tags:

views:

118

answers:

4

Is this even possible?

A: 

I never have seen anything like that. Normally you render a frame as fast as possible do some CPU frame post- or pre-processing and render the next one, so usage flaps between 0 and 100%. Only rarely the FPS are limited to a maximum number and only in this case would this be a meaningful number.

fschmitt
A: 

I think it's impossible, cos GPU consists of many parts and each of them could be a bottleneck.

You can test for these bottlenecks though, to see where you need to optimize.

Krom
+2  A: 

Nope. It's even hard to rigorously define in such a highly parallel environment. However you can approximate it with ARB_timer_query extension.

ybungalobill
could you give example of usage ARB_timer_query extension ?
Newbie
+4  A: 

Not really, but you can get differente performance counters using your vendor's utilities, for NVIDIA you have NVPerfKit and NVPerfHUD. Other vendors have similar utilities.

Matias Valdenegro