服务器之家

服务器之家 > 正文

Python多进程multiprocessing用法实例分析

时间:2020-12-02 00:53     来源/作者:水·域

本文实例讲述了Python多进程multiprocessing用法。分享给大家供大家参考,具体如下:

mutilprocess简介

像线程一样管理进程,这个是mutilprocess的核心,他与threading很是相像,对多核CPU的利用率会比threading好的多。

简单的创建进程:

?
1
2
3
4
5
6
7
8
9
10
11
import multiprocessing
def worker(num):
  """thread worker function"""
  print 'Worker:', num
  return
if __name__ == '__main__':
  jobs = []
  for i in range(5):
    p = multiprocessing.Process(target=worker, args=(i,))
    jobs.append(p)
    p.start()

确定当前的进程,即是给进程命名,方便标识区分,跟踪

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
import multiprocessing
import time
def worker():
  name = multiprocessing.current_process().name
  print name, 'Starting'
  time.sleep(2)
  print name, 'Exiting'
def my_service():
  name = multiprocessing.current_process().name
  print name, 'Starting'
  time.sleep(3)
  print name, 'Exiting'
if __name__ == '__main__':
  service = multiprocessing.Process(name='my_service',
                   target=my_service)
  worker_1 = multiprocessing.Process(name='worker 1',
                    target=worker)
  worker_2 = multiprocessing.Process(target=worker) # default name
  worker_1.start()
  worker_2.start()
  service.start()

守护进程就是不阻挡主程序退出,自己干自己的 mutilprocess.setDaemon(True)就这句等待守护进程退出,要加上join,join可以传入浮点数值,等待n久就不等了

守护进程:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
import multiprocessing
import time
import sys
def daemon():
  name = multiprocessing.current_process().name
  print 'Starting:', name
  time.sleep(2)
  print 'Exiting :', name
def non_daemon():
  name = multiprocessing.current_process().name
  print 'Starting:', name
  print 'Exiting :', name
if __name__ == '__main__':
  d = multiprocessing.Process(name='daemon',
                target=daemon)
  d.daemon = True
  n = multiprocessing.Process(name='non-daemon',
                target=non_daemon)
  n.daemon = False
  d.start()
  n.start()
  d.join(1)
  print 'd.is_alive()', d.is_alive()
  n.join()

最好使用 poison pill,强制的使用terminate()注意 terminate之后要join,使其可以更新状态

终止进程:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
import multiprocessing
import time
def slow_worker():
  print 'Starting worker'
  time.sleep(0.1)
  print 'Finished worker'
if __name__ == '__main__':
  p = multiprocessing.Process(target=slow_worker)
  print 'BEFORE:', p, p.is_alive()
  p.start()
  print 'DURING:', p, p.is_alive()
  p.terminate()
  print 'TERMINATED:', p, p.is_alive()
  p.join()
  print 'JOINED:', p, p.is_alive()

①. == 0 未生成任何错误 
②. 0 进程有一个错误,并以该错误码退出
③. < 0 进程由一个-1 * exitcode信号结束

进程的退出状态:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
import multiprocessing
import sys
import time
def exit_error():
  sys.exit(1)
def exit_ok():
  return
def return_value():
  return 1
def raises():
  raise RuntimeError('There was an error!')
def terminated():
  time.sleep(3)
if __name__ == '__main__':
  jobs = []
  for f in [exit_error, exit_ok, return_value, raises, terminated]:
    print 'Starting process for', f.func_name
    j = multiprocessing.Process(target=f, name=f.func_name)
    jobs.append(j)
    j.start()
  jobs[-1].terminate()
  for j in jobs:
    j.join()
    print '%15s.exitcode = %s' % (j.name, j.exitcode)

方便的调试,可以用logging

日志:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
import multiprocessing
import logging
import sys
def worker():
  print 'Doing some work'
  sys.stdout.flush()
if __name__ == '__main__':
  multiprocessing.log_to_stderr()
  logger = multiprocessing.get_logger()
  logger.setLevel(logging.INFO)
  p = multiprocessing.Process(target=worker)
  p.start()
  p.join()

利用class来创建进程,定制子类

派生进程:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
import multiprocessing
class Worker(multiprocessing.Process):
  def run(self):
    print 'In %s' % self.name
    return
if __name__ == '__main__':
  jobs = []
  for i in range(5):
    p = Worker()
    jobs.append(p)
    p.start()
  for j in jobs:
    j.join()

python进程间传递消息:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
import multiprocessing
class MyFancyClass(object):
  def __init__(self, name):
    self.name = name
  def do_something(self):
    proc_name = multiprocessing.current_process().name
    print 'Doing something fancy in %s for %s!' % \
      (proc_name, self.name)
def worker(q):
  obj = q.get()
  obj.do_something()
if __name__ == '__main__':
  queue = multiprocessing.Queue()
  p = multiprocessing.Process(target=worker, args=(queue,))
  p.start()
  queue.put(MyFancyClass('Fancy Dan'))
  # Wait for the worker to finish
  queue.close()
  queue.join_thread()
  p.join()
import multiprocessing
import time
class Consumer(multiprocessing.Process):
  def __init__(self, task_queue, result_queue):
    multiprocessing.Process.__init__(self)
    self.task_queue = task_queue
    self.result_queue = result_queue
  def run(self):
    proc_name = self.name
    while True:
      next_task = self.task_queue.get()
      if next_task is None:
        # Poison pill means shutdown
        print '%s: Exiting' % proc_name
        self.task_queue.task_done()
        break
      print '%s: %s' % (proc_name, next_task)
      answer = next_task()
      self.task_queue.task_done()
      self.result_queue.put(answer)
    return
class Task(object):
  def __init__(self, a, b):
    self.a = a
    self.b = b
  def __call__(self):
    time.sleep(0.1) # pretend to take some time to do the work
    return '%s * %s = %s' % (self.a, self.b, self.a * self.b)
  def __str__(self):
    return '%s * %s' % (self.a, self.b)
if __name__ == '__main__':
  # Establish communication queues
  tasks = multiprocessing.JoinableQueue()
  results = multiprocessing.Queue()
  # Start consumers
  num_consumers = multiprocessing.cpu_count() * 2
  print 'Creating %d consumers' % num_consumers
  consumers = [ Consumer(tasks, results)
         for i in xrange(num_consumers) ]
  for w in consumers:
    w.start()
  # Enqueue jobs
  num_jobs = 10
  for i in xrange(num_jobs):
    tasks.put(Task(i, i))
  # Add a poison pill for each consumer
  for i in xrange(num_consumers):
    tasks.put(None)
  # Wait for all of the tasks to finish
  tasks.join()
  # Start printing results
  while num_jobs:
    result = results.get()
    print 'Result:', result
    num_jobs -= 1

