| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Thu Dec 9 20:50:00 2021 +0100 Timestamp: 1639079400 Adjust singular extension depth restriction This patch is a modification of original idea by lonfom169 which had a good yellow run - do singular extension search with depth threshold 6 unless this is a PvNode with is a part of a PV line - for them set threshold to 8 instead. Passed STC https://tests.stockfishchess.org/tests/view/61b1080406b4c2dcb1b1128c LLR: 2.95 (-2.94,2.94) <0.00,2.50> Total: 84352 W: 21917 L: 21555 D: 40880 Elo +1.49 Ptnml(0-2): 288, 9524, 22185, 9896, 283 Passed LTC https://tests.stockfishchess.org/tests/view/61b1860a06b4c2dcb1b134a1 LLR: 2.95 (-2.94,2.94) <0.50,3.00> Total: 63520 W: 16575 L: 16237 D: 30708 Elo +1.85 Ptnml(0-2): 27, 6519, 18350, 6817, 47 https://github.com/official-stockfish/Stockfish/pull/3840 bench 4729473 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Stefan Geschwentner
Date: Wed Dec 8 17:16:17 2021 +0100 Timestamp: 1638980177 Improve transposition table remplacement strategy Increase chance that PV node replaces old entry in transposition table. STC: LLR: 2.93 (-2.94,2.94) <0.00,2.50> Total: 46744 W: 12108 L: 11816 D: 22820 Elo +2.17 Ptnml(0-2): 156, 5221, 12344, 5477, 174 https://tests.stockfishchess.org/tests/view/61ae068356fcf33bce7d99d0 LTC: LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 88464 W: 22912 L: 22513 D: 43039 Elo +1.57 Ptnml(0-2): 84, 9133, 25393, 9544, 78 https://tests.stockfishchess.org/tests/view/61ae973656fcf33bce7db3e1 closes https://github.com/official-stockfish/Stockfish/pull/3839 Bench: 5292488 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Tue Dec 7 18:15:06 2021 +0100 Timestamp: 1638897306 Introduce post-lmr extensions This idea is somewhat similar to extentions in LMR but has a different flavour. If result of LMR was really good - thus exceeded alpha by some pretty big given margin, we can extend move after LMR in full depth search with 0 window. The idea is that this move is probably a fail high with somewhat of a big probability so extending it makes a lot of sense passed STC https://tests.stockfishchess.org/tests/view/61ad45ea56fcf33bce7d74b7 LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 59680 W: 15531 L: 15215 D: 28934 Elo +1.84 Ptnml(0-2): 193, 6711, 15734, 6991, 211 passed LTC https://tests.stockfishchess.org/tests/view/61ad9ff356fcf33bce7d8646 LLR: 2.95 (-2.94,2.94) <0.50,3.00> Total: 59104 W: 15321 L: 14992 D: 28791 Elo +1.93 Ptnml(0-2): 53, 6023, 17065, 6364, 47 closes https://github.com/official-stockfish/Stockfish/pull/3838 bench 4881329 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Tomasz Sobczyk
Date: Tue Dec 7 18:08:54 2021 +0100 Timestamp: 1638896934 Optimize FT activation and affine transform for NEON. This patch optimizes the NEON implementation in two ways. The activation layer after the feature transformer is rewritten to make it easier for the compiler to see through dependencies and unroll. This in itself is a minimal, but a positive improvement. Other architectures could benefit from this too in the future. This is not an algorithmic change. The affine transform for large matrices (first layer after FT) on NEON now utilizes the same optimized code path as >=SSSE3, which makes the memory accesses more sequential and makes better use of the available registers, which allows for code that has longer dependency chains. Benchmarks from Redshift#161, profile-build with apple clang george@Georges-MacBook-Air nets % ./stockfish-b82d93 bench 2>&1 | tail -4 (current master) =========================== Total time (ms) : 2167 Nodes searched : 4667742 Nodes/second : 2154011 george@Georges-MacBook-Air nets % ./stockfish-7377b8 bench 2>&1 | tail -4 (this patch) =========================== Total time (ms) : 1842 Nodes searched : 4667742 Nodes/second : 2534061 This is a solid 18% improvement overall, larger in a bench with NNUE-only, not mixed. Improvement is also observed on armv7-neon (Raspberry Pi, and older phones), around 5% speedup. No changes for architectures other than NEON. closes https://github.com/official-stockfish/Stockfish/pull/3837 No functional changes. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Joost VandeVondele
