Advertisement

Assessing Kokkos Performance on Selected Architectures

  • Chang PhuongEmail author
  • Noman Saied
  • Craig Tanis
Conference paper
  • 24 Downloads
Part of the Communications in Computer and Information Science book series (CCIS, volume 1087)

Abstract

Performance Portability frameworks allow developers to write code for familiar High-Performance Computing (HPC) architecture and minimize development effort over time to port it to other HPC architectures with little to no loss of performance. In our research, we conducted experiments with the same codebase on a Serial, OpenMP, and CUDA execution and memory space and compared it to the Kokkos Performance Portability framework. We assessed how well these approaches meet the goals of Performance Portability by solving a thermal conduction model on a 2D plate on multiple architectures (NVIDIA (K20, P100, V100, XAVIER), Intel Xeon, IBM Power 9, ARM64) and collected execution times (wall-clock) and performance counters with perf and nvprof for analysis. We used the Serial model to determine a baseline and to confirm that the model converges on both the native and Kokkos code. The OpenMP and CUDA models were used to analyze the parallelization strategy as compared to the Kokkos framework for the same execution and memory spaces.

Keywords

Performance Portability OpenMP CUDA Kokkos High-Performance Computing HPC Parallel programming 

References

  1. 1.
    Buchanan, J.L., Turner, P.R.: Numerical Methods and Analysis. McGraw-Hill, New York (1992)Google Scholar
  2. 2.
    De Melo, A.C.: The new Linux ‘perf’ tools. In: Slides from Linux Kongress, vol. 18 (2010)Google Scholar
  3. 3.
    Edwards, H.C., Trott, C.R.: Kokkos: enabling performance portability across manycore architectures. In: 2013 Extreme Scaling Workshop (XSW), pp. 18–24. IEEE (2013)Google Scholar
  4. 4.
    Lopez, M.G., et al.: Towards achieving performance portability using directives for accelerators. In: 2016 Third Workshop on Accelerator Programming Using Directives (WACCPD), pp. 13–24. IEEE (2016)Google Scholar
  5. 5.
    Martineau, M., McIntosh-Smith, S., Boulton, M., Gaudin, W., Beckingsale, D.: A performance evaluation of Kokkos & Raja using the TeaLeaf mini-app. In: The International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2015 (2015)Google Scholar
  6. 6.
    Martineau, M., McIntosh-Smith, S., Gaudin, W.: Assessing the performance portability of modern parallel programming models using TeaLeaf. Concurrency Comput.: Practice Exp. 29(15), e4117 (2017)CrossRefGoogle Scholar
  7. 7.
    NVIDIA developer (2019). https://developer.nvidia.com/. Accessed 16 Jan 2019
  8. 8.
    OpenMP (2019). https://www.openmp.org/. Accessed 12 Jan 2019
  9. 9.
    Portability across DOE office of science HPC facilities (2019). http://performanceportability.org/. Accessed 14 Jan 2019
  10. 10.
    RAJA: Managing application portability for next-generation platforms, January 2019. https://computation.llnl.gov/projects/raja-managing-application-portability-next-generation-platforms
  11. 11.
    Tanis, C., Sreenivas, K., Newman, J.C., Webster, R.: Performance portability of a multiphysics finite element code. In: 2018 Aviation Technology, Integration, and Operations Conference, p. 2890 (2018)Google Scholar
  12. 12.
    Videau, B., et al.: BOAST: a metaprogramming framework to produce portable and efficient computing kernels for HPC applications. Int. J. High Perform. Comput. Appl. 32(1), 28–44 (2018)CrossRefGoogle Scholar
  13. 13.
    Wiki, K.: Kokkos: The C++ performance portability programming model (2019). https://github.com/kokkos/kokkos/wiki/. Accessed 14 Jan 2019

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.University of TennesseeChattanoogaUSA

Personalised recommendations