[MPIWG Fortran] MPI_SIZEOF: redux

[Bill Long]

On Jun 9, 2014, at 7:24 PM, Jeff Hammond <jeff.science at gmail.com> wrote:

> On Mon, Jun 9, 2014 at 5:21 PM, Jeff Squyres (jsquyres)
> <jsquyres at cisco.com> wrote:
>> Yep -- looks like it was quoted as being part of F90 incorrectly.
>> 
>> So -- ya: storage_size is F2008 and later.  But I still think it would be ok to deprecate MPI_SIZEOF, since the language now supports this.  And as Bill Long pointed out, having MPI deprecate MPI_SIZEOF might actually pressure compiler vendors to implement storage_size().
> 
> Which Fortran compiler vendors that do not currently implement this
> intrinsic are going to be swayed by MPI-4 deprecations to implement
> it?

Vendors put off implementing language features if no customer has expressed an interest, with the resistance higher if the feature is hard to implement. Implementing STORAGE_SIZE in a compiler is trivial, so any expression of interest is likely move it to the top of the list.  I’d note that the primary motivation for vendors to implement the TS 29113 features is the fact that MPI is using them.  Even though this is comparatively a lot of work to implement, compiler vendors (at least ones catering to parallel systems) do pay attention to MPI.  

Cheers,
Bill


Bill Long                                                                       longb at cray.com
Fortran Technical Suport  &                                  voice:  651-605-9024
Bioinformatics Software Development                     fax:  651-605-9142
Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101