[MPIWG Fortran] MPI_SIZEOF: redux

[Jeff Squyres (jsquyres)]

On Jun 5, 2014, at 11:35 PM, Malcolm Cohen <malcolm at nag-j.co.jp> wrote:

> Anyway, since the so-called F77 binding requires the compiler to have Fortran 90 features, whether Fortran 77 had STORAGE_SIZE is hardly relevant to the question of what to do with MPI_SIZEOF.  What being relevant would seem to be whether there is an alternative that is universally available, whether that alternative is actually better to use, and whether it is time to start steering users towards it, by means of deprecating MPI_SIZEOF.

Good info; thanks.

> But since STORAGE_SIZE is far from universally available, this falls at the very first hurdle.  (And alternative ways of computing sizeof in e.g. pure Fortran 90 are, although easy to write, not particularly convenient to use, so those fall at the second hurdle.)

Fair enough; but just to clarify -- deprecating is not removing.  Deprecating in MPI is a signal "you should try to not use this any more; it may disappear some day."

Would it still be ok to deprecate MPI_SIZEOF in MPI-4 (which is at least ~2 years away)?  We don't have to have any plan for removing it.

I only ask because it seems silly to have a Fortran-only MPI function that was specifically added to compensate for the language's lack of functionality -- but the Fortran has now provided this functionality for quite some time (regardless of the implementation status of various modern Fortran compilers).

-- 
Jeff Squyres
jsquyres at cisco.com
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/