[MPI3 Fortran] MPI draft 2011-04-21

Rolf Rabenseifner rabenseifner at hlrs.de
Thu Apr 21 17:04:40 CDT 2011


Dear all,

at MPI ticket 229
https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/229

I attached the draft for the MPI Form meeting May 9-11, 2011.

For online reading, I recommend 
https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/ticket/229/mpi-report-F2008-2011-04-21-changeonlyplustickets.pdf

(For the MPI Forum all information was in a separate mail.)

For the Fortran J3/WG5:

Most important pages are 13, 16, 545-558, 561-565

For me, it looks like that most problems are solved,
because in MPI-draft-2011-04-21, Section 16.2.16, we make additional 
requirements for the combination of MPI & compiler.

As you can see, the ASYNCHRONOUS in combination with the 
temporary data movements in MPI-draft-2011-04-21, Section 16.2.10
is kept unsolved if parts of an array are used for pending communication
and other parts for computation.
This may be solved in the future, not now.
Therefore I withdraw our ASYNCHRONOUS-comment on the ballot 1846.

More important for as is the handling of sequence derived types 
within BIND(C) derived types and BIND(C) routines in the dummy 
argument list.

It would be helpful, if the requirements on page 564 line 19-31 
and page 564 line 44 - page 565 line 3 are guaranteed by the TR 29113.

Best regards
Rolf

-- 
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: Allmandring 30)



More information about the mpiwg-fortran mailing list