Hi,
I am struggling finding resources that would exemplify how to handle pinned memory buffers (creation/destruction) and async memcpy (with queue selection) using HSA for AMD multi-GPU program. Resources exist for HIP and OpenCL ways of doing it, but couldn't find any docs on how to proceed using HSA.
The HIP functions I'd like to implement using HSA would be
hipHostMalloc
hipHostFree
hipMemcpyAsync
Thanks for help!