12
DESIGN OF FLOATING POINT MULTIPLIER

floating point multiplier

Embed Size (px)

DESCRIPTION

 

Citation preview

Page 1: floating point multiplier

DESIGN OF FLOATING POINT MULTIPLIER

Page 2: floating point multiplier

Representation of Floating Point NumberDecimal Number: -7.5

1 10000001 11100000000000000000000

Sign Exponent Mantissa(Significand)

Page 3: floating point multiplier

Why Floating Point???It can support a much wider range of values

as compared to fixed point.

As compared to Integer, Floating Point is more precised.

Page 4: floating point multiplier

How to Multiply?? Adding the exponent of the two numbers

then subtracting the bias from their result.

Multiplying the significand of the two numbers.

Calculating the sign by XORing the sign of the two numbers.

Page 5: floating point multiplier

Block Diagram of Floating Point Multiplier

Page 6: floating point multiplier

SIGNIFICAND MULTIPLICATION

Multiplication of Significand is done in two steps:

1. Partial Product generation2. Accumulation of generated Partial Products.

Page 7: floating point multiplier

PARTIAL PRODUCTS GENERATIONPartial Products generation can be done by various methods, but Radix-4 Booth Multiplication is one of the best technique

xi xi-1 xi-2 Operation

0 0 0 +0

0 0 1 +A

0 1 0 +A

0 1 1 +2A

1 0 0 -2A

1 0 1 -A

1 1 0 -A

1 1 1 +0

Page 8: floating point multiplier

ACCUMULATION OF PARTIAL PRODUCTSAccumulation of Partial products can be done by various adders, such as:1. Ripple Carry Adder (RCA)2. Carry Select Adder (CSLA)3. Carry Look Ahead Adder (CLA)4. Carry Save Adder (CSA)

Carry Save Adder & Ripple Carry Adder are used for Accumulation.

Page 9: floating point multiplier

Floating Point FormatsSingle precision

Double Precision

Quadruple Precision

Computer Memory

4 bytes (32 bits)

8 bytes (64 bits)

16 bytes (128 bits)

Sign bit 1 1 1

Exponent Width 8 11 15

Significand Precision

23 52 112

Page 10: floating point multiplier

Divide & Conquer Technique

Page 11: floating point multiplier

Divide & Conquer Technique in Quadruple Precision Multiplication

Page 12: floating point multiplier

THANK YOU…..