Abstract
Traditional serial explicit numerical codes, when applied for simulating flows over a large domain for long time periods, require a large amount of computational time. To reduce the simulation time, it is advantageous to make use of a parallel computing technology. A parallel code can speedup an application by working on multiple portions of the computation simultaneously. In this work, which is targeted to the application audience, we discuss the tools that can be used for formulating a parallel code. Then we demonstrate the advantages of using the parallel code by imbedding all the high performance tools into an existing serial two-dimensional finite volume code. The parallel code is based on domain decomposition principles and uses Message Passing Interface (MPI) for all inter-processor communication. Computational results are given to illustrate the efficiency of the parallel code, which simulates flow in transition. On a 16 processor shared memory platform, it has been observed that the parallel code accelerates the transient solution by a factor of 15.5.
Original language | English |
---|---|
Pages | 721-737 |
Number of pages | 17 |
State | Published - 2000 |
Event | Proceedings of the 6th International Conference on Estuarine Coastal Modeling - New Orleans, LA, USA Duration: Nov 3 1999 → Nov 5 1999 |
Conference
Conference | Proceedings of the 6th International Conference on Estuarine Coastal Modeling |
---|---|
City | New Orleans, LA, USA |
Period | 11/3/99 → 11/5/99 |
ASJC Scopus subject areas
- Civil and Structural Engineering
- Modeling and Simulation
- Geotechnical Engineering and Engineering Geology