Why is multiprocessing module not producing the desired result?

Question

import multiprocessing as mp
import os

def cube(num):
    print(os.getpid())
    print("Cube is {}".format(num*num*num))

def square(num):
    print(os.getpid())
    print("Square is {}".format(num*num))

if __name__ == "__main__":
    p1 = mp.Process(target = cube, args = (3,))
    p2 = mp.Process(target = square, args = (4,))
    p1.start()
    p2.start()
    p1.join()
    p2.join()
    print("Done")

I was using the multiprocessing module, but I am not able to print any output from a function using that.

I even tried flushing the stdout using the sys module.

`if _name_ == "_main_"` There should be _two_ underscores before and after `name` and `main`, not just one. — John Gordon, May 28 '20 at 14:29
`_name_` should have produced an error, as no such variable exists. That tells me this isn't your actual code. Please post your real code. — John Gordon, May 28 '20 at 15:20
This works on my Linux machine with Python 3.8.3. Maybe see https://stackoverflow.com/questions/21198857/python-multiprocessing-example-not-working — Alex, May 28 '20 at 18:11

user3666197 · Accepted Answer · 2020-05-29T18:03:13.320

Q : "Why is multiprocessing module not producing the desired result?"

Why?

Because it crashes.

The MWE/MCVE-representation of the problem has a wrong code. It crashes & it has nothing to do with the sys.stdout.flush() :

>>> cube(  4 )
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 2, in cube
NameError: global name 'os' is not defined

Solution :

>>> import os   # be it in the __main__ or in the def()-ed functions...

>>> cube(  4 )
14165
Cube is 64

and your mp.Process()-based replicas of the python-process instances will stop crashing too.

MCVE that works :

(base) Fri May 29 14:29:33 $ conda activate py3
(py3) Fri May 29 14:34:55 $ python StackOverflow_mp.py
This is ____6745::__main__
This is ____6746::PID
This is ____6747::PID
Cube(__3) is _______27.
Square(__4) is _______16.
Done.
Works.
Q.E.D.

import multiprocessing as mp
import os
import sys
import time

def cube( num ):
    print( "This is {0:_>8d}::PID".format( os.getpid() ) )
    print( "Cube({0:_>3d}) is {1:_>9d}.".format( num, num*num*num ) )
    sys.stdout.flush()

def square( num ):
    print( "This is {0:_>8d}::PID".format( os.getpid() ) )
    print( "Square({0:_>3d}) is {1:_>9d}.".format( num, num*num ) )
    sys.stdout.flush()

if __name__ == "__main__":
    print( "This is {0:_>8d}::__main__".format( os.getpid() ) )
    p1 = mp.Process( target = cube,   args = (3, ) )
    p2 = mp.Process( target = square, args = (4, ) )
    p1.start()
    p2.start()
    p1.join()
    p2.join()

    time.sleep( 1 )

    print( "Done.\nWorks.\nQ.E.D." )

_{I copied and pasted your exact code. But I still didn't get the output from the called functions using the multiprocessing libraries
– Kartikeya Agarwal 47 mins ago}

So,
- I opened a new Terminal process,
- I copied the conda activate py3 command and
- I hit Enter to let it run, so as to make python3 ecosystem go live.
- I re-launched the proof-of-solution again python StackOverflow_mp.py and
- I hit Enter to let it run
- I saw it working the very same way as it worked last time.
- I doubt the problem is on the provided twice (re)-validated proof-of-solution side, is it?
Q.E.D.

(py3) Fri May 29 19:53:58 $ python StackOverflow_mp.py
This is ___27202::__main__
This is ___27203::PID
Cube(__3) is _______27.
This is ___27204::PID
Square(__4) is _______16.
Done

Sir, it is working without the multiprocessing libraries but as soon as I try to divide the work between the CPU's. it stops giving the o/p. I am using JupiterNotebook to interpret python if it has anything to do with it — Kartikeya Agarwal, May 29 '20 at 10:16
I copied and pasted your exact code. But I still didn't get the output from the called functions using the multiprocessing libraries — Kartikeya Agarwal, May 29 '20 at 17:03

Why is multiprocessing module not producing the desired result?

1 Answers1

Solution :

MCVE that works :