The overall objective of this research is to develop instructional modules demonstrating the use of performance software in programming numerical algorithms and their implementations on concurrent processor systems.

Developing correct and efficient software is far more complex for parallel systems than it is for sequential processors.

As part of this development and implementation, performance software (such as IBM's VT, VAMPIR, PARAVER, etc.) will be used to help analyse and understand the execution behaviour. The goal is to identify program regions that do not perform well. With these, communication patterns, processor load balance, computation versus comunication ratios, timing characteristics, and processor idle times will be obtained and studied. Furthermore, an evaluation of these performance software tools will be made with the goal of testing their effectiveness and developing extensions to them.

Acknowledgment and Disclaimer: This material is based upon work supported by the National Science Foundation under Grant No. 0086262. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.