Please use this identifier to cite or link to this item: https://hdl.handle.net/11147/14911
Full metadata record
DC FieldValueLanguage
dc.contributor.authorKaya,E.-
dc.contributor.authorKaradaş,O.F.-
dc.contributor.authorÖz,I.-
dc.date.accessioned2024-10-25T23:27:20Z-
dc.date.available2024-10-25T23:27:20Z-
dc.date.issued2021-
dc.identifier.issn1613-0073-
dc.identifier.urihttps://hdl.handle.net/11147/14911-
dc.description.abstractWith the widespread use of GPU architectures in general-purpose computations, evaluating the soft error vulnerability of GPGPU programs and employing efficient fault tolerance techniques for more reliable execution becomes more prominent. Performing full redundancy, based on the redundant execution of the complete program, results in resource consumption and performance loss as well as energy inefficiency. Therefore, determining the most error-prone regions of the target program code and replicating only those parts maintains both high performance and acceptable error rates. In this study, we propose a partial redundant multithreading mechanism based on the soft error vulnerability of GPGPU applications and perform a trade-off analysis between performance and reliability. Firstly, we perform fault injection experiments to evaluate the SDC rates for each kernel function. Then, based on the outcome of the fault injection experiments, we determine the kernel function to-be-replicated. According to the pragmas denoting the redundancy points in the source code, our custom LLVM pass generates the code that enables the redundant execution for the specified code region. We evaluate both the reliability and performance of the redundant execution scenarios measuring the execution time of the redundant program generated by our compiler-managed redundancy technique. Our results demonstrate that protecting only the most vulnerable kernel functions enables high reliability without hurting the performance significantly. © 2021 The Authors.en_US
dc.description.sponsorshipCERCIRAS COST, (CA19135); COST Association; TÜBÝTAK; Türkiye Bilimsel ve Teknolojik Araştirma Kurumu, TÜBITAK, (119E011)en_US
dc.language.isoenen_US
dc.publisherCEUR-WSen_US
dc.relation.ispartofCEUR Workshop Proceedings -- 1st Workshop on Connecting Education and Research Communities for an Innovative Resource Aware Society, CERCIRAS 2021 -- 2 September 2021 -- Novi Sad -- 179621en_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectFault injectionen_US
dc.subjectGPGPU programsen_US
dc.subjectRedundant executionen_US
dc.subjectSoft error reliabilityen_US
dc.titleEvaluating Performance and Reliability of Selective Redundant Multithreading for GPGPU Applicationsen_US
dc.typeConference Objecten_US
dc.departmentIzmir Institute of Technologyen_US
dc.identifier.volume3145en_US
dc.identifier.scopus2-s2.0-85131328680-
dc.relation.publicationcategoryKonferans Öğesi - Uluslararası - Kurum Öğretim Elemanıen_US
dc.identifier.doi[SCOPUS-DOI-BELIRLENECEK-9]-
dc.authorscopusid57727235800-
dc.authorscopusid57236778300-
dc.authorscopusid37097877800-
dc.identifier.wosqualityN/A-
dc.identifier.scopusqualityQ4-
item.fulltextNo Fulltext-
item.grantfulltextnone-
item.languageiso639-1en-
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.cerifentitytypePublications-
item.openairetypeConference Object-
Appears in Collections:Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
Show simple item record



CORE Recommender

Page view(s)

6
checked on Nov 18, 2024

Google ScholarTM

Check




Altmetric


Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.