A system for the automated finding of inefficiencies and errors in parallel programs

Authors

  • D.Yu. Andreev
  • A.S. Antonov
  • Vad.V. Voevodin
  • S.A. Zhumatii
  • D.A. Nikitenko
  • K.S. Stefanov
  • P.A. Shvets

Keywords:

supercomputer
performance
efficiency study
parallel computing
parallel programs
dynamic program characteristics
high performance computing
profiling
monitoring PDF (in Russian) (163KB) PDF. zip (in Russian) (141KB)

Abstract

The first version of a system for the automated finding of inefficiencies and errors in parallel programs is considered. The developed algorithms devoted for the automated finding of logic and semantic errors in MPI parallel programs are implemented in this system. The system realizes the search for templates of inefficient behavior of parallel programs written on the UPC language. The user interaction is performed via a specialized web portal that provides the access to available high performance computing resources. This work was supported by the Ministry of Education and Science of Russian Federation (contract 14.514.11.4062). The paper is recommended for publication by the Program Committee of the International Scientific Conference «Scientific Services and Internet: all bounds of parallelism» ((http://agora.guru.ru/abrau2013)).

New computational technologies

Downloads

Published

2013-08-28

Issue

Vol. 14 (2013): Issue 3.

Section

Section 2. Programming

Author Biographies

D.Yu. Andreev

Lomonosov Moscow State University,
Research Computing Center
• Junior Researcher

A.S. Antonov

Lomonosov Moscow State University,
Research Computing Center
• Leading Researcher

Vad.V. Voevodin

Lomonosov Moscow State University,
Research Computing Center
• Researcher

S.A. Zhumatii

Lomonosov Moscow State University,
Research Computing Center
• Leading Researcher

D.A. Nikitenko

Lomonosov Moscow State University,
Research Computing Center
• Researcher

K.S. Stefanov

Lomonosov Moscow State University,
Research Computing Center
• Senior Researcher

P.A. Shvets

Lomonosov Moscow State University,
Research Computing Center
• Programmer

References

  1. Антонов А.С., Воеводин Вад.В., Жуматий С.А., Никитенко Д.А., Стефанов К.С., Швец П.А. Автоматизация поиска ошибок и неэффективностей в параллельных программах // Вычислительные методы и программирование. 2013. 14, раздел 2. 11-17.
  2. Афанасьев К.Е., Власенко А.Ю. Семантические ошибки в параллельных программах для систем с распределенной памятью и методы их обнаружения современными средствами отладки // Вестник КемГУ. 2009. Вып. 2.
  3. Adinets A.V., Bryzgalov P.A., Voevodin Vad.V., Zhumatii S.A., Nikitenko D.A., Stefanov K.S. Job Digest: an approach to dynamic analysis of job characteristics on supercomputers // Numerical Methods and Programming: Advanced Computing. 2012. 13, section 2. 160-166.
  4. Язык программирования UPC (http://upc.lbl.gov/).
  5. Cantonnet F., El-Ghazawi T., Lorenz P., Gaber J. Fast Address Translation Techniques for Distributed Shared Memory Compilers // Proc. 19th IEEE Int. Parallel and Distributed Processing Symp. Denver, 2005. 52-53.
  6. Руководство по языку программирования UPC (http://upc.gwu.edu/downloads/Manual-1.2.pdf).
  7. Спецификация GASP интерфейса (http://gasp.hcs.ufl.edu/gasp-1.5-61606.pdf).
  8. Тестовый пример, реализующий операцию Собеля (http://upc.gwu.edu/download.html (раздел «Benchmarking/GWU_Examples’’)).

License

Copyright (c) 2013 Вычислительные методы и программирование

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.