An efficient GPU version of the preconditioned GMRES method

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: An efficient GPU version of the preconditioned GMRES method
Συγγραφείς: Universitat Politècnica de València. Departamento de Informática de Sistemas y Computadores - Departament d'Informàtica de Sistemes i Computadors, Aliaga, Jose I., Dufrechou, Ernesto, Ezzatti, Pablo, Quintana-Orti, Enrique S.
Στοιχεία εκδότη: Springer-Verlag Hispana 2019-03
Τύπος εγγράφου: Mixed materials
Περίληψη: [EN] In a large number of scientific applications, the solution of sparse linear systems is the stage that concentrates most of the computational effort. This situation has motivated the study and development of several iterative solvers, among which preconditioned Krylov subspace methods occupy a place of privilege. In a previous effort, we developed a GPU-aware version of the GMRES method included in ILUPACK, a package of solvers distinguished by its inverse-based multilevel ILU preconditioner. In this work, we study the performance of our previous proposal and integrate several enhancements in order to mitigate its principal bottlenecks. The numerical evaluation shows that our novel proposal can reach important run-time reductions.
Όροι ευρετηρίου: GPUs, GMRES, Sparse triangular solver, MGSO, ARQUITECTURA Y TECNOLOGIA DE COMPUTADORES, info:eu-repo/semantics/article
Σύνδεσμος: http://hdl.handle.net/10251/160601
Διαθεσιμότητα: Open access content. Open access content
http://rightsstatements.org/vocab/InC/1.0
info:eu-repo/semantics/openAccess
Σημείωση: TEXT
English
Other Numbers: UPV oai:riunet.upv.es:10251/160601
urn:issn:0920-8542
info:doi:10.1007/s11227-018-2658-1
https://riunet.upv.es/bitstream/10251/160601/8/Aliaga;Dufrechou;Ezzatti - An efficient GPU version of the preconditioned GMRES method.pdf.jpg
1258890671
Πηγή συνεισφοράς: UNIVERSITAT POLITECNICA DE VALENCIA
From OAIster®, provided by the OCLC Cooperative.
Αριθμός Καταχώρησης: edsoai.on1258890671
Βάση Δεδομένων: OAIster
Περιγραφή
Η περιγραφή δεν είναι διαθέσιμη