IDENTIFYING PERFORMANCE BOTTLENECKS IN WORK-STEALING COMPUTATIONS.

The article discusses the authors' implementation of performance analysis methods and tools to resolve a work stealing computation's performance obstacles. Their strategy was implemented in HPCToolkit and it addresses the problem of pinpointing scalability bottlenecks and providing user-level insigh...

Descripción completa

Detalles Bibliográficos
Publicado en:Computer (00189162) Vol. 42; no. 12; pp. 44 - 51
Autores principales: Tallent, Nathan R., Mellor-Crummey, John M.
Formato: Artículo
Publicado: IEEE Dec2009
Materias:
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=hlh&AN=47400247&site=ehost-live
header:
  @attributes:
    shortDbName: hlh
    uiTerm: 47400247
    longDbName: Humanities International Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00189162
        PUT
      jtl: Computer (00189162)
      issn: 00189162
      maglogo: N
    pubinfo:
      dt: Dec2009
      vid: 42
      iid: 12
      pid: 13605
      pub: IEEE
    artinfo:
      ui:
        47400247
        10.1109/MC.2009.396
      ppf: 44
      ppct: 7
      formats:
      tig:
        atl: IDENTIFYING PERFORMANCE BOTTLENECKS IN WORK-STEALING COMPUTATIONS.
      aug:
        au:
          Tallent, Nathan R.
          Mellor-Crummey, John M.
        affil: Rice University
      su:
        Bottlenecks (Manufacturing)
        Programming languages
        Computer networks
        Electronic data processing
        Computers
      sug:
        subj:
          Bottlenecks (Manufacturing)
          Programming languages
          Computer networks
          Electronic data processing
          Computers
      ab: The article discusses the authors' implementation of performance analysis methods and tools to resolve a work stealing computation's performance obstacles. Their strategy was implemented in HPCToolkit and it addresses the problem of pinpointing scalability bottlenecks and providing user-level insight on their causes. The key properties of a parallel programming language are noted. The strategy also quantifies parallel idleness and overhead and incurs no measurement overhead.
      pubtype: Academic Journal
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: Y
      dt:
        @attributes:
          year: 2009
    holdings:
      @attributes:
        islocal: N