Re: Generating 4 channels on one iR at 60 Hz?!?

New Message Reply Date view Thread view Subject view Author view

Hendrik-Jan van Veen (veen++at++chicory.mpik-tueb.mpg.de)
Fri, 22 Nov 1996 10:26:57 +0000

On Nov 19, 15:16, Bernard Leclerc wrote:
> Do you believe it's possible to generate 4 medium resolution
(800x600)
> channels at 60 Hz on a Onyx2 InfiniteReality Deskside with 1 pipe, 2
> RMs, DG5-8 option and 4 CPUs?
> According to the specs and using a depth complexity of 2 for low
level
> flight, the fill rate would be sufficient (230/388 Mpixels/sec) and
> the polygon capacity would be large enough (11M/4/60 = 45800 polygons
> per frame per channel -- 10000 would be more than enough).
> However, I'm a bit uncomfortable with the idea of having only 4 msec
> to draw one channel knowing the field of view would be 60 by 45
> degrees.

Since you complain not getting any reply's at all I'll give you one. I
don't have an IR (just RE2) but we're about to order a 3-pipe IR ONYX2.
In fact, I would be interested to hear if I made any errors below...
Just from the specifications we can learn a few things.

Four times 800*600(SVGA) amounts to 1.92M pixels. This means that
a> Frame Buffer size, which is 160MB with two RM's, is large enough to
support 512bits pixels on all four displays.
b> Pixel Fill Rate for 2RM's is approximately 400M pixels/sec, or 6.6M
pixels/frame at 60Hz. Thus, an average depth complexity of 3 should be
possible. This is not much, but it depends strongly on your application
whether it's sufficient or not.

My experience so far is limited to a 4 processor 4RM RE2 ONYX with
MultiChannelOption. We use it to project 3 images of 1024*768 on a 180
by 55 degrees projection screen. It's limited everywhere, and
especially the drawing process (performer APP_CULL_DRAW mode) takes a
lot of time. I guess 4 channels makes life even worse, even with a IR.
My guess is that you will be limited by the CPU's. But if you have a
smart database and application then it might work. For us this was
reason enough to go to three pipes and 8 or 10 processors, splitting
the drawing process in parts and dividing it over three processors. I
guess the field of view is primarily of importance for the culling
process, a smaller frustrum makes life easier. Having three culling
processors in our case seems reasonable (180*55 deg FOV). In your case
(60*45 deg FOV in total?!) this might be unnecessary.

Any comments, anyone?

Hendrik-Jan van Veen

Max-Planck-Institute for Biological Cybernetics
Spemannstrasse 38 -- 72076 Tuebingen -- GERMANY
E-MAIL: veen++at++mpik-tueb.mpg.de
EUROPE: http://www.mpik-tueb.mpg.de/people/personal/veen/veen.html
USA:
http://www.cog.brown.edu/mpik-mirror/people/personal/veen/veen.html
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
            Submissions: info-performer++at++sgi.com
        Admin. requests: info-performer-request++at++sgi.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:53:59 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.