Open
Conversation
Passed STC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 21856 W: 5709 L: 5471 D: 10676 Ptnml(0-2): 77, 2466, 5589, 2734, 62 https://tests.stockfishchess.org/tests/view/68862ea4966ed85face24a27 Passed LTC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 108828 W: 27893 L: 27763 D: 53172 Ptnml(0-2): 59, 11860, 30435, 12012, 48 https://tests.stockfishchess.org/tests/view/68863046966ed85face24a3b closes #6181 Bench: 3069051
closes #6182 No functional change.
As this is tried only for cutNode, and all children of cutNodes will be nonCutNodes, the guard condition is redundant. Simplification just removes the unnecessary condition (always true). closes #6187 No functional change
Passed simplification STC LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 254464 W: 65774 L: 65793 D: 122897 Ptnml(0-2): 699, 30087, 65638, 30150, 658 https://tests.stockfishchess.org/tests/view/68863283966ed85face24a59 Passed simplification LTC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 144750 W: 37111 L: 37018 D: 70621 Ptnml(0-2): 79, 15676, 40764, 15785, 71 https://tests.stockfishchess.org/tests/view/6886655f7b562f5f7b731359 closes #6189 Bench: 3071078
This patch simplifies the maximum depth formula that reductions in LMR can allow. Passed STC LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 35616 W: 9358 L: 9139 D: 17119 Ptnml(0-2): 101, 4051, 9278, 4284, 94 https://tests.stockfishchess.org/tests/view/688608cf966ed85face24882 Passed LTC LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 22284 W: 5833 L: 5613 D: 10838 Ptnml(0-2): 8, 2367, 6173, 2585, 9 https://tests.stockfishchess.org/tests/view/68860d7f966ed85face248d5 closes #6191 Bench: 2902492
about a 0.7% speedup for the x86-64-avx512icl ARCH closes #6186 No functional change
Passed STC: LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 150240 W: 39305 L: 38819 D: 72116 Ptnml(0-2): 536, 17541, 38494, 17999, 550 https://tests.stockfishchess.org/tests/view/6887ce577b562f5f7b731a85 Passed LTC: LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 159120 W: 41025 L: 40461 D: 77634 Ptnml(0-2): 96, 16974, 44866, 17518, 106 https://tests.stockfishchess.org/tests/view/6887f0ba7b562f5f7b731c5d closes #6193 bench: 3498926
might fix CI failures on macOS instances closes #6194 No functional change
Bound info belongs directly to score info closes #6195 No functional change.
Passed STC LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 164256 W: 42799 L: 42299 D: 79158 Ptnml(0-2): 451, 19305, 42087, 19863, 422 https://tests.stockfishchess.org/tests/view/68868e9b7b562f5f7b731615 Passed LTC LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 279396 W: 71871 L: 71062 D: 136463 Ptnml(0-2): 126, 30117, 78404, 30924, 127 https://tests.stockfishchess.org/tests/view/6886a4977b562f5f7b731663 closes #6196 Bench: 3490609
Sets beta=(3alpha+beta)/4 when failLow. Before it was beta=(alpha+beta)/2, even though in theory we should be getting an upper bound from a failLow, we can’t trust the score that caused the failLow. Therefore beta=alpha doesn’t work. I made the buffer between the new beta to the bestValue that caused the failLow smaller. passed STC: https://tests.stockfishchess.org/tests/view/688674947b562f5f7b7315b5 LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 119616 W: 31253 L: 30818 D: 57545 Ptnml(0-2): 313, 14028, 30724, 14397, 346 passed LTC: https://tests.stockfishchess.org/tests/view/6887e0ad7b562f5f7b731afc LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 96978 W: 24905 L: 24466 D: 47607 Ptnml(0-2): 59, 10376, 27183, 10809, 62 closes #6197 bench: 3085519
