基本信息

 

刘 松

工学博士,计算机科学与技术学院副教授,博士生导师。

隶属高性能计算团队,新型计算机研究所。

研究方向:并行程序性能优化、高性能计算、AI赋能和优化技术、边缘计算等。

联系方式

办公室:兴庆校区西一楼405

             创新港泓理楼-6042

邮箱:liusong@mail.xjtu.edu.cn

站点计数器

科研项目

国家级和省部级纵向科研项目:

  • 国家重点研发计划项目子课题“基于国产芯群架构的机器学习算子优化技术的研究”负责人2023.03-2026.02
  • 国家自然科学基金青年科学基金项目模板计算程序在GPU上基于多面体模型的编译优化关键技术研究”,负责2021.01-2023.12
  • 国家级项目子课题负责2020.11-2025.11
  • 地理信息工程国家重点实验室2019年度开放基金课题GPU架构上GIS应用的并行优化技术研究”负责2020.01-2022.10
  • 陕西省自然科学基础研究计划一般项目(青年)面向加速器的Stencil计算性能优化方法研究”,负责2020.01-2021.12
  • 国防基础科研核科学挑战专题模式驱动的商用CPUGPU数值内核浮点优化方法子课题主要参与2019.09-2020.12
  • 国家自然科学基金重点项目“商用客机气动噪声大规模并行计算的建模、算法与软件”,主要参与2017.01-2019.12
  • 国家自然科学基金培育项目“面向P量级系统计算节点层次存储架构和并行粒度优化的多级循环分块方法研究”,主要参与2014.01-2016.12

 

企业合作横向项目:

  • 华为技术有限公司横向项目子课题“通过对推理、训练任务的分割、调度多种异构算力芯片实现计算效率最优的技术研究”,负责2021.10-2022.12
  • 华为技术有限公司横向课题“实时海量低压缩数据传输及GPU云化系统调度”,主要参与人2022.01-2022.12
  • 杭州海康威视数字技术股份有限公司横向课题“Ceph存储技术深度开发”,负责2019.07-2020.01
  • 西安汉格尚华网络科技有限公司技术(专利)转让“一种基于机器学习的循环分块大小选择方法”,负责人,2020.12

