Browse Prior Art Database

Floating Point Feature On The IBM Type 1620

IP.com Disclosure Number: IPCOM000097092D
Original Publication Date: 1962-May-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 4 page(s) / 71K

Publishing Venue

IBM

Related People

Jones, FB: AUTHOR [+2]

Abstract

In the type 1620 automatic floating point operations, a floating point number is a field consisting of a variable length mantissa and a two digit exponent. The exponent is in the two low order positions of the field, and the mantissa is in the remaining high order positions, bar M.....Mbar EE.

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 39% of the total text.

Page 1 of 4

Floating Point Feature On The IBM Type 1620

In the type 1620 automatic floating point operations, a floating point number is a field consisting of a variable length mantissa and a two digit exponent. The exponent is in the two low order positions of the field, and the mantissa is in the remaining high order positions, bar M.....Mbar EE.

The most significant digit positions are marked by flags and the algebraic signs are marked by flags over the least significant digit positions. The exponent is established on the premise that the mantissa is less than 1.0 and equal to or greater than 0.1, and has a range of -99 to +99. The smallest positive quantity that can be represented is thus 00.... 099. The mantissa may have from two to one hundred digits. However, when two fields are operands involved in the same instruction, i.e., quantities being added together, they must have mantissas of the same length and the result will have this same length.

Eight floating-point instructions are provided. All are in the type 1620 format. The High/Positive indicator is turned on when the mantissa result from a computation is greater than zero, The Equal/Zero indicator is turned on to indicate a zero mantissa result from a computation. Four instructions are for arithmetic computations: add, subtract, multiply, and divide; three control field size and location: shift right, shift left, and transmit; one provides for branch and transmit. The following abbreviations are employed: M(p) is the mantissa of the field at the P address (P); M(q) is the mantissa of the field at the Q address (Q); E(p) is the exponent of the field at the P address; Eq is the exponent of the field at the Q address; L is the number of digits in the mantissa; d is the difference between exponents.

When adding, the appropriate mantissa is effectively shifted, if necessary, to align decimal points. M(q) is added to M(p) and the result replaces M(p). The number of low-order digits of the shifted mantissa equal to d is truncated. If either mantissa is above the range of the other, no addition is performed. After addition is completed, the number of M(p) digits is checked to determine if it exceeds L. If so, this is an overflow condition, the least significant digit of M(p) is truncated and the mantissa is shifted one position to the right. A one is entered in the most significant digit position of the mantissa, and a one is added to E(p). When an overflow does not exist, M(p) is scanned for zeros, beginning with the most significant digit. High order digit positions having zeros are counted and M(p) is shifted positions to the left equal to the count, Vacated digit positions are set to zero. Flag bits in M(p) are not altered or moved. The number of high order zeros is subtracted from E(p).

When subtracting, the operation is the same as the add except that sign control procedures for M(q) are reversed.

In multiplying, M(p) is multiplied by M(q), and the L high order significant digits o...