Advanced Computer Architecture and Systems
Performance principles, architectural design, and the effective use of computing resources.
ACADEMIC RESEARCH GROUP
Yuhang Liu Research Group
Advanced Computer Architecture and Systems
Led by Yuhang Liu at the Institute of Computing Technology, Chinese Academy of Sciences, the group investigates performance mechanisms, architectural design, and practical efficiency in computer systems. It also pursues historical and source-critical studies of computing, as well as science and technology strategy and talent development.
RESEARCH & ACADEMIC PROFILE
Research Areas & Academic Profile
Technology innovation · Historical inquiry · Strategic thinking
Historical and Source-Critical Studies of Computing
Examining original sources, interpreting landmark contributions, and studying the evolution of computing technology.
Science and Technology Strategy & Talent Development
Which research problems matter—and how can we organize research and train people to address them effectively?
Grounded in computer systems research, informed by historical evidence, and attentive to consequential questions in science and education.
The three areas have distinct methods and research objects while informing one another.
About the Research Group · Read more
Our primary research area is advanced computer architecture and systems. We investigate scientific and engineering challenges that limit computing performance and effectiveness, spanning theoretical models, design methods, and implemented systems. Our work covers memory access concurrency, storage systems, processor architectures, datacenter resource management, and hardware–software co-optimization.
Alongside technical research, we pursue historical and source-critical studies of computing. By examining foundational publications, historical figures, and the evolution of computing devices, we seek to understand past technical choices and draw carefully grounded lessons for contemporary research.
We also study science and technology strategy and talent development, including the identification and organization of research problems, research institutions and practices, innovation ecosystems, and graduate education. We advocate Problemology as a counterpart to Methodology and explore how a stronger focus on research questions can improve scientific inquiry.
Our approach combines theoretical reasoning with experimental validation and encourages dialogue between technical innovation and historical scholarship. We aim to solve practical systems problems while developing a deeper understanding of computing science and technology.
SELECTED PUBLICATIONS
Selected Publications
Ten representative works spanning technical research and broader scholarly inquiry.
RESEARCH TRAJECTORIES
Research Trajectories
Memory concurrency, data locality, and memory-system optimization
A continuing investigation of how memory capacity, access concurrency, and data locality influence performance, with implications for modeling and optimization.
- C²-boundSC · 2015
Models manycore performance using capacity and concurrency.
- CaLIEEE TBigData · 2018
Extends locality analysis to account for concurrent accesses.
- LPMIEEE TPDS · 2019
Guides optimization through layered demand–supply matching.
Unifying question: How can we explain and reduce memory-related performance bottlenecks beyond cache hit rates and conventional data-reuse metrics?
Measuring and mitigating interference among co-located applications
From system-level interference measurement and resource scheduling toward a more general analysis of QoS, tail latency, and throughput.
- Ah-QHPCA · 2023
Early work on system entropy, interference measurement, and scheduling.
- Suppressing the InterferenceIEEE TPDS · 2024
Develops more general theorems, metrics, and strategies building on that work.
Documented extension: The TPDS paper explicitly identifies the conference work as preliminary and develops more general principles.
Advanced Computer Architecture and Systems
8 works · Models, architecture, and systemsAh-Q: Quantifying and Handling the Interference within a Datacenter from a System Perspective
Quantifies co-located workload interference using system entropy and coordinates resource sharing and isolation.
Suppressing the Interference Within a Datacenter: Theorems, Metric and Strategy
Connects datacenter interference, resource management capability, and service quality through theorems, metrics, and suppression strategies.
C²-bound: A Capacity and Concurrency Driven Analytical Model for Manycore Design
Builds a manycore performance model that accounts for memory capacity and access concurrency.
CaL: Extending Data Locality to Consider Concurrency for Performance Optimization
Extends data locality to include concurrency-aware performance optimization.
LPM: A Systematic Methodology for Concurrent Data Access Pattern Optimization from a Matching Perspective
Uses layered matching between memory demand and supply to improve performance.
Planaria: Pattern Directed Cross-page Composite Prefetcher
Learns within-page and cross-page access patterns to improve prefetching accuracy and coverage.
HCDA: From Computational Thinking to a Generalized Thinking Paradigm
Brings historical, computational, data-driven, and structural thinking together as a design framework.
Sapling: Quantifying and Measuring the Maturity of the RISC-V Software Ecosystem
Introduces multidimensional metrics for RISC-V software ecosystem maturity and identifies ecosystem gaps.
Historical Inquiry & Research Strategy
2 works · Computing history and ProblemologySelected Episodes in the Evolution of Computing Devices and Reflections on Establishing Source-Critical Studies of Computing
计算装置历史演变选析及对构建“计算技术考据学”的若干思考
Uses historical computing devices to examine source-critical methods and the evolution of computing technology.
Building Problemology as a Counterpart to Methodology: A Case Study of Computer Science
构建与方法学对称的问题学——以计算机学科为例
Advocates Problemology as a counterpart to Methodology for identifying and organizing meaningful scientific problems.
BOOKS & SCHOLARSHIP
Books & Other Scholarship
Selected books on systems design, classic papers, and computing performance.
Computer Systems Design Methodology
Computer Systems Design Methodology: History, Computation, Data, Structure
Yuhang Liu and Yungang Bao · Higher Education Press · 2022
Foundational Papers in Computer Systems
Foundational Classic Papers in Computer Systems: Readings and Analyses
Yuhang Liu and Yungang Bao · China Machine Press · 2023 / revised edition 2025
Classical & Quantum Parallelism
Classical and Quantum Parallelism: Unlocking Potential Computing Performance
Yuhang Liu · China Machine Press · 2024
Engineering Strategy 2040
China Engineering Science and Technology Development Strategy 2040: Information and Electronics
Contributor · Science Press · 2025
GROUP LEADER
Group Leader
Research leadership and scholarly profile.
Yuhang Liu
Associate Research Professor · ICT, CAS · CCF Senior Member
Yuhang Liu is an Associate Research Professor and master’s supervisor at the Institute of Computing Technology, Chinese Academy of Sciences, and a Senior Member of the China Computer Federation (CCF). His research centers on computer architecture, performance modeling, memory-level concurrency, storage systems, and datacenter resource management. His work includes C²-bound, CaL, LPM, and Ah-Q. He also pursues historical and source-critical studies of computing and advocates Problemology as a complement to methodology.
RECENT UPDATES
Recent Updates
Selected research, publications, and academic activities.
A research article on open-source production and new quality productive forces appeared in the Bulletin of the Chinese Academy of Sciences.
An article in CCF Newsletter reviewed Qian Xuesen’s contributions to the development of computing in China.
A historical analysis of Min Naida’s 1955 article on computing appeared in Computing.
The Sapling paper appeared at ICSE 2026, introducing a quantitative approach to RISC-V software ecosystem maturity.
More updatesExpand
People, Exchange & Collaboration
We welcome interest in advanced computer architecture and systems, historical studies of computing, and science and technology strategy. Learn more about the group, admissions, and collaboration.
Admissions & CollaborationView details
We welcome motivated students through graduate admission, recommendation-based admission, and research internships. Applicants should demonstrate:
- Enthusiasm and determination for research, integrity, self-discipline, and sound physical and mental health.
- Strong programming ability.
- A solid foundation and practical experience in computer systems and computer organization; competition experience such as the Loongson Cup is a plus.
- Strong analytical ability.
- Good command of English.
- Clear oral and written communication skills.
Contactliuyuhang@ict.ac.cn