GNU Chess

GNU Chess is a free chess-playing program developed as part of the GNU project of the Free Software Foundation. The first version of GNU Chess was written by Stuart Cracraft. Fabien Letouzey is the primary author of GNU Chess 6, which is based on Fabien's chess engine Fruit 2.1.
http://ftp.gnu.org/pub/gnu/

Gnuchess.bin
http://hardy.uhasselt.be/Toga/gnuchess-release/gnuchess.bin

Gnuchess5 5.07.174b

- opening book: 'gnuchess.bin' updated
- various tunings (piece values, mobility)
- replace old PST's by those of Fruit 2.1
- now about the same strength as Fruit and hence as GNU Chess 6

http://hardy.uhasselt.be/Toga/gnuchess-release/

Gnuchess5 5.07.174.1b
http://hardy.uhasselt.be/Toga/gnuchess-release/
- this version is identical to 5.07.174b but with the EGTB code of Miguel Ballicora included
- opening book: 'gnuchess.bin' again updated

GNUchess v6.02 for Android
http://freeshell.de/~oesser/gnuchess-for-android/

GnuChess 6.0.3
http://ftp.gnu.org/pub/gnu/chess/

Version 6.0.3 is primarily a bug fix release. What is fixed in v6.0.3:

- If the program is in manual mode, an 'undo' action will keep it in manual mode.
- 'configure' and Makefile.in have been generated with autoconf 2.69 and automake 1.13.
- Remove some unused code.

Gnuchess5 5.0.175b (13-Apr-2013)

http://hardy.uhasselt.be/Toga/gnuchess-release/

The .zip file contains the source as well as binaries for linux32/64, windows32/64 and android.

* SwapOffSign implemented and used almost everywhere instead of
SwapOff (GNU Chess parlance for SEE).
* Optimized futility pruning (do pruning without doing the move) both
in main search and quiescence search (big elo gain).
* --sig and --version print full name.
* Only extend checks on non-negative SEE.
* SEE pruning (Lucas Braesch). Nice elo gain.
* CLOP tuning of king center control in endgame.
* CLOP tuning of aspiration search margins.
* History ageing (another nice elo gain).
* Null threat extension (Lucas Braesch).
* Evaluation cache implemented (size still needs to be tuned).
* Very rudimentary implementation of KRKP.
* Rudimentary implementation of KRKRP.
* CLOP tuning of futility margins.
* CLOP tuning of passed pawn eval.
* Free passer bonus (nice elo gain).
* Reduced delta margin.

would be really cool if GNU was multi processing, to see how it would fare with other engines MP ... best

GNU Chess 5.50
http://hardy.uhasselt.be/Toga/gnuchess-release/

Michel Van den Bergh wrote:

According to my testing GNU Chess 5.50 is about 120 elo (+-20) stronger than Fruit 2.1 at 40/4 time control, 128Mb hash.

I think it is a good milestone since it is the first GNU Chess 5 which is substantially stronger than GNU Chess 6.

GnuChess6 6.1.0
- available as a tar.gz file only
- bugfixes and translation/internationalisation work
- added 'pgnreplay' command
http://ftp.gnu.org/pub/gnu/chess/

GnuChess6 6.1.1
- available as a tar.gz file only
- include translations to other languages
- minor corrections in translatable strings
- migrate frontend to C++
- use autoconf 2.69 and require it in configure.ac
- use automake 1.14
http://ftp.gnu.org/pub/gnu/chess/

GnuChess6 6.1.1 TM w32
http://tmokonen.net84.net/gnuchess-6.1.1.rar

GnuChess6 6.1.2

- available as a tar.gz file only
- include translations to other languages
- minor corrections in translatable strings
- migrate frontend to C++
- use autoconf 2.69 and require it in configure.ac
- use automake 1.14

http://ftp.gnu.org/pub/gnu/chess/

Hi all,
indeed Gnu Chess 6 is Fruit+Polyglot...

Extract from manual
'PolyGlot 1.4 has been adapted and incorporated to GNU Chess v6 as chess engine protocol adapter. It connects Fruit-based GNU Chess engine to the good old GNU Chess frontend'

And this is how it has been incorporated: 'PolyGlot and Fruit run as two independent single-threaded processes. PolyGlot starts first, and it forks Fruit. Both processes get communicated by means of pipes: PolyGlot captures Fruit’s standard input and output.'

As a result, watch your core usage: I tested GC 6 under Arena and got one core permanently used by Polyglot, the other by the engine was either permanently used (ponder on) or only while thinking.

Running Fruit and GC 6 with one core each, testing identical positions, I got the same play with GC 6 showing half the nodes per second of Fruit... Frustrating. The Androïd version of GC 6 behaves the same way. And that is perhaps why GC 5 can show itself stronger than GC 6? 

Thanks a lot for the information.

Werner

GnuChess6 6.1.2 TM
http://tmokonen.net84.net/gnuchess-6.1.2x32TM.rar

Is it possible to use tbs with this version?

How can I set the path?

Werner

Michel Van den Bergh wrote:

Here is GNU Chess 5.60.

http://hardy.uhasselt.be/Toga/gnuchess-release/

The main point is the inclusion of a .eng file to support HGM's idea for seamless GUI/engines integration on Linux. To the other engine and GUI authors: please support this! Even if you believe the standard isn't perfect yet. The wrinkles can be sorted out later.

I stopped working on GNU Chess. However digging through my old testing results (which are now a mess) I think this is about 50 Elo stronger that 5.50.

Here is a battle between GNU Chess 5 and GNU Chess 6 (the official version).

GnuChess6 6.2.1 (available as a tar.gz file only)
http://ftp.gnu.org/pub/gnu/chess/gnuchess-6.2.1.tar.gz

GNUchess6 6.2.1 JA
mega.co.nz #!CVF0AZCL!nkqWgheK7hx0QVKRJHVIbzDWHxnEwECGx7ZMxfG1Q3U