Book Image

C++17 STL Cookbook

By : Jacek Galowicz
Book Image

C++17 STL Cookbook

By: Jacek Galowicz

Overview of this book

C++ has come a long way and is in use in every area of the industry. Fast, efficient, and flexible, it is used to solve many problems. The upcoming version of C++ will see programmers change the way they code. If you want to grasp the practical usefulness of the C++17 STL in order to write smarter, fully portable code, then this book is for you. Beginning with new language features, this book will help you understand the language’s mechanics and library features, and offers insight into how they work. Unlike other books, ours takes an implementation-specific, problem-solution approach that will help you quickly overcome hurdles. You will learn the core STL concepts, such as containers, algorithms, utility classes, lambda expressions, iterators, and more, while working on practical real-world recipes. These recipes will help you get the most from the STL and show you how to program in a better way. By the end of the book, you will be up to date with the latest C++17 features and save time and effort while solving tasks elegantly using the STL.
Table of Contents (18 chapters)
Title Page
Credits
About the Author
About the Reviewer
www.PacktPub.com
Customer Feedback
Preface
Index

Implementing a tiny automatic parallelization library with std::future


Most complex tasks can be broken down into subtasks. From all subtasks, we can draw an directed acyclic graph (DAG) that describes which subtask depends on what other subtasks in order to finish the higher level task. Let us, for example, imagine that we want to produce the string "foo bar foo bar this that ", and we can only do this by creating single words and concatenate those with other words, or with themselves. Let's say this functionality is provided by three primitive functions create, concat, and twice.

Taking this into account, we can draw the following DAG that visualizes the dependencies between them in order to get the final result:

When implementing this in code, it is clear that everything can be implemented in a serial manner on one CPU core. Alternatively, all subtasks that depend on no other subtasks or other subtasks that already have been finished, can be executed concurrently on multiple CPU cores.

It...