Optimized structure for hexadecimal and binary multiplier array

Patent metadata

Field	Value
Publication number	US-9720648-B2
Application number	US-201414578879-A
Country	US
Kind code	B2
Filing date	Dec 22, 2014
Priority date	Sep 18, 2014
Publication date	Aug 1, 2017
Grant date	Aug 1, 2017

Abstract

A method for hiding implicit bit corrections in a partial product adder array in a binary and hexadecimal floating-point multiplier such that no additional adder stages are needed for the implicit bit corrections. Two leading-one correction terms are generated for the fraction in the multiplier floating-point number and two leading-one correction terms are generated for the fraction in the multiplicand floating-point number. The floating-point numbers may be single-precision or double-precision. Each leading-one correction term for the single-precision case is appended to the left of an intermediate partial product sum in the adder array that is an input to an adder so as to not to extend the bits in the input further to the left than the bits in another input to the adder. Each leading-one correction term for the double-precision case replaces an adder input that is unused when base-2 floating-point numbers are multiplied.

First claim

What is claimed is:

1. A method for hiding implicit bit corrections in a partial product adder array in a binary and hexadecimal floating-point multiplier such that no additional adder stages are needed for the implicit bit corrections, the method comprising:
generating a first implicit bit correction term for a base-2 multiplier with an implicit bit that is one and a second implicit bit correction term for the base-2 multiplier with an implicit bit that is zero;
generating a third implicit bit correction term for a base-2 multiplicand with an implicit bit that is one and a fourth implicit bit correction term for the base-2 multiplicand with an implicit bit that is zero;
determining a value of a first implicit bit of the base-2 multiplier and a value of a second implicit bit of the base-2 multiplicand;
receiving a plurality of partial products of a product of the base-2 multiplier and the base-2 multiplicand;
adding, the plurality of partial products of the base-2 multiplier and the base-2 multiplicand in an adder array;
appending a first actual implicit bit correction term to the immediate left of a first addend to a first adder in the adder array such that a number of bits in the first actual implicit bit correction term does not exceed a number of bit positions between a left-most bit of the first addend and a left-most bit of a second addend to the first adder; and
appending a second actual implicit bit correction term to the immediate left of a third addend to a second adder in the adder array such that a number of bits in the second actual implicit bit correction term does not exceed a number of bit positions between a left-most bit of the first addend and a left-most bit of a fourth addend to the second adder.

2. The method of claim 1 , wherein the first implicit bit correction term is the first actual implicit correction term if the value of the first implicit bit of the base-2 multiplier is one.

3. The method of claim 1 , wherein the second implicit bit correction term is the first actual implicit correction term if the value of the first implicit bit of the base-2 multiplier is zero.

4. The method of claim 1 , wherein the third implicit bit correction term is the second actual implicit correction term if the value of the second implicit bit of the base-2 multiplicand is one.

5. The method of claim 1 , wherein the fourth implicit bit correction term is the second actual implicit correction term if the value of the second implicit bit of the base-2 multiplicand is zero.

6. The method of claim 1 , wherein the adder array is an array of carry-save adders.

7. The method of claim 6 , wherein the adders in the array of adders are (4:2) carry save adders.

8. The method of claim 1 , wherein the base-2 multiplier is a fraction in a single-precision floating-point number and the base-2 multiplicand is a fraction in a single-precision floating-point number.

9. The method of claim 1 , wherein the base-2 multiplier is a fraction in a double-precision floating-point number and the base-2 multiplicand is a fraction in a double-precision floating-point number.

10. The method of claim 1 , wherein the first, second, third, and fourth implicit bit correction terms are generated before the value of the first implicit bit of the base-2 multiplier and the value of the second implicit bit of the base-2 multiplicand are determined.

11. The method of claim 1 , wherein the first adder and the second adder are the same adder.

Assignees

• IBM

Inventors

• Mueller Silvia M
• Dao Trong Son

Classifications

• G06F7/4876 Primary

  Multiplying · CPC title