We don't need to actually call the executable. We can implement this in Cython and call __cu_demangle:
https://docs.nvidia.com/cuda/cuda-binary-utilities/index.html#library-availability
TBD: Should we place this in cuda.bindings.utils or cuda.core.utils?
We don't need to actually call the executable. We can implement this in Cython and call
__cu_demangle:https://docs.nvidia.com/cuda/cuda-binary-utilities/index.html#library-availability
TBD: Should we place this in
cuda.bindings.utilsorcuda.core.utils?