• DocumentCode
    3028007
  • Title

    A Dynamic Fuzzy Logic Based Routing Scheme for Bufferless NoCs

  • Author

    Chrysostomou, Charalambos ; Tatas, K. ; Runcan, A.R.

  • Author_Institution
    Dept. of Comput. Sci. & Eng., Frederick Univ., Nicosia, Cyprus
  • fYear
    2012
  • fDate
    5-7 Dec. 2012
  • Firstpage
    295
  • Lastpage
    302
  • Abstract
    This paper proposes a novel, adaptive routing scheme for Network-on-Chip (NoC). In particular, the proposed routing scheme decides on the output port of an incoming flit by taking into account the dynamic traffic and power consumption on neighboring router links. For that, a simple, generic, and efficient nonlinear control law is built, based on fuzzy logic control, to dynamically calculate the input links cost. Although we focus on buffer less NoCs, it can also be applied to buffered NoCs as well. The additional digital logic required for the proposed scheme is not in the router critical path, and, therefore, imposes no additional latency. We demonstrate, via simulative evaluation of light to heavy congestion conditions, that the proposed scheme outperforms representative conventional counterparts in terms of throughput and latency. Thus, the effectiveness of the proposed scheme is shown, as the performance is consistently better.
  • Keywords
    fuzzy control; network routing; network-on-chip; nonlinear control systems; adaptive routing scheme; buffered NoC; bufferless NoC; digital logic; dynamic fuzzy logic; dynamic traffic; flit; fuzzy logic control; input link cost; light-to-heavy congestion conditions; network-on-chip; nonlinear control law; on-chip communications; power consumption; router links; Fuzzy logic; Measurement; Ports (Computers); Power demand; Pragmatics; Routing; Switches; Fuzzy Logic; NoC; Routing Algorithms;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Science and Engineering (CSE), 2012 IEEE 15th International Conference on
  • Conference_Location
    Nicosia
  • Print_ISBN
    978-1-4673-5165-2
  • Electronic_ISBN
    978-0-7695-4914-9
  • Type

    conf

  • DOI
    10.1109/ICCSE.2012.48
  • Filename
    6417307