[Python] Multiprocessing on python

I have a task which takes about 10 seconds to run on a single core. I had the idea of using parallelization to improve the performance (if I can run 4 instances of the task in ten seconds I will be able to make some smarter decisions).

So finally I write the following code:

from multiprocessing import Pool, Process, Queue
from vrp import task
import time

def task_paralela(queue, parameter):
ret = task(parameter)
queue.put(ret)

if __name__ == '__main__':
start = time.time()
cores = 4
parameter = 1
processes = []
q = Queue()
rets = []
for i in range(cores):
p = Process(target=task_pararela, args=(q, parameter))
processes.append(p)
p.start()
for p in processes:
ret = q.get()
rets.append(ret)
for p in processes:
p.join()
end = time.time()
total_time = end - start
print(total_time)


And it works, but the time is now near 20 seconds (twice as much). Why is that? Where is my mistake?

Continue reading...