Kun Li (李琨), MSRA

Kun Li （李琨）
Senior Researcher, Systems and Networking Research Group, Microsoft Research Asia

Research interests:
scientific AI high-performance parallel algorithm & LLM system

Brief Biography

Dr. Kun Li is a Senior Researcher at the Systems and Networking Research Group, Microsoft Research Asia since Jul. 2022. He earned his Ph.D. degree with the State Key Laboratory of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences (ICT, CAS) in 2022. His dissertation, titled Research and Application on Multi-level Discontinuous and Nonlinear Scalability for Massively Parallelism, was recognized with CCF Outstanding Doctoral Dissertation Award and ACM SIGHPC China Outstanding Doctoral Dissertation Award. He has also received the CCF HPC Young Scientist Award, the ACM SIGHPC China Rising Star Award, among others. He is a keynote speaker at the CCF HPCChina 2024 conference, an Executive Member of the CCF Technical Committee on High-Performance Computing and a member of the CCF Technical Committee on Computer Architecture.

Dr. Kun Li has published extensively in top-tier international conferences and journals. Notably, his Cloud4Science series of work has been continuously featured in premier HPC conferences such as SC and PPoPP, and received the Best Paper Award at PPoPP 2024. Currently, He leads the Scientific AI & LLM System project. His research addresses key challenges in scientific reasoning, scalability, and LLM inference efficiency, including breakthroughs in long-text generation and low-latency systems. He actively collaborates with academic and industry partners to bridge HPC and AI to enable scalable, high-performance solutions for scientific intelligence and large-scale AI systems. If you're interested, feel free to join him in exploring these exciting frontiers!

News

[Nov. 2024] Our paper "FlashFFTStencil: Bridging Fast Fourier Transforms to Memory-Efficient Stencil Computations on Tensor Core Units" is accepted by PPoPP'25. Congratulations to Haozhi!

[Nov. 2024] Our paper "Jigsaw: Toward Conflict-free Vectorized Stencil Computation by Tessellating Swizzled Registers" is accepted by PPoPP'25. Congratulations to Yiwei!

[Oct. 2024] Awarded with 2024 ACM SIGHPC 中国新星奖! [More]

[Sep. 2024] Awarded with 2024 CCF高性能计算青年科技人才奖! [More]

[Aug. 2024] Our paper "LONG EXPOSURE: Accelerating Parameter-Efficient Fine-Tuning for LLMs under Shadowy Sparsity" is accepted by SC'24. Congratulations to Tuowei!

[Aug. 2024] Our paper "LoRAStencil: Low-Rank Adaptation of Stencil Computation on Tensor Cores" is accepted by SC'24. Congratulations to Yiwei!

[Mar. 2024] Our paper "ConvStencil: Transform Stencil Computation to Matrix Multiplication on Tensor Cores" wins PPOPP'24 Best Paper Award!

Selected Publications

*: Corresponding author.

[To be appeared] Tuowei Wang, Kun Li *, Donglin Bai, Fusong Ju, Leo Xia, Ju Ren, Yaoxue Zhang, Ting Cao, Mao Yang. Matryoshka: Optimization of Dynamic Diverse Quantum Chemistry Systems via Elastic Parallelism Transformation. [Paper]

[PPOPP'25] Haozhi Han, Kun Li *, Wei Cui, Donglin Bai, Yiwei Zhang, Liang Yuan, Yifeng Chen, Yunquan Zhang, Ting Cao, Mao Yang. FlashFFTStencil: Bridging Fast Fourier Transforms to Memory-Efficient Stencil Computations on Tensor Core Units. [Paper]

[PPOPP'25] Yiwei Zhang, Kun Li *, Liang Yuan, Haozhi Han, Yunquan Zhang, Ting Cao, Mao Yang. Jigsaw: Toward Conflict-free Vectorized Stencil Computation by Tessellating Swizzled Registers. [Paper]

