I’m writing a reaction-diffusion shader and managed to get it running but I would like to accelerate it.
I would like to run the algorithm several times before spitting out of the GLSL node.
I think what I need to do is process the entire image out to an FBO and then repeat as desired before rendering it out.
Creating/writing to an FBO seems really complicated though! Can anyone point me to some examples?
There is a “passes” parameter for all filter TOPs, including GLSL tops. I believe this simply takes the result and puts it back through the same shader again as many times as you specify. I recall using this very feature for a reactive diffusion implementation and getting pretty cool results.
Not at a TD computer right now, but I think that parameter is on the Common page…?