Scribd
Upload
148 views2 pages

Booth Radix 4

The document summarizes Booth's algorithm for multiplying binary numbers more efficiently in hardware. It observes that strings of 1's in a binary number can be replaced by a carry out with subtraction. This allows expressing the multiplier in radix-4 to reduce the number of partial products from two to one. Booth's algorithm further recodes the radix-4 digits to 0, 1, -1, -2 to avoid multiplying by 3, with negative values handled by subtraction instead of addition. An example 4-bit multiplication shows how this halves the number of partial products needed.

Uploaded by

ishaanksheth2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
148 views2 pages

Booth Radix 4

The document summarizes Booth's algorithm for multiplying binary numbers more efficiently in hardware. It observes that strings of 1's in a binary number can be replaced by a carry out with subtraction. This allows expressing the multiplier in radix-4 to reduce the number of partial products from two to one. Booth's algorithm further recodes the radix-4 digits to 0, 1, -1, -2 to avoid multiplying by 3, with negative values handled by subtraction instead of addition. An example 4-bit multiplication shows how this halves the number of partial products needed.

Uploaded by

ishaanksheth2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Booth’s Algorithms for Multiplication

There is a family of algorithms for speeding multiplication in hardware that are all based on
Booth’s original work. He observed that a string of 1’s in a binary number could be replaced by
a carry out with a subtraction at the least significant bit of the string. Put in mathematical
terms,

2 2 1

As a concrete example, 01112 = 10002 ‐ 00012. Obviously strings of 1’s in the middle of a
number can be represented in the same way by left shifting this result appropriately.

This can be exploited in a variety of ways to reduce the number of partial products one has to
form and accumulate to calculate a product. In class I talked about a particularly common
implementation of this in radix‐4 multiplication. The basic idea is to express the multiplier in
radix‐4, that is, as a set of digits 0 – 3 instead of just 0,1. This cuts the number of partial
products in half because you are multiplying by two binary bits at once. However, it requires
multiplying by 3 which is difficult. (Multiplication by 0, 1, or 2 is trivial because they only
involve simple shifts. Three is the hard one.) To avoid multiplying by 3, we use Booth’s
observation and recode the digit set to be 2, 1, 0, ‐1, and ‐2. The partial products with the
positive digits are trivial to form while the negative values can be done by subtracting instead of
adding their partial products.

Here is an example: +610 * +610 = +36 where the numbers are 4‐bit unsigned binary. As needed
the negative partial products are extended to a 9‐bit 2’s complement number. (8‐bits is
adequate to express any 4 x 4 product unsigned product and the 9th bit allows for worst case
sign extension. The 9th bit is lost in the product because the product is never negative.
Generally only minor adjustment is needed in this algorithm for 2’s complement operands.)
Only two partial products are needed and the second is left shifted by two bits because of the
radix‐4 coding.

0110 ‐ multiplicand = +6
0110 – multiplier as base 2 = +6
+2 ‐2 – multiplier recoded with Booth’s algorithm per table below = +2*4 + (‐2) = +6
111110100 – first partial product = ‐2*(+6) = ‐12
0001100 ‐ second partial product = +2*(+6) left shift by 2 = +48
00100100 – final product = +4 + +32 = +36
In this table, the bits i and i+1 are the bits being turned into a radix‐4 digit and the i‐1 th bit
contributes to the recoded value for the reasons given in the comment column.
Xi+1 Xi Xi‐1 Zi Comment
0 0 0 0 0
0 0 1 1 End of string of 1's ‐ carry in
0 1 0 1 Isolated 1
End of string of 1's ‐ carry into
0 1 1 2 second bit
1 0 0 ‐2 Beginning of 1's in 2nd bit
1 0 1 ‐1 ‐2 + 1
1 1 0 ‐1 Beginning of 1's in 1st bit
1 1 1 0 String goes through ‐ all zero

You might also like