The utils-multiprocessing.ipynb example does not work on Mac OS X without MKL

[ppeder08]

This bug is not caused by magni but rather numpy (or multiprocessing). It is present on Mac OS X for versions of Python and numpy that have not been compiled against MKL. The following piece of code illustrates the problem:

import multiprocessing as mp

import numpy as np


def compute(n):
    print('Enter')
    np.dot(np.eye(n), np.eye(n))
    print('Exit')


print('\nWithout multiprocessing:')
compute(1000)

print('\nWith multiprocessing:')
workers = mp.Pool(1)
results = workers.map(compute, (1000, 1000))

The above code hangs in the workers.map call when executing np.dot(np.eye(n), np.eye(n)). Strangely, there is no problem if the non-multiprocessing call to compute is ommitted, or the value of n is reduced sufficiently.

[ppeder08]

An issue has been submitted upstream (numpy/numpy#4776).