Array of Textures in driver API

Hello,
I need to have an array of 2D textures in kernel, something like this:

texture<float, 2, cudaReadModeNormalizedFloat> *tex_array;

or

texture<float, 2, cudaReadModeNormalizedFloat> tex_array[N];

simply - an array of textures. Both above lines are compilable (tried), so I think there should be a way. The problem is, how do I load data into them? Just making array of CUarrays? The other problem - linking them with texture references, cuModuleGetTexRef needs the texture name, but they would be technically in an array, so in that case I need handles to particular array members. How to do this stuff? Thanks in advance…

I’m using driver API and Tesla C2075 (Fermi).

Search for ‘texture arrays cuda’.
Gives e.g.

https://devtalk.nvidia.com/default/topic/376700/an-array-of-texture-references-/

I would use runtime API instead driver API, and as far as i know one can mix both in one program.

And for Kepler generation and above i would use texture objects, they are simpler to use.