Advanced search
1 file | 82.52 KB
Author
Organization
Abstract
Low-power, high-performance computing nowadays relies on accelerator cards to speed up the calculations. Combining the power of GPUs with the flexibility of FPGAs enlarges the scope of problems that can be accelerated. We describe the performance analysis of a desktop equipped with a GPU Tesla 2050 and an FPGA Virtex- 6 LX 240T. The balance between the I/O and the raw peak performance is analyzed using the roofline model. A well-tuned accelerator- based codesign, identifying the parallelism, the computation and data patterns of different classes of algorithms, will enable to maximize the performance of the combined GPU/FPGA system.
Keywords
FPGA, high-level synthesis, GPU, roofline model

Downloads

  • fpga167-dasilva.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 82.52 KB

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 Toolchain of a Combined GPU/FPGA Desktop.” In FPGA  ’13 Proceedings of the ACM/SIGDA International Symposium on Field Programmable Gate Arrays, ed. Brad Hutchings and Vaughn Betz, 274–274. New York, NY, USA: ACM.
APA
da Silva, B., Braeken, A., D’Hollander, E., Touhafi, A., Cornelis, J. G., & Lemeire, J. (2013). Performance and toolchain of a combined GPU/FPGA desktop. In B. Hutchings & V. Betz (Eds.), FPGA  ’13 Proceedings of the ACM/SIGDA international symposium on Field programmable gate arrays (pp. 274–274). Presented at the ACM/SIGDA international symposium on Field programmable gate arrays, New York, NY, USA: ACM.
Vancouver
1.
da Silva B, Braeken A, D’Hollander E, Touhafi A, Cornelis JG, Lemeire J. Performance and toolchain of a combined GPU/FPGA desktop. In: Hutchings B, Betz V, editors. FPGA  ’13 Proceedings of the ACM/SIGDA international symposium on Field programmable gate arrays. New York, NY, USA: ACM; 2013. p. 274–274.
MLA
da Silva, Bruno, An Braeken, Erik D’Hollander, et al. “Performance and Toolchain of a Combined GPU/FPGA Desktop.” FPGA  ’13 Proceedings of the ACM/SIGDA International Symposium on Field Programmable Gate Arrays. Ed. Brad Hutchings & Vaughn Betz. New York, NY, USA: ACM, 2013. 274–274. Print.
@inproceedings{3148367,
  abstract     = {Low-power, high-performance computing nowadays relies on accelerator cards to speed up the calculations. Combining the power of GPUs with the flexibility of FPGAs enlarges the scope of problems that can be accelerated. We describe the performance analysis of a desktop equipped with a GPU Tesla 2050 and an FPGA Virtex- 6 LX 240T. The balance between the I/O and the raw peak performance is analyzed using the roofline model. A well-tuned accelerator- based codesign, identifying the parallelism, the computation and data patterns of different classes of algorithms, will enable to maximize the performance of the combined GPU/FPGA system.},
  author       = {da Silva, Bruno and Braeken, An and D'Hollander, Erik and Touhafi, Abdellah and Cornelis, Jan G.  and Lemeire, Jan},
  booktitle    = {FPGA '13 Proceedings of the ACM/SIGDA international symposium on Field programmable gate arrays},
  editor       = {Hutchings, Brad and Betz, Vaughn},
  isbn         = {9781450318877},
  keyword      = {FPGA,high-level synthesis,GPU,roofline model},
  language     = {eng},
  location     = {Monterey, CA, USA},
  pages        = {274--274},
  publisher    = {ACM},
  title        = {Performance and toolchain of a combined GPU/FPGA desktop},
  year         = {2013},
}