In this chapter, we looked at atomic operations and exactly how they are integrated into compilers to allow one's code to work as closely with the underlying hardware as possible. The reader will now be familiar with the types of atomic operations, the use of a memory barrier (fencing), as well as the various types of memory ordering and their implications.
The reader is now capable of using atomic operations in their own code to accomplish lock-free designs and to make proper use of the C++11 memory model.
In the next chapter, we will take everything we have learned so far and move away from CPUs, instead taking a look at GPGPU, the general-purpose processing of data on video cards (GPUs).
