High-performance generation of the Hamiltonian and overlap matrices in FLAPW methods

DOI10.1016/J.CPC.2016.10.003zbMATH Open1376.65067arXiv1602.06589OpenAlexW2282110028MaRDI QIDQ1686969FDOQ1686969

Authors: Edoardo Di Napoli, Elmar Peise, Markus Hrywniak, Paolo Bientinesi 

Publication date: 18 December 2017

Published in: Computer Physics Communications (Search for Journal in Brave)

Abstract: One of the greatest efforts of computational scientists is to translate the mathematical model describing a class of physical phenomena into large and complex codes. Many of these codes face the difficulty of implementing the mathematical operations in the model in terms of low level optimized kernels offering both performance and portability. Legacy codes suffer from the additional curse of rigid design choices based on outdated performance metrics (e.g. minimization of memory footprint). Using a representative code from the Materials Science community, we propose a methodology to restructure the most expensive operations in terms of an optimized combination of dense linear algebra kernels. The resulting algorithm guarantees an increased performance and an extended life span of this code enabling larger scale simulations.

Full work available at URL: https://arxiv.org/abs/1602.06589

Recommendations

• Parallelization of the FLAPW method
• Efficient computation of Hamiltonian matrix elements between non-orthogonal Slater determinants
• High Performance Computing for Eigenvalue Solver in Density-Matrix Renormalization Group Method: Parallelization of the Hamiltonian Matrix-Vector Multiplication
• Efficient calculation of the Coulomb matrix and its expansion around k=0 within the FLAPW method
• Elimination of the linearization error and improved basis-set convergence within the FLAPW method
• Efficient computation of the coupling matrix in time-dependent density functional theory
• Analytic spin-orbit coupling matrix element formulae in FLAPW calculations

zbMATH Keywords

density functional theoryhigh-performance computingperformance portabilitydense linear algebraFLAPW methodsmatrix generation

Mathematics Subject Classification ID

Complexity and performance of numerical algorithms (65Y20) Computational methods for problems pertaining to quantum theory (81-08) Many-body theory; quantum Hall effect (81V70) Applications to the sciences (65Z05)

Cites Work

• Iterative diagonalization in augmented plane wave based methods in electronic structure calculations
• Improving the efficiency of FP-LAPW calculations
• Solid state physics.
• Elemental, a new framework for distributed memory dense matrix computations
• Block iterative eigensolvers for sequences of correlated eigenvalue problems
• A primer in density functional theory. Lectures from the second Coimbra school on computational physics, Caramulo Mountains, Portugal, August 28 --September 1, 2001
• The WY Representation for Products of Householder Matrices
• High-performance generation of the Hamiltonian and overlap matrices in FLAPW methods
• Title not available (Why is that?)
• Parallelization of the FLAPW method
• Accumulating Householder transformations, revisited

Cited In (5)

• Materials by design and the exciting role of quantum computation/simulation
• Efficient calculation of the Coulomb matrix and its expansion around k=0 within the FLAPW method
• The LAPW method with eigendecomposition based on the Hari-Zimmermann generalized hyperbolic SVD
• High-performance generation of the Hamiltonian and overlap matrices in FLAPW methods
• Parallelization of the FLAPW method

Uses Software

• WIEN2k
• Elemental
• wien-speedup
• FLEUR

This page was built for publication: High-performance generation of the Hamiltonian and overlap matrices in FLAPW methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1686969)