Mathew,
>I want to use ifort on a non windows machine (x86_64) but when I look at
>the errata the instructions seem to be only for windows. Are there
>instructions for my architecture?
>
>I read somewhere on the site that gfortran should give optimal
>performance but I am forced to use ifort. How much will I be losing in
>speed?
Scope:
http://math-atlas.sourceforge.net/atlas_install/atlas_install.html#SECTION00042000000000000000Just throw the flags you usually use in the code that will link to this
library.
In general, changing the fortran compiler should have minimal or no impact
on performance. If you could see a difference it would be in lapack
performance, for routines ATLAS doesn't natively provide, and that don't
get all their performance from the BLAS: I'm not sure if such routines exist :)
In such cases, I would guess ifort to be faster than gfortran, albeit with
possibly unbounded error.
Cheers,
Clint
**************************************************************************
** R. Clint Whaley, PhD ** Assist Prof, UTSA ** www.cs.utsa.edu/~whaley **
**************************************************************************
------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword_______________________________________________
Math-atlas-devel mailing list
Math-atlas-devel@...
https://lists.sourceforge.net/lists/listinfo/math-atlas-devel
