From: Joel K. <jj...@ya...> - 2004-09-09 10:39:49
|
Jonathan Brandmeyer wrote: >Cluster? No, I think that would >be a bad idea for VPython. Bad as in "not cost-effective" (sort of your next point), or bad as in "no significant speed increase at all in a cluster"? I'd be interested in finding out more about whether VPython can benefit from any kind of parallel processing. >Any recent video card will be able >to handle pretty much anything >you can throw at it from VPython. >A scene with several translucent >and textured objects, some using a >source image of 1024x2048 pixels, >can be rendered in only a few ms on >common PC hardware. You really >can't notice that it takes any time at >all, and the UI remains buttery-smooth. I can speak only from my own experience with my 2Ghz Athlon processor, 256MB system RAM, and S3 Graphics ProSavageDDR card (32MB graphics RAM), usually kept in 1024x768 true-color mode, all operating under Windows XP. In this environment, I have been able to put together objects that caused unacceptably slow renderings, "low virtual memory" warnings, and even the odd system crash. When I animate some complex scenes (usually using "rotate" commands in loops), this can produce more issues. Mostly the problems have resulted when I was working with either curves or faces, and running up into thousands of points, *and* doing fancy vector-based rotations as well. I suspect that this is just what happens when a nut like me starts trying to turn out artworks with a language designed for doing physics. :-) Seriously, though--VPython is a great system, and I fully intend to keep exploring it; but everything has its limits, and more powerful hardware *can* help. Joel __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com |