Preconditioned GMRES solver on multiple GPU architecture

Simple item page
dc.contributor.authorChen, Zhangxing (John)
dc.contributor.authorYang, B.
dc.contributor.authorLiu, Hui
dc.date.accessioned2017-03-13T21:42:47Z
dc.date.available2017-03-13T21:42:47Z
dc.date.issued2016-08
dc.description.abstractIn this paper, we analyze the preconditioned GMRES algorithm in detail and decompose it into components to implement on multiple-GPU architecture. The operations of vector updates, dot products and Sparse Matrix-Vector multiplication (SpMV) are implemented in parallel. In addition, a specific communication mechanism for SpMV is designed. The preconditioner is established on the host (CPU) and solved on the devices (GPUs). Validated by a series of numerical experiments, the GPU-based GMRES solver is effective and favorable parallel performance is achieved.en_US
dc.description.grantingagencyNSERCen_US
dc.description.refereedYesen_US
dc.description.sponsorshipIndustrial Consortium in Reservoir Simulation and Modelling, Foundation CMG, Alberta Innovates.en_US
dc.identifier.citationhttp://hdl.handle.net/11023/2512en_US
dc.identifier.doihttp://dx.doi.org/10.11575/PRISM/30183
dc.identifier.grantnumberNSERC: IRC PJ365863-12; AITF: G203000197; AIEES: 3130.en_US
dc.identifier.urihttp://hdl.handle.net/1880/51849
dc.language.isoenen_US
dc.publisherComputers and Mathematics with Applicationsen_US
dc.publisher.departmentChemical & Petroleum Engineeringen_US
dc.publisher.facultySchulich School of Engineeringen_US
dc.publisher.institutionUniversity of Calgaryen_US
dc.relation.ispartofseries72;1076-1095
dc.subjectMathematicsen_US
dc.subjectComputer Scienceen_US
dc.subjectEngineering-Petroleumen_US
dc.titlePreconditioned GMRES solver on multiple GPU architectureen_US
dc.typejournal article
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
ucalgary_2015_yang_bo.pdf
Size:
973.85 KB
Format:
Adobe Portable Document Format
Description:
Electronic Theses
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.95 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Schulich School of Engineering Research & Publications