Advertisement

Analysis and Optimization of CFS Scheduler on NUMA-Based Systems

  • Hongyun Tian
  • Kun Zhang
  • Li Ruan
  • Mingfa Zhu
  • Limin Xiao
  • Xiuqiao Li
  • Yuhang Liu
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 236)

Abstract

Non Uniform Memory Access (NUMA) architecture becomes more and more popular as it has better scalability than Uniform Memory Access (UMA). However, all previous work on the operation system scheduler assumed that the underlying system is UMA. As a result, the performance degrades on NUMA machines due to lacking of consideration to the underlying hardware. Researchers discover that the Completely Fair Scheduler (CFS) does not work smoothly on NUMA machines and even interfere performance relative to the O (1) scheduler. In this paper researchers investigate the causes for the performance decline and devise an architecture aware task-bound approach for NUMA system, which can help the CFS scheduler works efficiently on NUMA platforms. The evaluation shows that the approach can upgrade the system performance by more than 60 % on average. The research has great significance to the development and popularity of domestic operating system.

Keywords

NUMA CFS scheduler Operating system High-performance computer 

Notes

Acknowledgments

Our research is sponsored by the National “Core electronic devices high-end general purpose chips and fundamental software” project under Grant No.2010ZX01036-001-001, the Hi-tech Research and Development Program of China (863 Program) under Grant NO.2011AA01A205, the National Natural Science Foundation of China under Grant NO.60973007, the Doctoral Fund of Ministry of Education of China under Grant NO.20101102110018, the Beijing Natural Science Foundation under Grant NO.4122042, the fund of the State Key Laboratory of Software Development Environment under Grant NO.SKLSDE-2012ZX-07.

References

  1. 1.
    Burkhardt, H.: KSR. June 2011 | TOP500 Supercomputing Sites (2011)Google Scholar
  2. 2.
    Jones, T.: Inside the Linux scheduler——The latest version© Copyright IBM Corporation (2006)Google Scholar
  3. 3.
    Jones, T.: Inside the Linux 2.6 Completely Fair Scheduler© Copyright IBM Corporation (2009)Google Scholar
  4. 4.
    Hewlett-Packard, Intel, Toshiba.: Advanced configuration and power interface (2011)Google Scholar
  5. 5.
    Ma, T., Bosilca, G., Bouteiller, A., Dongarra, J.J.: Locality and topology aware intra-node communication among multicore CPUs. In: Proceedings for EuroMPI, pp. 265–274 (2010)Google Scholar
  6. 6.
    Li, T., Baumberger D, et al.: Efficient operating system scheduling for performance-asymmetric multi-core architectures. In: Proceedings of Supercomputing (2007)Google Scholar
  7. 7.
    Azimi, T., et al.: Thread clustering: sharing-aware scheduling on multiprocessors. In: Proceedings of Eurosys (2007)Google Scholar
  8. 8.
    Corbalan, J., Martorell, X.: Evaluation of the memory page migration influence in the system performance. In: Proceedings of super computing, pp. 121–129 (2003)Google Scholar
  9. 9.
    Blagodurov, S., et al.: User-level scheduling on NUMA multicore systems under Linux. ACM Trans. Comput. Syst. 28(4), Article 8 (2010)Google Scholar
  10. 10.
    Li, X.: NUMA scheduling algorithm based on affinitive node. Comput. Eng. 32(1), 99–101 (2006)Google Scholar
  11. 11.
    Blagodurov, S., Zhuravle, S., et al.: A case for NUMA-aware contention management on multicore systems. In: PaCT, pp. 557–558 (2010)Google Scholar
  12. 12.
    Kamali, A: Sharing Aware Scheduling on Multicore Systems. Simon Fraser University, Burnaby (2010)Google Scholar
  13. 13.
    Schmidl, D.: Towards NUMA support with distance information. In: IWOMP’11 Proceedings of the 7th International Conference on OpenMP, pp. 69–79. Springer, Berlin, Heidelberg © (2011)Google Scholar

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  1. 1.State Key Laboratory of Software Development EnvironmentBeijingChina
  2. 2.School of Computer Science and EngineeringBeihang UniversityBeijingChina

Personalised recommendations