Python: Why 2 calls of the same function don't run in parallel?

Question

I'm trying to call the same function with different parameters in parallel using multiprocessing module.

Here is example of my code:

from django.apps import AppConfig

class CurrencyRatesConfig(AppConfig):
    name = 'currency_rates'

    def ready(self):
        import time
        from multiprocessing import Process

        import currency_rates.ws_get_rates

        p1 = Process(currency_rates.ws_get_rates.get_rates('1m', 'tBTCUSD', 'BTC2USD'))
        p1.start()
        p2 = Process(currency_rates.ws_get_rates.get_rates('1m', 'tIOTUSD', 'IOT2USD'))
        p2.start()

        p1.join()
        p2.join()

But it doesn't work. I see just p1's execution result.

NOTE: get_rates() uses websocket module.

Could somebody describe what is wrong?

Thanks.

I believe p1.join() will block until p1 returns. Then p2 would resume. — Ron Norris, Oct 18 '17 at 12:23
have a look at this answer https://stackoverflow.com/a/19138327/5476782 if you don't need join(), remove it from your code, it blocks the thread in execution — Evhz, Oct 18 '17 at 12:25
@Karlos there's no code after the joins, nothing is being blocked from running. There was an answer here a minute ago that seems plausible about the work carried out by get_rates which is called in the current process and is therefore not parallel — jonatan, Oct 18 '17 at 12:28
the code after doesn't matter. With join you block the thread which lauched the p2, so no answer from p2. Have a look at https://pymotw.com/2/multiprocessing/basics.html — Evhz, Oct 18 '17 at 12:30

Ron Norris · Accepted Answer · 2017-10-18T12:52:59.807

The problem is with the Process arguments. Modify it as follows:

This demo does work. Not sure why the specific issue is not resolved for OP.

from multiprocessing import Process
import time
import sys

def outputString(thestring):
    for i in range(3):
        print(thestring)
        time.sleep(2)
    return

def main():
    p1 = Process(target=outputString, args=('P1 is processing',))
    p1.start()
    p2 = Process(target=outputString, args=('P2 is processing',))
    p2.start()

if __name__ == "__main__":
    main()

OP original code (modified)

from django.apps import AppConfig

class CurrencyRatesConfig(AppConfig):
    name = 'currency_rates'

    def ready(self):
        import time
        from multiprocessing import Process

        import currency_rates.ws_get_rates

        p1 = Process(target=currency_rates.ws_get_rates.get_rates, args=('1m', 'tBTCUSD', 'BTC2USD'))
        p1.start()
        p2 = Process(target=currency_rates.ws_get_rates.get_rates, args=('1m', 'tIOTUSD', 'IOT2USD'))
        p2.start()

Doesn't work. Probably because I'm using `websocket` module in `get_rates()` — Raman, Oct 18 '17 at 12:40
Meaning it still blocks? Not sure how unless all resources on the machine are consumed by `get_rates()` because they run in completely different process space. But you can't have the join() in their, or it will not work in parallel. — Ron Norris, Oct 18 '17 at 12:50

Python: Why 2 calls of the same function don't run in parallel?

1 Answers1