In this episode of the Computer Architecture Podcast, hosts Dr. Suvinay Subramanian and Dr. Lisa Hsu welcome Professor Fred Chang, the Seymour Goodman Professor in the Department of Computer Science at the University of Chicago and Chief Scientist of Supertech, a quantum software startup. Professor Chang, also a lead principal investigator for the NSF's Epic project, brings his extensive experience in quantum computing, architecture, and system stacks to the discussion.
The conversation delves into the rapidly evolving landscape of quantum computing. Professor Chang provides an overview of the field, explaining how quantum computation fundamentally differs from classical computing, particularly highlighting its unique property where adding a single device can double computational power—an exponential scaling not seen in classical systems. He discusses the current "historic time" for quantum computing, with real machines being built that can perform computations beyond the simulation capabilities of classical computers. This progress opens up new scientific frontiers, as understanding these machines now requires direct experimentation.
The episode explores the significant challenges and opportunities within quantum computing, including error correction, the nature of quantum algorithms, and the development of the full system stack. Professor Chang shares insights from his work and the broader research community, touching on topics from the fidelity of quantum gates and the limitations of current quantum machines to the ongoing efforts in developing benchmarks and debugging tools for this novel paradigm. He also discusses his personal journey into quantum computing, offering advice for students and researchers interested in this exciting and interdisciplinary field.
Chapters
00:00:00 — Introduction of Professor Fred Chang
00:01:16 — What Excites Professor Chang About Quantum Computing
00:02:10 — Broad Overview of Quantum Computing and its Differences from Classical Computing
00:04:33 — The Finicky Nature of Quantum Devices and Redundancy
00:05:37 — Error Correction and Mitigation in Quantum Systems
00:09:03 — Near-Term and Long-Term Quantum Algorithms and Applications
00:16:22 — Hardware vs. Software in Quantum Error Correction
00:19:19 — The QPU Concept and Partitioning Problems for Quantum Acceleration
00:24:01 — Challenges in Mapping Programs to Quantum Hardware
00:30:09 — The Race for Quantum Supremacy: Better Machines vs. Better Stacks
00:34:23 — Debugging and Verification in Quantum Computing
00:40:52 — An Open Question: Marrying Error Mitigation with Error Correction
00:43:27 — The Inflection Point in Quantum Computing and Future Outlook
00:48:13 — Essential Skills for Systems and Architecture Researchers Entering Quantum Computing
00:52:01 — Professor Chang's Serendipitous Path into Quantum Computing
00:56:00 — Words of Advice for Listeners
Takeaways
Exponential Scaling Potential: Quantum computing is unique in that adding a single device (like a qubit) can theoretically double computational power, offering true exponential scaling, a feat unachievable by classical computers.
Current Era of Experimentation: We are at a point where real quantum machines exist that classical computers cannot simulate, making direct experimentation crucial for understanding and advancing the field.
Error Correction is a Fundamental Challenge: Due to the "no-cloning theorem" and the inherent fragility of quantum states, error correction and mitigation are paramount. Current approaches range from running short, reliable computations to developing complex error correction codes and mitigation techniques.
The Software Stack is Critical: While hardware advancements are key, the software stack—from high-level programming languages down to pulse-level control—plays a vital role. Optimizing this stack, sometimes by breaking traditional abstractions to be "physics-aware," can yield significant performance improvements (e.g., 2-3 orders of magnitude).
Debugging and Verification are Immense Hurdles: Unlike classical systems, directly observing the internal state of a quantum computation without collapsing it is impossible. This makes debugging and verifying quantum programs and hardware incredibly challenging, necessitating new formal methods and tools.