I’m trying to a programming guide like this here http://developer.download.nvidia.com/GPU_Programming_Guide/GPU_Programming_Guide_G80.pdf
except for AMD 7000’s series (or newer NVIDIA/ATI hardware). G80 architecture is a little old and computation cores have increased in capabilities since then.
Another thing that could help is programming hints. An example would be how ,ost boards say that display lists are the FASTEST way to render static geometry on NVIDIA-class hardware.
I’m also wondering if texture fill-rate is affected by the fragment shader. If a fragment shader works only on textures does that affect the texel fill-rate OR does texel fill-rate affect things such as the speed of a texture fetch or something. Maybe things will be clear when I understand and experiment with OpenGL but I have no clue (what I DO KNOW is pixel fill-rate).