elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Algorithmic Developments and Software Engineering for Scalable Sparse Eigensolvers in the DFG Project ESSEX

Basermann, Achim and Thies, Jonas and Alvermann, Andreas and Fehske, Holger and Lang, Bruno (2017) Algorithmic Developments and Software Engineering for Scalable Sparse Eigensolvers in the DFG Project ESSEX. In: JST/CREST International Symposium on Post Petascale System Software. JST/CREST International Symposium on Post Petascale System Software, 11.-12. Dez. 2017, Tokio, Japan.

[img] PDF
3MB

Official URL: http://www.hal.ipc.i.u-tokyo.ac.jp/isp2s2-2017/index.html

Abstract

In the German Research Foundation (DFG) project ESSEX (Equipping Sparse Solvers for Exascale), we develop scalable sparse eigensolver libraries for large quantum physics problems. Partners in ESSEX are the Universities of Erlangen, Greifswald, Wuppertal, Tokyo and Tsukuba as well as DLR. The ESSEX Sparse Solver Repository (ESSR) follows a distributed software development strategy using the distributed version control system Git and supports application driven fault tolerance. ESSR includes the kernel library GHOST (General, Hybrid, and Optimized Sparse Toolkit) and the flexible software framework PHIST for implementing iterative methods on HPC systems. PHIST (Pipelined Hybrid Iterative Solver Toolkit) has been developed containing an interface to the existing numerical software framework Trilinos originally. PHIST also includes adapters to basic building block libraries so that high-level algorithm developments can benefit from high-performance kernel implementations, e.g. sparse matrix-vector multiplication kernels. For solving huge sparse eigenproblems on extreme computing systems, we present our algorithmic approaches regarding Jacobi-Davidson methods with preconditioning, FEAST-like methods and Chebyshev filter diagonalization. The former method is applied to determine extreme eigenvalues and -vectors, while the latter two methods are well suited for solving interior eigenproblems. Moreover, we describe the software architecture of our solver toolkit PHIST.

Item URL in elib:https://elib.dlr.de/117989/
Document Type:Conference or Workshop Item (Speech)
Title:Algorithmic Developments and Software Engineering for Scalable Sparse Eigensolvers in the DFG Project ESSEX
Authors:
AuthorsInstitution or Email of AuthorsAuthors ORCID iD
Basermann, AchimAchim.Basermann (at) dlr.dehttps://orcid.org/0000-0003-3637-3231
Thies, JonasJonas.Thies (at) dlr.deUNSPECIFIED
Alvermann, AndreasErnst-Moritz-Arndt Universität GreifswaldUNSPECIFIED
Fehske, HolgerErnst-Moritz-Arndt University of GreifswaldUNSPECIFIED
Lang, Brunolang (at) math.uni-wuppertal.deUNSPECIFIED
Date:11 December 2017
Journal or Publication Title:JST/CREST International Symposium on Post Petascale System Software
Refereed publication:No
Open Access:Yes
Gold Open Access:No
In SCOPUS:No
In ISI Web of Science:No
Status:Published
Keywords:Scalable Sparse Eigensolvers; Software Engineering; Exascale Computing; Jacobi-Davidson; FEAST; Performance Engineering; Chebyshev filter diagonalization; preconditioning.
Event Title:JST/CREST International Symposium on Post Petascale System Software
Event Location:Tokio, Japan
Event Type:international Conference
Event Dates:11.-12. Dez. 2017
Organizer:JST
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Space Technology
DLR - Research area:Raumfahrt
DLR - Program:R SY - Technik für Raumfahrtsysteme
DLR - Research theme (Project):R - Vorhaben SISTEC
Location: Köln-Porz
Institutes and Institutions:Institut of Simulation and Software Technology > High Performance Computing
Institut of Simulation and Software Technology
Deposited By: Basermann, Dr.-Ing. Achim
Deposited On:12 Jan 2018 13:51
Last Modified:31 Jul 2019 20:15

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.