Why function called by multiprocessing is not printing the messages?

问题

Why in the example below myFunct() is not printing any messages is supposed to when it is ran by 'multiprocessing'? And how to solve it?

import multiprocessing as mp
poolDict=mp.Manager().dict()

def myFunct(arg):
    print 'myFunct():', arg
    for i in range(110):
        for n in range(500000):
            pass
        poolDict[arg]=i
    print 'myFunct(): completed', arg, poolDict


from multiprocessing import Pool
pool = Pool(processes=2)
myArgsList=['arg1','arg2','arg3']

pool.map_async( myFunct, myArgsList)
print 'completed'

回答1:

This isn't printing because the main process is exiting immediately after calling map_async, so none of the child processes have a chance to actually run. It prints as expected if you make the script wait for the children to finish:

pool.map_async( myFunct, myArgsList)
pool.close()
pool.join()
print 'completed'

Output from bash prompt:

dan@dan:~> ./mult.py 
myFunct(): arg1
myFunct(): arg2
myFunct(): completed arg2 {'arg1': 108, 'arg2': 109}
myFunct(): arg3
myFunct(): completed arg1 {'arg1': 109, 'arg2': 109}
myFunct(): completed arg3 {'arg1': 109, 'arg2': 109, 'arg3': 109}
completed

回答2:

It works fine if you use apply_async() instead of map_async() and add the close() and join() calls:

import multiprocessing as mp
poolDict=mp.Manager().dict()

def myFunct(arg):
    print 'myFunct():', arg
    for i in range(110):
        for n in range(500000):
            pass
        poolDict[arg]=i
    print 'myFunct(): completed', arg, poolDict

from multiprocessing import Pool
pool = Pool(processes=2)
myArgsList=['arg1','arg2','arg3']

pool.apply_async( myFunct, myArgsList)
pool.close()
pool.join()
print 'completed'

Result:

myFunct(): arg1
myFunct(): arg2
myFunct(): completedmyFunct(): completed arg1 {'arg1': 109, 'arg2': 109}
 arg2 {'arg1': 108, 'arg2': 109}
myFunct(): arg3
myFunct(): completed arg3 {'arg1': 109, 'arg2': 109, 'arg3': 109}
completed

来源：https://stackoverflow.com/questions/23961316/why-function-called-by-multiprocessing-is-not-printing-the-messages

标签

python

multiprocessing