Fast Tensor Product Schwarz Smoothers for High-Order Discontinuous Galerkin Methods

TL;DR

This paper introduces efficient tensor product Schwarz smoothers for high-order discontinuous Galerkin methods, significantly reducing computational effort by exploiting operator separability and low rank matrix representations.

Contribution

It presents a novel implementation of domain decomposition smoothers that efficiently invert local matrices for high-order DG methods using tensor product structures.

Findings

Reduced local matrix inversion complexity from O(k^{3d}) to O(dk^{d+1})

Demonstrated efficiency of smoothers in high-order DG discretizations

Exploited low rank representations for fast local solvers

Abstract

In this article, we discuss the efficient implementation of powerful domain decomposition smoothers for multigrid methods for high order discontinuous Galerkin (DG) finite element methods. In particular, we study the inversion of matrices associated to mesh cells and to the patches around a vertex, respectively, in order to obtain fast local solvers for additive and multiplicative subspace correction methods. The effort of inverting local matrices for tensor product polynomials of degree $k$ is reduced from $\mathcal O(k^{3d})$ to $\mathcal O(dk^{d+1})$ by exploiting the separability of the differential operator and resulting low rank representation of its inverse as a prototype for more general low rank representations.