[mpich-discuss] Is it okay to remove F77 support?

[Jeff Squyres (jsquyres)]

On Jan 27, 2021, at 9:45 AM, Raffenetti, Kenneth J. via discuss <discuss at mpich.org> wrote:
> 
> The MPICH team is asking if it would be OK to drop support for building MPICH with an F77 only compiler.

MPI -- MPICH, Open MPI, etc. -- has never been able to be built with a F77-only compiler because the MPI standard has never adhered to the Fortran 77 standard.

Additionally, MPI-3.1 17.1.1 quite clearly states that a conformant MPI-3.1 implementation must support either the mpi_f08 module (which implies support for much of the Fortran'08 standard) and/or mpif.h + the mpi module (which requires elements of the Fortran 90 standard).  This requirement has been there since MPI-3.0, published in 2012.

Hence, it has not been possible to build any conformant MPI implementation with a "F77-only" compiler for nearly a decade.

And let's not forget that Fortran 77 limits symbols -- variables, subroutines, and functions -- to 6 characters.  Hence, it has *never* been possible to build any MPI implementation or application with a Fortran 77-only compiler.

Håkon's answer earlier in the thread was also quite correct: "fixed form Fortran" is not the same thing as "Fortran 77".  Please stop conflating the two.

> MPICH will still support all standard Fortran language bindings: mpif.h, use mpi, and use mpi_f08. Does the code you speak of not build with modern Fortran compilers? If so, that is the kind of feedback we are looking for.

That is a good question, much better formed than the original.  It's a question that the entire MPI Forum would benefit hearing answers to!

-- 
Jeff Squyres
jsquyres at cisco.com