The Art of Writing Efficient Programs, 2nd Edition
English | 2026 | ISBN: 1835465188 | 702 pages | True PDF,EPUB | 62.48 MB
Gain practical skills to optimize C++ programs using modern CPU architectures, memory models, and concurrency toolsKey FeaturesApply advanced CPU, memory, and concurrency techniques to real-world codeDesign efficient thread-safe data structures using lock-free and atomic strategiesLeverage compiler behavior and modern C++ standards for performance gainsBook DescriptionHardware stopped making your programs faster for free a long time ago. What it does instead is reward code that understands it, code written with a clear picture of how the CPU executes instructions, how memory really moves through caches, and what the compiler can and cannot do on your behalf. This book teaches that understanding, from first principles to the most advanced techniques, and shows you how to turn it into measurably faster software.This edition is roughly half new and >60% longer. The coverage of CPU architecture (now including ARM), instruction processing, and the memory hierarchy, from caches to paging and virtual addressing, is expanded throughout. But the largest change is concurrency, which now makes up nearly half the book and is entirely rewritten: the memory model, atomic operations and atomic shared pointers, coroutine-based execution, NUMA-aware design, and a completely rewritten chapter on concurrent data structures that shows how cache coherency on modern hardware changes the established rules of lock-free programming, including advanced techniques such as helping.Written by Fedor G. Pikus, industry leader in high-performance C++, with a foreword by Herb Sutter, this book equips you with both conceptual clarity and the hands-on techniques needed to make your software faster, leaner, and smarter.What you will learnAnalyze and optimize CPU resource usageImprove memory access patterns across architecturesDesign performant multithreaded and lock-free systemsUse atomic operations and smart pointers effectivelyApply memory order guarantees for correctness and speedEvaluate compiler decisions and leverage vectorizationImplement coroutine-based execution modelsBuild APIs that scale in concurrent environmentsWho this book is forExperienced C++ developers, performance-focused programmers, and engineers working in high-speed systems like trading platforms, games, or scientific computing. Prior knowledge of multithreading and memory management is required to make the most of this book.Table of ContentsIntroduction to Performance and ConcurrencyPerformance MeasurementsCPU Architecture, Resources, and PerformanceMemory Architecture and PerformanceThreads, Memory, and ConcurrencyConcurrency and PerformanceData Structures for ConcurrencyConcurrency in C++High-Performance C++Compiler Optimizations in C++Undefined Behavior and PerformanceDesign for Performance
Gain practical skills to optimize C++ programs using modern CPU architectures, memory models, and concurrency tools
Apply advanced CPU, memory, and concurrency techniques to real-world code
Design efficient thread-safe data structures using lock-free and atomic strategies
Leverage compiler behavior and modern C++ standards for performance gains
Hardware stopped making your programs faster for free a long time ago. What it does instead is reward code that understands it, code written with a clear picture of how the CPU executes instructions, how memory really moves through caches, and what the compiler can and cannot do on your behalf. This book teaches that understanding, from first principles to the most advanced techniques, and shows you how to turn it into measurably faster software.
This edition is roughly half new and >60% longer. The coverage of CPU architecture (now including ARM), instruction processing, and the memory hierarchy, from caches to paging and virtual addressing, is expanded throughout. But the largest change is concurrency, which now makes up nearly half the book and is entirely rewritten: the memory model, atomic operations and atomic shared pointers, coroutine-based execution, NUMA-aware design, and a completely rewritten chapter on concurrent data structures that shows how cache coherency on modern hardware changes the established rules of lock-free programming, including advanced techniques such as helping.
Written by Fedor G. Pikus, industry leader in high-performance C++, with a foreword by Herb Sutter, this book equips you with both conceptual clarity and the hands-on techniques needed to make your software faster, leaner, and smarter.
Analyze and optimize CPU resource usage
Improve memory access patterns across architectures
Links are Interchangeable - Single Extraction
