5

I'm fairly new to python programming and need some help understanding the python interpreter flow, especially in the case of multiprocessing. Please note that I'm running python 3.7.1 on Windows 10. Here is my simple experimental code and the output.

import multiprocessing import time def calc_square(numbers, q): for n in numbers: q.put(n*n) time.sleep(0.2) q.put(-1) print('Exiting function') print('Now in the main code. Process name is: ' + __name__) if __name__ == '__main__': numbers = [2, 3, 4, 5] q = multiprocessing.Queue() p = multiprocessing.Process(target=calc_square, args=(numbers, q)) p.start() while True: if q.empty() is False: nq = q.get() print('Message is:' + str(nq)) if nq == -1: break print('Done') p.join() Program Output: Now in the main code. Process name is: __main__ Now in the main code. Process name is: __mp_main__ Message is:4 Message is:9 Message is:16 Message is:25 Exiting function Message is:-1 Done Process finished with exit code 0

I've seen code examples that do not contain the if name qualifier. I first tried running one of those and got several errors that pointed to trying to trying to start a process before process init had completed. According to the author of that code, it works on linux because it forks, while Windows does not (I learned later that Windows uses spawn as the process start method.).

My question centers around why the process seems to execute instructions that are outside of the target. I was surprised to see the 2nd line of the output which printed before the target sent a message to the parent. Why does the sub-process interpreter run that code that is outside of the target? Is it really running that code before the target code? Why?

Improve this question
0

1 Answer 1

Reset to default
11

On Windows, the new processes import the main script, which executes it. This is why, on Windows, you need if __name__ == '__main__': ... it prevents code under that if from running in every process.

Your Now in main code. line is outside the if, so it will run in every process. Move it inside the if:

import multiprocessing import time def calc_square(numbers, q): for n in numbers: q.put(n*n) time.sleep(0.2) q.put(-1) print('Exiting function') if __name__ == '__main__': print('Now in the main code. Process name is:', __name__) numbers = [2, 3, 4, 5] q = multiprocessing.Queue() p = multiprocessing.Process(target=calc_square, args=(numbers, q)) p.start() while True: nq = q.get() print('Message is:', nq) if nq == -1: break print('Done') p.join()
Improve this answer
Sign up to request clarification or add additional context in comments.

2 Comments

Thank you for your explanation. I had the "in the main code " print outside of the "if" intentionally to confirm that the code was actually being executed by the child process interpreter, and when (before the target was executed). I see that with the dynamic nature of python, that a fresh interpreter would need to execute the code before the target would be in the correct environment.
What if it's not possible to use the name clause? For example, what if I want to call a function in multiprocessing mode from a class method?