Computer Chess Club Archives




Subject: Re: is the

Author: Tom Kerrigan

Date: 10:37:00 08/04/98

Go up one level in this thread

On August 04, 1998 at 11:02:17, Komputer Korner wrote:

>The point is that Tom Kerrigan wants proof that hash tables will improve a
>program by 50 points or a doubling in overall speed. I say they will but he

Yes, I do.
The rule of thumb is that doubling speed gives you 50 points. This is for
tournament time controls, but it's the best we can do for now.
Here's the experiment: take an average computer (say, a PII/266) with 64 MB RAM
and use a popular program (Fritz, maybe?) to analyze a few positions overnight.
Then increase the RAM to 128 MB and repeat. Use the largest hash tables possible
with each amount of RAM. Use positions that somebody would want to analyze
overnight, too. Perhaps positions from tournament games at crucial points? Be
sure to have a fair distribution of middlegame and endgame positions.


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.