Browse Prior Art Database

Hybrid Carry Save Adder Modules

IP.com Disclosure Number: IPCOM000092448D
Original Publication Date: 1966-Nov-01
Included in the Prior Art Database: 2005-Mar-05
Document File: 3 page(s) / 49K

Publishing Venue

IBM

Related People

Smith, JL: AUTHOR [+2]

Abstract

In high-speed scientific computers carry-save adder modules other than the usual 3-input, 2-output adder are advantageous for use in fast multiplier networks and other assemblies of batch adders. Some of these have the potentiality of providing higher rates of adding, variously termed encoding, counting, or eliminating, per logic level than the 3-to-2 adder module.

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 53% of the total text.

Page 1 of 3

Hybrid Carry Save Adder Modules

In high-speed scientific computers carry-save adder modules other than the usual 3-input, 2-output adder are advantageous for use in fast multiplier networks and other assemblies of batch adders. Some of these have the potentiality of providing higher rates of adding, variously termed encoding, counting, or eliminating, per logic level than the 3-to-2 adder module.

A characteristic of these adder modules is the existence of one or more carry outputs and corresponding carry inputs. These are functions only of primary inputs and carry inputs that originate on earlier logic levels as carry outputs from other modules of the same or different type. As a result, true carry assimilation takes place over a fixed number of digit orders. For this reason the modules are termed as being hybrid, having both carry-save features and carry-assimilate features. Some examples follow, which illustrate some of the design principles of hybrid carry-save adder modules.

A functional block is shown for a 4-to-2 adder module, including the necessary inputs and outputs. The logic for the module is described in terms of Boolean equations. As implied by the equations, the final outputs S1 and S2 are obtained in two steps. During the first step the intermediate functions E1, F1, G4, and X2 are formed. The X2 is a true carry and is transmitted to a higher order module to the X1 carry input. The X1 of this instance comes from the next lower order module. During the second step the functions S1 and S2 are formed. These are respectively the sum and saved carry. X2 must not be a function of X1 so that the carry can ripple only a fixed distance, in this example one digit order.

The primary inputs and carry input have unit weight 1, indicated by the subscripts in the Boolean equations. The primary outputs have weights 1 and 2. The weight of the carry-out must be such that the maximum possible excess of all inputs not represented by a carry-out is representable by the primary outputs. This...