Computation Efficiency Analysis of Multiple GPUs and Multiple CPUs Based Cluster Computing Environments

  • Bongjae Kim
  • Boseon Hong
  • Jeong-Dong KimEmail author
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 536)


Multiple CPUs and multiple GPUs based cluster computing environments are popularly used and very attractive approach in various computing areas because those computing environments provide very high computing performance when compared to a typical single node based computing environments. In this paper, we compare and evaluate the performance of multiple CPUs and multiple GPUs based on cluster computing environment with MPI (Message Passing Interfaces). In the performance evaluations, we evaluate and analyze the performance of sparse matrix-vector multiply (SpMV). SpMV is one of the most widely used operations in many scientific, computational, and mathematical applications. Based on the performance evaluation results, generally, the execution time of SpMV is decreased as the number of CPUs and GPUs increase. However, there were cases that were not so in the case of GPU.


Multiple CPUs Multiple GPUs High performance computing Message passing interface Cluster computing 



This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No.2017R1C1B5017476). The corresponding author is Jeong-Dong Kim.


  1. 1.
    Kim, B., Jung, H.: A case study of data transfer efficiency optimization for GPU- and infiniband-based clusters. In: the 2015 Conference on Research in Adaptive and Convergent Systems, pp. 247–250. ACM, New York (2015)Google Scholar
  2. 2.
    Awan, A.A., Hamidouche, K., Hashmi, J.M., Panda, D.K.: S-caffe: co-designing mpi runtimes and caffe for scalable deep learning on modern gpu clusters. In: ACM Sigplan Notices, vol. 52, no. 8, pp. 193–205 (2017)CrossRefGoogle Scholar
  3. 3.
    Stone, J.E., Saam, J., Hardy, D.J., Vandivort, K. L., Hwu, W.-M.W., Schulten, K.: High performance computation and interactive display of molecular orbitals on GPUs and multi-core CPUs. In: 2nd Workshop on General Purpose Processing on Graphics Processing Units, pp. 9–18. ACM, New York (2009)Google Scholar
  4. 4.
    Kim, B., Jung, J., Min, H., Heo, J., Jung, H.: Performance evaluations of multiple GPUs based on MPI environments. In: the 2017 Conference on Research in Adaptive and Convergent Systems, pp. 303–204. ACM, New York (2017)Google Scholar
  5. 5.
    Panda, D.K., Tomko, K., Schulz, K., Majumdar, A.: The MVAPICH project: evolution and sustainability of an open source production quality MPI library for HPC. In: Workshop on Sustainable Software for Science: Practice and Experiences, Held in Conjunction with Int’l Conference on Supercomputing (2012)Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  1. 1.Division of Computer Science and EngineeringSun Moon UniversityAsan-siSouth Korea
  2. 2.Department of Computer and Electronics Convergence EngineeringSun Moon UniversityAsan-siSouth Korea

Personalised recommendations