Acknowledgments
This book grew from work and conversations across computer architecture, machine-learning systems, electronic design automation, and benchmarking. I am grateful to the students, collaborators, colleagues, and members of the broader research community who questioned its framing, challenged weak claims, and kept the argument grounded in evidence. Their questions and examples repeatedly brought the discussion back to what architects can measure, verify, and defend.
I am especially grateful to Parthasarathy (Partha) Ranganathan, Vice President and Engineering Fellow at Google, who delivered the inaugural keynote at the online Architecture 2.0 workshop. That workshop gave the effort its early momentum and helped turn an emerging idea into a broader community conversation.
Srivatsan Krishnan (NVIDIA), my former student at Harvard, led the research project that generated many of the core ideas discussed in this book. I am particularly grateful for his leadership and contributions throughout that effort.
Many other colleagues helped test and sharpen the vision behind this book. I am grateful to Saman Amarasinghe (Massachusetts Institute of Technology), David Brooks (Harvard University), Dan Connors (NVIDIA), Siddharth Garg (New York University), Brian Hirano (Micron), Qijing Jenny Huang (NVIDIA), Ravi Iyer (Intel), David Kanter (MLCommons), Christos Kozyrakis (Stanford University), Tushar Krishna (Georgia Institute of Technology), Benjamin C. Lee (University of Pennsylvania), Hsien-Hsin Sean Lee (Intel), Jae W. Lee (Seoul National University), Yingyan (Celine) Lin (Georgia Institute of Technology), Jason Lowe-Power (University of California, Davis), Martin Maas (Google DeepMind), Divya Mahajan (Georgia Institute of Technology), Ankita Nayak (Qualcomm), Phitchaya Mangpo Phothilimthana (Google DeepMind), Matt Sinclair (University of Wisconsin-Madison), Srinivas Sridharan (NVIDIA), Thierry Tambe (Stanford University), Zishen Wan (Columbia University), Carole-Jean Wu (Meta), Amir Yazdanbakhsh (Google DeepMind), Hongil Yoon (Google), and Cliff Young (Google DeepMind).
The book is stronger because of their questions and contributions. Any errors that remain are my own.