Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: 88.86 ns on a Dual Xeon 2.8/533 ScienceMark 2.0 beta

Author: Matt Taylor

Date: 15:20:22 03/21/03

Go up one level in this thread


On March 20, 2003 at 18:39:21, Matt Taylor wrote:

<snip>
>I'll try disabling the ECC on my ram. There isn't anything I can do about it
>being registered, and I can't tweak any of the settings to get a fair registered
>vs. non-registered on AMD 760MPX benchmark.

LMBench reports 201.6 with ECC disabled and 201.7 with ECC enabled. I don't
think ECC makes a difference, or at least in my case. (It is possible that
disabling ECC really doesn't.)

I don't have anything to compare my ram with since I can't control timings on
this board. LMBench & ScienceMark usually coincide very nicely. My old figures
from ScienceMark were ~133 ns. I retested and now get 198-199 ns. I ran
ScienceMark on my work machine which also uses an AMD 760MPX chipset. The
latency I got from ScienceMark was about equal (197-198 ns). So it seems the 133
ns figure was wrong. I get 200 ns on the 760 MPX.

Of course, both machines are using Registered/ECC ram. The figure could be lower
if I wanted to minimize latency.

-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.