Comparison of execution time decomposition methods for performance evaluation

  • Authors:

  • Jan Kwiatkowski;Marcin Pawlik;Dariusz Konieczny

  • Affiliations:

  • Institute of Applied Informatics, Wrocław University of Technology, Wrocław, Poland;Institute of Applied Informatics, Wrocław University of Technology, Wrocław, Poland;Institute of Applied Informatics, Wrocław University of Technology, Wrocław, Poland

  • Venue:
  • PPAM'07 Proceedings of the 7th international conference on Parallel processing and applied mathematics
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

With the advent of multi-core processors and the growing popularity of local cluster installations, better understanding of parallel applications behavior becomes a necessity. It can be argued that the raising popularity of parallelization results in the dare need of methods and tools capable of automatic analysis and prediction of parallel applications efficiency. Traditional methods of performance evaluation based on wall-clock time measurements require consecutive application executions or, when the detailed application profile is created, involves a time-consuming data analysis. In the paper an alternative approach is analyzed. Utilizing the execution time decomposition, a separate analysis of the computations and overhead time is performed to determine the analyzed application efficiency.