Preprints
https://doi.org/10.5194/gmd-2020-340
https://doi.org/10.5194/gmd-2020-340

Submitted as: development and technical paper 30 Oct 2020

Submitted as: development and technical paper | 30 Oct 2020

Review status: a revised version of this preprint was accepted for the journal GMD and is expected to appear here in due course.

LISFLOOD-FP 8.0: the new discontinuous Galerkin shallow water solver for multi-core CPUs and GPUs

James Shaw1, Georges Kesserwani1, Jeffrey Neal2, Paul Bates2, and Mohammad Kazem Sharifian1 James Shaw et al.
  • 1Department of Civil and Structural Engineering, The University of Sheffield, Western Bank, Sheffield, UK
  • 2School of Geographical Sciences, University of Bristol, Bristol, UK

Abstract. LISFLOOD-FP 8.0 includes second-order discontinuous Galerkin (DG2) and first-order finite volume (FV1) solvers of the two-dimensional shallow water equations for modelling a wide range of flows, including rapidly-propagating, supercritical flows, shock waves, or flows over very smooth surfaces. Alongside the existing local inertia solver (called ACC), the new solvers are parallelised on multi-core CPU and Nvidia GPU architectures and run existing LISFLOOD-FP modelling scenarios without modification. The predictive capabilities and computational scalability of the new solvers are studied for two Environment Agency benchmark tests and a real-world fluvial flood simulation driven by rainfall across a 2500 km2 catchment. DG2's second-order-accurate, piecewise-planar representation of topography and flow variables enables predictions on coarse grids that are competitive with FV1 and ACC predictions on 2–4 × finer grids, particularly where river channels are wider than half the grid spacing. Despite the simplified formulation of the local inertia solver, ACC is shown to be spatially second-order-accurate and yields predictions that are close to DG2. The DG2-CPU and FV1-CPU solvers achieve near-optimal scalability up to 16 CPU cores and achieve greater efficiency on grids with fewer than 0.1 million elements. The DG2-GPU and FV1-GPU solvers are most efficient on grids with more than 1 million elements, where the GPU solvers are 2.5–4 × faster than the corresponding 16-core CPU solvers. LISFLOOD-FP 8.0 therefore marks a new step towards operational DG2 flood inundation modelling at the catchment scale.

James Shaw et al.

 
Status: closed
Status: closed
AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment
Printer-friendly Version - Printer-friendly version Supplement - Supplement
 
Status: closed
Status: closed
AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment
Printer-friendly Version - Printer-friendly version Supplement - Supplement

James Shaw et al.

Data sets

LISFLOOD-FP 8.0 results of Environment Agency and Storm Desmond simulations Shaw, J., Kesserwani, G., Neal, J., Bates, P., and Sharifian, M. K. https://doi.org/10.5281/zenodo.4066824

Model code and software

LISFLOOD-FP 8.0 hydrodynamic model LISFLOOD-FP developers https://doi.org/10.5281/zenodo.4073011

James Shaw et al.

Viewed

Total article views: 884 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
601 276 7 884 8 8
  • HTML: 601
  • PDF: 276
  • XML: 7
  • Total: 884
  • BibTeX: 8
  • EndNote: 8
Views and downloads (calculated since 30 Oct 2020)
Cumulative views and downloads (calculated since 30 Oct 2020)

Viewed (geographical distribution)

Total article views: 619 (including HTML, PDF, and XML) Thereof 618 with geography defined and 1 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 

Cited

Latest update: 11 May 2021
Download
Short summary
LISFLOOD-FP has been extended with new shallow water solvers – DG2 and FV1 – for modelling all types of slow- or fast-moving waves over any smooth or rough surface. Using GPU parallelisation, the FV1 solver is faster than the simpler ACC solver on grids with millions of elements. The DG2 solver is notably effective on coarse grids where river channels are hard to capture, improving predicted river levels and flood water depths. This marks a new step towards real-world DG2 flood inundation modelling.