Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2007-04-02 08:01:34

On Apr 1, 2007, at 12:50 PM, de Almeida, Valmor F. wrote:

> I would be interested in hearing folk's experiences with gfortran and
> ompi-1.2. Is gfortran good enough for prime time?

It is my understanding that gfortran is the GNU path forward -- they
are not spending any time on g77. I've been using gfortran (for
testing only; I'm not an apps guy) for a few years and have been
relatively pleased with it.

> I have built ompi-1.2 with gfortran-4.1.1 but no luck in testing it
> because my application of interest (legacy fortran code) will not
> compile with gfortran.

Bummer. FWIW, we have some internal testing OMPI Fortran codes that
are difficult to get to compile uniformly across all Fortran
compilers (the most difficult point is usually fortran pointers).
This could well be because of my lack of knowledge of Fortran,
however -- I am nowhere close to being a Fortran expert.

> In addition I would be interested in hearing about the use of
> different
> fortran compilers. For instance, how about building ompi with gfortran
> but compiling mpi fortran codes with the intel compiler?

As Rainer said, we don't recommend this. It may be possible, but I
would not suggest this course of action unless you have a strong need
to do so. Additionally, we tend to only test the single-compiler-
suite scenarios (e.g., gcc/g++/gfortran, pgcc/pgCC/pgf77/pgf90, icc/
icpc/ifort, etc.), so I don't have much data about mixing compilers
for you; sorry.

Jeff Squyres
Cisco Systems