Computer architecture has evolved over time. Today’s cheap yet powerful computer systems are a result of hardware miniaturization and optimization techniques, among many other factors. Progress will continue as user needs change over time. In order to contribute to any future evolution, one needs to first understand current computer design. Computer architecture: a quantitative approach is a good start. Many universities have already adopted it as a core reference. This review looks at the sixth edition.

Chapter 1 is a very informative introduction to computer architecture. It defines different types of computers and explains in detail the hardware and software evolution driving the changes in computer architecture.

Chapter 2 covers memory technology and optimizations, memory hierarchy design, and protection via virtual memory and virtual machines.

Chapter 3 details instruction-level parallelism, including dynamic parallelism managed by hardware and static parallelism managed by software at compile time.

Chapter 4 explains data-level parallelism in vector, graphics processing unit (GPU), and multimedia single instruction, multiple data (SIMD) architectures.

Chapters 5 and 6 discuss thread-level and request-level parallelisms, respectively.

Chapter 7, “Domain-Specific Architectures,” is a new addition to the sixth edition. It includes Google’ tensor processing unit (TPU), Microsoft Catapult, Intel Crest, and Pixel Visual Core.

Overall, the text is quite comprehensive, reflecting the authors’ deep knowledge of, and significant experience with, the subject. Each new concept is presented in historical context. This makes it easy to understand the design decisions involved in building computers.

Each chapter includes a section on the main “fallacies and pitfalls” related to the covered concepts, and a set of case studies and questions for assessment. Exercises are rated with minutes (10, 15, and so on) to give readers a sense of the amount of time required.

The book’s companion website offers historical perspective resources, instructor slides in PowerPoint, and links to related material on the web. This makes the current edition a good candidate as a course reference.

In summary, Computer architecture is definitely a textbook that students, teachers, and researchers need to have on their shelf.

More reviews about this item: Amazon, Goodreads