mapping scientific data to a texture

hi, I have been writing a shader to perform some vector math on scientific data. My inexperience with OpenGL is causing some problems in the texture setup. I want to have a 128 bit per pixel texture with 4 32 bit floating point components. The book I have doesn’t have that as an option when setting up the texture, does OpenGL support this kind of texture? Also what kind of pointer should I use to point to this texture? unsigned double double (is that possible)? Any help would be greatly appreciated.

thx.

ATI and NVIDIA both support 32bit/channel textures(single precision) using their own extensions. Check out the OpenGL registry for exact details on them. For float data, a float pointer is always nice.

thx. yeah, the float pointer makes sense now, I had some weird picture of the array structure packing all 4 channels into one element…doh.