Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: L2-Cache and chess programs

Author: Robert Hyatt

Date: 11:00:47 01/24/00

Go up one level in this thread


On January 24, 2000 at 09:09:50, Tom Kerrigan wrote:

>On January 24, 2000 at 03:49:37, David Blackman wrote:
>
>>We've heard claims of "full speed" or "core speed" L2 caches before from various
>>companies. Most often these caches have not been impressively fast. I'll wait
>>until i get hold of a coppermine for testing before i believe it this time.
>
>Can you explain this more? I always assumed e.g. the Celeron's cache is twice as
>fast as the Pentium II's.
>
>-Tom


It is more.  The celeron cache is faster than the xeon cache if you look at
the time it takes to process a cache miss.  The cache cycle time is the same
once data starts streaming from the L2 to L1 cache, but the xeon has something
over 12 clocks for the miss penalty, while the celery is 6 IIRC.  New PIII's
(copper) apparently stole the celery's L2 cache but made it bigger).



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.