This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: -ffast-math and gfortran
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Toon Moene <toon at moene dot indiv dot nluug dot nl>
- Cc: fortran at gcc dot gnu dot org, fxcoudert at gmail dot com, howarth at bromo dot msbb dot uc dot edu
- Date: Sun, 4 Jun 2006 13:21:46 -0700
- Subject: Re: -ffast-math and gfortran
- References: <18686767.1149411692110.JavaMail.root@dtm1eusosrv72.dtm.ops.eu.uu.net> <E1Fmpnr-0005sk-Ec@laptop.moene.indiv.nluug.nl>
On Sun, Jun 04, 2006 at 12:25:51PM +0200, Toon Moene wrote:
> > For an example of -ffast-math's potential problems, try
> > compiling netlib's lapack and run the tests. This option
> > should be renamed to -frisky-math.
>
> Yes, -ffast-math is definitely *not* the option you want to use
> to compile software off-the-net that you don't know the algorithmic
> basis of.
If an option is available, people will use it regardless of
their level of expertise. Look at richi's Polyhedron benchmark
page as well as others who post benchmark scores. -ffast-math
is used and people aren't checking whether the benchmarks actually
pass.
This is going to become a Support issue as gfortran gains users.
I hope your onboard to answer all the questions.
> However, if you compile code you yourself wrote and know the limitations
> of, it can be useful.
It isn't my code or even the code written by Jack Dongarra that's
the problem. Where is the documentation of what -ffast-math does?
Oh yeah, you found it too complicated to document.
--
Steve