Moore’s law states that the number of transistors that can be placed on an integrated circuit (IC) doubles approximately every two years. This exponential growth in IC technology has led to advancements in everything digital, from central processing units (CPUs) and memory to digital cameras. Since computers are made up of CPUs, memory, and input/output (I/O) devices, it is a logical consequence that computers have also experienced tremendous improvements.
This drastic change in computers makes it difficult--if not impossible--for a textbook on computer architecture to include every new technology. Often, when a computer architecture textbook hits the counter, it is already out of date. Therefore, it is no wonder that this book is in its fifth edition.
The main part of the book contains six chapters. The focus is on parallelism. Besides a chapter on the fundamentals of quantitative methods and a chapter on memory hierarchy, the other four chapters deal with parallelism at various levels. It is explained as it relates to cloud computing in chapter 6, “Warehouse-Scale Computers to Exploit Request-Level and Data-Level Parallelism.” However, not everyone will agree with the authors’ decisions regarding which topics to include or exclude. For example, traditional computer architecture textbooks would include designs of CPU, memory, and I/O. In this book, I/O systems are rarely touched on at all.
Moore’s law tells us that computer industries and technologies are still quickly evolving. To chase the newest technology in a textbook is unrealistic. Going back to the basics may be the solution. We have to teach computer science students the basic principles that have applied since the computer was invented.