Search⌘ K
Home
Courses
Modern C++ Concurrency: Get the most out of any machine

Thread Local Data

Explore how thread local data enables each thread to have its own separate copy of variables in C++ concurrency. Understand the use of the thread_local keyword to avoid data races and safely port single-threaded programs to multithreaded environments. This lesson includes practical examples to illustrate memory isolation for thread-safe code.

We'll cover the following...

Thread-local data, also known as thread-local storage, will be created for each thread separately. It behaves like static data because it’s bound for the lifetime of the thread and it will be created at its first usage. Also, thread-local data belongs exclusively to the thread.

C++
// threadLocal.cpp
#include <iostream>
#include <string>
#include <mutex>
#include <thread>
std::mutex coutMutex;
thread_local std::string s("hello from ");
void addThreadLocal(std::string const& s2){
  s += s2;
  // protect std::cout
  std::lock_guard<std::mutex> guard(coutMutex);
  std::cout << s << std::endl;
  std::cout << "&s: " << &s << std::endl;
  std::cout << std::endl;
}
int main(){
  std::cout << std::endl;
  std::thread t1(addThreadLocal,"t1"); 
  std::thread t2(addThreadLocal,"t2"); 
  std::thread t3(addThreadLocal,"t3"); 
  std::thread t4(addThreadLocal,"t4"); 
  t1.join();
  t2.join();
  t3.join();
  t4.join();
}

By using the keyword ...