[Mpi-21] Implementation of MPI-2.1

Supalov, Alexander alexander.supalov at [hidden]
Thu Jun 26 09:09:22 CDT 2008

Hi everybody,

I'm going to miss this meeting, so let me use this reply as a chance to
express my opinion ahead of the vote:

- We need to produce official versions of the MPI-1.3 and MPI-2.1
documents with change mark-up, either in change bars or in color coding.
At the moment, as far as I understand, only the official documents are
going to be published, and they won't have clear indication, apart from
the MPI-2.1 change log, of what was actually changed in the text. Many
standard bodies I know routinely publish marked-up versions of their new
documents to facilitate smooth transition. If we don't follow this
practice, we risk implementors overlooking some small but important
changes, and then claiming MPI-1.3 and MPI-2.1 compliance by mistake.

- The desire to have only Forum members edit the standards is a good
one, but it may not be practical. In the particular case of the language
binding Annex, we'd not be able to complete the massive chances on time
without delegating this to someone else. This would make the standard
process slip, which was considered a bigger risk at the time. What is
necessary here is rather more careful supervision and review of the
results, which I think we should take into account going forward.

- If there are any material changes needed to Intel MPI implementation
to make it MPI-1.3 and MPI-2.1 compliant, we'll target 2009. If we can
claim MPI-1.3 and MPI-2.1 compliance in the current code without
noticeable changes, either to the code or to the current draft during
the pending voting, we may decide to go for this year's fall release.

Best regards.


-----Original Message-----
From: mpi-21-bounces_at_[hidden]
[mailto:mpi-21-bounces_at_[hidden]] On Behalf Of Rolf
Sent: Thursday, June 26, 2008 3:29 PM
To: MPI 2.1 Mailing List
Subject: [Mpi-21] Implementation of MPI-2.1

Dear MPI-implementor,

if you are not yet on the list at slide 8 in

or if you are on the list and you have not yet given an
estimate when your MPI-2.1 implemenatation will be available,
then please give us this estimate at June meeting in San Francisco
(or with a short reply to this list before the meeting).

Thanks and best regards

Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden]
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)
mpi-21 mailing list
Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen Germany
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
Registergericht: Muenchen HRB 47456 Ust.-IdNr.
VAT Registration No.: DE129385895
Citibank Frankfurt (BLZ 502 109 00) 600119052

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.

More information about the Mpi-21 mailing list