stringify(7);
}
/// OptionsMap::print_all() returns a string with all the UCI options in chronological
/// insertion order (the idx field) and in the format defined by the UCI protocol. /// insertion order (the idx field) and in the format defined by the UCI protocol.
string OptionsMap::print_all() const { std::ostream& operator<<(std::ostream& os, const OptionsMap& om) {
std::stringstream s; for (size_t idx = 0; idx < om.size(); ++idx)
for (const auto& it : om)
for (size_t i = 0; i <= size(); i++) if (it.second.idx == idx)
for (OptionsMap::const_iterator it = begin(); it != end(); ++it)
if (it->second.idx == i)
{ {
const UCIOption& o = it->second; const Option& o = it.second;
s << "\noption name " << it->first << " type " << o.type; os << "\noption name " << it.first << " type " << o.type;
if (o.type != "button") if (o.type != "button")
s << " default " << o.defaultValue; os << " default " << o.defaultValue;
if (o.type == "spin") if (o.type == "spin")
s << " min " << o.minValue << " max " << o.maxValue; os << " min " << o.min << " max " << o.max;
break; break;
} }
return s.str();
return os;
} }
/// Option class c'tors /// Option class constructors and conversion operators
UCIOption::UCIOption(const char* def) : type("string"), minValue(0), maxValue(0), idx(Options.size()) Option::Option(const char* v, OnChange f) : type("string"), min(0), max(0), on_change(f)
{ defaultValue = currentValue = def; } { defaultValue = currentValue = v; }
UCIOption::UCIOption(bool def, string t) : type(t), minValue(0), maxValue(0), idx(Options.size()) Option::Option(bool v, OnChange f) : type("check"), min(0), max(0), on_change(f)
{ defaultValue = currentValue = (def ? "true" : "false"); } { defaultValue = currentValue = (v ? "true" : "false"); }
UCIOption::UCIOption(int def, int minv, int maxv) : type("spin"), minValue(minv), maxValue(maxv), idx(Options.size()) Option::Option(OnChange f) : type("button"), min(0), max(0), on_change(f)
{ defaultValue = currentValue = stringify(def); } {}
Option::Option(int v, int minv, int maxv, OnChange f) : type("spin"), min(minv), max(maxv), on_change(f)
{ defaultValue = currentValue = std::to_string(v); }
Option::operator int() const {
assert(type == "check" || type == "spin");
return (type == "spin" ? stoi(currentValue) : currentValue == "true");
}
Option::operator std::string() const {
assert(type == "string");
return currentValue;
}
/// set_value() updates currentValue of the Option object. Normally it's up to /// operator<<() inits options and assigns idx in the correct printing order
/// the GUI to check for option's limits, but we could receive the new value
/// directly from the user by teminal window. So let's check the bounds anyway.
void UCIOption::set_value(const string& v) { void Option::operator<<(const Option& o) {
static size_t insert_order = 0;
*this = o;
idx = insert_order++;
}
/// operator=() updates currentValue and triggers on_change() action. It's up to
/// the GUI to check for option's limits, but we could receive the new value from
/// the user by console window, so let's check the bounds anyway.
