Dev Builds » 20230806-1922

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 06:53:46 583504 4006 1336 711 1959 +54.65 ± 5.06 2 191 999 802 9 +110.81 ± 10.72
ncm-dbt-02 06:53:42 586408 4008 1332 737 1939 +51.96 ± 5.15 4 212 979 803 6 +105.98 ± 10.86
ncm-dbt-03 06:52:46 584602 3988 1337 624 2027 +62.79 ± 4.98 1 159 967 860 7 +128.78 ± 10.89
ncm-dbt-04 06:52:25 568945 4000 1393 722 1885 +58.84 ± 5.0 3 175 971 850 1 +121.65 ± 10.88
ncm-dbt-05 06:52:55 580287 3998 1339 723 1936 +53.96 ± 5.06 2 192 1003 792 10 +109.13 ± 10.69
20000 6737 3517 9746 +56.43 ± 2.26 12 929 4919 4107 33 +115.19 ± 4.83

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
424385 ncm-dbt-01 584034 6 1 1 4 -0.0 ± 173.92 0 1 1 1 0 -0.0 ± 561.2 ↓
424384 ncm-dbt-02 586139 8 3 2 3 +43.58 ± 76.58 0 0 3 1 0 +88.62 ± 172.94 ↓
424383 ncm-dbt-03 586901 488 151 61 276 +64.82 ± 14.51 0 20 116 106 2 +131.21 ± 31.6 ↓
424382 ncm-dbt-05 581028 498 155 95 248 +42.06 ± 14.79 0 32 127 88 2 +82.44 ± 30.21 ↓
424381 ncm-dbt-04 570068 500 176 87 237 +62.51 ± 14.85 1 24 110 115 0 +130.94 ± 32.59 ↓
424380 ncm-dbt-02 585042 500 155 101 244 +37.67 ± 14.48 1 31 131 87 0 +77.71 ± 29.68 ↓
424379 ncm-dbt-01 584411 500 168 95 237 +51.09 ± 14.31 1 22 132 93 2 +102.97 ± 29.38 ↓
424378 ncm-dbt-03 583405 500 170 76 254 +66.1 ± 13.74 0 17 123 109 1 +135.76 ± 30.5 ↓
424377 ncm-dbt-05 576823 500 166 91 243 +52.51 ± 14.52 1 25 122 102 0 +109.07 ± 30.84 ↓
424376 ncm-dbt-04 570228 500 169 83 248 +60.36 ± 13.5 0 18 128 104 0 +124.6 ± 29.8 ↓
424375 ncm-dbt-02 585084 500 174 90 236 +58.93 ± 14.56 0 24 120 104 2 +118.33 ± 31.1 ↓
424374 ncm-dbt-01 582485 500 160 89 251 +49.67 ± 14.64 0 27 128 92 3 +96.94 ± 30.02 ↓
424373 ncm-dbt-03 587113 500 163 83 254 +56.07 ± 14.43 0 25 121 103 1 +113.68 ± 30.97 ↓
424372 ncm-dbt-05 582861 500 161 92 247 +48.25 ± 14.03 1 21 138 88 2 +96.94 ± 28.52 ↓
424371 ncm-dbt-04 569430 500 177 92 231 +59.64 ± 13.32 1 14 134 101 0 +124.6 ± 28.81 ↓
424370 ncm-dbt-01 584958 500 161 94 245 +46.84 ± 14.09 0 27 129 94 0 +95.44 ± 29.88 ↓
424369 ncm-dbt-02 586689 500 163 100 237 +44.01 ± 14.47 0 31 125 94 0 +89.48 ± 30.48 ↓
424368 ncm-dbt-03 580862 500 167 72 261 +66.82 ± 13.77 0 17 122 110 1 +137.37 ± 30.65 ↓
424367 ncm-dbt-05 583070 500 171 85 244 +60.36 ± 14.49 0 25 114 111 0 +124.6 ± 31.99 ↓
424366 ncm-dbt-04 569869 500 174 98 228 +53.22 ± 14.82 0 29 117 103 1 +107.54 ± 31.57 ↓
424365 ncm-dbt-01 583279 500 169 86 245 +58.21 ± 14.11 0 23 121 106 0 +119.89 ± 30.94 ↓
424364 ncm-dbt-02 586604 500 172 90 238 +57.5 ± 14.22 0 23 123 103 1 +116.78 ± 30.65 ↓
