Communication-Aware Processor Allocation for Supercomputers

  • Michael A. Bender
  • David P. Bunde
  • Erik D. Demaine
  • Sándor P. Fekete
  • Vitus J. Leung
  • Henk Meijer
  • Cynthia A. Phillips
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3608)

Abstract

We give processor-allocation algorithms for grid architectures, where the objective is to select processors from a set of available processors to minimize the average number of communication hops.

The associated clustering problem is as follows: Given n points in \(\mathcal{R}^d\), find a size-k subset with minimum average pairwise L1 distance. We present a natural approximation algorithm and show that it is a \(\frac{7}{4}\)-approximation for 2D grids. In d dimensions, the approximation guarantee is 2 - \(\frac{1}{2d}\), which is tight. We also give a polynomial-time approximation scheme (PTAS) for constant dimension d and report on experimental results.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Ahmadinia, A., Bobda, C., Fekete, S., Teich, J., der Veen, J.: Optimal routing-conscious dynamic placement for reconfigurable computing. In: Becker, J., Platzner, M., Vernalde, S. (eds.) FPL 2004. LNCS, vol. 3203, pp. 847–851. Springer, Heidelberg (2004)CrossRefGoogle Scholar
  2. 2.
    Bartal, Y., Charikar, M., Raz, D.: Approximating min-sum k-clustering in metric spaces. In: Proc. 33rd Symp. on Theory of Computation, pp. 11–20 (2001)Google Scholar
  3. 3.
    Baylor, S., Benveniste, C., Hsu, Y.: Performance evaluation of a massively paralel I/O subsystem. In: Jain, R., Werth, J., Browne, J. (eds.) Synthesizing Synchronous Systems by Static Scheduling in Space-Time. The Kluwer International Series in Engineering and Computer Science, ch. 13, vol. 362, pp. 293–311. Kluwer Academic Publishers, Dordrecht (1996)Google Scholar
  4. 4.
    Bender, C.M., Bender, M.A., Demaine, E.D., Fekete, S.P.: What is the optimal shape of a city? J. Physics A: Mathematical and General 37, 147–159 (2004)MATHCrossRefMathSciNetGoogle Scholar
  5. 5.
    Bender, M.A., Bunde, D.P., Demaine, E.D., Fekete, S.P., Leung, V.J., Meijer, H., Phillips, C.A.: Communication-aware processor allocation for supercomputers. Technical Report cs.DS/0407058, Computing Research Repository (2004), http://arxiv.org/abs/cs.DS/0407058
  6. 6.
    Bhattacharya, S., Tsai, W.-T.: Lookahead processor allocation in mesh connected massively parallel computers. In: Proc. 8th International Parallel Processing Symposium, pp. 868–875 (1994)Google Scholar
  7. 7.
    Brightwell, R., Fisk, L.A., Greenberg, D.S., Hudson, T., Levenhagen, M., Maccabe, A.B., Riesen, R.: Massively parallel computing using commodity components. Parallel Computing 26(2-3), 243–266 (2000)MATHCrossRefGoogle Scholar
  8. 8.
    Chang, C., Mohapatra, P.: Improving performance of mesh connected multicomputers by reducing fragmentation. Journal of Parallel and Distributed Computing 52(1), 40–68 (1998)MATHCrossRefGoogle Scholar
  9. 9.
    Chuang, P.-J., Tzeng, N.-F.: An efficient submesh allocation strategy for mesh computer systems. In: Proc. Int. Conf. Dist. Comp. Systems, pp. 256–263 (1991)Google Scholar
  10. 10.
    Feitelson, D.: The parallel workloads archive, http://www.cs.huji.ac.il/labs/parallel/workload/index.html
  11. 11.
    Fekete, S.P., Meijer, H.: Maximum dispersion and geometric maximum weight cliques. Algorithmica 38, 501–511 (2004)MATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Guttmann-Beck, N., Hassin, R.: Approximation algorithms for minimum sum p-clustering. Disc. Appl. Math. 89, 125–142 (1998)MATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Indyk, P.: A sublinear time approximation scheme for clustering in metric spaces. In: Proc. 40th Ann. IEEE Symp. Found. Comp. Sci (FOCS), pp. 154–159 (1999)Google Scholar
  14. 14.
