MetaCartSign in to MyCiteSeer

Include Citations | Advanced Search | Help

Include Citations | Advanced Search | Help

  A Comparative Analysis of Four Parallelisation Schemes (1999) [3 citations — 0 self]

Download:
pdf | ps
by N. Mukherjee, J. R. Gurd
in Proceedings of the 1999 ACM International Conference on Supercomputing, ACM
http://www.cs.man.ac.uk/cnc/docs/1999/icspaper82.ps
Add To MetaCart

Abstract:

An experimental study of four different schemes for parallelisation of FORTRAN codes is presented. One scheme is manual (performed by the programmer), the other three are automatic (performed entirely by software). The performance of code generated for two parallel computers from seven different test cases is compared, and reasons for differences in achieved performance between the four parallelisation schemes are analysed. It is concluded that, even using sophisticated techniques, high performance parallelised code cannot be generated by automatic tools unless they take into account feedback about execution-time behaviour. Both post-execution performance analysis and interaction with the programmer are necessary for success. This observation argues for user-centred, feedback-driven parallelisation tools that aid the manual process.

Citations

294 Computer Simulation of Liquids – Allen, Tildesley - 1987
114 Performance analysis of parallelizing compilers on the Perfect Benchmarks programs – Blume, Eigenmann - 1992
80 A hierarchical O(NlogN) forcecalculation algorithm – Barnes, Hut - 1986
59 Parallel Performance Prediction Using Lost Cycles Analysis”, SuperComputing – Crovella, LeBlanc - 1994
9 et al. Advanced Program Restructuring for HighPerformance Computers with Polaris – Blume - 1996
9 et al. The CMU Task Parallel Program Suite – Dinda - 1994
7 et al. Automatic Program Parallelization – Banerjee - 1993
6 A Comparison Study of Automatically Vectorizing Fortran Compilers – Nobayashi, Eoyang - 1989
6 Fortran Accelerator User's Guide – POWER - 1993
5 Expert Programmer versus Parallelizing Compiler: A Comparative Study of Two Approaches for Distributed Shared Memory – O'Boyle, Bull - 1996
3 et al. Experiences using the ParaScope Editor: an interactive parallel programming tool – Hall - 1993
3 A Hierarchical Micromagnetic Model of Longitudinal Thin Film Recording – Miles, Middleton - 1991
2 Parallelising Serial Code: a comparison of three high performance parallel programming methods – MacLaren - 1997
1 et al. Aspects of using multiprocessors in meteorological modeling – Hoffmann - 1988
1 Korn et al. Parallelisation of a Three-dimensional Shallow Water Estuary Model on the KSR-1 – Falc'o - 1995
1 et al. Compiling for heterogeneous systems: A survey and an approach – McKinley - 1995
1 A mapping study of the shallow water equations – O'Boyle, Nisbet - 1992
1 et al. Performance Improvement Through Overhead Analysis: a case study in molecular dynamics – Riley - 1997
View or Download | Add to My Collection | Correct Errors