Dev Builds » 20230612-1841

You are viewing an old NCM Stockfish dev build test. You may find the most recent dev build tests using Stockfish 15 as the baseline here.

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 14. 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 09:52:24 1222445 3332 1408 309 1615 +119.04 ± 5.39 0 59 479 1098 30 +264.39 ± 15.58
ncm-dbt-02 09:52:18 1231541 3328 1401 277 1650 +122.14 ± 5.35 2 42 488 1094 38 +271.69 ± 15.43
ncm-dbt-03 09:56:20 1236964 3338 1432 276 1630 +125.51 ± 5.28 0 42 469 1118 40 +280.84 ± 15.74
ncm-dbt-04 09:55:33 1230787 3338 1402 299 1637 +119.28 ± 5.33 0 45 515 1070 39 +261.95 ± 15.01
ncm-dbt-05 09:56:14 1231044 3322 1393 283 1646 +120.73 ± 5.16 1 28 527 1070 35 +268.05 ± 14.79
ncm-dbt-06 09:53:15 1230188 3342 1410 283 1649 +121.93 ± 5.34 0 47 489 1096 39 +270.03 ± 15.42
20000 8446 1727 9827 +121.44 ± 2.17 3 263 2967 6546 221 +269.43 ± 6.25

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
192647 ncm-dbt-02 1225415 328 135 30 163 +115.27 ± 16.98 0 6 49 107 2 +256.47 ± 49.35
192646 ncm-dbt-05 1221883 322 136 19 167 +132.28 ± 15.8 0 1 46 110 4 +302.6 ± 50.68
192645 ncm-dbt-01 1178377 332 147 26 159 +132.72 ± 16.54 0 5 38 120 3 +308.83 ± 56.29
192644 ncm-dbt-03 1237651 338 138 29 171 +116.19 ± 16.03 0 5 51 112 1 +262.86 ± 48.35
192643 ncm-dbt-04 1223860 338 145 29 164 +124.28 ± 16.86 0 5 47 113 4 +277.13 ± 50.47
192642 ncm-dbt-06 1244952 342 147 30 165 +123.85 ± 18.2 0 7 47 110 7 +265.35 ± 50.35
192635 ncm-dbt-03 1262611 500 221 43 236 +129.35 ± 13.9 0 6 68 168 8 +288.06 ± 41.76
192634 ncm-dbt-05 1213689 500 205 44 251 +116.0 ± 13.2 0 5 83 158 4 +256.44 ± 37.51
192633 ncm-dbt-02 1231649 500 208 40 252 +121.45 ± 14.01 0 7 75 161 7 +265.78 ± 39.69
192632 ncm-dbt-01 1239800 500 201 46 253 +111.37 ± 14.22 0 10 80 155 5 +240.82 ± 38.39
192631 ncm-dbt-04 1224699 500 206 43 251 +117.55 ± 12.82 0 6 76 167 1 +268.17 ± 39.39
192630 ncm-dbt-06 1235405 500 215 43 242 +124.6 ± 13.45 0 5 74 165 6 +277.93 ± 39.92
192365 ncm-dbt-05 1235316 500 203 44 253 +114.45 ± 13.21 0 3 91 150 6 +247.41 ± 35.47
192364 ncm-dbt-03 1237670 500 214 39 247 +126.97 ± 12.66 0 2 76 167 5 +288.06 ± 39.11
192363 ncm-dbt-04 1238896 500 205 40 255 +119.11 ± 14.03 0 8 75 161 6 +261.07 ± 39.69
192362 ncm-dbt-02 1229592 500 221 42 237 +130.15 ± 14.57 1 5 68 166 10 +288.06 ± 41.76
192361 ncm-dbt-06 1220975 500 208 36 256 +124.6 ± 13.97 0 6 74 162 8 +273.0 ± 39.95
192360 ncm-dbt-01 1187910 500 216 41 243 +126.97 ± 12.85 0 4 71 171 4 +290.66 ± 40.76
192353 ncm-dbt-05 1237652 500 211 47 242 +118.33 ± 13.52 0 5 82 157 6 +258.75 ± 37.76
192352 ncm-dbt-04 1221573 500 198 54 248 +102.97 ± 13.38 0 6 98 142 4 +219.87 ± 34.26
192351 ncm-dbt-02 1236871 500 208 38 254 +123.02 ± 13.11 0 5 74 167 4 +277.93 ± 39.92
192350 ncm-dbt-03 1227314 500 209 47 244 +116.77 ± 13.71 0 7 79 159 5 +256.44 ± 38.62
192349 ncm-dbt-06 1234201 500 204 45 251 +114.45 ± 13.72 0 8 79 159 4 +251.89 ± 38.63
192348 ncm-dbt-01 1222434 500 204 42 254 +116.77 ± 13.36 0 6 80 160 4 +258.75 ± 38.32
192341 ncm-dbt-05 1221867 500 208 49 243 +114.45 ± 13.03 1 5 78 166 0 +263.42 ± 38.85
192340 ncm-dbt-04 1222163 500 222 46 232 +127.76 ± 13.75 0 8 63 174 5 +290.66 ± 43.37
192339 ncm-dbt-03 1220477 500 216 34 250 +132.54 ± 13.47 0 6 62 176 6 +304.07 ± 43.81
192338 ncm-dbt-02 1222348 500 217 35 248 +132.54 ± 14.01 1 6 59 178 6 +306.84 ± 44.88
192337 ncm-dbt-01 1247549 500 214 49 237 +119.11 ± 14.84 0 11 71 160 8 +256.44 ± 40.73
192336 ncm-dbt-06 1218973 500 206 42 252 +118.33 ± 14.03 0 11 67 169 3 +265.78 ± 41.9
192329 ncm-dbt-05 1222021 500 217 43 240 +126.17 ± 13.95 0 5 75 161 9 +275.45 ± 39.63
192328 ncm-dbt-02 1240380 500 206 48 246 +113.68 ± 14.05 0 9 79 157 5 +247.41 ± 38.64
192327 ncm-dbt-04 1246261 500 212 46 242 +119.89 ± 14.51 0 9 74 159 8 +258.75 ± 39.96
192326 ncm-dbt-01 1241978 500 215 50 235 +119.11 ± 13.34 0 9 68 172 1 +273.0 ± 41.7
192325 ncm-dbt-03 1234224 500 217 41 242 +127.76 ± 13.75 0 7 66 171 6 +288.06 ± 42.4
192324 ncm-dbt-06 1231720 500 217 43 240 +126.17 ± 13.42 0 6 69 170 5 +285.49 ± 41.44
192317 ncm-dbt-05 1264881 500 213 37 250 +127.76 ± 13.21 0 4 72 168 6 +288.06 ± 40.46
192316 ncm-dbt-01 1239067 500 211 55 234 +112.14 ± 14.85 0 14 71 160 5 +243.0 ± 40.56
192315 ncm-dbt-02 1234534 500 206 44 250 +116.77 ± 13.01 0 4 84 158 4 +258.75 ± 37.2
192314 ncm-dbt-03 1238802 500 217 43 240 +126.17 ± 14.62 0 9 67 165 9 +275.45 ± 42.01
192313 ncm-dbt-06 1225092 500 213 44 243 +122.24 ± 13.3 0 4 79 161 6 +270.57 ± 38.47
192312 ncm-dbt-04 1238058 500 214 41 245 +125.38 ± 13.96 0 3 82 154 11 +268.17 ± 37.61

