Newsgroups: comp.parallel.mpi,comp.parallel.pvm
From: Lutz Laemmer <laemmer@sect.mce.hw.ac.uk>
Subject: Re: PVM and MPI
Organization: Heriot-Watt University
Date: Mon, 02 Feb 1998 12:46:41 +0000
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <34D5C031.7621@sect.mce.hw.ac.uk>

James Cownie wrote:
> 
> I'm glad to see that you managed to come up with an inoffensive
> way of levelling the playing field a bit. It does irritate me
> when Al's paper is presented as the last word on MPI/PVM issues
> without any comment that Al is (was ?) heavily involved in PVM...
> 
> - Jim
> 
> James Cownie                    <jcownie@dolphinics.com>
> Dolphin Interconnect Solutions  +44 117 9071438
> http://www.dolphinics.com/toolworks

I really like the definitions found in Al's paper. I am just wondering
about the statements on portability and interoperability. 

Would somebody from the MPI developers comment on the lack of
interoperability of MPI implementations as stated in chapter 3 of Als
paper

"Nothing in the MPI standard describes cooperation across heterogeneous
networks and architectures. [...] None of the existing MPI
implementations can interoperate."


Lutz Laemmer                          <laemmer@sect.mce.hw.ac.uk>

Heriot-Watt University   Department of Chemical & Mechanical Engn
                  Structural Engineering Computational Technology