help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: speed of octave


From: Michael Goffioul
Subject: Re: speed of octave
Date: Fri, 14 Sep 2007 22:30:29 +0200

On 9/14/07, Tatsuro MATSUOKA <address@hidden> wrote:
> Hi Michael.
>
> I have a question about ATLAS.  There shoud be differences in the L2 cash 
> size between several pentium
> 4 core cpu's.  How much size did you set for your ALTAS in your binary you 
> distribute?  I am using two
> computers: the cpu of them is HT pentium 3.4 GHz and the other is celeron M 
> 1.2GHz.  If the L2 cash
> size is not the same as you build ALTAS, will it be better to select ATLAS 
> when the cpu is pentium4
> core? In the installing your binary, the user will select either the pentium4 
> with sse2 or the
> generic.  Until now, I have not been care about the ATLAS.
>
> Of course the features which is not related lapack, it is worth to select 
> sse2.  However is the
> selection of sse2 better even if the L2 cash size of the computer is 
> different from what you designed
> for?

Theoretically, it's suboptimal. In practice, I ran a benchmark script in octave
with 2 versions of ATLAS: one compiled for P4 (512KB L2 cache) and one
compiled for Centrino (2MB L2 cache). The test was run on the Centrino
CPU. Both ATLAS versions brought big speed-up compared to the generic
Fortran libraries, but the gain of using a Centrino-compiled ATLAS instead
of a P4-compiled ATLAS was not that high (IIRC it was < 10%). I didn't test
the other possibility: running the Centrino-based ATLAS on the P4 machine.
The binary installer only provides the P4-compiled ATLAS, which probably
provides good results on most CPU out there (P4, AMD, Centrino...)

Michael.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]