Asynchronous memory transfer

[LWisteria]

Currently, all async memory operation are implemented as sync.

• clpy/clpy/backend/memory.pyx

  Lines 300 to 301 in 3fd3b23

  	# TODO(LWisteria): Impelment async
  	self.copy_from_device(src, size)

• clpy/clpy/backend/memory.pyx

  Lines 335 to 336 in 3fd3b23

  	# TODO(LWisteria): Impelment async
  	self.copy_from_host(mem, size)

• clpy/clpy/backend/memory.pyx

  Lines 407 to 408 in 3fd3b23

  	# TODO(LWisteria): Impelment async
  	self.copy_to_host(mem, size)

• clpy/clpy/backend/memory.pyx

  Lines 432 to 433 in 3fd3b23

  	# TODO(LWisteria): Impelment async
  	self.memset(value, size)

They are should be implemented as async.

[ybsh]

@LWisteria When the source and destination of copy_from_device_async() are both device memory objects (cl_mem), should it raise an error, or execute EnqueueCopyBuffer just as in copy_from_device()?

[LWisteria]

@ybsh I understand that your question is about clpy itself but not about this issue. If you suggest raising an error on copy_from_device() (and also on _async), please make an issue for it.

[ybsh]

@LWisteria
Thank you for your advice, and sorry that was off-topic.
Aside from the error handling issue above, I think it is necessary to migrate CuPy's Event/Stream classes first, as those async family take a CUDA stream as one of their arguments.
Do you agree? If you do, could I open an issue for that?

[ybsh]

Regarding the migration of CUDA Stream, @LWisteria told me an issue for that has already been created: #5

[LWisteria]

This issue would be pending because _async functions are not used in cupy tests and even in chainer.