02: Data Representation and Arithmetic

IT1206- Computer Systems

Level I - Semester 1

© 2020 e-Learning Centre, UCSC

 List of Sub Topics

• Positinal Numbering System

• Decimal-Binary Conversion
• Unsigned whole number conversion
• Fraction conversion
• Power of two Radices conversion
• Signed Integer Representation
• Signed Magnitude
• Two’s complement
• Floating Point Representation
• Character Codes

© 2020 e-Learning Centre, UCSC 2

 What is a bit ?

• A bit is the smallest unit of data on a machine.

• A bit can hold only one of two values: 0 or 1
• In a computer, a bit represents one of possible two
states.
• Because bits are so small, you rarely work with
information one bit at a time.

© 2020 e-Learning Centre, UCSC 3

 What is a byte ?

• Conventionally, a group of 8 bits are named as a byte.

Counting starts with zero

7 6 5 4 3 2 1 0 Position

1 1 0 1 1 0 1 0
A byte represents a group of bits

Most Least
Significant Bit Significant Bit

© 2020 e-Learning Centre, UCSC 4

 Word

• A word is a fixed-sized piece of data which is handled

as a unit by the CPU of a computer.
• E.g. 32 bit, 64 bit, and so on.
• A word may consist with several bytes.

© 2020 e-Learning Centre, UCSC 5

 Units of Data Measurement

• 8 Bits = 1 Byte
• 1024 Bytes = 1 Kilobyte (KB)
• 1024 KB = 1 Megabyte (MB)
• 1024 MB = 1 Gigabyte (GB)
• 1024 GB = 1 Terabytes (TB)
• 1024 TB = 1 Petabytes (PB)
• 1024 PB = 1 Exabyte
• 1024 EB = 1 Zettabyte
• 1024 ZB = 1 Yottabyte

© 2020 e-Learning Centre, UCSC 6

 Positional Numbering System

• A positional numbering system is the representation of

numbers by an ordered set of numerals symbols
(digits) in which the value of a numeral symbol
depends on its position.
• We all are familiar with base (radix) 10 number system.
• In based 10 number system, number 283.75 means;
2 × 102 + 8 × 101 + 3 × 100 + 7 × 10−1 + 5 × 10−2

2 8 3 7 5
102 101 100 10-1 10-2

© 2020 e-Learning Centre, UCSC 7

 Number Systems

• There are many other number systems

• Decimal [0,1,2,3,4,5,6,7,8,9]
• Binary [0,1]
• Hexa Decimal [0,1,2,3,4,5,6,7,8,9,A,B,C,D,E,F]
• Octal [0,1,2,3,4,5,6,7]

© 2020 e-Learning Centre, UCSC 8

 Converting Decimal to Binary (whole numbers)

• To convert decimal whole number to binary, start with

the integer part in the given number and divide it by 2
• Keep the resulting quotient and the remainder
separately.
• Continue dividing the quotient by 2 until you get a
quotient of zero.
• At the end just write out the remainders in the reverse
order.

© 2020 e-Learning Centre, UCSC 9

Converting Decimal to Binary (whole numbers)

123 11110112
÷2
61 → remainder 1
÷2
30 → remainder 1
÷2
15 → remainder 0
÷2
7 → remainder 1
÷2
3 → remainder 1
÷2
1 → remainder 1
÷2
0 → remainder 1

© 2020 e-Learning Centre, UCSC 10

 Converting Decimal to Binary (fractional numbers)

• To convert decimal fraction to binary, start with the

fractional part in the given number and multiply it by 2.
• Keep the resulting integer and fractional parts separately.
• Continue multiplying by 2 until you get a resulting
fractional part equal to zero.
• At the end just write out the integer parts from the results
of each multiplication starting from the first multiplication.
0.375 x 2 = 0 + 0.75
0.75 x 2 = 1 + 0.5
0.5 x 2 = 1 + 0

