1
0
Fork 0
mirror of https://github.com/sockspls/badfish synced 2025-05-08 12:19:36 +00:00

Commit graph

  • 0405f35403 Double probability of using classical eval SFisGOD 2020-09-07 04:54:26 +08:00
  • d2562cde12 Always re-enable NNUE after "bench". Gian-Carlo Pascutto 2020-09-08 15:37:53 +02:00
  • fc27d158c0 Bug fix in do_null_move() and NNUE simplification. syzygy1 2020-09-06 17:29:12 +02:00
  • d539da19d2 Use classical eval more often SFisGOD 2020-09-04 17:14:50 +08:00
  • 9a063fc3cb Adjust penalty on refuted early quiet moves Vizvezdenec 2020-09-04 15:53:59 +03:00
  • 9cc482c788 Update default net to nn-308d71810dff.nnue Sergio Vieri 2020-09-03 20:22:51 +08:00
  • 2a69611509 LMR Simplification VoyagerOne 2020-09-02 16:58:44 -04:00
  • d6530f7d49 Simplify singularQuietLMR Unai Corzo 2020-09-03 12:18:42 +02:00
  • 0e1f734b05 Less pruning in qsearch Unai Corzo 2020-09-02 16:45:49 +02:00
  • 571c2d6d8d Restore development version Joost VandeVondele 2020-09-04 07:46:06 +02:00
  • c306d83869 Stockfish 12 sf_12 Joost VandeVondele 2020-09-02 09:12:04 +02:00
  • aa2de71230 Update CPU contributors list Joost VandeVondele 2020-09-02 08:05:08 +02:00
  • be87517734 Only use MADV_RANDOM if defined Joost VandeVondele 2020-09-01 12:22:47 +02:00
  • a8bbaa1795 LMR Root Node Simplification VoyagerOne 2020-08-30 13:57:57 -04:00
  • 61381372ec Always print an info line before a bestmove Joost VandeVondele 2020-08-31 22:53:20 +02:00
  • a057f170c6 Use llvm linker with clang on windows for LTO. mstembera 2020-08-30 20:48:10 -07:00
  • a0afe32d16 Use stable sort to make sure bench with TB yields same results everywhere. mstembera 2020-08-30 18:40:49 -07:00
  • e0bafa1911 Update parametes in classical evaluation. FauziAkram 2020-08-30 13:58:05 +03:00
  • 9b5b9ec9a6 QS Pruning Simplification VoyagerOne 2020-08-29 21:13:05 -04:00
  • c02b3a4c7a Add / remove leaves from search tree ttPv MJZ1977 2020-08-28 12:06:36 +02:00
  • d90d893b5e Reintroduce depth reduction Unai Corzo 2020-08-28 09:27:15 +02:00
  • e4ed7d3dd7 Cleaner make help Joost VandeVondele 2020-08-26 18:00:54 +02:00
  • 406979ea12 Embed default net, and simplify using non-default nets Stéphane Nicolet 2020-08-23 13:43:38 +02:00
  • 242a7d9fea Simplify MCP in QS VoyagerOne 2020-08-25 09:10:47 -04:00
  • 95b8f3f800 Remove Reduce Depth VoyagerOne 2020-08-23 12:04:50 -04:00
  • 9b4967071e Remove EvalList syzygy1 2020-08-24 02:29:38 +02:00
  • b0b4ca17db Check ARCH=.... variable Joost VandeVondele 2020-08-24 21:32:04 +02:00
  • 530fccbf27 Allow for VNNI256 compilation with g++-8 mstembera 2020-08-24 03:38:01 -07:00
  • 843a961a8c Introduce countermove based pruning for qsearch Vizvezdenec 2020-08-24 08:04:16 +03:00
  • f7b3f0e842 Allow TT entries with key16==0 to be fetched Sami Kiminki 2020-08-21 12:12:39 +03:00
  • 701b2427bd Support VNNI on 256bit vectors mstembera 2020-08-20 16:59:27 -07:00
  • e453f09f06 armv8 AArch64 does not require -mfpu=neon George Sobala 2020-08-24 06:37:42 +01:00
  • d5f86b6359 Introduce movecount pruning for qsearch() Vizvezdenec 2020-08-23 14:22:32 +03:00
  • cc9d503dde Skip the alignment bug workaround for Clang syzygy1 2020-08-22 13:36:34 +02:00
