1. Personal Profile

Liuyang Zhang, Ph.D., Assistant Researcher, and Master's Supervisor. I have long been engaged in research on MRAM integrated circuit design, RISC-V processor microarchitecture, and edge accelerator chips for AI large models. He has participated in several major research projects funded by the Ministry of Science and Technology, Beijing Municipality, and Guangdong Province. He has published multiple papers, co-authored one book ("Analysis and Verification of Open-Source RISC-V Processor Architectures", Tsinghua University Press), applied for several invention patents, and one authorized invention patent. He has contributed to the development of two industrial processors based on RISC-V architecture and led the design and tape-out of two SRAM-based computing-in-memory chips. Currently, he supervises four graduate students, has assisted in mentoring two doctoral students and three master’s students, and has guided students to win awards, including an Outstanding Student Paper Award at an international conference and a SUSTech Climbing Program grant for undergraduate research projects.

2. Research Areas

2.1 Computing-in-Memory Architectures

With the rapid development of data-intensive applications such as artificial intelligence, traditional computer architectures increasingly struggle to meet the computing demands due to power and memory bottlenecks. Computing-in-memory (CIM) architectures, which integrate "storage" with parts of "computation," significantly reduce the time and energy consumption caused by data movement. CIM architectures are typically classified into analog and digital types, differing in computing parallelism, power consumption, and accuracy. Analog CIM is more suitable for power-constrained scenarios, while digital CIM excels in applications requiring higher precision. Our research focuses on achieving higher computing accuracy through advanced data encoding, improving energy efficiency via optimized input data encoding methods, and enhancing parallelism through innovative computing paradigms. Specific research mediums include SRAM, DRAM, and MRAM technologies.

2.2 Magnetic RAM (MRAM) Circuits

Spin, as an intrinsic property of electrons, plays a vital role in information storage. MRAM, a product of spintronics, offers excellent read/write speed, endurance, retention, non-volatility, reliability, radiation resistance, and CMOS compatibility, making it a promising candidate for general-purpose memory. The first-generation MRAM is already used in fields such as aerospace, aircraft control systems, and low-power IoT scenarios. The next-generation MRAM is rapidly commercializing, and key challenges to be tackled include improving reliability under PVT variations, storage density, data interfaces, error correction coding, memory architecture, and testing.

2.3 RISC-V Processor Microarchitecture

RISC-V is an open-source processor microarchitecture that allows anyone to develop processors based on it without licensing fees, providing a pathway for China's independent control over processors. Our research focuses on designing high-performance processor microarchitectures (e.g., superscalar out-of-order execution), architectures for real-time applications, and microarchitecture-level hardware security.

2.4 Computing Optimization for Large AI Models

The emergence of large language models has propelled AI into a new phase, with attention mechanisms being the core of these models. However, attention computation accounts for a significant portion of the workload. Our research addresses reducing memory consumption and data movement in large language model computation on CIM or other architectures.

3. Admissions Information

We recruits approximately four master's students annually, including domestic and international students. We emphasizes that students should not only have a solid theoretical foundation but also strong practical and hands-on skills. For integrated circuit design, every student has the opportunity to participate in a tape-out process during their studies.

For students with different career aspirations, we tailor training plans. Students aiming to work after graduation are provided with internship opportunities and participation in major projects, while students intending to pursue PhD degree receive support for international exchanges or joint training with renowned scholars in the field, as well as opportunities to publish high-quality papers. The laboratory fosters a friendly and collaborative environment. we welcome students with dreams, strong planning abilities, critical thinking, and the courage to practice. 

4. Educational Background

2014.09–2019.11: Ph.D., Beihang University

2015.09–2016.09: Joint Ph.D. Program, University of Montpellier / CNRS LIRMM Laboratory, France

2011.09–2014.06: M.Sc., Jiangxi University of Science and Technology

2007.09–2011.07: B.Sc., ***Information Engineering University

5. Professional Experience

2024.07–Present:  Assistant Researcher, Hangzhou International Innovation Institute, Beihang University

2023.07–2024.07: Research Assistant Professor, Southern University of Science and Technology

2020.04–2023.07: Assistant Researcher, Peng Cheng Laboratory

2016.09–2017.04: Intern, Huawei Hangzhou Research Institute, Storage Media Application Laboratory

6. Contact Me

Email: liuyang.zhang@buaa.edu.cn

Office Phone: 0571-28881220

Office Address: 166 Shuanghongqiao Street, Pingyao Town, Yuhang District, Hangzhou, Zhejiang Province, China