[MPI3 Fortran] What to do with mpif.h in MPI-3?

[Jeff Squyres]

On Jan 25, 2010, at 11:19 PM, Torsten Hoefler wrote:

> > I think the key issue here is that there is no compiler in use anymore
> > that *only* supports F77 syntax (even those that have "77" in their
> > executable name).  There hasn't been a only-supports-fortran-77-syntax
> > compiler for many, many years (since the early-/mid-90s? That's a
> > guess).
> That doesn't matter, if it's not in the F77 standard then we can't have
> it in the F77 bindings.

There are no F77 bindings.  There are only implicit Fortran bindings in mpif.h.

> > So using (KIND=foo) syntax, even though that didn't debut until F90 or
> > F95 (or whatever; I'm no Fortran expert) is perfectly acceptable here.
> Not without renaming the bindings (I suppose that would require some
> wording changes that I'm not totally opposed to).

We already have some KIND types in the MPI Fortran bindings (MPI_AINT_KIND is my favorite :-) ).

-- 
Jeff Squyres
jsquyres at cisco.com