home *** CD-ROM | disk | FTP | other *** search
- Path: nntp.teleport.com!sschaem
- From: sschaem@teleport.com (Stephan Schaem)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Texturemapping speedu
- Date: 1 Jan 1996 23:25:31 GMT
- Organization: Teleport - Portland's Public Access (503) 220-1016
- Message-ID: <4c9qhb$v0@maureen.teleport.com>
- References: <4b433m$mei@sinsen.sn.no> <4bee9a$99j@news01.uni-trier.de> <38231780@kone.fipnet.fi> <4bu23o$h02@oreig.uji.es>
- NNTP-Posting-Host: julie.teleport.com
- X-Newsreader: TIN [version 1.2 PL2]
-
- Jorge Acereda (acereda@vents.uji.es) wrote:
- : Jyrki Saarinen (jsaarinen@kone.fipnet.fi) wrote:
-
- : > Argh. How about a 3d transformations code? 9 muls per pixel,
- : > not enough registers? Why not put those 9 values to the code
- : > and do a CacheClearU once a frame. Much faster..
-
- : 9 muls/pixel? AFAIK it can be done with 7 muls/pixel (or less if
- : you use spherical coordinates)
-
- : And back to the topic...IMHO SM code is fine when caches are
- : flushed only once per frame. Things like once per polygon and
- : the like seem too slow on A4000.
-
- Unless the polygon in question is a floor or ceiling.. Lets just
- say BIG .
-
- Stephan
-
-