[SC'24] Yiwei Zhang, Kun Li *, Liang Yuan, Jiawen Cheng, Yunquan Zhang, Ting Cao, Mao Yang. LoRAStencil: Low-Rank Adaptation of Stencil Computation on Tensor Cores. [Paper]

[SC'24] Tuowei Wang, Kun Li *, Zixu Hao, Donglin Bai, Ju Ren, Yaoxue Zhang, Ting Cao, Mao Yang. LONG EXPOSURE: Accelerating Parameter-Efficient Fine-Tuning for LLMs under Shadowy Sparsity. [Paper]

[IPDPS'24] Luhan Wang, Haipeng Jia, Lei xu, Cunyang Wei, Kun Li , Xianmeng Jiang, Yunquan Zhang. VNEC: A Vectorized Non-Empty Column Format for SpMV on CPUs.

[PPOPP'24, [Best Paper Award] ] Yuetao Chen, Kun Li *, Yuhao Wang, Donglin Bai, Lei Wang, Lingxiao Ma, Liang Yuan, Yunquan Zhang, Ting Cao, Mao Yang. ConvStencil: Transform Stencil Computation to Matrix Multiplication on Tensor Cores. [Paper]

[ICS'23] Tun Chen, Haipeng Jia, Yunquan Zhang, Kun Li, Zhihao Li, Xiang Zhao, Jianyu Yao. OpenFFT: An Adaptive Tuning Framework for 3D FFT on ARM Multicore CPUs.

[TPDS'23] Hang Cao, Liang Yuan, He Zhang, Yunquan Zhang, Baodong Wu, Kun Li, Shigang Li, Minghua Zhang, Pengqi Lu, and Junmin Xiao. AGCM-3DLF: Accelerating Atmospheric General Circulation Model via 3D Parallelization and Leap-Format.

[HPCC'22] Luhan Wang, Haipeng Jia, Yunquan Zhang, Kun Li, and Cunyang Wei. EgpuIP: An Embedded GPU Accelerated Library for Image Processing.

[HPCC'22] Cunyang Wei, Haipeng Jia, Yunquan Zhang, Kun Li, and Luhan Wang. LBBGEMM: A Load-Balanced Batch GEMM Framework on ARM CPUs.

[IPDPS'22] Kun Li, Liang Yuan, Yunquan Zhang, Yue Yue, and Hang Cao. An Efficient Vectorization Scheme for Stencil Computation. [Paper]

[TPDS'22] Kun Li, Liang Yuan, Yunquan Zhang, and Gongwei Chen. An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering. [Paper]

[SC'21] Kun Li, Liang Yuan, Yunquan Zhang, and Yue Yue. Reducing Redundancy in Data Organization and Arithmetic Calculation for Stencil Computations. [Paper]

[SC'21] Liang Yuan, Hang Cao, Yunquan Zhang, Kun Li, Pengqi Lu, and Yue Yue. Temporal Vectorization for Stencils. [Paper]

[SC'19] Kun Li, Honghui Shang, Yunquan Zhang, Shigang Li, Baodong Wu, Dong Wang, Libo Zhang, Fang Li, Dexun Chen, and Zhiqiang Wei. OpenKMC : a KMC design for hundred-billion-atom simulation using millions of cores on Sunway Taihulight. (Acceptance rate: 22.7%, 78/344) [Paper]

[ISPA'19] Kun Li, Shigang Li, Bei Wang, Yifeng Chen, and Yunquan Zhang. swMD: Performance Optimizations for Molecular Dynamics Simulation on Sunway Taihulight. [Paper]

[JSUPERCOMPUT'19] Kun Li, Shigang Li, Shan Huang, Yifeng Chen, and Yunquan Zhang. FastNBL: fast neighbor lists establishment for molecular dynamics simulation based on bitwise operations. The Journal of Supercomputing (2019): 1-20. [Paper]

[ICPP'18] Junmin Xiao, Shigang Li, Baodong Wu, He Zhang, Kun Li, Erlin Yao, Yunquan Zhang, and Guangming Tan. Communication-Avoiding for Dynamical Core of Atmospheric General Circulation Model. [Paper]

