Newsgroups: comp.parallel.mpi
From: salo@mrjones.engr.sgi.com (Eric Salo)
Subject: Re: MPI Non blocking Sends, send buffer not readable or unclear standard?
Organization: Silicon Graphics, Inc.  Mountain View, CA
Date: 29 May 1996 17:25:35 GMT
Message-ID: <4oi1af$i3l@murrow.corp.sgi.com>

> > I'll try my best not to use your implementation.

Well, that's unfortunate because we have the fastest host-to-host
MPI bandwidth in the world. This restriction on send buffers was
put in place to allow for hardware optimizations such as the kind
we are using by people with significant knowledge about message
passing. (No, I wasn't one of them.) I'll defer to one of the
original MPI-1 members for more background on this decision.

Are you saying that we are incorrect to follow the standard?

> BTW, I think it's a telling point that you use the term "culprit."

I used that term because it is a very recent addition to our
implementation and so it slipped my mind. Our host-to-host bandwidth
is the single best feature of our implementation.

Eric Salo         Silicon Graphics Inc.             "Do you know what the
(415)933-2998     2011 N. Shoreline Blvd, 7L-802     last Xon said, just
salo@sgi.com      Mountain View, CA   94043-1389     before he died?"

