opelly/Stockfish
1
0
Fork 0
mirror of https://github.com/opelly27/Stockfish.git synced 2026-05-20 15:37:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

Use NNUE complexity in search, retune related parameters

Browse Source 
This builds on ideas of xoto10 and mstembera to use more output from NNUE in the search algorithm.

passed STC:
https://tests.stockfishchess.org/tests/view/62ae454fe7ee5525ef88a957
LLR: 2.95 (-2.94,2.94) <0.00,2.50>
Total: 89208 W: 24127 L: 23753 D: 41328
Ptnml(0-2): 400, 9886, 23642, 10292, 384

passed LTC:
https://tests.stockfishchess.org/tests/view/62acc6ddd89eb6cf1e0750a1
LLR: 2.93 (-2.94,2.94) <0.50,3.00>
Total: 56352 W: 15430 L: 15115 D: 25807
Ptnml(0-2): 44, 5501, 16782, 5794, 55

closes https://github.com/official-stockfish/Stockfish/pull/4088

bench 5332964
This commit is contained in:
Dubslow
2022-06-10 08:11:02 -05:00
committed by Joost VandeVondele
parent 5304b561ab
commit 442c40b43d
5 changed files with 38 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/evaluate.cpp
+20 -12
View File
@@ -1080,35 +1080,39 @@ make_v:
/// evaluate() is the evaluator for the outer world. It returns a static
/// evaluation of the position from the point of view of the side to move.
Value Eval::evaluate(const Position& pos) {
Value Eval::evaluate(const Position& pos, int* complexity) {
Value v;
Color stm = pos.side_to_move();
Value psq = pos.psq_eg_stm();
// Deciding between classical and NNUE eval (~10 Elo): for high PSQ imbalance we use classical,
// but we switch to NNUE during long shuffling or with high material on the board.
bool useClassical = (pos.this_thread()->depth > 9 || pos.count<ALL_PIECES>() > 7) &&
abs(eg_value(pos.psq_score())) * 5 > (856 + pos.non_pawn_material() / 64) * (10 + pos.rule50_count());
bool useClassical = (pos.this_thread()->depth > 9 || pos.count<ALL_PIECES>() > 7)
&& abs(psq) * 5 > (856 + pos.non_pawn_material() / 64) * (10 + pos.rule50_count());
// Deciding between classical and NNUE eval (~10 Elo): for high PSQ imbalance we use classical,
// but we switch to NNUE during long shuffling or with high material on the board.
if (!useNNUE || useClassical)
{
v = Evaluation<NO_TRACE>(pos).value(); // classical
v = Evaluation<NO_TRACE>(pos).value();
useClassical = abs(v) >= 297;
}
// If result of a classical evaluation is much lower than threshold fall back to NNUE
if (useNNUE && !useClassical)
{
int complexity;
int scale = 1048 + 109 * pos.non_pawn_material() / 5120;
Color stm = pos.side_to_move();
int nnueComplexity;
int scale = 1092 + 106 * pos.non_pawn_material() / 5120;
Value optimism = pos.this_thread()->optimism[stm];
Value psq = (stm == WHITE ? 1 : -1) * eg_value(pos.psq_score());
Value nnue = NNUE::evaluate(pos, true, &complexity); // NNUE
complexity = (137 * complexity + 137 * abs(nnue - psq)) / 256;
optimism = optimism * (255 + complexity) / 256;
v = (nnue * scale + optimism * (scale - 848)) / 1024;
Value nnue = NNUE::evaluate(pos, true, &nnueComplexity);
// Blend nnue complexity with (semi)classical complexity
nnueComplexity = (104 * nnueComplexity + 131 * abs(nnue - psq)) / 256;
if (complexity) // Return hybrid NNUE complexity to caller
*complexity = nnueComplexity;
optimism = optimism * (269 + nnueComplexity) / 256;
v = (nnue * scale + optimism * (scale - 754)) / 1024;
if (pos.is_chess960())
v += fix_FRC(pos);
@@ -1120,6 +1124,10 @@ Value Eval::evaluate(const Position& pos) {
// Guarantee evaluation does not hit the tablebase range
v = std::clamp(v, VALUE_TB_LOSS_IN_MAX_PLY + 1, VALUE_TB_WIN_IN_MAX_PLY - 1);
// When not using NNUE, return classical complexity to caller
if (complexity && (!useNNUE || useClassical))
*complexity = abs(v - psq);
return v;
}