    Karp, R.M., McKellar, A.C., Wong, C.K.: Near-optimal solutions to a 2-dimensional placement problem. SIAM Journal on Computing 4, 271–286 (1975)MATHCrossRefMathSciNetGoogle Scholar
  15. 15.
    Krueger, P., Lai, T.-H., Dixit-Radiya, V.: Job scheduling is more important than processor allocation for hypercube computers. IEEE Trans. on Parallel and Distributed Systems 5(5), 488–497 (1994)CrossRefGoogle Scholar
  16. 16.
    Krumke, S., Marathe, M., Noltemeier, H., Radhakrishnan, V., Ravi, S., Rosenkrantz, D.: Compact location problems. Th. Comp. Sci. 181, 379–404 (1997)MATHCrossRefMathSciNetGoogle Scholar
  17. 17.
    Leung, V., Arkin, E., Bender, M., Bunde, D., Johnston, J., Lal, A., Mitchell, J., Phillips, C., Seiden, S.: Processor allocation on Cplant: achieving general processor locality using one-dimensional allocation strategies. In: Proc. 4th IEEE International Conference on Cluster Computing, pp. 296–304 (2002)Google Scholar
  18. 18.
    Li, K., Cheng, K.-H.: 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
  19. 19.
    Lo, V., Windisch, K., Liu, W., Nitzberg, B.: Non-contiguous processor allocation algorithms for mesh-connected multicomputers. IEEE Transactions on Parallel and Distributed Computing 8(7) (1997)Google Scholar
  20. 20.
    Mache, J., Lo, V.: Dispersal metrics for non-contiguous processor allocation. Technical Report CIS-TR-96-13, University of Oregon (1996)Google Scholar
  21. 21.
    Mache, J., Lo, V.: The effects of dispersal on message-passing contention in processor allocation strategies. In: Proc. Third Joint Conf. on Information Sciences, Sessions on Parallel and Distributed Processing, vol. 3, pp. 223–226 (1997)Google Scholar
  22. 22.
    Mache, J., Lo, V., Windisch, K.: Minimizing message-passing contention in fragmentation-free processor allocation. In: Proc. 10th Intern. Conf. Parallel and Distributed Computing Systems, pp. 120–124 (1997)Google Scholar
  23. 23.
    Moore, S., Ni, L.: The effects of network contention on processor allocation strategies. In: Proc. 10th Int. Par. Proc. Symp., pp. 268–274 (1996)Google Scholar
  24. 24.
    Sahni, S., Gonzalez, T.: p-complete approximation problems. JACM 23(3), 555–565 (1976)Google Scholar
  25. 25.
    Sandia National Laboratories. The Computational Plant Project, http://www.cs.sandia.gov/cplant
  26. 26.
    Subramani, V., Kettimuthu, R., Srinivasan, S., Johnson, J., Sadayappan, P.: Selective buddy allocation for scheduling parallel jobs on clusters. In: Proc. 4th IEEE International Conference on Cluster Computing (2002)Google Scholar
  27. 27.
    University of Oregon Resource Allocation Group. Procsimity. http://www.cs.uoregon.edu/research/DistributedComputing/ProcSimity.html
  28. 28.
    Windisch, K., Miller, J., Lo, V.: Procsimity: An experimental tool for processor allocation and scheduling in highly parallel systems. In: Proc. Fifth Symp. on the Frontiers of Massively Parallel Computation, pp. 414–421 (1995)Google Scholar
  29. 29.
    Zhu, Y.: Efficient processor allocation strategies for mesh-connected parallel computers. J. Parallel and Distributed Computing 16, 328–337 (1992)MATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Michael A. Bender
    • 1
  • David P. Bunde
    • 2
  • Erik D. Demaine
    • 3
  • Sándor P. Fekete
    • 4
  • Vitus J. Leung
    • 5
  • Henk Meijer
    • 6
  • Cynthia A. Phillips
    • 5
  1. 1.Department of Computer ScienceSUNY Stony BrookStony BrookUSA
  2. 2.Department of Computer ScienceUniversity of IllinoisUrbanaUSA
  3. 3.MIT Computer Science and Artificial Intelligence LaboratoryCambridgeUSA
  4. 4.Dept. of Mathematical OptimizationBraunschweig University of TechnologyBraunschweigGermany
  5. 5.Discrete Algorithms & Math DepartmentSandia National LaboratoriesAlbuquerqueUSA
  6. 6.Dept. of Computing and Information ScienceQueen’s UniversityKingstonCanada

Personalised recommendations