Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: chess computer ratings

Author: Robert Hyatt

Date: 06:53:13 09/24/01

Go up one level in this thread


On September 24, 2001 at 04:14:30, Jouni Uski wrote:

>On September 24, 2001 at 03:12:18, Tony Hedlund wrote:
>
>>On September 23, 2001 at 09:04:34, gregory j capace wrote:
>>
>>>If you have a faster processor, how much strength does this dd to the program ?
>>>Does Fritz 6 run any stronger on my 566 mghz, versus a 450 nghz. , like the
>>>rating says. How strong would it be on a 1.2 mghz. computer ?
>>
>>The step from 450 to 1200 MHz have gained 75 points so far. The step from 200 to
>>450 gained 79 points.
>>
>>Tony
>
>Very interesting! Because 1200/450 = 2.67 and 450/200 = 2.25 there seems to be
>diminishing return now?! Also I expect Fritz/Tiger to get 2725 rating in next
>list. Hmm. What says Bob?
>
>Jouni


First, Bob says 2725 has nothing to do with reality.

Second, not enough data.  Doubling the clock rate does _not_ double the cpu
speed.  Which means this is an apples and oranges discussion.  I've always
used the expression "doubling the cpu speed produces about 60-70 rating
points."  That does not mean "doubling the cpu clock speed produces ..."

Just compare your favorite program on a 200mhz machine vs a 400 mhz machine.
In some cases, the 400 might be > 2x faster (ie an old pentium 200/mmx compared
to a PII/400 (which is based on the pentium pro core, much better processor than
the original pentium core).

So the clock speed isn't very interesting.  The program's NPS is a better
measure of speed improvement and rating increase.



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.