Advanced search
1 file | 993.21 KB

Towards automatic program partitioning

Sean Rul (UGent) , Hans Vandierendonck (UGent) and Koen De Bosschere (UGent)
Author
Organization
Abstract
There is a trend towards using accelerators to increase performance and energy efficiency of general-purpose processors. Adoption of accelerators, however, depends on the availability of tools to facilitate programming these devices. In this paper, we present. techniques for automatically partitioning programs for execution on accelerators. We call the off-loaded code regions sub-algorithms, which are parts of the program that are loosely connected to the remainder of the program. We present three heuristics for automatically identifying sub-algorithms based oil control flow and data flow properties. Analysis of SPECint and MiBench benchmarks shows that, on average 12 sub-algorithms are identified (up to 54), covering the full execution time for 27 out, of 30 benchmarks. We show that these sub-algorithms are suitable for off-loading them to accelerators by manually implementing sub-algorithms for 2 SPECint benchmarks on the Cell processor.
Keywords
accelerators, off-loading, partitioning, sub-algorithms, CODE

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 993.21 KB

Citation

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

Chicago
Rul, Sean, Hans Vandierendonck, and Koen De Bosschere. 2009. “Towards Automatic Program Partitioning.” In CF’09: CONFERENCE ON COMPUTING FRONTIERS & WORKSHOPS, 89–98. New York, NY, USA: Association for Computing Machinery (ACM).
APA
Rul, S., Vandierendonck, H., & De Bosschere, K. (2009). Towards automatic program partitioning. CF’09: CONFERENCE ON COMPUTING FRONTIERS & WORKSHOPS (pp. 89–98). Presented at the 6th ACM International Conference on Computing Frontiers and Workshops, New York, NY, USA: Association for Computing Machinery (ACM).
Vancouver
1.
Rul S, Vandierendonck H, De Bosschere K. Towards automatic program partitioning. CF’09: CONFERENCE ON COMPUTING FRONTIERS & WORKSHOPS. New York, NY, USA: Association for Computing Machinery (ACM); 2009. p. 89–98.
MLA
Rul, Sean, Hans Vandierendonck, and Koen De Bosschere. “Towards Automatic Program Partitioning.” CF’09: CONFERENCE ON COMPUTING FRONTIERS & WORKSHOPS. New York, NY, USA: Association for Computing Machinery (ACM), 2009. 89–98. Print.
@inproceedings{908726,
  abstract     = {There is a trend towards using accelerators to increase performance and energy efficiency of general-purpose processors. Adoption of accelerators, however, depends on the availability of tools to facilitate programming these devices.
In this paper, we present. techniques for automatically partitioning programs for execution on accelerators. We call the off-loaded code regions sub-algorithms, which are parts of the program that are loosely connected to the remainder of the program. We present three heuristics for automatically identifying sub-algorithms based oil control flow and data flow properties.
Analysis of SPECint and MiBench benchmarks shows that, on average 12 sub-algorithms are identified (up to 54), covering the full execution time for 27 out, of 30 benchmarks. We show that these sub-algorithms are suitable for off-loading them to accelerators by manually implementing sub-algorithms for 2 SPECint benchmarks on the Cell processor.},
  author       = {Rul, Sean and Vandierendonck, Hans and De Bosschere, Koen},
  booktitle    = {CF'09: CONFERENCE ON COMPUTING FRONTIERS \& WORKSHOPS},
  isbn         = {9781605584133},
  language     = {eng},
  location     = {Ischia, Italy},
  pages        = {89--98},
  publisher    = {Association for Computing Machinery (ACM)},
  title        = {Towards automatic program partitioning},
  url          = {http://dx.doi.org/10.1145/1531743.1531759},
  year         = {2009},
}

Altmetric
View in Altmetric
Web of Science
Times cited: