Thread and shared lock

在看programing python 4th，第5張parallel system tool 192頁開始，書中講到thread知識，如下作個筆記，以便後期學習

1.主線程執行，開啓5個子線程進行計數，沒有使用mutex鎖住，因此線程沒有lock住資源，每一個線程對全局變量的操做錯亂，結果以下：

 1 """
 2 synchronize access to stdout: because it is shared global
 3 thread outputs may be intermixed if not syschronized
 4 """
 5 import thread,time
 6 global num                                  #global var to be used by many threads
 7 num=0
 8 
 9 def cnt(id,count):                          # function run in threads
10     for i in range(count):
11         global num
12         #mutex.acquire()                     # lock the share var before execute 
13         num +=1
14         time.sleep(0.5)                     # simulate read work
15         print('[%s] num= %s\n' %(id,num))   #print isn't interrupted now
16         #mutex.release()                     #release the lock for the other thread
17 
18 if __name__ =="__main__":
19     #mutex=thread.allocate_lock()            #make a global mutex for lock
20     for i in range(5):                      #spawm 5 threads
21         thread.start_new_thread(cnt,(i,3))  #start threads
22     time.sleep(8)                          # wait for spawn thread work done,don't exit too early
23    
24     print('main thread exitting')

View Code

2.把mutex 註釋打開，有了mutex變量，每個線程進入都會獨佔num變量，結果以下：

 1 """
 2 synchronize access to stdout: because it is shared global
 3 thread outputs may be intermixed if not syschronized
 4 """
 5 import thread,time
 6 global num                                  #global var to be used by many threads
 7 num=0
 8 
 9 def cnt(id,count):                          # function run in threads
10     for i in range(count):
11         global num
12         mutex.acquire()                     # lock the share var before execute 
13         num +=1
14         time.sleep(0.5)                     # simulate read work
15         print('[%s] num= %s\n' %(id,num))   #print isn't interrupted now
16         mutex.release()                     #release the lock for the other thread
17 
18 if __name__ =="__main__":
19     mutex=thread.allocate_lock()            #make a global mutex for lock
20     for i in range(5):                      #spawm 5 threads
21         thread.start_new_thread(cnt,(i,3))  #start threads
22     time.sleep(8)                          # wait for spawn thread work done,don't exit too early
23    
24     print('main thread exitting')

View Code

3.若是把time.sleep(6)註釋掉或者子線程沒有執行完畢，而主線程sleep的時間一到，主線程直接退出而不等待子線程執行完畢，結果以下：

a.主線程不等待，則直接退出

b.主線程只等待3s，而5個子線程須要7.5s，因此num只計數5.

4.設定有效等待時間和鎖以後，主線程等待全部子線程執行結束才退出，結果以下：

6.無需在主線程設置等待時間，而是設定單獨的鎖或者變量來記錄每一個子線程的執行狀態，每執行完一個線程，設定狀態鎖，而後在主線程判斷全部狀態鎖的狀態便可

 1 """
 2 used mutexex to know when threads are done in parent/main thread,
 3 instead of time.sleep;lock stdout to avoid comingled prints
 4 """
 5 import thread,time
 6 global num   
 7 num =0
 8 
 9 def cnt(id,count):
10     for i in range(count):
11         global num
12         stdoutmutex.acquire()
13         num +=1
14         time.sleep(0.5)
15         print('[%s] num= %s time:[%s]\n' %(id,num,time.ctime()))   #print isn't interrupted now
16         stdoutmutex.release()
17     #exitmutexs[id].acquire() # signal main thread
18     exitFlags[id] = True     #signal main thread
19     
20 if __name__ =="__main__":           
21     stdoutmutex = thread.allocate_lock()                           #make a global mutex for lock
22     #exitmutexs = [thread.allocate_lock() for i in range(5)] 
23     exitFlags=[False]*5
24     for i in range(5):                                             #spawm 5 threads
25         thread.start_new_thread(cnt,(i,3))                         #start threads
26     #for mutex in exitmutexs:
27     #    while not mutex.locked():                                  
28     #        pass
29     while False in exitFlags:pass
30     print('main thread exitting')

View Code