The High Performance Scalable Computing (HPSC) architecturesupports distributed implementations of applications requiring highperformance, both in terms of processor cycles and data bandwidth.This architecture uses clusters of multi-processor nodes intercon- Nected by a network using the MYRINET protocol. This paper Describesthe simulation capability developed to model and predict Theperformance of the HPSC architecture. This capability was cre- Atedby developing various models and extensions to Ptolemy, a soft- Waremodelling environment developed at the University of Califor- Nia,Berkeley.
展开▼