Dev Builds » 20230806-1922

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 09:55:19 1192759 3338 1439 312 1587 +122.09 ± 5.36 0 48 486 1095 40 +270.14 ± 15.47
ncm-dbt-02 09:51:15 1199684 3334 1408 302 1624 +119.79 ± 5.32 1 49 491 1095 31 +266.66 ± 15.39
ncm-dbt-03 09:56:13 1229339 3342 1432 280 1630 +124.88 ± 5.34 0 44 475 1108 44 +277.35 ± 15.65
ncm-dbt-04 09:56:21 1225096 3336 1422 291 1623 +122.64 ± 5.32 1 40 495 1091 41 +271.83 ± 15.32
ncm-dbt-05 09:51:16 1225909 3304 1397 283 1624 +121.91 ± 5.41 1 45 486 1079 41 +269.44 ± 15.46
ncm-dbt-06 09:56:11 1237033 3346 1407 269 1670 +123.07 ± 5.25 0 47 472 1123 31 +276.49 ± 15.7
20000 8505 1737 9758 +122.4 ± 2.18 3 273 2905 6591 228 +271.97 ± 6.32

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
199065 ncm-dbt-05 1245023 304 129 31 144 +116.14 ± 17.12 0 3 51 95 3 +254.73 ± 48.09
199064 ncm-dbt-02 1190193 334 141 34 159 +115.36 ± 17.09 0 4 57 101 5 +246.74 ± 45.46
199063 ncm-dbt-01 1169837 338 154 30 154 +133.69 ± 17.29 0 4 44 114 7 +296.15 ± 52.28
199062 ncm-dbt-04 1229627 336 147 27 162 +129.8 ± 14.39 0 1 47 119 1 +307.07 ± 50.15
199061 ncm-dbt-03 1232121 342 153 31 158 +129.63 ± 16.62 0 7 37 125 2 +302.53 ± 56.58
199060 ncm-dbt-06 1222900 346 147 26 173 +126.85 ± 16.55 0 6 43 121 3 +289.41 ± 52.75
199059 ncm-dbt-05 1219363 500 220 35 245 +134.95 ± 12.64 0 3 64 178 5 +315.35 ± 43.03
199058 ncm-dbt-02 1188826 500 218 47 235 +123.81 ± 12.92 0 4 75 167 4 +280.42 ± 39.57
199057 ncm-dbt-03 1250401 500 216 43 241 +125.38 ± 14.13 0 9 65 170 6 +280.42 ± 42.65
199056 ncm-dbt-01 1182401 500 214 49 237 +119.11 ± 13.52 0 4 84 155 7 +258.75 ± 37.2
199055 ncm-dbt-04 1206404 500 211 39 250 +124.6 ± 14.64 0 8 72 160 10 +268.17 ± 40.54
199054 ncm-dbt-06 1250722 500 206 41 253 +119.11 ± 12.99 0 5 78 164 3 +268.17 ± 38.8
199053 ncm-dbt-05 1213364 500 211 34 255 +128.55 ± 12.82 0 5 66 176 3 +298.62 ± 42.41
199052 ncm-dbt-02 1197185 500 213 52 235 +116.0 ± 13.37 1 6 75 167 1 +265.78 ± 39.69
199051 ncm-dbt-04 1222591 500 225 52 223 +125.38 ± 13.96 0 5 76 160 9 +273.0 ± 39.35
199050 ncm-dbt-03 1208163 500 217 47 236 +123.02 ± 14.49 0 6 79 154 11 +261.07 ± 38.58
199049 ncm-dbt-01 1202424 500 212 50 238 +116.77 ± 14.04 0 8 78 158 6 +254.16 ± 38.89
199048 ncm-dbt-06 1257471 500 199 42 259 +112.91 ± 14.05 0 12 71 165 2 +251.89 ± 40.68
199047 ncm-dbt-05 1226261 500 199 53 248 +104.49 ± 15.54 0 14 86 140 10 +211.87 ± 36.97
199046 ncm-dbt-04 1234480 500 209 39 252 +123.02 ± 13.99 0 9 67 169 5 +275.45 ± 42.01
199045 ncm-dbt-02 1198305 500 206 45 249 +116.0 ± 14.53 0 13 67 166 4 +256.44 ± 41.76
199044 ncm-dbt-03 1218536 500 209 39 252 +123.02 ± 13.65 0 5 77 161 7 +270.57 ± 39.07
199043 ncm-dbt-01 1208654 500 215 44 241 +123.81 ± 13.81 0 9 65 172 4 +280.42 ± 42.65
199042 ncm-dbt-06 1237231 500 225 37 238 +137.37 ± 13.15 0 2 67 172 9 +312.48 ± 41.9
199041 ncm-dbt-05 1218451 500 211 40 249 +123.81 ± 14.15 0 6 76 159 9 +268.17 ± 39.39
199040 ncm-dbt-02 1196475 500 207 47 246 +115.22 ± 13.38 0 8 76 164 2 +258.75 ± 39.42
199039 ncm-dbt-03 1222857 500 213 42 245 +123.81 ± 13.98 0 6 75 161 8 +270.57 ± 39.67
199038 ncm-dbt-04 1213710 500 216 42 242 +126.18 ± 14.12 1 6 67 170 6 +285.49 ± 42.08
199037 ncm-dbt-01 1194397 500 208 54 238 +110.6 ± 13.73 0 8 84 154 4 +240.82 ± 37.4
199036 ncm-dbt-06 1226649 500 221 39 240 +132.54 ± 13.83 0 5 67 169 9 +295.94 ± 42.08
199035 ncm-dbt-02 1214207 500 210 40 250 +123.02 ± 14.65 0 10 68 164 8 +268.17 ± 41.66
199034 ncm-dbt-03 1235927 500 211 33 256 +129.35 ± 13.17 0 5 67 173 5 +295.94 ± 42.08
199033 ncm-dbt-05 1240787 500 212 45 243 +120.67 ± 14.35 1 7 72 164 6 +268.17 ± 40.54
199032 ncm-dbt-04 1246041 500 207 51 242 +112.14 ± 13.73 0 7 85 153 5 +243.0 ± 37.13
199031 ncm-dbt-01 1213370 500 220 42 238 +129.35 ± 13.54 0 6 66 172 6 +293.29 ± 42.41
199030 ncm-dbt-06 1217113 500 200 39 261 +116.0 ± 13.71 0 10 71 167 2 +261.07 ± 40.78
199029 ncm-dbt-01 1178234 500 216 43 241 +125.38 ± 14.13 0 9 65 170 6 +280.42 ± 42.65
199028 ncm-dbt-02 1212602 500 213 37 250 +127.76 ± 13.39 0 4 73 166 7 +285.49 ± 40.16
199027 ncm-dbt-05 1218117 500 215 45 240 +123.02 ± 13.65 0 7 71 167 5 +275.45 ± 40.84
199026 ncm-dbt-06 1247150 500 209 45 246 +118.33 ± 13.35 0 7 75 165 3 +265.78 ± 39.69
199025 ncm-dbt-03 1237369 500 213 45 242 +121.45 ± 13.49 0 6 75 164 5 +270.57 ± 39.67
199024 ncm-dbt-04 1222821 500 207 41 252 +119.89 ± 13.16 0 4 81 160 5 +265.78 ± 37.95

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 2011–2024 Next Chess Move LLC