[Beowulf] MPI ABI

Toon Knapen toon.knapen at fft.be
Mon Oct 10 06:54:38 PDT 2005

Patrick Geoffray wrote:
> The Fortran interface is actually worse than the C interface. Instead of
> using pointers to opaque structures, the Fortran interface may use
> integers as indexes into array of structures, into array of pointers, as
> pointers casted to integers, etc.

I can imagine that implementors of MPI libraries do not like the Fortran
interface very much but the standard requires that it's there. As a user
however I do not care that the MPI implementation of the fortran
interface is more complicated. That's stuff to worry about for the MPI
implementors only.

> Even if you implement all of the wrappers in a MorphMPI approach,
> nothing guarantee you that you can take an Fortran MorphMPI compliant
> application and link it with a MorhMPI compliant MPI library compiled
> with a different Fortran compiler. Fortran is evil.

As long as the calling convention (who and how the arguments are put on
the stack) is the same the library can be linked with programs compiled
with a different compilers AFAICT.

I'm into C++ and do not like Fortran neither very much but I'm just
being pragmatic here.

More information about the Beowulf mailing list