This is perhaps the best and most thorough digital arithmetic design book in print. While it is mostly oriented toward practicing engineers, the well-engineered blend between basic theory, algorithm descriptions, and real-life implementations will be beneficial to students having to deal with these matters. For instance, while textbooks by Ercegovac and Lang [1], Parhami [2], and Koren [3] are extremely well written, it would take some advanced engineering skills to turn the design ideas inside of them into field-programmable gate arrays (FPGAs) or application-specific integrated circuit (ASIC) working modules. This book can serve as a guide for acquiring these skills in a consistent and fast manner. The above-cited textbooks would be a good supplement if more theoretical insight is desired.
The book has 16 chapters. In many technical books, the introductory chapter is present because it has to be there, even if its content could usually be! skipped. That is not so with this book. Right after the second page, there is an example design of a digital signature system. Although simple, it gives a feel for how the various topics to follow will be treated--meticulously. Chapter 2 serves as a mathematical refresher and background. The emphasis is on abstract algebra concepts and introductory number theory. Even though the notation can be disorienting at times, this well-written chapter contains perhaps the shortest and clearest Chinese remainder theorem proof in an engineering text. The function approximation section is very brief, and perhaps should be expanded in a future edition.
Chapter 3 is the key to the arithmetic algorithms, since it provides a detailed treatment of number representations. What follows is a series of chapters on the fundamental arithmetic procedures from an algorithmic viewpoint. Almost no hardware is introduced, even though detailed block diagrams are provided throughout, which ma!y be thought of as a form of high-level hardware representation. These chapters cover addition and subtraction (chapter 4), multiplication and squaring (chapter 5), division (chapter 6), and operations that are a superset of the basic ones, such as base conversions and special functions (chapter 7). Basic details about the CORDIC and square root algorithms are also provided. Chapter 8 gives an insightful account of finite field arithmetic operations.
Chapter 9 provides general information about circuit design, software, processors, and FPGA and ASIC structures. Chapter 10 is an important one, although short. It introduces the concepts of pipelining, scheduling, and even self-timing, which should probably be given expanded coverage in a subsequent edition.
Chapters 11 through 15 mirror the arithmetic algorithms chapters. These chapters alone are worth the purchase price of the book. It is virtually impossible to convey in a short review what exactly “happens!” in them. Basically, they contain everything you ever wanted to know about arithmetic circuitry. In addition, Sutter8217;s Web site contains very-high-speed integrated circuit (VHSIC) hardware description language (VHDL) code for everything mentioned in the book. Of course, there are quite a few sites dealing with VHDL or Verilog arithmetic algorithm implementations, but the level of detail and the clarity of the explanations in this part of the book are rarely matched. The final chapter presents the design of a floating point arithmetic unit. This book is warmly recommended to anyone having to design or understand how computer arithmetic operates at almost every conceivable level of detail.
