Dev Builds » 20230822-0914

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 15. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo
ncm-dbt-01 06:54:25 583926 3998 1315 692 1991 +54.58 ± 4.98 2 181 1014 796 6 +111.24 ± 10.62
ncm-dbt-02 06:53:43 586538 4000 1326 715 1959 +53.49 ± 5.07 3 193 1003 792 9 +108.49 ± 10.7
ncm-dbt-03 06:54:13 587692 4002 1358 718 1926 +56.04 ± 5.06 3 184 993 812 9 +114.0 ± 10.75
ncm-dbt-04 06:52:11 570321 4000 1387 761 1852 +54.82 ± 5.05 4 183 1004 801 8 +111.75 ± 10.68
ncm-dbt-05 06:54:34 583423 4000 1367 693 1940 +59.11 ± 5.03 2 175 978 837 8 +120.67 ± 10.84
20000 6753 3579 9668 +55.61 ± 2.25 14 916 4992 4038 40 +113.22 ± 4.79

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
403763 ncm-dbt-03 587622 2 1 0 1 +189.7 ± 55.98 0 0 0 1 0 +1129.65 ± 376.02 ↓
403762 ncm-dbt-01 586139 498 170 83 245 +61.33 ± 13.43 0 17 128 104 0 +126.73 ± 29.75 ↓
403761 ncm-dbt-05 585801 500 185 89 226 +67.54 ± 13.36 0 14 127 108 1 +138.99 ± 29.78 ↓
403760 ncm-dbt-02 587494 500 165 84 251 +56.78 ± 13.91 1 18 131 99 1 +116.78 ± 29.4 ↓
403759 ncm-dbt-03 587834 500 166 90 244 +53.22 ± 14.56 0 27 121 101 1 +107.54 ± 30.99 ↓
403758 ncm-dbt-04 569669 500 169 80 251 +62.51 ± 13.74 0 17 129 102 2 +126.18 ± 29.62 ↓
403757 ncm-dbt-01 583154 500 160 82 258 +54.65 ± 14.49 0 27 118 105 0 +112.14 ± 31.42 ↓
403756 ncm-dbt-05 584117 500 178 89 233 +62.51 ± 14.72 1 22 115 111 1 +129.35 ± 31.83 ↓
403755 ncm-dbt-02 587579 500 160 90 250 +48.96 ± 14.61 0 28 126 94 2 +96.94 ± 30.31 ↓
403754 ncm-dbt-04 570068 500 180 96 224 +58.93 ± 13.72 0 18 132 98 2 +118.33 ± 29.22 ↓
403753 ncm-dbt-03 588600 500 170 93 237 +53.93 ± 15.12 0 30 115 103 2 +107.54 ± 31.86 ↓
403752 ncm-dbt-01 584369 500 157 83 260 +51.8 ± 13.94 0 23 131 95 1 +104.49 ± 29.52 ↓
403751 ncm-dbt-05 587028 500 166 86 248 +56.07 ± 14.01 0 22 127 100 1 +113.68 ± 30.06 ↓
403750 ncm-dbt-02 586858 500 162 93 245 +48.25 ± 14.3 0 27 128 94 1 +96.94 ± 30.02 ↓
403749 ncm-dbt-04 571190 500 174 103 223 +49.67 ± 14.77 1 27 123 98 1 +101.46 ± 30.72 ↓
403748 ncm-dbt-03 587028 500 174 79 247 +66.82 ± 13.77 0 15 128 104 3 +134.15 ± 29.68 ↓
403747 ncm-dbt-01 583614 500 150 82 268 +47.55 ± 14.53 1 26 128 94 1 +96.94 ± 30.02 ↓
403746 ncm-dbt-05 581444 500 161 79 260 +57.5 ± 14.22 0 23 123 103 1 +116.78 ± 30.65 ↓
403745 ncm-dbt-04 571190 500 183 102 215 +56.79 ± 13.77 1 18 130 101 0 +118.33 ± 29.54 ↓
403744 ncm-dbt-02 583112 500 179 74 247 +74.06 ± 14.18 0 17 113 118 2 +152.18 ± 32.04 ↓
403743 ncm-dbt-03 587494 500 169 83 248 +60.36 ± 13.93 0 20 125 104 1 +123.02 ± 30.3 ↓
