Basically I want to share vertex buffer data from one process to another through CUDA shared memory.
There’s similar questions to this one, but none of them have complete answers so imma ask again.
The basics of this would be to get the device pointer of the RHIBuffer (say for example a vertex buffer), register it as a cuda dx12 interop buffer (or whatever graphics API I might be using), then copy the data GPU → GPU to the IPC shared buffer.
The problem I’m finding myself in is that I can’t seem to find a way to leave the RHI abstraction layer and go into the specific graphic API to retrieve the device pointer. I found out that there’s a function already available for FRHITexture called GetD3D12TextureFromRHITexture(), but nothing similar is available for FBufferRHI.
Thanks for any help provided