Näytä suppeat kuvailutiedot

adBoost: Thermal Aware Performance Boosting through Dark Silicon Patterning

Kanduri A.; Liljeberg P.; Haghbayan M.; Jantsch A.; Rahmani A.; Shafique M.

dc.contributor.authorKanduri A.
dc.contributor.authorLiljeberg P.
dc.contributor.authorHaghbayan M.
dc.contributor.authorJantsch A.
dc.contributor.authorRahmani A.
dc.contributor.authorShafique M.
dc.date.accessioned2022-10-27T12:21:08Z
dc.date.available2022-10-27T12:21:08Z
dc.identifier.urihttps://www.utupub.fi/handle/10024/158008
dc.description.abstract<p>Limitation on power budget in many-core systems leaves a fraction of on-chip resources inactive, referred to as dark silicon. An efficient run-time system that manages critical interlinked parameters of power, performance and temperature can enhance resource utilization and mitigate dark silicon. In this paper, we present a run-time resource management system for thermal aware performance boosting. The run-time system employs a dark silicon aware run-time application mapping approach that patterns active cores alongside the inactive cores in order to evenly distribute power density across the chip. This provides enough thermal headroom for boosting the frequency of active cores upon performance surges. Lower operating temperatures from patterning also allows sustaining the boosting for longer periods, improving the performance further. We design a thermal aware controller for performance boosting that decides on allocation and efficient utilization of power budget and thermal headroom resources, obtained from patterning. Our strategy yields up to 37% better throughput, 29% lower waiting time and up to 2X longer boosting periods, in comparison with other state-of-the-art run-time mapping policies.<br /></p>
dc.language.isoen
dc.publisherIEEE Computer Society
dc.titleadBoost: Thermal Aware Performance Boosting through Dark Silicon Patterning
dc.identifier.urnURN:NBN:fi-fe2021042719001
dc.relation.volume67
dc.contributor.organizationfi=PÄÄT Tietojenkäsittelytiede|en=PÄÄT Computer Science|
dc.contributor.organizationfi=PÄÄT Terveysteknologia|en=PÄÄT Health Technology|
dc.contributor.organization-code2606803
dc.contributor.organization-code2606808
dc.converis.publication-id30709398
dc.converis.urlhttps://research.utu.fi/converis/portal/Publication/30709398
dc.format.pagerange1077
dc.format.pagerange1062
dc.identifier.eissn1557-9956
dc.identifier.jour-issn0018-9340
dc.okm.affiliatedauthorLiljeberg, Pasi
dc.okm.affiliatedauthorHaghbayan, Mohammadhashem
dc.okm.affiliatedauthorKanduru, Srinivasa
dc.okm.discipline113 Tietojenkäsittely ja informaatiotieteetfi_FI
dc.okm.discipline113 Computer and information sciencesen_GB
dc.okm.internationalcopublicationinternational co-publication
dc.okm.internationalityInternational publication
dc.okm.typeJournal article
dc.publisher.countryYhdysvallat (USA)fi_FI
dc.publisher.countryUnited Statesen_GB
dc.publisher.country-codeUS
dc.relation.doi10.1109/TC.2018.2805683
dc.relation.ispartofjournalIEEE Transactions on Computers
dc.relation.issue8
dc.year.issued2018


Aineistoon kuuluvat tiedostot

Thumbnail

Aineisto kuuluu seuraaviin kokoelmiin

Näytä suppeat kuvailutiedot