lichess-org / lila

♞ lichess.org: the forever free, adless and open source chess server ♞
https://lichess.org
GNU Affero General Public License v3.0
15.64k stars 2.28k forks source link

Stockfish loses by time - Stockfish NEVER loses by time #8252

Closed psebastiani2021 closed 3 years ago

psebastiani2021 commented 3 years ago

Hello. In this game https://lichess.org/RG8AuoPWvdMc Stockfish loses by time. Is this common? please @ornicar fix this bug.

TBestLittleHelper commented 3 years ago

Stockfish be like: IDK En-Passant

https://lichess.org/pzzTDPle#49

https://lichess.org/forum/general-chess-discussion/you-know-it-feels-bad-when-stockfish-level-7-does-not-know-en-passant

Seems like Antichess en passant causes stockfish to time out. Haven't tried to repo this myself, so can't confirm. Also, usually no need to ping a developer when making an issue.

ddugovic commented 3 years ago

You're right, merging changes from official-stockfish broke my fork yet again.

$ gdb -ex run --args ./stockfish bench giveaway 16 1 13 test.fen depth classical

info string variant giveaway startpos rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1

Position: 1/1 (8/1r1ppk2/8/5n1P/4pP2/2P5/3P4/8 b - f3 0 25)
stockfish: ./bitboard.h:189: Bitboard pawn_attacks_bb(Color, Square): Assertion `is_ok(s)' failed.

Thread 1 "stockfish" received signal SIGABRT, Aborted.
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49
49      ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bt
#0  __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49
#1  0x00007ffff7a65864 in __GI_abort () at abort.c:79
#2  0x00007ffff7a65749 in __assert_fail_base (fmt=0x7ffff7bf1458 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n", assertion=0x4a8bbc "is_ok(s)", file=0x18c8339 "./bitboard.h", line=189,
    function=<optimized out>) at assert.c:92
#3  0x00007ffff7a77a96 in __GI___assert_fail (assertion=0x4a8bbc "is_ok(s)", file=0x18c8339 "./bitboard.h", line=189, function=0x4a8bc5 "Bitboard pawn_attacks_bb(Color, Square)")
    at assert.c:101
#4  0x000000000046f706 in Position::attackers_to (this=this@entry=0x1b3f1e8, s=SQ_NONE, occupied=16326236112685056) at position.cpp:842
#5  0x000000000046c566 in Position::attackers_to (this=0x1b3f1e8, s=SQ_A1) at ./position.h:542
#6  Position::set (this=<optimized out>, fenStr=..., isChess960=<optimized out>, v=<optimized out>, si=<optimized out>, th=0x1b3f100) at position.cpp:379
#7  0x0000000000487164 in ThreadPool::start_thinking (this=<optimized out>, pos=..., states=std::unique_ptr<class std::deque<StateInfo, std::allocator<StateInfo> >> = {...}, limits=...,
    ponderMode=<optimized out>) at thread.cpp:228
#8  0x000000000048a45d in (anonymous namespace)::go (pos=..., is=..., states=std::unique_ptr<class std::deque<StateInfo, std::allocator<StateInfo> >> = {...}) at uci.cpp:230
#9  0x000000000048b46d in (anonymous namespace)::bench (pos=..., args=..., states=std::unique_ptr<class std::deque<StateInfo, std::allocator<StateInfo> >> = {...}) at uci.cpp:258
#10 0x0000000000489a12 in UCI::loop (argc=argc@entry=9, argv=argv@entry=0x7fffffffe478) at uci.cpp:356
#11 0x000000000043a1b1 in main (argc=9, argv=0x7fffffffe478) at main.cpp:49
niklasf commented 3 years ago

Should we revert to the last merge before NNUE, and stay there until we can upgrade to Fairy Stockfish?

ddugovic commented 3 years ago

Yes please, I give up on maintaining my fork. Sorry.

niklasf commented 3 years ago

Updated to fishnet 2.2.6.

TBestLittleHelper commented 3 years ago

Same thing in horde? Not sure if the same thing or something different

https://lichess.org/forum/general-chess-discussion/i-beated-stockfish-8

ddugovic commented 3 years ago

My testing of a possible patch for that https://github.com/ddugovic/Stockfish/commit/0b4402a3c1a651d36e1bf099c180ed7f482bad53 and https://github.com/ddugovic/Stockfish/commit/4e875080ad0f20a93cbae245d6e7c3d7d9a62c50 has been quite hasty.