Open
Description
Follow-up of #208.
We should revisit the simple wrapper over cudaMalloc
once we have better MemoryResource
, and for that we might need pool_memory_resource
from RMM or perhaps CCCL if it's moved there by then.
Follow-up of #208.
We should revisit the simple wrapper over cudaMalloc
once we have better MemoryResource
, and for that we might need pool_memory_resource
from RMM or perhaps CCCL if it's moved there by then.