Path: chuka.playstation.co.uk!news From: "Russ Williams" Newsgroups: scee.yaroze.programming.3d_graphics Subject: Re: 3D engines Date: 15 Apr 1997 21:56:37 GMT Organization: PlayStation Net Yaroze (SCEE) Lines: 17 Message-ID: <01bc49e8$7996d660$b756dec2@algorithm> References: <3347dd19.3248416@news.playstation.co.uk> <3348D38C.1DC1@interactive.sony.com> <334A3648.1873@micronetics.com> <3349266B.7BD4@interactive.sony.com> <3349a345.33045335@news.playstation.co.uk> <334A3058.218B@interactive.sony.com> <334b3a68.199522@news.playstation.co.uk> <334B7033.44F8@interactive.sony.com> <01bc46a2$06428400$01000080@solabh51.dial.pipex.com> <33527210.971450248@news.playstation.co.uk> NNTP-Posting-Host: algorithm.demon.co.uk X-Newsreader: Microsoft Internet News 4.70.1155 SCEE Developer Support wrote in article <33527210.971450248@news.playstation.co.uk>... > As an idea, you could probally get a massive performance increase, by > using a variation of the idea below, where a 'dummy' TMD is set up on > the scratchpad... When you register this TMD with LibPS, all of the > accesses to it's points/primitive structures will come from fast > scratchpad ram. Yes, but this requires that you either: a) Draw the TMD as soon as you calculate it using drawprim() or whatever. or b) Use <1k of 3d data ('b' isn't practical, and 'a' is a *minimum* of 50% slower than parallel DrawOT / calculate OT systems, since you cannot calculate a second frame while drawing the first) --- Russ