Advertisement

A batch scheduler for the Intel Paragon with a non-contiguous node allocation algorithm

  • Michael Wan
  • Reagan Moore
  • George Kremenek
  • Ken Steube
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1162)

Abstract

As the system usage model for scalable parallel processors evolves from the singleuser, dedicated access model to a multi-user production environment, a versatile batch scheduler is needed to match user requirements to system resources. This paper describes the design and performance of a batch scheduler for the Intel Paragon system that addresses the issues associated with a multi-user production environment, including scheduling for heterogeneous nodes, scheduling for long-running jobs, scheduling for large jobs, prime/non-prime time modes, and node allocation schemes. The Modified 2-D Buddy system (M2DB) for non-contiguous node allocation is introduced and studied in this paper.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    B. Kingsbury, H. Walter, M. Bridge and T. Carver. “NQS, Network Queuing System, Version 2.0”, Cosmic Program # ARC-13179.Google Scholar
  2. [2]
    L. M. Ni and P. K. McKinley. “A survey of wormhole routing techniques in direct networks”. IEEE Trans. Computers, 1993.Google Scholar
  3. [3]
    V. Lo, K. Windisch, W. Liu and B. Nitzberg. “Non-contiguous Processor Allocation Algorithm for Mesh-connected Multicomputers”, paper to be published.Google Scholar
  4. [4]
    K. Li and K. Cheng. “A two-dimensional buddy system for dynamic resource allocation in a partitionable mesh connected system”. Journal of Parallel and Distributed Computing, 12:79–83, 1991.CrossRefGoogle Scholar
  5. [5]
    FTP://ftp.sdsc.edu//pub/sdsc/parallel/ paragon/paragon95.txt.ZGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Michael Wan
    • 1
  • Reagan Moore
    • 1
  • George Kremenek
    • 1
  • Ken Steube
    • 1
  1. 1.San Diego Supercomputer CenterSan Diego

Personalised recommendations