Workload decomposition for particle simulation applications on hierarchical distributed-shared memory parallel systems with integration of HPF and OpenMP

Authors:
Sergio Briguglio;Beniamino Di Martino;Gregorio Vlad
Affiliations:
Associazione Euratom-ENEA, sulla Fusione, C.R. Frascati, C.P. 65-I-00044, Frascati, Rome, Italy;Dip. Ingegneria, dell'Informazione, Second University of Naples, Italy;Associazione Euratom-ENEA, sulla Fusione, C.R. Frascati, C.P. 65-I-00044, Frascati, Rome, Italy
Venue:
ICS '01 Proceedings of the 15th international conference on Supercomputing
Year:
2001

Citing 7
Cited 1

Solving problems on concurrent processors. Vol. 1: General techniques and regular problems

Solving problems on concurrent processors. Vol. 1: General techniques and regular problems
A general concurrent algorithm for plasma particle-in-cell simulation codes

Journal of Computational Physics
Dynamic load balancing for a 2D concurrent plasma PIC code

Journal of Computational Physics
Object-oriented parallel computation for plasma simulation

Communications of the ACM - Special issue on object-oriented experiences and future trends
Particle-in-cell simulation codes in High Performance Fortran

Supercomputing '96 Proceedings of the 1996 ACM/IEEE conference on Supercomputing
Plasma Physics Via Computer

Plasma Physics Via Computer
Parallel Plasma Simulation in High Performance FORTRAN

HPCN Europe 1998 Proceedings of the International Conference and Exhibition on High-Performance Computing and Networking

Workload decomposition strategies for shared memory parallel systems with OpenMP

Scientific Programming

Quantified Score

Hi-index	0.00

Visualization

Abstract

A crucial issue in programming hierarchical distributed-shared memory systems is the workload decomposition. In this paper we address this issue in the framework of porting typical particle in cell (PIC) applications on hierarchical distributed-shared memory parallel systems. The workload decomposition we have devised consists in a two-stage procedure: a higher-level decomposition among the computational nodes, and a lower-level one among the processors of each computational node. We have implemented the described work decomposition without large programming effort by using and integrating the high-level languages High Performance Fortran and OpenMP.