Python中单线程、多线程和多进程的效率对比实验实例_Python

python的多进程性能要明显优于多线程，因为cpython的gil对性能做了约束。

python是运行在解释器中的语言，查找资料知道，python中有一个全局锁（gil），在使用多进程(thread)的情况下，不能发挥多核的优势。而使用多进程(multiprocess)，则可以发挥多核的优势真正地提高效率。

对比实验

资料显示，如果多线程的进程是cpu密集型的，那多线程并不能有多少效率上的提升，相反还可能会因为线程的频繁切换，导致效率下降，推荐使用多进程；如果是io密集型，多线程进程可以利用io阻塞等待时的空闲时间执行其他线程，提升效率。所以我们根据实验对比不同场景的效率

操作系统	cpu	内存	硬盘
windows 10	双核	8gb	机械硬盘

(1)引入所需要的模块

				?

									import requests

									import time

									from threading import thread

									from multiprocessing import process

(2)定义cpu密集的计算函数

				?

									def count(x, y):

									  # 使程序完成150万计算

									  c = 0

									  while c < 500000:

									    c += 1

									    x += x

									    y += y

（3）定义io密集的文件读写函数

				?

									def write():

									  f = open("test.txt", "w")

									  for x in range(5000000):

									    f.write("testwrite\n")

									  f.close()

									def read():

									  f = open("test.txt", "r")

									  lines = f.readlines()

									  f.close()

(4) 定义网络请求函数

				?

									_head = {

									      'user-agent': 'mozilla/5.0 (windows nt 10.0; wow64) applewebkit/537.36 (khtml, like gecko) chrome/48.0.2564.116 safari/537.36'}

									url = "http://www.tieba.com"

									def http_request():

									  try:

									    webpage = requests.get(url, headers=_head)

									    html = webpage.text

									    return {"context": html}

									  except exception as e:

									    return {"error": e}

(5)测试线性执行io密集操作、cpu密集操作所需时间、网络请求密集型操作所需时间

				?

									# cpu密集操作

									t = time.time()

									for x in range(10):

									  count(1, 1)

									print("line cpu", time.time() - t)

									# io密集操作

									t = time.time()

									for x in range(10):

									  write()

									  read()

									print("line io", time.time() - t)

									# 网络请求密集型操作

									t = time.time()

									for x in range(10):

									  http_request()

									print("line http request", time.time() - t)

输出

cpu密集：95.6059999466、91.57099986076355 92.52800011634827、 99.96799993515015
io密集：24.25、21.76699995994568、21.769999980926514、22.060999870300293
网络请求密集型: 4.519999980926514、8.563999891281128、4.371000051498413、4.522000074386597、14.671000003814697

(6)测试多线程并发执行cpu密集操作所需时间

				?

									counts = []

									t = time.time()

									for x in range(10):

									  thread = thread(target=count, args=(1,1))

									  counts.append(thread)

									  thread.start()

									e = counts.__len__()

									while true:

									  for th in counts:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print(time.time() - t)

output: 99.9240000248 、101.26400017738342、102.32200002670288

(7)测试多线程并发执行io密集操作所需时间

				?

									def io():

									  write()

									  read()

									t = time.time()

									ios = []

									t = time.time()

									for x in range(10):

									  thread = thread(target=count, args=(1,1))

									  ios.append(thread)

									  thread.start()

									e = ios.__len__()

									while true:

									  for th in ios:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print(time.time() - t)

output: 25.69700002670288、24.02400016784668

(8)测试多线程并发执行网络密集操作所需时间

				?

									t = time.time()

									ios = []

									t = time.time()

									for x in range(10):

									  thread = thread(target=http_request)

									  ios.append(thread)

									  thread.start()

									e = ios.__len__()

									while true:

									  for th in ios:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print("thread http request", time.time() - t)

output: 0.7419998645782471、0.3839998245239258、0.3900001049041748

(9)测试多进程并发执行cpu密集操作所需时间

				?

									counts = []

									t = time.time()

									for x in range(10):

									  process = process(target=count, args=(1,1))

									  counts.append(process)

									  process.start()

									e = counts.__len__()

									while true:

									  for th in counts:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print("multiprocess cpu", time.time() - t)

output: 54.342000007629395、53.437999963760376

(10)测试多进程并发执行io密集型操作

				?

									t = time.time()

									ios = []

									t = time.time()

									for x in range(10):

									  process = process(target=io)

									  ios.append(process)

									  process.start()

									e = ios.__len__()

									while true:

									  for th in ios:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print("multiprocess io", time.time() - t)

output: 12.509000062942505、13.059000015258789

(11)测试多进程并发执行http请求密集型操作

				?

									t = time.time()

									httprs = []

									t = time.time()

									for x in range(10):

									  process = process(target=http_request)

									  ios.append(process)

									  process.start()

									e = httprs.__len__()

									while true:

									  for th in httprs:

									    if not th.is_alive():

									      e -= 1

									  if e <= 0:

									    break

									print("multiprocess http request", time.time() - t)

output: 0.5329999923706055、0.4760000705718994

实验结果

cpu密集型操作	io密集型操作	网络请求密集型操作
线性操作	94.91824996469	22.46199995279	7.3296000004
多线程操作	101.1700000762	24.8605000973	0.5053332647
多进程操作	53.8899999857	12.7840000391	0.5045000315