A data-parallel LU relaxation method for the Navier-Stokes equations

Michael J. Wright, Graham V. Candler, Marco Prampolini

Research output: Contribution to conferencePaperpeer-review

Abstract

The LU-SGS method of Yoon and Jameson is modified for the simulation of viscous flows on massively parallel computers. The resulting diagonal data- parallel lower-upper relaxation method (DP-LUR) is shown to have good convergence properties on many problems. However, the convergence rate decreases on the high cell aspect ratio grids required to simulate high Reynolds number flows. Therefore, the diagonal approximation is relaxed and a full matrix DP-LUR method is derived. It retains the data-parallel properties of the original method and reduces the sensitivity of the convergence rate to the grid aspect ratio. Both methods are implemented on the Thinking Machines CM-5, and a large fraction of the peak theoretical performance of the machine is obtained. The low memory use and high parallel efficiency of the methods make them attractive for large-scale simulation of viscous flows.

Original languageEnglish (US)
Pages1200-1209
Number of pages10
StatePublished - Jan 1 1995
Event12th Computational Fluid Dynamics Conference, 1995 - San Diego, United States
Duration: Jun 19 1995Jun 22 1995

Other

Other12th Computational Fluid Dynamics Conference, 1995
CountryUnited States
CitySan Diego
Period6/19/956/22/95

Fingerprint Dive into the research topics of 'A data-parallel LU relaxation method for the Navier-Stokes equations'. Together they form a unique fingerprint.

Cite this