opelly/Stockfish
1
0
Fork 0
mirror of https://github.com/opelly27/Stockfish.git synced 2026-05-20 10:57:43 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
e5f450cf0bfe5a34dd4ea51a5592a71be4514601
Stockfish/src/nnue
T
History
mstembera dd63b98fb0 Add support for VNNI 
Adds support for Vector Neural Network Instructions (avx512), as available on Intel Cascade Lake

The _mm512_dpbusd_epi32() intrinsic (vpdpbusd instruction) is taylor made for NNUE.

on a cascade lake CPU (AWS C5.24x.large, gcc 10) NNUE eval is at roughly 78% nps of classical
(single core test)

bench 1024 1 24 default depth:
target 	classical 	NNUE 	ratio
vnni 	2207232 	1725987 	78.20
avx512 	2216789 	1671734 	75.41
avx2 	2194006 	1611263 	73.44
modern 	2185001 	1352469 	61.90

closes https://github.com/official-stockfish/Stockfish/pull/2987

No functional change
2020-08-13 07:39:52 +02:00
..
architectures
Add NNUE evaluation
2020-08-06 16:37:45 +02:00
features
Add NNUE evaluation
2020-08-06 16:37:45 +02:00
layers
Add support for VNNI
2020-08-13 07:39:52 +02:00
evaluate_nnue.cpp
Add NNUE evaluation
2020-08-06 16:37:45 +02:00
evaluate_nnue.h
Add NNUE evaluation
2020-08-06 16:37:45 +02:00
nnue_accumulator.h
Cleanup and optimize SSE/AVX code
2020-08-10 14:38:17 +02:00
nnue_architecture.h
Add NNUE evaluation
2020-08-06 16:37:45 +02:00
nnue_common.h
Restrict avx2 hack to windows target
2020-08-11 23:35:02 +02:00
nnue_feature_transformer.h
Provide vectorized NNUE code for SSE2 and MMX targets
2020-08-10 19:17:57 +02:00