Abstract
The aim of this contribution is to discuss possible algorithmic choices and hardware configurations for the implementation of efficient particle-based simulation programs. By using a population decomposition scheme, we modified the scalar version of the algorithm in order to improve the efficiency of our hybrid particle-based simulation engine. Using a Beowulf-class computer cluster, we measured parallel speed-up with different algorithmic configurations, and related it to the inter-process communication hardware.
Original language | English (US) |
---|---|
Pages (from-to) | 215-218 |
Number of pages | 4 |
Journal | Journal of Computational Electronics |
Volume | 1 |
Issue number | 1-2 |
DOIs | |
State | Published - Jul 1 2002 |
Keywords
- charge transport simulation
- Monte Carlo
- parallel computing
- particle-based methods
ASJC Scopus subject areas
- Electronic, Optical and Magnetic Materials
- Atomic and Molecular Physics, and Optics
- Modeling and Simulation
- Electrical and Electronic Engineering