424363 ncm-dbt-03 584790 500 174 88 238 +60.36 ± 14.35 1 20 122 106 1 +124.6 ± 30.76 ↓
424362 ncm-dbt-05 573325 500 161 81 258 +56.07 ± 14.7 0 27 117 105 1 +113.68 ± 31.56 ↓
424361 ncm-dbt-04 569151 500 176 80 244 +67.55 ± 13.95 0 19 116 115 0 +140.62 ± 31.61 ↓
424360 ncm-dbt-01 583866 500 168 93 239 +52.51 ± 14.66 0 28 120 101 1 +106.01 ± 31.14 ↓
424359 ncm-dbt-02 587494 500 166 85 249 +56.79 ± 14.99 1 27 112 110 0 +118.33 ± 32.29 ↓
424358 ncm-dbt-03 585464 500 178 85 237 +65.38 ± 13.42 0 15 128 106 1 +134.15 ± 29.68 ↓
424357 ncm-dbt-05 578712 500 177 95 228 +57.5 ± 14.49 0 24 122 102 2 +115.23 ± 30.81 ↓
424356 ncm-dbt-04 565115 500 168 96 236 +50.38 ± 13.72 0 23 132 95 0 +102.97 ± 29.38 ↓
424355 ncm-dbt-01 581693 500 167 82 251 +59.64 ± 14.32 0 23 120 106 1 +121.46 ± 31.09 ↓
424354 ncm-dbt-02 585379 500 185 90 225 +66.82 ± 13.92 0 19 117 114 0 +138.99 ± 31.46 ↓
424353 ncm-dbt-05 583950 500 183 92 225 +63.94 ± 13.66 0 16 129 103 2 +129.35 ± 29.58 ↓
424352 ncm-dbt-03 585970 500 162 76 262 +60.36 ± 13.93 0 21 122 107 0 +124.6 ± 30.76 ↓
424351 ncm-dbt-01 583028 500 182 94 224 +61.79 ± 14.13 1 17 127 103 2 +126.18 ± 29.94 ↓
424350 ncm-dbt-04 569151 500 177 98 225 +55.36 ± 14.12 0 24 123 103 0 +113.68 ± 30.67 ↓
424349 ncm-dbt-02 586435 500 158 96 246 +43.3 ± 14.82 2 27 129 91 1 +89.48 ± 29.91 ↓
424348 ncm-dbt-03 582318 500 172 83 245 +62.51 ± 14.44 0 24 113 113 0 +129.35 ± 32.14 ↓
424347 ncm-dbt-05 582527 500 165 92 243 +51.09 ± 13.76 0 22 134 93 1 +102.97 ± 29.07 ↓
424346 ncm-dbt-04 568554 500 176 88 236 +61.79 ± 14.82 1 24 111 114 0 +129.35 ± 32.44 ↓
424345 ncm-dbt-02 588813 500 156 83 261 +51.09 ± 14.98 0 30 119 99 2 +101.46 ± 31.3 ↓
424344 ncm-dbt-01 583782 500 160 77 263 +58.21 ± 14.11 0 23 121 106 0 +119.89 ± 30.94 ↓

Commit

Commit ID a6d9a302b867a76c3df5b658de6206e77b649a4d
Author AndrovT
Date 2023-08-06 19:22:37 UTC
Implement AffineTransformSparseInput for armv8 Implements AffineTransformSparseInput layer for the NNUE evaluation for the armv8 and armv8-dotprod architectures. We measured some nice speed improvements via 10 runs of our benchmark: armv8, Cortex-X1 : 18.5% speed-up armv8, Cortex-A76 : 13.2% speed-up armv8-dotprod, Cortex-X1 : 27.1% speed-up armv8-dotprod, Cortex-A76 : 12.1% speed-up armv8, Cortex-A72, Raspberry Pi 4 : 8.2% speed-up (thanks Torom!) closes https://github.com/official-stockfish/Stockfish/pull/4719 No functional change
Copyright 2011–2025 Next Chess Move LLC
Feedback & Support Terms of Service Privacy Policy Desktop Version