Option& Option::operator=(const string& v) {
assert(!type.empty()); assert(!type.empty());
if (v.empty()) if ( (type != "button" && v.empty())
return; || (type == "check" && v != "true" && v != "false")
|| (type == "spin" && (stoi(v) < min || stoi(v) > max)))
return *this;
if ((type == "check" || type == "button") != (v == "true" || v == "false")) if (type != "button")
return; currentValue = v;
if (type == "spin") if (on_change)
{ on_change(*this);
int val = atoi(v.c_str());
if (val < minValue || val > maxValue)
return;
}
currentValue = v; return *this;
} }
} // namespace UCI
-85
View File
@@ -1,85 +0,0 @@
/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
Copyright (C) 2008-2010 Marco Costalba, Joona Kiiski, Tord Romstad
Stockfish is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Stockfish is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#if !defined(UCIOPTION_H_INCLUDED)
#define UCIOPTION_H_INCLUDED
#include <cassert>
#include <cstdlib>
#include <map>
#include <string>
class UCIOption {
public:
UCIOption() {} // To be used in a std::map
UCIOption(const char* defaultValue);
UCIOption(bool defaultValue, std::string type = "check");
UCIOption(int defaultValue, int minValue, int maxValue);
void set_value(const std::string& v);
template<typename T> T value() const;
private:
friend class OptionsMap;
std::string defaultValue, currentValue, type;
int minValue, maxValue;
size_t idx;
};
/// Custom comparator because UCI options should not be case sensitive
struct CaseInsensitiveLess {
bool operator() (const std::string&, const std::string&) const;
};
/// Our options container is actually a map with a customized c'tor
class OptionsMap : public std::map<std::string, UCIOption, CaseInsensitiveLess> {
public:
OptionsMap();
std::string print_all() const;
};
extern OptionsMap Options;
/// Option::value() definition and specializations
template<typename T>
T UCIOption::value() const {
assert(type == "spin");
return T(atoi(currentValue.c_str()));
}
template<>
inline std::string UCIOption::value<std::string>() const {
assert(type == "string");
return currentValue;
}
template<>
inline bool UCIOption::value<bool>() const {
assert(type == "check" || type == "button");
return currentValue == "true";
}
#endif // !defined(UCIOPTION_H_INCLUDED)
+123
View File
@@ -0,0 +1,123 @@
#!/bin/bash
# check for errors under valgrind or sanitizers.
error()
{
echo "instrumented testing failed on line $1"
exit 1
}
trap 'error ${LINENO}' ERR
# define suitable post and prefixes for testing options
case $1 in
--valgrind)
echo "valgrind testing started"
prefix=''
exeprefix='valgrind --error-exitcode=42'
postfix='1>/dev/null'
threads="1"
;;
--valgrind-thread)
echo "valgrind-thread testing started"
prefix=''
exeprefix='valgrind --error-exitcode=42'
postfix='1>/dev/null'
threads="2"
;;
--sanitizer-undefined)
echo "sanitizer-undefined testing started"
prefix='!'
exeprefix=''
postfix='2>&1 | grep "runtime error:"'
threads="1"
;;
--sanitizer-thread)
echo "sanitizer-thread testing started"
prefix='!'
exeprefix=''
postfix='2>&1 | grep "WARNING: ThreadSanitizer:"'
threads="2"
cat << EOF > tsan.supp
race:TTEntry::move
race:TTEntry::depth
race:TTEntry::bound
race:TTEntry::save
race:TTEntry::value
race:TTEntry::eval
race:TranspositionTable::probe
race:TranspositionTable::hashfull
EOF
export TSAN_OPTIONS="suppressions=./tsan.supp"
;;
*)
echo "unknown testing started"
prefix=''
exeprefix=''
postfix=''
threads="1"
;;
esac
# simple command line testing
for args in "eval" \
"go nodes 1000" \
"go depth 10" \
"go movetime 1000" \
"go wtime 8000 btime 8000 winc 500 binc 500" \
"bench 128 $threads 10 default depth"
do
echo "$prefix $exeprefix ./stockfish $args $postfix"
eval "$prefix $exeprefix ./stockfish $args $postfix"
done
