Newsgroups: comp.parallel.pvm
From: gmj@crab.pha.jhu.edu (Geir Magnusson)
Subject: Re: PVM under Windows?
Organization: JHU Dept Physics
Date: 1 Sep 1994 14:08:41 GMT
Message-ID: <gmj-010994100655@windward.pha.jhu.edu>

In article <aorndorff.268.2E64781E@nhtsa.dot.gov>, aorndorff@nhtsa.dot.gov
(Andrew Orndorff) wrote:
> 
> In corporate and gov't environments, you probably (almost guaranteed) will not
> find any support for Linux - it's nearly unsupportable in networks of
> hundreds or thousands of machines,

UNIX is unsupportable in large networks?  Huh?  And you claim that Windows
is fine?  I have trouble keeping Windows from hosing itself in a _single_
machine environment.

> and it's a completely different environment
> to train users to.

But a parallel message passing library (that works...) isn't?  It appears
to me that what the original poster wanted was to utilize his considerable
investment in hardware.  To want Joe User (or Jo-Anne User...) to use it
seamlessly in Windows(tm) is, I believe, an _entirely_ different story.

> If that sort of network has computational needs that can be filled by
> using the existing CPU's, i.e. PC's running Windows and PVM/Windows, with 
> littlle or no additional investment, then that could be very attractive.

Hey, I think that a network of cheap 486's running linux and pvm is the
ideal solution for a cheap parallel machine...  (in my deranged opinion...)

But it would seem to me that it would be an incredible investment to get
something as "minimally featured" (to be charitable) as Windows(tm) to
support PVM.  

As someone else pointed out, why not run OS/2 for Windows?  There is
allegedly a port for PVM, it runs Windows(tm) applications, and you can buy
it retail for $25.00.


Geir Magnusson Jr.                       while (ms--)
Dept of Physics and Astronomy            {
Johns Hopkins University                     linux++;
gmj@crab.pha.jhu.edu                     }

