Estimating and Optimizing Performance for Parallel Programs

Estimating and Optimizing Performance for Parallel Programs

Abstract

P3T is an interactive performance estimator that assists users in performance tuning of scientific Fortran programs. It detects performance bottlenecks in the program, identifies the causes of performance problems, and guides the user in selecting effective program transformations in order to gain performance. Four of the most critical performance aspects of parallel programs are estimated: load balance, cache locality, communication and computation overhead. P3T is an integrated tool of the Vienna Fortran Compilation System, which enables the estimator to aggressively exploit considerable knowledge about the compiler's analysis information and code restructuring strategies. We evaluate existing features and describe substantial enhancements in three key areas: graphical user interface, performance parameters and input programs. P3T's graphical user interface directs the user to bottlenecks in a computation that prevent the program from performing well. In addition, it allows filtering and visualizing of performance data at various levels of detail. A new performance parameter, which models computation time is discussed. Furthermore, P3T has been extended to handle a subset of HPF programs. Experiments show the ability of P3T to successfully guide the user in performance tuning of parallel programs. This includes the selection of performance efficient data distribution strategies and profitable program transformations. The results are very encouraging, and demonstrate the feasibility of automatic performance prediction for parallel programs.

Grafik Top
Authors
  • Fahringer, T.
Grafik Top
Shortfacts
Category
Technical Report (Technical Report)
Divisions
Scientific Computing
Publisher
Institute for Software Technology and Parallel Systems, University of Vienna
Date
March 1996
Official URL
http://www.par.univie.ac.at/publications/download/...
Export
Grafik Top