Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: ChessMaster 9000 & hash table settings

Author: Mike Byrne

Date: 16:58:36 10/12/02

Go up one level in this thread


On October 12, 2002 at 18:22:00, James T. Walker wrote:

>Hello Kurt,
>Thanks for filling me in but the explanation does not seem to cover what I'm
>talking about.  I understand the "solution time" may vary with different size
>hash tables but why play different moves at the same depth?  This makes no sense
>to me unless the hash tables are filling up quickly.  This is not likely with
>ChessMaster.  It's like CM has a built in randomizer to choose between moves
>which are close to the same score.
>Again thanks for the response.
>Jim

Hi Jim,

Different hash tables sizes are a built in randomizer to a certain extent.  I
see this issue all the time when compiling different crafty versions.  Factor in
SMP and there is even more variation on solution times.  That is why the best
way to evaluate different versions of the same program is not through chess
problems but from games.  I have seen cases where solution times for certain may
go up, but in game situations it plays better.  It's just the way it is.

Mike

Mike



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.