Mixed-precision preconditionders in parallel domain decomposition solvers

MaRDI QIDQ3500076zbMATH OpenFDO

Authors L. Giraud, A. Haidar, Layne T. Watson

Publication date 3 June 2008

zbMATH Keywords

32-bit arithmetic32-bit operations64-bit floating-point arithmeticconvergencedirect methodsdomain decompositionfactorizationinexact Newtoniterative refinementKrylov subspace methodslinear iterative methodsmixed-precision algorithmsnumerial examplesparallel computationpreconditioning

Mathematics Subject Classification ID

Iterative numerical methods for linear systems (65F10) Numerical computation of matrix norms, conditioning, scaling (65F35) Multigrid methods; domain decomposition for boundary value problems involving PDEs (65N55) Parallel numerical computation (65Y05)

Recommendations

• Mixed precision algorithms in numerical linear algebra
• A fast and robust mixed-precision solver for the solution of sparse symmetric linear systems
• Newton's method in mixed precision
• Performance and accuracy of hardware-oriented native-, emulated- and mixed-precision solvers in FEM simulations
• Mixed Precision Iterative Refinement with Sparse Approximate Inverse Preconditioning

Cited in

(12)

• Mixed precision bisection
• Parallel decomposition version in the AISM preconditioner
• A fast and robust mixed-precision solver for the solution of sparse symmetric linear systems
• The impact of data distribution in accuracy and performance of parallel linear algebra subroutines
• Impact of mixed precision and storage layout on additive Schwarz smoothers
• RCHOL: Randomized Cholesky Factorization for Solving SDD Linear Systems
• Newton's method in mixed precision
• Mixed precision algorithms in numerical linear algebra
• M-IHS: an accelerated randomized preconditioning method avoiding costly matrix decompositions
• The stability of split-preconditioned FGMRES in four precisions
• Multigrid methods using block floating point arithmetic
• Accelerating scientific computations with mixed precision algorithms

This page was built for publication: Mixed-precision preconditionders in parallel domain decomposition solvers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3500076)