Title :
High-bandwidth trace collection for multicomputer performance monitoring
Author :
Hudnall, Charles E., Jr. ; Ledlow, Dallas L.
Author_Institution :
NSF Eng. Res. Center for Comput. Field Simulation, Mississippi State Univ., MS, USA
Abstract :
Developing parallel applications requires sophisticated tools for tuning program performance. These tools require large quantities of data to be collected for analysis. A design for the collection and storage of trace data is presented in this paper. The design provides sufficient bandwidth and storage capacity for the collection of performance data from a 32-processor parallel computer using hybrid performance monitoring. A custom token-ring network is used to collect performance traces at a maximum bandwidth of 20 MBytes/sec. A 16 GByte disk array is utilized in the design for the storage of collected trace data. The resulting design provides a processor-independent method for collecting performance data from moderately parallel computing systems
Keywords :
arrays; parallel machines; parallel programming; software performance evaluation; system monitoring; token networks; 32-processor parallel computer; custom token-ring network; data analysis; data collection; data storage; disk array; high-bandwidth trace collection; hybrid performance monitoring; multicomputer performance monitoring; parallel applications development; processor independent method; program performance tuning; storage capacity; Bandwidth; Computational modeling; Computer interfaces; Computer networks; Computerized monitoring; Concurrent computing; Hardware; High performance computing; Instruments; Probes;
Conference_Titel :
System Theory, 1995., Proceedings of the Twenty-Seventh Southeastern Symposium on
Conference_Location :
Starkville, MS
Print_ISBN :
0-8186-6985-3
DOI :
10.1109/SSST.1995.390609
