Author: Jim Monaghan
Date: 07:27:34 08/08/02
Go up one level in this thread
Hi Victor, My PC has 256 MB of ram, and I was using 32 MB for hash tables. I should clarify a point. On my machine the curve started to flatten out at about 15 seconds. There is of course nothing absolute about that number. On a slower machine, say a P2-800 that point might occur at 30 seconds and on a faster machine, say an AMD 1.8 Ghz, it might occur at 10 seconds. It's all relative, but the pattern should be the same. I think I gave the impression that 15 seconds per move was an "optimal" testing time for engines (verses time spent). This is true perhaps for my Cel 1.3 Ghz only and not for other machines. I'm going to run the test through Crafty in early exit mode. First at 10 seconds per move. Then take the missed ones and run them through for 2 minutes again in early exit mode. The log files should give me better information as to ply depth reached verses time spent. I might be able to improve the table that way. I will post a follow-up if I get better information. Thanks for the interest. Cheers, Jim
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.