multiprocessing - Python: Using Multiprocess to switch functions -

i'm trying learn multiprocessing in python.

what want happen have x increase time, when hit 'enter' decrease time , can keep hitting enter switch increasing decreasing.

all i've managed make far this

import time multiprocessing import pool  def f(x):     return (x+1) def f1(x):     return (x-1)  if __name__ == '__main__':     pool = pool(processes=2)     x=0     while x<100:         x = pool.apply_async(f, [10])         time.sleep(0.05) 

i'm having trouble looking though documentation because don't give explicit examples. please have no idea

threw example give ideas.

#!/usr/bin/python3 import time multiprocessing import pool, manager threading import thread  num_processes = 4  def consumer(q):     out = []     while true:         val = q.get()         if val none:             #poison pill             break         #doing work here         time.sleep(0.01)         out.append(val)     return out  def producer(queue):     flip = true     val = 5      def flipper():         nonlocal flip         input('enter flip it!')         while true:             flip = not flip             txt = 'up' if flip else 'down'             input('val {}, counting {}'.format(val, txt))      t = thread(target=flipper, args=(), daemon=true)     t.start()      while val > 0:         _ in range(num_processes):             queue.put(val)         val = val + (1 if flip else -1)         time.sleep(0.2)      print()     print('hit zero, shutting down.')      _ in range(num_processes):         #poison pills         queue.put(none)  if __name__ == '__main__':     pool = pool(processes=num_processes)     m = manager() # n.b.: multiprocessing.queue doesn't work pool.apply_async     q = m.queue()      results = pool.apply_async(consumer, args=(q,))     producer(q) # running in main thread because i'm lazy      print(results.get()) 

output:

ben@nixbox:~$ python3 multithread.py  enter flip it! val 13, counting down val 4, counting val 8, counting down hit zero, shutting down. [5, 5, 5, 5, 6, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 8, 9, 9, 9, 9, 10, 10, 10, 10, 11, 11, 11, 11, 12, 12, 12, 12, 13, 13, 13, 13, 12, 12, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10, 9, 9, 9, 9, 8, 8, 8, 8, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 8, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1] 

the way original code set up, parallelizing didn't make whole lot of sense. switched producer/consumer paradigm: producer produces "work units" increase until hit enter, decrease until hits zero, loads work queue poison pills tell consumers shut down. there's few tricks input/flipping thing work, i'll let @ code that. note used threads wait on blocking i/o operations , processes (simulated) cpu-heavy work. on purpose.

there lots of ways structure multiprocessing code find producer/consumer useful pattern. 1 thing don't want doing rapidly creating processes teeny tiny amount of work , tearing them down. slow! (i note using pool correctly avoids - keeps workers alive you. still, worth mentioning since lot of people start out firing off bunch of processes in loop, later wonder why parallel code slower single-threaded code :-))