Ref: Python3 多線程html
Ref: Python3之多進程 # python中的多線程沒法利用多核優點python
更多的提升效率的策略,請參見:[Pandas] 01 - A guy based on NumPygit
線程在執行過程當中與進程仍是有區別的。 1. 每一個獨立的線程有一個程序運行的入口、順序執行序列和程序的出口。 2. 可是線程不可以獨立執行,必須依存在應用程序中,由應用程序提供多個線程執行控制。 3. 每一個線程都有他本身的一組CPU寄存器,稱爲線程的上下文,該上下文反映了線程上次運行該線程的CPU寄存器的狀態。 4. 指令指針 和 堆棧指針寄存器 是線程上下文中兩個最重要的寄存器,線程老是在進程獲得上下文中運行的,這些地址都用於標誌擁有線程的進程地址空間中的內存。
Ref: https://github.com/giampaolo/psutilgithub
from multiprocessing import cpu_count
print(cpu_count())
Python3 經過兩個標準庫 _thread 和 threading 提供對線程的支持。安全
_thread 提供了低級別的、原始的線程以及一個簡單的鎖,它相比於 threading 模塊的功能仍是比較有限的。多線程
提供了低級別,原始的線程以及一個簡單的鎖。app
#!/usr/bin/python3 import _thread import time # 爲線程定義一個函數 def print_time( threadName, delay): count = 0 while count < 5: time.sleep(delay) count += 1 print ("%s: %s" % ( threadName, time.ctime(time.time()) ))
----------------------------------------------------------------
# 建立兩個線程,參數是:函數名 以及對應的參數 try: _thread.start_new_thread( print_time, ("Thread-1", 2, ) ) _thread.start_new_thread( print_time, ("Thread-2", 4, ) ) except: print ("Error: 沒法啓動線程")
# 讓主線程不要提早結束 while 1: pass
採用了線程類的手法,該方法比較 engineering。dom
#!/usr/bin/python3 import threading import time exitFlag = 0
# 線程類 class myThread (threading.Thread):
def __init__(self, threadID, name, counter): threading.Thread.__init__(self) self.threadID = threadID self.name = name self.counter = counter
def run(self): print ("開始線程:" + self.name) print_time(self.name, self.counter, 5) print ("退出線程:" + self.name)
----------------------------------------------------------------
def print_time(threadName, delay, counter): while counter: if exitFlag: threadName.exit() time.sleep(delay) print ("%s: %s" % (threadName, time.ctime(time.time()))) counter -= 1
----------------------------------------------------------------
# (1) 建立新 線程'類‘ thread1 = myThread(1, "Thread-1", 1) thread2 = myThread(2, "Thread-2", 2) # (2) 啓動新線程 thread1.start() thread2.start()
# (3) 等待全部線程結束 thread1.join() thread2.join() print ("退出主線程")
使用 Thread 對象的 Lock 和 Rlock 能夠實現簡單的線程同步,這兩個對象都有 acquire 方法和 release 方法;ide
對於那些須要每次只容許一個線程操做的數據,能夠將其操做放到 acquire 和 release 方法之間。函數
#!/usr/bin/python3 import threading import time class myThread (threading.Thread):
def __init__(self, threadID, name, counter): threading.Thread.__init__(self) self.threadID = threadID self.name = name self.counter = counter
def run(self): print ("開啓線程: " + self.name)
------------------------------------------------------- threadLock.acquire() # <---- print_time(self.name, self.counter, 3)
threadLock.release() # <---- -------------------------------------------------------
# 做爲線程共享資源 def print_time(threadName, delay, counter): while counter: time.sleep(delay) print ("%s: %s" % (threadName, time.ctime(time.time()))) counter -= 1
threadLock= threading.Lock() threads = []
# (1) 建立新線程 thread1 = myThread(1, "Thread-1", 1) thread2 = myThread(2, "Thread-2", 2) # (2) 開啓新線程 thread1.start() thread2.start() # (3) 等待線程 threads.append(thread1) threads.append(thread2) for t in threads: t.join() print ("退出主線程")
不添加setDaemon時,主線程和子線程分別在執行,約在主線程執行完5秒後子線程也執行完畢。
添加setDaemon的話,主進程執行完後不會等待 「做爲守護線程」 的子進程,以下代碼中,不會給child thread留有運行的機會。