Date: Tue Dec 7 12:40:48 2021 +0100 Timestamp: 1638877248 Update default net to nn-63376713ba63.nnue. same data set as previous trained nets, tuned the wdl model slightly for training. https://github.com/vondele/nnue-pytorch/tree/wdlTweak1 passed STC: https://tests.stockfishchess.org/tests/view/61abe9e456fcf33bce7d2834 LLR: 2.93 (-2.94,2.94) <0.00,2.50> Total: 31720 W: 8385 L: 8119 D: 15216 Elo +2.91 Ptnml(0-2): 117, 3534, 8273, 3838, 98 passed LTC: https://tests.stockfishchess.org/tests/view/61ac293756fcf33bce7d36cf LLR: 2.96 (-2.94,2.94) <0.50,3.00> Total: 136136 W: 35255 L: 34741 D: 66140 Elo +1.31 Ptnml(0-2): 114, 14217, 38894, 14727, 116 closes https://github.com/official-stockfish/Stockfish/pull/3836 Bench: 4667742 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Mon Dec 6 07:42:04 2021 +0100 Timestamp: 1638772924 Assign extra bonus for previous move that caused a fail low more often This patch allows to assign extra bonus for previous move that caused a fail low not only for PvNodes and cutNodes but also fo some allNodes - namely if the best result we could've got from the search is still far below alpha. passed STC https://tests.stockfishchess.org/tests/view/61aa26a49e8855bba1a36d96 LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 73808 W: 19183 L: 18842 D: 35783 Elo +1.61 Ptnml(0-2): 251, 8257, 19564, 8564, 268 passed LTC https://tests.stockfishchess.org/tests/view/61aa7dc29e8855bba1a3814f LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 142416 W: 36717 L: 36192 D: 69507 Elo +1.28 Ptnml(0-2): 106, 14799, 40862, 15346, 95 closes https://github.com/official-stockfish/Stockfish/pull/3835 bench 4724181 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Stefan Geschwentner
Date: Sun Dec 5 18:13:49 2021 +0100 Timestamp: 1638724429 Tweak history initialization Initialize continuation history with a slighlty negative value -71 instead of zero. The idea is, because the most history entries will be later negative anyway, to shift the starting values a little bit in the "correct" direction. Of course the effect of initialization dimishes with greater depth so I had the apprehension that the LTC test would be difficult to pass, but it passed. STC: LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 34520 W: 9076 L: 8803 D: 16641 Elo +2.75 Ptnml(0-2): 136, 3837, 9047, 4098, 142 https://tests.stockfishchess.org/tests/view/61aa52e39e8855bba1a3776b LTC: LLR: 2.93 (-2.94,2.94) <0.50,3.00> Total: 75568 W: 19620 L: 19254 D: 36694 Elo +1.68 Ptnml(0-2): 44, 7773, 21796, 8115, 56 https://tests.stockfishchess.org/tests/view/61aa87d39e8855bba1a383a5 closes https://github.com/official-stockfish/Stockfish/pull/3834 Bench: 4674029 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Stefan Geschwentner
Date: Sun Dec 5 17:56:54 2021 +0100 Timestamp: 1638723414 Tweak time management Use for adjustment of the falling eval time factor now also the difference between previous best average score and current best score. STC: LLR: 2.95 (-2.94,2.94) <0.00,2.50> Total: 109216 W: 28296 L: 27900 D: 53020 Elo +1.26 Ptnml(0-2): 312, 11759, 30148, 11999, 390 https://tests.stockfishchess.org/tests/view/61aafa8d1b31b85bcfa29d9c LTC: LLR: 2.93 (-2.94,2.94) <0.50,3.00> Total: 54096 W: 14091 L: 13787 D: 26218 Elo +1.95 Ptnml(0-2): 29, 5124, 16447, 5410, 38 https://tests.stockfishchess.org/tests/view/61abbbbd56fcf33bce7d1d64 closes https://github.com/official-stockfish/Stockfish/pull/3833 Bench: 4829419 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: bmc4
