Sloppy 0.1.1 released

[ilari]

http://koti.mbnet.fi/~ilaripih/sloppy/

It's the second version of my open source chess engine. Unlike the first version Sloppy 0.1.1 compiles also with Microsoft's compiler (makefile included). Configuration (hash size, etc.) is now also easier because there's a config file.

As for playing strength, this version should definitely be stronger by 0 - 100 ELO, thanks to improvements in search and a complete rewrite of king-attack evaluation. I haven't run enough test games to be absolutely sure though.

As always, I'd very much appreciate faster binaries, feedback, and especially a binary for Windows x86-64.

Regards,
Ilari

[Jim Ablett]

Sloppy 0.1.1 win32 Intel compiler 10 pgo build

http://www.2xupload.de/file/21284f53867 ... a.zip.html

Jim.

[Tony Thomas]

I was waiting for the release of this version. Let's hope I have found a new comer that can out do bugchess.

[Tony Thomas]

I wish I had a logo to go with it.


Found one at the page of JA.

[David Dahlem]

I wish I had a logo to go with it.


Found one at the page of JA.

You can find my Sloppy logo here ...

http://wbec-ridderkerk.nl/html/Dave_logosA.html

Regards
Dave

[ilari]

Sloppy 0.1.1 win32 Intel compiler 10 pgo build

Thanks! I've uploaded the binary to my site.

I was waiting for the release of this version. Let's hope I have found a new comer that can out do bugchess.

I just ran 100 bullet games against BugChess2 V1.5.2. Sloppy 0.1.1 (64-bit) got 55% of the score (41-31-28). Being a bitboard engine, the 32-bit version would of course be a bit weaker.

You can find my Sloppy logo here ...

Nicely done. Logos seem to be important in the computer chess world.

[Tony Thomas]

Nicely done. Logos seem to be important in the computer chess world.

You arent kidding, my Alfil program gained about 30 points because I used a Graham Banks logo.

[Tony Thomas]

I just ran 100 bullet games against BugChess2 V1.5.2. Sloppy 0.1.1 (64-bit) got 55% of the score (41-31-28). Being a bitboard engine, the 32-bit version would of course be a bit weaker.

In a 40 moves in 6 minutes match Sloppy is leading BugChess 7-3-4 here (32-bit).

Ended 14-6-12. For Sloppy I used Jim Ablett's slightly faster compile on a P4 3GHz machine.

At my time controls, sloppy is still lagging behind Bugchess after around 80 games. May be my time control (1m+1sec) is too fast for sloppy, and simple tricks like the ones Uri recently did, printing the pv etc, could help it a lot.

Code: Select all

2 BugChess2_V1_5_2               : 2641   57  56   127    61.8 %   2557   18.1 %
24 Sloppy-0.1.1 JA                : 2561   59  59    97    50.0 %   2561   27.8 

[ilari]

At my time controls, sloppy is still lagging behind Bugchess after around 80 games. May be my time control (1m+1sec) is too fast for sloppy...

It's very likely. I haven't done proper testing with the "whole game in X minutes" time control. The time Sloppy allocates for each move is roughly "(time_left / 30) + extra_time". It's probably a good idea to replace "30" with a significantly bigger value. It's line 848 of search.c, in case someone wants to experiment. I sure do.

and simple tricks like the ones Uri recently did, printing the pv etc, could help it a lot.

Can you elaborate, how could printing the pv help?

[Jim Ablett]

I've compiled a 64-bit GCC-Mingw64 version >

http://www.2xupload.de/file/e9fbbccf557 ... a.zip.html

This is untested - I don't have a 64-bit Windows environment to test, so don't complain if it doesn't work

Jim.