Event提供一种简单的方法,可以在进程间传递状态信息。事件可以切换设置和未设置状态。通过使用一个可选的超时值,时间对象的用户可以等待其状态从未设置变为设置。

进程间信号传递:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
import multiprocessing
import time
def wait_for_event(e):
  """Wait for the event to be set before doing anything"""
  print 'wait_for_event: starting'
  e.wait()
  print 'wait_for_event: e.is_set()->', e.is_set()
def wait_for_event_timeout(e, t):
  """Wait t seconds and then timeout"""
  print 'wait_for_event_timeout: starting'
  e.wait(t)
  print 'wait_for_event_timeout: e.is_set()->', e.is_set()
if __name__ == '__main__':
  e = multiprocessing.Event()
  w1 = multiprocessing.Process(name='block',
                 target=wait_for_event,
                 args=(e,))
  w1.start()
  w2 = multiprocessing.Process(name='nonblock',
                 target=wait_for_event_timeout,
                 args=(e, 2))
  w2.start()
  print 'main: waiting before calling Event.set()'
  time.sleep(3)
  e.set()
  print 'main: event is set'

Python多进程,一般的情况是Queue来传递。

Queue:

?
1
2
3
4
5
6
7
8
9
from multiprocessing import Process, Queue
def f(q):
  q.put([42, None, 'hello'])
if __name__ == '__main__':
  q = Queue()
  p = Process(target=f, args=(q,))
  p.start()
  print q.get()  # prints "[42, None, 'hello']"
  p.join()

多线程优先队列Queue:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
import Queue
import threading
import time
exitFlag = 0
class myThread (threading.Thread):
  def __init__(self, threadID, name, q):
    threading.Thread.__init__(self)
    self.threadID = threadID
    self.name = name
    self.q = q
  def run(self):
    print "Starting " + self.name
    process_data(self.name, self.q)
    print "Exiting " + self.name
def process_data(threadName, q):
  while not exitFlag:
    queueLock.acquire()
    if not workQueue.empty():
      data = q.get()
      queueLock.release()
      print "%s processing %s" % (threadName, data)
    else:
      queueLock.release()
    time.sleep(1)
threadList = ["Thread-1", "Thread-2", "Thread-3"]
nameList = ["One", "Two", "Three", "Four", "Five"]
queueLock = threading.Lock()
workQueue = Queue.Queue(10)
threads = []
threadID = 1
# Create new threads
for tName in threadList:
  thread = myThread(threadID, tName, workQueue)
  thread.start()
  threads.append(thread)
  threadID += 1
# Fill the queue
queueLock.acquire()
for word in nameList:
  workQueue.put(word)
queueLock.release()
# Wait for queue to empty
while not workQueue.empty():
  pass
# Notify threads it's time to exit
exitFlag = 1
# Wait for all threads to complete
for t in threads:
  t.join()
print "Exiting Main Thread"

多进程使用Queue通信的例子

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
import time
from multiprocessing import Process,Queue
MSG_QUEUE = Queue(5)
def startA(msgQueue):
  while True:
    if msgQueue.empty() > 0:
      print ('queue is empty %d' % (msgQueue.qsize()))
    else:
      msg = msgQueue.get()
      print( 'get msg %s' % (msg,))
    time.sleep(1)
def startB(msgQueue):
  while True:
    msgQueue.put('hello world')
    print( 'put hello world queue size is %d' % (msgQueue.qsize(),))
    time.sleep(3)
if __name__ == '__main__':
  processA = Process(target=startA,args=(MSG_QUEUE,))
  processB = Process(target=startB,args=(MSG_QUEUE,))
  processA.start()
  print( 'processA start..')

主进程定义了一个Queue类型的变量,并作为Process的args参数传给子进程processA和processB,两个进程一个向队列中写数据,一个读数据。

希望本文所述对大家Python程序设计有所帮助。

原文链接:http://www.cnblogs.com/IPYQ/p/5573628.html

相关文章

热门资讯

2020微信伤感网名听哭了 让对方看到心疼的伤感网名大全
2020微信伤感网名听哭了 让对方看到心疼的伤感网名大全 2019-12-26
Intellij idea2020永久破解,亲测可用!!!
Intellij idea2020永久破解,亲测可用!!! 2020-07-29
歪歪漫画vip账号共享2020_yy漫画免费账号密码共享
歪歪漫画vip账号共享2020_yy漫画免费账号密码共享 2020-04-07
电视剧《琉璃》全集在线观看 琉璃美人煞1-59集免费观看地址
电视剧《琉璃》全集在线观看 琉璃美人煞1-59集免费观看地址 2020-08-12
最新idea2020注册码永久激活(激活到2100年)
最新idea2020注册码永久激活(激活到2100年) 2020-07-29
返回顶部