Hybrid parallelization of CFD applications with dynamic thread balancing

Authors:
Alexander Spiegel;Dieter an Mey;Christian Bischof
Affiliations:
Center for Computing and Communications, RWTH Aachen University, Germany;Center for Computing and Communications, RWTH Aachen University, Germany;,Center for Computing and Communications, RWTH Aachen University, Germany
Venue:
PARA'04 Proceedings of the 7th international conference on Applied Parallel Computing: state of the Art in Scientific Computing
Year:
2004

Citing 1
Cited 2

Parallel programming with message passing and directives

Computing in Science and Engineering

Nested parallelization with OpenMP

International Journal of Parallel Programming
Nested parallelization of the flow solver TFS using the ParaWise parallelization environment

IWOMP'05/IWOMP'06 Proceedings of the 2005 and 2006 international conference on OpenMP shared memory parallel programming

Quantified Score

Hi-index	0.00

Visualization

Abstract

SMP Clusters with fat nodes offer an interesting capability for large applications that employ a hybrid parallelization model: to improve load balance, the number of threads can be increased in order to speed-up busy MPI processes or decreased to slow down idle MPI processes, provided these processes reside on the same SMP node. We developed a library which performs this thread adjustment automatically during program execution. Experimental results demonstrate remarkable speed-ups with minimal programming effort.