Author: Uri Blass
Date: 10:30:55 12/06/03
Go up one level in this thread
On December 06, 2003 at 12:31:41, Alex Szabo wrote: >On December 06, 2003 at 11:22:15, Alex Szabo wrote: > >>My self-play experiment between crafty versions 19.6 and 17.14 resulted in the >>19.6 version being stronger by 11 (plus or minus 2) rating points. >> >>The experiment was for 20,000 games at a 20+1 time control. It was run on a >>2x2GHz pentium4 linux system with xboard 4.2.6. I can include other details if >>there is interest. > >Here are some more details: >(1) 4 matches were run at the same time >(2) pondering was on (ie normally 8 craftys were competing for CPU time) 8 crafties on one cpu? I wonder if we can trust the results. Do all crafty use the cpu all the time? I will be happy to do matches on one cpu with ponder on if I have a way to force the engines to use only half of the cpu even if the opponent waits. If there is a tool to force engines to use half of the cpu then it will be good to use it for games. The tool should notice if one of the engines does not use half of the cpu and immediately use it in case that the engine wait because of having no move to ponder or finding a forced mate. I do not know the exact changes from 17.14 to 19.6 but if one of the changes means that a problem of not pondering in rare cases is corrected then the result is not reliable. In your case you need a tool to tell the 8 engines to use not more than 1/8 of the cpu time but I am going to be happy even with having something that tell the engines always to use not more than 1/2 of the time when the tool use the rest of the cpu time if it does not happen. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.