Concurrent Thinking and Fundamentals
I believe that concurrency is a dividing line for C++ engineering skills. As single-core frequency gains hit the power wall, modern software performance improvements rely almost entirely on two directions: better algorithms, and better parallelization. The foundation of parallelization is concurrency—coordinating multiple flows of execution without breaking correctness, to fully leverage the computational power of multi-core hardware. Frankly, if the programs you write always execute sequentially in a single thread, no matter how beautiful the code is, you are wasting most of the transistors on the CPU in your hands.
However, before writing any multithreaded code, we must first answer three questions: why do we need concurrency? What problems can concurrency cause? And how do the CPU and operating system support multithreading? Most tutorials might start by teaching you std::thread, but I prefer not to do that. I want to establish the mindset of the concurrency domain first—correctness before performance, which is the principle of this entire volume.