Newsgroups: comp.parallel.mpi
From: bruce@liverpool.ac.uk (Bruce Stephens)
Subject: Re: Uniqueness of Messages in MPI
Organization: Centre for Mathematical Software Research, Univ. Liverpool
Date: Fri, 26 Aug 1994 13:39:32 GMT
Message-ID: <BRUCE.94Aug26143933@cmsr3.liverpool.ac.uk>

>>>>> On Thu, 25 Aug 1994 02:32:53 GMT, spimchon@cs.anu.edu.au (M. Spinks) said:

> Does anyone know if, semantically, messages are considered unique in
> MPI?  That is, given two messages with the same sender and
> recipient, same tag and communicator, and same contents, if the
> messages are distinct?

They must be distinct, because the order in which they will be
received is defined.  That is, if I send two identical messages, and
have two appropriate receives, then both receives will succeed (and,
of course, the first of the identical messages will be received by the
first receive, and the second message by the second receive, although
I couldn't tell that except by using extra information).



-- 
Bruce                    Institute of Advanced Scientific Computation
bruce@liverpool.ac.uk    University of Liverpool

