home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.graphics:9385 comp.graphics.visualization:1295 comp.sys.ibm.pc.hardware:23255
- Newsgroups: comp.graphics,comp.graphics.visualization,comp.sys.ibm.pc.hardware
- Path: sparky!uunet!orca!orca!pmartz
- From: pmartz@dsd.es.com (Paul Martz)
- Subject: Re: Looking for polygon engine hardware
- Message-ID: <1992Sep2.192347.27239@dsd.es.com>
- Sender: usenet@dsd.es.com
- Nntp-Posting-Host: bambam-gw
- Reply-To: pmartz@dsd.es.com (Paul Martz)
- Organization: Evans & Sutherland Computer Corp., Salt Lake City, UT
- References: <1992Aug27.140202.21662@eng.umd.edu> <34709@cbmvax.commodore.com> <1992Sep2.112611.29917@news.uni-stuttgart.de>
- Date: Wed, 2 Sep 92 19:23:47 GMT
- Lines: 18
-
- In article <1992Sep2.112611.29917@news.uni-stuttgart.de>, andreas@awstar.rus.uni-stuttgart.de (Andreas Wierse) writes:
- > I think Mike didn't just speak about fill-rate, but meant full 3D
- > Gouraud shaded triangles. This means 20,000 * 3 matrix calculations to
- > perform every second (if it's a triangle mesh ~ 20,000 +), the whole
- > clipping stuff, color interpolation, polygon breakdown to spans and
- > pixels etc.. I'm quite sure that even a fast 486 is not fast enough to
- > render so much triangles.
-
- So what he really wants is a card that will do the 3d transform, color
- interp, and scan convert. Then all the cpu would have to generate
- would be the transform description and the list of WC vertices, and
- send that info to the graphics board for processing. Is there such a
- card at all in the pc market? I know this type of stuff is widely
- available on workstations.
- --
-
- -paul pmartz@dsd.es.com
- Evans & Sutherland
-