• DocumentCode
    2331383
  • Title

    Simultaneous input vector selection and dual threshold voltage assignment for static leakage minimization

  • Author

    Yuan, Lin ; Qu, Gang

  • Author_Institution
    Synopsys. Inc., Mountain View
  • fYear
    2007
  • fDate
    4-8 Nov. 2007
  • Firstpage
    548
  • Lastpage
    551
  • Abstract
    Dual Vt assignment and input vector control are two tightly coupled leakage reduction techniques. We study how to apply them effectively to a circuit to minimize the static leakage power. We argue that simply combining them in a serial fashion will not reach their full potential in leakage reduction. To show this, we propose a heuristic algorithm that integrates them into a single optimization loop by assigning the value for primary inputs and Vt for logic gates simultaneously. Our algorithm leverages the fact that both input vector and threshold voltage Vt have great impact on a gate´s leakage at standby mode and avoids to assign a gate both low Vt and input vector that results high leakage. The selection of input vector and the assignment of Vt are integrated seamlessly through the concepts of leakage observability, worst leakage state, and path factor. The proposed algorithm has a low run time complexity and achieves an average 15% leakage reduction on all the ISCAS and MCNC benchmarks over the serial combination of input vector selection and dual Vt assignment.
  • Keywords
    heuristic programming; logic gates; minimisation; coupled leakage reduction techniques; dual threshold voltage assignment; heuristic algorithm; input vector; input vector selection; leakage observability; logic gates; path factor; static leakage minimization; worst leakage state; Coupling circuits; Delay; Educational institutions; Flip-flops; Heuristic algorithms; Minimization; Optimization methods; Runtime; Threshold voltage; Timing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer-Aided Design, 2007. ICCAD 2007. IEEE/ACM International Conference on
  • Conference_Location
    San Jose, CA
  • ISSN
    1092-3152
  • Print_ISBN
    978-1-4244-1381-2
  • Electronic_ISBN
    1092-3152
  • Type

    conf

  • DOI
    10.1109/ICCAD.2007.4397322
  • Filename
    4397322