Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Bus Speed V CPU Speed--?`

Author: Matthew Hull

Date: 10:17:48 08/06/02

Go up one level in this thread


On August 06, 2002 at 12:30:15, Mike S. wrote:

>On August 06, 2002 at 10:08:00, Rex wrote:
>
>>Do you receive better performance upgrading from a PC133Mhz bus speed to 333Mhz
>>FSB -V- keeping the PC133 FSB and upgrading CPU 300 or so Mhz?
>>
>>Better performance meaning Chess Programs.
>
>I did some Fritzmark tests with different FSB clock rates and/or* memory clock
>rates, with an Athlon XP. The impact of memory clock speed has been max. 6%.
>I've read from others, the impact of FSB speed & -settings can be up to (but not
>more) than 10%, for chess performance.
>
>*) I can choose mem clock asynchronous to FSB speed, i.e. FSB 133, mem 166.
>
>Also, I have noticed that DDR266 (133 MHz) with *small* latency settings
>provides better Fritzmarks than DDR333 (166 MHz) with *high* (longer) latency
>settings. 166 MHz and small latencies were not running stable, unfortunately.
>
>The CPU clock speed has *much* more impact on the chess performance. The speed
>increase is virtually identical to the clock rate ratio (with *very small*
>differences, depending on the program, position, hash size etc.).
>
>Regards,
>M.Scheidl


I would suspect that a nice backside cache minimizes slow memory performance
hits.  But if you did the same test with a Duron (small backside cache), then
the differences in memory speed might become more apparent.

Regards,
Matt



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.