[JCST'17] Kun Li, Haipeng Jia, Ting Cao, and Yunquan Zhang. The Implementation and Optimization of Multidimensional FFT Algorithm on Large-scale Clusters. The Journal of Frontiers of Computer Science and Technology, 2017. [Paper]

[HPCChina'16] Kun Li, Yan Li, Ting Cao, Haipeng Jia, and Yunquan Zhang. An MPI-based 3D FFT Implementation on CPUGPU Heterogeneous Clusters. National Annual Conference on High Performance Computing 2016.

Selected Awards & Position

CCF High-Performance Computing Young Talent Award , 中国计算机学会高性能计算青年科技人才奖
ACM SIGHPC China Rising Star Award , 美国计算机协会SIGHPC中国新星奖
Microsoft Research Asia Science Craftsman
CCF Outstanding Doctoral Dissertation Award , 中国计算机学会优秀博士学位论文奖
ACM SIGHPC China Outstanding Doctoral Dissertation Award , 美国计算机协会SIGHPC中国优秀博士学位论文奖
Microsoft Star of Tomorrow
Executive Committee Member, CCF High-Performance Computing Committee , 中国计算机学会高性能计算委员会执行委员
中国计算机学会(CCF)体系结构委员会委员

Media

Dec.23, 2024. Featured by Microsoft Research, Toward Zettascale Computing: Accelerating Scientific Discovery with the Cloud4Science”.
Feb.24, 2023. Featured by Microsoft Research, 科学匠人 | 李琨：执著于高性能计算研究的“别人家的孩子”.
Jan.10, 2023. Featured by ICT, CAS, 学术科研 | 计算所两篇论文入选2022年“CCF优秀博士学位论文激励计划”.
Jul.20, 2022. Featured by ICT, CAS, 毕业生故事 | 与你相见，千万次不曾放弃.

Talks

Dec.23, 2024. Invited by Honghui Shang@University of Science and Technology of China. Evolution of HPC Paradigm through Unified Matrix Computation: Computing the Future in the AI Era
Dec.22, 2024. Invited by Hong An@University of Science and Technology of China. Evolution of HPC Paradigm through Unified Matrix Computation: Computing the Future in the AI Era
Sep.25, 2024. Keynote @ HPCChina 2024. Evolving the HPC Paradigm with Unified Matrix Computation on AI Accelerators
Sep.20, 2024. Invited by Yan Li@Lenovo Research. A New High-Performance Computing Paradigm on Tensor Core Units
Aug.30, 2024. Invited by Jue Wang @ CNIC, CAS. Evolving the HPC Paradigm with Unified Matrix Computation on AI Accelerators
Aug.29, 2024. Invited by En Shao @ ICT,CAS. Evolving the HPC Paradigm with Unified Matrix Computation on AI Accelerators
Nov.19, 2023. Invited by Chen Ding @ University of Rochester. ConvStencil
Nov.19, 2023. Invited by GuoMeng Studio@University of Chinese Academy of Sciences. Star Roundtable
Oct.28, 2023. Invited by PKU Linux Club@Peking University. AI4Science Salon
Oct.11, 2023. Invited by CCF. CCF SPP Live
May.18, 2023. Invited by Jianfei Chen@Tsinghua University. CCF YEF 2023
Apr.26, 2023. Invited by Liang Yuan@ICT,CAS. The Young Scholars Forum
Mar.23, 2023. Invited by Haisen Zhao@Shandong University. The Young Scholars Forum
Dec.15, 2022. Invited by Jidong Zhai@Tsinghua University. HPC China 2022
Jun.21, 2019. Invited by Xinfu He@China Institute of Atomic Energy. Nuclear Reactors Prototype System Workshop
Dec.24, 2017. Invited by Mingmin Chi@Fudan University. Square Kilometre Array Annual Conference. [Poster]

Last updated on 12/30/2024.