• DocumentCode
    643772
  • Title

    A multi-application mapping case study for NoC-based MPSoCs

  • Author

    Guolei Zhu ; Yingmin Wang

  • Author_Institution
    Sch. of Marine Eng., Northwestern Polytech. Univ., Xi´an, China
  • fYear
    2013
  • fDate
    5-8 Aug. 2013
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    Contemporary researches on Network-on-Chip (NoC) have seen advancements in both software application mapping and hardware implementation techniques. However, two important issues remain untouched: (1) while single-application mapping algorithms are focused, few multi-application mapping methodologies exist in the literature, and (2) with HW and SW technologies developed in nearly orthogonal tracks, deficiencies in real application+platform development have been observed. This paper presents a platform-involved design flow for multi-application mapping onto tiled NoC. The methodology allows accurate mapping design space exploration by integrating various single-application heuristics. As a real implementation, a case study on mapping two multimedia applications, JPEG and AVC/H.264, onto an FPGA-based NoC (FbNoC) is presented. Different mapping outcomes are evaluated in experimental results for selecting the most appropriate mapping heuristics.
  • Keywords
    integrated circuit design; multiprocessing systems; network-on-chip; AVC/H.264; FPGA-based NoC; FbNoC; HW technology; JPEG; NoC-based MPSoCs; SW technology; hardware implementation techniques; mapping design space exploration; mapping heuristics; multiapplication mapping case study; multimedia applications; network-on-chip; platform-involved design flow; single-application heuristics; single-application mapping algorithms; software application mapping; Algorithm design and analysis; Nickel; Partitioning algorithms; Ports (Computers); Program processors; Resource management; Transform coding; FPGA; Mapping; NoC; algorithms;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing, Communication and Computing (ICSPCC), 2013 IEEE International Conference on
  • Conference_Location
    KunMing
  • Type

    conf

  • DOI
    10.1109/ICSPCC.2013.6664092
  • Filename
    6664092