Author: blass uri
Date: 11:16:12 06/01/99
Go up one level in this thread
On June 01, 1999 at 13:00:48, Melvin S. Schwartz wrote: > >On May 31, 1999 at 20:06:48, Manuel J. Petit de Gabriel wrote: > >> >>On May 31, 1999 at 19:33:28, Melvin S. Schwartz wrote: >> >> >>>best. As for ratings, I also believe testing programs on equal hardware is the >>>optimum way to achieve the most accurate results. >> >>The question then could be, on what hardware? >> >>There are a lot of knowledgable here than can correct me if I'm wrong, >>but I think that the relative strength of engines is dependant on >>the box you run them. Program A can be better than program B when running >>on a given computer, but when run on a box twice as fast program B can be >>stronger. >> >>I feel that root processors will lose relative strength to programs that >>does full evaluations at the end of the search when the CPU speed is doubled. >>Am I right? Or am I missing something? >> >> >>manuel, > >Hi Manuel! > >When I speak of comparing program A against program B on equal hardware - I am >referring to running both programs on identical computers! > >Mel The problem is that A may be better than B on pentium200MMX when B is better than A on pentium90 because some programs earn more from time relative to other programs. This is what Manual means by saying that the relative strength of engines is dependant on the box you run them. Uri
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.