Towards local-failure local-recovery in PDE frameworks: the case of linear solvers

DOI10.1007/978-3-030-67077-1_2MaRDI QIDQ2233545zbMATH OpenOpenAlexFDO

Authors Mirco Altenbernd, Nils-Arne Dreier, Christian Engwer, Dominik Göddeke

Publication date 18 October 2021

Full work available at URL https://doi.org/10.1007/978-3-030-67077-1_2

zbMATH Keywords

parallel computationfault tolerancelossy compressionPDE frameworks

Mathematics Subject Classification ID

Parallel numerical computation (65Y05) Iterative numerical methods for linear systems (65F10) Software, source code, etc. for problems pertaining to numerical analysis (65-04)

Recommendations

• Finite element recovery techniques for local quantities of linear problems using fundamental solutions
• Failures of local approximation in finite-element methods
• Fault resilient domain decomposition preconditioner for PDEs
• Numerical recovery strategies for parallel resilient Krylov linear solvers.
• Local Piecewise Hyperbolic Reconstruction of Numerical Fluxes for Nonlinear Scalar Conservation Laws
• A massively-parallel, fault-tolerant solver for high-dimensional PDEs
• Local structure-preserving algorithms for partial differential equations
• Local convergence of an inexact-restoration method and numerical experiments
• Solver composition across the PDE/linear algebra barrier

Cites work

• A generic grid interface for parallel and adaptive scientific computing. I: Abstract framework
• A generic grid interface for parallel and adaptive scientific computing. II: Implementation and tests in DUNE
• A minimally intrusive low-memory approach to resilience for existing transient solvers
• Algorithm-Based Fault Tolerance for Matrix Operations
• Exaflop/s: the why and the how
• Fault-tolerance techniques for high-performance computing
• GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems
• Interpolation-restart strategies for resilient eigensolvers
• Methods of conjugate gradients for solving linear systems
• Numerical recovery strategies for parallel resilient Krylov linear solvers.
• Recovery Patterns for Iterative Methods in a Parallel Unstable Environment
• Resilience for massively parallel multigrid solvers

Cited in

(2)

• Scalable failure masking for stencil computations using ghost region expansion and cell to rank remapping
• Fault-tolerant parallel multigrid method on unstructured adaptive mesh

Describes a project that uses

Uses Software

• ISTL
• DUNE

This page was built for publication: Towards local-failure local-recovery in PDE frameworks: the case of linear solvers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2233545)