1. 程式人生 > >C++11多執行緒程式設計 第七章: 條件變數及其使用方法

C++11多執行緒程式設計 第七章: 條件變數及其使用方法

C++11 Multithreading – Part 7: Condition Variables Explained

Varun June 2, 2015 C++11 Multithreading – Part 7: Condition Variables Explained2018-08-18T15:21:29+00:00C++C++ 11c++11 ThreadsMultithreading 9 Comments

In this article we will discuss the usage of Condition Variable in C++11 Multi-threading with example.

 

Condition Variables

Condition Variable is a kind of Event used for signaling between two or more threads. One or more thread can wait on it to get signaled, while an another thread can signal this.

Header file required for condition Variable in C++11 is ,

 

1

#include <condition_variable>

A mutex is required along with condition variable.

How things actually work with condition variable,

  • Thread 1 calls the wait on condition variable, which internally acquires the mutex and check if required condition is met or not.
  • If not then it releases the lock and waits for Condition Variable to get signaled ( thread gets blocked). Condition Variable’s wait() function provides both these operations in atomic manner.
  • Another Thread i.e. like Thread 2 signals the Condition Variable when condition is met
  • Once Conditional Variable get signaled the the Thread 1 which was waiting for it resumes. It then acquires the mutex lock again and checks if the condition associated with Condition Variable is actually met or if it is superiors call. If more than one thread was waiting then notify_one will unblock only one thread.
  • If it was a superiors call then it again calls the wait() function.

Main member functions for std::condition_variable are,

Wait()

It makes the current thread to block until the condition variable get signaled or a spurious wake up happens.

It atomically releases the attached mutex, blocks the current thread, and adds it to the list of threads waiting on the current condition variable object. The thread will be unblocked when some thread calls notify_one() or notify_all() on same condition variable object. It may also be unblocked spuriously, therefore after every unblocking it needs to check condition again.

A callback is passed as an parameter to this function, which will be called to check if it is a spurious call or actually condition is met.

When threads get unlocked,

wait() function reacquires the mutex lock and checks that actually condition is met or not. If condition is not met then again it atomically releases the attached mutex, blocks the current thread, and adds it to the list of threads waiting on the current condition variable object.

notify_one()

If any threads are waiting on same conditional variable object then  notify_one unblocks one of the waiting threads.

notify_all()

If any threads are waiting on same conditional variable object then  notify_all unblocks all of the waiting threads.

Let’s see how we can handle previously discussed multi threaded scenario with condition variable i.e.

Need of Event Handling in Multi-threading (In this Article we handled the problem without Event Handling)

Problem Scenario

Suppose we are building a network based application. This application does following tasks,

  1. Perform some handshaking with server
  2. Load Data from XML files.
  3. Do processing on data loaded from XML.

As we can see that Task 1 is not dependent on any other Tasks but Task 3 is dependent on Task 2. So, it means Task 1 and Task 2 can be run in parallel by different Threads to improve the performance of application. So, let’s break this into a multi threaded application,

Responsibilities of Thread 1 are,

  • Perform some handshaking with server.
  • Wait for data to be loaded from XML by Thread 2
  • Do processing on data loaded from XML.

Responsibilities of Thread 2 are,

  • Load data from XML
  • Notify another Thread i.e. waiting for the message.

Code to achieve this with condition variable is as follows,

 

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

#include <iostream>

#include <thread>

#include <functional>

#include <mutex>

#include <condition_variable>

using namespace std::placeholders;

class Application

{

  std::mutex m_mutex;

  std::condition_variable m_condVar;

  bool m_bDataLoaded;

public:

  Application()

  {

    m_bDataLoaded = false;

  }

  void loadData()

  {

   // Make This Thread sleep for 1 Second

   std::this_thread::sleep_for(std::chrono::milliseconds(1000));

   std::cout<<"Loading Data from XML"<<std::endl;

   // Lock The Data structure

   std::lock_guard<std::mutex> guard(m_mutex);

   // Set the flag to true, means data is loaded

   m_bDataLoaded = true;

   // Notify the condition variable

   m_condVar.notify_one();

  }

  bool isDataLoaded()

  {

    return m_bDataLoaded;

  }

  void mainTask()

  {

    std::cout<<"Do Some Handshaking"<<std::endl;

    // Acquire the lock

    std::unique_lock<std::mutex> mlock(m_mutex);

    // Start waiting for the Condition Variable to get signaled

    // Wait() will internally release the lock and make the thread to block

    // As soon as condition variable get signaled, resume the thread and

    // again acquire the lock. Then check if condition is met or not

    // If condition is met then continue else again go in wait.

    m_condVar.wait(mlock, std::bind(&Application::isDataLoaded, this));

    std::cout<<"Do Processing On loaded Data"<<std::endl;

  }

};

int main()

{

   Application app;

   std::thread thread_1(&Application::mainTask, &app);

   std::thread thread_2(&Application::loadData, &app);

   thread_2.join();

   thread_1.join();

   return 0;

}

 

相對比直接用鎖需要while迴圈多次加鎖/去鎖多次, 條件變數只用加鎖兩次, 極大的降低了CPU的佔用率.