A Course in Physical Design of Integrated Circuits

Springer Your dedication goes here Preface Physical design of integrated circuits is a fascinating and rapidly advancing discipline, situated between the obvious and the impossible. Its roots are in hand-drawn circuit layout, where experienced engineers are capable of performing any operation, but lack the productivity to handle multimillion-gate circuits. Design automation is achieved by developing efficient algorithms for the " obvious " layout tasks and implementing them in software that can successfully operate at large scale. Several key layout tasks require solving NP-hard optimization problems in graph theory or computational geometry, which is why scalable algorithms are not easy to find. Furthermore, the graphs and collections of shapes arising in applications are so large that algorithms with cubic runtime are often impractical , and square matrices storing data for each pair of vertices do not fit into the memory of a typical workstation. Thus, physical design algorithms must solve NP-hard problems (approximate solutions are often acceptable) using only near-linear memory and near-linear runtime, at least on the instances arising in practical applications. In this book we attempted to bridge the gap between the obvious and the impossible by first introducing key layout tasks and studying their important aspects, then demonstrating the development of efficient computational techniques. This approach is reflected in the structure of the book, where each part first introduces key concepts and main challenges in its first chapter, and offers algorithmic solutions to these challenges in subsequent chapters. Every chapter ends with a brief review of the material covered and offers a set of exercises to check comprehension. We hope that you find the book both enjoyable and useful,