Commit

Commit ID 38e61663d836e062af0bc002814ad5149c4b7729
Author AndrovT
Date 2023-06-12 18:41:27 UTC
Use block sparse input for the first layer. Use block sparse input for the first fully connected layer on architectures with at least SSSE3. Depending on the CPU architecture, this yields a speedup of up to 10%, e.g. ``` Result of 100 runs of 'bench 16 1 13 default depth NNUE' base (...ockfish-base) = 959345 +/- 7477 test (...ckfish-patch) = 1054340 +/- 9640 diff = +94995 +/- 3999 speedup = +0.0990 P(speedup > 0) = 1.0000 CPU: 8 x AMD Ryzen 7 5700U with Radeon Graphics Hyperthreading: on ``` Passed STC: https://tests.stockfishchess.org/tests/view/6485aa0965ffe077ca12409c LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 8864 W: 2479 L: 2223 D: 4162 Ptnml(0-2): 13, 829, 2504, 1061, 25 This commit includes a net with reordered weights, to increase the likelihood of block sparse inputs, but otherwise equivalent to the previous master net (nn-ea57bea57e32.nnue). Activation data collected with https://github.com/AndrovT/Stockfish/tree/log-activations, running bench 16 1 13 varied_1000.epd depth NNUE on this data. Net parameters permuted with https://gist.github.com/AndrovT/9e3fbaebb7082734dc84d27e02094cb3. closes https://github.com/official-stockfish/Stockfish/pull/4612 No functional change
Copyright 2011–2024 Next Chess Move LLC