What do you guys thing about the P4

Greg Lindahl lindahl at conservativecomputer.com
Wed Apr 11 12:08:51 PDT 2001


On Wed, Apr 04, 2001 at 01:01:21PM -0600, Scott Delinger wrote:

> I found the P4 in testing computational chemistry codes to be such a poor
> performer in comparison to Athlons, P3s, and even Durons for some code
> that the expense was unreasonable for almost all uses. This is the proof
> of the rule "Your own code is your only meaningful benchmark.".

Scott,

Did you use the Intel compiler? If not, then there's no surprise that
the P4 did poorly. The SPECfp scores with an old compiler are probably
similarly terrible.

I wish people would mention what compiler they used when they make
comments like yours, because in this case the details really matter.

-- g





More information about the Beowulf mailing list