Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!maverick.ksu.ksu.edu!uafhp!kcampbel
From: kcampbel@uafhp.uark.edu (Keith Alan Campbell)
Newsgroups: comp.sys.amiga.hardware
Subject: Re: New hardware idea!!!!!
Summary: Graphics coprocessors...
Message-ID: <6308@uafhp.uark.edu>
Date: 16 Apr 91 23:00:58 GMT
References: <1991Apr4.015718.14559@sbcs.sunysb.edu>
Distribution: na
Organization: College of Engineering, University of Arkansas, Fayetteville
Lines: 16



Ever since we started playing with DSP56001's a year ago when developing the
SunRize 16 bit digital audio system, we wondered if 56001's (plural) might be 
usable as specialized graphics coprocessors. A friend is going to move some of
the math routines from DBWRender over to 56001 asm to test the concept. Common
sense would imply that for fixed point operations they should be at least as 
fast as an '882 at the same clock speed, if not faster. With multiple DSP chips
the system could subdivide images into chunks that could be rendered  
simultaneously, achieving significant speedups by using parallel processing.
If this idea can be proven successful, we will try to see if any of the 
commercial renderers would be willing to test some of their code on the system.
I could forsee a 4 DSP board very easily...

Don Kennedy
Vision Quest Systems