This book presents a hands-on approach for solving electronic design automation problems with modern machine intelligence techniques by including step-by-step development of commercial grade design applications including resistance estimation, capacitance estimation, cell classification and others using dataset extracted from designs at 20nm. It walks the reader step by step in building solution flow for EDA problems with Python and Tensorflow.Intended audience includes design automation engineers, managers, executives, research professionals, graduate students, Machine learning enthusiasts, EDA and CAD developers, mentors, and the merely inquisitive. It is organized to serve as a compendium to a beginner, a ready reference to intermediate and source for an expert.
Every time I read a new research paper in machine learning and deep learning, my excitement used to peak on the possibilities of applying the same techniques to design automation. That gave me early motivation to start applying these techniques to design automation problems including:
- Capacitance Estimation
- Resistance Estimation
- Building effective wire load models
- Predicting RC trees with no routing for accurate delay calculation
- Build a fast parasitic extractor
- Improve early timing/power analysis accuracy.
- Speed up circuit simulator.
- Compress liberty libraries with dimensional reduction.
- Speed up Monte Carlo simulation for variation etc.
- Identify and label AMS/Analog circuits.
I've outlined some of these techniques in this book and released source code in srohit0.github.io/mida.
Who Should Read This Book?Intended audience includes design automation engineers, managers, executives, research professionals, graduate students, Machine learning enthusiasts, EDA and CAD developers, mentors, and the merely inquisitive. It is organized to serve as a compendium to a beginner, a ready reference to intermediate and source for an expert.
How To Read This Book?Beginner Quick first pass by overlooking math and programming.
General Reading Chapter 1, 2, 4 and 8.
Developing an Application Chapter 5, 6, 7 and 9.
Optional chapter 4, Appendix A and Appendix B.
Clarifying Concepts with Math Chapter 4, 5, 6, 7 and 9.
EDA industry has a long way to catch up with fast moving research in Machine Learning and Deep Learning. This book will serve to motivate and initiate my peers and professionals working in design automation.
