Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!tut.cis.ohio-state.edu!ucbvax!dlgm.daresbury.ac.uk!RJA
From: RJA@dlgm.daresbury.ac.uk ("R.J." ALLAN)
Newsgroups: comp.sys.transputer
Subject: (none)
Message-ID: <9103131436.AA27071@prg.ox.ac.uk>
Date: 13 Mar 91 14:36:41 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 22

Science on Massively Parallel Machines

In reply to Bill Moses' question about who wants massively parallel machines.

I think if there were no headaches it would be obvious that everyone does.
But of course the real world is different, and you first need to define
what you mean by massively parallel.
     Yes we have been experimenting and playing with small systems for a
few years now, well actually developing virtual machine interfaces which
are portable and will allow numerical algorithms to be developed quickly
with an underlying message-passing philosophy. We know how to do that now
and I can supply printed reports to anyone who is interested in our 
particular solution.
    There is however also a lot of real science and engineering happening,
and its growing at an exponential rate as people talk to each other, and
to manufacturers (convince them of the tools we want !). Both Daresbury
Laboratory and Edinburgh University Parallel Computing Groups can supply
printed reports whowing the vast range of current REAL applications of
MIMD computers.
 
R.J.Allan, Daresbury Laboratory, SERC, Warrington, WA4 4AD, U.K.
rja@uk.ac.dl.cxa on U.K. JANET system