Newsgroups: comp.parallel.pvm
From: don@hp71550n.neuronet.pitt.edu (Donald Krieger)
Subject: Re: Assessing Workstation clusters
Organization: University of Pittsburgh
Date: 7 May 1997 12:27:23 GMT
Message-ID: <5kpsfb$o0v@usenet.srv.cis.pitt.edu>

Gareth Evans (gevans) wrote:
> Hi all,

> I looking for fairly recent papers/reports describing work concentrating on
> assessing the potential for using a Workstation cluster as a single
> computational resource (via PVM or similar) for the implementation of parallel
> algorithms. I'm particularly interested in poor results - bad speedup - due to
> expensive inter-processor communication perhaps or results showing
> linear/super-linear relative speedup.

> Anyone have any pointers to such work?

> Thanks

> Gareth.

	The stuff we have been working on is reasonably well documented at
 http://www.neuronet.pitt.edu/~don/papers.html.  The most complete description
is entitled: Workstation Cluster Use in a Medical . . .

								Don
--
  Don Krieger  Neurosurgery  Children's Hospital of Pittsburgh
  don@neuronet.pitt.edu  http://www.neuronet.pitt.edu/~don/