Abstract
This paper explores the federated modeling of an optical data path that interconnects processors and memory in a multicomputer system. By federated modeling, we mean the use of two pre-existing simulation models of distinct subsystems to form a resulting overall model of the entire system. In this experimental federation, we investigate mechanisms that facilitate the distinct models working together while explicitly minimizing the model integration effort required. This focus yields tradeoffs in both model execution time and (potentially) the fidelity of the overall performance results. These tradeoffs are investigated quantitatively in the context of a particular system under investigation, a multicomputer optical data path.
Original language | English |
---|---|
Pages (from-to) | 264-274 |
Number of pages | 11 |
Journal | Proceedings of the IASTED International Conference on Modelling, Simulation, and Optimization |
State | Published - 2004 |
Event | Proceedings of the Fourth IASTED International Conference on Modelling, Simulation, and Optimization - Kauai, HI, United States Duration: Aug 17 2004 → Aug 19 2004 |
Keywords
- Federated simulation models
- Optical data path
- Performance evaluation
- Processor/memory gap