Advertisement

Invasive Computing: An Application Assisted Resource Management Approach

  • Andreas Hollmann
  • Michael Gerndt
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7303)

Abstract

Efficient use of computational resources is essential to get a high performance and a low energy consumption. The operating system (OS) is responsible for managing all system resources in a fair and efficient way. This complex task is, however, accomplished without having any notion about the characteristics of the running applications. The lack of information can cause severe performance degradations, especially on non-uniform memory access (NUMA) multicore architectures with their sensitivity to memory placement.

In this paper we present an approach how to augment applications with performance critical properties that can be used for optimizing the resource management within the OS or by user-level tools. We illustrate our concept using the OpenMP parallel programming model.

Keywords

resource management resource awareness numa parallel programming OpenMP 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Terboven, C., Mey, D.A., Schmidl, D., Jin, H., Reichstein, T.: Data and Thread Affinity in OpenMP Programs. In: Proceedings of the 2008 Workshop on Memory Access on Future Processors, MAW 2008 (2008)Google Scholar
  2. 2.
    Teich, J., Henkel, J., Herkersdorf, A., Schmitt-Landsiedel, D., Schröder-Preikschat, W., Snelting, G.: Invasive Computing: An Overview. In: Multiprocessor System-on-Chip - Hardware Design and Tool Integration. Springer (2011)Google Scholar
  3. 3.
    Kleen, A.: A NUMA API for LINUX, Novell, Technical Linux Whitepaper (April 2005)Google Scholar
  4. 4.
    Lankes, S., Bierbaum, B., Bemmerl, T.: Affinity-On-Next-Touch: An Extension to the Linux Kernel for NUMA Architectures. In: Wyrzykowski, R., Dongarra, J., Karczewski, K., Wasniewski, J. (eds.) PPAM 2009. LNCS, vol. 6067, pp. 576–585. Springer, Heidelberg (2010)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Andreas Hollmann
    • 1
  • Michael Gerndt
    • 1
  1. 1.Department of Computer ScienceTechnische Universität MünchenGarchingGermany

Personalised recommendations