Gull 3 Linux+Syzygy and Fathom released.

[basil00]

[d] 4k3/8/8/8/7q/8/3Q4/R3K3 w Q - 3 1 if playing white, does not pick up mate score. Komodo here will show mate score. I have syzygy table base in two different directories - could that be an issue?

This is actually not a syzygy TB position because of the castling rights.

[basil00]

After some first tests it looks that syzygy for now only works up to 5men, initializing 6men makes the engine crash, at least in Chessbase and Arena gui

It seems there is a problem with 6-man, also here. Unfortunately I could not test 6-man. The 5-man case passed 1000s of games (play-to-mate) so seems stable.

Hard to say what the problem without being able to reproduce it. Possible buffer overflow? Perhaps try to re-compile with -fsanitize=address and see what it says.

[MikeB]

[d] 4k3/8/8/8/7q/8/3Q4/R3K3 w Q - 3 1 if playing white, does not pick up mate score. Komodo here will show mate score. I have syzygy table base in two different directories - could that be an issue?

This is actually not a syzygy TB position because of the castling rights.

Good catch , this position works fine.

[d]4k3/8/8/8/7q/8/3Q4/1R2K3 w - - 3 1

[stuwph]

Hard to say what the problem without being able to reproduce it. Possible buffer overflow? Perhaps try to re-compile with -fsanitize=address and see what it says.

-fsanitize=address doesn't work on my build, but mingw600 raises these warnings:

tbprobe.c: In function 'unsigned int tb_probe_wdl_impl(uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, unsigned int, bool)':
tbprobe.c:1822:0: warning: narrowing conversion of 'ep' from 'unsigned int' to 'uint8_t {aka unsigned char}' inside { } [-Wnarrowing]
};

tbprobe.c: In function 'unsigned int tb_probe_root_impl(uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, unsigned int, unsigned int, bool, unsigned int*)':
tbprobe.c:1857:0: warning: narrowing conversion of 'rule50' from 'unsigned int' to 'uint8_t {aka unsigned char}' inside { } [-Wnarrowing]
};

[Dan Cooper]

Has anyone been able to get the new Windows binary to work properly? Gull loads and runs fun but it doesn't detect the Syzygy bases nor does it report any TB hits.

[ernest]

Has anyone been able to get the new Windows binary to work properly?

Where is that new Windows binary ?

[basil00]

Has anyone been able to get the new Windows binary to work properly? Gull loads and runs fun but it doesn't detect the Syzygy bases nor does it report any TB hits.

The current version does not report TB hits yet. I've experimented with adding this functionality in but nothing is complete yet.

Where is that new Windows binary ?

There is a build here: https://github.com/basil00/Gull/releases
But I have had reports that it does not work with various GUIs. This will be fixed as part of a bigger patch to move the UCI interface into its own thread.

[Dan Cooper]

I checked task manager and saw disk I/O from all of the threads, so I assume tb probing is working.

And yes, it seems to have difficulties running in ChessGUI, but Arena is fine.

[Ozymandias]

There is a build here: https://github.com/basil00/Gull/releases
But I have had reports that it does not work with various GUIs. This will be fixed as part of a bigger patch to move the UCI interface into its own thread.

Is it possible to get a bmi2 compile? I tried doing it myself, but when running "make", it says there's nothing to be done with "main". (x86_64-5.3.0-release-posix-seh-rt_v4-rev0 with MSYS 1.0.11, under Win8).

[stuwph]

try this one, compiled with mingw64 5.30 -march=native on a 5960x

http://1drv.ms/1Sy1FsC