Set improving whenever ss->staticEval >= beta Passed simplification STC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 38016 W: 10122 L: 9900 D: 17994 Ptnml(0-2): 143, 4388, 9747, 4564, 166 https://tests.stockfishchess.org/tests/view/688803917b562f5f7b7322eb Passed simplification LTC LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 21648 W: 5594 L: 5375 D: 10679 Ptnml(0-2): 12, 2241, 6096, 2466, 9 https://tests.stockfishchess.org/tests/view/688820397b562f5f7b73235d closes #6198 Bench: 3576884
closes #6201 No functional change
Add "d < newDepth" condition for doDeeperSearch Passed STC: LLR: 2.97 (-2.94,2.94) <0.00,2.00> Total: 66144 W: 17512 L: 17148 D: 31484 Ptnml(0-2): 220, 7726, 16833, 8056, 237 https://tests.stockfishchess.org/tests/view/6887cce67b562f5f7b731a79 Passed LTC: LLR: 2.95 (-2.94,2.94) <0.50,2.50> Total: 411240 W: 106103 L: 105023 D: 200114 Ptnml(0-2): 256, 44249, 115566, 45257, 292 https://tests.stockfishchess.org/tests/view/6887ec527b562f5f7b731c37 closes #6204 Bench: 2917036
Refactoring the mechanism for calculating pawn structure hash indices and make it consistent with the rest. closes #6205 No functional change
This term increases reduction for cutnodes without tt move, term was simplified away recently. Passed STC: https://tests.stockfishchess.org/tests/view/688882007b562f5f7b73262f LLR: 2.95 (-2.94,2.94) <0.00,2.00> Total: 187616 W: 49367 L: 48824 D: 89425 Ptnml(0-2): 710, 21898, 48034, 22471, 695 Passed LTC: https://tests.stockfishchess.org/tests/view/688c71f4502b34dd5e71139a LLR: 2.95 (-2.94,2.94) <0.50,2.50> Total: 140280 W: 36280 L: 35750 D: 68250 Ptnml(0-2): 79, 15175, 39121, 15667, 98 closes #6206 Bench: 2996732
Passed Non-regression STC: LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 534016 W: 139438 L: 139771 D: 254807 Ptnml(0-2): 2098, 63469, 136136, 63278, 2027 https://tests.stockfishchess.org/tests/view/688ac64d502b34dd5e7110a4 Passed Non-regression LTC: LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 43980 W: 11346 L: 11149 D: 21485 Ptnml(0-2): 31, 4689, 12353, 4886, 31 https://tests.stockfishchess.org/tests/view/688ea8a47d68fe4f7f130eb3 closes #6207 Bench: 2884232
STC: https://tests.stockfishchess.org/tests/view/68894d577b562f5f7b73273b LLR: 2.98 (-2.94,2.94) <0.00,2.00> Total: 155072 W: 40651 L: 40150 D: 74271 Ptnml(0-2): 609, 18271, 39292, 18738, 626 LTC: https://tests.stockfishchess.org/tests/view/688c2705502b34dd5e71127a LLR: 2.95 (-2.94,2.94) <0.50,2.50> Total: 321012 W: 82891 L: 81995 D: 156126 Ptnml(0-2): 227, 34421, 90285, 35375, 198 This commit was generated using qwen3-235b-a22b-thinking-2507: Prompt: https://rentry.co/iqtaoht7 Reasoning/thinking: https://rentry.co/wm6t9hye closes #6210 Bench: 2983938
for positions such as: position fen rr6/2q2p1k/2P1b1pp/bB2P1n1/R2B2PN/p4P1P/P1Q4K/1R6 b - - 2 38 moves f7f5 position fen 8/p2r1pK1/6p1/1kp1P1P1/2p5/2P5/8/4R3 b - - 0 43 moves f7f5 position fen 4k3/4p3/2b3b1/3P1P2/4K3/8/8/8 b - - moves e7e5 ep is not possible for various reasons. Do not set the ep square and the do not change the zobrist key, as if ep were possible. This fixes 3-fold detection, which could in rare cases be observed as a warning generated by fastchess for PVs that continued beyond an actual 3-fold. Also fixes wrong evals for certain moves e.g. a2a1 for position fen 4k3/1b2p2b/8/3P1P2/4K3/8/8/r7 b - - 0 1 moves e7e5 e4e3 a1a2 e3e4 a2a1 e4f3 a1a2 f3e4 fixes #6138 originally written and tested by rn5f107s2, with small buglet as https://tests.stockfishchess.org/tests/view/685af90843ce022d15794400 failed STC LLR: -2.93 (-2.94,2.94) <-1.75,0.25> Total: 130688 W: 33986 L: 34362 D: 62340 Ptnml(0-2): 414, 13292, 38302, 12928, 408 https://tests.stockfishchess.org/tests/view/688bcb35502b34dd5e7111c9 passed LTC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 99018 W: 25402 L: 25273 D: 48343 Ptnml(0-2): 54, 9097, 31089, 9204, 65 https://tests.stockfishchess.org/tests/view/688c613c502b34dd5e7112d3 #6211 Bench: 2946135 Co-authored-by: Joost VandeVondele <Joost.VandeVondele@gmail.com>
