[Mpi3-tools] Ticket #354

Rolf Rabenseifner rabenseifner at hlrs.de
Wed Jan 9 03:36:44 CST 2013

If you want to do #354, i.e., to remove the MPI_T_... constants
from the Fortran interfaces (mpif.h, mpi module, mpi_f08 module),
then you should do it now and rename your ticket title to

   MPI-3.0 Errata: MPI_T constants erroneously listed for Fortran

You should put it on the agenda and vote on it in the next meeting.

Keep in mind:
 - If you want to have these constants also for Fortran-written wrappers
   available, you can re-introduce them in MPI-next,
   but then also with the appropriate interface for the MPI_T_... handles.
 - If you do not have both, then it is only very partially useful
   in Fortran-written profiling layers.
 - Therefore, i expect that it is only a copy&paste bug that
   MPI-3.0 defines the constants also for Fortran.
 - You should do it as fast as possible as MPI-3.0 errata,
   because a removal in MPI-next would be a non-backward-compatible

Best regards

Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner at hlrs.de
High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530
University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832
Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner
Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307)

More information about the mpiwg-tools mailing list