Dev Builds » 20230223-1222

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 07:01:33 583717 4000 1223 865 1912 +31.18 ± 5.14 2 292 1056 646 4 +62.51 ± 10.44
ncm-dbt-02 07:01:11 585621 3998 1158 755 2085 +35.14 ± 5.11 6 258 1066 665 4 +71.38 ± 10.37
ncm-dbt-03 07:02:08 585440 4004 1247 872 1885 +32.64 ± 5.12 5 269 1082 638 8 +65.32 ± 10.28
ncm-dbt-04 07:00:15 569357 4000 1225 831 1944 +34.33 ± 5.13 3 275 1050 669 3 +69.35 ± 10.47
ncm-dbt-05 07:01:52 581433 3998 1212 864 1922 +30.32 ± 5.22 8 289 1055 641 6 +61.46 ± 10.44
20000 6065 4187 9748 +32.72 ± 2.3 24 1383 5309 3259 25 +66.0 ± 4.65

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
411962 ncm-dbt-03 585885 4 2 0 2 +190.27 ± 27.79 0 0 0 2 0 +1199.83 ± 312.71 ↓
411961 ncm-dbt-05 582152 498 147 98 253 +34.3 ± 14.45 1 30 139 77 2 +67.82 ± 28.58 ↓
411960 ncm-dbt-02 587664 498 151 97 250 +37.82 ± 15.19 0 39 117 93 0 +76.56 ± 31.57 ↓
411959 ncm-dbt-01 584958 500 155 102 243 +36.97 ± 14.57 1 31 133 84 1 +74.79 ± 29.41 ↓
411958 ncm-dbt-04 570268 500 144 115 241 +20.17 ± 15.33 1 46 127 75 1 +40.48 ± 30.28 ↓
411957 ncm-dbt-03 585548 500 155 101 244 +37.67 ± 14.22 1 29 135 85 0 +77.71 ± 29.11 ↓
411956 ncm-dbt-05 578712 500 150 107 243 +29.95 ± 14.76 0 40 127 83 0 +60.36 ± 30.26 ↓
411955 ncm-dbt-02 585843 500 144 98 258 +32.05 ± 14.51 1 33 136 79 1 +64.66 ± 29.04 ↓
411954 ncm-dbt-04 570188 500 143 96 261 +32.76 ± 14.82 0 39 125 86 0 +66.1 ± 30.52 ↓
411953 ncm-dbt-01 584916 500 151 100 249 +35.56 ± 14.22 0 32 136 81 1 +70.44 ± 29.01 ↓
411952 ncm-dbt-03 586223 500 169 99 232 +48.96 ± 14.47 2 21 134 91 2 +99.95 ± 29.1 ↓
411951 ncm-dbt-05 578588 500 145 104 251 +28.55 ± 14.92 1 38 131 79 1 +57.5 ± 29.74 ↓
411950 ncm-dbt-02 588260 500 142 91 267 +35.56 ± 15.12 1 36 125 87 1 +71.89 ± 30.51 ↓
411949 ncm-dbt-04 569151 500 158 98 244 +41.89 ± 14.74 0 34 122 94 0 +85.04 ± 30.9 ↓
411948 ncm-dbt-03 584327 500 143 106 251 +25.76 ± 14.85 1 38 136 73 2 +50.38 ± 29.09 ↓
411947 ncm-dbt-01 583531 500 155 99 246 +39.08 ± 15.21 0 39 116 95 0 +79.17 ± 31.71 ↓
411946 ncm-dbt-02 582402 500 151 90 259 +42.6 ± 12.84 0 20 149 81 0 +86.52 ± 26.94 ↓
411945 ncm-dbt-05 580323 500 155 104 241 +35.56 ± 14.74 2 31 131 86 0 +74.79 ± 29.69 ↓
411944 ncm-dbt-04 568792 500 151 102 247 +34.16 ± 14.52 0 34 135 79 2 +66.1 ± 29.17 ↓
411943 ncm-dbt-03 581070 500 153 111 236 +29.25 ± 14.06 0 35 138 77 0 +58.93 ± 28.79 ↓
411942 ncm-dbt-01 585295 500 143 120 237 +15.99 ± 15.51 0 54 119 77 0 +32.05 ± 31.27 ↓
411941 ncm-dbt-02 586266 500 137 92 271 +31.35 ± 13.52 0 30 145 75 0 +63.23 ± 27.78 ↓
411940 ncm-dbt-05 579455 500 152 111 237 +28.56 ± 14.41 2 32 139 77 0 +60.36 ± 28.64 ↓
411939 ncm-dbt-04 567165 500 158 106 236 +36.27 ± 13.44 1 24 147 78 0 +74.79 ± 27.38 ↓
411938 ncm-dbt-03 586901 500 132 116 252 +11.12 ± 14.75 0 50 135 64 1 +20.87 ± 29.27 ↓
411937 ncm-dbt-01 583698 500 154 108 238 +32.05 ± 13.71 0 31 142 77 0 +64.66 ± 28.2 ↓
411936 ncm-dbt-02 584916 500 158 84 258 +51.8 ± 14.75 1 26 122 100 1 +106.01 ± 30.85 ↓
411935 ncm-dbt-05 583614 500 150 114 236 +25.06 ± 15.06 2 39 130 79 0 +53.22 ± 29.88 ↓
411934 ncm-dbt-04 570829 500 162 108 230 +37.67 ± 14.08 1 28 137 84 0 +77.71 ± 28.83 ↓
411933 ncm-dbt-03 585801 500 176 114 210 +43.3 ± 14.02 0 28 132 90 0 +88.0 ± 29.49 ↓
411932 ncm-dbt-01 584285 500 155 117 228 +26.46 ± 15.03 1 39 133 75 2 +51.8 ± 29.48 ↓
411931 ncm-dbt-02 584495 500 135 88 277 +32.76 ± 13.9 0 32 139 79 0 +66.1 ± 28.61 ↓
411930 ncm-dbt-05 582444 500 158 116 226 +29.25 ± 14.84 0 40 129 80 1 +57.5 ± 30.0 ↓
411929 ncm-dbt-04 570509 500 162 104 234 +40.48 ± 14.52 0 33 126 91 0 +82.1 ± 30.36 ↓
411928 ncm-dbt-01 581693 500 161 120 219 +28.55 ± 14.14 0 36 137 77 0 +57.5 ± 28.93 ↓
411927 ncm-dbt-03 585548 500 164 111 225 +36.97 ± 14.17 0 31 136 82 1 +73.34 ± 29.0 ↓
411926 ncm-dbt-02 585126 500 140 115 245 +17.39 ± 15.37 3 42 133 71 1 +37.67 ± 29.51 ↓
411925 ncm-dbt-05 586181 500 155 110 235 +31.35 ± 14.98 0 39 129 80 2 +60.36 ± 30.0 ↓
411924 ncm-dbt-04 567958 500 147 102 251 +31.35 ± 14.47 0 37 131 82 0 +63.23 ± 29.72 ↓
411923 ncm-dbt-03 587664 500 153 114 233 +27.16 ± 14.83 1 37 136 74 2 +53.22 ± 29.08 ↓
411922 ncm-dbt-01 581361 500 149 99 252 +34.86 ± 13.76 0 30 140 80 0 +70.44 ± 28.44 ↓

