Simultaneous Multithreading and Hard Real Time: Can it be Safe? (Artifact)

Authors Sims Hill Osborne, Joshua J. Bakita, James H. Anderson



PDF
Thumbnail PDF

Artifact Description

DARTS.6.1.1.pdf
  • Filesize: 353 kB
  • 3 pages

Document Identifiers

Author Details

Sims Hill Osborne
  • University of North Carolina at Chapel Hill, Chapel Hill, North Carolina U.S.A.
Joshua J. Bakita
  • University of North Carolina at Chapel Hill, Chapel Hill, North Carolina U.S.A.
James H. Anderson
  • University of North Carolina at Chapel Hill, Chapel Hill, North Carolina U.S.A.

Acknowledgements

We wish to thank the anonymous reviewers for their time and suggested improvements.

Cite As Get BibTex

Sims Hill Osborne, Joshua J. Bakita, and James H. Anderson. Simultaneous Multithreading and Hard Real Time: Can it be Safe? (Artifact). In Special Issue of the 32nd Euromicro Conference on Real-Time Systems (ECRTS 2020). Dagstuhl Artifacts Series (DARTS), Volume 6, Issue 1, pp. 1:1-1:3, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2020) https://doi.org/10.4230/DARTS.6.1.1

Artifact

  MD5 Sum: 03dd6255f2cee5f39b0977e0cacddd5b (Get MD5 Sum)

Abstract

The applicability of Simultaneous Multithreading (SMT) to real-time systems has been hampered by the difficulty of obtaining reliable execution costs in an SMT-enabled system. This problem is addressed from two directions. A scheduler is introduced, CERT-MT, that minimizes SMT-related timing variations, and two new timing analysis methods - one based on the binomial distribution and one based on Cantelli’s Inequality - are given. Both methods estimate probabilistic WCETs and attach statistical confidence levels to those estimates. The timing analyses are applied to tasks executing with and without SMT, and it is found that in some cases, two tasks utilizing SMT can be safely executed in less time than would be needed for either task by itself. A large-scale schedulability study is conducted, showing that CERT-MT can schedule systems with total utilizations twice what could otherwise be achieved. This artifact includes benchmark experiments used to compare execution times with and without SMT and code to analyze the benchmark experiments and duplicate the reported schedulability experiments.

Subject Classification

ACM Subject Classification
  • Computer systems organization → Real-time systems
  • Computer systems organization → Real-time system specification
  • Software and its engineering → Scheduling
  • Software and its engineering → Multithreading
Keywords
  • real-time systems
  • simultaneous multithreading
  • real-time
  • scheduling algorithms
  • timing analysis
  • probability
  • statistics

Metrics

  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    0
    PDF Downloads

References

  1. H. Falk, S. Altmeyer, P. Hellinckx, B. Lisper, W. Puffitsch, C. Rochange, M. Schoeberl, R. B. Sorensen, P. Wagemann, and S. Wegener. TACLeBench: A Benchmark Collection to Support Worst-Case Execution Time Research. In WCET 2016, pages 2:1-2:10, 2016. Google Scholar
  2. S. Osborne and J. H. Anderson. Simultaneous multithreading and hard real time: Can it be safe? In ECRTS 2020 (conditionally accepted), 2020. Google Scholar
  3. S. Osborne, J. Bakita, and J. H. Anderson. Simultaneous multithreading applied to real time. In ECRTS '19, 2019. Google Scholar
  4. S. Osborne, J. Bakita, and J. H. Anderson. Simultaneous Multithreading Applied to Real Time (Artifact). Dagstuhl Artifacts Series, 5(1):8:1-8:2, 2019. URL: https://doi.org/10.4230/DARTS.5.1.8.
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail