Concurrent Programming in C++

Understanding concurrency: pitfalls and evolutions in programming

The concurrency support in C++ makes it possible for a program to execute multiple tasks concurrently. As mentioned earlier, writing a correct concurrent C++ program is, in general, a lot harder than writing a program that executes all tasks sequentially in one thread. This section will also demonstrate some common pitfalls to make you aware of all the difficulties involved in writing concurrent programs.

Concurrency support was first introduced in C++11 and has since been extended into C++14, C++17, and C++20. Before concurrency was part of the language, it was implemented with native concurrency support from the operating system, POSIX Threads (pthreads), or some other library.

With concurrency support directly in the C++ language, we can write cross-platform concurrent programs, which is great! Sometimes, however, we have to reach for platform-specific functionality when dealing with concurrency on our platform. For example, there is no support in the C++ standard library for setting thread priorities, configuring CPU affinity (CPU pinning), or setting the stack size of new threads.

It should also be said that the thread support library has been extended quite a bit with the release of C++20, and more features are likely to be added in future versions of the language. The need for good concurrency support is increasing because of the way hardware is being developed, and there is a lot yet to be discovered regarding the efficiency, scalability, and correctness of highly concurrent programs.

The thread support library

We will now take a tour through the C++ thread support library and cover its most important components.

Threads

A running program contains at least one thread. When our main function is called, it is executed on a thread usually referred to as the main thread. Each thread has an identifier, which can be useful when debugging a concurrent program. The following program prints the thread identifier of the main thread: