Author: Stuart Cracraft
Date: 09:03:22 09/05/04
Go up one level in this thread
On September 04, 2004 at 21:20:59, David B Weller wrote: >Hi Stuart, > >I am getting just a shade over 240 also. [1sec/pos on 1.1gz AMD ATHLON XP] > >-David I've expeirmented with 1/2 second per position and it does save time in testing with a surprisingly not-too-distant result. If you don't clear your hash table after every search, that can save time, since a memory clear on a large block of memory takes a significant fraction of a second on these boxes. Your hashing algorithm needs to be be okay with that though. Mine is single tier replace if length.candidate >= depth.stored. And that is okay with not clearing hash table between moves. My hash table for all my testing is only 500,000 entries or so. Quite small. When I try to increase it much, WINDOZE bolixes up and starts thrashing to disk. I am running Windows ME and really need to chuck this box but am waiting (and waiting) for a decent 64-bit box with a decent full 64-bit Windows. Stuart
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.