Re: Floating-point performance

Nicolas Neuss <[email protected]> Wed, 11 May 2005 21:34:55 +0200
Newsgroups gmane.lisp.allegro
Organization IWR
Message-ID <[email protected]>
"Ernst van Waning" <[email protected]> writes:

> Hello Nicolas,

Hello Ernst,

> ...
> ; real time  2,113 msec
> ...
> ; real time  2,794 msec

Your results look at least better than mine (I have a Pentium M with about
1.5? GHz, I must admit that I don't know the exact value).  In the moment,
I am in contact with Duane Rettig, who is examining the problem.  As much
as I understood the culprit might be some floating point instruction which
is extraordinarily slow on Pentium 4 (and probably also Centrino, because I
have one).  I guess that Duane will report on this when everything is
clear.

Yours, Nicolas.