# more general testing, following an uci protocol exchange
cat << EOF > game.exp
set timeout 10
spawn $exeprefix ./stockfish
send "uci\n"
expect "uciok"
send "setoption name Threads value $threads\n"
send "ucinewgame\n"
send "position startpos\n"
send "go nodes 1000\n"
expect "bestmove"
send "position startpos moves e2e4 e7e6\n"
send "go nodes 1000\n"
expect "bestmove"
send "position fen 5rk1/1K4p1/8/8/3B4/8/8/8 b - - 0 1\n"
send "go depth 30\n"
expect "bestmove"
send "quit\n"
expect eof
# return error code of the spawned program, useful for valgrind
lassign [wait] pid spawnid os_error_flag value
exit \$value
EOF
for exps in game.exp
do
echo "$prefix expect $exps $postfix"
eval "$prefix expect $exps $postfix"
rm $exps
done
rm -f tsan.supp
echo "instrumented testing OK"
Executable
+32
View File
@@ -0,0 +1,32 @@
#!/bin/bash
# verify perft numbers (positions from https://chessprogramming.wikispaces.com/Perft+Results)
error()
{
echo "perft testing failed on line $1"
exit 1
}
trap 'error ${LINENO}' ERR
echo "perft testing started"
cat << EOF > perft.exp
set timeout 10
lassign \$argv pos depth result
spawn ./stockfish
send "position \$pos\\ngo perft \$depth\\n"
expect "Nodes searched? \$result" {} timeout {exit 1}
send "quit\\n"
expect eof
EOF
expect perft.exp startpos 5 4865609 > /dev/null
expect perft.exp "fen r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -" 5 193690690 > /dev/null
expect perft.exp "fen 8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - -" 6 11030083 > /dev/null
expect perft.exp "fen r3k2r/Pppp1ppp/1b3nbN/nP6/BBP1P3/q4N2/Pp1P2PP/R2Q1RK1 w kq - 0 1" 5 15833292 > /dev/null
expect perft.exp "fen rnbq1k1r/pp1Pbppp/2p5/8/2B5/8/PPP1NnPP/RNBQK2R w KQ - 1 8" 5 89941194 > /dev/null
expect perft.exp "fen r4rk1/1pp1qppp/p1np1n2/2b1p1B1/2B1P1b1/P1NP1N2/1PP1QPPP/R4RK1 w - - 0 10" 5 164075551 > /dev/null
rm perft.exp
echo "perft testing OK"
+61
View File
@@ -0,0 +1,61 @@
#!/bin/bash
# verify reproducible search
error()
{
echo "reprosearch testing failed on line $1"
exit 1
}
trap 'error ${LINENO}' ERR
echo "reprosearch testing started"
# repeat two short games, separated by ucinewgame.
# with go nodes $nodes they should result in exactly
# the same node count for each iteration.
cat << EOF > repeat.exp
set timeout 10
spawn ./stockfish
lassign \$argv nodes
send "uci\n"
expect "uciok"
send "ucinewgame\n"
send "position startpos\n"
send "go nodes \$nodes\n"
expect "bestmove"
send "position startpos moves e2e4 e7e6\n"
send "go nodes \$nodes\n"
expect "bestmove"
send "ucinewgame\n"
send "position startpos\n"
send "go nodes \$nodes\n"
expect "bestmove"
send "position startpos moves e2e4 e7e6\n"
send "go nodes \$nodes\n"
expect "bestmove"
send "quit\n"
expect eof
EOF
# to increase the likelyhood of finding a non-reproducible case,
# the allowed number of nodes are varied systematically
for i in `seq 1 20`
do
nodes=$((100*3**i/2**i))
echo "reprosearch testing with $nodes nodes"
# each line should appear exactly an even number of times
expect repeat.exp $nodes 2>&1 | grep -o "nodes [0-9]*" | sort | uniq -c | awk '{if ($1%2!=0) exit(1)}'
done
rm repeat.exp
echo "reprosearch testing OK"
+27
View File
@@ -0,0 +1,27 @@
#!/bin/bash
# obtain and optionally verify Bench / signature
# if no reference is given, the output is deliberately limited to just the signature
error()
{
echo "running bench for signature failed on line $1"
exit 1
}
trap 'error ${LINENO}' ERR
# obtain
signature=`./stockfish bench 2>&1 | grep "Nodes searched : " | awk '{print $4}'`
if [ $# -gt 0 ]; then
# compare to given reference
if [ "$1" != "$signature" ]; then
echo "signature mismatch: reference $1 obtained $signature"
exit 1
else
echo "signature OK: $signature"
fi
else
# just report signature
echo $signature
fi