The paragon performance monitoring environment

Parallel systems present a special challenge to debugging and performance analysis given the massive amounts of data and physically distributed state of the application. The authors describe an integrated toolset that provides an extensible and consistent programming environment for monitoring and visualizing the performance of parallel applications. It uses a scalable distributed monitoring system to gather performance data that drive profiling and graphical performance visualization tools. Performance data are analyzed by both application and system oriented performance tools which are tied together by a common graphical interface.