Scalable Fast Multipole Method for Large-Scale Electromagnetic Scattering Problems on Heterogeneous CPU-GPU Clusters

被引：4
|
作者：
Vinh Dang [1 ]
Tran, Nghia [1 ]
Kilic, Ozlem [1 ]
机构：
[1] Catholic Univ Amer, Dept Elect Engn & Comp Sci, Washington, DC 20064 USA
来源：
IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS | 2016年 / 15卷
关键词：
Fast multipole method; graphics processing unit; heterogeneous clusters; message passing interface (MPI); portable operating system interface (POSIX); ALGORITHM;
D O I：
10.1109/LAWP.2016.2537779
中图分类号：
TM [电工技术]; TN [电子技术、通信技术];
学科分类号：
0808 ; 0809 ;
摘要：
This letter investigates a hybrid framework for the solution of large-scale electromagnetic scattering problems by using the fast multipole method (FMM) on a heterogeneous CPU-GPU system. Enabling the use of both CPU and GPU resources available in the cluster allows for solving significantly larger problem sizes than using only CPU or GPU resources. The performance is evaluated on a 13-node cluster equipped with NVIDIA Tesla M2090 GPUs. The experimental results demonstrate that our FMM implementation on CPU-GPU is up to 72.3x faster than that of the 12-core eight-node CPU implementation. The scalability of the CPU-GPU implementation is very close to the theoretical expectations.
引用
下载
收藏
页码：1807 / 1810
页数：4
相关论文
共 50 条
  • [41] MLaaS in the Wild: Workload Analysis and Scheduling in Large-Scale Heterogeneous GPU Clusters
    Weng, Qizhen
    Xiao, Wencong
    Yu, Yinghao
    Wang, Wei
    Wang, Cheng
    He, Jian
    Li, Yong
    Zhang, Liping
    Lin, Wei
    Ding, Yu
    PROCEEDINGS OF THE 19TH USENIX SYMPOSIUM ON NETWORKED SYSTEMS DESIGN AND IMPLEMENTATION (NSDI '22), 2022, : 945 - 960
  • [42] Fast multipole boundary element method for large-scale numerical computation
    Department of Mechanics, Huazhong University of Science and Technology, Wuhan 430074, China
    J. Mech. Strength, 2008, 2 (229-234):
  • [43] An automatic load-balanced parallel multilevel fast multipole method for large scale electromagnetic scattering problem
    Dong, J.
    Chai, S. L.
    Mao, J. J.
    2005 ASIA-PACIFIC MICROWAVE CONFERENCE PROCEEDINGS, VOLS 1-5, 2005, : 1795 - 1798
  • [44] Solution of large-scale scattering problems with the multilevel fast multipole algorithm parallelized on distributed-memory architectures
    Ergul, Ozgur
    Gurel, Levent
    2007 22ND INTERNATIONAL SYMPOSIUM ON COMPUTER AND INFORMATION SCIENCES, 2007, : 63 - 68
  • [45] AIM: Adaptive integral method for solving large-scale electromagnetic scattering and radiation problems
    North American Aircraft, Rockwell International Corporation, Los Angeles
    CA, United States
    不详
    CA
    91360, United States
    不详
    CA
    90064, United States
    不详
    CA
    90740-7644, United States
    Radio Sci, 5 (1225-1251):
  • [46] AIM: Adaptive integral method for solving large-scale electromagnetic scattering and radiation problems
    Bleszynski, E
    Bleszynski, M
    Jaroszewicz, T
    RADIO SCIENCE, 1996, 31 (05) : 1225 - 1251
  • [47] Fast-multipole accelerated regularized meshless method for large-scale isotropic heat conduction problems
    Wang, Zhaoyang
    Gu, Yan
    Chen, Wen
    INTERNATIONAL JOURNAL OF HEAT AND MASS TRANSFER, 2016, 101 : 461 - 469
  • [48] Broadband Multilevel Fast Multipole Algorithm For Large-Scale Problems With Nonuniform Discretizations
    Ergul, Ozgur
    Karaosmanoglu, Bariscan
    Takrimi, Manouchehr
    Erturk, Vakur B.
    2016 URSI INTERNATIONAL SYMPOSIUM ON ELECTROMAGNETIC THEORY (EMTS), 2016, : 284 - 287
  • [49] GPU Cluster Implementation of FMM-FFT for Large-Scale Electromagnetic Problems
    Vinh Dang
    Nguyen, Quang M.
    Kilic, Ozlem
    IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS, 2014, 13 : 1259 - 1262
  • [50] Fast multipole method based model order reduction for large scattering problems
    Lukashevich, Dzianis
    Tuncer, Ozgur
    Russer, Peter
    2006 IEEE MTT-S INTERNATIONAL MICROWAVE SYMPOSIUM DIGEST, VOLS 1-5, 2006, : 1057 - 1060
12345