Browse Prior Art Database

Low Overhead Array Delete Function

IP.com Disclosure Number: IPCOM000103854D
Original Publication Date: 1993-Feb-01
Included in the Prior Art Database: 2005-Mar-18
Document File: 2 page(s) / 37K

Publishing Venue

IBM

Related People

Quinn, LL: AUTHOR [+2]

Abstract

Disclosed is an effective, low-cost method to mask off failing locations in an array.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 92% of the total text.

Low Overhead Array Delete Function

      Disclosed is an effective, low-cost method to mask off failing
locations in an array.

      A common method used to mask off failing locations in an array
is to have one bit representing each array location.  If this bit is
active, then the corresponding array location is considered masked
off and will not be used.  These bits are often implemented as
another separate array called a delete array.

      Having such a delete array, which contains a bit for every
location in the main array, is sometimes too costly as far as logic
cells and power.

      A method was needed so that if an array error did occur then
the failing location could be masked off.  This had to be done with
as few cells as possible due to cell and power restraints.

      The arrays that needed this delete function were organized into
four sets.  Each set was 32 rows deep by 16 locations wide (see
figure).  Thus, a delete array would have required 4 x 32 x 16 bits.

      To save cells and power, the following method for a delete
function was designed:

      For each set, a 5-bit delete register specifies which of the 32
rows have a deleted location, and another 16-bit delete register
specifies which locations in that row are bad.  For example,

     S_SETA_DEL_ROW(0:4)   = '00100'B
     S_SETA_DEL_LOC(015)   = '0010000001000000'B

      This coverage is not as complete as a delete array, but it
works effectively...