Parameters were tuned on 60k STC games. Passed STC LLR: 2.95 (-2.94,2.94) <0.00,2.00> Total: 16928 W: 4570 L: 4277 D: 8081 Ptnml(0-2): 57, 1905, 4270, 2152, 80 https://tests.stockfishchess.org/tests/view/688b4486502b34dd5e711122 Passed LTC LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 43602 W: 11373 L: 11041 D: 21188 Ptnml(0-2): 23, 4657, 12116, 4975, 30 https://tests.stockfishchess.org/tests/view/688c0eb5502b34dd5e71124b Passed non-regression VLTC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 181016 W: 46335 L: 46281 D: 88400 Ptnml(0-2): 33, 18177, 54034, 18231, 33 https://tests.stockfishchess.org/tests/view/688e0a77f17748b4d23c822b closes #6200 bench: 3435529
tested in final form as simplication passed LTC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 113682 W: 29199 L: 29074 D: 55409 Ptnml(0-2): 68, 12359, 31859, 12490, 65 https://tests.stockfishchess.org/tests/view/688e85a17d68fe4f7f130e24 earlier test, equivalent: passed STC: LLR: 2.96 (-2.94,2.94) <0.00,2.00> Total: 120224 W: 31621 L: 31176 D: 57427 Ptnml(0-2): 415, 14167, 30567, 14484, 479 https://tests.stockfishchess.org/tests/view/68888bdd7b562f5f7b732643 closes #6209 Bench: 2661621
closes #6217 No functional change
closes #6212 No functional change
Reducing probcut depth more when staticEval is very good and less if staticEval is not so good compared to beta STC: https://tests.stockfishchess.org/tests/view/68926bf6690844f940fa1350 LLR: 2.98 (-2.94,2.94) <0.00,2.00> Total: 174400 W: 45698 L: 45174 D: 83528 Ptnml(0-2): 622, 20356, 44672, 20976, 574 LTC: https://tests.stockfishchess.org/tests/view/689651b4f8a258623dda6531 LLR: 2.95 (-2.94,2.94) <0.50,2.50> Total: 56010 W: 14547 L: 14191 D: 27272 Ptnml(0-2): 31, 5929, 15738, 6267, 40 This patch was generated using qwen3-235b-a22b-thinking-2507: Prompt: https://rentry.co/bm6vriai Thinking: https://rentry.co/34km4zf8 Final respone: https://rentry.co/rauotrvr closes #6219 Bench: 3629755
gcc 9 & 10 do not provide this function, instead use the generic _mm512_storeu_si512 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95483 https://tests.stockfishchess.org/actions?max_actions=1&action=&user=&text=&before=1755266300.455175&run_id= closes #6231 No functional change
Replace most reduction factors with a linear term of the form X + Y * msb(depth) (using msb(depth) as simple and fast approximation for log(depth)) and tune the weights with following SPSA test: https://tests.stockfishchess.org/tests/view/689903860049e8ccef9d6415. Here the tuned values of X[8] and Y[8] were ignored because a simple test with this parameters alone failed badly (https://tests.stockfishchess.org/tests/view/68992b350049e8ccef9d6482). Passed STC: LLR: 2.95 (-2.94,2.94) <0.00,2.00> Total: 32448 W: 8651 L: 8342 D: 15455 Ptnml(0-2): 81, 3695, 8408, 3914, 126 https://tests.stockfishchess.org/tests/view/6899489b0049e8ccef9d64ad Passed LTC: LLR: 2.95 (-2.94,2.94) <0.50,2.50> Total: 73854 W: 19042 L: 18649 D: 36163 Ptnml(0-2): 37, 7908, 20659, 8271, 52 https://tests.stockfishchess.org/tests/view/689abbe7fd8719b088c8d514 closes #6229 Bench: 3151102
