• DocumentCode
    2855120
  • Title

    Fault detection and location in reconfigurable VLSI arrays

  • Author

    Wang, K. ; Kuo, S.-Y.

  • Author_Institution
    Dept. of Electr. & Comput. Eng., Arizona Univ., Tucson, AZ, USA
  • fYear
    1989
  • fDate
    5-9 Nov. 1989
  • Firstpage
    234
  • Lastpage
    237
  • Abstract
    A systematic and efficient fault diagnosis methodology in reconfigurable VLSI array architectures is presented. This methodology utilizes the output data path independence of subsets of processing elements (PEs) based on the topology of the arrays. The ´divide the conquer´ technique is applied to reduce testing complexity and enhances the controllability and observability of arrays. An array under test is divided into several nonoverlapping parallel partitions. Those PEs in the same partition can be diagnosed simultaneously. The problem to find parallel partitions is shown equivalent to a generalized Eight Queens problem. Three types of easily testable PEs are designed to illustrate this approach. The main contribution of this paper is a novel PE fault diagnosis approach which speeds up the testing by at least O( mod V mod /sup 1/2/) for the arrays considered, where mod V mod is the number of PEs. This approach requires little or no hardware overhead depending on the types of architectures and can diagnose multiple PE faults.<>
  • Keywords
    VLSI; controllability; digital integrated circuits; fault location; logic arrays; observability; controllability; divide the conquer; fault detection; fault location; generalized Eight Queens problem; observability; output data path independence; processing elements; reconfigurable VLSI arrays; testing complexity; Electrical fault detection; Fault detection; Fault diagnosis; Hardware; Multiplexing; Switches; Systolic arrays; Testing; Very large scale integration; Voting;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer-Aided Design, 1989. ICCAD-89. Digest of Technical Papers., 1989 IEEE International Conference on
  • Conference_Location
    Santa Clara, CA, USA
  • Print_ISBN
    0-8186-1986-4
  • Type

    conf

  • DOI
    10.1109/ICCAD.1989.76943
  • Filename
    76943