Pretražite po imenu i prezimenu autora, mentora, urednika, prevoditelja

Napredna pretraga

Pregled bibliografske jedinice broj: 980374

Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework


Bosner, Nela; Bujanović, Zvonimir; Drmač, Zlatko
Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework // SIAM journal on scientific computing, 40 (2018), 4; C605-C633 doi:10.1137/17m1144465 (međunarodna recenzija, članak, znanstveni)


CROSBI ID: 980374 Za ispravke kontaktirajte CROSBI podršku putem web obrasca

Naslov
Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework

Autori
Bosner, Nela ; Bujanović, Zvonimir ; Drmač, Zlatko

Izvornik
SIAM journal on scientific computing (1064-8275) 40 (2018), 4; C605-C633

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
GPU ; Hessenberg matrix ; interpolatory model reduction ; parallel solver ; pseudospectrum ; shifted linear systems ; transfer function

Sažetak
This paper proposes a combination of a hybrid CPU--GPU and a pure GPU software implementation of a direct algorithm for solving shifted linear systems $(A-\sigma I)X=B$ with a large number of complex shifts $\sigma$ and multiple right-hand sides. Such problems often appear, e.g., in control theory when evaluating the transfer function, or as a part of an algorithm performing interpolatory model reduction, as well as when computing pseudospectra and structured pseudospectra, or solving large linear systems of ordinary differential equations. The proposed algorithm first jointly reduces the general full $n\times n$ matrix $A$ and the $n\times m$ full right-hand side matrix $B$ to the controller Hessenberg canonical form that facilitates efficient solution: $A$ is transformed to a so-called $m$-Hessenberg form, and $B$ is made upper triangular. This is implemented as a blocked highly parallel CPU--GPU hybrid algorithm ; individual blocks are reduced by the CPU, and the necessary updates of the rest of the matrix are split among the cores of the CPU and the GPU. To enhance parallelization, the reduction and the updates are overlapped. In the next phase, the reduced $m$-Hessenberg--triangular systems are solved entirely on the GPU, with shifts divided into batches. The benefits of such load distribution are demonstrated by numerical experiments. In particular, we show that our proposed implementation provides an excellent basis for efficient implementations of computational methods in systems and control theory, from evaluation of transfer function to the interpolatory model reduction.

Izvorni jezik
Engleski

Znanstvena područja
Matematika



POVEZANOST RADA


Projekti:
HRZZ-IP-2013-11-9345 - Matematičko modeliranje, analiza i računanje s primjenama na kompleksne mehaničke sustave (MMACACMS) (Drmač, Zlatko, HRZZ - 2013-11) ( CroRIS)

Ustanove:
Prirodoslovno-matematički fakultet, Matematički odjel, Zagreb,
Prirodoslovno-matematički fakultet, Zagreb

Profili:

Avatar Url Zlatko Drmač (autor)

Avatar Url Nela Bosner (autor)

Avatar Url Zvonimir Bujanović (autor)

Poveznice na cjeloviti tekst rada:

doi epubs.siam.org doi.org

Citiraj ovu publikaciju:

Bosner, Nela; Bujanović, Zvonimir; Drmač, Zlatko
Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework // SIAM journal on scientific computing, 40 (2018), 4; C605-C633 doi:10.1137/17m1144465 (međunarodna recenzija, članak, znanstveni)
Bosner, N., Bujanović, Z. & Drmač, Z. (2018) Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework. SIAM journal on scientific computing, 40 (4), C605-C633 doi:10.1137/17m1144465.
@article{article, author = {Bosner, Nela and Bujanovi\'{c}, Zvonimir and Drma\v{c}, Zlatko}, year = {2018}, pages = {C605-C633}, DOI = {10.1137/17m1144465}, keywords = {GPU, Hessenberg matrix, interpolatory model reduction, parallel solver, pseudospectrum, shifted linear systems, transfer function}, journal = {SIAM journal on scientific computing}, doi = {10.1137/17m1144465}, volume = {40}, number = {4}, issn = {1064-8275}, title = {Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework}, keyword = {GPU, Hessenberg matrix, interpolatory model reduction, parallel solver, pseudospectrum, shifted linear systems, transfer function} }
@article{article, author = {Bosner, Nela and Bujanovi\'{c}, Zvonimir and Drma\v{c}, Zlatko}, year = {2018}, pages = {C605-C633}, DOI = {10.1137/17m1144465}, keywords = {GPU, Hessenberg matrix, interpolatory model reduction, parallel solver, pseudospectrum, shifted linear systems, transfer function}, journal = {SIAM journal on scientific computing}, doi = {10.1137/17m1144465}, volume = {40}, number = {4}, issn = {1064-8275}, title = {Parallel Solver for Shifted Systems in a Hybrid CPU--GPU Framework}, keyword = {GPU, Hessenberg matrix, interpolatory model reduction, parallel solver, pseudospectrum, shifted linear systems, transfer function} }

Časopis indeksira:


  • Current Contents Connect (CCC)
  • Web of Science Core Collection (WoSCC)
    • Science Citation Index Expanded (SCI-EXP)
    • SCI-EXP, SSCI i/ili A&HCI
  • Scopus


Citati:





    Contrast
    Increase Font
    Decrease Font
    Dyslexic Font