import threading import time from datetime import datetime class MyThread(threading.Thread): def __init__(self, id): threading.Thread.__init__(self) self.id = id def run(self): time.sleep(5) print "子線程動做",threading.current_thread().name, datetime.now() if __name__ == "__main__": t1 = MyThread(999) t1.setDaemon(True) # 添加守護線程! t1.start() for i in range(5): print "主線程動做",threading.current_thread().name, datetime.now()
只是添加了join函數一行代碼,咱們發現主線程和子線程執行的順序就改變了。
主線程會等待子線程。
if __name__ == "__main__": t1 = MyThread(999) t1.start() t1.join() # 添加join函數! for i in range(5): print "主線程動做",threading.current_thread().name, datetime.now()
Output: 等待child執行完,再執行join()以後main thread的內容。
child thread Thread-4 2019-09-26 17:50:16.049128 main thread MainThread 2019-09-26 17:50:16.050622 main thread MainThread 2019-09-26 17:50:16.050930 main thread MainThread 2019-09-26 17:50:16.051079 main thread MainThread 2019-09-26 17:50:16.051915 main thread MainThread 2019-09-26 17:50:16.05206
主線程一直等待所有的子線程結束以後,主線程自身才結束,程序退出。(其實守護線程的設置就沒用了)
if __name__ == "__main__": t1 = MyThread(999) t1.setDaemon(True) # 添加守護線程! t1.start() t1.join() # 添加join函數! for i in range(5): print "主線程動做",threading.current_thread().name, datetime.now()
GIL 的全名是 the Global Interpreter Lock (全局解釋鎖),是常規 python 解釋器(固然,有些解釋器沒有)的核心部件。
GIL 是 Python 解釋器正確運行的保證,Python 語言自己沒有提供任何機制訪問它。但在特定場合,咱們仍有辦法下降它對效率的影響。
經過cpython啓動多進程,能 "繞過" GIL。
from multiprocessing import Process def spawn_n_processes(n, target): threads = [] for _ in range(n): thread = Process(target=target) thread.start() threads.append(thread) for thread in threads: thread.join()
經過 cpython 執行以上程序。
def test(target, number=10, spawner=spawn_n_threads): """ 分別啓動 1, 2, 3, 4 個控制流,重複 number 次,計算運行耗時 """ for n in (1, 2, 3, 4, ): start_time = time() for _ in range(number): spawner(n, target) end_time = time() print('Time elapsed with {} branch(es): {:.6f} sec(s)'.format(n, end_time - start_time)) test(fib, spawner=spawn_n_processes)
Python 的 「Queue 模塊」 中提供了同步的、線程安全的隊列類,包括
三種隊列均提供以下方法:
這些隊列都實現了鎖原語,可以在多線程中直接使用,可使用隊列來實現線程間的同步。
Queue 模塊中的經常使用方法:
import Queue
From: python多線程-queue隊列類型優先級隊列,FIFO,LIFO
默認隊列:Queue.Queue()
#coding=utf8 import Queue queuelist = Queue.Queue() for i in range(5): if not queuelist.full(): queuelist.put(i) print "put list : %s ,now queue size is %s "%(i,queuelist.qsize()) while not queuelist.empty(): print "get list : %s , now queue size is %s"%(queuelist.get(),queuelist.qsize())
Output:
put list : 0 ,now queue size is 1 put list : 1 ,now queue size is 2 put list : 2 ,now queue size is 3 put list : 3 ,now queue size is 4 put list : 4 ,now queue size is 5 get list : 0 , now queue size is 4 get list : 1 , now queue size is 3 get list : 2 , now queue size is 2 get list : 3 , now queue size is 1 get list : 4 , now queue size is 0
原本是個stack,非要叫成是LIFO隊列,汗~
#coding=utf8 import Queue queuelist = Queue.LifoQueue() for i in range(5): if not queuelist.full(): queuelist.put(i) print "put list : %s ,now queue size is %s "%(i,queuelist.qsize()) while not queuelist.empty(): print "get list : %s , now queue size is %s"%(queuelist.get(),queuelist.qsize())
Output:
put list : 0 ,now queue size is 1 put list : 1 ,now queue size is 2 put list : 2 ,now queue size is 3 put list : 3 ,now queue size is 4 put list : 4 ,now queue size is 5 get list : 4 , now queue size is 4 get list : 3 , now queue size is 3 get list : 2 , now queue size is 2 get list : 1 , now queue size is 1 get list : 0 , now queue size is 0
