Not sure if anyone can answer this, but I was wondering about something…
Are there any plans to expose the Tensor cores through the DXR Callable Shaders, in some way akin to how it’s done experimentally in CUDA with the Sub-Byte stuff?
This would make the DX12/DXR flow much more potent, with access to the RT, CUDA, and Tensor cores, all done through the call graph represented with the DXR Shader Tables, would it not?
Looking forward to any information that anyone has, thanks!
-Real-Time Chris