Scribd
Upload
47 views19 pages

Floating Point Representation

The document discusses floating point representation in computers. It explains that floating point numbers are represented in scientific notation as a sign, mantissa, and exponent. For binary floating point, numbers use a sign, mantissa stored as a binary fraction, and exponent. The IEEE 754 standard defines common floating point formats, including a 32-bit single precision format with 1 sign bit, 8 exponent bits, and 23 mantissa bits. Special values like infinity and NaN are represented with reserved exponent values.

Uploaded by

Muhammad Ahmad Raza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
47 views19 pages

Floating Point Representation

The document discusses floating point representation in computers. It explains that floating point numbers are represented in scientific notation as a sign, mantissa, and exponent. For binary floating point, numbers use a sign, mantissa stored as a binary fraction, and exponent. The IEEE 754 standard defines common floating point formats, including a 32-bit single precision format with 1 sign bit, 8 exponent bits, and 23 mantissa bits. Special values like infinity and NaN are represented with reserved exponent values.

Uploaded by

Muhammad Ahmad Raza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 19

Floating Point Representation

06/07/21 1
Floating Decimal Point : Scientific Form

256.78 is written as  2.5678 10 2

3
0.003678 is written as  3.678 10
 256.78 is written as  2.5678 10 2

2
Example

The form is
sign  mantissa 10exponent
or
  m  10e
Example: For
 2.5678  10 2
  1
m  2.5678
e2
3
Floating Point Format for Binary
Numbers

y   m2 e

  sign of number  0 for  ve, 1 for - ve


m  mantissa  1 2  m  10 2 
1 is not stored as it is always given to be 1.
e  integer exponent

4
Example
9 bit-hypothetical word
the first bit is used for the sign of the number,
the second bit for the sign of the exponent,
the next four bits for the mantissa, and
the next three bits for the exponent

 54.7510  110110 .11 2  1.1011011  2  25


 1.1011 2  101 2
We have the representation as

0 0 1 0 1 1 1 0 1
mantissa exponent
Sign of the Sign of the
number exponent

5
Machine Epsilon
Defined as the measure of accuracy and found
by difference between 1 and the next number
that can be represented

6
Example
Ten bit word
Sign of number
Sign of exponent
Next four bits for exponent
Next four bits for mantissa
0 0 0 0 0 0 0 0 0 0  1 10
Next
number 0 0 0 0 0 0 0 0 0 1  1.0001 2  1.0625 10

mach  1.0625  1  2 4

7
Relative Error and Machine Epsilon
The absolute relative true error in representing
a number will be less then the machine epsilon
Example
 0.0283210  1.1100  2  25
 1.1100  2  2  0110  2

10 bit word (sign, sign of exponent, 4 for exponent, 4 for mantissa)

0 1 0 1 1 0 1 1 0 0
Sign of the exponent mantissa
Sign of the
number
exponent

1.1100  2  2  0110  2
 0.0274375
0.02832  0.0274375
a 
0.02832
 0.034472  2  4  0.0625
8
IEEE 754 Standards for Single
Precision Representation

http://numericalmethods.eng.usf.edu
IEEE-754 Floating Point
Standard
• Standardizes representation of
floating point numbers on
different computers in single and
double precision.

• Standardizes representation of
floating point operations on
different computers.
One Great Reference
What every computer scientist (and even if
you are not) should know about floating point
arithmetic!

http://www.validlab.com/goldberg/paper.pdf
IEEE-754 Format Single Precision

32 bits for single precision


0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Sign Biased Mantissa (m)


(s) Exponent (e’)

s
.
Value  ( 1)  1 m  2  2 e ' 127

12
Example#1
1 1 0 1 0 0 0 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Sign Biased Mantissa (m)


(s) Exponent (e’)

Value    1  1. m  2  2e ' 127


s

   1  1.10100000 2  2 (10100010) 2 127


1

   1  1.625  2162127
   1  1.625  235  5.5834 1010

13
Example#2
Represent -5.5834x1010 as a single
precision floating point number.
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

Sign Biased Mantissa (m)


(s) Exponent (e’)

 5.5834  10    1  1. ?   2
10 1 ?

14
Exponent for 32 Bit IEEE-754
8 bits would represent
0  e  255
Bias is 127; so subtract 127 from
representation
 127  e  128

15
Exponent for Special Cases e
Actual range of
1  e  254
e  0 and e  255 are reserved for special numbers

Actual range of e
 126  e  127
Special Exponents and Numbers
e  0 all zeros
e  255 all ones
s e m Represents
0 all zeros all zeros 0
1 all zeros

all zeros -0
0 all ones all zeros
1 all ones all zeros 
0 or 1 all ones non-zero NaN
IEEE-754 Format

The largest number by magnitude

1.1........1 2  2
127
 3.40 10 38

The smallest number by magnitude


1.00......0 2  2 126
 2.18  10 38

Machine epsilon
 mach  2 23
 1.19  10 7

18
THE END

You might also like