如何在Python 2.7中并行调用不同实例化的类方法？

Question

I have a class which contains functions that perform calculations. If I have several instances of these objects, how do I make the calculations to go in parallel?

class SomeClass:
    def __init__(self, arg):
    ....
    def compute(self):
    ....

And then in a different script:

from multiprocessing import Pool
from wherever import SomeClass
    g1 = SomeClass(arg1)
    g2 = SomeClass(arg2)
    pool = Pool(processes = 2)

How do I make one of the workers do g1.compute() and the other g2.compute()?

Answer 1

On python2.7 you'll have to define a worker function that calls the compute method on the given argument, then you can use it with pool.map() :

def call_compute(o):
    return o.compute()
...
result = pool.map(call_compute, [g1, g2])

On python3 (tested on 3.5) it's possible to use pool.map(SomeClass.comute, [g1, g2]) instead because pickling instance methods is supported there.