Efficient Computing Lab

Home

The Efficient Computing Lab (ECL) is led by Prof. Lin Zhong and is located in Arthur K. Watson Hall. The research laboratory is located in AKW 408. Our current research focuses on designing low-latency, high-throughput systems in the context of AI and Quantum Computing. In AI, we are rethinking the foundation model serving stack. Our work on Prompt Cache has already become a standard practice in industry. In Quantum Computing, we address one of the most computationally demanding classical challenges for fault-tolerant quantum computers: quantum error correction (QEC) decoding. We have developed some of the world's most advanced QEC decoders, including Helios, Fusion Blossom, and Micro Blossom, which have influenceed both commercial QEC decoder development and scientific experiments.

Pie, Programmable Inference Engine, for foundation models, has been released. (SOSP paper)

We also have a lasting interest in novel operating systems (OS) design, from Linux-based K2 to Theseus OS Theseus and Hopter, both written from scratch in Rust. We continue to explore how generative AI could reshape the OS stack.

All our research is open-source. If you have trouble locating a code repository, feel free to email us.

☆Graduate research opportunities available in a broad range of experimental topics. Many of the opportunities are collaborative with colleagues from the NSF AI Institute and Yale Quantum Institute.

Selected Recent Publications (Full list)

Pie: a programmable serving system for emerging LLM applications (SOSP 2025)(PDF, Code release)
Blindfold: Confidential memory management by untrusted operating system (NDSS 2025) (arXiv) (Distinguished Paper)
Prompt Cache: Modular attention reuse for low-latency inference (MLSys 2024) (arXiv) (impact: Anthropic, Google, and OpenAI)
Fusion Blossom: fast MWPM decoders for QEC (IEEE QCE 2023) (arXiv) (Best Paper, 1st Place, Quantum Systems Software)
Theseus: an experiment in operating system structure and state management (OSDI 2020) (PDF)

[Home] [People] [Publications] [Projects] [Downloads] [News] [Acknowledgments] [Contact us]