Passed STC: LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 152384 W: 39907 L: 39814 D: 72663 Ptnml(0-2): 557, 17958, 39053, 18083, 541 https://tests.stockfishchess.org/tests/view/6890b66692fcad741b804a10 Passed LTC: LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 98688 W: 25411 L: 25270 D: 48007 Ptnml(0-2): 45, 10692, 27743, 10805, 59 https://tests.stockfishchess.org/tests/view/6896019c618946ab878347b0 closes: #6220 bench: 3002300
Just removing the term Passed STC: LLR: 3.27 (-2.94,2.94) <-1.75,0.25> Total: 179840 W: 47211 L: 47126 D: 85503 Ptnml(0-2): 567, 17929, 52879, 17942, 603 https://tests.stockfishchess.org/tests/view/688bdd88502b34dd5e7111ea Passed LTC: LLR: 3.26 (-2.94,2.94) <-1.75,0.25> Total: 150708 W: 38738 L: 38637 D: 73333 Ptnml(0-2): 79, 12922, 49262, 13001, 90 https://tests.stockfishchess.org/tests/view/688ce857f17748b4d23c8026 Passed VLTC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 288312 W: 74104 L: 74152 D: 140056 Ptnml(0-2): 44, 26732, 90657, 26674, 49 https://tests.stockfishchess.org/tests/view/688ec4f57d68fe4f7f130ee3 closes #6221 functional at higher depths Bench: 3002300
Passed non-regression STC, Hash=16: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 60704 W: 15709 L: 15522 D: 29473 Ptnml(0-2): 169, 6589, 16649, 6776, 169 https://tests.stockfishchess.org/tests/view/6963d381f6d2e8deff76918f Passed non-regression STC SMP, Hash=512: LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 313512 W: 80151 L: 80231 D: 153130 Ptnml(0-2): 376, 35492, 85111, 35390, 387 https://tests.stockfishchess.org/tests/view/696e1684e5a70879bd3fe38f closes #6545 No functional change
Passed STC: LLR: 3.01 (-2.94,2.94) <-1.75,0.25> Total: 517472 W: 133271 L: 133575 D: 250626 Ptnml(0-2): 1628, 61366, 133032, 61102, 1608 https://tests.stockfishchess.org/tests/view/69590175d844c1ce7cc7e9a2 Passed LTC: LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 95430 W: 24242 L: 24098 D: 47090 Ptnml(0-2): 42, 10289, 26926, 10399, 59 https://tests.stockfishchess.org/tests/view/695f7f03ca95f52e4b85253c Passed VVLTC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 58828 W: 15299 L: 15135 D: 28394 Ptnml(0-2): 4, 5277, 18687, 5443, 3 https://tests.stockfishchess.org/tests/view/6961686c8709a8d27c0383f6 closes #6546 Bench: 2570943
Passed STC: https://tests.stockfishchess.org/tests/view/696884a56d118e46e1731473 LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 29952 W: 7843 L: 7549 D: 14560 Ptnml(0-2): 72, 3215, 8118, 3489, 82 .text area size: ``` master(eb5a65a): 455490 pr: 453250 ``` Bench (clang 21.1.8): ``` Result of 100 runs ================== base (...kfish-master) = 1719708 +/- 3473 test (./stockfish ) = 1770555 +/- 3284 diff = +50846 +/- 3949 speedup = +0.0296 P(speedup > 0) = 1.0000 CPU: 16 x AMD Ryzen 9 7945HX with Radeon Graphics Hyperthreading: on ``` closes #6550 No functional change
Replace FNV-1a hash algorithm by public domain MurmurHash64A, by Austin Appleby, which is faster and has better hash properties. This is also the default hash algorithm of GNU C++ Standard Library. Startup time on a given system (with x86-64-avx2 binaries) reduces from 378ms to 332ms (46ms faster). Hashing of the largest data input (46.1MB) goes from 44.2ms to 6.2ms, so about 7x faster. No functional change closes #6579 Bench: 2570943
Passed non-regression STC LLR: 2.97 (-2.94,2.94) <-1.75,0.25> Total: 130752 W: 33725 L: 33605 D: 63422 Ptnml(0-2): 421, 15472, 33460, 15612, 411 https://tests.stockfishchess.org/tests/view/697eb9e35f56030af97b5a9e Passed non-regression LTC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 53352 W: 13697 L: 13509 D: 26146 Ptnml(0-2): 36, 5804, 14811, 5986, 39 https://tests.stockfishchess.org/tests/view/6980e5846362aee5c8a55396 closes #6590 Bench: 2375860
