Reputation: 5492
When I was reading up on measuring Frames Per Second in OpenGL I came across this:
How to measure true FPS performance of OpenGL ES?
What I got from it was that you can get the FPS but its not a true indication of performance, and that Frame Duration would be a better choice. Are their any other ways to measure rendering performances for OpenGL? And if so why are they better (or worse) than the last?
I'm not sure I understood why FPS is not a good indication for performance too.
Upvotes: 1
Views: 920
Reputation: 35933
It's not so much that it's not a 'true indication of performance', it's just that if you don't know what you're talking about it can be easily misleading.
If you really want to target a certain FPS, then there's nothing wrong with looking at it as a metric, but you just have to remember that it's not a linear measure of performance.
For example, if your FPS goes from 30 to 20, then that's a pretty significant decrease in frame performance time (33 ms to 50 ms)
However if your FPS goes from 2000 to 400, that's a miniscule difference in real time (only 2ms difference).
Using FPS is discouraged because then you get new developers worked up in a panic saying "my FPS decreased by 1600, what did I do wrong??"
Upvotes: 3