Ares 1.0 in honor of the NASA Ares I-X Space Rocket
Posted: Sat Oct 31, 2009 4:38 pm
Happy Halloween!
I felt inspired by the coming holidays and the NASA efforts.
So, I am releasing a chess engine in honor of the NASA Ares I-X space rocket.
It runs under MS OS's and was developed under Linux.
I may release a Linux executable, if there is a demand.
It is called Ares 1.0 and it should be available on a few sites today.
Thorsten's site: http://www.thorstenczub.de/ares.html (btw, neat looking)
Leo's site: http://wbec-ridderkerk.nl/html/details1/Ares.html
Thanks go to these men for hosting it!
Ares supports the UCI protocol, but it doesn't support multiPV mode. It does support a minimal set of the Xboard protocol. Pondering works
in UCI but not in Xboard and it does not have an opening book. I've tested the UCI mode in Fritz, Arena, SCID, Polyglot,
ChessPartner. Dasher and ChessAssistant 10. In CA 10, you get sporadic output from the engine in Analysis mode and while playing.
It doesn't seem to effect the playing strength.
A week ago. I sent email to Convekta and they told me a programmer would get in touch with me. No word yet.
Does anybody know how to monitor the communications between CA 10 and an engine?
This is the 32 bit windows executable; the 64 bit version is faster and stronger by 2x and about 120 Elo. What is Ares' playing strength?
Ares 1.0 is stronger than NoonianChess by about 600 Elo. That is a rough estimate. A quick test would be to run the internal benchmark.
It uses the same benchmark technique as NoonianChess. Just type: Ares1.0.exe -bench -fd 11 . If you want to compare NoonianChess performance to Ares,
I would use -fd 9 as the test depth because NoonianChess doesn't search as deep or as fast.
It was designed with particular openings in mind. The KIA as white and Sicilian as black. It seems best at closed and semi-open
positions. It may be better at other openings. I don't have enough test results with all openings to say.
While this version of Ares is being distributed without charge, I forbid cloning it in any way.
I felt inspired by the coming holidays and the NASA efforts.
So, I am releasing a chess engine in honor of the NASA Ares I-X space rocket.
It runs under MS OS's and was developed under Linux.
I may release a Linux executable, if there is a demand.
It is called Ares 1.0 and it should be available on a few sites today.
Thorsten's site: http://www.thorstenczub.de/ares.html (btw, neat looking)
Leo's site: http://wbec-ridderkerk.nl/html/details1/Ares.html
Thanks go to these men for hosting it!
Ares supports the UCI protocol, but it doesn't support multiPV mode. It does support a minimal set of the Xboard protocol. Pondering works
in UCI but not in Xboard and it does not have an opening book. I've tested the UCI mode in Fritz, Arena, SCID, Polyglot,
ChessPartner. Dasher and ChessAssistant 10. In CA 10, you get sporadic output from the engine in Analysis mode and while playing.
It doesn't seem to effect the playing strength.
A week ago. I sent email to Convekta and they told me a programmer would get in touch with me. No word yet.
Does anybody know how to monitor the communications between CA 10 and an engine?
This is the 32 bit windows executable; the 64 bit version is faster and stronger by 2x and about 120 Elo. What is Ares' playing strength?
Ares 1.0 is stronger than NoonianChess by about 600 Elo. That is a rough estimate. A quick test would be to run the internal benchmark.
It uses the same benchmark technique as NoonianChess. Just type: Ares1.0.exe -bench -fd 11 . If you want to compare NoonianChess performance to Ares,
I would use -fd 9 as the test depth because NoonianChess doesn't search as deep or as fast.
It was designed with particular openings in mind. The KIA as white and Sicilian as black. It seems best at closed and semi-open
positions. It may be better at other openings. I don't have enough test results with all openings to say.
While this version of Ares is being distributed without charge, I forbid cloning it in any way.