|prev : next : index||SPEW|
I've been writing some OpenGL performance testing code. The goal is to measure a given implementation looking for "optimal" approaches to certain tasks on that implementation (e.g. when's the best time to download textures). I will put all this stuff in the public domain when I'm happier with its stability.
I worry about people trying to use it as a benchmark. I never intended it as a realistic comparison between different OpenGL implementations--just as a comparison within a single implementation. There might be useful overall things to learn, of the form "on all cards, approach XYZ is optimal"... or, more likely, "There is no approach that is optimal on all cards."
Wrote some spew about PIPE, a project I'll never get around to working on. So I'll convert it to HTML and post it instead.