Mutex and Atomics

by okanakyuz · 2020-03-15

To protect value with a simple mutex.

Mutex

C++

#include <iostream> #include <thread> #include <mutex> long sum{0}; std::mutex m_tex{}; void calcSum(int start, int end){ for(int x=start; x <= end; x++ ){ m_tex.lock(); sum += x; m_tex.unlock(); } } int main() { calcSum(0, 10000000); std::cout << "Normal Calculation " << sum << std::endl; sum = 0; std::thread t1(calcSum, 0, 5000000); std::thread t2(calcSum, 5000001, 10000000); t1.join(); t2.join(); std::cout << "Threat Calculation " << sum << std::endl; }

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

#include <iostream> #include <thread> #include <mutex>   long sum{0}; std::mutex m_tex{};   void calcSum(int start, int end){     for(int x=start; x <= end; x++ ){         m_tex.lock();         sum += x;         m_tex.unlock();     } }   int main() {     calcSum(0, 10000000);     std::cout << "Normal Calculation " <<  sum << std::endl;       sum = 0;     std::thread t1(calcSum, 0, 5000000);     std::thread t2(calcSum, 5000001, 10000000);     t1.join();     t2.join();       std::cout << "Threat Calculation " <<  sum << std::endl; }

The other way is using unique_lock but it locks threat critically

std::unique_lock

C++

#include <iostream> #include <thread> #include <mutex> long sum{0}; std::mutex m_tex{}; void calcSum(int start, int end){ std::unique_lock<std::mutex> unilock(m_tex); for(int x=start; x <= end; x++ ){ sum += x; } } int main() { calcSum(0, 10000000); std::cout << "Normal Calculation " << sum << std::endl; sum = 0; std::thread t1(calcSum, 0, 5000000); std::thread t2(calcSum, 5000001, 10000000); t1.join(); t2.join(); std::cout << "Threat Calculation " << sum << std::endl; }

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

#include <iostream> #include <thread> #include <mutex>   long sum{0}; std::mutex m_tex{};   void calcSum(int start, int end){     std::unique_lock<std::mutex> unilock(m_tex);     for(int x=start; x <= end; x++ ){         sum += x;     } }   int main() {     calcSum(0, 10000000);     std::cout << "Normal Calculation " <<  sum << std::endl;       sum = 0;     std::thread t1(calcSum, 0, 5000000);     std::thread t2(calcSum, 5000001, 10000000);     t1.join();     t2.join();       std::cout << "Threat Calculation " <<  sum << std::endl; }

New Memory model Atomic

Atomic brings us a secure model to reach data on memory. However, it is slower than the simple mutex operation, atomic types make us sure about memory sharing.

Atomic

C++

#include <iostream> #include <thread> #include <atomic> std::atomic<long> sum{0}; void calcSum(int start, int end){ for(int x=start; x <= end; x++ ){ sum += x; } } int main() { calcSum(0, 10000000); std::cout << "Normal Calculation " << sum.load() << std::endl; sum = 0; std::thread t1(calcSum, 0, 5000000); std::thread t2(calcSum, 5000001, 10000000); t1.join(); t2.join(); std::cout << "Threat Calculation " << sum.load() << std::endl; }

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

#include <iostream> #include <thread> #include <atomic>   std::atomic<long> sum{0};   void calcSum(int start, int end){     for(int x=start; x <= end; x++ ){         sum += x;     } }   int main() {     calcSum(0, 10000000);     std::cout << "Normal Calculation " <<  sum.load() << std::endl;       sum = 0;     std::thread t1(calcSum, 0, 5000000);     std::thread t2(calcSum, 5000001, 10000000);     t1.join();     t2.join();       std::cout << "Threat Calculation " << sum.load() << std::endl; }

Tags: AtomicCppMutex

You may also like...

• 0

  Smart pointers

  2020-03-03

• 0

  Non-type Template in Array

  2020-03-08

• 1

  Delete, Default and Delegating Constructors

  2020-03-02