Next:
Introduction
Up:
Parallelisation of CABARET
Previous:
Parallelisation of CABARET
Contents
Introduction
Application Background
Why Use CABARET?
Outline of this project
CABARET Implementation for HECToR
Introduction
Code structure
Input Grid
Output Data
Implementing Data Parallelism
Data decomposition
Partitioning the Unstructured grid
Performance of the code
Efficiency of Partitioning Methods
Comparison of the Efficiency of Partitioners
Partitioning and Load Balancing
Compiler comparison
Main CABARET Loops
Loop Structure
Multi-core Data Parallelism
Distributed Data Parallelism in CABARET
Transisition of CABARET to HECToR Phase 2b
Hybrid Parallelism in CABARET
Performance of the Hybrid Parallel CABARET code
CABARET Demonstration
Conclusion
Acknowledgments
Bibliography
Phil Ridley 2011-02-01