Work

Preconditioning elliptic operators in high-performance all-scale atmospheric models on unstructured meshes

Year: 2024

Type: article

Abstract: Effective simulation of all-scale atmospheric flows – e.g., cloud-resolvingglobal weather – involves semi-implicit integration of the non-hydrostaticcompressible Euler equations under gravity on a rotating sphere. Suchintegrations depend on complex non-symmetric elliptic solvers. The conditionnumber of the underlying sparse linear operator is O(1010), which necessitatesbespoke operator preconditioning. This paper highlights the developmentand implementation on unstructured meshes of specialised preconditionersfor the non-symmetric Krylov-subspace solver. These developments are set inthe context of a massively-parallel high-performance computing environment,aimed at architectures evolving towards exascale.The baroclinic instability benchmark bearing representative features relevantto numerical weather prediction (NWP) has been selected to study theperformance of the preconditioning options. The reported results illustrate theimproved performance with the new preconditioning options. In particular,the Jacobi based option, for the computational meshes tested in this study,provides an excellent time to solution improvement.

Source: Journal of Computational Physics

DOI: 10.1016/j.jcp.2024.113503

Authors: Mike Gillard, Joanna Szmelter, Francesco Cocetta

Institutions: Loughborough University, CMCC Foundation - Euro-Mediterranean Center on Climate Change

Language: English

Topic: Matrix Theory and Algorithms

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $3,860

Funder: Horizon 2020 Framework Programme