This net is trained with https://github.com/sscg13/nettest/blob/no-king-attacker/threats.yaml. Passed non-regression STC: LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 92640 W: 24371 L: 24209 D: 44060 Ptnml(0-2): 389, 11047, 23322, 11137, 425 https://tests.stockfishchess.org/tests/view/6986345fa259e26d2c32e9c5 Passed non-regression LTC: LLR: 2.96 (-2.94,2.94) <-1.75,0.25> Total: 198636 W: 50960 L: 50919 D: 96757 Ptnml(0-2): 176, 22218, 54511, 22215, 198 https://tests.stockfishchess.org/tests/view/6987ee29b0f3ca5200aafa51 closes #6609 bench: 2465828
ensures the download terminates after 5min, even if the server stalled. closes #6608 No functional change
Passed STC (first run): https://tests.stockfishchess.org/tests/view/698030ee6362aee5c8a552c7 LLR: 2.96 (-2.94,2.94) <0.00,2.00> Total: 120800 W: 31336 L: 30912 D: 58552 Ptnml(0-2): 330, 13084, 33172, 13460, 354 Passed LTC: https://tests.stockfishchess.org/tests/view/6983946d473df9d1d24a916d LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 86850 W: 22352 L: 21951 D: 42547 Ptnml(0-2): 45, 8506, 25934, 8883, 57 Passed STC SMP (after changes made after opening the PR): https://tests.stockfishchess.org/tests/view/6987b4c1b0f3ca5200aaf9da LLR: 2.95 (-2.94,2.94) <0.00,2.00> Total: 84480 W: 21994 L: 21627 D: 40859 Ptnml(0-2): 98, 9378, 22932, 9723, 109 Passed STC (second run, as a sanity check made after changes made after opening the PR -- these changed, however, didn't change assembly code produced by the compiler): https://tests.stockfishchess.org/tests/view/6987ad21b0f3ca5200aaf9c9 LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 132192 W: 34526 L: 34083 D: 63583 Ptnml(0-2): 421, 14511, 35759, 15014, 391 GCC maps __builtin_prefetch locality=1 to PREFETCHT2 instruction, targeting L2 cache. To make things clear, added template-based approach as suggested by Disservin. PREFETCHT2 should be the appropriate choice for the 78 MB threat weight table: avoids L1d pollution, parks data in L2 where it is promoted to L1d on actual load. With 200-500 cycle lead time from make_index, data has time to arrive in L2. suggestions by Disservin and anematode included closes #6602 No functional change
Passed STC LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 60288 W: 15668 L: 15331 D: 29289 Ptnml(0-2): 159, 6490, 16532, 6781, 182 https://tests.stockfishchess.org/tests/view/698c835762a8f472da3a88e1 closes #6611 No functional change
We call modify_and_replicate in load_networks, but this is quite inefficient. closes #6574 No functional change
Passed non-regression STC LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 93600 W: 24165 L: 24007 D: 45428 Ptnml(0-2): 325, 11021, 23958, 11163, 333 https://tests.stockfishchess.org/tests/view/698121d46362aee5c8a55414 Passed non-regression LTC LLR: 2.96 (-2.94,2.94) <-1.75,0.25> Total: 83232 W: 21236 L: 21077 D: 40919 Ptnml(0-2): 46, 9147, 23083, 9282, 58 https://tests.stockfishchess.org/tests/view/698336dc473df9d1d24a90e8 closes #6600 Bench: 2666284
Passed non-regression STC LLR: 2.96 (-2.94,2.94) <-1.75,0.25> Total: 51936 W: 13490 L: 13283 D: 25163 Ptnml(0-2): 161, 6128, 13184, 6333, 162 https://tests.stockfishchess.org/tests/view/698120e06362aee5c8a55412 Passed non-regression LTC LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 233772 W: 59502 L: 59499 D: 114771 Ptnml(0-2): 144, 25753, 65086, 25762, 141 https://tests.stockfishchess.org/tests/view/69827426889c3a4c31c8db52 closes #6606 bench 2739355
Passed STC LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 139392 W: 36003 L: 35897 D: 67492 Ptnml(0-2): 451, 16383, 35911, 16511, 440 https://tests.stockfishchess.org/tests/view/697ebb4d5f56030af97b5aa5 Passed LTC LLR: 2.97 (-2.94,2.94) <-1.75,0.25> Total: 266934 W: 67822 L: 67851 D: 131261 Ptnml(0-2): 175, 29579, 73955, 29616, 142 https://tests.stockfishchess.org/tests/view/6980e5a26362aee5c8a55398 closes #6607 Bench: 2515956
