Pipelined, flexible Krylov subspace methods
From MaRDI portal
Abstract: We present variants of the Conjugate Gradient (CG), Conjugate Residual (CR), and Generalized Minimal Residual (GMRES) methods which are both pipelined and flexible. These allow computation of inner products and norms to be overlapped with operator and nonlinear or nondeterministic preconditioner application.The methods are hence aimed at hiding network latencies and synchronizations which can become computational bottlenecks in Krylov methods on extreme-scale systems or in the strong-scaling limit. The new variants are not arithmetically equivalent to their base flexible Krylov methods, but are chosen to be similarly performant in a realistic use case, the application of strong nonlinear preconditioners to large problems which require many Krylov iterations. We provide scalable implementations of our methods as contributions to the PETSc package and demonstrate their effectiveness with practical examples derived from models of mantle convection and lithospheric dynamics with heterogeneous viscosity structure. These represent challenging problems where multiscale nonlinear preconditioners are required for the current state-of-the-art algorithms, and are hence amenable to acceleration with our new techniques. Large-scale tests are performed in the strong-scaling regime on a contemporary leadership supercomputer, where speedups approaching, and even exceeding can be observed. We conclude by analyzing our new methods with a performance model targeted at future exascale machines.
Recommendations
- The numerical stability analysis of pipelined conjugate gradient methods: historical context and methodology
- Hiding global communication latency in the GMRES algorithm on massively parallel machines
- Flexible Inner-Outer Krylov Subspace Methods
- Scalable linear solvers based on enlarged Krylov subspaces with dynamic reduction of search directions
- A Comparison of Preconditioned Nonsymmetric Krylov Methods on a Large-Scale MIMD Machine
Cites work
- A Black Box Generalized Conjugate Gradient Solver with Inner Iterations and Variable-Step Preconditioning
- A Flexible Inner-Outer Preconditioned GMRES Algorithm
- A generalized conjugate gradient, least square method
- A Restricted Additive Schwarz Preconditioner for General Sparse Linear Systems
- A scalable, matrix-free multigrid preconditioner for finite element discretizations of heterogeneous Stokes flow
- A stabilized finite element method for the Stokes problem based on polynomial pressure projections
- Accurately measuring overhead, communication time and progression of blocking and nonblocking collective operations at massive scale
- Comparison of different propagation steps for lattice Boltzmann methods
- Composing scalable nonlinear algebraic solvers
- Flexible conjugate gradients
- GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems
- Hiding global communication latency in the GMRES algorithm on massively parallel machines
- scientific article; zbMATH DE number 3120437 (Why is no real title available?)
- scientific article; zbMATH DE number 1953444 (Why is no real title available?)
- scientific article; zbMATH DE number 949303 (Why is no real title available?)
- Inexact Preconditioned Conjugate Gradient Method with Inner-Outer Iteration
- Iterative Krylov Methods for Large Linear Systems
- Methods of conjugate gradients for solving linear systems
- NOTES ON THE SOLUTION OF ALGEBRAIC LINEAR SIMULTANEOUS EQUATIONS
- Pipelined iterative solvers with kernel fusion for graphics processing units
- Preconditioning
- s-step iterative methods for symmetric linear systems
- Some History of the Conjugate Gradient and Lanczos Algorithms: 1948–1976
- The Conjugate Residual Method for Constrained Minimization Problems
- Variational Iterative Methods for Nonsymmetric Systems of Linear Equations
Cited in
(10)- GMRES algorithms over 35 years
- Blendenpik: Supercharging LAPACK's Least-Squares Solver
- Flexible Inner-Outer Krylov Subspace Methods
- The numerical stability analysis of pipelined conjugate gradient methods: historical context and methodology
- Hiding global communication latency in the GMRES algorithm on massively parallel machines
- Scalable linear solvers based on enlarged Krylov subspaces with dynamic reduction of search directions
- Analyzing the effect of local rounding error propagation on the maximal attainable accuracy of the pipelined conjugate gradient method
- Efficient AMG reduction-based preconditioners for structural mechanics
- A multigrid reduction framework for domains with symmetries
- On applying deflation and flexible preconditioning to the adaptive simpler GMRES method for Sylvester tensor equations
This page was built for publication: Pipelined, flexible Krylov subspace methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2818249)
