Mixed-Precision Algorithm for Finding Selected Eigenvalues and Eigenvectors of Symmetric and Hermitian Matrices1

The multi-precision methods commonly follow approximate-iterate scheme by first obtaining the approximate solution from a low-precision factorization and solve. Then, they iteratively refine the solution to the desired accuracy that is often as high as what is possible with traditional approaches. W...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2022 IEEE/ACM Workshop on Latest Advances in Scalable Algorithms for Large-Scale Heterogeneous Systems (ScalAH) S. 43 - 50
Hauptverfasser: Tsai, Yaohung M., Luszczek, Piotr, Dongarra, Jack
Format: Tagungsbericht
Sprache:Englisch
Japanisch
Veröffentlicht: IEEE 01.11.2022
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract The multi-precision methods commonly follow approximate-iterate scheme by first obtaining the approximate solution from a low-precision factorization and solve. Then, they iteratively refine the solution to the desired accuracy that is often as high as what is possible with traditional approaches. While targeting symmetric and Hermitian eigenvalue problems of the form Ax = λx, we revisit the SICE algorithm proposed by Dongarra et al. By applying the Sherman-Morrison formula on the diagonally-shifted tridiagonal systems, we propose an updated SICE-SM algorithm. By incorporating the latest two-stage algorithms from the PLASMA and MAGMA software libraries for numerical linear algebra, we achieved up to 3.6× speedup using the mixed-precision eigensolver with the blocked SICE-SM algorithm for iterative refinement when compared with full double complex precision solvers for the cases with a portion of eigenvalues and eigenvectors requested. 1
AbstractList The multi-precision methods commonly follow approximate-iterate scheme by first obtaining the approximate solution from a low-precision factorization and solve. Then, they iteratively refine the solution to the desired accuracy that is often as high as what is possible with traditional approaches. While targeting symmetric and Hermitian eigenvalue problems of the form Ax = λx, we revisit the SICE algorithm proposed by Dongarra et al. By applying the Sherman-Morrison formula on the diagonally-shifted tridiagonal systems, we propose an updated SICE-SM algorithm. By incorporating the latest two-stage algorithms from the PLASMA and MAGMA software libraries for numerical linear algebra, we achieved up to 3.6× speedup using the mixed-precision eigensolver with the blocked SICE-SM algorithm for iterative refinement when compared with full double complex precision solvers for the cases with a portion of eigenvalues and eigenvectors requested. 1
Author Luszczek, Piotr
Dongarra, Jack
Tsai, Yaohung M.
Author_xml – sequence: 1
  givenname: Yaohung M.
  surname: Tsai
  fullname: Tsai, Yaohung M.
  email: ytsai2@icl.utk.edu
  organization: University of Tennessee,Innovative Computing Laboratory,Knoxville,TN,USA
– sequence: 2
  givenname: Piotr
  surname: Luszczek
  fullname: Luszczek, Piotr
  email: luszczek@icl.utk.edu
  organization: University of Tennessee,Innovative Computing Laboratory,Knoxville,TN,USA
– sequence: 3
  givenname: Jack
  surname: Dongarra
  fullname: Dongarra, Jack
  email: dongarra@icl.utk.edu
  organization: University of Tennessee,Innovative Computing Laboratory,Knoxville,TN,USA
BookMark eNo1jF9rwjAUxTPYHqbbNxgjX6DuJk1i-yiiU1A20He5TW_chTaVtBvz28_9ezmH8_vBGYnr2EUS4lHBRCkon3Yem9nKOqf1RMMlAECpKzFSzlkztVNwt-K05U-qs9dEnnvuopw1xy7x8NbK0CW55FhzPModNeQHquWCjxQ_sHmnXmL83xfXpV52Qe7ObUtDYv9jV5RaHhij3OI3pF7diZuATU_3fz0W--ViP19lm5fn9Xy2yVgpazKDUyiCLm1VGqjQO4KgC5WTtyYn53MfdEDjCvAlFsEUpvRgKoUaC6SQj8XD7y0T0eGUuMV0PigAbawr8y8zkVkq
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/ScalAH56622.2022.00011
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Xplore POP ALL
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
EISBN 1665475706
9781665475709
EndPage 50
ExternalDocumentID 10024569
Genre orig-research
GroupedDBID 6IE
6IL
CBEJK
RIE
RIL
ID FETCH-LOGICAL-i1154-4a708f295b940bac6e0f2813ec543e6c3cf2fa4680c9a8f4849c04b1a2a8aef3
IEDL.DBID RIE
IngestDate Thu Jan 18 11:13:30 EST 2024
IsPeerReviewed false
IsScholarly false
Language English
Japanese
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i1154-4a708f295b940bac6e0f2813ec543e6c3cf2fa4680c9a8f4849c04b1a2a8aef3
PageCount 8
ParticipantIDs ieee_primary_10024569
PublicationCentury 2000
PublicationDate 2022-11
PublicationDateYYYYMMDD 2022-11-01
PublicationDate_xml – month: 11
  year: 2022
  text: 2022-11
PublicationDecade 2020
PublicationTitle 2022 IEEE/ACM Workshop on Latest Advances in Scalable Algorithms for Large-Scale Heterogeneous Systems (ScalAH)
PublicationTitleAbbrev SCALAH
PublicationYear 2022
Publisher IEEE
Publisher_xml – name: IEEE
Score 1.8117849
Snippet The multi-precision methods commonly follow approximate-iterate scheme by first obtaining the approximate solution from a low-precision factorization and...
