shared between processes multi-dimensional arrays for Python 3

Question

As far as I understand sharedmem-numpy package (numpy-sharedmem) is only for Python 2. Does it exist something similar (shared sort of numpy arrays) for Python 3?
The purpose is to write to one multi-dimensional array simultaneously by several processes. By default, it is known that every process will write only in its own part of array, so there will be no any conflicts.

Answer 1

Have a look at Blaze and Dask, developed as a successor to Numpy.

Multi-core NumPy arrays can be achieved with Dask.array.