Improving Performance and Energy Efficiency on OpenPower Systems Using Scalable Hardware-Software Co-design

  • Miloš Puzović
  • Vadim ElisseevEmail author
  • Kirk Jordan
  • James Mcdonagh
  • Alexander Harrison
  • Robert Sawko
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11203)


Exascale level of High Performance Computing (HPC) implies performance under stringent power constraints. Achieving power consumption targets for HPC systems requires hardware-software co-design to manage static and dynamic power consumption. We present extensions to the open source Global Extensible Open Power Manager (GEOPM) framework, which allows for rapid prototyping of various power and performance optimization strategies for exascale workloads. We have ported GEOPM to OpenPower\({^{\textregistered }}\) architecture and have used our modifications to investigate performance and power consumption optimization strategies for real-world scientific applications.


OpenPOWER Energy efficiency Performance optimization 



Authors would like to acknowledge J. Eastep and C. Cantalupo, Intel, S. Bhat and T. Rosedahl, IBM Systems and D. Graham, STFC.


  1. 1.
    Scientific Grand Challenges: Architectures and Technology for Extreme Scale Computing, San Diego, CA. U.S. Department of Energy, Office of Science, Washington, D.C., 8–10 December 2009Google Scholar
  2. 2.
    Ang, J.: The DOE exascale computing project: overview of relevant energy/power efforts. In: 8th Annual Workshop for Energy Efficient HPC Working Group at SC (2017)Google Scholar
  3. 3.
    Bhat, S.: Programming on-chip components to retrieve sensor data. In: OpenPOWER Summit (2016)Google Scholar
  4. 4.
    Bhat, S.: Openpower based Inband OCC sensors (2017).
  5. 5.
    Vermeire, B.C., et al.: On the utility of GPU accelerated high-order methods for unsteady flow simulations: a comparison with industry-standard tools. J. Comput. Phys. 334, 497–521 (2017)MathSciNetCrossRefGoogle Scholar
  6. 6.
    Eranian, S.: Perfmon2: a flexible performance monitoring interface for Linux. In: Proceedings of the Ottawa Linux Symposium (2006)Google Scholar
  7. 7.
    Eastep, J., et al.: Global extensible open power manager: a vehicle for HPC community collaboration on co-designed energy management solutions. In: ISC (2017)Google Scholar
  8. 8.
    Karlin, I., Keasler, J., Neely, R.: Lulesh 2.0 updates and changes. Technical report LLNL-TR-641973, August 2013Google Scholar
  9. 9.
    Labasan, S., et al.: Variorum: extensible framework for hardware monitoring and contol. In: E2SC at SC (2017)Google Scholar
  10. 10.
    OpenPower Foundation: Openpower technical resources.
  11. 11.
    Plimpton, S.: Fast parallel algorithms for short-range molecular dynamics. J. Comput. Phys. 117, 1–19 (1995)CrossRefGoogle Scholar
  12. 12.
    Rosedahl, T., et al.: Power/performance controlling techniques in OpenPOWER. In: ISC (2017)Google Scholar
  13. 13.
  14. 14.
  15. 15.
    LLNL: MSR-SAFE (2018).
  16. 16.
    NVIDIA: NVIDIA Management Library (2018).
  17. 17.
    READEX: READEX project (2017).
  18. 18.
    Ahmad, W., et al.: Design of an energy aware petaflops class high performance cluster based on power architecture. In: 2017 IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPS Workshops 2017, Orlando/Buena Vista, FL, USA, 29 May–2 June 2017, pp. 964–973 (2017)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Miloš Puzović
    • 1
  • Vadim Elisseev
    • 2
    Email author
  • Kirk Jordan
    • 3
  • James Mcdonagh
    • 2
  • Alexander Harrison
    • 2
  • Robert Sawko
    • 2
  1. 1.The Hartree Centre, STFC Daresbury LaboratorySci-Tech DaresburyCheshireUK
  2. 2.IBM Research, STFC Daresbury LaboratorySci-Tech DaresburyCheshireUK
  3. 3.Data Centric Solutions, IBM T. J. Watson ResearchCambridgeUSA

Personalised recommendations