Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!cs.uoregon.edu!ogicse!ucsd!sdcc6!rblewitt
From: rblewitt@sdcc6.ucsd.edu (Richard Blewitt)
Newsgroups: comp.sys.amiga.graphics
Subject: Re: Molec3D
Keywords: molecular graphics, amiga
Message-ID: <20882@sdcc6.ucsd.edu>
Date: 22 Jun 91 07:18:57 GMT
References: <1991Jun21.181649.25505@menudo.uh.edu> <20843@sdcc6.ucsd.edu> <1991Jun21.222643.14055@menudo.uh.edu>
Organization: University of California, San Diego
Lines: 28

In article <1991Jun21.222643.14055@menudo.uh.edu> jet@karazm.math.uh.edu (J Eric Townsend) writes:

>Har.  This package is a general parallel computing package written
>in a language invented here for that very purpose.  Currently it
>runs in an Intel iPSC/860.  We see > Cray-2 performance with
>16 i860 nodes.

>Now if a friend of mine gets off his but and builds is dream
>parallel machine (bunch of '030s in a box with an ethernet port) maybe
>we'll see it running on a machine that costs < $300,000. :-)

I know this is diverging from the Amiga a bit, but have you looked
at the algorythms put out by Klaus Schulten's group?  They were
working with a connections machine, and a homebrew setup consisting
of 20 t-800 transputers sitting in a PC.  Got about 1/4th cray speed
for ~20K, the connections setup was much faster than the cray
(actually, they said that the cray would not even work with large
systems).  Getting back to the Amiga, a few multiprocessing boards
stuffed in Zorro III slots,  along with a 34020 graphics board setup
could make on hell of a killer modeling workstation, for a mere
$25-30K.

Rick

_____________________________________________________________________
_____________________________.sig____________________________________
_____________________________________________________________________
The generic .sig          Rick Blewitt     rblewitt@ucsd.edu