Integrated Circuits and Systems Series Editor Anantha Chandrakasan, Massachusetts Institute of Technology Cambridge, Massachusetts For other titles published in this series, go to http://www.springer.com/series/7236 Yuan Xie · Jason Cong · Sachin Sapatnekar Editors Three-Dimensional Integrated Circuit Design EDA, Design and Microarchitectures 123 Editors Yuan Xie Jason Cong Department of Computer Science and Department of Computer Science Engineering University of California, Los Angeles Pennsylvania State University [email protected] [email protected] Sachin Sapatnekar Department of Electrical and Computer Engineering University of Minnesota [email protected] ISBN 978-1-4419-0783-7 e-ISBN 978-1-4419-0784-4 DOI 10.1007/978-1-4419-0784-4 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2009939282 © Springer Science+Business Media, LLC 2010 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com) Foreword We live in a time of great change. In the electronics world, the last several decades have seen unprecedented growth and advancement, described by Moore’s law. This observation stated that transistor density in integrated circuits doubles every 1.5–2 years. This came with the simultaneous improvement of individual device perfor- mance as well as the reduction of device power such that the total power of the resulting ICs remained under control. No trend remains constant forever, and this is unfortunately the case with Moore’s law. The trouble began a number of years ago when CMOS devices were no longer able to proceed along the classical scaling trends. Key device parameters such as gate oxide thickness were simply no longer able to scale. As a result, device off- state currents began to creep up at an alarming rate. These continuing problems with classical scaling have led to a leveling off of IC clock speeds to the range of several GHz. Of course, chips can be clocked higher but the thermal issues become unmanageable. This has led to the recent trend toward microprocessors with multi- ple cores, each running at a few GHz at the most. The goal is to continue improving performance via parallelism by adding more and more cores instead of increasing speed. The challenge here is to ensure that general purpose codes can be efficiently parallelized. There is another potential solution to the problem of how to improve CMOS technology performance: three-dimensional integrated circuits (3D ICs). By moving to a technology with multiple active “tiers” in the vertical direction, a number of significant benefits can be realized. Global wires become much shorter, interconnect bandwidth can be greatly increased, and latencies can be significantly decreased. Large amounts of low-latency cache memory can be utilized and intelligent physical design can help mitigate thermal and power delivery hotspots. Three-dimensional IC technology offers a realistic path for maintaining the progress defined by Moore’s Law without requiring classical scaling. This is a critical opportunity for the future. The Defense Advanced Research Project Agency (DARPA) has recognized the significance of 3D IC technology a number of years ago and began carefully tar- geted investments in this area based on the potential for military relevance and applications. There are also many potential commercial benefits from such a tech- nology. The Microsystems Technology Office at DARPA has launched a number of 3D IC-based Programs in recent years targeting areas such as intelligent imagers, v vi Foreword heterogeneously integrated 3D stacks, and digital performance enhancement. The research results in a number of the chapters in this book were made possible by DARPA-sponsored programs in the field of 3D IC. Three-dimensional IC technology is currently at an early stage, with several pro- cesses just becoming available and more in the early development stage. Still, its potential is so great that a dedicated community has already begun to seriously study the EDA, design, and architecture issues associated with 3D IC, which are well summarized in this book. Chapter 1 provides a good introduction to this field by an expert from IBM well versed in both design and technology aspects. Chapter 2 pro- vides an excellent overview of key 3D IC technology issues by process technology researchersfrom IBM and can be beneficial to any designer or architect. Chapters 3– 6 cover important 3D IC electronic design automation (EDA) issues by researchers from the University of California, Los Angeles and the University of Minnesota. Key issues covered in these chapters include methods for managing the thermal, electrical, and layout challenges of a multi-tier electronic stack during the modeling and physical design processes. Chapters 7–9 deal with 3D design issues, including the 3D processor design by authors from the Georgia Institute of Technology, a 3D network-on-chip (NoC) architecture by authors from Pennsylvania State University, and a 3D architectural approach to energy efficient server designs by authors from the University of Michigan and Intel. The book concludes with a system–level analysis of the potential cost advantages of 3D IC technology by researchers at Pennsylvania State University. As I mentioned in the beginning we live at a time of great change. Such change can be viewed as frightening, as long-held assumptions and paradigms, such as Moore’s Law, lose relevance. Challenging times are also important opportunities to try new ideas. Three-dimensional IC technology is such a new idea and this book will play an important and pioneering role in ushering in this new technology to the research community and the IC industry. DARPA Microsystems Technology Office Michael Fritze, Ph.D. Arlington, Virginia March 2009 Preface To the observer, it would appear that New York city has a special place in the hearts of integrated circuit (IC) designers. Manhattan geometries, which mimic the blocks and streets of the eponymous borough, are routinely used in physical design: under this paradigm, all shapes can be decomposed into rectangles, and each wire is either parallel and perpendicular to any other. The advent of 3D circuits extends the anal- ogy to another prominent feature of Manhattan – namely, its skyscrapers – as ICs are being built upward, with stacks of active devices placed on top of each other. More precisely, unlike conventional 2D IC technologies that employ a single tier with one layer of active devices and several layers of interconnect above this layer, 3D ICs stack multiple tiers above each other. This enables the enhanced use of silicon real estate and the use of efficient communication structures (analogous to elevators in a skyscraper) within a stack. Going from the prevalent 2D paradigm to 3D is certainly not a small step: in more ways than one, this change adds a new dimension to IC design. Three-dimensional design requires novel process and manufacturing technologies to reliably, scalably, and economically stack multiple tiers of circuitry, design methods from the circuit level to the architectural level to exploit the promise of 3D, and computer-aided design (CAD) techniques that facilitate circuit analysis and optimization at all stages of design. In the past few years, as process technologies for 3D have neared maturity and 3D circuits have become a reality, this field has seen a flurry of research effort. The objective of this book is to capture the current state of the art and to provide the readers with a comprehensive introduction to the underlying manufacturing technol- ogy, design methods, and computer-aided design (CAD) techniques. This collection consists of contributions from some of the most prominent research groups in this area, providing detailed insights into the challenges and opportunities of designing 3D circuits. The history of 3D circuits goes back many years, and some of its roots can be traced to a major government-funded program in Japan from a couple of decades ago. It is only in the past few years that the idea of 3D has gained major traction, so that it is considered a realistic option today. Today, most major players in the semiconductor industry have dedicated significant resources and effort to this area. As a result, 3D technology is at a stage where it is poised to make a major leap. The context and motivation for this technology are provided in Chapter 1. vii viii Preface The domain of 3D circuits is diverse, and various 3D technologies available today provide a wide range of tradeoffs between cost and performance. These include silicon-carrier-like technologies with multiple dies mounted on a substrate, wafer stacking with intertier spacings of the order of hundreds of microns, and thinned die/wafer stacks with intertier distances of the order of ten microns. The former two have the advantage of providing compact packaging and higher levels of integration but often involve significant performance overheads in communications from one tier to another. The last, with small intertier distances, not only provides increased levels of integration but also facilitates new architectures that can actually improve significantly upon an equivalent 2D implementation. Such advanced technologies are the primary focus of this book, and a cutting-edge example within this class is described in detail in Chapter 2.