Hi,
I’d like to code a kernel to do image processing in a high definition video. I want to test the interoperability between the output of a directshow sample (vmr9allocator) and CUDA. But the result shows that the texture from playback video can’t pass it to CUDA when using DirectX interoperability API (cudaD3D9RegisterResource, cudaD3D9MapResources, cudaD3D9ResourceGetMappedPointer). Any ideas how to do that? Could you explain in detail please. Thank you!