Date: Sun Dec 5 17:50:04 2021 +0100 Timestamp: 1638723004 Simplifies bestMoveChanges from LMR As bestMoveChanges is only reset on mainThread and it could change how other threads search, a multi-threads test was made. STC: LLR: 2.94 (-2.94,2.94) <-2.25,0.25> Total: 146776 W: 37934 L: 37941 D: 70901 Elo -0.02 Ptnml(0-2): 477, 15644, 41173, 15597, 497 https://tests.stockfishchess.org/tests/view/61a8f9f34ed77d629d4ea2d6 LTC: LLR: 3.11 (-2.94,2.94) <-2.25,0.25> Total: 114040 W: 29314 L: 29269 D: 55457 Elo +0.14 Ptnml(0-2): 50, 10584, 35722, 10599, 65 https://tests.stockfishchess.org/tests/view/61a9d4bf9e8855bba1a35c4f (SMP, 8 threads) STC: LLR: 2.95 (-2.94,2.94) <-2.25,0.25> Total: 23888 W: 6308 L: 6143 D: 11437 Elo +2.40 Ptnml(0-2): 36, 2557, 6600, 2708, 43 https://tests.stockfishchess.org/tests/view/61ac27a756fcf33bce7d3677 closes https://github.com/official-stockfish/Stockfish/pull/3831 bench: 4829419 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Joost VandeVondele
Date: Sat Dec 4 10:31:22 2021 +0100 Timestamp: 1638610282 Update default net to nn-cdf1785602d6.nnue Same process as in https://github.com/official-stockfish/Stockfish/commit/e4a0c6c75950bf27b6dc32490a1102499643126b with the training started from the current master net. passed STC: LLR: 2.95 (-2.94,2.94) <0.00,2.50> Total: 38224 W: 10023 L: 9742 D: 18459 Elo +2.55 Ptnml(0-2): 133, 4328, 9940, 4547, 164 https://tests.stockfishchess.org/tests/view/61a8611e4ed77d629d4e836e passed LTC: LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 115176 W: 29783 L: 29321 D: 56072 Elo +1.39 Ptnml(0-2): 68, 12039, 32936, 12453, 92 https://tests.stockfishchess.org/tests/view/61a8963e4ed77d629d4e8d9b closes https://github.com/official-stockfish/Stockfish/pull/3830 Bench: 4829419 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Fri Dec 3 08:54:46 2021 +0100 Timestamp: 1638518086 Do more aggressive pruning for some node types This patch allows more aggressive futility/see based pruning for PV nodes with low delta and non-pv nodes. Fixes some white space issues. Passed STC https://tests.stockfishchess.org/tests/view/61a5ed33d16c530b5dcc27cc LLR: 2.95 (-2.94,2.94) <0.00,2.50> Total: 182088 W: 47121 L: 46584 D: 88383 Elo +1.02 Ptnml(0-2): 551, 20687, 48037, 21212, 557 Passed LTC https://tests.stockfishchess.org/tests/view/61a74dfdbd5c4360bcded0ac LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 87136 W: 22494 L: 22103 D: 42539 Elo +1.56 Ptnml(0-2): 38, 8918, 25272, 9295, 45 closes https://github.com/official-stockfish/Stockfish/pull/3828 closes https://github.com/official-stockfish/Stockfish/pull/3829 bench 4332259 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Gian-Carlo Pascutto
Date: Fri Dec 3 08:51:06 2021 +0100 Timestamp: 1638517866 Add AVX-VNNI support for Alder Lake and later. In their infinite wisdom, Intel axed AVX512 from Alder Lake chips (well, not entirely, but we kind of want to use the Gracemont cores for chess!) but still added VNNI support. Confusingly enough, this is not the same as VNNI256 support. This adds a specific AVX-VNNI target that will use this AVX-VNNI mode, by prefixing the VNNI instructions with the appropriate VEX prefix, and avoiding AVX512 usage. This is about 1% faster on P cores: Result of 20 runs ================== base (./clang-bmi2 ) = 3306337 +/- 7519 test (./clang-vnni ) = 3344226 +/- 7388 diff = +37889 +/- 4153 speedup = +0.0115 P(speedup > 0) = 1.0000 But a nice 3% faster on E cores: Result of 20 runs ================== base (./clang-bmi2 ) = 1938054 +/- 28257 test (./clang-vnni ) = 1994606 +/- 31756 diff = +56552 +/- 3735 speedup = +0.0292 P(speedup > 0) = 1.0000 This was measured on Clang 13. GCC 11.2 appears to generate worse code for Alder Lake, though the speedup on the E cores is similar. It is possible to run the engine specifically on the P or E using binding, for example in linux it is possible to use (for an 8 P + 8 E setup like i9-12900K): taskset -c 0-15 ./stockfish taskset -c 16-23 ./stockfish where the first call binds to the P-cores and the second to the E-cores. closes https://github.com/official-stockfish/Stockfish/pull/3824 No functional change see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: bmc4
Date: Wed Dec 1 18:22:44 2021 +0100 Timestamp: 1638379364 Correctly reset bestMoveChanges for searches not using time management (e.g. analysis, fixed node game play etc), bestMoveChanges was not reset during search iterations. As LMR uses this quantity, search was somewhat weaker. Tested using fixed node playing games: ``` ./c-chess-cli -each nodes=10000 option.Hash=16 -engine cmd=../Stockfish/src/fix -engine cmd=../Stockfish/src/master -concurrency 6 -openings file=../books/UHO_XXL_+0.90_+1.19.epd -games 10000 Score of Stockfish Fix vs Stockfish Master: 3187 - 3028 - 3785 [0.508] 10000 ./c-chess-cli -each nodes=30000 option.Hash=16 -engine cmd=../Stockfish/src/fix -engine cmd=../Stockfish/src/master -concurrency 6 -openings file=../books/UHO_XXL_+0.90_+1.19.epd -games 10000 Score of Stockfish Fix vs Stockfish Master: 2946 - 2834 - 4220 [0.506] 10000 ``` closes https://github.com/official-stockfish/Stockfish/pull/3818 bench: 5061979 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: bmc4
Date: Wed Dec 1 18:10:51 2021 +0100 Timestamp: 1638378651 Simplify reduction on rootNode when bestMoveChanges is high The reduction introduced in #3736 also consider on rootNode, so we don't have to reduce again. STC: LLR: 2.94 (-2.94,2.94) <-2.25,0.25> Total: 28736 W: 7494 L: 7329 D: 13913 Elo +1.99 Ptnml(0-2): 95, 3247, 7503, 3444, 79 https://tests.stockfishchess.org/tests/view/61a3abe01b7fdf52228e74d8 LTC: LLR: 2.94 (-2.94,2.94) <-2.25,0.25> Total: 47816 W: 12434 L: 12308 D: 23074 Elo +0.92 Ptnml(0-2): 37, 4972, 13755, 5116, 28 https://tests.stockfishchess.org/tests/view/61a3c3e39f0c43dae1c71d71 closes https://github.com/official-stockfish/Stockfish/pull/3817 bench: 6331638 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Ortmann
Date: Wed Dec 1 18:07:30 2021 +0100 Timestamp: 1638378450 Fix typos in comments, adjust readme closes https://github.com/official-stockfish/Stockfish/pull/3822 also adjusts readme as requested in https://github.com/official-stockfish/Stockfish/pull/3816 No functional change see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: hengyu
Date: Wed Dec 1 17:59:20 2021 +0100 Timestamp: 1638377960 Small clean-up remove unneeded calculation. closes https://github.com/official-stockfish/Stockfish/pull/3807 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: pb00067
Date: Wed Dec 1 17:55:23 2021 +0100 Timestamp: 1638377723 Remove depth dependence and use same limit (2000) as stat_bonus STC: https://tests.stockfishchess.org/tests/view/619df59dc0a4ea18ba95a424 LLR: 2.96 (-2.94,2.94) <-2.25,0.25> Total: 83728 W: 21329 L: 21242 D: 41157 Elo +0.36 Ptnml(0-2): 297, 9669, 21847, 9752, 299 LTC: https://tests.stockfishchess.org/tests/view/619e64d7c0a4ea18ba95a475 LLR: 2.94 (-2.94,2.94) <-2.25,0.25> Total: 79888 W: 20238 L: 20155 D: 39495 Elo +0.36 Ptnml(0-2): 57, 8391, 22980, 8444, 73 closes https://github.com/official-stockfish/Stockfish/pull/3806 bench: 6792010 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: noobpwnftw
