• DocumentCode
    2447514
  • Title

    Comparison and Analysis of Novel Architectures for Media Compression

  • Author

    Yan, Ming ; Li, Sikun

  • Author_Institution
    Nat. Univ. of Defense Technol., Changsha
  • fYear
    2007
  • fDate
    15-18 Oct. 2007
  • Firstpage
    509
  • Lastpage
    512
  • Abstract
    Image and video compression is a typical data-intensive and computation-intensive application. As general processors are inefficient and power consumptive in processing such tasks, more and more novel architectures appeared. This paper first gave an overview about the media compression standards, and then introduced several novel architecture processors, such as VIRAM, DTVA, Imagine, Cell, RAW and TRIPS. Some of these processors used PIM technology, and some are based on stream processing, the others are multicore and polymorph architecture. Finally compared and analyzed the design parameters of these processors, such as the performance and power consumptions. According to the media compression features which have constraints on architectural design, we concluded some useful suggestions for media compression processor architecture designers, that include execution model, memory hierarchies, performance and dissipation, etc.
  • Keywords
    data compression; image coding; video coding; Cell; DTVA; Imagine; PIM technology; RAW; TRIPS; VIRAM; computation-intensive application; data-intensive application; execution model; image compression; media compression; memory hierarchies; multicore architecture; polymorph architecture; stream processing; video compression; Computer applications; Computer architecture; Digital TV; Energy consumption; Image coding; Multicore processing; Performance analysis; Process design; Streaming media; Video compression;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer-Aided Design and Computer Graphics, 2007 10th IEEE International Conference on
  • Conference_Location
    Beijing
  • Print_ISBN
    978-1-4244-1579-3
  • Electronic_ISBN
    978-1-4244-1579-3
  • Type

    conf

  • DOI
    10.1109/CADCG.2007.4407942
  • Filename
    4407942