From: mklapman@adsl (Matthew Klapman)
Subject: Re: Highest Performance VR System
Date: Mon, 5 Aug 1991 21:32:16 GMT
Organization:  Future Vision Technologies & University of Illinois, ADSL



In article <1991Aug5.180546.23821@milton.u.washington.edu> edsr!jem@uunet.UU.
NET (John E. Mendenhall) writes:

>I have to disagree with you concerning the highest polygon rates.  Of course, 
>you may have been referring to systems that are for sale.  The PixelPlanes5 
>from UNC @ Chapel Hill is a custom designed machine that UNC stated had a 2M+ 
>polygon rate.
>
>I am in no way connected with UNC, except that I am interested in the highest 
>performance machine available.
>
>_______________________________________________________________________
>
>John E. Mendenhall                            Research Systems Engineer
>Electronic Data Systems   7223 Forest Lane MS RD1      Dallas, TX 75230
>EMail: jem@edsr.eds.com   Office: (214) 661-6058   Home: (214) 964-0938
>_______________________________________________________________________


The PixelPlanes5 is the highest performance polygon engine around. 
The 2 million polygon rate is for phong shaded polygons of arbitrary size...
yes- arbitrary.  Each pixel has its own processor (massively parallel!!!), and
the processor would decide whether it should be on or off.  Each pixel
receives the same set of parameters at the same time as all the rest.

Anyway, I'm not an expert on it, but was very impressed with their research
team.  I hear that version 6 will have its design completed this month.


In the following are my comments with Siggraph.

(The following is my own personal opinion.)

The VR exhibit was very poor overall, but had some great highlights:

Fake Space Labs had the best display unit at the show (LCD screens are
unbearable) and reinforces my view that high quality wire-frame is better than
"see-the-space-between-the-pixels" color LCD screens

LEEP optics portable telepresence unit made the LCD screen slightly bearable.
Their lens system is the best around (as far as I've seen).

UNC had the best graphics processor around (even over E&S), but it was better
looking at the output on a desktop monitor than in a VPL eyephone...
it would be great to see the PixelPlanes5 output in Mark Bolas' (Fake Space)
display unit.... also, the design for the rendering card is a work of art.

The rest of the show was very impressive from the software point of view
except for the majority of the VR worlds which were rehashed cities.
A few small firms had extremely impressive demos (such as a "real-time"
Mandelbrot engine that would allow you to travel into, or zoom into, the
fractal.

The "Electronic Theatre" was the highlight of the show with an impressive
montage of mathematical visualizations and flowing art.  I would like to see
the red-green audience participation piece be used at sporting events such as
the Superbowl or the Olympics.  The audience could vote on different issues
with instant tallies...could be interesting.

Now for some sleep...work...sleep.

-Matthew Klapman
 mklapman@adsl.uiuc.edu

 Future Vision Technologies