Prof. Xuehai Qian
Tsinghua University, China
In this talk, I will first briefly review my research journal from multi-core architectures to domain-specific systems and architectures for AI and graph analytics applications, following the modern SW/HW co-designed approach. I will spend most time of the talk introducing our new research project on building a whole-stack AI system with Temporal-Spatial Deterministic Architecture named Timely Token Processing Unit (TTPU). Specifically, we will discuss the concept and design principles of the new type of architecture, followed research challenges and problems on single-TTPU, compiler, and multi-TTPU. Finally, we outline the full system organization of hybrid GPU/NPU-TTPU system and edge AI system based on TTPU.
Xuehai Qian is professor of the Department of Computer Science and Technology at Tsinghua University. He previously taught at Purdue University and University of Southern California. His current research focuses on developing efficient systems and architectures for the future AI applications. In the past, he developed scalable and efficient coherence protocols and architectural techniques to support the debugging and enhance programmability of shared-memory parallel programs on multicores. He was the recipient of W.J Poppelbaum Memorial Award at UIUC, NSF CRII and CAREER Award, and the inaugural ACSIC (American Chinese Scholar In Computing) Rising Star Award. He has been inducted into the "Hall of Fames" of all of the top-4 architecture conferences (ISCA, MICRO, ASPLOS, HPCA).
Web site is up.
Call for Papers published.
Paper Submission Due
Paper Submission Due
Paper Submission Due
Notification of Acceptance
Final Manuscript (Camera Ready)
Conference Date
Copyright © 2026 International Conference on Metaverse Computing, Networking and Applications