Browse Prior Art Database

Multiplex Code

IP.com Disclosure Number: IPCOM000083767D
Original Publication Date: 1975-Jul-01
Included in the Prior Art Database: 2005-Mar-01
Document File: 5 page(s) / 171K

Publishing Venue

IBM

Related People

Bossen, DC: AUTHOR [+3]

Abstract

Described is an economical and effective code for a single bit and single bus error correction, multiple bit error detection. The assumption in generating the code system is that the size of a code word (152 bits) is double that of the bus width (76 bits), and a word has to be transferred through the same bus in two cycles. The encoding and decoding processes can be performed sequentially on one half the code word at a time. Structure.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 98% of the total text.

Page 1 of 5

Multiplex Code

Described is an economical and effective code for a single bit and single bus error correction, multiple bit error detection. The assumption in generating the code system is that the size of a code word (152 bits) is double that of the bus width (76 bits), and a word has to be transferred through the same bus in two cycles. The encoding and decoding processes can be performed sequentially on one half the code word at a time. Structure.

A multiplex code is a code which is constructed by combining a b-adjacent code and an odd weight column code. A (144,136) 2-adjacent code can be generated by the parity matrix H(144) shown in Fig. 1, where I(8) is an identity matrix with dimension 8.

By rearranging the columns in Fig. 1, a reshaped matrix of H(144), H'(144) is obtained and shown in Fig. 2. Thus, the column-by-column interleaving of A(68) and A'(68) is a 2-adjacent code, which can correct any double error occurring at the corresponding bits of A(68) and A'(68). Note that both A(68) and A'(68) have an explicit form of byte-parity.

Let F(ij) be the right shift operation on a matrix M with the column m(1),m(2),..., and, m(n) where 1 < or = i < j < or = n is such a way that:

(Image Omitted)

(Image Omitted)

1

Page 2 of 5

2

[This page contains 4 pictures or other non-text objects]

Page 3 of 5

3

[This page contains 3 pictures or other non-text objects]

Page 4 of 5

4

[This page contains 6 pictures or other non-text objects]

Page 5 of 5

5

[This page contains 1...