GPUDirect RDMA access through DirectX

I am switching from AMD gpu to Nvidia gpu and have some queries to use DirectX in Nvidia gpus.

  1. I found online that Nvidia GPUDirect RDMA supports CUDA/OpenGL/DirectX but could not find any examples where DirectX was used to access RDMA. So can you let me know if I can use DirectX APIs to access RDMA , if yes then can you provide the steps to perform this.
  2. What is the maximum & minimum memory size which can be allocated for GPUDirect RDMA in Nvidia gpus??

Please provide appropriate response.

Where you get info opengl/driectx related with GDR(GPU Driect RDMA)?

OpenGL/DriectX is graph libs. GDR is for communication between node(through RDMA NIC HCA), We support GDR through CUDA/infiniband Verbs, base on Tesla GPU and NVIDIA infiniband/RoCE.

I found the info from this official document - https://developer.download.nvidia.com/devzone/devcenter/cuda/docs/GPUDirect_Technology_Overview.pdf
Slide 2
Please confirm if my understanding is correct or not.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.