Date: Tue Nov 30 20:57:47 2021 +0100 Timestamp: 1638302267 Enable compilation on older Windows systems Improve compatibility of the last NUMA patch when running under older versions of Windows, for instance Windows Server 2003. Reported by user "g3g6" in the following comments: https://github.com/official-stockfish/Stockfish/commit/7218ec4df9fef1146a451b71f0ed3bfd8123c9f9 Closes https://github.com/official-stockfish/Stockfish/pull/3821 No functional change see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Joost VandeVondele
Date: Mon Nov 29 12:56:01 2021 +0100 Timestamp: 1638186961 Update default net to nn-4f56ecfca5b7.nnue New net trained with nnue-pytorch, started from a master net on a data set of Leela (T60.binpack+T74.binpck) Stockfish data (wrongIsRight_nodes5000pv2.binpack), and Michael Babigian's conversion of T60 Leela data (including TB7 rescoring) (farseer.binpack) available as a single interleaved binpack: https://drive.google.com/file/d/1_sQoWBl31WAxNXma2v45004CIVltytP8/view?usp=sharing The nnue-pytorch branch used is https://github.com/vondele/nnue-pytorch/tree/wdl passed STC: https://tests.stockfishchess.org/tests/view/61a3cc729f0c43dae1c71f1b LLR: 2.95 (-2.94,2.94) <0.00,2.50> Total: 49152 W: 12842 L: 12544 D: 23766 Elo +2.11 Ptnml(0-2): 154, 5542, 12904, 5804, 172 passed LTC: https://tests.stockfishchess.org/tests/view/61a43c6260afd064f2d724f1 LLR: 2.96 (-2.94,2.94) <0.50,3.00> Total: 25528 W: 6676 L: 6425 D: 12427 Elo +3.42 Ptnml(0-2): 9, 2593, 7315, 2832, 15 closes https://github.com/official-stockfish/Stockfish/pull/3816 Bench: 6885242 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Sun Nov 28 14:25:06 2021 +0100 Timestamp: 1638105906 Refine futility pruning for parent nodes This patch is a result of refining of tuning vondele did after new net passed and some hand-made values adjustements - excluding changes in other pruning heuristics and rounding value of history divisor to the nearest power of 2. With this patch futility pruning becomes more aggressive and history influence on it is doubled again. passed STC https://tests.stockfishchess.org/tests/view/61a2c4c1a26505c2278c150d LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 33848 W: 8841 L: 8574 D: 16433 Elo +2.74 Ptnml(0-2): 100, 3745, 8988, 3970, 121 passed LTC https://tests.stockfishchess.org/tests/view/61a327ffa26505c2278c26d9 LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 22272 W: 5856 L: 5614 D: 10802 Elo +3.78 Ptnml(0-2): 12, 2230, 6412, 2468, 14 closes https://github.com/official-stockfish/Stockfish/pull/3814 bench 6302543 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Sat Nov 27 14:47:46 2021 +0100 Timestamp: 1638020866 Adjust usage of history in futility pruning This patch refines 0ac8aca893dd2052f8433e0b4a3d65073266b00f that uses history heuristics in futility pruning. Now it adds main history of the move to in and also increases effect by factor of 2. passed STC https://tests.stockfishchess.org/tests/view/61a156829e83391467a2b2c9 LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 68464 W: 17920 L: 17587 D: 32957 Elo +1.69 Ptnml(0-2): 239, 7711, 18025, 7992, 265 passed LTC https://tests.stockfishchess.org/tests/view/61a1bde99e83391467a2b305 LLR: 2.95 (-2.94,2.94) <0.50,3.00> Total: 26088 W: 6926 L: 6674 D: 12488 Elo +3.36 Ptnml(0-2): 18, 2619, 7531, 2845, 31 closes https://github.com/official-stockfish/Stockfish/pull/3812 bench 6804653 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Joost VandeVondele
Date: Sat Nov 27 14:26:35 2021 +0100 Timestamp: 1638019595 Tune NNUE scaling params passed STC: https://tests.stockfishchess.org/tests/view/61a156f89e83391467a2b2cc LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 22816 W: 5896 L: 5646 D: 11274 Elo +3.81 Ptnml(0-2): 55, 2567, 5961, 2723, 102 passed LTC: https://tests.stockfishchess.org/tests/view/61a1cf3d9e83391467a2b30b LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 17904 W: 4658 L: 4424 D: 8822 Elo +4.54 Ptnml(0-2): 6, 1821, 5079, 2025, 21 closes https://github.com/official-stockfish/Stockfish/pull/3811 Bench: 7218806 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Joost VandeVondele
