- Home /
What's a good method for measuring shader performance?
I've recently started to write shaders and while I try my best to keep performance in mind. I really have no clue how they compare with any other shader. I've put a FPS counter in my scene to help measure, but then I was thinking there must be a better way.
For my test, I created 4 scenes that contain planes with my shader applied to them. The first scene has 1 plane, the second has 10 planes, the third has 100 planes, and the last has 1000 planes. I log the average framerate and switch scenes every couple of minutes. I then compare the results with other shaders that have gone through the same process. I guess this kind of works or gives me ballpark figures, but I'd still like to know if there's a more accurate or accepted method of testing shaders.
I also check to see if the shader is batching, but other than batching and framerate, I'm not sure what else to look out for. Is my simple setup good enough? How do you accurately test shaders?
I found some interesting blog posts related to this subject. $$anonymous$$ostly focusing on functionality but still very cool. Big thanks to Aras Pranckevičius =D
http://aras-p.info/blog/2011/06/17/testing-graphics-code-4-years-later/
Your answer
Follow this Question
Related Questions
Batch texture sampling in shader? 1 Answer
Need help optimizing grid initialization. 1 Answer
What's killing my rendering performance? (pic included) 1 Answer
What is more optimized: use manually the shader variables or declare it and then use 1 Answer
Performance Question: Baked textures vs Procedural shaders 0 Answers