Newsgroups: comp.parallel.mpi
From: eddemain@neumann.uwaterloo.ca (Erik Demaine)
Subject: Re: Performance problems using mpi on the SGI Power Challenge
Organization: University of Waterloo
Date: Fri, 27 Sep 1996 20:56:10 GMT
Message-ID: <DyEstM.6Kw@undergrad.math.uwaterloo.ca>

Eric Salo (salo@mrjones.engr.sgi.com) wrote:
: As soon as you have more MPI processes than CPUs on which to run them,
: performance drops in a very big way. This was a deliberate trade-off
: that we made early on because we believed that the common case would be
: to not oversubscribe the CPUs in this way.

This seems to be a shame.  Granted, you get "better" parallel performance
this way.  But it really kills any attempt at concurrent programming using
MPI as a basis.

Oh well, I'm sure it will be fixed when there is sufficient need.

Erik
-- 
Erik Demaine                 ()  e-mail: eddemain@neumann.uwaterloo.ca
Dept. of Computer Science    \/  URL: http://barrow.uwaterloo.ca/~eddemain/
University of Waterloo       /\  PGP key: finger me.  "Flying is the art of
Waterloo, ON Canada N2L 3G1  ()  falling and not hitting the ground" -Adams

