Tensor-product vertex patch smoothers for biharmonic problems
Multilinear algebra, tensor calculus (15A69) Biharmonic, polyharmonic functions and equations, Poisson's equation in two dimensions (31A30) Higher-order elliptic systems (35J48) Preconditioners for iterative methods (65F08) Iterative numerical methods for linear systems (65F10) Stability and convergence of numerical methods for boundary value problems involving PDEs (65N12) Error bounds for boundary value problems involving PDEs (65N15) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30) Mesh generation, refinement, and adaptive methods for boundary value problems involving PDEs (65N50) Multigrid methods; domain decomposition for boundary value problems involving PDEs (65N55) Numerical algorithms for specific classes of architectures (65Y10) Plates (74K20)
This paper presents a carefully engineered GPU implementation of tensor-product vertex patch smoothers for the biharmonic equation. The core achievement lies in its effective adaptation of mathematical algorithms to modern hardware constraints. The use of shared memory, conflict-free access patterns, and strategies like atomic operations or coloring to handle write conflicts demonstrates a deep understanding of GPU programming. A significant performance boost of 59--70 is achieved through a mixed-precision model, where the multigrid preconditioner runs in single precision while the outer solver retains double precision. Algorithmically, the study highlights a clear trade-off. The multiplicative smoother (MVS) proves superior in numerical efficiency, often converging in just a few iterations. However, its more sequential nature leads to a significant performance drop in 3D, achieving only about one-third the throughput of the additive smoother (AVS). While the use of an inexact, fast-diagonalization-based local solver is crucial for performance, it introduces a numerical approximation that slightly increases iteration counts, especially for MVS at high polynomial degrees. Overall, the work successfully demonstrates that complex finite element solvers can achieve high computational intensity on GPUs. This research provides valuable insights for high-performance finite element computing in engineering applications.
- \(C^0\) interior penalty methods for fourth order elliptic boundary value problems on polygonal domains
- \(C^{0}\) interior penalty methods
- A generic interface for parallel cell-based finite element operator application
- A Scalable and Robust Vertex-Star Relaxation for High-Order FEM
- An implementation of tensor product patch smoothers on GPUs
- An Interior Penalty Finite Element Method with Discontinuous Elements
- An optimal order multigrid method for biharmonic, \(C^ 1\) finite element equations
- Convergence of Multigrid Algorithms for Interior Penalty Methods
- Convergence of V- and F-cycle multigrid methods for the biharmonic problem using the Hsieh-Clough-Tocher element
- Direct solution of partial difference equations by tensor product methods
- Divergence-conforming discontinuous Galerkin methods and C^0 interior penalty methods
- Fast matrix-free evaluation of discontinuous Galerkin finite element operators
- Fast tensor product Schwarz smoothers for high-order discontinuous Galerkin methods
- Fully discrete hp-finite elements: Fast quadrature
- Geometric multigrid for Darcy and Brinkman models of flows in highly heterogeneous porous media: a numerical study
- Isogeometric Schwarz preconditioners for the biharmonic problem
- Multigrid methods for \(H^{\mathrm{div}}\)-conforming discontinuous Galerkin methods for the Stokes equations
- Performance and accuracy of hardware-oriented native-, emulated- and mixed-precision solvers in FEM simulations
- Preconditioning in H(đđđŁ) and applications
- The \texttt{deal.II} library, version 9.5
- Two-level additive Schwarz preconditioners for \(C^0\) interior penalty methods
- Two-level non-overlapping Schwarz preconditioners for a discontinuous Galerkin approximation of the biharmonic equation
- Ăber ein Variationsprinzip zur Lösung von Dirichlet-Problemen bei Verwendung von TeilrĂ€umen, die keinen Randbedingungen unterworfen sind. (On a variational principle for solving Dirichlet problems less boundary conditions using subspaces)
This page was built for publication: Tensor-product vertex patch smoothers for biharmonic problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6945987)
