I am evaluating how fast calculations on the i.MX53 3D GPU can be.
Now I am a bit stuck since all the examples (and the original target of OpenGL) is always rendering to a display.
What I want to do is
Input data (2D, pixel data) --> GPU (Some per-pixel calculations) --> Output data
I need the output data again in my application.
Can you suggest a way to do this with OpenGL ES 2.0?
Solved! Go to Solution.
I implemented the virtual framebuffer approach.
If you care of some restrictions (regarding alignment, scaling etc.) I could manage to get a 1:1 mapping from input-(texture-)pixels to output-(framebuffer-)pixels.
See also here: Turn off anti-aliasing in i.MX53 GPU with OpenGL ES 2.0