Browse Prior Art Database

Unified Check Sum Algorithm for N-Bit Processors

IP.com Disclosure Number: IPCOM000111547D
Original Publication Date: 1994-Mar-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 2 page(s) / 54K

Publishing Venue

IBM

Related People

Gross, AL: AUTHOR [+3]

Abstract

A check sum algorithm that effectively considers key processor attributes, such as word size and byte ordering of memory, is disclosed. This allows for common check sum program source across a variety of processors within a complex system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Unified Check Sum Algorithm for N-Bit Processors

      A check sum algorithm that effectively considers key processor
attributes, such as word size and byte ordering of memory, is
disclosed.  This allows for common check sum program source across a
variety of processors within a complex system.

      A new check sum algorithm has been developed that effectively
abstracts key processor attributes that typically create the
requirements for unique algorithm implementations.  The following key
attributes are considered for each processor in the set of processors
that are supported by the algorithm:

o   Inherent processor word width (i.e., 8, 16, or 32 bits, etc.)

o   Method processor uses to order bytes in memory (little vs.  bit
    endian)

      The algorithm sets the width of the check sum equal to the
widest inherent word width of all processors in the set.  All
processors in the set will compute the same check sum value
regardless of their own inherent word size.  As an example, suppose
processors with 16 and 32 bit words comprise the set supported by a
specific implementation of the algorithm.  The associated check sum
width used is 32 bits.  The algorithm requires that the data used for
the check sum computation be aligned on a multiple of the maximum
inherent word width of the processor set.

      The algorithm requires a constant non-zero value of width equal
to the smallest inherent word width of all processors in the set.
For each processor in the set, the algorithm derives a cons...