Newsgroups: comp.parallel.pvm
From: Simon Thorpe <thorpe@cerco.ups-tlse.fr>
Reply-To: thorpe@cerco.ups-tlse.fr
Subject: PVM in Byte Magazine
Organization: CNRS
Date: Fri, 07 Mar 1997 10:50:02 +0000
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <331FF2D9.37E3@cerco.ups-tlse.fr>

Some of you may have seen the article in the International section of
March's Byte magazine on Mainstreaming Parallel Computing. Although
generally encouraging it does contain the following passage:

"Some developers described the PVM (Parallel Virtual Machine)
message-passing library as unstable and unreliable. Other developers
felt release updates were issued too frequently. Several consortia
reported that PVM didn't fully support C++ features on some platforms,
so they had to spend time programming around missing funtions."

We are currently thinking of using PVM ourselves, and so I would be most
interested to hear the reactions of other members of this newsgroup. Is
PVM really as unstable and unreliable as the article seems to suggest?

Thanks for your comments

Simon Thorpe
Centre de Recherche Cerveau & Cognition
Toulouse, France