This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Two glitches while testing gcc version 4.3.0 20070216
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Dominique Dhumieres <dominiq at lps dot ens dot fr>
- Cc: fortran at gcc dot gnu dot org
- Date: Sat, 17 Feb 2007 10:23:50 -0800
- Subject: Re: Two glitches while testing gcc version 4.3.0 20070216
- References: <20070217180630.EA67D5BB46@mailhost.lps.ens.fr>
On Sat, Feb 17, 2007 at 07:06:30PM +0100, Dominique Dhumieres wrote:
> Thanks for the answers.
>
> > It's simply a timing problem. secnds.exe is trying to make a timing
> > measurement and the tolerance is IMHO too strict.
>
> >From the list I got the impression it has been fixed. Apparently not!
>
> > You need to edit LAPACK's second.f code. Remove "external etime"
> > and replace "call etime" with "call cpu_time".
>
> I'll try it later, but is not ETIME among the intrinsic extensions?
>
Sigh. Yes, there is an ETIME intrinsic procedure. If you
have an "EXTERNAL ETIME" statement in your program, you
are telling gfortran that you are supplying an *external*
user-defined etime routine. gfortran will not use its
intrinsic. This is easy to see in that ld failed looking
for _etime_. All gfortran intrinsics are name mangled to
include a _gfortran_ prefix, ie., _gfortran_etime_.
--
Steve