What this book covers

Chapter 1, Revisiting Multithreading, summarizes multithreading in C++, revisiting all the concepts you should already be familiar with and going through a basic example of multithreading using the native threading support added in the 2011 revision of C++.

Chapter 2, Multithreading Implementation on the Processor and OS, builds upon the fundamentals provided by the hardware implementations discussed in the preceding chapter, showing how OSes have used the capabilities to their advantage and made them available to applications. It also discusses how processes and threads are allowed to use the memory and processor in order to prevent applications and threads from interfering with each other.

Chapter 3, C++ Multithreading APIs, explores the wide variety of multithreading APIs available as OS-level APIs (for example, Win32 and POSIX) or as provided by a framework (for example, Boost, Qt, and POCO). It briefly runs through each API, listing the differences compared to the others as well as the advantages and disadvantages it may have for your application.

Chapter 4, Thread Synchronization and Communication, takes the topics learned in the previous chapters and explores an advanced multithreading implementation implemented using C++ 14's native threading API, which allows multiple threads to communicate without any thread-safety issues. It also covers the differences between the many types of synchronization mechanisms, including mutexes, locks, and condition variables.

Chapter 5, Native C++ Threads and Primitives, includes threads, concurrency, local storage, as well as thread-safety supported by this API. Building upon the example in the preceding chapter, it discusses and explores extending and optimizing thread-safty using the features offered by the full feature set in C++ 11 and C++ 14.

Chapter 6, Debugging Multithreaded Code, teaches you how to use tools such as Valgrind (Memcheck, DRD, Helgrind, and so on) to analyze the multithreaded performance of an application, find hotspots, and resolve or prevent issues resulting from concurrent access.

Chapter 7, Best Practices, covers common pitfalls and gotchas and how to spot them before they come back to haunt you. It also explores a number of common and less common scenarios using examples.

Chapter 8, Atomic Operations – Working with the Hardware, covers atomic operations in detail: what they are and how they are best used. Compiler support is looked at across CPU architectures and an evaluation is made of when it is worth to invest time in implementing atomic operations in your code. It also looks at how such optimizations will limit the portability of your code.

Chapter 9, Multithreading with Distributed Computing, takes many of the lessons learned in the preceding chapters and applies them on a multi-system, cluster-level scale. Using an OpenMPI-based example, it shows how multithreading can be done across multiple systems, such as the nodes in a computer cluster.

Chapter 10, Multithreading with GPGPU, shows the use of multithreading in GPGPU applications (for example, CUDA and OpenCL). Using an OpenCL-based example, a basic multithreaded application is explored that can execute tasks in parallel. This chapter takes lessons learned in the preceding chapters and applies them to processing on video cards and derived hardware (for example, rack-mounted vector processor hardware).