Newsgroups: comp.parallel
From: rcarter@best.com (Russell Carter)
Subject: Re: Parallel platform
Organization: Best Internet Communications, Inc. (info@best.com)
Date: Thu, 3 Aug 1995 03:20:57 GMT
Message-ID: <3vpfap$pui@shell2.best.com>

In article <3vbv3h$t4l@kadath.zeitgeist.net>,
Kevin Cameron  <kevin@chronologic.com> wrote:
>In article <3tte8h$qda@convex.convex.com>, Jens Brauer <brauer@convex.convex.com> writes:
>|> Dear Netter:
>|> 
>|> Our group plans to procure a distributed-memory parallel
>|> computer platform to perform large-scale MD simulations.
>|> Initially we may not be able to afford many nodes, but we
>|> expect to get more money to buy more nodes in the coming 
>|> years. So the ideal system should be expandable. Could
>|> someone give us some information what kind system may be
>|> a wise choice at current time and in future and where we
>|> can get quotation for such system? Thanks in advance.
>
>Personal opinion: just buy Sun multiprocessor sparc stations and try to get
>FDDI net between them, and write your code in C/g++ so it's easily ported to 
>other hardware later. I tried doing a distributed simulator on transputers 
>once - the tools where awful, it's much easier doing it on general purpose 
>hardware.
>
>If you don't like that idea try: http://www.parsytec.de/

Or if you don't like that idea try: http://www.geli.com/
and see what can be done for about 1/2 the price/flop.

Best regards,
Russell