发表论文

  • Weiduo Chen, Xiaoshe Dong*, Xinhang Chen, Song Liu, Qin Xia, Qiang Wang[J]. pommDNN: Performance optimal GPU memory management for deep neural network training[J]. Future Generation Computer Systems. 2024, vol.152, pp.160-169.
  • Chi Zhang, Fangxing Yu, Shiqiang Nie, Wei Tang, Fei Liu, Song Liu*, Weiguo Wu*. Amphisbaena: A Novel Persistent Buffer Management Strategy to Improve SMR Disk Performance[J]. Applied Sciences. 2024, 14(630): 1-26.
  • Xiaoya An, Ziming Wang, Ding Wang, Song Liu*, Cheng Jin, Xinpeng Xu, Jianjun Cao. STRP-DBSCAN: A Parallel DBSCAN Algorithm Based on Spatial-Temporal Random Partitioning for Clustering Trajectory Data[J]. Applied Sciences. 2023, 13(20), 11122: 1-23.
  • Chi Zhang, Song Liu, Fangxing Yu, Menghan Li, Wei Tang, Fei Liu, Weiguo Wu*. Balloon: An Elastic Data Management Strategy for Interlaced Magnetic Recording[J]. Applied Sciences. 2023, 13(17), 9767: 1-19.
  • Song Liu, Zengyuan Zhang, Weiguo Wu. DHTS: A Dynamic Hybrid Tiling Strategy for Optimizing Stencil Computation on GPUs[J]. IEEE Transactions on Computers. 2023, vol.72, no.10, pp.2795-2807.
  • Song Liu, Xinhe Wan, Zengyuan Zhang, Bo Zhao, Weiguo Wu. TurboStencil: You Only Compute Once for Stencil Computation[J]. Future Generation Computer Systems. 2023, vol.146, pp.260-272.
  • Xiangjun Zhang, Weiguo Wu, Zhihe Zhao, Jinyu Wang, Song Liu. RMDDQN-Learning: Computation Offloading Algorithm Based on Dynamic Adaptive Multi-Objective Reinforcement Learning in Internet of Vechicles[J]. IEEE Transactions on Vehicular Technology. 2023 (Accepted).
  • 屈彬, 刘松, 张增源,马洁, 伍卫国. 一种六边形循环分块的Jacobi计算优化方法[J]. 软件学报, 2023(已录用).
  • Chi Zhang, Shiqiang Nie, Jinyu Wang, Song Liu, Weiguo Wu. MCB: A Multidevice Cooperative Buffer Management Strategy for Boosting the Write Performance of the SSD-SMR Hybrid Storage[J]. The Journal of Supercomputing. 2023.
  • Song Liu, Chen Zhang, Shiqiang Nie, Keqiang Duan, Weiguo Wu. PC-Allocation: Performance Cliff-aware Two-level Cache Resource Allocation Scheme for Storage System[J]. Applied Sciences. 2023, 13(6), 3556: 1-12.
  • Song Liu, Shiyuan Yang, Hanze Zhang, Weiguo Wu. A Federated Learning and Deep Reinforcement Learning-Based Method with Two-Agent for Computation Offload[J]. Sensors. 2023, 23(4), 2243.
  • Song Liu, Xiong Wang, Longshuo Hui, and Weiguo Wu. Blockchain-Based Decentralized Federated Learning Method in Edge Computing Environment[J]. Applied Sciences. 2023, 13(3):1677.
  • Song Liu, Jie Ma, Chenyu Zhao, Xinhe Wan, Weiguo Wu. LFWS: Long-Operation First Warp Scheduling Algorithm to Effectively Hide the Latency for GPUs[J]. IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences. 2023,106(8).
  • Song Liu, Yuxiang Chai, Longshuo Hui, Weiguo Wu. Blockchain-Based Anonymous Authentication in Edge Computing Environment[J]. Electronics, 2023, 12(1): 219.
  • Xiangjun Zhang, Weiguo Wu, Song Liu, Jinyu Wang. An Efficient Computation Offloading and Resource Allocation Algorithm in RIS Empowered MEC[J]. Computer Communications, 2023, 197: 113-123.
  • Xiangjun Zhang, Weiguo Wu, Jinyu Wang, Song Liu. BiLSTM-based Federated Learning Computation Offloading and Resource Allocation Algorithm in MEC[J]. ACM Transactions on Sensor Networks, 2023 (accepted).
  • Xinpeng Xu, Hongfei Tao, Weiguo Wu, Song Liu. An Instant Discovery Method for Companion Vehicals based on Incremental and Parallel Calculation[J]. Physica A, 610 (2023), 128420.
  • Song Liu, Chenyu Zhao, Xiaoya An, Kai Gao, Yingjie Ji, Weiguo Wu. Accelerating Ordinary Kriging Interpolation Algorithm on GPUs[C]//Proceedings of IEEE International Conference on Advances in Electrical Engineering and Computer Applications, Dalian, China, August 20-21, 2022: 225-231.  
  • Jinyu Wang, Yifei Kang, Yiwen Li, Weiguo Wu, Song Liu, Longxiang Wang. Hexagonal Tiling based Multiple FPGAs Stencil Computation Acceleration and Optimization Methodology[C]// Proceedings of IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA), 2021: 697-705.
  • Wei Song, Song Liu, Xiaochun Wang, Weiguo Wu. An Improved Sparrow Search Algorithm[C]// Proceedings of 18th IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA), Exeter, UK: IEEE, December 17-19, 2020: 537-543.
  • 崔元桢, 刘松, 王倩, 伍卫国. 格子玻尔兹曼方法计算程序的循环优化技术研究[J]. 计算机学报, 2020, 43(6): 1086-1102. 
  • 柴晓菲, 刘松, 屈彬, 王倩, 伍卫国. 向量化友好的循环分块因子选择算法[J]. 计算机工程与应用, 2020, 56(15): 37-42.
  • Bin Qu, Song Liu, Hailong Huang, Jiajun Yuan, Qian Wang, Weiguo Wu. Accelerating Lattice Boltzmann Method by Fully Exposing Vectorizable Loops[C]//Proceedings of International Conference on Algorithms and Architectures for Parallel Processing (ICA3PP), Melbourne, Australia: Springer, December 9-11, 2019: 107-121. 
  • Song Liu, Yuanzhen Cui, Nianjun Zou, Wenhao Zhu, Dong Zhang, Weiguo Wu. Revisiting the Parallel Strategy for DOACROSS Loops[J]. Journal of Computer Science and Technology, 2019, 34(2): 456-475.
  • Song Liu, Yuanzhen Cui, Qing Jiang, Qian Wang, Weiguo Wu. An Efficient Tile Size Selection Model Based on Machine Learning[J]. Journal of Parallel and Distributed Computing, 2018, 121(2018): 27-41. 
  • Yuanzhen Cui, Song Liu, Nianjun Zou, Weiguo Wu. A Dynamic Parallel Strategy for DOACROSS Loops[C]//Proceedings of the International Conference on High Performance Computing in Asia-Pacific Region (HPC Asia), Tokyo, Japan: ACM, January 28-31, 2018: 108-115.
  • 刘松, 赵博, 蒋庆, 伍卫国. 一种面向循环优化和非规则代码段的粗粒度半自动并行化方法[J]. 计算机学报, 2017, 40(9): 2127-2147.
  • 刘松, 伍卫国, 赵博, 蒋庆. 面向局部性和并行优化的循环分块技术[J]. 计算机研究与发展, 2015, 52(5): 1160-1176.
  • Song Liu, Nianjun Zou, Yuanzhen Cui, Weiguo Wu. Accelerating the Parallelization of Lattice Boltzmann Method by Exploiting the Temporal Locality[C]//Proceedings of 15th IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA), Guangzhou, China: IEEE, December 13-15, 2017: 1186-1193.
  • Song Liu, Xiao Xie, Yuanzhen Cui, Weiguo Wu. An Efficient Locality-Aware Task Assignment Algorithm for Minimizing Shared Cache Contention[C]//Proceedings of 18th International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT), Taipei, Taiwan: IEEE, December 18-20, 2017: 44-51.

