Acknowledgements

This project would not have been possible without the guidance, support, and inspiration of several remarkable individuals.

My deepest thanks go to Sara Achour, whose teaching in CS349H: Software Techniques for Emerging Hardware Platforms provided both the conceptual foundation and methodological rigor that shaped this work. Her passion for emerging hardware—along with the software techniques that help mitigate their challenges and improve their usability—and her clarity in bridging computer science and electrical engineering made this project intellectually exciting and technically feasible.

I am equally grateful to Pu (Luke) Yi, whose support as the course Teaching Assistant was invaluable. His responsiveness, thoughtful feedback, and encouragement throughout the quarter helped me navigate challenges, scope the design of the simulator, and bring the project to completion.

I would also like to thank Philip Levis for his time, generosity, and broader influence on my research direction. His encouragement to explore the energy footprint of modern machine learning infrastructure played a meaningful role in motivating this line of inquiry.

To all of them—thank you for helping shape this project and for inspiring a deeper exploration of sustainable, energy-efficient computing.