Date: Wed, 23 Oct 91 10:22:20 -0600
From: dirish@math.utah.edu
Message-Id: <9110231622.AA28492@alfred.math.utah.edu>
To: glove-list@karazm.math.uh.edu
In-Reply-To: Dave Stampe-Psy+Eng's message of Tue, 22 Oct 91 22:53:13 -0400 <9110230253.AA18321@watserv1.uwaterloo.ca>
Subject:  Transputers...
I have also considered transputers for high speed rendering.  I was
interested in complex scenes with thousands of polygons rather than
stereo, but they should work for either.  My idea was one transputer
with lots of ram to store the display list and n transputers suitably
connected to video ram to do the rendering.  Then the PC would be used
to control the whole system and to talk to the user.  After seeing the
specs on the new T9000 (Byte, August 1991) I realized that these would
make a very powerful system.  However, at this point in time I don't
have the money to buy a bunch of transputers.  (Just an aside, there
really isn't an assembly language for the transputer.  If the C
compiler doesn't generate fast enough code for you then you would
write in OCCAM.  However, from what I have heard of the quality of the
C compiler, you are unlikely to need to do this.)

However, I haven't given up on the transputer.  If you have addresses
for companies which are selling PC transputer cards I would love it if
you sent them to me.

I really think that people who are interested in VR are going to have
to look at some kind of parallel architecture.  It is very unlikely
that you will be able to get a single serial CPU that is fast enough.
Also, VR is quite amenable to parallelizing.  A cpu for input devices,
a cpu (or more) for stereo output, a cpu for sound, a cpu for the
motion platform, a big cpu to coordinate things, and a collection of
cpu's to model objects.  The only question is whether 100Mbps
interconnect is fast enough.  Anybody know of a transputer board that
has NTSC output and another with a/d and parallel inputs?

Dudley Irish
________________________________________________________________________
Dudley Irish / dirish@math.utah.edu / Manager Computer Operations
Center for Scientific Computing, Dept of Mathematics, University of Utah

The views expressed in this message do not reflect the views of the
Dept of Mathematics, the University of Utah, or the State of Utah.