Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: In man vs. machine, doubling speed added about 30 FIDE rating points

Author: Brian Kostick

Date: 13:25:43 03/01/01

Go up one level in this thread


On March 01, 2001 at 15:26:00, Jonathan Lee wrote:

>With all due respect to John Merlino, upon the actual ratings of parallel
>processors against the top ten chess playing humans in the world seemed to add
>30 FIDE rating points for double the speed.
>
>Fritz on 4 GHZ (game in 30 minutes) had a FIDE rating of about 2720. (Without
>Shirov's 2 games it had 2750 FIDE)
>
>Junior on 5.6 GHZ (tournament time control) had a FIDE rating of about 2700.
>(Without Leko's loss it had 2690 FIDE)
>
>If you do the math, you get
>450 MHZ  2600 FIDE  (more than 2600 if game in 30 minutes)
>1 GHZ    2630 FIDE  ("     "   2630 "   "    "  "  "     )
>2 GHZ    2660 FIDE  ("     "   2660 "   "    "  "  "     )
>4 GHZ    2690 FIDE  ("     "   2690 "   "    "  "  "     )
>If you read my article, on parallel processors and their losses the GM's will
>learn from their mistakes and the rating of Fritz and Junior should slide down a
>few (or a lot of) points.
>Jonathan (104 message)

  I am just curious. I have a limited understanding of multiple CPU and their
co-operation. But who is the manufacturer of the 4GHz CPU?? Thanks in advance,
Brian K.



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.