• DocumentCode
    2067492
  • Title

    Models of Computation for Networks on Chip

  • Author

    Jantsch, Axel

  • Author_Institution
    R. Inst. of Technol., Stockholm
  • fYear
    2006
  • fDate
    28-30 June 2006
  • Firstpage
    165
  • Lastpage
    178
  • Abstract
    Networks on chip platforms offer the opportunity to introduce a new abstraction level that defines a set of platform services with performance and power characteristics. By making the implementation of these services entirely irrelevant for system design, an effective separation of system design from component design can be achieved. We discuss the principles to formulate network-on-chip services to establish an abstract computational model that exposes all relevant properties of the platform´s functionality, performance and power consumption while hiding all irrelevant implementation details. As in many other successful abstractions, these principles are based on separating functionality from time and power aspects to allow for reasoning about these properties at the system level. As a concrete example we formulate a MoC for the Nostrum NoC. It is based on guaranteed bandwidth (GB) and best effort (BE) traffic. The MoC characterizes both GB and BE traffic in terms of closed formulas and allows for efficient composition of traffic
  • Keywords
    logic design; network-on-chip; MoC; Nostrum NoC; abstract computational model; best effort traffic; component design; guaranteed bandwidth; network-on-chip service; power consumption; system design; Application software; Bandwidth; Computational modeling; Computer networks; Concrete; Delay; Energy consumption; Hardware; Network-on-a-chip; Power system modeling;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Application of Concurrency to System Design, 2006. ACSD 2006. Sixth International Conference on
  • Conference_Location
    Turku
  • ISSN
    1550-4808
  • Print_ISBN
    0-7695-2556-3
  • Type

    conf

  • DOI
    10.1109/ACSD.2006.14
  • Filename
    1640234