• DocumentCode
    324018
  • Title

    Instruction level power model and its application to general purpose processors

  • Author

    Sridhar, Ramalingam ; Schindler, Kaspar

  • Author_Institution
    State Univ. of New York, Buffalo, NY, USA
  • Volume
    1
  • fYear
    1997
  • fDate
    2-5 Nov. 1997
  • Firstpage
    753
  • Abstract
    Increased use of portable applications has placed severe limitations on the power consumed by processors and systems. This is particularly true for digital signal processors. Many researchers are now considering power minimization through modification of the high level software and the algorithms. This however will be more effective if a realistic power model for the instructions and the various types of memory accesses were developed. The power consumed in the fetch and the execution of any instruction is dependent on a number of factors, including the state of various functional units associated with the operation and their input patterns. This requires a careful model for the power consumption for any instruction. This paper presents a metric for such a model and its use in determining the power consumption for a given program. This is then used to rewrite the code to achieve considerable power reduction.
  • Keywords
    digital signal processing chips; power consumption; software engineering; algorithms; code; digital signal processors; functional units; general purpose processors; high level software; input patterns; instruction level power model; memory access; mobile computing; portable applications; power consumption; power minimization; program; Application software; Costs; Digital signal processing; Digital signal processors; Energy consumption; Minimization methods; Power system modeling; Signal processing algorithms; Software algorithms; Testing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signals, Systems & Computers, 1997. Conference Record of the Thirty-First Asilomar Conference on
  • Conference_Location
    Pacific Grove, CA, USA
  • ISSN
    1058-6393
  • Print_ISBN
    0-8186-8316-3
  • Type

    conf

  • DOI
    10.1109/ACSSC.1997.680545
  • Filename
    680545