why so much differnt performance in python 2.7.10, 2.7.14 and 3.6.4

[jiangtianyong]

Hi ALL

I have a question when i develop a testcase to run in different python version
1: Python 2.7.10 (default, Feb 7 2017, 00:08:15) [GCC 4.2.1 Compatible Apple LLVM 8.0.0 (clang-800.0.34)] on darwin (here is python)


2: Python 2.7.14 (default, Jan 6 2018, 12:15:00) [GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.39.2)] on darwin (here is python2 in my env)


3:Python 3.6.4 (v3.6.4:d48ecebad5, Dec 18 2017, 21:07:28) [GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin (here is my python3)

the same code run on the three different python version, i take much different time

Here is the code:

#!/usr/bin/env python
#coding:utf-8

'''
test GIL global interpreter lock
'''

from threading import Thread
import time
from multiprocessing import Process
from multiprocessing import Pool
import multiprocessing


def counter():
    i = 0 
    for _ in range(50000000):
        i += 1
    return True

def multiprocessPool_main():
    start = time.time()
    pool = Pool(processes=4)
    for x in range(10):
        pool.apply_async(counter, args=())

    pool.close()
    pool.join()

    end = time.time()
    print('in pool processes take %s seconds' % (end - start))

def multiprocess_main():
    start = time.time()
    p1 = Process(target=counter, args=())
    p2 = Process(target=counter, args=())
    p3 = Process(target=counter, args=())

    p1.start()
    p2.start()
    p3.start()

    print("The number of CPU is:" + str(multiprocessing.cpu_count()))
    for p in multiprocessing.active_children():
        print("child p.name: " + p.name + "\tp.id: " + str(p.pid))

    p1.join()
    p2.join()
    p3.join()

    end = time.time()
    print('3 processes take %s seconds' % (end - start))

def main():
    start_time = time.time()
    for tid in range(2):
        t = Thread(target=counter)
        t.start()
        t.join()
    end_time = time.time()
    print ("total time of single is: {}".format(end_time - start_time))

def multi_main():
    thread_all = []
    start_time = time.time()
    for tid in range(2):
        t = Thread(target=counter)
        t.start()
        thread_all.append(t)
    for i in range(2):
        thread_all[i].join()
    end_time = time.time()
    print ("total time of multi is: {}".format(end_time -start_time))

if __name__ == '__main__':
    main()
    multi_main()
    multiprocess_main()
    multiprocessPool_main()

Output:
bogon:Documents mobike$ python3 GIL.py
total time of single is: 5.48216438293457
total time of multi is: 6.442492961883545
The number of CPU is:8
child p.name: Process-3	p.id: 8492
child p.name: Process-2	p.id: 8491
child p.name: Process-1	p.id: 8490
3 processes take 3.0569851398468018 seconds
in pool processes take 8.76835322380066 seconds
bogon:Documents mobike$ python GIL.py
total time of single is: 4.60499000549
total time of multi is: 6.05071091652
The number of CPU is:8
child p.name: Process-1	p.id: 8499
child p.name: Process-2	p.id: 8500
child p.name: Process-3	p.id: 8501
3 processes take 3.91811919212 seconds
in pool processes take 9.27075195312 seconds
bogon:Documents mobike$ python2 GIL.py
total time of single is: 13.2945930958
total time of multi is: 17.866697073
The number of CPU is:8
child p.name: Process-2	p.id: 8509
child p.name: Process-1	p.id: 8508
child p.name: Process-3	p.id: 8510
3 processes take 8.49618196487 seconds
in pool processes take 25.7352690697 seconds

---

output in python 3.6.4

Output:
bogon:Documents mobike$ python3 GIL.py
total time of single is: 5.48216438293457
total time of multi is: 6.442492961883545
The number of CPU is:8
child p.name: Process-3	p.id: 8492
child p.name: Process-2	p.id: 8491
child p.name: Process-1	p.id: 8490
3 processes take 3.0569851398468018 seconds
in pool processes take 8.76835322380066 seconds

output in python 2.7.10

Output:
bogon:Documents mobike$ python GIL.py
total time of single is: 4.60499000549
total time of multi is: 6.05071091652
The number of CPU is:8
child p.name: Process-1	p.id: 8499
child p.name: Process-2	p.id: 8500
child p.name: Process-3	p.id: 8501
3 processes take 3.91811919212 seconds
in pool processes take 9.27075195312 seconds

output in the python 2.7.14

Output:
bogon:Documents mobike$ python2 GIL.py
total time of single is: 13.2945930958
total time of multi is: 17.866697073
The number of CPU is:8
child p.name: Process-2	p.id: 8509
child p.name: Process-1	p.id: 8508
child p.name: Process-3	p.id: 8510
3 processes take 8.49618196487 seconds
in pool processes take 25.7352690697 seconds