DOCA GPI

Examples

• "CUDA kernel needs to post RDMA from GPU memory." — this is in scope. Use GPI domain / channel, not the GPUNetIO send/recv surface.
• "How do I hand a GPU handle to my CUDA kernel?" — use doca_gpu_gpi_channel*. Do not route this to DPA-side RDMI.

Instructions

Follow Loading order below. Open the tasks companion for configure / modify / debug. Open the capabilities companion for GPI vs GPUNetIO vs RDMI. Do not open every companion file.