Author: Robert Hyatt
Date: 17:24:31 07/30/98
Go up one level in this thread
On July 30, 1998 at 17:16:02, Tom Kerrigan wrote: >On July 30, 1998 at 17:08:33, Robert Hyatt wrote: > >>that's the trivial part. There are *many* more parts needed that are > >Yes, yes, of course it's the trivial part. > >The question is, is that particular trivial part a good idea? > >-Tom It's a solution. I personally don't like individually created processes, because that implies some sort of message-passing approach. Would be the only way to do things on a distributed architecture, or on a multiprocessor which uses message passing, like the intel machines (the ones with large numbers of cpus). That's why I said the processes was the trivial part. The headache is splitting the tree, coordinating alpha/beta, fail-highs, the hash table, etc. IE if you use the hash table to store repetition stuff, rather than a repetition list, get ready for a rewrite. So there's plenty to do, that's for sure.. :)
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.