Advertisement

Automatic Data Distribution Method Using First Touch Control for Distributed Shared Memory Multiprocessors

  • Takashi Hirooka
  • Hiroshi Ohta
  • Takayoshi Iitsuka
Conference paper
  • 300 Downloads
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2624)

Abstract

We propose an interprocedural automatic data distribution method for a parallelizing compiler on distributed shared memory multiprocessors. In this paper we present the algorithm and its implementation and evaluation. This method combines “first touch control (FTC)” with data distribution directives. The characteristics of the FTC method are that our compiler controls first touch data distribution of the operating system and accurately determines a complex distribution of data. This method appropriately distributes data for those program patterns which conventional data distribution methods are unable to properly treat. In addition, we implemented an interprocedural analysis which improves data locality for the program as a whole. We evaluated the parallel performance, on SGI Origin2000, of applying this method to NPB2.3serial/FT, SP, and CG, and SPECfp95/tomcatv. On 16 processors, these benchmarks ran 2.1 times faster (on average) than they did without our method.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Chandra, R., Chen, D., Cox, R., Maydan, D.E., Nedeljkovic, N., Anderson, J.: Data Distribution Support on Distributed Shared Memory Multiprocessors, Proc. PLDI’97, pp. 334–345 (1997).Google Scholar
  2. [2]
    Kennedy, K., Kremer, U.: Automatic Data Layout for High Performance Fortran, Proc. Supercomputing’95, (1995).Google Scholar
  3. [3]
    Gupta, M., Banerjee, P.: PARADIGM: A Compiler for Automatic Data Distribution on Multicomputers, Proc. ICS’93, pp. 87–96 (1993).Google Scholar
  4. [4]
    High Performance Fortran Forum: High Performance Fortran Language Specification Version 2.0, (1997).Google Scholar
  5. [5]
    SGI MIPSpro Fortran77 Programmer’s Guide, Silicon Graphics Inc.Google Scholar
  6. [6]
    Tatsumi, S., Kubota, A., Goshima, M., Mori, S., Nakajima, H., Tomita, S.: An Implementation of the Automatic Data Distribution Phase of the Parallelizing Compiler TINPAR, IPSJ SIG Notes, 96-PRO-8, pp. 25–30 (1996). (in Japanese).Google Scholar
  7. [7]
    Matsuura, K., Murai, H., Suehiro, K., Seo, Y.: Fast Automatic Data Layout for Data Parallel Programs, IPSJ Journal, Vol. 41, No. 5, pp. 1420–1429 (2000). (in Japanese).Google Scholar
  8. [8]
    Aoki, Y., Sato, M., Iitsuka, T., Sato, S., Kikuchi, S.: Prototyping of Interprocedural Parallelizing Compiler “WPP” — Performance Evaluation-, IPSJ SIG Notes, 98-ARC-130, pp. 43–48 (1998). (in Japanese).Google Scholar
  9. [9]
    Hirooka, T., Ohta, H., Kikuchi, S.: Automatic Data Distribution Method by First Touch Control for Distributed Shared Memory, IPSJ Journal, Vol. 41, No. 5, pp. 1430–1438 (2000). (in Japanese).Google Scholar
  10. [10]
    The NAS Parallel Benchmarks, http://www.nas.nasa.gov/Software/NPB/
  11. [11]
    SPEC Benchmarks, http://www.specbench.org/

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Takashi Hirooka
    • 1
    • 2
  • Hiroshi Ohta
    • 3
  • Takayoshi Iitsuka
    • 1
    • 2
  1. 1.Advanced Parallelizing Compiler ProjectJapan
  2. 2.Systems Development LaboratoryHitachi, Ltd.Japan
  3. 3.Information & Computer SystemsHitachi, Ltd.Japan

Personalised recommendations