[net.graphics] Parallel Ray-Tracing

ksbooth@watcgl.UUCP (10/26/84)

[Response to net.graphics request for info]

A PhD student of Jim Kajiya's at Caltech has written an analysis of a number
of ways in which ray-tracing (and other) algorithms can be run on parallel
architectures.  His thesis is:

Parallel Machines for Computer Graphics
Michael K. Ullner
California Institute of Technology (1983)

John Cleary and Brian Wyville of the University of Calgary are working on
parallel machines for ray-tracing.  Some related work of theirs is reported
in the Graphics Interface '84 (Ottawa) conference proceedings.

	Kelly Booth (Computer Graphics Laboratory, University of Waterloo)

ken@turtlevax.UUCP (Ken Turkowski) (10/29/84)

> [Response to net.graphics request for info]
> 
> A PhD student of Jim Kajiya's at Caltech ...
> 
> John Cleary and Brian Wyville of the University of Calgary ...

Don't forget the LINKS project, fruit of a liason between the
University of Osaka (I think) and industry.  All of the colorful
surrealistic images produced by Yochiro Kawaguchi are produced on this
machine, an array of 64 Z8000's (bet you thought they had been
surpassed by the 68000 and 32016, eh?).  The LINKS-2 is supposed to be
an array of custom-designed processors.  Both of them are designed to
do nothing but ray-traced images.
-- 
Ken Turkowski @ CADLINC, Palo Alto, CA
UUCP: {amd,decwrl,flairvax,nsc}!turtlevax!ken
ARPA: turtlevax!ken@DECWRL.ARPA

drforsey@watcgl.UUCP (Dave Forsey) (10/31/84)

Actually I believe the LINKS system has been using 256 of
the "unit computers" for some time and hope to have a 1024 version
running shortly.

Each "unit computer" consists of the Z8000, an 8068, an 8087 and
1.7 Meg of memory.


David Forsey
Computer Graphics Laboratory
University of Waterloo
Waterloo Ontario Canada