Date: Fri Nov 26 18:16:04 2021 +0100 Timestamp: 1637946964 Update default net to nn-3678835b1d3d.nnue New net trained with nnue-pytorch, started from the master net on a data set of Leela (T60.binpack+T74.binpck) and Stockfish data (wrongIsRight_nodes5000pv2.binpack), available as a single interleaved binpack: https://drive.google.com/file/d/12uWZIA3F2cNbraAzQNb1jgf3tq_6HkTr/view?usp=sharing The nnue-pytorch branch used is https://github.com/vondele/nnue-pytorch/tree/wdl, which has the new feature to filter positions based on the likelihood of the current evaluation leading to the game outcome. It should make it less likely to try to learn from misevaluated positions. Standard options have been used, starting from the master net: --gpus 1 --threads 4 --num-workers 4 --batch-size 16384 --progress_bar_refresh_rate 300 --smart-fen-skipping --random-fen-skipping 12 --features=HalfKAv2_hm^ --lambda=1.0 Testing with games shows neutral Elo at STC, and good performance at LTC: STC: https://tests.stockfishchess.org/tests/view/619eb597c0a4ea18ba95a4dc ELO: -0.44 +-1.8 (95%) LOS: 31.2% Total: 40000 W: 10447 L: 10498 D: 19055 Elo -0.44 Ptnml(0-2): 254, 4576, 10260, 4787, 123 LTC: https://tests.stockfishchess.org/tests/view/619f6e87c0a4ea18ba95a53f ELO: 3.30 +-1.8 (95%) LOS: 100.0% Total: 33062 W: 8560 L: 8246 D: 16256 Elo +3.30 Ptnml(0-2): 54, 3358, 9352, 3754, 13 passed LTC SPRT: https://tests.stockfishchess.org/tests/view/61a0864e8967bbf894416e65 LLR: 2.94 (-2.94,2.94) <0.50,3.00> Total: 29376 W: 7663 L: 7396 D: 14317 Elo +3.16 Ptnml(0-2): 67, 3017, 8205, 3380, 19 closes https://github.com/official-stockfish/Stockfish/pull/3808 Bench: 7011501 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Michael Chaly
Date: Thu Nov 25 19:38:03 2021 +0100 Timestamp: 1637865483 Use fraction of history heuristics in futility pruning This idea is somewhat of a respin of smth we had in futility pruning and that was simplified away - dependence of it not only on static evaluation of position but also on move history heuristics. Instead of aborting it when they are high there we use fraction of their sum to adjust static eval pruning criteria. passed STC https://tests.stockfishchess.org/tests/view/619bd438c0a4ea18ba95a27d LLR: 2.93 (-2.94,2.94) <0.00,2.50> Total: 113704 W: 29284 L: 28870 D: 55550 Elo +1.27 Ptnml(0-2): 357, 12884, 30044, 13122, 445 passed LTC https://tests.stockfishchess.org/tests/view/619cb8f0c0a4ea18ba95a334 LLR: 2.96 (-2.94,2.94) <0.50,3.00> Total: 147136 W: 37307 L: 36770 D: 73059 Elo +1.27 Ptnml(0-2): 107, 15279, 42265, 15804, 113 closes https://github.com/official-stockfish/Stockfish/pull/3805 bench 6777918 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers + AVX2 Windows x64 for modern computers Windows x64 + SSSE3 Windows x64 Linux x64 for Haswell CPUs Linux x64 for modern computers + AVX2 Linux x64 for modern computers Linux x64 + SSSE3 Linux x64 | Author: Stefan Geschwentner
Date: Tue Nov 23 21:23:28 2021 +0100 Timestamp: 1637699008 Less futility pruning. Disable futility pruning at former PV nodes stored in the transposition table. STC: LLR: 2.96 (-2.94,2.94) <0.00,2.50> Total: 102256 W: 25708 L: 25318 D: 51230 Elo +1.33 Ptnml(0-2): 276, 11511, 27168, 11893, 280 https://tests.stockfishchess.org/tests/view/61990b3135c7c6348cb602db LTC: LLR: 2.96 (-2.94,2.94) <0.50,3.00> Total: 183304 W: 46027 L: 45408 D: 91869 Elo +1.17 Ptnml(0-2): 96, 19029, 52778, 19658, 91 https://tests.stockfishchess.org/tests/view/619a0d1b35c7c6348cb603bc closes https://github.com/official-stockfish/Stockfish/pull/3804 Bench: 7334766 see source |