Using Simple Abstraction to Reinvent Computing for Parallelism.

The article presents a method for teaching computer programmers to think in terms of parallel algorithms, designed to enable improved programming for multi-processor machines. It involves an Immediate Concurrent Execution (ICE) abstraction, which is supported by explicit multi-threaded (XMT) compute...

Descripción completa

Detalles Bibliográficos
Publicado en:Communications of the ACM Vol. 54; no. 1; pp. 75 - 86
Autor principal: VISHKIN, UZI
Formato: Artículo
Publicado: Association for Computing Machinery Jan2011
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=56676293&site=ehost-live
header:
  @attributes:
    shortDbName: hlh
    uiTerm: 56676293
    longDbName: Humanities International Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00010782
        ACM
      jtl: Communications of the ACM
      issn: 00010782
      maglogo: N
    pubinfo:
      dt: Jan2011
      vid: 54
      iid: 1
      pid: 68
      pub: Association for Computing Machinery
    artinfo:
      ui:
        56676293
        10.1145/1866739.1866757
      ppf: 75
      ppct: 11
      formats:
      tig:
        atl: Using Simple Abstraction to Reinvent Computing for Parallelism.
      aug:
        au: VISHKIN, UZI
        affil: Professor, University of Maryland Institute for Advanced Computer Studies.
      su:
        Parallel computers
        Computer multitasking
        Computer programming
        Computer architecture
        Algorithms
        Random access memory
      sug:
        subj:
          Parallel computers
          Computer multitasking
          Computer programming
          Computer architecture
          Algorithms
          Random access memory
      ab: The article presents a method for teaching computer programmers to think in terms of parallel algorithms, designed to enable improved programming for multi-processor machines. It involves an Immediate Concurrent Execution (ICE) abstraction, which is supported by explicit multi-threaded (XMT) computer architecture and parallel random-access machine (PRAM) algorithmics. A workflow using these techniques is described, and the advantages of such an approach over serial algorithms are explicated. The major steps involved in the ranking/merging algorithm are discussed, and a block diagram of the XMT architecture is included.
      pubtype: Periodical
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: Y
      dt:
        @attributes:
          year: 2011
    holdings:
      @attributes:
        islocal: N