Amit Parghi (amit++at++discreet.qc.ca)
Wed, 29 Jan 1997 21:47:26 -0500
As Brian Corrie suggested, some potential snags lie within the reduced
frame buffer capacity and the available pixel formats using the 40MB raster
manager (RM8) on the Reality pipe. However, as Trina Roy (hi, Trina!)
pointed out, there is also a potential issue in the reduced geometry
throughput on the GE boards ("reduced" relative to an InfiniteReality, that
is).
(Please note that I don't know much at all about the ImmersaDesk and I'm
not addressing its requirements specifically here!)
RM8: For double-buffered stereo at 1280x1024, my (somewhat preliminary)
information tells me that you can get:
- RGBA 12/12/12/12 pixels with either
- 31 bits of Z and one bit of stencil; or
- 32 bits of Z.
- RGB 10/10/10 pixels (no destination alpha planes) with one of:
- four-sample multisampling;
- 32/32/32 accumulation; or
- 1 aux, 32 bits of Z, and 8 bits of stencil.
While it would be possible to add a second RM8 to the Reality pipe to get
you back into the world of Medium pixel depth, you should compare the
pricing of this option against a single-RM7 InfiniteReality system.
GE: Due to the design of the GE on the Reality pipeline, you should expect
its geometry throughput (as opposed to pixel-fill) to be approximately half
of what you would get on an InfiniteReality pipeline. I think this is also
made clear on the table on the back page of the Onyx2 short brochure (the
little eight-page thing).
Ciao,
Amit
--
AMIT PARGHI DISCREET LOGIC
amit++at++discreet.com 5505, boul. St.-Laurent, bureau 5200
+1.514.272.0525 Montreal (Quebec) Canada H2T 1S6
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:54:30 PDT