Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning

Kolev, T V; Vassilevski, P S

doi:10.2172/900179

Title: Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning

Full Record
Other Related Research

Abstract

This report describes an application of the recently developed H{sup 1}-auxiliary space preconditioner for H(curl) problems to the Maxwell eigenvalue problem. The auxiliary space method based on the new (HX) finite element space decomposition introduced in [7], was implemented in the hypre library, [10, 11] under the name AMS. The eigensolver considered in the present paper, referred to as the AME, is an extension of the AMS. It is based on the locally optimal block eigensolver LOBPCG [9] and the parallel AMG (algebraic multigrid) solver BoomerAMG [2] from the hypre library. AME is designed to compute a block of few minimal nonzero eigenvalues and eigenvectors, for general unstructured finite element discretizations utilizing the lowest order Nedelec elements. The main goal of the current report is to document the usage of AME and to illustrate its parallel scalability.

Authors:: Kolev, T V; Vassilevski, P S

Publication Date:: Wed Nov 15 00:00:00 EST 2006

Research Org.:: Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)

Sponsoring Org.:: USDOE

OSTI Identifier:: 900179

Report Number(s):: UCRL-TR-226197
TRN: US200709%%568

DOE Contract Number:: W-7405-ENG-48

Resource Type:: Technical Report

Country of Publication:: United States

Language:: English

Subject:: 99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; EIGENVALUES; EIGENVECTORS; LAWRENCE LIVERMORE NATIONAL LABORATORY

Citation Formats


                    Kolev, T V, and Vassilevski, P S. Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning.  United States: N. p., 2006. 
        Web.  doi:10.2172/900179.

Copy to clipboard


                    Kolev, T V, & Vassilevski, P S. Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning.  United States.  https://doi.org/10.2172/900179

Copy to clipboard


                    Kolev, T V, and Vassilevski, P S. 2006.  
        "Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning".  United States.  https://doi.org/10.2172/900179.  https://www.osti.gov/servlets/purl/900179.

Copy to clipboard


                    
@article{osti_900179,

  title        = {Parallel eigensolver for H(curl) problems using H1-auxiliary space AMG preconditioning},

  author       = {Kolev, T V and Vassilevski, P S},

  abstractNote = {This report describes an application of the recently developed H{sup 1}-auxiliary space preconditioner for H(curl) problems to the Maxwell eigenvalue problem. The auxiliary space method based on the new (HX) finite element space decomposition introduced in [7], was implemented in the hypre library, [10, 11] under the name AMS. The eigensolver considered in the present paper, referred to as the AME, is an extension of the AMS. It is based on the locally optimal block eigensolver LOBPCG [9] and the parallel AMG (algebraic multigrid) solver BoomerAMG [2] from the hypre library. AME is designed to compute a block of few minimal nonzero eigenvalues and eigenvectors, for general unstructured finite element discretizations utilizing the lowest order Nedelec elements. The main goal of the current report is to document the usage of AME and to illustrate its parallel scalability.},

  doi          = {10.2172/900179},

  url          = {https://www.osti.gov/biblio/900179},
  journal      = {},
number       = ,

  volume       = ,

  place        = {United States},

  year         = {Wed Nov 15 00:00:00 EST 2006},

  month        = {Wed Nov 15 00:00:00 EST 2006}

}

Copy to clipboard

Technical Report:

View Technical Report (2.79 MB)

https://doi.org/10.2172/900179

Save / Share:

Export Metadata

Save to My Library

Similar records in OSTI.GOV collections:

Parallel H1-based auxiliary space AMG solver for H(curl) problems

Technical Report Kolev, T; Vassilevski, P

This report describes a parallel implementation of the auxiliary space methods for definite Maxwell problems proposed in [4]. The solver, named AMS, extends our previous study [7]. AMS uses ParCSR sparse matrix storage and the parallel AMG (algebraic multigrid) solver BoomerAMG [1] from the hypre library. It is designed for general unstructured finite element discretizations of (semi)definite H(curl) problems discretized by Nedelec elements. We document the usage of AMS and illustrate its parallel scalability and overall performance.
https://doi.org/10.2172/897951

Full Text Available
Parallel Auxiliary Space AMG for H(Curl) Problems

Journal Article Kolev, Tzanio; Vassilevski, Panayot - Journal of Computational Mathematics

In this paper we review a number of auxiliary space based preconditioners for the second order definite and semi-definite Maxwell problems discretized with the lowest order Nedelec finite elements. In addition, we discuss the parallel implementation of the most promising of these methods, the ones derived from the recent Hiptmair-Xu (HX) auxiliary space decomposition. An extensive set of numerical experiments demonstrate the scalability of our implementation on large-scale H(curl) problems.
https://doi.org/10.4208/jcm.2009.27.5.013

Full Text Available
Parallel auxiliary space AMG for definite Maxwell problems

Conference Kolev, T; Vassilevski, P

Motivated by the needs of large multi-physics simulation codes, we are interested in algebraic solvers for the linear systems arising in time-domain electromagnetic simulations. Our focus is on finite element discretization, and we are developing scalable parallel preconditioners which employ only fine-grid information, similar to algebraic multigrid (AMG) for diffusion problems. In the last few years, the search for efficient algebraic preconditioners for H(curl) bilinear forms has intensified. The attempts to directly construct AMG methods had some success, see [12, 1, 7]. Exploiting available multilevel methods on auxiliary mesh for the same bilinear form led to efficient auxiliary mesh preconditionersmore »« less
Full Text Available
On ParELAG's Parallel Element-based Algebraic Multigrid and its MFEM Miniapps for H(curl) and H(div) Problems: a report including lowest and next to the lowest order numerical results

Technical Report Kalchev, Delyan; Vassilevski, Panayot; Villa, Umberto

This paper presents the utilization of element-based algebraic multigrid (AMGe) hierarchies, implemented in the ParELAG (Parallel Element Agglomeration Algebraic Multigrid Upscaling and Solvers) library, to produce multilevel preconditioners and solvers for H(curl) and H(div) formulations. This involves the construction of hierarchies of compatible nested spaces, forming an exact de Rham sequence on each level. This allows the application of hybrid smoothers on all levels and AMS (Auxiliary-space Maxwell Solver) or ADS (Auxiliary-space Divergence Solver) on the coarsest levels, obtaining complete multigrid cycles. Numerical results are presented, showing the parallel performance of the proposed methods. As a part of the exposition,more »« less
https://doi.org/10.2172/1807757

Full Text Available
A comparison of eigensolvers for large-scale 3D modal analysis using AMG-preconditioned iterative methods.

Journal Article Arbenz, Peter; Lehoucq, Richard; Hetmaniuk, Ulrich; ... - Proposed for publication in International Journal for Numerical Methods in Engineering.

The goal of our paper is to compare a number of algorithms for computing a large number of eigenvectors of the generalized symmetric eigenvalue problem arising from a modal analysis of elastic structures. The shift-invert Lanczos algorithm has emerged as the workhorse for the solution of this generalized eigenvalue problem; however, a sparse direct factorization is required for the resulting set of linear equations. Instead, our paper considers the use of preconditioned iterative methods. We present a brief review of available preconditioned eigensolvers followed by a numerical comparison on three problems using a scalable algebraic multigrid (AMG) preconditioner.
https://doi.org/10.1002/nme.1365

Similar Records