zoukankan      html  css  js  c++  java
  • python多线程

    详细学一下python中的多线程。

    首先需要掌握的是进程和线程的概念,

    线程是执行的指令集,是进程的最小执行单位,进程是资源的集合(线程是一个个工人,而进程是一个房子);

    线程共享内存空间,进程的内存是独立的(画图软件不能访问QQ)

    同一个进程的线程之间可以直接交流,两个进程想通信,必须通过一个中间代理来实现;

    创建新线程很简单,创建新进程需要对其父进程进行一次克隆(进程开销大,线程开销小)

    一个线程可以控制和操作同一进程里的其他线程,但是进程只能操作子进程;

    不能直接比较进程和线程的快慢,因为这两个概念不是一个维度的,进程的启动至少需要一个线程;

    python中的多线程

    直接看代码

    import time
    import threading
    
    start_time = time.time()
    t_objs = []
    
    def run(n):
        print('running task ',n ) 
        time.sleep(2)
        print("task done,",n)
        
    for i in range(50):
        t = threading.Thread(target=run,args=("t-%s"%i ,))
        #t.setDaemon(True)#把当前线程变成守护线程
        t.start()
        t_objs.append(t)
        
    for t in t_objs:
        t.join()
        
    print("-"*30+"all threads has finished",threading.current_thread(),threading.activeCount())
    print("cost:",time.time()-start_time,threading.currentThread())

    ###

    running task t-0
    running task t-1
    running task t-2
    running task t-3
    running task t-4
    running task t-5
    running task t-6
    running task t-7
    running task t-8
    running task t-9
    running task t-10
    running task t-11
    running task t-12
    running task t-13
    running task t-14
    running task t-15
    running task t-16
    running task t-17
    running task t-18
    running task t-19
    running task t-20
    running task t-21
    running task t-22
    running task t-23
    running task t-24
    running task t-25
    running task t-26
    running task t-27
    running task t-28
    running task t-29
    running task t-30
    running task t-31
    running task t-32
    running task t-33
    running task t-34
    running task t-35
    running task t-36
    running task t-37
    running task t-38
    running task t-39
    running task t-40
    running task t-41
    running task t-42
    running task t-43
    running task t-44
    running task t-45
    running task t-46
    running task t-47
    running task t-48
    running task t-49
    task done, t-1
    task done, t-2
    task done, t-0
    task done, t-4
    task done, t-5
    task done, t-3
    task done, t-6
    task done, t-7
    task done, t-9
    task done, t-8
    task done, t-10
    task done, t-11
    task done, t-12
    task done, t-13
    task done, t-14
    task done, t-16
    task done, t-15
    task done, t-17
    task done, t-19
    task done, t-18
    task done, t-20
    task done, t-22
    task done, t-21
    task done, t-23
    task done, t-24
    task done, t-25
    task done, t-27
    task done, t-26
    task done, t-29
    task done, t-28
    task done, t-30
    task done, t-32
    task done, t-31
    task done, t-33
    task done, t-34
    task done, t-36
    task done, t-35
    task done, t-37
    task done, t-38
    task done, t-40
    task done, t-39
    task done, t-42
    task done, t-41
    task done, t-44
    task done, t-43
    task done, t-47
    task done, t-46
    task done, t-45
    task done, t-49
    task done, t-48
    ------------------------------all threads has finished <_MainThread(MainThread, started 25376)> 1
    cost: 2.03932785987854 <_MainThread(MainThread, started 25376)>

    每一个run函数都是子线程,t.join是等待每一个子线程执行完毕,接着才会执行之后的主线程(所以采用列表存储线程的方式),也就是两个print函数;如果去掉join,则主线程不会等子线程执行完毕再执行,而是在中间某个时段就执行了;

    接着是守护线程的概念

    守护线程,顾名思义,是为主线程服务的,程序不会等所有守护线程执行完毕再退出(如上去掉join的时候),而是等主线程执行完毕就立刻退出,效果如下:

    import time
    import threading
    
    start_time = time.time()
    t_objs = []
    
    def run(n):
        print('running task ',n ) 
        time.sleep(2)
        print("task done,",n)
        
    for i in range(50):
        t = threading.Thread(target=run,args=("t-%s"%i ,))
        t.setDaemon(True)#把当前线程变成守护线程
        t.start()
        t_objs.append(t)
        
    # for t in t_objs:
    #     t.join()
        
    print("-"*30+"all threads has finished",threading.current_thread(),threading.activeCount())
    print("cost:",time.time()-start_time,threading.currentThread())

    ###

    running task t-0
    running task t-1
    running task t-2
    running task t-3
    running task t-4
    running task t-5
    running task t-6
    running task t-7
    running task t-8
    running task t-9
    running task t-10
    running task t-11
    running task t-12
    running task t-13
    running task t-14
    running task t-15
    running task t-16
    running task t-17
    running task t-18
    running task t-19
    running task t-20
    running task t-21
    running task t-22
    running task t-23
    running task t-24
    running task t-25
    running task t-26
    running task t-27
    running task t-28
    running task t-29
    running task t-30
    running task t-31
    running task t-32
    running task t-33
    running task t-34
    running task t-35
    running task t-36
    running task t-37
    running task t-38
    running task t-39
    running task t-40
    running task t-41
    running task t-42
    running task t-43
    running task t-44
    running task t-45
    running task t-46
    running task t-47
    running task t-48
    running task t-49
    ------------------------------all threads has finished <_MainThread(MainThread, started 26016)> 51
    cost: 0.035875558853149414 <_MainThread(MainThread, started 26016)>

    全局解释器锁GIL

    保证同一时刻只有一个线程在执行

     Last but not least:

    我去NMLGB的中国移动杭研,虽然老子学艺不精,可你那些破烂音响真算不上什么牛逼AI

    人生苦短,何不用python
  • 相关阅读:
    Alien Security (BFS+DFS)
    HDU 1495 喝可乐(暴力BFS)
    Tempter of the Bone(奇偶剪枝)の反面教材
    Fire Game (双起点bfs)
    Linux学习
    c# 实现IComparable、IComparer接口、Comparer类的详解
    DataTable 基本转换简单实例
    C#集合
    递归方法
    存储过程生成复杂的随机编号
  • 原文地址:https://www.cnblogs.com/yqpy/p/9796446.html
Copyright © 2011-2022 走看看