On Thu, Feb 18, 2016 at 11:06 AM, Bill Long <[email protected]> wrote:

On Feb 18, 2016, at 11:55 AM, Jeff Hammond <[email protected]> wrote:

...

From MPI 3.1, page 674:

"If an accompanying C++ compiler is missing, then the MPI datatypes in this table are not defined."

Please tell me why we want to do something different for Fortran.

I agree that this is a good point. Although an increasingly strange one in the C++ case since the C and C++ compilers are often the identical same compiler these days. The choice of cc or CC just sets a different mode internally.

<tangent> We still need to be careful, because C99 and C11 are not strict subsets of any ISO C++. And at least one MPI Forum member (Microsoft) has the audacity to not support ISO C at all, except insofar as it is a subset of C++, which is why lots of folks (e.g. MPICH) are limited to C89, which is - as far as I know - a proper subset of C++. While we do not have C++ bindings and our C bindings are compliant with C89, there is some potential for this to change in the future, particularly if Squyres and Hammond follow-through on their intent to solve the large-count problem with C11/C++98 polymorphism. </tangent>

Cheers, Bill

...

Bill Long [email protected] Fortran Technical Support & voice: 651-605-9024 Bioinformatics Software Development fax: 651-605-9142 Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101

_______________________________________________ mpiwg-fortran mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran

-- Jeff Hammond [email protected] http://jeffhammond.github.io/