On the Strong Scalability of Maritime CFD

Hawkes, J.; Vaz, G.; Phillips, A.B.; Cox, S.J.; Turnock, S.R.. 2017 On the Strong Scalability of Maritime CFD. Journal of Marine Science and Technology, 23 (1), JMST-D-16-00183R1. 10.1007/s00773-017-0457-7

Before downloading, please read NORA policies.

Preview

Text © Springer Science+Business Media B.V. 2017 This document is the author’s final manuscript version of the journal article, incorporating any revisions agreed during the peer review process. Some differences between this and the publisher’s version remain. You are advised to consult the publisher’s version if you wish to cite from this article. The final publication is available at link.springer.com On_the_Strong_Scalability_of_Maritime_CFD_revised_submitted.pdf - Accepted Version Download (3MB) | Preview

Abstract/Summary

Since 2004, supercomputer growth hasbeen constrained by energy efficiency rather than raw hardware speeds. Tomaintain exponential growth of overall computing power, a massive growth inparallelization is under way. To keep up with these changes, computationalfluid dynamics (CFD) must improve its strong scalability – its ability tohandle lower cells-per-core ratios and achieve finer-grain parallelization. Amaritime-focused, unstructured, finite-volume code (ReFRESCO) is used toinvestigate the scalability problems for incompressible, viscous CFD using two classicaltest-cases. Existing research suggests that the linear equation-system solveris the main bottleneck to incompressible codes, due to the stiff Poisson pressure equation. Here, these results are expandedby analysing the reasons for this poor scalability. In particular, a number ofalternative linear solvers and preconditioners are tested to determine if thescalability problem can be circumvented, including GMRES, Pipelined-GMRES,Flexible-GMRES and BCGS. Conventional block-wise preconditioners are tested,along with multi-grid preconditioners and smoothers in various configurations.Memory-bandwidth constraints and global communication patterns are found to bethe main bottleneck, and no state-of-the-art solution techniques which solve thestrong-scalability problem satisfactorily could be found. There is significantincentive for more research and development in this area.

Item Type:	Publication - Article
Digital Object Identifier (DOI):	10.1007/s00773-017-0457-7
ISSN:	0948-4280
Additional Keywords:	High-Performance Computing, Strong Scalability, Software Profiling, Linear Solvers
Date made live:	18 May 2017 09:03 +0 (UTC)
URI:	https://nora.nerc.ac.uk/id/eprint/517045

Actions (login required)

View Item

Document Downloads

More statistics for this item...