Different sized textures - one algorithm?

Hi all,

I have two textures. T1 is e.g. 1000x8. Each of the 1000 “rows” represents an index set into the other 1331xn floating point texture T2. I need to perform certain ops depending on the values of T1 and I may look up different texels of T2 multiple times. What I want to do is use a GLSL fragment shader to render into a floating point P-buffer and get the results back using pbo and vbo. Final results from T2 will then be used to render geometry (in a vbo) and for further iterations through the algorithm (i.e. repeat the whole process ad nauseum).

My problem is this - to access each element of T1 I need a rectangular P-buffer of e.g. 1000x8 bound to a viewport sized quad. But then to read and “write” T2 I need a different sized quad. I need one-to-one mapping of pixels to texels for both textures obviously. Any ideas? Am I missing something simple? I’m tired so the answer may be right in front of me and I just can’t grasp it.

Thanks in advance.