Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Memory Error Detection Using Error Correction Code Hardware

IP.com Disclosure Number: IPCOM000059801D
Original Publication Date: 1986-Jan-01
Included in the Prior Art Database: 2005-Mar-08
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Hoffman, H: AUTHOR

Abstract

A method of detecting uncorrectable (double-bit) errors in memory is described. The method provides a means of testing random-access memory (RAM) memory to determine if the memory has failed in such a way that uncorrectable errors will occur. So that the error detection runs quickly, this should be done without specifically testing all possible data patterns for each memory address. The hardware used to read memory contains automatic single error correcting code logic (ECC) which hides the exact failing bit. This logic cannot be disabled. However, if a multiple bit (uncorrectable) error occurs, the hardware will return the exact data as read from memory. It will also indicate if error correction is performed due to a single bit error, or if an uncorrectable error has occurred.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 1 of 1

Memory Error Detection Using Error Correction Code Hardware

A method of detecting uncorrectable (double-bit) errors in memory is described. The method provides a means of testing random-access memory (RAM) memory to determine if the memory has failed in such a way that uncorrectable errors will occur. So that the error detection runs quickly, this should be done without specifically testing all possible data patterns for each memory address. The hardware used to read memory contains automatic single error correcting code logic (ECC) which hides the exact failing bit. This logic cannot be disabled. However, if a multiple bit (uncorrectable) error occurs, the hardware will return the exact data as read from memory. It will also indicate if error correction is performed due to a single bit error, or if an uncorrectable error has occurred. The steps of the method are as follows: First, determine if any data lines are dotted together using a "walking O" test. A single pair of dotted data lines can cause errors on two different data bits in two different data patterns, but not in the same pattern; therefore, a single pair of dotted data lines is not an uncorrectable error. However, dotted data lines must be identified for two reasons. First, if two pairs of dotted data lines exist, or if a pair of dotted data lines and a stuck data bit exist, then this is an uncorrectable error. Second, dotted data lines will appear as stuck data bits in later portions of the test and must be detected so that they are not identified as an uncorrectable stuck bit error. The stuck d...