SourceID ieee
SourceType Publisher
StartPage 43
SubjectTerms eigenvalue solver
Eigenvalues and eigenfunctions
hardware accelerators
Linear algebra
mixed-precision algorithms
Parallel processing
Runtime
Software algorithms
Software libraries
Symmetric matrices
Title Mixed-Precision Algorithm for Finding Selected Eigenvalues and Eigenvectors of Symmetric and Hermitian Matrices1
URI https://ieeexplore.ieee.org/document/10024569
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlZ3NT8MgGMaJWzx4UuOM3-HgtQ4oo3BczJYd3LJki9ltofCiTVy7rJvR_95CO_XiwVspBwJvCC8fv-dB6N4RR6sqFgnl4oiDVZG2HuMS2hkSs1gEnYLnp2QykYuFmjawemBhACA8PoMH_xnu8m1hdv6orEvDRaFQLdRKElHDWg31S4nqzqpe9UdVfsI8YcWCEqd3BvplmxJWjeHxP9s7QZ0f_g5Pv1eWU3QA-Rlaj7MPsNF007ji4P7bS1Ft7V9XuEo88TALfAqeBWMbsHjgdTa9ljeUWOf7cjikL3Hh8OxztfJ2WibUjvyjGD_b8TiI9kNJO2g-HMwfR1HjlxBlXlQn4joh0jHVSxUnqTYCiGOSxmB6PAZhYuOY01xIYpSWjkuuDOEp1UxLDS4-R-28yOECYZpKn9qB0anklloptDWMmyqMRpjEXaKOH63lulbEWO4H6uqP_9foyAekZvhuUHu72cEtOjTv26zc3IU4fgGQ6aMu
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlZ3NT8IwGMYbRRM9qRHjtz14nbZdKe2RGMiMQEgghhvp2rdKIhthYPS_dy1DvXjwtq6Hpn3T9O3H73kQunXE0bKKRUK5OOJgVaStx7iEdobELBZBp-C52-z35XisBhWsHlgYAAiPz-DOf4a7fJublT8qu6fholCobbTT4JyRNa5Vcb-UqPth2a9WUmYozDNWLGhxem-gX8YpYd3oHPyzxUNU_yHw8OB7bTlCW5Ado3lv-gE2GiwqXxzcenvJy8396wyXqSfuTAOhgofB2gYsbnulTa_mDQXW2aYcjukLnDs8_JzNvKGWCbWJfxbj5zvuBdl-KGgdjTrt0UMSVY4J0dTL6kRcN4l0TDVSxUmqjQDimKQxmAaPQZjYOOY0F5IYpaXjkitDeEo101KDi09QLcszOEWYptInd2B0KrmlVgptDeOmDKQRpunOUN2P1mS-1sSYbAbq_I__N2gvGfW6k-5j_-kC7fvgrIm-S1RbLlZwhXbN-3JaLK5DTL8AOGamdQ
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=2022+IEEE%2FACM+Workshop+on+Latest+Advances+in+Scalable+Algorithms+for+Large-Scale+Heterogeneous+Systems+%28ScalAH%29&rft.atitle=Mixed-Precision+Algorithm+for+Finding+Selected+Eigenvalues+and+Eigenvectors+of+Symmetric+and+Hermitian+Matrices1&rft.au=Tsai%2C+Yaohung+M.&rft.au=Luszczek%2C+Piotr&rft.au=Dongarra%2C+Jack&rft.date=2022-11-01&rft.pub=IEEE&rft.spage=43&rft.epage=50&rft_id=info:doi/10.1109%2FScalAH56622.2022.00011&rft.externalDocID=10024569