On Sat, 29 May 1999 12:55:26 -0600, "Blosser, Jeremy"
<[EMAIL PROTECTED]> wrote:
>Hrm... I just looked at the rate my Quad Xeons are doing (2-Compaq 6500s and
>a 7000), and they are getting .31sec/iteration. Yet my peiced together
>Dual-412 (OC'd 300s) gets in the .21-.22 range... (same size FFT) Any
>thoughts as to why the Xeons are doing so poorly? Their affinity is for
>seperate processors (so they aren't jumping around). And I don't think that
>cache differences are a problem (cuz they got 512k and 1024k respectively).
>
>Any thoughts?
I sort of remember something being said about Prime95's timing being
incorrect for Xenon CPUs. It was either that the speed
auto-detection was wrong (check that Options->CPU matches what you
think it should be, or that the number of clock ticks wasn't counted
correctly. If both of these seem OK, try timing a few hundred
iterations by hand and see what you get.
>
>The only thing that came to mind was ECC cache or something?
--
Kevin Jaget (or an FDA approved generic equivalent)
kcjaget at mindspring.com
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
