A projected preconditioned conjugate gradient algorithm for computing many extreme eigenpairs of a Hermitian matrix
From MaRDI portal
Publication:349770
DOI10.1016/J.JCP.2015.02.030zbMATH Open1349.65133arXiv1407.7506OpenAlexW1963825920MaRDI QIDQ349770FDOQ349770
Authors: Eugene Vecharynski, Chao Yang, John E. Pask 
Publication date: 5 December 2016
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: We present an iterative algorithm for computing an invariant subspace associated with the algebraically smallest eigenvalues of a large sparse or structured Hermitian matrix A. We are interested in the case in which the dimension of the invariant subspace is large (e.g., over several hundreds or thousands) even though it may still be small relative to the dimension of A. These problems arise from, for example, density functional theory based electronic structure calculations for complex materials. The key feature of our algorithm is that it performs fewer Rayleigh--Ritz calculations compared to existing algorithms such as the locally optimal precondition conjugate gradient or the Davidson algorithm. It is a block algorithm, hence can take advantage of efficient BLAS3 operations and be implemented with multiple levels of concurrency. We discuss a number of practical issues that must be addressed in order to implement the algorithm efficiently on a high performance computer.
Full work available at URL: https://arxiv.org/abs/1407.7506
Recommendations
- Computing several eigenpairs of Hermitian problems by conjugate gradient iterations
- Preconditioned conjugate gradient method for the sparse generalized eigenvalue problem in electronic structure calculations
- scientific article; zbMATH DE number 1507105
- Preconditioned eigensolvers for large-scale nonlinear Hermitian eigenproblems with variational characterizations. I. Extreme eigenvalues
- A novel algorithm for calculation of the extreme eigenvalues of large Hermitian matrices
Hermitian, skew-Hermitian, and related matrices (15B57) Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Preconditioners for iterative methods (65F08)
Cites Work
- ScaLAPACK Users' Guide
- Toward the optimal preconditioned eigensolver: Locally optimal block preconditioned conjugate gradient method
- The Gradient Projection Method for Nonlinear Programming. Part I. Linear Constraints
- Matrix Analysis
- Numerical Optimization
- The Gradient Projection Method for Nonlinear Programming. Part II. Nonlinear Constraints
- Numerical methods for large eigenvalue problems
- Convex programming in Hilbert space
- Some Metric Inequalities in the Space of Matrices
- Title not available (Why is that?)
- Title not available (Why is that?)
- Elemental, a new framework for distributed memory dense matrix computations
- A Trace Minimization Algorithm for the Generalized Eigenvalue Problem
- The iterative calculation of a few of the lowest eigenvalues and corresponding eigenvectors of large real-symmetric matrices
- The trace minimization method for the symmetric generalized eigenvalue problem
- A spectrum slicing method for the Kohn-Sham problem
- Block Locally Optimal Preconditioned Eigenvalue Xolvers (BLOPEX) in Hypre and PETSc
- A Simultaneous Iteration Algorithm for Real Matrices
- Elements of structural optimization.
Cited In (27)
- Gradient flow based Kohn-Sham density functional theory model
- Convergence analysis of direct minimization and self-consistent iterations
- Triangularized orthogonalization-free method for solving extreme eigenvalue problems
- Convergence theory for preconditioned eigenvalue solvers in a nutshell
- Solution of the \(k\)-th eigenvalue problem in large-scale electronic structure calculations
- Numerical methods for Kohn–Sham density functional theory
- A Fast Iterative Algorithm for Near-Diagonal Eigenvalue Problems
- TRPL+K: Thick-Restart Preconditioned Lanczos+K Method for Large Symmetric Eigenvalue Problems
- Fast eigenpairs computation with operator adapted wavelets and hierarchical subspace correction
- Computation of Large Invariant Subspaces Using Polynomial Filtered Lanczos Iterations with Applications in Density Functional Theory
- ELSI -- an open infrastructure for electronic structure solvers
- Gaussian basis implementation of the charge patching method
- On Chebyshev-Davidson method for symmetric generalized eigenvalue problems
- Coordinatewise descent methods for leading eigenvalue problem
- On the computation of large-scale self-consistent-field iterations
- A thick-restart Lanczos algorithm with polynomial filtering for Hermitian eigenvalue problems
- On flexible block Chebyshev-Davidson method for solving symmetric generalized eigenvalue problems
- Projective block Lanczos algorithm for dense, Hermitian eigensystems
- An unconditionally energy-stable and orthonormality-preserving iterative scheme for the Kohn-Sham gradient flow based model
- Weighted Trace-Penalty Minimization for Full Configuration Interaction
- Efficient block preconditioned eigensolvers for linear response time-dependent density functional theory
- A conjugate gradient method for electronic structure calculations
- On relaxed filtered Krylov subspace method for non-symmetric eigenvalue problems
- Computing several eigenpairs of Hermitian problems by conjugate gradient iterations
- KSSOLV 2.0: an efficient Matlab toolbox for solving the Kohn-Sham equations with plane-wave basis set
- Title not available (Why is that?)
- Preconditioned eigensolvers for large-scale nonlinear Hermitian eigenproblems with variational characterizations. I. Extreme eigenvalues
Uses Software
This page was built for publication: A projected preconditioned conjugate gradient algorithm for computing many extreme eigenpairs of a Hermitian matrix
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q349770)