  • 5f1843c9cb Small trivial cleanups Joost VandeVondele 2020-07-11 16:59:33 +02:00
  • 3542033342 Instructions to build on older Macintosh Stéphane Nicolet 2020-08-22 11:37:53 +02:00
  • 34f67c5722 Explicitly rely on pthreads if possible Joost VandeVondele 2020-08-21 22:10:55 +02:00
  • cbcb05ca09 Display classic and NNUE evaluation in trace mode MJZ1977 2020-08-21 10:57:34 +02:00
  • e64b957274 Simplify away internal iterative deepening Unai Corzo 2020-08-21 09:24:25 +02:00
  • 15abcaedc1 Update Makefile for macOS gsobala 2020-08-21 11:28:53 +01:00
  • 8b45b1c490 Deal with very old linux kernels Joost VandeVondele 2020-08-21 07:42:19 +02:00
  • daac86691d Set Use NNUE by default to true Joost VandeVondele 2020-08-20 14:24:49 +02:00
  • a1ad8604a1 Send error message as an UCI info string Joost VandeVondele 2020-08-19 19:21:41 +02:00
  • 2deb08a529 Reintroduce last captures extension SFisGOD 2020-08-18 18:54:28 +08:00
  • 19129473f2 Merge NNUE (master) in the cluster branch. Joost VandeVondele 2020-08-19 21:26:22 +02:00
  • 42e8789f0b Expanded support for x86-32 architectures. syzygy1 2020-08-18 01:56:12 +02:00
  • 384d684484 Better error message on missing curl/wget Joost VandeVondele 2020-08-18 18:06:28 +02:00
  • fbae5614eb Fix Makefile typo Joost VandeVondele 2020-08-18 08:49:06 +02:00
  • 1bcc981a5a Fallback to NNUE mstembera 2020-08-16 15:23:50 -07:00
  • 581b92e4a7 Remove last captures extension Unai Corzo 2020-08-17 18:22:32 +02:00
  • 1c0b7bdf4f Remove history bonus from Eval VoyagerOne 2020-08-17 08:58:03 -04:00
  • 65b976439f Support building for Android using NDK notruck 2020-08-16 08:59:13 -07:00
  • 0e17a89e4d Simplify away the passed pawn extension Unai Corzo 2020-08-17 09:22:15 +02:00
  • 81d716f5cc Reformat code in little-endian patch Stéphane Nicolet 2020-08-16 21:46:54 +02:00
  • 65572de4a7 Add further targets to travis testing Joost VandeVondele 2020-08-16 13:21:07 +02:00
  • 72dc7a5c54 Assume network file is in little-endian byte order syzygy1 2020-08-15 16:50:39 +02:00
  • 8cf43c6317 Display NEON in compiler string Daylen Yang 2020-08-14 19:53:46 -07:00
  • cd0b8b4cf2 Use NNUE more for fortresses Joost VandeVondele 2020-08-14 22:18:12 +02:00
  • 6eb186c97e Try to match relative magnitude of NNUE eval to classical mstembera 2020-08-14 04:49:33 -07:00
  • e5f450cf0b Also dampen NNUE eval with 50 move rule Miguel Lahoz 2020-08-10 22:57:11 +08:00
  • ce009ea1aa Verify SHA of downloaded net file Joost VandeVondele 2020-08-13 22:54:13 +02:00
  • e8ea215a13 Clean-up Makefile help Stéphane Nicolet 2020-08-13 13:40:06 +02:00
  • 67e48418af Update default net to nn-82215d0fd0df.nnue Sergio Vieri 2020-08-12 23:21:21 +08:00
  • 69cfe28f31 Output the SSE2 flag in compiler_info Joost VandeVondele 2020-08-12 17:21:12 +02:00
  • dd63b98fb0 Add support for VNNI mstembera 2020-08-11 12:59:39 -07:00
  • 6bc0256292 Use posix_memalign for Apple Silicon instead of _mm_malloc Daylen Yang 2020-08-11 12:02:48 -07:00
  • 992f549ae7 Restrict avx2 hack to windows target Joost VandeVondele 2020-08-11 21:11:17 +02:00
