Newsgroups: comp.parallel,comp.parallel.mpi,comp.parallel.pvm
From: carlos@triton.lpac.ac.uk (Carlos Falco Korn)
Subject: Re: HPF (was:  MPI vs. PVM)
Organization: London Parallel Applications Centre
Date: Tue, 07 May 1996 11:33:19 +0100
Message-ID: <4mnff3$c2v@usenet.srv.cis.pitt.edu>

> Dave Kohr <kohr@mcs.anl.gov> writes:
> > I've spent some time browsing DEC's HPF page, at
> > http://www.digital.com/info/hpc/f90/f90.html, looking for performance
> > results in particular.  Other than the numbers from the DEC Technical
> > Journal paper (which I also looked at), and on the page
> > http://www.digital.com/info/hpc/f90/optimizations.html, is there anything
> > else?

Take a look at following paper:

Practical Experience of HPF for Financial Applications, 
Carlos Falco Korn, Adriaan Joubert and Peter Dzwig, 
in E.H. D'Hollander, G.R. Joubert, F.J. Peters, D. Trystam (editors):
Parallel Computing: State-of-the Art Perspective (PARCO '95), 
Elsevier, to appear 1996. 

which can be found in
 
     ftp://ftp.lpac.ac.uk/pub/users/adriaan/Parco95.ps.gz


There we present a complex problem which we have implemented using
Digital's HPF compiler and several tricks. The paper contains tables 
with measured run times (of the HPF code and the *best* serial code).

Enjoy,

  Carlos


=======================================
Dr. Carlos Falco Korn
The London Parallel Applications Centre
Queen Mary & Westfield College
Mile End Road
London, E1 4NS
United Kingdom
e-mail:  C.Falco-Korn@lpac.ac.uk 

