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

Re: gfortran transition release goal proposal


I suppose everyone is aware that -ffloat-store is a real performance
loser, especially on slower machines.  Would it be preferable to fix
the testers to not be fooled by extra precision?

Take care,

Riku Voipio <riku.voipio@iki.fi> writes:

> Camm,
> your email is bounding for me!
> On Mon, Jul 16, 2007 at 06:14:46PM -0400, Camm Maguire wrote:
> > Sure.  I suppose we still have testsuite failures in
> > blas/lapack/atlas, yes?  I'll try to take a look.  Is refblas3 as good
> > a place to start as any?
> you need to you -ffloat-store gfortran option. see man gcc-4.2 for
> exoplanation. With attached patch, I got refblas3 testsuite succeed
> on i386 and it was told to me that it also works for amd64.
> -- 
> "rm -rf" only sounds scary if you don't have backups

Camm Maguire			     			camm@enhanced.com
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

Reply to: