Browse Prior Art Database

Dynamic Delete of a Failing Array Element

IP.com Disclosure Number: IPCOM000122075D
Original Publication Date: 1991-Oct-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 1 page(s) / 47K

Publishing Venue

IBM

Related People

McGilvray, BL: AUTHOR [+2]

Abstract

Typically in an array that has some form of Error Correcting Code (ECC), the reporting of single corrections is not logged, especially, if there is either a high (rapid) access to a single location that is corrected or if there are multiple locations to be corrected. There is a need to know that a correction has taken place, so as to remove that location, once it is solid, so that a two-bit failure which is uncorrectable, does not occur. The reporting of the failing location is generally through a single register. If the element reporting has to be stopped or quiesced to collect the failing address location, it inhibits the relative performance of the function being executed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 82% of the total text.

Dynamic Delete of a Failing Array Element

      Typically in an array that has some form of Error Correcting
Code (ECC), the reporting of single corrections is not logged,
especially, if there is either a high (rapid) access to a single
location that is corrected or if there are multiple locations to be
corrected.  There is a need to know that a correction has taken
place, so as to remove that location, once it is solid, so that a
two-bit failure which is uncorrectable, does not occur.  The
reporting of the failing location is generally through a single
register.  If the element reporting has to be stopped or quiesced to
collect the failing address location, it inhibits the relative
performance of the function being executed.

      In this design, the functional array has a locator array
containing address locations of the array.  The failing location gets
reported to a Service Processor.  When the Service Processor
determines the failure mechanism is solid, it returns the failing
location to the locator array. The functional array logic then
compares the next failing location to the locations within the
locator array, if a match occurs, that location does not get reported
to the Service Processor.  In this manner, the element reporting does
not have to be stopped or quiesced and there is limited reporting, to
the Service Processor, of already reported locations.  It also
provides a mechanism to identify specific element addresses within an
array, at minimum...