0.37510 = 0.0112

© 2020 e-Learning Centre, UCSC 11

 Activities

Convert following numbers to binary

• 2273
• 12478
• 0.2487
• 0.99724
• 347.35
• 437.639

© 2020 e-Learning Centre, UCSC 12

 Converting Binary to Decimal

• If we want to convert 101.112 to decimal;

1 0 1 1 1
22 21 20 2-1 2-2

1 × 22 + 0 × 21 + 1 × 20 + 1 × 2−1 + 1 × 2−2

4 + 0 + 1 + 0.5 + 0.25

5.75

© 2020 e-Learning Centre, UCSC 13

 Converting Between Power-of-Two Radices

• Hexadecimal and Octal are two number systems where

their radices are of power of two (16 = 24 and 8 = 23).
• We can easily convert numbers between power of two
radices number systems.

© 2020 e-Learning Centre, UCSC 14

 Hexadecimal

HEX Bit Pattern HEX Bit Pattern

0 0000 8 1000
1 0001 9 1001
2 0010 A 1010
3 0011 B 1011
4 0100 C 1100
5 0101 D 1101
6 0110 E 1110
7 0111 F 1111

© 2020 e-Learning Centre, UCSC 15

 Octal

Octal Bit Pattern

0 000
1 001
2 010
3 011
4 100
5 101
6 110
7 111

© 2020 e-Learning Centre, UCSC 16

 Converting Between Power-of-Two Radices

• Ex. 1100100111012
• Conversion to hexadecimal
1 1 0 0 1 0 0 1 1 1 0 1 Binary

C 9 D Hexadecimal

• Conversion to octal
1 1 0 0 1 0 0 1 1 1 0 1 Binary

6 2 3 5 Octal

© 2020 e-Learning Centre, UCSC 17

 Binary Addition

• 0+0=0
• 0+1=1
• 1+0=1
• 1 + 1 = 10 (carry: 1)

• E.g.
11 1 1 1 (carry)
0 1 1 0 1
+ 1 0 1 1 1
= 10 0 1 0 0

© 2020 e-Learning Centre, UCSC 18

 Binary Subtraction

• 0-0=0
• 0 - 1 = 1 (with borrow)
• 1-0=1
• 1-1=0

• E.g.
* * * (borrow)
1 0 1 1 0 1
- 0 1 0 1 1 1
= 0 1 0 1 1 0

© 2020 e-Learning Centre, UCSC 19

 Binary Multiplication

• E.g.
1 0 1 1
x 1 0 1 0
0 0 0 0
+ 1 0 1 1
+ 0 0 0 0
+ 1 0 1 1
= 1 1 0 1 1 1 0

© 2020 e-Learning Centre, UCSC 20

 Binary Division

• E.g.
1 0 1 Quotient

1 0 1 1 1 0 1 1
- 1 0 1
1 1
- 0 0
1 1 1
- 1 0 1
1 0 Remainder

© 2020 e-Learning Centre, UCSC 21

 Representing Numbers

• Problems of number representation

• Positive and negative
• Radix point
• Range of representation

• Different ways to represent numbers

• Unsigned representation: non-negative integers
• Signed representation: integers
• Floating-point representation: fractions

© 2020 e-Learning Centre, UCSC 22

 Unsigned Binary Numbers

• Unsigned binary numbers

• Have 0 and 1 to represent numbers
• Only positive numbers stored in binary
• The Smallest binary number that can be stored
using 8 bits would be 0 0 0 0 0 0 0 0
which equals to 0
• The largest binary number would be 1 1 1 1 1 1 1 1

which equals to 128 + 64 + 32 + 16 + 8 + 4 + 2 + 1

= 255 = 28-1

• Therefore the range is 0 - 255 (256 numbers)