专利

 

  • 刘松,伍卫国,屈彬,王倩,马洁.一种开发循环代码中潜在可向量化循环的方法[P].申请号:201911243343.X.
  • 刘松,王倩,伍卫国,屈彬,张增源.一种针对模板计算菱形分块并行优化方法[P].申请号:202010136404.9.
  • 刘松,伍卫国,谢骁,屈彬.一种多核系统基于局部性量化的并行任务分配调度方法[P].申请号:201910894969.0.
  • 刘松,伍卫国,柴晓菲,屈彬,马洁.一种多维循环自动向量化分块因子分块方法及装置[P].专利号:ZL202010706144.4,授权号:CN111857727B.
  • 伍卫国,刘松,崔元桢,蒋庆,谢骁,邹年俊.一种基于机器学习的循环分块大小选择方法[P].专利号:ZL201710175139.3,授权号:CN106990995B.
  • 伍卫国,刘松,谢骁,崔元桢,邹年俊.多级共享高速缓冲存储器架构下的任务分配方法与系统[P].专利号:ZL201711298943.7,授权号:CN108132834B.
  • 伍卫国,刘松,邹年俊.一种利用时间局部性的格子玻尔兹曼方法并行加速方法[P].专利号:201711297745.9,授权号:CN108038304B.
  • 伍卫国,崔元桢,刘松,柴晓菲,聂世强,邹年俊.一种DOACROSS循环的并行优化方法[P].专利号:ZL201610851036.X,授权号:CN106445666B.
  • 聂世强,伍卫国,崔金华,刘松,胡壮,薛尚山,邹年俊.一种针对由异构存储设备组成的对象存储系统的对象分布算法[P].专利: ZL201610933662.3,授权号:CN106527982B.
  • 针对多重嵌套循环优化的源到源代码转换系统V1.0[CP/CD].著作权登记号:2016SR250826,登记日期:201656.

招生信息

招收计算机专业的学术型博士生、学术型和专业型硕士生、以及本科生实习生。

 

招生要求:

1. 对计算机专业,特别是程序性能优化、高性能计算机体系结构、AI赋能优化技术等相关领域,有较强的科研兴趣;

2. 具备良好的计算机专业知识和编程能力,熟悉和掌握CC++JavaPython等一种以上的编程语言,熟悉Linux操作系统,了解计算机体系结构、操作系统等基本原理,英语能力良好;

3. 具备良好的学习和沟通能力,主观能动性强,性格开朗,有团队协作精神。