This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: GCC 3.0 Release Criteria
- To: Mumit Khan <khan at NanoTech dot Wisc dot EDU>
- Subject: Re: GCC 3.0 Release Criteria
- From: Toon Moene <toon at moene dot indiv dot nluug dot nl>
- Date: Thu, 27 Apr 2000 01:03:20 +0200
- CC: Jonathan Thornburg <jthorn at galileo dot thp dot univie dot ac dot at>,GCC Mailing List <gcc at gcc dot gnu dot org>
- Organization: Moene Computational Physics, Maartensdijk, The Netherlands
- References: <200004262205.RAA14874@hp2.xraylith.wisc.edu>
Mumit Khan wrote:
> Toon Moene <toon@moene.indiv.nluug.nl> writes:
> > Jonathan Thornburg wrote:
... the Kagstrom, Ling, and Van Loan GEMM-based implementation ...
> > Hmmm, I never tried anything else than the default BLAS - eeks, more
> > options to test ...
>
> The default BLAS tends to be rather bad on most machines, with just a
> few exceptions.
Well, the premier problem is that I, in my *real* job, do not need BLAS
[n]or LAPACK (those who think that all computational physics problems
can be reduced to an exercise in solving linear systems - BZZZZZZZZT,
thank you for playing). No, I do not think that either Jonathan or
Mumit believe this :-)
I think LAPACK is a useful package to test the compiler (because of its
test and timing programs). However, I've no opinions on its intrinsic
"optimality" :-)
> Toon, how do you feel about using Atlas http://www.netlib.org/atlas/
> instead of LAPACK-3.0 (or in addition to LAPACK 2.0)?
I've downloaded it in the past, because I found the idea promising, but
never got around trying it, sorry ...
--
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://xena.eas.asu.edu/~andy (under construction)