put方法的參數是個元組 (<優先級> ,<value>)。
#coding=utf8 import queue as Queue import random queuelist = Queue.PriorityQueue() for i in range(5): if not queuelist.full(): x=random.randint(1,20) y=random.randint(1,20) print x queuelist.put((x,y)) while not queuelist.empty(): print "get list : %s , now queue size is %s"%(queuelist.get(),queuelist.qsize())
Output:
11 5 10 7 10 get list : (5, 10) , now queue size is 4 get list : (7, 10) , now queue size is 3 get list : (10, 10) , now queue size is 2 get list : (10, 10) , now queue size is 1 get list : (11, 10) , now queue size is 0
內容:一個隊,三個檢票口 (三個線程)
鎖機制:不能同時「取」,因此取的過程須要加「鎖」。
#coding=utf8 import Queue import threading import time exitsingle = 0 class myThread(threading.Thread):
def __init__(self, threadname, queuelist): threading.Thread.__init__(self) self.threadname = threadname self.queuelist = queuelist def run(self): print "Starting queue %s"%self.threadname queue_enter(self.threadname, self.queuelist) # 每個線程從管道中」取數據「 time.sleep(1) print "close " + self.threadname
def queue_enter(threadname, queuelist): while not exitsingle: queueLock.acquire() if not workQueue.empty(): data = queuelist.get() queueLock.release() # 取完就能夠釋放「鎖」
print "%s check ticket %s" % (threadname, data) else: queueLock.release() time.sleep(1)
####################################################
# 初始化
#################################################### threadList = ["list-1", "list-2", "list-3"] queueLock = threading.Lock() workQueue = Queue.Queue() threads = [] queueLock.acquire() for num in range(100001,100020): workQueue.put(num) # 計入「票的編號」 queueLock.release()
print "start .."
# 三個線程從一個管道里取數據,但不能同時取 for name in threadList: thread = myThread( name, workQueue) thread.start() threads.append(thread) while not workQueue.empty(): pass exitsingle = 1 for t in threads: t.join() print "stop enter.."
但這裏貌似少了lock相關,具體可參考以上兩個栗子。
#!/usr/bin/python3 # -*- coding: utf-8 -*- # @Author : # @File : text.py # @Software : PyCharm # @description : XXX from queue import Queue import random import threading import time # Producer thread class Producer(threading.Thread):
def __init__(self, t_name, queue): threading.Thread.__init__(self, name=t_name) self.data = queue
def run(self): for i in range(5): print("%s: %s is producing %d to the queue!" % (time.ctime(), self.getName(), i)) self.data.put(i) time.sleep(random.randrange(10) / 5) print("%s: %s finished!" % (time.ctime(), self.getName())) # Consumer thread class Consumer(threading.Thread):
def __init__(self, t_name, queue): threading.Thread.__init__(self, name=t_name) self.data = queue
def run(self): for i in range(5): val = self.data.get() print("%s: %s is consuming. %d in the queue is consumed!" % (time.ctime(), self.getName(), val)) time.sleep(random.randrange(10)) print("%s: %s finished!" % (time.ctime(), self.getName()))
# Main thread def main(): queue = Queue() producer = Producer('Pro.', queue) consumer = Consumer('Con.', queue)
producer.start() consumer.start() producer.join() consumer.join() print('All threads terminate!') if __name__ == '__main__': main()
End.