Dev Builds » 20211203-0751

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 7. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games Wins Losses Draws Elo
ncm-et-3 08:26:13 1959677 3331 2884 3 444 +456.01 ± 16.17
ncm-et-4 08:26:30 1961211 3335 2912 7 416 +464.69 ± 16.73
ncm-et-9 08:26:08 1966079 3359 2837 8 514 +426.91 ± 15.03
ncm-et-10 08:26:25 1955615 3317 2808 5 504 +430.32 ± 15.17
ncm-et-13 08:26:30 1960339 3333 2853 6 474 +441.74 ± 15.65
ncm-et-15 08:26:26 1955025 3325 2885 2 438 +459.01 ± 16.28
20000 17179 31 2790 +445.91 ± 6.44

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
154363 ncm-et-10 2021-12-04 03:32 00:49:12 1953444 317 259 2 56 +392.3 ± 46.28
154362 ncm-et-15 2021-12-04 03:32 00:49:14 1955722 325 278 0 47 +443.28 ± 50.63
154361 ncm-et-3 2021-12-04 03:32 00:49:35 1968157 331 278 0 53 +424.13 ± 47.49
154360 ncm-et-4 2021-12-04 03:30 00:51:18 1967694 335 298 1 36 +488.37 ± 58.54
154359 ncm-et-13 2021-12-04 03:30 00:51:22 1959091 333 290 0 43 +464.4 ± 53.12
154358 ncm-et-9 2021-12-04 03:27 00:55:04 1969088 359 300 1 58 +416.03 ± 45.42
154357 ncm-et-3 2021-12-04 02:15 01:16:02 1961228 500 428 0 72 +444.08 ± 40.59
154356 ncm-et-15 2021-12-04 02:15 01:16:37 1955279 500 442 0 58 +484.24 ± 45.47
154355 ncm-et-10 2021-12-04 02:14 01:17:20 1949792 500 429 0 71 +446.7 ± 40.89
154354 ncm-et-13 2021-12-04 02:14 01:15:47 1961397 500 409 2 89 +395.65 ± 36.43
154353 ncm-et-4 2021-12-04 02:13 01:16:19 1964320 500 443 0 57 +487.45 ± 45.89
154352 ncm-et-9 2021-12-04 02:11 01:14:31 1965843 500 419 5 76 +410.58 ± 39.55
154351 ncm-et-3 2021-12-04 00:58 01:16:46 1950094 500 435 1 64 +460.32 ± 43.24
154350 ncm-et-4 2021-12-04 00:58 01:15:06 1961414 500 430 2 68 +444.09 ± 41.92
154349 ncm-et-10 2021-12-04 00:57 01:16:33 1958524 500 420 0 80 +424.27 ± 38.4
154348 ncm-et-15 2021-12-04 00:57 01:17:32 1938638 500 437 0 63 +468.95 ± 43.54
154347 ncm-et-13 2021-12-04 00:56 01:17:01 1959543 500 419 1 80 +419.61 ± 38.47
154346 ncm-et-9 2021-12-04 00:56 01:15:09 1966004 500 423 1 76 +429.05 ± 39.52
154345 ncm-et-3 2021-12-03 23:42 01:15:12 1966932 500 432 0 68 +454.76 ± 41.82
154344 ncm-et-15 2021-12-03 23:41 01:14:45 1959853 500 433 0 67 +457.52 ± 42.15
154343 ncm-et-10 2021-12-03 23:41 01:15:52 1953904 500 419 0 81 +421.93 ± 38.15
154342 ncm-et-4 2021-12-03 23:40 01:16:34 1941493 500 435 1 64 +460.32 ± 43.24
154341 ncm-et-9 2021-12-03 23:40 01:15:12 1970013 500 429 1 70 +444.09 ± 41.26
154340 ncm-et-13 2021-12-03 23:40 01:15:47 1962767 500 438 0 62 +471.92 ± 43.9
154339 ncm-et-3 2021-12-03 22:26 01:15:05 1961411 500 437 0 63 +468.95 ± 43.54
154338 ncm-et-10 2021-12-03 22:25 01:14:56 1962613 500 426 2 72 +433.94 ± 40.69
154337 ncm-et-4 2021-12-03 22:24 01:15:21 1959241 500 435 2 63 +457.52 ± 43.63
154336 ncm-et-15 2021-12-03 22:24 01:16:39 1949050 500 429 1 70 +444.09 ± 41.26
154335 ncm-et-13 2021-12-03 22:24 01:15:29 1959252 500 435 1 64 +460.32 ± 43.24
154334 ncm-et-9 2021-12-03 22:23 01:16:40 1955431 500 409 0 91 +399.81 ± 35.89
154333 ncm-et-10 2021-12-03 21:09 01:15:04 1960783 500 428 1 71 +441.5 ± 40.95
154332 ncm-et-3 2021-12-03 21:09 01:16:47 1954351 500 439 1 60 +471.92 ± 44.73
154331 ncm-et-4 2021-12-03 21:09 01:15:07 1964302 500 436 1 63 +463.16 ± 43.6
154330 ncm-et-13 2021-12-03 21:09 01:14:17 1962000 500 434 0 66 +460.32 ± 42.48
154329 ncm-et-9 2021-12-03 21:08 01:13:57 1970334 500 432 0 68 +454.76 ± 41.82
154328 ncm-et-15 2021-12-03 21:07 01:16:10 1960020 500 446 0 54 +497.39 ± 47.22
154327 ncm-et-9 2021-12-03 19:52 01:15:35 1965843 500 425 0 75 +436.43 ± 39.73
154326 ncm-et-3 2021-12-03 19:51 01:16:46 1955569 500 435 1 64 +460.32 ± 43.24
154325 ncm-et-4 2021-12-03 19:51 01:16:45 1970013 500 435 0 65 +463.15 ± 42.83
154324 ncm-et-13 2021-12-03 19:51 01:16:47 1958323 500 428 2 70 +438.95 ± 41.29
154323 ncm-et-10 2021-12-03 19:51 01:17:28 1950246 500 427 0 73 +441.5 ± 40.29
154322 ncm-et-15 2021-12-03 19:51 01:15:29 1966614 500 420 1 79 +421.93 ± 38.73

