Performance Evaluation and Prediction

Topic 02


Parallel systems exist solely to achieve better performance than is possible on sequential systems.  To meet this objective,  it is critical that users can both measure performance on a given system and predict the performance for other systems. This workshop provides a forum for research in all areas of performance measurement and prediction.  Of particular interest are tools and systems that work with large-scale applications or large numbers (hundreds to thousands) of processors.


In addition, approaches to understanding memory system performance are encouraged.



  • instrumentation for measurement and prediction
  • predicting performance as applications scale-up
  • measurement and modelling of grid middleware & applications
  • performance data analysis and visualization
  • evaluation and benchmarking
  • system and hardware monitoring
  • predictive performance models
  • case studies involving tuning of real applications
  • automatic performance analysis


Global Chair

Prof. Jeff Hollingsworth

Computer Science Department

University of Maryland, USA



Vice Chairs

Prof. Allen D. Malony

Department of Computer and

Information Science

University of Oregon, USA


Prof. Jesús Labarta

European Center for Parallelism of Barcelona

Technical University of Catalonia, Spain



Local Chair

Prof. Thomas Fahringer

Institute for Software Science

University of Vienna, Austria