DiscoCheck 4.0.0 (release candidate)

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

lucasart
Posts: 3242
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

DiscoCheck 4.0.0 (release candidate)

Post by lucasart »

New version of my engine DiscoCheck. You can find the source code (licensed GPL v3+) and binaries here:
https://github.com/lucasart/chess

Binaries are in the 'bin' folder:
* Only Linux and Windows 64-bit binaries for the moment.
* MacOSX users who have gcc installed (recent version supporting ISO C++11) should be able to compile it using 'make.sh', although I don't have a Mac, so I haven't tried.
* Android compile is planned, stay tuned!

This is a release candidate, so I'd like to know about bugs primarly, so I can fix them and make a 4.0.1 release with less bugs and so on.

In terms of playing strength, it's hard to say, as I haven't conducted any experiment that is comparable to a rating lists (mostly self-play at super fast time control). My guess is 2760 ELO by CCRL 40/4 standard.

In terms of playing style, it really plays like you would expect: pathetic and stupid materialistic computer. And it's not even very strong at tactics (relatively to other programs, not by human standard obviously), but somehow it plays well in the endgame, and can be quite aggressive and speculative with passed pawns.
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
mar
Posts: 2667
Joined: Fri Nov 26, 2010 2:00 pm
Location: Czech Republic
Full name: Martin Sedlak

Re: DiscoCheck 4.0.0 (release candidate)

Post by mar »

Thanks & cg Lucas!
lucasart wrote:And it's not even very strong at tactics (relatively to other programs, not by human standard obviously)
You're kidding us, right? :wink:
lucasart
Posts: 3242
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

Re: DiscoCheck 4.0.0 (release candidate)

Post by lucasart »

I meant that it's not strong at tactics, *compared* to other computers of the same ELO level. It "somehow" compensates by playing well in the endgame. But as an opponent to human players, it certainly is strong enough at tactics.
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
Pablo Vazquez
Posts: 154
Joined: Thu May 31, 2007 9:05 pm
Location: Madrid, Spain

Re: DiscoCheck 4.0.0 (release candidate)

Post by Pablo Vazquez »

Thanks Lucas, I hope you keep improving it.
User avatar
Werner
Posts: 2997
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: DiscoCheck 4.0.0 (release candidate)

Post by Werner »

Hi Lucas,
thanks for the engine.
Sorry, does not run here - missing dll
libgcc_s_sjlj

can you please compile once more - or help how to use the exe?

best wishes
Werner
Werner
User avatar
velmarin
Posts: 1600
Joined: Mon Feb 21, 2011 9:48 am

Re: DiscoCheck 4.0.0 (release candidate)

Post by velmarin »

Thanks for relase.

Someone tried it does not work in windows 7,
missing dll, libgcc_s_sjlj-1.dll
I download by http://dll-files.org/7947/libgcc_s_sjlj-1.dll.html

Now I get this error.
the application failed to initialize
(0xc000007b)

Thanks.
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: DiscoCheck 4.0.0 (release candidate)

Post by zullil »

lucasart wrote:New version of my engine DiscoCheck. You can find the source code (licensed GPL v3+) and binaries here:
https://github.com/lucasart/chess

Binaries are in the 'bin' folder:
* Only Linux and Windows 64-bit binaries for the moment.
* MacOSX users who have gcc installed (recent version supporting ISO C++11) should be able to compile it using 'make.sh', although I don't have a Mac, so I haven't tried.
Needed to use gcc-4.7 installed via MacPorts, but it compiled without a warning on my MacBook running 10.7 (Lion). Gave it Marek's

[d]1Rb1r3/2P4k/6p1/pP2qp1p/4p2P/8/B4PP1/3Q1K2 w - -

as a test:

