Browse Prior Art Database

Partitionable Error Correction Code

IP.com Disclosure Number: IPCOM000063099D
Original Publication Date: 1985-Feb-01
Included in the Prior Art Database: 2005-Feb-18

Publishing Venue

IBM

Related People

Authors:
Villante, FJ [+details]

Abstract

Many double-error detection, single-error correction techniques use a modified Hamming Code. When applied to large width byte (or multi-byte) organized paths, this results in check bit generation requirements which are not consistent for each byte (or multi-byte). The error syndrome is also complex and does not directly decode into which byte (or multi-byte) and which bit within that byte (or multi-byte) is in error. These factors make it difficult to partition the required error correction code (ECC) logic so there is a repeatable logic set associated with each byte (or multi-byte) interface to the device which is error correction coded. Wide data paths are normally partitioned into smaller bit groupings for functional and/or physical and/or notational purposes.