On the strong scalability of maritime CFD
On the strong scalability of maritime CFD
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.
High-Performance Computing, Strong Scalability, Software Profiling, Linear Solvers
Hawkes, James
3882428e-838b-4e54-9fbf-ba85ba080dc9
Vaz, Guilherme
a053069d-9831-4b28-a9c0-6503ddaab25d
Phillips, Alexander
f565b1da-6881-4e2a-8729-c082b869028f
Cox, Simon
0e62aaed-24ad-4a74-b996-f606e40e5c55
Turnock, Stephen
d6442f5c-d9af-4fdb-8406-7c79a92b26ce
Hawkes, James
3882428e-838b-4e54-9fbf-ba85ba080dc9
Vaz, Guilherme
a053069d-9831-4b28-a9c0-6503ddaab25d
Phillips, Alexander
f565b1da-6881-4e2a-8729-c082b869028f
Cox, Simon
0e62aaed-24ad-4a74-b996-f606e40e5c55
Turnock, Stephen
d6442f5c-d9af-4fdb-8406-7c79a92b26ce
Hawkes, James, Vaz, Guilherme, Phillips, Alexander, Cox, Simon and Turnock, Stephen
(2017)
On the strong scalability of maritime CFD.
Journal of Marine Science and Technology, [JMST-D-16-00183R1].
(doi:10.1007/s00773-017-0457-7).
Abstract
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.
Text
On the strong scalability of maritime CFD_revised_submitted
- Accepted Manuscript
More information
Accepted/In Press date: 16 May 2017
e-pub ahead of print date: 27 May 2017
Keywords:
High-Performance Computing, Strong Scalability, Software Profiling, Linear Solvers
Organisations:
Professional Services - COO, Computational Engineering & Design Group, National Oceanography Centre, Civil Maritime & Env. Eng & Sci Unit, Fluid Structure Interactions Group, Southampton Marine & Maritime Institute
Identifiers
Local EPrints ID: 408301
URI: http://eprints.soton.ac.uk/id/eprint/408301
ISSN: 0948-4280
PURE UUID: 8c462b22-8b8e-4aa1-9361-8e9284ae2c7a
Catalogue record
Date deposited: 19 May 2017 04:02
Last modified: 26 Nov 2021 05:44
Export record
Altmetrics
Contributors
Author:
James Hawkes
Author:
Guilherme Vaz
Author:
Alexander Phillips
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