© 2020 e-Learning Centre, UCSC 23

 Signed Magnitude Representation

• Signed Magnitude (Signed binary) numbers

• Have 0 and 1 to represent numbers

• By convention, the leftmost bit is used as the sign bit

• 0 for positive
• 1 for negative

Sign bit

© 2020 e-Learning Centre, UCSC 24

 Signed Magnitude (Signed Binary) Numbers

• Signed Magnitude (Signed binary) numbers

• The Smallest positive binary number is
0 0 0 0 0 0 0 0 =0
• The largest positive binary number is
0 1 1 1 1 1 1 1 = 127
• Therefore the range for positive numbers is 0 - 127
(the range is 128 numbers)

© 2020 e-Learning Centre, UCSC 25

Challenges with Negative Numbers in Signed Binary

Problems with simple signed representation

• Two representation of zero: + 0 and – 0
0 0 0 0 0 0 0 0 and 1 0 0 0 0 0 0 0

• Need to consider both sign and magnitude in

arithmetic
• E.g. 5 – 3
= 5 + (-3)
= 00000101+10000011
= 10001000
= -8

© 2020 e-Learning Centre, UCSC 26

 Problems with Simple Signed Representation

• Need to consider both sign and magnitude in

arithmetic
E.g. = 18 + (-18)
= 00010010 +10010010
= 10100100
= -36
• We have to change the way these arithmetic
operations are performed based on the signs of two
numbers.

© 2020 e-Learning Centre, UCSC 27

 Two’s Complement Number System

• Two's complement is the most common method of

representing signed integers on computers.
• In this scheme most significant bit indicates the sign.
• In order to represent the negative number, the positive
value of the number will be two’s complemented.
Two’s Complement
Bits Unsigned Value
Value
0
000 0 0 -1 1
001 1 1
010 2 2 Two’s
-2 2
011 3 3 Complement
100 4 -4
101 5 -3 -3 3
110 6 -2 -4
111 7 -1
© 2020 e-Learning Centre, UCSC 28
 Two’s Complement

• Start from the signed binary representation of its

positive value
• Copy the bit pattern from right to left until a 1 has
been copied
• Complement the remaining bits: all the 1’s with 0’s,
and all the 0’s with 1’s
• Ex. -10510 in Two’s complement (10510 = 0110 10012)
0110 1001 ⟹ 1001 0111
−10510 = 1001 0111
• An exception: 1 0 0 0 0 0 0 0 = -12810

© 2020 e-Learning Centre, UCSC 29

 Example: Conversion

• Ex. Convert decimal -104 to two’s complement

• First signed magnitude representation of 104;
10410 = 0110 10002
• Find the first digit1 from right to left
0110 1000
• Copy everything till first digit 1 from right to left
_ _ _ _ 1000
• Then complement the remaining bits
10011000
• Therefore, two’s complement of -10410 is 10011000

© 2020 e-Learning Centre, UCSC 30

 Your Turn…!

• What is the SMALLEST and LARGEST signed binary

numbers that can be stored in 1 BYTE ?

© 2020 e-Learning Centre, UCSC 31

 Positive

Negative

Source: https://en.wikipedia.org/wiki/Two%27s_complement
© 2020 e-Learning Centre, UCSC 32
 Two’s Complement - Benefits

• One representation of zero

• Arithmetic works easily

• Negating is fairly easy

© 2020 e-Learning Centre, UCSC 33

 Ranges of Integer Representation

• 8-bit unsigned binary representation

• Largest number: 1 1 1 1 1 1 1 12 = 25510
• Smallest number: 0 0 0 0 0 0 0 02 = 010

• 8-bit two’s complement representation

• Largest number: 0 1 1 1 1 1 1 12 = 12710
• Smallest number: 1 0 0 0 0 0 0 02 = -12810

• The problem of overflow

• 13010 = 1 0 0 0 0 0 1 02
• 0 0 0 0 0 1 02 in two’s complement

© 2020 e-Learning Centre, UCSC 34

Geometric Depiction of Two’s Complement Integers

© 2020 e-Learning Centre, UCSC 35

 Integer Data Types in C++

Type Size in Bits Range

unsigned int 16 0 – 65535

int 16 -32768 - 32767

unsigned long int 32 0 to 4,294,967,295

long int 32 -2,147,483,648 to 2,147,483,647

© 2020 e-Learning Centre, UCSC 36

 Activities

• Convert following two’s complement numbers to

Decimal representation.
• 0010 1101
• 1101 1010
• 0101 0101
• 1010 1100

© 2020 e-Learning Centre, UCSC 37

 Addition with Two’s Complement

• Add 910 to -2310

9 0000 1001
-23 1110 1001
? 1111 0010

• Check whether 1111 0010 represents expected -1410 in

two’s complement representation.

