Ampehre: An Open Source Measurement Framework for Heterogeneous Compute Nodes

  • Achim LöschEmail author
  • Alex Wiens
  • Marco Platzner
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10793)


Profiling applications on a heterogeneous compute node is challenging since the way to retrieve data from the resources and interpret them varies between resource types and manufacturers. This holds especially true for measuring the energy consumption. In this paper we present Ampehre, a novel open source measurement framework that allows developers to gather comparable measurements from heterogeneous compute nodes, e.g., nodes comprising CPU, GPU, and FPGA. We explain the architecture of Ampehre and detail the measurement process on the example of energy measurements on CPU and GPU. To characterize the probing effect, we quantitatively analyze the trade-off between the accuracy of measurements and the CPU load imposed by Ampehre. Based on this analysis, we are able to specify reasonable combinations of sampling periods for the different resource types of a compute node.


Heterogeneous computing Measurement Energy Open source 



This work has been partially supported by the German Research Foundation (DFG) within the Collaborative Research Center 901 “On-The-Fly Computing”.


  1. 1.
    Linux Kernel: perf: Linux Profiling with Performance Counters (2017).
  2. 2.
    Eulisse, G., Tuura, L.: IgProf, the Ignominous Profiler (2013).
  3. 3.
    Roehl, T.: Performance Monitoring and Benchmarking Suite (2017).
  4. 4.
    Intel Corporation: Intel VTune Amplifier (2017).
  5. 5.
    Nvidia Corporation: Nvidia Nsight (2017).
  6. 6.
    Khan, K.N., Nybäck, F., Ou, Z., Nurminen, J.K., Niemi, T., Eulisse, G., Elmer, P., Abdurachmanov, D.: Energy profiling using IgProf. In: 2015 15th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, May 2015Google Scholar
  7. 7.
    Innovative Computing Laboratory, University of Tennessee: Performance Application Programming Interface (PAPI) (2016).
  8. 8.
    McCraw, H., Ralph, J., Danalis, A., Dongarra, J.: Power monitoring with PAPI for extreme scale architectures and dataflow-based programming models. In: 2014 IEEE International Conference on Cluster Computing (CLUSTER), September 2014Google Scholar
  9. 9.
    Lösch, A., Knorr, C., El-Ali, A., Wiens, A.: Ampehre: Accurately Measuring Power and Energy for Heterogeneous Resource Environments (2017).
  10. 10.
    Intel Corporation: Intelligent Platform Management Interface (IPMI), IPMI Technical Resources (2015).
  11. 11.
    Nvidia Corporation: Nvidia Management Library (NVML) (2017).
  12. 12.
    Intel Corporation: Intel 64 and IA-32 Architectures Software Developer Manuals, October 2017.
  13. 13.
    Vlasenko, D.: BusyBox: The Swiss Army Knife of Embedded Linux (2017).
  14. 14.
    Lösch, A., Beisel, T., Kenter, T., Plessl, C., Platzner, M.: Performance-centric scheduling with task migration for a heterogeneous compute node in the data center. In: 2016 Design, Automation Test in Europe Conference Exhibition (DATE), pp. 912–917, March 2016Google Scholar
  15. 15.
    Lösch, A., Platzner, M.: reMinMin: a novel static energy-centric list scheduling approach based on real measurements. In: 2017 IEEE 28th International Conference on Application-specific Systems, Architectures and Processors (ASAP), pp. 149–154, July 2017Google Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Paderborn UniversityPaderbornGermany

Personalised recommendations