Newsgroups: comp.parallel.mpi
From: llewins@msmail4.hac.com (Lloyd J Lewins)
Subject: Re: MPI collective comm question
Organization: Hughes Aerospace Electronics Co.
Date: Thu, 11 May 1995 07:51:51 -0800
Message-ID: <llewins-1105950751510001@x-147-16-95-58.es.hac.com>

In article <3oe4io$261@fido.asd.sgi.com>, salo@mrjones.engr.sgi.com (Eric
Salo) wrote:

> For collective operations on a given communicator, all members of that
> communicator must participate. If any are left out, the operation blocks.
                                                                     ^^^^

Not so, read the standard, section 4.1 page 92 lines 10-16.

   "Thus, a collective communication call may, or may not, have the effect of
    synchronizing all calling processes." 

Also see the examples in section 4.12. In particular, note that the root
node in a broadcast operation may (or may not) proceed before any or all
of the leaf nodes participate in that broadcast.

--------------------------------------------------------------------------
Lloyd J Lewins                                  Mail Stop: RE/R1/B507
Hughes Aerospace and Electronics Co.            P.O. Box 92426
                                                Los Angeles, CA 90009-2426
Email: llewins@msmail4.hac.com                  USA
Tel: 1 (310) 334-1145
Any opinions are not neccessarily mine, let alone my employers!!