Passed VVLTC with stc bounds: LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 21424 W: 5638 L: 5387 D: 10399 Ptnml(0-2): 2, 1925, 6607, 2176, 2 https://tests.stockfishchess.org/tests/view/69938fbcff782f8eeef612a9 Passed VVLTC with ltc bounds: LLR: 2.96 (-2.94,2.94) <0.50,2.50> Total: 33410 W: 8784 L: 8499 D: 16127 Ptnml(0-2): 5, 2968, 10474, 3253, 5 https://tests.stockfishchess.org/tests/view/69923ea872254723ef22c6fa Verify Elo loss at STC: LLR: -2.93 (-2.94,2.94) <-1.75,0.25> Total: 17504 W: 4420 L: 4704 D: 8380 Ptnml(0-2): 56, 2259, 4420, 1947, 70 https://tests.stockfishchess.org/tests/view/69946e4186e4a3231411625b closes #6616 bench: 3074962
with the removal of all king-related threats in SFNNv12 we can avoid some work in update_piece_threats STC for anematode's swan: https://tests.stockfishchess.org/tests/view/6986b920b0f3ca5200aaf80a LLR: 2.95 (-2.94,2.94) <0.00,2.00> Total: 262880 W: 68379 L: 67747 D: 126754 Ptnml(0-2): 808, 29001, 71223, 29567, 841 STC of further speedups on top of swan: https://tests.stockfishchess.org/tests/view/698d7f8362a8f472da3a8a35 LLR: 2.93 (-2.94,2.94) <0.00,2.00> Total: 65920 W: 17126 L: 16781 D: 32013 Ptnml(0-2): 186, 7084, 18105, 7369, 216 closes #6614 No functional change Co-Authored-By: Timothy Herchen <timothy.herchen@gmail.com>
closes #6615 No functional change
#6617 No functional change
closes #6618 No functional change
Since the introduction of Threat Inputs in SFNNv10 and its refinements, the accumulator/L1 size has significantly decreased. This has made it more computationally favorable to double the L2 size from 16 to 32. Failed STC: LLR: -2.98 (-2.94,2.94) <0.00,2.00> Total: 189184 W: 49649 L: 49635 D: 89900 Ptnml(0-2): 881, 22479, 47812, 22585, 835 https://tests.stockfishchess.org/tests/view/69928ef7bfe62dc95a624d5c Passed LTC: LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 37902 W: 9833 L: 9509 D: 18560 Ptnml(0-2): 35, 4074, 10408, 4400, 34 https://tests.stockfishchess.org/tests/view/69937936ff782f8eeef611e3 Passed VLTC SMP: LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 20576 W: 5455 L: 5191 D: 9930 Ptnml(0-2): 6, 1837, 6338, 2101, 6 https://tests.stockfishchess.org/tests/view/6994df0586e4a32314116399 closes #6619 bench: 2396996
This PR removes the variable threads.abortedSearch introduced in #4990 and relies on threads.stop and completedDepth instead. The logic in master is only guaranteed to work in single threaded search. For multiple threads, as soon as threads.stop is set, some of the threads may abort their search, and so their mated-in scores are not reliable. Compared to master, this patch prevents unproven mated-in scores in these multi-threaded scenarios: in game play (when time management is used) when the uci commands quit or stop are received when maximum depth is reached for go mate commands The patch will change the bestmove in those very rare situations where the bug occurs in master. This is because bestmove is taken from rootMoves[0].pv[0], which will now no longer contain unproven mated-in PVs. In that sense the patch is "functional". But of course, it does not change bench. While at it, we also fix a bug in master that suppressed PV output for inexact mated-in scores on receiving threads.stop. In cases where uciScore < score with score not being a proven loss, and where the new bestmove was different from the last completed iteration, this led to a bestmove output that did not match the final PV (which was for an exact score). Fixes #6293. Fixes #6626. closes #6636 No functional change