403742 ncm-dbt-01 583028 500 175 99 226 +53.22 ± 13.15 0 18 138 94 0 +109.07 ± 28.38 ↓
403741 ncm-dbt-05 582527 500 163 100 237 +44.01 ± 14.6 0 31 126 92 1 +88.0 ± 30.34 ↓
403740 ncm-dbt-04 572276 500 170 84 246 +60.36 ± 15.02 0 28 109 112 1 +123.02 ± 32.74 ↓
403739 ncm-dbt-02 586520 500 177 102 221 +52.51 ± 14.11 0 25 125 100 0 +107.54 ± 30.4 ↓
403738 ncm-dbt-03 587579 500 176 101 223 +52.51 ± 14.11 1 22 128 99 0 +109.07 ± 29.94 ↓
403737 ncm-dbt-01 582110 500 159 83 258 +53.22 ± 13.87 0 23 128 99 0 +109.07 ± 29.94 ↓
403736 ncm-dbt-05 582360 500 178 70 252 +76.25 ± 13.51 0 13 117 119 1 +158.93 ± 31.27 ↓
403735 ncm-dbt-04 567244 500 168 100 232 +47.54 ± 14.92 0 32 119 98 1 +95.44 ± 31.3 ↓
403734 ncm-dbt-02 586223 500 155 89 256 +46.13 ± 14.18 0 27 131 91 1 +92.46 ± 29.6 ↓
403733 ncm-dbt-03 587452 500 162 98 240 +44.72 ± 14.24 1 26 131 92 0 +92.46 ± 29.6 ↓
403732 ncm-dbt-01 583028 500 169 89 242 +56.07 ± 14.56 1 23 122 103 1 +115.23 ± 30.81 ↓
403731 ncm-dbt-04 569989 500 167 105 228 +43.31 ± 14.56 2 26 130 92 0 +90.97 ± 29.76 ↓
403730 ncm-dbt-05 579702 500 178 76 246 +71.88 ± 13.96 0 16 118 114 2 +147.19 ± 31.23 ↓
403729 ncm-dbt-03 588302 500 176 95 229 +56.78 ± 14.19 1 20 127 101 1 +116.78 ± 30.03 ↓
403728 ncm-dbt-02 587070 500 164 97 239 +46.84 ± 14.36 0 28 128 93 1 +93.95 ± 30.03 ↓
403727 ncm-dbt-04 570949 500 176 91 233 +59.64 ± 13.47 0 17 132 100 1 +121.46 ± 29.18 ↓
403726 ncm-dbt-01 585970 500 175 91 234 +58.93 ± 14.69 0 24 121 102 3 +116.78 ± 30.96 ↓
403725 ncm-dbt-02 587452 500 164 86 250 +54.65 ± 14.89 2 23 121 103 1 +113.68 ± 30.97 ↓
403724 ncm-dbt-05 584411 500 158 104 238 +37.67 ± 14.87 1 34 125 90 0 +77.71 ± 30.5 ↓
403723 ncm-dbt-03 587325 500 164 79 257 +59.64 ± 14.46 0 24 118 107 1 +121.46 ± 31.39 ↓

Commit

Commit ID c6f62363a657263a567a0cc9bae09f3c4016156d
Author Gian-Carlo Pascutto
Date 2023-08-22 09:14:19 UTC
Simplify Square Clipped ReLU code. Squared numbers are never negative, so barring any wraparound there is no need to clamp to 0. From reading the code, there's no obvious way to get wraparound, so the entire operation can be simplified away. Updated original truncated code comments to be sensible. Verified by running ./stockfish bench 128 1 24 and by the following test: STC: https://tests.stockfishchess.org/tests/view/64da4db95b17f7c21c0eabe7 LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 60224 W: 15425 L: 15236 D: 29563 Ptnml(0-2): 195, 6576, 16382, 6763, 196 closes https://github.com/official-stockfish/Stockfish/pull/4751 No functional change
Copyright 2011–2025 Next Chess Move LLC
Feedback & Support Terms of Service Privacy Policy