What I am trying to do is process a series of captured images using CUDA and then display the results using opengl without transferring the images back to the host. This is my first time attempting to use opengl and I am quite lost. I basically copied the simplegl sample but used my image processing kernel. The issue that arises is that the sample uses cuda to calculate new vertex locations but I am using cuda to calculate new pixel colors at fixed vertex locations.
How would I use the cuda/gl interop to do this? Does using glDrawPixels make sense (I have RGBA imges) and would it be easy to implement using the framework already available in samglegl? If so how would I go about doing it?
Also, I have read that are some serious performance hits doing this with cuda 1.1, but I’m hoping that will be fixed in the next release