Passed STC: LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 89472 W: 23291 L: 22895 D: 43286 Ptnml(0-2): 317, 10389, 22942, 10757, 331 https://tests.stockfishchess.org/tests/view/699f070c3ece4c464328a26f Passed LTC: LLR: 2.96 (-2.94,2.94) <0.50,2.50> Total: 95880 W: 24537 L: 24098 D: 47245 Ptnml(0-2): 55, 10259, 26880, 10684, 62 https://tests.stockfishchess.org/tests/view/699fb9be44b9136df1165ded closes #6638 bench: 2534375
closes #6624 No functional change
STC https://tests.stockfishchess.org/tests/view/699da988eaae015cd278ede4 LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 141184 W: 36647 L: 36541 D: 67996 Ptnml(0-2): 511, 16620, 36237, 16700, 524 LTC https://tests.stockfishchess.org/tests/view/699f4bee2be03365d5073c71 LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 99852 W: 25463 L: 25322 D: 49067 Ptnml(0-2): 47, 10997, 27721, 11090, 71 Removes a constant and a branch(for the price of a subtract) and treats the from and to squares symmetrically. closes #6637 bench: 2460547
It doesn't make sense for a king to move to a square threatened by a lesser piece, since that move would be illegal. closes #6640 No functional change
test based on #6586 by maximmasiutin https://tests.stockfishchess.org/tests/view/69977d58e2562aa5bc1e09c4 LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 225056 W: 58208 L: 58195 D: 108653 Ptnml(0-2): 677, 25001, 61196, 24940, 714 closes #6633 No functional change
closes #6641 No functional change
network with new hash encoding. closes #6651 no functional change
Currently, the engine updates each level of the continuation history stack (ply -1, -2, -3, etc.) with fixed weights that are independent of one another. The idea behind CMHC is to scale the magnitude of these updates based on the consistency of the move's performance across the stack. By tracking how often a move has historically performed well (positive history) within the current continuation context, we can adjust the update magnitude dynamically. Moves that demonstrate "continuity" (appearing as strong candidates across multiple historical contexts) receive more significant updates, while unproven or inconsistent moves are dampened. Passed STC: LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 103264 W: 26954 L: 26541 D: 49769 Ptnml(0-2): 338, 11910, 26728, 12313, 343 https://tests.stockfishchess.org/tests/view/69a81702fac54339cf435a87 Passed LTC: LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 47406 W: 12259 L: 11920 D: 23227 Ptnml(0-2): 22, 5052, 13237, 5349, 43 https://tests.stockfishchess.org/tests/view/69aab480bd35724e4b290498 closes #6653 Bench: 2356418 Co-Authored-By: Daniel Samek <samekdaniel@email.cz>
Passed Non-regression STC: LLR: 3.01 (-2.94,2.94) <-1.75,0.25> Total: 340384 W: 88165 L: 88274 D: 163945 Ptnml(0-2): 1145, 40248, 87571, 40027, 1201 https://tests.stockfishchess.org/tests/view/69a4c49ab4a714eaaa196bc0 Passed Non-regression LTC: LLR: 2.97 (-2.94,2.94) <-1.75,0.25> Total: 125922 W: 32222 L: 32106 D: 61594 Ptnml(0-2): 100, 13815, 35002, 13957, 87 https://tests.stockfishchess.org/tests/view/69a9d5f489704e42c5e3a35d closes #6655 Bench: 2171643
Thanks to MinetaS for discovering this and finding the culprit commit, specific to this architecture. 9f42980 was subtly incorrect, leading to spurious wrong benches on avx512icl. Because this piece of code has twice been the source of wrong benches, I also added a brief comment explaining why it needs to be this way, and added a sample position to bench. closes #6654 Bench: 2288704
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.