Computer Chess Club Archives




Subject: Re: Couple of chess programming questions - MDT and parallel

Author: Robert Hyatt

Date: 08:44:03 09/11/02

Go up one level in this thread

On September 10, 2002 at 23:46:01, scott farrell wrote:

>What are people's opionions on MTD and parrallel.
>Platt states its a good idea. I think's its a good idea.
>If MTD can be shown to be - equally as good as PVS, and it scales better accross
>CPU/multiple machines, it might work better the more CPU/machines added.
>I understand the problem that MTD relies more heavily on hashtable then other
>algorithms, but I think they all rely on hashtables pretty heavily.

it has been done.  Don D. had a parallel version of mtd(f) running and
it has some interesting characteristics.  But, if you have to do a bunch
of the searches, it suffers from the same problem it would on a single
cpu...  I think this works for certain approaches only, and large positional
scores are not a good idea, at least from my experience, which is not as
much as several others here that are really using mtd(f).

This page took 0.09 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.