In Search of a Standards-Based Approach to Hybrid Performance Monitoring

The authors descirbe a new multicomputer performance monitoring system that employs portable performance monitoring instrumentation technology and leverages previous work wherever possible. Its trace-event acquisition is hardware assisted and based on the MultiKron, a single-chip measurement solution developed at the U.S. National Institute for Standards and Technology. The user interface is based on the Pablo Performance Analysis Environment, a visualization and sonification toolkit developed at the University of Illinois. The SPIscope component bridges the gap between these emerging standard interfaces, providing a high-bandwidth path to a large secondary storage for recording performance data. Connectivity to the user's performance analysis workstation is via a TCP/IP LAN. The system provides facilities to support application-specific trace events, breakpoint-style debugging, on-line transmission of selected data, and dynamic acquisition-rate control.By using existing technology and interfaces, the authors hope to contribute a performance monitoring component with greater utility throughout the parallel-processing community. The high bandwidth and large storage capacity of the SPIscope possess immediate significance in refining current limits on practical performance data resolution. The system also provides a foundation for further research in on-line observation, dynamic rate control, and the natural integration of debugging facilities with performance-monitoring tools.