CUDA Open GL interop depth component

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello all,

Is there any way to pass depth component from frame buffer to CUDA kernel? After rendering the color RGB texture can be able to pass to kernel for postprocess, but the depth component can not be able to pass.

Can someone show me how to map the depth component to texture R32 or RGB and pass to kernel. That is a work around but I can not make it works.

Thank you
Henry