python多处理和核心数

Question

I have a mac , and it has 2 physical cores and 4 logical cores. I am trying to figure out the python multiprocessing . As far as I know, each process in the processing uses one core in the computer, so I believe the number of processes in a python script can not exceed the number of the core of computer. Is this correct ?

But if I execute the following code (it has 8 functions ) , it runs same time with the version that has 4 functions.

from multiprocessing import Process
import time 
def func1():
  for i in range(0,500):
    print "11111"
    time.sleep(0.1)


def func2():
  for i in range(500,1000):
    print "22222"
    time.sleep(0.1)

def func3():
  for i in range(1000,1500):
    print "33333"
    time.sleep(0.1)


def func4():
   for i in range(1500,2000):
    print "444444"
    time.sleep(0.1)

def func5():
   for i in range(2000,2500):
    print "555555"
    time.sleep(0.1)


def func6():
   for i in range(2500,3000):
    print "666666666"
    time.sleep(0.1)


def func7():
  for i in range(3500,4000):
    print "7777777777"
    time.sleep(0.1)

def func8():
  for i in range(4500,5000):
    print "8888888888 "
    time.sleep(0.1)

if __name__=='__main__':
  start_time = time.time()
  p1 = Process(target = func1)
  p1.start()
  p2 = Process(target = func2)
  p2.start()
  p3 = Process(target = func3)
  p3.start()
  p4 = Process(target = func4)
  p4.start()
  p5 = Process(target = func5)
  p5.start()
  p6 = Process(target = func6)
  p6.start()
  p7 = Process(target = func7)
  p7.start()
  p8 = Process(target = func8)
  p8.start()
  p1.join()
  p2.join()
  p3.join()
  p4.join()
  p5.join()
  p6.join()
  p7.join()
  p8.join()
  print "done"
  print("--- %s seconds ---" % (time.time() - start_time))

If I run this code with 4 functions or 8 functions, the execution time is same, but the number of cores is 4, so I think the 8 functions version should take longer, but it did not. I am missing something here ?

Thank you !

Answer 1

You can start as many processes from your Python script as you like and the operating system scheduler will schedule them to run on your CPU cores. If those processes are CPU bound, then you won't get better performance by starting more processes than cores. But in your example, the processes are spending most of their time sleeping, and therefore they aren't competing for the cores, which is why 8 processes can run in the same time as 4.