  • ee06046412 Tweak castling extension SFisGOD 2020-08-10 21:13:56 +08:00
  • ea6220f381 This commit enables a mixed bench, to improve CI and allow for PGO (profile-build) of the NNUE part of the code. Guy Vreuls 2020-08-07 17:07:46 +02:00
  • f46c73040c Fix AVX512 build with older compilers mstembera 2020-08-10 12:52:46 -07:00
  • 399cddf444 More aligned_alloc changes to support Android Joost VandeVondele 2020-08-10 16:14:17 +02:00
  • 4ab8b0b738 Fix parallel LTO issues on Windows Guy Vreuls 2020-08-11 04:38:38 +02:00
  • a72cec1ff8 Add comments to probCut code Vizvezdenec 2020-07-18 16:30:00 +03:00
  • 220ef1d27d Assorted search parameter tune Unai Corzo 2020-08-10 15:38:44 +02:00
  • 21df37d7fd Provide vectorized NNUE code for SSE2 and MMX targets Fanael Linithien 2020-08-09 16:20:45 +02:00
  • f948cd008d Cleanup and optimize SSE/AVX code mstembera 2020-08-09 16:23:33 -07:00
  • cb0504028e Makefile rework/cleanup sf-x 2020-08-09 18:01:18 +03:00
  • ad2ad4c657 Modify castling extension SFisGOD 2020-08-10 15:39:22 +08:00
  • 875183b310 Workaround using unaligned loads for gcc < 9 mstembera 2020-08-09 23:50:59 -07:00
  • a54f9011c3 simplying hybrid condition jjoshua2 2020-08-09 16:16:04 -04:00
  • bcdf41dadc Update default net to nn-112bb1c8cdb5.nnue Sergio Vieri 2020-08-10 08:47:52 +08:00
  • 651ec3b31e Revert "Avoid special casing for MinGW" Joost VandeVondele 2020-08-10 07:18:15 +02:00
  • 27b593a944 Fix a data race for NNUE Joost VandeVondele 2020-08-09 18:11:38 +02:00
  • a6e89293df Avoid special casing for MinGW Dariusz Orzechowski 2020-08-09 14:32:24 -07:00
  • 2bfde55429 Adjust NNUE usage based on number of pawns in position Vizvezdenec 2020-08-09 21:39:46 +03:00
  • cd1bb27dd4 Fix aligned_alloc on MinGW Joost VandeVondele 2020-08-09 19:08:47 +02:00
  • f16face915 Fix aligned_alloc on MinGW alignedAllocMinGW Joost VandeVondele 2020-08-09 19:08:47 +02:00
  • 320fa1b2f0 Improve error message on missing net. Joost VandeVondele 2020-08-09 11:05:07 +02:00
  • d7a26899a9 Use fallback implementation for C++ aligned_alloc Daniel Dugovic 2020-08-08 15:39:29 -05:00
  • add890a10b LMR search tweak Unai Corzo 2020-08-08 22:08:40 +02:00
  • 1949eb8604 Singular extension search tweak Unai Corzo 2020-08-08 22:03:37 +02:00
  • 6d6267c378 Parallelize Link Time Optimization for GCC, CLANG and MINGW Guy Vreuls 2020-08-08 12:45:10 +02:00
  • e663bc5330 Do more aggressive futility pruning for captures Vizvezdenec 2020-08-08 17:43:41 +03:00
  • 3368d03285 update Null Move Pruning parameters Moez Jellouli 2020-08-08 12:35:34 +02:00
  • 450b60a303 Remove unnecessay legality check mckx00 2020-08-08 03:07:07 -07:00
  • 23ecf3d5c6 simplified and increased threshold to switch between NNUE and classical U-DESKTOP-3900\Mark 2020-08-07 19:53:18 -04:00
  • 910f779eb1 Do more futility pruning for parent nodes. Vizvezdenec 2020-08-08 05:51:26 +03:00
  • f4c27cda1a Reintroduce late irreversible move extension SFisGOD 2020-08-08 03:45:08 +08:00
  • 5ccff25df2 Expand outposts to minors shielded by pawns Lolligerhans 2020-08-07 11:24:37 +02:00
  • dc5af66ead Tweak futility pruning depth. Unai Corzo 2020-08-08 08:24:20 +02:00