Advanced search
1 file | 2.08 MB

Performance and programming environment of a combined GPU/FPGA desktop

Author
Organization
Abstract
The performance and the versatility of today's PCs exceeds many times the power of the fastest number crunchers in the 90s. Yet the computational hunger of many scientific applications has led to the development of GPU- and FPGA-accelerator cards. In this paper the programming environment and the performance analysis of a super desktop with a combined GPU/FPGA architecture is presented. A unified roofline model is used to compare the performance of the GPU and the FPGA taking into account the computational intensity of the algorithm and the resource consumption. The model is validated by two image processing kernels which are compiled using OpenCL for the GPU and a C-to-VHDL compiler for the FPGA. It is shown that an FPGA compiler outperforms handwritten code and is highly productive, but also uses more resources. While both the GPU and FPGA excel in particular applications, both devices suffer from the limited I/O bandwidth to the processor.
Keywords
FPGA, performance modeling, programming environment, GPU, desktop

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 2.08 MB

Citation

Please use this url to cite or link to this publication:

Chicago
da Silva, Bruno, An Braeken, Erik D’Hollander, Abdellah Touhafi, Jan G Cornelis, and Jan Lemeire. 2013. “Performance and Programming Environment of a Combined GPU/FPGA Desktop.” In Transition of HPC Towards Exascale Computing, ed. Erik D’Hollander, Jack J Dongarra, Ian T Foster, Lucio Grandinetti, and Gerhard R Joubert, 24:177–193. Amsterdam, The Netherlands: IOS Press.
APA
da Silva, B., Braeken, A., D’Hollander, E., Touhafi, A., Cornelis, J. G., & Lemeire, J. (2013). Performance and programming environment of a combined GPU/FPGA desktop. In E. D’Hollander, J. J. Dongarra, I. T. Foster, L. Grandinetti, & G. R. Joubert (Eds.), Transition of HPC towards exascale computing (Vol. 24, pp. 177–193). Amsterdam, The Netherlands: IOS Press.
Vancouver
1.
da Silva B, Braeken A, D’Hollander E, Touhafi A, Cornelis JG, Lemeire J. Performance and programming environment of a combined GPU/FPGA desktop. In: D’Hollander E, Dongarra JJ, Foster IT, Grandinetti L, Joubert GR, editors. Transition of HPC towards exascale computing. Amsterdam, The Netherlands: IOS Press; 2013. p. 177–93.
MLA
da Silva, Bruno, An Braeken, Erik D’Hollander, et al. “Performance and Programming Environment of a Combined GPU/FPGA Desktop.” Transition of HPC Towards Exascale Computing. Ed. Erik D’Hollander et al. Vol. 24. Amsterdam, The Netherlands: IOS Press, 2013. 177–193. Print.
@incollection{4227346,
  abstract     = {The performance and the versatility of today's PCs exceeds many times the power of the fastest number crunchers in the 90s. Yet the computational hunger of many scientific applications has led to the development of GPU- and FPGA-accelerator cards. In this paper the programming environment and the performance analysis of a super desktop with a combined GPU/FPGA architecture is presented. A unified roofline model is used to compare the performance of the GPU and the FPGA taking into account the computational intensity of the algorithm and the resource consumption. The model is validated by two image processing kernels which are compiled using OpenCL for the GPU and a C-to-VHDL compiler for the FPGA. It is shown that an FPGA compiler outperforms handwritten code and is highly productive, but also uses more resources. While both the GPU and FPGA excel in particular applications, both devices suffer from the limited I/O bandwidth to the processor.},
  author       = {da Silva, Bruno and Braeken, An and D'Hollander, Erik and Touhafi, Abdellah and Cornelis, Jan G and Lemeire, Jan},
  booktitle    = {Transition of HPC towards exascale computing},
  editor       = {D'Hollander, Erik and Dongarra, Jack J and Foster, Ian T and Grandinetti, Lucio and Joubert, Gerhard R},
  isbn         = {9781614993230},
  issn         = {0927-5452},
  keyword      = {FPGA,performance modeling,programming environment,GPU,desktop},
  language     = {eng},
  pages        = {177--193},
  publisher    = {IOS Press},
  series       = {Advances in Parallel Computing},
  title        = {Performance and programming environment of a combined GPU/FPGA desktop},
  url          = {http://dx.doi.org/10.3233/978-1-61499-324-7-177},
  volume       = {24},
  year         = {2013},
}

Altmetric
View in Altmetric