Fast OpenGL interop for CUDA anyday?

Accelerated Computing CUDA CUDA Programming and Performance
1

I see since ages in release notes that OGL interop is slow goes through host and as seems OpenCL on MacOS provides fast OGL interop (stays in device mem) I ask if the are strong plans on getting this limitation away or not?

2

I see since ages in release notes that OGL interop is slow goes through host and as seems OpenCL on MacOS provides fast OGL interop (stays in device mem) I ask if the are strong plans on getting this limitation away or not?

3

This is a very good question. I have a client which is hoping to use OpenGL interop on a

MAC sometime in the near future. What is the prognosis on this ?

Thank you.

MW