Preface

Solving problems is a large part of writing code. Sometimes these are small problems that we barely acknowledge and sometimes they are grand challenges that seem insurmountable at first. Whatever domain you work in, and whatever the problem, solving problems is an iterative process with false starts and dead ends until you eventually find the “right” path – this is normal. After solving a few big problems, you’ll start to recognize some common features in this process: breaking the problem down into smaller pieces, formulating or finding good abstractions, recognizing patterns that appear in other kinds of problems, and formulating a list of steps to follow to construct solutions to the problem. In computer science, this framework for solving problems is called computational thinking.

Of course, formulating abstract solutions and algorithms is only part of actually solving the problem. The other part is actually realizing the solution as software and delivering it, whatever form that might take. This means you must first select the right tool to build this software, which includes selecting the programming language to use. Sometimes this choice is made for you, but other times, the choice depends on many factors. When performance is paramount, C++ is an obvious choice.

Modern C++ is very powerful and comes with a large ecosystem of high-performance libraries. Many substantial updates in the past decade have brought C++ into the modern era of programming languages. More than anything, what C++ provides is control. Other languages can achieve performance in some places, but they often fall down in other areas. Few languages deliver the same performance as C++ without sacrificing control, flexibility, or ease of use.

Being proficient at solving problems in C++, of course, requires you to have a good knowledge of C++ and standard techniques. Delivering the right balance of performance and flexibility often also requires an understanding of the broader context in which the code will run: the operating system, hardware, processor features, and specific accelerator hardware.

Having all this in mind when choosing data structures and designing algorithms can give you the edge. But it’s more than that. Understanding these concepts will also help you select good abstractions and identify useful patterns more easily. The connection between C++ and problem-solving is two-way, and that is the topic of this book.