Well, that is exactly how it uses the loadGameFile. It forces the moves until they run out, and then the engines take over.Michel wrote:Like any GUI uses a book. The GUI uses the book until it runs out and then gives control to the engine. This is easy to implement using force mode. I did this for my own FICS client (an icsdrone derivative).
Arena or similar testing program for Linux
Moderators: hgm, Rebel, chrisw
-
- Posts: 27894
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: Arena or similar testing program for Linux
-
- Posts: 2273
- Joined: Mon Sep 29, 2008 1:50 am
Re: Arena or similar testing program for Linux
Hmm that seems to be quite different from an opening book. In xboard 4.2.7 (4.3 has not been packaged for Ubuntu) it seems that one of the games from the gamefile is selected, the moves are played and then the engines take over.Well, that is exactly how it uses the loadGameFile. It forces the moves until they run out, and then the engines take over
This may be fine for engine-engine matches but it is not what you want against a human.
It is also not suitable for testing different opening books against each other.
Michel
-
- Posts: 224
- Joined: Mon Mar 12, 2007 7:31 pm
- Location: Bonn, Germany
Re: Arena or similar testing program for Linux
Shredder GUI runs fine under Wine. Arena does not.OliverBr wrote: I am urgently looking for a program that let chess engines play each other multiple times (like Arena for Windows), but that it running on Linux.
-
- Posts: 27894
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: Arena or similar testing program for Linux
For testing opening books against each other, each side should have its own (different) book, not? What would be the point of having the GUI handle the book? I thought the whole idea of having the GUI handle the openings was to make them come from a single source.
Likewise, if a Human plays an engine, you would want the engine to use a book, not the Human. So what is against having the engine hande the book?
Likewise, if a Human plays an engine, you would want the engine to use a book, not the Human. So what is against having the engine hande the book?
-
- Posts: 2273
- Joined: Mon Sep 29, 2008 1:50 am
Re: Arena or similar testing program for Linux
Nothing. But I prefer the GUI handling the book. I have the impression that this is what most people prefer but I maybe wrong. Often you don't like the book that comes with the engine.Likewise, if a Human plays an engine, you would want the engine to use a book, not the Human. So what is against having the engine hande the book?
Under UCI you have the "OwnBook" option to select between the two possibilities.
-
- Posts: 27894
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: Arena or similar testing program for Linux
Under WinBoard, UCI engines run through Polyglot, and use their own book or the Polyglot book just as you mention. If I want WinBoard engines to run with an external book, I run them through WB2UCI and Polyglot. I don't think there is much that you cannot do this way.
-
- Posts: 2273
- Joined: Mon Sep 29, 2008 1:50 am
Re: Arena or similar testing program for Linux
.If I want WinBoard engines to run with an external book, I run them through WB2UCI and Polyglot
Well that's a bit of a hack isn't it, using two(!) adapters to do what the GUI could be doing?
Not to mention that WB2UCI only runs on windows and is closed source AFAICS.
-
- Posts: 27894
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: Arena or similar testing program for Linux
Most engines also are closed source and run only under Windows. I thought that Linux users could still run them, using an emulator.
One cold als say, of course, that it is a hack to do something the engine could (should?) be doing. Al these engines for thinking up moves are a hassle anyway. It would be much more eficient to have the GUI do that too...
I still don't see much application for external books outside engine-engine play. If a Human would want to switch engines after the opening, copying the game to the clipboard and pasting it back after starting the other engine is just as easy as setting up the GUI to handle the book. And it would work with any book for which you can run an engine that understands it, not just for a format the GUI happens to know.
If your purpose is to compare books, the most logical way seems to run two teams of UCI engines against each other, each team using a different Polyglot book. The only reason I can imagine why you would want to use WinBoard engine for this, it if you are developing a book for that particular engine. In which case you'd better write it in a format that the engine can handle itself.
For forcing engine-engine matches to have the opening variety provided by a given book, I would still recommend the use of the -loadGameFile option, after writing the openings to a PGN file. This allows you to easily redo the match. (E.g. after you made an improved version of one of the engines.)
One cold als say, of course, that it is a hack to do something the engine could (should?) be doing. Al these engines for thinking up moves are a hassle anyway. It would be much more eficient to have the GUI do that too...
I still don't see much application for external books outside engine-engine play. If a Human would want to switch engines after the opening, copying the game to the clipboard and pasting it back after starting the other engine is just as easy as setting up the GUI to handle the book. And it would work with any book for which you can run an engine that understands it, not just for a format the GUI happens to know.
If your purpose is to compare books, the most logical way seems to run two teams of UCI engines against each other, each team using a different Polyglot book. The only reason I can imagine why you would want to use WinBoard engine for this, it if you are developing a book for that particular engine. In which case you'd better write it in a format that the engine can handle itself.
For forcing engine-engine matches to have the opening variety provided by a given book, I would still recommend the use of the -loadGameFile option, after writing the openings to a PGN file. This allows you to easily redo the match. (E.g. after you made an improved version of one of the engines.)
-
- Posts: 725
- Joined: Tue Dec 18, 2007 9:38 pm
- Location: Munich, Germany
- Full name: Dr. Oliver Brausch
Re: Arena or similar testing program for Linux - Own proggie
ChessDM works fine. If anyone is interested in having a Testing Suite that runs platform independent without any GUI and without any CPU usage, please write here.
This is the result after some hours running:
This is the result after some hours running:
Code: Select all
olithink - crafty : 130.5/472 96-69-307 27% (00000000100==11==001=00=0000100010000100===0000100=00=0001000=00010010011000000000=00000=100100000=10=0010000000=00110=000=011000100110=00100000=00000==00=0=11000=100=0010100000=010011100001=0110101000010000011=0010000=011=0==000000000000000==01000010000=01=1=0100001000000010=000=001===10100010000010000110=11000100=100000000010011====0=001000==000=000000000=00011000000=010=110====0=00000010010001011010=0010=00000==010001010000=1001100100010110100110010000=0100=1100000)
-
- Posts: 27894
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: Arena or similar testing program for Linux - Own proggie
Was it too difficult to make it print linefeeds?