[Mpi-forum] Timeline for MPI 3.1

Jeff Hammond jeff.science at gmail.com
Wed Sep 3 10:11:16 CDT 2014

I support getting it right late as opposed to the timely alternative.
There is no point in having a 3.1 release that doesn't cover all the
critical errata.


On Tue, Sep 2, 2014 at 5:03 PM, Schulz Martin <schulzm at llnl.gov> wrote:
> Hi all,
> After discussing with the steering committee, I wanted to bring up the
> following:
> In one of the previous meetings we decided on the timeline for MPI 3.1: the
> intent is to ratify the complete MPI 3.1 standard by the December meeting,
> which makes the upcoming September meeting the last meeting to allow for
> errata votes (actual additions had to be read in March). The main goal of
> the MPI 3.1 document is/was to create a “clean” standard that integrates
> outstanding errata items and the small additions we worked on. While the
> latter seem to be more or less done, we have multiple outstanding errata
> items (in particular in RMA, but also a few in tools, potentially also in
> other areas) that won’t get done by the September meeting (the respective WG
> chairs felt their wasn’t enough time or participation in the Japan meeting
> for their WG to allow for enough discussion).
> If we stick with the current timeline, we’ll have an MPI 3.1 and yet again a
> good number of errata items next to it, which is counter the intention for
> publishing an intermediate version before MPI 4.0. This means we are either
> back in the same situation as now with MPI 3.0 for quite a while longer or
> we have to go for an MPI 3.2. Neither option seems good.
> Therefore, I would like to propose to not rush it and to slip the schedule
> by one (or if needed by two) meetings to make sure we conclude the
> outstanding errata items. Otherwise, we’ll end up with a standard that we
> know is broken by the time we vote for it (we would have the final document
> vote and the MPI 3.1 errata votes in the same meeting).
> Opinions?
> Thanks,
> Martin
> ________________________________________________________________________
> Martin Schulz, schulzm at llnl.gov, http://scalability.llnl.gov/
> CASC @ Lawrence Livermore National Laboratory, Livermore, USA
> _______________________________________________
> mpi-forum mailing list
> mpi-forum at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

Jeff Hammond
jeff.science at gmail.com

More information about the mpi-forum mailing list