Author: ujecrh
Date: 11:29:42 01/12/01
Go up one level in this thread
On January 12, 2001 at 13:58:17, John Wentworth wrote: >This may be a silly question but, in order to determine the strength of chess >software , wouldn't it be more accurate to use slower hardware? It seems to me >that when faster hardware is used, it just increases the tactical ability of the >software and does not really show what the software "knows" or "understands", >seems like it makes it harder to gage the knowledge level. If we were to take a >program like Rebel, playing against a well known GM using a slow machine, if the >program could still win, then I would be impressed. >Other than that hardware has a lot to do with it, so let's get the old Cray >Blitz software off the shelf running on a Cray Supercomputer, it would destroy >every PC software running and be at least as competitive against the GM. Two things come to mind: First, running a program on a slower hardware is about the same as running it on fast hardware with less time to search (excepting ponder of course because it depends on your thinking time). Second, some knowledge is sometimes removed from a program because with a decent hardware we know it will find it by pure tactics so what you're saying is probably right for engines designed to run on slower computers (genius comes to mind) but might be wrong for recent engines. ujecrh
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.