Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: Re: [OMPI devel] Failure to make progress
From: Ken Olum (kdo_at_[hidden])
Date: 2009-02-25 15:21:19

   From: George Bosilca <bosilca_at_[hidden]>
   Date: Mon, 23 Feb 2009 16:54:47 -0500

   Your interpretation of the MPI standard is way too optimistic.

The text I quoted seemed completely clear to me. However, it was
pointed out to me that MPI 2.0, 6.7.2 (in a section about one-sided
communication that is not otherwise relevant) says

   A similar issue is whether such progress must occur while a process is
   busy computing, or blocked in a non- MPI call.... Then, according to
   one interpretation of the MPI standard, process 0 must return from the
   complete call after a bounded delay, even if process 1 does not reach
   any MPI call in this period of time. According to another
   interpretation, the complete call may block until process 1 reaches
   the wait call, or reaches another MPI call.... The MPI forum decided
   not to decide which interpretation of the standard is the correct one,
   since the issue is very contentious, and a decision would have much
   impact on implementors but less impact on users.

So I guess one could say that the 2.0 standard is now explicitly vague
on this subject.