Jeff Squyres wrote:

Hmm. Ok, that's the opposite advice we got for MPI-3.0 (which was now admittedly several years ago).

Soo... which way should it go these days?

- use native Fortran datatypes - use C datatypes (via BIND(C))

I think that MPI Fortran Interface should support native Fortran datatypes in pair types ! (*) We (NEC) don't know any relevant Fortran application which doesn't use native Fortran datatypes. If you switch for C Datatypes in MINLOC/MAXLOC, all application programmers must solve the portability problems (possibly loosing accuracy in reduce operations). (*) Nearly all Fortran compilers support compile time options such as "-r8" or "-size real64" which would cause problems with the C datatypes. And this options are used by application programmers ! (*) C datatypes would be inconsistent with functions such as MPI_Type_create_f90_real. (*) Supporting Fortran subarrays without supporting native Fortran datatypes doesn't make sense for me. Hubert