.. currentmodule:: playdoh

.. _example-map_shared:

Example: map_shared
===================

Example usage of the :func:`map` function with shared data.

::

    from playdoh import *
    from numpy.random import rand
    
    
    # The function to parallelize. The extra argument ``shared_data`` is a
    # read-only dictionary
    # residing is shared memory on the computer. It can contain large NumPy arrays
    # used by
    # all the CPUs to execute the function.
    def fun(x, shared_data):
        return x + shared_data['x0']
    
    
    # This line is required on Windows, any call to a Playdoh function
    # must be done after this line on this OS.
    # See http://docs.python.org/library/multiprocessing.html#windows
    if __name__ == '__main__':
        # Execute two function evaluations with a large NumPy array in shared data.
        map(fun,
            [rand(100000, 2), rand(100000, 2)],
            cpu=2,
            shared_data={'x0': rand(100000, 2)})