A scalable performance analysis tool for PowerPC based MPP systems

This paper introduces a tool for optimizing programs on massively parallel computing systems. The tool has been implemented for a PowerPC based parallel computing platform. It is scalable with respect to its implementation and an the way it presents performance data. A major feature contributing to the scalable representation of performance data is the ability to focus measurements on points of interest in the program execution by specifying behavioral attributes. Behavioral attributes are given as thresholds to the results of other measurements. Thus a direct link between results of different measurements can be made which enables the user to link global system behavior to the execution of individual program parts