It's usually better to run remote matches than local matches with remote engines.
Yes I found that out the hard way and I am going to do that, I am just in the process of writing ftp scritps to 7-zip the various books send them (along with polyglot.ini files and cutechess bat files) to the remote machines (I have over 30 engines dotted around the place), unzip them to the right places and then start the cutechess matches by starting a windows scheduled task. I found a windows scheduled task works but running a batch file via ssh just wouldn't work to start cutechess for some reason (even with cmd.exe /k invocation). Then I have batch scripts to end the cutechess matches and get the games back to the central book machine. I am doing this because of the timeout problems, running things from a central machine makes it easier on logistics, but as you say it won't work in the long run over a poor network connection.
The project is testing my aquarium book against other polyglot books, the polyglot books are called A00.bin, A01.bin. all the way to...E99.bin, made in batch using pgn-extract on a big database by Eco code, this way it ensures my book gets to play against all Ecos. Good winning games go into my book.
Quick question - does cutechess-cli support the "old" winboard/ xboard move syntax? This was originally in place for gnuchess, but other engines may use it.
Essentially, instead of outputting
move X
(where X is a valid move, e.g. e2e4), it is permissible for an engine to output
Y ... X
(where Y is the move number and X is the move, e.g. 1 ... e2e4)
None of the backward compatibility stuff is supported by Cute Chess, and Xboard protocol 1 support is incomplete (eg. there's no "edit" command). Protocol 2 is well-supported, but the implementation is very strict, meaning that weird move notations or syntax aren't accepted.
When I have time I'll look into improving the protocol 1 implementation and backward compatibility, but it's not a top priority.
ilari wrote:None of the backward compatibility stuff is supported by Cute Chess, and Xboard protocol 1 support is incomplete (eg. there's no "edit" command). Protocol 2 is well-supported, but the implementation is very strict, meaning that weird move notations or syntax aren't accepted.
When I have time I'll look into improving the protocol 1 implementation and backward compatibility, but it's not a top priority.
well that's fair enough! thank you for the great little utility.
ilari wrote:None of the backward compatibility stuff is supported by Cute Chess, and Xboard protocol 1 support is incomplete (eg. there's no "edit" command). Protocol 2 is well-supported, but the implementation is very strict, meaning that weird move notations or syntax aren't accepted.
When I have time I'll look into improving the protocol 1 implementation and backward compatibility, but it's not a top priority.
I suggest to keep it strict, or at least, have a switch to have option to run it strict (--pedantic). All the flexibility that xboard has had did not encourage to detect/fix problems in engines. I think that it ended up being counter productive.
In fact, I would prefer that someone else writes an adapter WBII-strict/WB-relaxed+gnuchess and you do not bloat your software. If we had one like that before, maybe the development of WBII would have been faster and UCI would have never been needed.
michiguel wrote:I suggest to keep it strict, or at least, have a switch to have option to run it strict (--pedantic). All the flexibility that xboard has had did not encourage to detect/fix problems in engines. I think that it ended up being counter productive.
I fully agree. Whenever there has been a non-compliant engine, Winboard/Xboard (the application) has been hacked to support the engine. That is backwards, and I've complained about it before. I'm trying to keep the implementation as strict as possible, but not without sacrificing WB1 support. The second protocol version is better, but it too leaves so many important cases undefined that making a strict implementation is difficult.
Thank for this new release, I yet have to give cutechess-cli a try. Meanwhile I have trouble accessing the windows binary release site, would you mind to upload that in another place?
Thank for this new release, I yet have to give cutechess-cli a try. Meanwhile I have trouble accessing the windows binary release site, would you mind to upload that in another place?
Sorry to bother but I can not access original link for win32 download. Can some one upload this in another place like http://www.mediafire.com ?
I also decided to give up using Arena for test games. Unfortunately it sometimes seems to half load engines where they play at 2 to 10 times slower which really affects the outcome of matches.
Thank for this new release, I yet have to give cutechess-cli a try. Meanwhile I have trouble accessing the windows binary release site, would you mind to upload that in another place?
Sorry to bother but I can not access original link for win32 download. Can some one upload this in another place like http://www.mediafire.com ?
I also decided to give up using Arena for test games. Unfortunately it sometimes seems to half load engines where they play at 2 to 10 times slower which really affects the outcome of matches.
Ok, I finally managed to get it by some tricks, so dont mind. Not that any one in particular cared. I hope Dr Deeb would not read this though. First I was repeating myself, now talking to myself, what next?