© 2020 e-Learning Centre, UCSC 38

 Addition with Two’s Complement

• Add -910 to 2310

Carries 1 111 111
-9 1111 0111
23 0001 0111
? 0000 1110

• The carry going out (encircled) at the sign bit can be

discarded as we are working only with 8 bits.
• Check whether 1111 0010 still represents the expected
1410 in two’s complement representation.

© 2020 e-Learning Centre, UCSC 39

 Addition with Two’s Complement

• Add -910 to 2310

Carries 1 111 111
-9 1111 0111
23 0001 0111
? 0000 1110
• We do not consider this as an overflow as it gives the correct
answer.
• An overflow occurs if two positive numbers are added and the
result is negative, or if two negative numbers are added and the
result is positive.
• It is not possible to have overflow when using two's complement
notation if a positive and a negative number are being added
together. 40
© 2020 e-Learning Centre, UCSC
 Addition with Two’s Complement

• Add -910 to 2310

Carries 1 111 111
-9 1111 0111
23 0001 0111
? 0000 1110

• Carry going into the sign bit (highlighted in green) is the same as
the carry going out of the sign bit (highlighted in yellow).
• In such a scenario, no overflow occurs.
• When these carries are different, an overflow indicator is set in
the arithmetic logic unit, indicating the result is incorrect.

© 2020 e-Learning Centre, UCSC 41

 Overflow in Addition

• Add 12610 to 2310

Carries 0 1111 11
126 0111 1110
23 0001 0111
? 1001 0101

• Carry going into the sign bit (highlighted in green) is 1 and the
carry going out of the sign bit (highlighted in yellow) is 0.
• Addition of two positive numbers has resulted a negative number.
• Carries are different in here and an overflow has occurred.

© 2020 e-Learning Centre, UCSC 42

 Carry Vs. Overflow

• CPUs often have flags to indicate both carry and

overflow.
• The overflow flag is used only with signed numbers.
• For unsigned numbers carry flag is sufficient to detect
incorrect outcome after the arithmetic operation.
• However, carry flag is insufficient to detect incorrect
outcome in signed number arithmetic operations.
• When carry going into the sign bit and the carry
going out of the sign bit are different in signed
number arithmetic, then that is an overflow.

© 2020 e-Learning Centre, UCSC 43

 Carry Vs. Overflow

Correct
Operation Result Carry ? Overflow ?
Result ?
0100 + 0010 0110 No No Yes
([+4] + [-2] in decimal) (+6 in decimal)

0100 + 0110 1010 No Yes No

([+4] + [+6] in decimal) (-6 in decimal)

1100 + 1110 1010 Yes No Yes

([-4] + [-2] in decimal) (-6 in decimal)

1100 + 1010 0110 Yes Yes No

([-4] + [-6] in decimal) (+6 in decimal)

Carry – indicates there is a carry out from the sign bit

© 2020 e-Learning Centre, UCSC 44

 Fractional Numbers in Decimal

• 16.357 = the SUM of ...

7 x 10-3 = 7/1000
5 x 10-2 = 5/100
3 x 10-1 = 3/10
6 x 100 = 6
1 x 101 = 10

357
• 7/1000 + 5/100 + 3/10 + 6 + 10 =16
1000

© 2020 e-Learning Centre, UCSC 45

 Fractions in Binary

• 10.011 = the SUM of ...

