Skip to main content

Using remote access histories for thread scheduling in distributed shared memory systems

  • Contributed Papers
  • Conference paper
  • First Online:
Distributed Computing (DISC 1998)

Part of the book series: Lecture Notes in Computer Science ((LNCS,volume 1499))

Included in the following conference series:

Abstract

We developed the remote access histories mechanism (rahm). a technique that can be used to dynamically, adaptively. and transparently improve the locality of memory references in distributed shared memory systems (dsms). The rahm helps the run-time system make correct decisions concerning the redistribution of threads to hosts — operations that typically involve high overhead.

The rahm is independent of the memory consistency model, and comes into play only when the memory model fails to achieve locality of memory reference. The rahm piggybacks on existing communication and adds only a negligible overhead. It is designed to work in dynamically changing systems, such as non-dedicated distributed environments.

The rahm is fully implemented in user-space as part of the millipede system, a Virtual Parallel Machine (vpm) designed for non-dedicated NT/PC clusters. The performance evaluation shows a potential for dramatic improvement in the efficiency of shared-memory parallel computations in loosely coupled distributed environments, due to an enhanced hit-ratio for local memory references.

Much of this work was done while this author was still with the Department of Computer Science at the Technion.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Institutional subscriptions

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. M.C. Carlisle and A. Rogers. Software caching and computation migration in Olden. In Proc. of the 5th ACM SIGPLAN Symp. on Principles and Practice of Parallel Programming (PPOPP). July 1995.

    Google Scholar 

  2. R. Chandra, A. Gupta, and J.L. Hennessy. Data Locality and Load Balancing in COOL. In Proc. of the Fourth ACM SIGPLAN Symp. on Principles and Practice of Parallel Programming (PPOPP), pages 249–259, May 1993.

    Google Scholar 

  3. J. S. Chase, F. G. Amador, E. D. Lazowska, H. M. Levy, and R. J. Littlefield. The Amber System: Parallel Programming on a Network of Multiprocessors. In Proc. of the 12th ACM Symp. on Operating Systems Principles (SOSP), pages 147–158, December 1989.

    Google Scholar 

  4. T. Chiueh and M. Verma. A compiler-directed distributed shared memory system. In Proc. of the Int'l Conf. on Supercomputing, July 1995.

    Google Scholar 

  5. A. Dubrovski. R. Friedman, and A. Schuster. Load Balancing in Distributed Shared Memory Systems. International Journal of Applied Software Technology, 3, March 1998.

    Google Scholar 

  6. R. Friedman, M. Goldin, A. Itzkovitz, and A. Schuster. Millipede: Easy Parallel Programming in Available Distributed Environments. Software: Practice & Experience, 27(8):929–965. August 1997. Preliminary version appeared in Proc. Euro-Par, Lyon, August 1996, pp. 84–87.

    Article  Google Scholar 

  7. T. Hershman. Millipede Virtual Parallel Machine. BYTE, pages 3–4, May 1998.

    Google Scholar 

  8. W.C. Hsieh, M.F. Kaashoek, and W.E. Weihl. Dynamic Computation Migration in DSM systems. In Proc. of Supercomputing, November 1996.

    Google Scholar 

  9. A. Itzkovitz, A. Schuster, and L. Shalev. Millipede: Supporting Multiple Programming Paradigms on Top of a Single Virtual Parallel Machine. In Proc. HIPS Workshop, Geneve, April 1997.

    Google Scholar 

  10. A. Itzkovitz, A. Schuster, and L. Shalev. Thread Migration and its Applications in Distributed Shared Memory Systems. To appear in the Journal of Systems and Software, 1998. (Also: Technion TR LPCR-#9603).

    Google Scholar 

  11. J. Naor and L. Zosin. A 2-Approximation Algorithm for the Directed Multiway Cut Problem. Manuscript (naor@cs.technion.ac.il).

    Google Scholar 

  12. A. Schuster and L. Shalev. Using Remote Access Histories for Thread Scheduling in Distributed Shared Memory Systems. In 12th Intl. Symp. on Distributed Computing, Andros, September 1998. A Technion/LPCR TR-#9701, January 1997.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Editor information

Shay Kutten

Rights and permissions

Reprints and permissions

Copyright information

© 1998 Springer-Verlag Berlin Heidelberg

About this paper

Cite this paper

Schuster, A., Shalev, L. (1998). Using remote access histories for thread scheduling in distributed shared memory systems. In: Kutten, S. (eds) Distributed Computing. DISC 1998. Lecture Notes in Computer Science, vol 1499. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0056494

Download citation

  • DOI: https://doi.org/10.1007/BFb0056494

  • Published:

  • Publisher Name: Springer, Berlin, Heidelberg

  • Print ISBN: 978-3-540-65066-9

  • Online ISBN: 978-3-540-49693-9

  • eBook Packages: Springer Book Archive

Publish with us

Policies and ethics