Commit

Commit ID b4ad3a3c4b68f9c8736f444aeb3364f833247fdc
Author Sebastian Buchwald
Date 2023-02-23 12:22:03 UTC
Add support for ARM dot product instructions The sdot instruction computes (and accumulates) a signed dot product, which is quite handy for Stockfish's NNUE code. The instruction is optional for Armv8.2 and Armv8.3, and mandatory for Armv8.4 and above. The commit adds a new 'arm-dotprod' architecture with enabled dot product support. It also enables dot product support for the existing 'apple-silicon' architecture, which is at least Armv8.5. The following local speed test was performed on an Apple M1 with ARCH=apple-silicon. I had to remove CPU pinning from the benchmark script. However, the results were still consistent: Checking both binaries against themselves reported a speedup of +0.0000 and +0.0005, respectively. ``` Result of 100 runs ================== base (...ish.037ef3e1) = 1917997 +/- 7152 test (...fish.dotprod) = 2159682 +/- 9066 diff = +241684 +/- 2923 speedup = +0.1260 P(speedup > 0) = 1.0000 CPU: 10 x arm Hyperthreading: off ``` Fixes #4193 closes https://github.com/official-stockfish/Stockfish/pull/4400 No functional change
Copyright 2011–2025 Next Chess Move LLC
Feedback & Support Terms of Service Privacy Policy Desktop Version