Newsgroups: comp.parallel.mpi
From: salo@mrjones.engr.sgi.com (Eric Salo)
Subject: Re: [Q]MPICH on SGI Power Challenge
Organization: Silicon Graphics, Inc.  Mountain View, CA
Date: 29 Oct 1996 21:16:54 GMT
Message-ID: <555s86$a32@murrow.corp.sgi.com>

> Question 1: should I expect to see a significant difference in the
> performance of an application compiled with the ANL MPICH and then
> compiled using the vendor supplied MPI on a Power Challenge? (I have not
> seen a significant difference. Am I missing compiler flags?)

Well, that obviously depends on the application. But you should definitely
see an improvement in the performance of MPI itself. I don't have the raw
MPICH numbers, but on a 90MHz R8K Power Challenge our latency is ~18 usecs
and our peak bandwidth is ~64 MB/sec.

> I used the ANL/MPICH because the target machine has not got vendor MPI -
> that machine is 2xR10000@194MHz IRIX64

Our implementation is both free of charge and directly available from the
SGI web site. There should be no problems with the above configuration.

	http://www.sgi.com/Products/Evaluation/evaluation.html

Follow the "IRIX 6.2" link and select the "Array 2.0" product.

Eric Salo         Silicon Graphics Inc.             "Do you know what the
(415)933-2998     2011 N. Shoreline Blvd, 8U-802     last Xon said, just
salo@sgi.com      Mountain View, CA   94043-1389     before he died?"

