Hardware Accelaration

How can I know if a function is hardware accelerated ?

Is glClipPlane hwd accelerated ?
Is Stencil Buffer hwd accelerated ?

How can you know if a function is hardware deccelerated?

The point is, we don’t care where, what, who does it. You just want the software to run at a reasonable speed and not lock the users machines since its taking to much time to render a single frame.

Check out this forum and suggestions forum for other reasons why opengl doesn’t tell you anything about this.


Nobody cares about hardware acceleration !?
Thanks V-Man for your help.

I was about to use glClipPlane in my program with good results on a TNT2 but on my Geforce3 glClipPlane is too slow, so I care about hardware acceleration !

Just do a search of the forums for it. You will see what V-Man is talking about. To summarize, you only care if something is fast or not. Knowing that it is hardware accellerated or not tells you nothing about how fast it is. For instance, I was testing on a very old 3D card. The drivers told me it supported 1024x1024 textures in hardware. Cool. However, using a 1024x1024 texture forced the card to constantly swap the texture in and out. Result was about 2 FPS. It was still hardware accellerated, but that didn’t really tell me much, now did it.

…Knowing that it is hardware accellerated or not tells you nothing about how fast it is…

Really ?

In other words, you should run a short benchmarking function before your app decides what features to use…eg. create a clip plane and render a ton of polys intersecting it for a few frames, and time how long it took…compare this time value to what you have previously asserted to be acceptable, and if its as fast or faster than your threshold then use clip planes, otherwise don’t use them.

Ok, thanks. Everyone.