Code: Select all

uci
id name DiscoCheck 4.0.0
id author Lucas Braesch
option name Hash type spin default 32 min 1 max 8192
option name Clear Hash type button
uciok
setoption name Hash value 512
position fen 1Rb1r3/2P4k/6p1/pP2qp1p/4p2P/8/B4PP1/3Q1K2 w - -
isready
readyok
go infinite
info score cp 258 depth 1 nodes 65 time 0 pv d1c1
info score cp 230 depth 2 nodes 196 time 1 pv d1c1 a5a4
info score cp 246 depth 3 nodes 742 time 2 pv d1c1 h7g7 a2b3
info score cp 229 depth 4 nodes 1223 time 3 pv d1c1 h7g7 a2b3 f5f4
info score cp 245 depth 5 nodes 2196 time 5 lowerbound
info score cp 233 depth 5 nodes 2980 time 7 pv d1c1 h7g7 f1g1 e5f6 g2g3
info score cp 240 depth 6 nodes 6000 time 13 pv d1c1 e5f6 g2g3 h7g7 b5b6 a5a4
info score cp 256 depth 7 nodes 8617 time 16 lowerbound
info score cp 266 depth 7 nodes 12722 time 22 pv d1c1 h7g7 f1g1 e5f6 b5b6 f6h4 b6b7
info score cp 234 depth 8 nodes 19979 time 32 upperbound
info score cp 218 depth 8 nodes 22027 time 35 upperbound
info score cp 209 depth 8 nodes 25999 time 41 pv d1c1
info score cp 232 depth 9 nodes 47507 time 72 pv d1c1
info score cp 230 depth 10 nodes 71372 time 104 pv d1c1 h7g7 f1g1 f5f4 b5b6 e4e3 f2e3 f4e3 b6b7 e3e2
info score cp 214 depth 11 nodes 106841 time 150 upperbound
info score cp 198 depth 11 nodes 175858 time 246 upperbound
info score cp 183 depth 11 nodes 246996 time 342 pv d1c1 e4e3 f2e3 f5f4 a2f7 e8f8 e3f4 f8f7 b8c8 f7f4 f1g1
info score cp 141 depth 12 nodes 345603 time 471 pv d1c1 e4e3 f2e3 f5f4 e3f4 e5e2
info score cp 157 depth 13 nodes 389688 time 529 lowerbound
info score cp 173 depth 13 nodes 414758 time 560 lowerbound
info score cp 205 depth 13 nodes 455168 time 615 lowerbound
info score cp 241 depth 13 nodes 594357 time 791 pv d1c1
info score cp 135 depth 14 nodes 2298564 time 3032 pv b5b6 c8a6 f1g1 e8b8
info score cp 119 depth 15 nodes 2787037 time 3622 upperbound
info score cp 125 depth 15 nodes 3429178 time 4457 pv b5b6 c8a6 f1g1 e8b8
info score cp 116 depth 16 nodes 6578294 time 8415 pv b5b6 c8a6 f1g1 e8b8
info score cp 132 depth 17 nodes 7712975 time 9789 lowerbound
info score cp 148 depth 17 nodes 10619395 time 13478 lowerbound
info score cp 138 depth 17 nodes 11848047 time 14890 pv b5b6 c8a6 f1g1
info score cp 202 depth 18 nodes 20115703 time 24546 lowerbound
info score cp 218 depth 18 nodes 20773411 time 25207 lowerbound
info score cp 250 depth 18 nodes 21538805 time 26000 lowerbound
info score cp 314 depth 18 nodes 22644239 time 27225 lowerbound
info score cp 442 depth 18 nodes 26130687 time 31075 lowerbound
info score cp 498 depth 18 nodes 29610600 time 34874 pv b8c8 e5b5
info score cp 511 depth 19 nodes 32484188 time 37880 pv b8c8
info score cp 527 depth 20 nodes 36283169 time 41970 lowerbound
info score cp 543 depth 20 nodes 39107757 time 44977 lowerbound
info score cp 575 depth 20 nodes 44124822 time 50516 lowerbound
info score cp 639 depth 20 nodes 51703141 time 58972 lowerbound
info score cp 495 depth 20 nodes 52533156 time 59886 upperbound
info score cp 428 depth 20 nodes 54031274 time 61498 pv b8c8
info score cp 429 depth 21 nodes 63222449 time 70551 pv b8c8
info score cp 413 depth 22 nodes 70534574 time 78583 upperbound
info score cp 397 depth 22 nodes 73095431 time 81358 upperbound
info score cp 365 depth 22 nodes 90670626 time 101621 upperbound
info score cp 301 depth 22 nodes 102037297 time 114555 upperbound
info score cp 173 depth 22 nodes 121508110 time 137174 upperbound
stop
lucasart
Posts: 3242
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

Re: DiscoCheck 4.0.0 (release candidate)

Post by lucasart »

zullil wrote:
lucasart wrote:New version of my engine DiscoCheck. You can find the source code (licensed GPL v3+) and binaries here:
https://github.com/lucasart/chess

Binaries are in the 'bin' folder:
* Only Linux and Windows 64-bit binaries for the moment.
* MacOSX users who have gcc installed (recent version supporting ISO C++11) should be able to compile it using 'make.sh', although I don't have a Mac, so I haven't tried.
Needed to use gcc-4.7 installed via MacPorts, but it compiled without a warning on my MacBook running 10.7 (Lion). Gave it Marek's

[d]1Rb1r3/2P4k/6p1/pP2qp1p/4p2P/8/B4PP1/3Q1K2 w - -

as a test:

