Zobrazit v EDS

RADAR: Runtime-Assisted Dead Region Management for Last-Level Caches

Uloženo v:
Podrobná bibliografie
Název: RADAR: Runtime-Assisted Dead Region Management for Last-Level Caches
Autoři: Manivannan, Madhavan, 1986, Papaefstathiou, Vasileios, 1980, Pericas, Miquel, 1979, Stenström, Per, 1957
Témata: last level cache, task-based programming models, runtime system, dead block prediction
Popis: Last-level caches bridge the speed gap between processors and the off-chip memory hierarchy and reduce energy per access. Unfortunately, last-level caches are poorly utilized because of the relatively large occurrence of dead blocks; blocks that are not accessed before becoming evicted. In particular, dead-block prediction ischallenged by unpredictable scheduling decisions made in run-time systems supporting task parallel programming models.This paper presents RADAR, a hybrid hardware/software dead-block management scheme, that can accurately predict dead blocks. It does so by inferring dead blocks from data-flow information about adress regions through functionality built into the run-time system and uses hardware support to evict dead blocks belonging to such regions. RADAR utilizes semantic information about address regions provided by programmers, past access patterns, and future reuse information available to the runtime system to accuratelypredict dead regions during execution. Our evaluation shows that RADAR outperforms state of the art dead block prediction schemes that rely only on the past eviction patterns to predict dead blocks.
Přístupová URL adresa: https://research.chalmers.se/publication/218513
Databáze: SwePub
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=EBSCO&SrcAuth=EBSCO&DestApp=WOS&ServiceName=TransferToWoS&DestLinkType=GeneralSearchSummary&Func=Links&author=Manivannan%20M
    Name: ISI
    Category: fullText
    Text: Nájsť tento článok vo Web of Science
    Icon: https://imagesrvr.epnet.com/ls/20docs.gif
    MouseOverText: Nájsť tento článok vo Web of Science
Header DbId: edsswe
DbLabel: SwePub
An: edsswe.oai.research.chalmers.se.97142077.2213.462e.9dab.ca04c4032d66
RelevancyScore: 763
AccessLevel: 6
PubType: Report
PubTypeId: report
PreciseRelevancyScore: 763.456298828125
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: RADAR: Runtime-Assisted Dead Region Management for Last-Level Caches
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Manivannan%2C+Madhavan%22">Manivannan, Madhavan</searchLink>, 1986<br /><searchLink fieldCode="AR" term="%22Papaefstathiou%2C+Vasileios%22">Papaefstathiou, Vasileios</searchLink>, 1980<br /><searchLink fieldCode="AR" term="%22Pericas%2C+Miquel%22">Pericas, Miquel</searchLink>, 1979<br /><searchLink fieldCode="AR" term="%22Stenström%2C+Per%22">Stenström, Per</searchLink>, 1957
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22last+level+cache%22">last level cache</searchLink><br /><searchLink fieldCode="DE" term="%22task-based+programming+models%22">task-based programming models</searchLink><br /><searchLink fieldCode="DE" term="%22runtime+system%22">runtime system</searchLink><br /><searchLink fieldCode="DE" term="%22dead+block+prediction%22">dead block prediction</searchLink>
– Name: Abstract
  Label: Description
  Group: Ab
  Data: Last-level caches bridge the speed gap between processors and the off-chip memory hierarchy and reduce energy per access. Unfortunately, last-level caches are poorly utilized because of the relatively large occurrence of dead blocks; blocks that are not accessed before becoming evicted. In particular, dead-block prediction ischallenged by unpredictable scheduling decisions made in run-time systems supporting task parallel programming models.This paper presents RADAR, a hybrid hardware/software dead-block management scheme, that can accurately predict dead blocks. It does so by inferring dead blocks from data-flow information about adress regions through functionality built into the run-time system and uses hardware support to evict dead blocks belonging to such regions. RADAR utilizes semantic information about address regions provided by programmers, past access patterns, and future reuse information available to the runtime system to accuratelypredict dead regions during execution. Our evaluation shows that RADAR outperforms state of the art dead block prediction schemes that rely only on the past eviction patterns to predict dead blocks.
– Name: URL
  Label: Access URL
  Group: URL
  Data: <link linkTarget="URL" linkTerm="https://research.chalmers.se/publication/218513" linkWindow="_blank">https://research.chalmers.se/publication/218513</link>
PLink https://erproxy.cvtisr.sk/sfx/access?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsswe&AN=edsswe.oai.research.chalmers.se.97142077.2213.462e.9dab.ca04c4032d66
RecordInfo BibRecord:
  BibEntity:
    Languages:
      – Text: English
    Subjects:
      – SubjectFull: last level cache
        Type: general
      – SubjectFull: task-based programming models
        Type: general
      – SubjectFull: runtime system
        Type: general
      – SubjectFull: dead block prediction
        Type: general
    Titles:
      – TitleFull: RADAR: Runtime-Assisted Dead Region Management for Last-Level Caches
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Manivannan, Madhavan
      – PersonEntity:
          Name:
            NameFull: Papaefstathiou, Vasileios
      – PersonEntity:
          Name:
            NameFull: Pericas, Miquel
      – PersonEntity:
          Name:
            NameFull: Stenström, Per
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2015
          Identifiers:
            – Type: issn-locals
              Value: CTH_SWEPUB
ResultId 1