Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Requesting TSCP 1.41 benchmarks...

Author: Thorsten Czub

Date: 12:41:13 12/09/99

Go up one level in this thread


On December 09, 1999 at 12:01:16, James T. Walker wrote:

>>No.  I remember that when it happened, and there was some discussion.  Who
>>knows, maybe some machines get shipped with different cpu's than labeled.
>>Overall, the numbers give a good idea of what to expect for generic integer
>>programs.
>>
>>Will
>
>Thanks Will,
>The reason I asked is I had a K6-2-400 and I changed the motherboard to a
>K6-3-450.  Not one single program I have ran faster on the K6-2 and Programs
>like Hiarcs and CSTal-2 gained more than the Mhz would indicate.
>Jim Walker

If you have ALI-chipset, it is slower than using VIA-Apollo chipset.
Having 512 KB cache on MB is slower than 1MB.
Also: do you use 100 Mhz or maybe other bus-speed ?
Do you use PS/2 Ram 60ns or do you use DRAM PC100 ?
this all influences the speed.
also each program has different profit benchmarked on a special hardware.
it is right, hiarcs and cstal (as to very big programs) profit
much from the k6-3 very big internal cache.
small programs (fast-searcher) fit in the internal cache anyway.
so a bigger cache = k6-3 will not speed them up.
so k6-3 is the way for slow programs to get tuned :-))



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.