Tested module of multiprocessing at python 3.3.0. My mac is, * OS X 10.8.3 * Python 3.3.0 * 2.3 GHz Intel Core i7 * 16GB 1600 MHz DDR3

mlt_proc_test.py

import numpy as np
import multiprocessing as mlp
import time

def main(X):
    print("Data Length\t\t\t:", len(X))

    spd = time.clock()
    calc_map(X)
    spd = time.clock() - spd
    print("Multiprocess Map\t:", spd)

    spd = time.clock()
    calc_imap(X)
    spd = time.clock() - spd
    print("Multiprocess iMap\t:", spd)

    spd = time.clock()
    calc_apply(X)
    spd = time.clock() - spd
    print("Multiprocess Apply \t:", spd)

    spd = time.clock()
    calc_comp(X)
    spd = time.clock() - spd
    print("List Comprehensions\t:", spd)

def calc_map(X):
    pool = mlp.Pool()
    result = pool.map(func=calc_power, iterable=X, chunksize=3000)
    return result

def calc_imap(X):
    pool = mlp.Pool()
    result = [r for r in pool.imap(func=calc_power, iterable=X, chunksize=3000)]
    return result

def calc_apply(X):
    pool = mlp.Pool()
    result = pool.apply(func=calc_power, args=(X,))
    return result

def calc_comp(X):
    return [calc_power(X[t]) for t in range(len(X))]

def calc_power(X):
    return X ** 2

if __name__ == '__main__':
    print("CPUs\t\t\t\t:", mlp.cpu_count())
    main(np.random.rand(10))
    main(np.random.rand(100))
    main(np.random.rand(1000))
    main(np.random.rand(10000))
    main(np.random.rand(100000))
    main(np.random.rand(1000000))
    main(np.random.rand(10000000))

result

CPUs                : 8
Data Length         : 10
Multiprocess Map    : 0.027174000000000004
Multiprocess iMap   : 0.01109199999999999
Multiprocess Apply  : 0.011204999999999965
List Comprehensions : 9.100000000000774e-05

Data Length         : 100
Multiprocess Map    : 0.015120999999999996
Multiprocess iMap   : 0.01593100000000003
Multiprocess Apply  : 0.013925999999999994
List Comprehensions : 0.00022499999999997522

Data Length         : 1000
Multiprocess Map    : 0.025980000000000003
Multiprocess iMap   : 0.021921999999999997
Multiprocess Apply  : 0.018123
List Comprehensions : 0.001831000000000027

Data Length         : 10000
Multiprocess Map    : 0.087561
Multiprocess iMap   : 0.09058100000000008
Multiprocess Apply  : 0.019840999999999998
List Comprehensions : 0.017249999999999988

Data Length         : 100000
Multiprocess Map    : 0.705835
Multiprocess iMap   : 0.7403
Multiprocess Apply  : 0.03594000000000008
List Comprehensions : 0.17743100000000034

Data Length         : 1000000
Multiprocess Map    : 6.757089
Multiprocess iMap   : 6.983575000000002
Multiprocess Apply  : 0.19784699999999944
List Comprehensions : 1.8046490000000013

Data Length         : 10000000
Multiprocess Map    : 67.830874
Multiprocess iMap   : 70.555025
Multiprocess Apply  : 13.619205999999991
List Comprehensions : 17.82403400000001