Author: Anthony Cozzie
Date: 11:04:05 08/19/03
Go up one level in this thread
On August 19, 2003 at 13:02:33, David H. McClain wrote: >Has anyone ever calculated approximate rating differences in points between >programs run on a single processor versus the same program run on multiple >processor computers? It would be interesting to evaluate the Deep Junior >program played against Kasparov on a eight processor machine versus the same >single processor version program played on a single processor machine. What >percentage would a program's strength increase by adding processors? Is this >known? > >Would evaluations be similar for other "deep" multiple processor built programs >versus their same version built for single threads? > >Dave I seem to recall 50 ELO/2x processor speed as a rule of thumb. Considering search inefficiencies, I think that comes out to about 125 ELO. anthony
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.