A numerically stable communication-avoiding s-step GMRES algorithm
From MaRDI portal
A numerically stable communication-avoiding \(s\)-step GMRES algorithm
Recommendations
- Varying the \(s\) in your \(s\)-step GMRES
- Avoiding communication in nonsymmetric Lanczos-based Krylov subspace methods
- Mixed-Precision Orthogonalization Scheme and Adaptive Step Size for Improving the Stability and Performance of CA-GMRES on GPUs
- A Communication Optimization Scheme for Basis Computation of Krylov Subspace Methods on Multi-GPUs
- A residual replacement strategy for improving the maximum attainable accuracy of s-step Krylov subspace methods
Cites work
- s-Step Iterative Methods for (Non)Symmetric (In)Definite Linear Systems
- A Newton basis GMRES implementation
- A set of level 3 basic linear algebra subprograms
- Accuracy and Stability of Numerical Algorithms
- Block Gram-Schmidt algorithms and their stability properties
- Communication lower bounds and optimal algorithms for numerical linear algebra
- Communication-optimal parallel and sequential QR and LU factorizations
- GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems
- scientific article; zbMATH DE number 6159604 (Why is no real title available?)
- Implementation of the GMRES Method Using Householder Transformations
- Incremental Condition Estimation
- Mixed-Precision Cholesky QR Factorization and Its Case Studies on Multicore CPU with Multiple GPUs
- Numerical stability of GMRES
- On the generation of Krylov subspace bases
- Parallel iterative S-step methods for unsymmetric linear systems
- Parallelizable restarted iterative methods for nonsymmetric linear systems. II: parallel implementation
- Parallelizable restarted iterative methods for nonsymmetric linear systems. part I: Theory
- Reducing the effect of global communication in \(\text{GMRES} (m)\) and CG on parallel distributed memory computers
- Reorthogonalization and Stable Algorithms for Updating the Gram-Schmidt QR Factorization
- Reorthogonalized block classical Gram-Schmidt
- Roundoff error analysis of the CholeskyQR2 algorithm
- s-step iterative methods for symmetric linear systems
- SLEPc
- The adaptive s-step conjugate gradient method
- The condition number of real Vandermonde, Krylov and positive definite Hankel matrices
- The University of Florida sparse matrix collection
- Varying the \(s\) in your \(s\)-step GMRES
Cited in
(4)- Mixed-Precision Orthogonalization Scheme and Adaptive Step Size for Improving the Stability and Performance of CA-GMRES on GPUs
- Variable s-step technique for planar algorithms in solving indefinite linear systems
- Reorthogonalized Pythagorean variants of block classical Gram-Schmidt
- High-performance statistical computing (HPSC): challenges, opportunities, and future directions
This page was built for publication: A numerically stable communication-avoiding \(s\)-step GMRES algorithm
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6641763)
