AI for Science Infra

• WIP

HPC Performance Tools

• TraceFlow: Efficient Trace Analysis for Large-Scale Parallel Applications via Interaction Pattern-Aware Trace Distribution. SC. 2025.
• Leveraging Graph Analysis to Pinpoint Root Causes of Scalability Issues for Parallel Applications. IEEE TPDS. 2025.
• Graph-Centric Performance Analysis for Large-Scale Parallel Applications. IEEE TPDS. 2024.
• BoostN: Optimizing Imbalanced Neighborhood Communication on Homogeneous Many-Core System. ICPP. 2024.
• PerFlow: A Domain Specific Framework for Automatic Performance Analysis of Parallel Applications. PPoPP. 2022.
• Vapro: Performance Variance Detection and Diagnosis for Production-Run Parallel Applications. PPoPP. 2022.
• ScalAna: Automating Scaling Loss Detection with Graph Analysis. SC. 2020.
• Detecting Performance Variance for Parallel Applications Without Source Code. IEEE TPDS. 2022.

AI Infra

• UltraAttn: Efficiently Parallelizing Attention through Hierarchical Context-Tiling. SC. 2025.
• mTuner: Accelerating Parameter-Efficient FineTuning on Multi-GPU Servers with Elastic Tensor. ATC. 2025.
• FlashTensor: Optimizing Tensor Programs by Leveraging Fine-grained Tensor Property. PPoPP. 2025.
• WiseGraph: Optimizing GNN with Joint Workload Partition of Graph and Operations. EuroSys. 2024.
• Efficient Inference for Pruned CNN Models on Mobile Devices With Holistic Sparsity Alignment. IEEE TPDS. 2024.
• PUZZLE: Efficiently Aligning Large Language Models through Light-Weight Context Switch. USENIX ATC. 2024.

Monographs

• Research on Key Technologies of Performance Analysis and Optimization for Large-Scale Parallel Applications. Tsinghua University Press (清华大学出版社). 2024.
• Performance Analysis of Parallel Applications for HPC. Springer Nature. 2023.