Browse Prior Art Database

Exponent Difference and Preshifter

IP.com Disclosure Number: IPCOM000092574D
Original Publication Date: 1966-Dec-01
Included in the Prior Art Database: 2005-Mar-05
Document File: 2 page(s) / 37K

Publishing Venue

IBM

Related People

Earle, J: AUTHOR [+4]

Abstract

The drawing shows hardware for an extremely fast implementation for the determination of the difference between two exponents of seven binary bits each. Such is an operation necessary in floating point addition and subtraction. The following assumes that the data is in a sixty-four bit format with one sign bit, seven exponent bits and fifty-six fraction bits.

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 52% of the total text.

Page 1 of 2

Exponent Difference and Preshifter

The drawing shows hardware for an extremely fast implementation for the determination of the difference between two exponents of seven binary bits each. Such is an operation necessary in floating point addition and subtraction. The following assumes that the data is in a sixty-four bit format with one sign bit, seven exponent bits and fifty-six fraction bits.

The fifty-six bit fraction is grouped into fourteen hexadecimal digits, each of four bits. Data shifter 10 is fourteen hexadecimal digits wide and is capable of either a 0, 1, 2, or 3 digit right-shift in a first stage or a 0, 4, 8, or 12 digit right- shift in a second stage or both. Any data shift of fourteen digit positions or greater reduces the fraction to zero. For determination of the extent of shift it is therefore only necessary to decode the right four bits of the seven bit exponent difference,
i. e., these four bits can code a shift number up to sixteen. The exponents of the two numbers to be either added or subtracted are brought in parallel along busses 1 and 2 into difference adders 3 and 4. In adder 3 the exponent EA from bus 2 is subtracted from the exponent EB of bus 1. In difference adder 4 the exponent EB of bus 1 is subtracted from the exponent EA of bus 2. Two difference adders are employed to insure that at least one difference is in true form. The true result is selected by gates 16 and 17 which select difference adder 3 or 4, depending on which difference adder h...