Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: benchmark hardware setup with a standard position and counter

Author: Richard A. Fowell (fowell@netcom.com)

Date: 21:37:43 12/04/97

Go up one level in this thread


From what you've said, the software should be:
- freely available
- run on multiple operating systems
- capable of taking a given position to a fixed depth.

Alas, the speed of given machines as chess hardware varies with the
chess engine in questions - the fastest machine for Crafty isn't
necessarily the fastest one for HIARCS, for example. Also, the
speed even for the same software and position changes with the position:
In my 32-problem HIARCS benchmark, the relative speed of two hardware
platforms across the 32 problems varies from problem to problem.
For example, the speed ratio on a 180 MHz 604e to that on a 66 MHz
601 varies by 18% over the 32 problems. That is why I use 32 problems
chosen to span tactical, positional, and endgame positions (per Robert
Hyatt's recommendation).

Unfortunately, this means the benchmark takes longer to run and longer
to score, and therefore gets run less often.

fowell@netcom.com (Richard A. Fowell)



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.