About
I am a final-year undergraduate in Microelectronics at the South China University of Technology, advised by Prof. Simei Yang, and an exchange student at HKUST, where I am a research intern in Prof. Yuan Xie's FACT Lab. My previous research includes a GCN accelerator and a shared-LLC management policy. My research interests include architecture-level power/performance modeling, CPU microarchitecture design, and hardware-software co-design for emerging AI workloads.
Publications
View All →FlexVector: A SpMM Vector Accelerator with Flexible VRF for GCNs on Varying-Sparsity Graphs
Bohan Li, Shengmin Li, Xinyu Shi, Enyi Yao, Francky Catthoor, Simei Yang†
IEEE Transactions on Computers (IF: 4.9, Accepted)
FlexVector is a domain-specific vector processor that accelerates SpMM for GCN inference on power-law graphs via a software-managed flexible VRF, row-wise product dataflow, and coarse-grained ISA, achieving high utilization and reduced memory traffic under varying sparsity.
Not All Reuse Is Worth Keeping: Context-Aware Prefetch Retention for Shared Caches
Bohan Li, Xiangfeng Sun, Ceyu Xu, Yuan Xie†
HPCA 2027 (under rebuttal)
KEEPER is a shared-LLC prefetch-retention policy that separates per-fill reuse qualification from epoch-level source-core selection: a prefetch keeps its reuse-derived priority only when its line is reuse-qualified and its source core currently shows both high prefetch service and frequent retirement blocking. Under rebuttal at HPCA 2027.
News
FlexVector is accepted by IEEE Transactions on Computers 🎉
Joined the Hong Kong University of Science and Technology as an exchange student and research intern in Prof. Yuan Xie's FACT Lab
Our work has Won the National award 🎉 see https://github.com/brandon-lee-bo/fpgachina2025_47488
Starting my University life at the South China University of Technology