Commit

Commit ID c9977aa0a89c83bf21651bffd3b6f10c344ccc46
Author Gian-Carlo Pascutto
Date 2021-12-03 07:51:06 UTC
Add AVX-VNNI support for Alder Lake and later. In their infinite wisdom, Intel axed AVX512 from Alder Lake chips (well, not entirely, but we kind of want to use the Gracemont cores for chess!) but still added VNNI support. Confusingly enough, this is not the same as VNNI256 support. This adds a specific AVX-VNNI target that will use this AVX-VNNI mode, by prefixing the VNNI instructions with the appropriate VEX prefix, and avoiding AVX512 usage. This is about 1% faster on P cores: Result of 20 runs ================== base (./clang-bmi2 ) = 3306337 +/- 7519 test (./clang-vnni ) = 3344226 +/- 7388 diff = +37889 +/- 4153 speedup = +0.0115 P(speedup > 0) = 1.0000 But a nice 3% faster on E cores: Result of 20 runs ================== base (./clang-bmi2 ) = 1938054 +/- 28257 test (./clang-vnni ) = 1994606 +/- 31756 diff = +56552 +/- 3735 speedup = +0.0292 P(speedup > 0) = 1.0000 This was measured on Clang 13. GCC 11.2 appears to generate worse code for Alder Lake, though the speedup on the E cores is similar. It is possible to run the engine specifically on the P or E using binding, for example in linux it is possible to use (for an 8 P + 8 E setup like i9-12900K): taskset -c 0-15 ./stockfish taskset -c 16-23 ./stockfish where the first call binds to the P-cores and the second to the E-cores. closes https://github.com/official-stockfish/Stockfish/pull/3824 No functional change
Copyright 2011–2024 Next Chess Move LLC