A Fine-Grained Parallel Model for the Fast Iterative Method in Solving Eikonal Equations

Author

Dang, F. ; Emad, N. ; Fender, Alexandre

Author_Institution

Silkan & Lab. PRiSM, Univ. de Versailles, Versailles, France

fYear

2013

fDate

28-30 Oct. 2013

Firstpage

152

Lastpage

157

Abstract

In this paper we present a parallel strategy for solving Eikonal and related static (steady state) Hamilton-Jacobi equations using the Fast Iterative Method. The Fast Iterative Method is a variant of the Fast Marching Method which is more fitted for parallel computing since it is basically designed for graphic processing units (GPUs). We propose a parallel model based on front partitioning which particularly fits shared-memory architectures. We review Fast Marching Method parallel methods, we explain our parallel strategy and give a performance analysis.

Keywords

graphics processing units; iterative methods; mathematics computing; nonlinear differential equations; parallel processing; partial differential equations; performance evaluation; shared memory systems; Eikonal equations; GPU; fast iterative method; fast marching method; fine-grained parallel model; front partitioning; graphic processing units; parallel computing; parallel strategy; performance analysis; shared-memory architecture; static Hamilton-Jacobi equations; steady state Hamilton-Jacobi equations; Complexity theory; Computational modeling; Equations; Iterative methods; Mathematical model; Niobium; Program processors; Eikonal equation; Fast Iterative Method; Hamilton-Jacobi equations; multi-core processing;

fLanguage

English

Publisher

ieee

Conference_Titel

P2P, Parallel, Grid, Cloud and Internet Computing (3PGCIC), 2013 Eighth International Conference on

Conference_Location

Compiegne

Type

conf

DOI

10.1109/3PGCIC.2013.29

Filename

6681222