Code: Select all

uci
id name DiscoCheck 4.0.0
id author Lucas Braesch
option name Hash type spin default 32 min 1 max 8192
option name Clear Hash type button
uciok
setoption name Hash value 512
position fen 1Rb1r3/2P4k/6p1/pP2qp1p/4p2P/8/B4PP1/3Q1K2 w - -
isready
readyok
go infinite
info score cp 258 depth 1 nodes 65 time 0 pv d1c1
info score cp 230 depth 2 nodes 196 time 1 pv d1c1 a5a4
info score cp 246 depth 3 nodes 742 time 2 pv d1c1 h7g7 a2b3
info score cp 229 depth 4 nodes 1223 time 3 pv d1c1 h7g7 a2b3 f5f4
info score cp 245 depth 5 nodes 2196 time 5 lowerbound
info score cp 233 depth 5 nodes 2980 time 7 pv d1c1 h7g7 f1g1 e5f6 g2g3
info score cp 240 depth 6 nodes 6000 time 13 pv d1c1 e5f6 g2g3 h7g7 b5b6 a5a4
info score cp 256 depth 7 nodes 8617 time 16 lowerbound
info score cp 266 depth 7 nodes 12722 time 22 pv d1c1 h7g7 f1g1 e5f6 b5b6 f6h4 b6b7
info score cp 234 depth 8 nodes 19979 time 32 upperbound
info score cp 218 depth 8 nodes 22027 time 35 upperbound
info score cp 209 depth 8 nodes 25999 time 41 pv d1c1
info score cp 232 depth 9 nodes 47507 time 72 pv d1c1
info score cp 230 depth 10 nodes 71372 time 104 pv d1c1 h7g7 f1g1 f5f4 b5b6 e4e3 f2e3 f4e3 b6b7 e3e2
info score cp 214 depth 11 nodes 106841 time 150 upperbound
info score cp 198 depth 11 nodes 175858 time 246 upperbound
info score cp 183 depth 11 nodes 246996 time 342 pv d1c1 e4e3 f2e3 f5f4 a2f7 e8f8 e3f4 f8f7 b8c8 f7f4 f1g1
info score cp 141 depth 12 nodes 345603 time 471 pv d1c1 e4e3 f2e3 f5f4 e3f4 e5e2
info score cp 157 depth 13 nodes 389688 time 529 lowerbound
info score cp 173 depth 13 nodes 414758 time 560 lowerbound
info score cp 205 depth 13 nodes 455168 time 615 lowerbound
info score cp 241 depth 13 nodes 594357 time 791 pv d1c1
info score cp 135 depth 14 nodes 2298564 time 3032 pv b5b6 c8a6 f1g1 e8b8
info score cp 119 depth 15 nodes 2787037 time 3622 upperbound
info score cp 125 depth 15 nodes 3429178 time 4457 pv b5b6 c8a6 f1g1 e8b8
info score cp 116 depth 16 nodes 6578294 time 8415 pv b5b6 c8a6 f1g1 e8b8
info score cp 132 depth 17 nodes 7712975 time 9789 lowerbound
info score cp 148 depth 17 nodes 10619395 time 13478 lowerbound
info score cp 138 depth 17 nodes 11848047 time 14890 pv b5b6 c8a6 f1g1
info score cp 202 depth 18 nodes 20115703 time 24546 lowerbound
info score cp 218 depth 18 nodes 20773411 time 25207 lowerbound
info score cp 250 depth 18 nodes 21538805 time 26000 lowerbound
info score cp 314 depth 18 nodes 22644239 time 27225 lowerbound
info score cp 442 depth 18 nodes 26130687 time 31075 lowerbound
info score cp 498 depth 18 nodes 29610600 time 34874 pv b8c8 e5b5
info score cp 511 depth 19 nodes 32484188 time 37880 pv b8c8
info score cp 527 depth 20 nodes 36283169 time 41970 lowerbound
info score cp 543 depth 20 nodes 39107757 time 44977 lowerbound
info score cp 575 depth 20 nodes 44124822 time 50516 lowerbound
info score cp 639 depth 20 nodes 51703141 time 58972 lowerbound
info score cp 495 depth 20 nodes 52533156 time 59886 upperbound
info score cp 428 depth 20 nodes 54031274 time 61498 pv b8c8
info score cp 429 depth 21 nodes 63222449 time 70551 pv b8c8
info score cp 413 depth 22 nodes 70534574 time 78583 upperbound
info score cp 397 depth 22 nodes 73095431 time 81358 upperbound
info score cp 365 depth 22 nodes 90670626 time 101621 upperbound
info score cp 301 depth 22 nodes 102037297 time 114555 upperbound
info score cp 173 depth 22 nodes 121508110 time 137174 upperbound
stop
This is odd, I don't have the same node counts here. Can you run this command, and see what node count comes out ?

Code: Select all

./discocheck bench 12
Thank you
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: DiscoCheck 4.0.0 (release candidate)

Post by zullil »

lucasart wrote: This is odd, I don't have the same node counts here. Can you run this command, and see what node count comes out ?

Code: Select all

./discocheck bench 12
Thank you
Hi Lucas,

Code: Select all

signature = 7429915
time = 10.6312
0
0
lucasart
Posts: 3242
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

Re: DiscoCheck 4.0.0 (release candidate)

Post by lucasart »

Werner wrote:Hi Lucas,
thanks for the engine.
Sorry, does not run here - missing dll
libgcc_s_sjlj

can you please compile once more - or help how to use the exe?

best wishes
Werner
I think the cross compiler (mingw-w64) could be at fault here. Here's another attempt, this time compiling libgcc statically:

Code: Select all

$ x86_64-w64-mingw32-g++ ./src/*.cc -o ./bin/DC401_win64.exe -static-libgcc -DNDEBUG -std=c++0x -O3 -flto -fno-rtti -Wall
I don't have Windows, so I can't really verify myself if it works. But I've put the new executable in ./bin/DC401_win64.exe
If that trick still fails, then I'm out of tricks :cry:
Perhaps Jim Ablett can come and rescue me
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.