Hey i have done a project in OpenGL for a subject i have in college. Nothing huge…just a car with animation, evaluators, lamp posts, textures, tiling…and though i’ve used display lists to render some groups of objects, i obtained low fps…
I want to expand mt scene to a little town but i’d like performance as well…What is the major factor to consider in larger projects?
And BTW i would also like to see a nice tutorial on vbo’s.
Try shrinking your window. If it is still slow when the window is small, then you are probably geometry limited. Try reducing your scene complexity to see if one particular object is slowing you down.
I have a geforce 4 Ti4200. Well, Quake3 runs fine but of course i don’t want to build an application similar in complexity as Q3. But i get 30fps running two houses,evaluators…how can i get the best from the graphics card?
Can you try your app without evaluators to see how they’re affecting your framerate? I could be wrong, but I suspect that evaluators aren’t on a driver “fast path”.
If your control points are all static you should be able to do the “evaluation” once and put your object geometry in a static VBO. Of course, that won’t help if the evaluators or the geometry transfer isn’t your bottleneck, but I’d be really surprised if your app ran any slower with this change (and there’s a possibility it could run much faster).
Originally posted by secnuop: Can you try your app without evaluators to see how they’re affecting your framerate? I could be wrong, but I suspect that evaluators aren’t on a driver “fast path”.
I also point my finger at that. GLU evaluators are software by sure means and NV_evaluators are being deprecated anyway. I’m pretty sure nothing of that stuff is supported as a performance path.