[Mpi-forum] MPI_Abort - meaning

Richard Treumann treumann at us.ibm.com
Mon Apr 5 17:20:18 CDT 2010


Maybe - the standard does not say anything specific about the meaning of
MPI_Abort with a value of zero.

The "best effort" part and the fact that the value passed to MPI_Abort is
not required as a job RC both make MPI_Abort unsatisfactory as currently
described. The name "Abort" also sounds like an undesired outcome.  You do
not abort something when you have gotten what you wanted.

If the Forum chose to be explicit about the meaning of MPI_Abort with zero,
it would answer the need but it would still feel sort of ugly to me.

In part, I bring this up because now that it has been pointed out to me, I
am a bit surprised that I have never come across the issue before. That
made me wonder if wanting to end a parallel application this way is rare or
if people have just been calling ABORT and being content it seemed to work
OK.


Dick Treumann  -  MPI Team
IBM Systems & Technology Group
Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601
Tele (845) 433-7846         Fax (845) 433-8363



                                                                                                                          
  From:       Fab Tillier <ftillier at microsoft.com>                                                                        
                                                                                                                          
  To:         Main MPI Forum mailing list <mpi-forum at lists.mpi-forum.org>                                                 
                                                                                                                          
  Date:       04/05/2010 03:48 PM                                                                                         
                                                                                                                          
  Subject:    Re: [Mpi-forum] MPI_Abort - meaning                                                                         
                                                                                                                          
  Sent by:    mpi-forum-bounces at lists.mpi-forum.org                                                                       
                                                                                                                          





Richard Treumann wrote on Mon, 5 Apr 2010 at 10:45:00

> A call to MPI_Quit would let the application give a return code of 0 so
> a script or wrapper that checks the mpiexec return code would regard the
> application as a success.

How is this different than calling MPI_Abort with an error code of 0?

Is it just that MPI_Quit would do more than 'best effort'?

-Fab

_______________________________________________
mpi-forum mailing list
mpi-forum at lists.mpi-forum.org
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpi-forum/attachments/20100405/04e46f87/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://lists.mpi-forum.org/pipermail/mpi-forum/attachments/20100405/04e46f87/attachment-0002.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ecblank.gif
Type: image/gif
Size: 45 bytes
Desc: not available
URL: <http://lists.mpi-forum.org/pipermail/mpi-forum/attachments/20100405/04e46f87/attachment-0003.gif>


More information about the mpi-forum mailing list