1 * 2-3 = 1/8
1 * 2-2 = 1/4
0 * 2-1 = 0
0 * 20 = 0
1 * 21 = 2

3
• 1/8 + 1/4 + 2 =2
8
3
• (10.011)2 = 2 in Decimal (Base 10)
8
© 2020 e-Learning Centre, UCSC 46
 Activities

• What is Binary 011.0101 in Base 10?

• What is Binary 101.1101 in Base 10?

• What is Binary 111.111 in Base 10?

5
• What is Decimal 3 in Binary ?
16

© 2020 e-Learning Centre, UCSC 47

 Scientific Notation in Decimal

• Consider the following representation in decimal

number …
135.26 = .13526 x 103
13526000 = .13526 x 108
0.0000002452 = .2452 x 10-6

• .13526 x 103 has the following components:

• a Mantissa = .13526
• an Exponent = 3
• a Base = 10

© 2020 e-Learning Centre, UCSC 48

 Scientific Normalized Notation for Binary Numbers

• Scientific notation for binary. Examples …

• 11011.101 = 1.1011101 x 24
• -10110110000 = -1.011011 x 210
• 0.00000010110111 = 1.0110111 x 2-7

© 2020 e-Learning Centre, UCSC 49

 Floating Point Format in 1 Byte

RADIX POINT

MANTISSA
EXPONENT

SIGN

SIGN = 0 (+ve) | 1 (-ve)

EXPONENT in EXCESS FOUR Notation

© 2020 e-Learning Centre, UCSC 50

 Floating Point Example

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

1. STORE the SIGN BIT

© 2020 e-Learning Centre, UCSC 51

 Floating Point Example

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

2. STORE the MANTISSA BITS

0 0 0 1 0

© 2020 e-Learning Centre, UCSC 52

 Floating Point Example

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

3. STORE the EXPONENT BITS

0 X X X 0 0 1 0

© 2020 e-Learning Centre, UCSC 53

 Floating Point Example

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

3. STORE the EXPONENT BITS

0 X X X 0 0 1 0

Exponent is represented in Biased representation

© 2020 e-Learning Centre, UCSC 54
 Biased (Excess-k) Representation

• EXCESS THREE NOTATION

An excess notation system using bit pattern of length three

Bit Pattern Value Representation

111 4
110 3
101 2
100 1
011 0
010 -1
001 -2
000 -3

© 2020 e-Learning Centre, UCSC 55

 Biased (Excess-k) Representation

• Usually, for N bit numbers, k is (2N-1-1)

• E.g. for 4-bit integers, k is 7
for 3-bit integers, k is 3
for 8-bit exponent, k is 127
• The actual value of each bit string is its unsigned value
minus k
• To represent a number in excess-k, add k

© 2020 e-Learning Centre, UCSC 56

 Biased (Excess-k) Representation with 4 Bits

Unsigned Excess-k
0000.................... 0 -7
0001.................... 1 -6
0010.................... 2 -5
0011.................... 3 k=7 -4
0100.................... 4 -3
0101.................... 5 -2
0110.................... 6 -1
sliding
0111.................... 7 0
ruler 1000.................... 8 1
1001.................... 9 2
1010.................... 10 3
1011.................... 11 4
1100.................... 12 5
1101.................... 13 6
1110.................... 14 7
1111.................... 15 8
© 2020 e-Learning Centre, UCSC 57
 Floating Point Example - Exponent

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

3. STORE the EXPONENT BITS

0 X X X 0 0 1 0

Let’s bring Excess-3 representation for the Exponent

© 2020 e-Learning Centre, UCSC 58
 Floating Point Example - Exponent

• To STORE the number ...

1
+1 = 1.001 × 20
8
in FLOATING POINT NOTATION ...

3. STORE the EXPONENT BITS

0 0 1 1 0 0 1 0

Unsigned 0 = 011 in Excess-3

© 2020 e-Learning Centre, UCSC 59
 Floating Point Example 2

• To STORE the number ...

1
−3 = −11.01
4
in FLOATING POINT NOTATION ...

- First write the number in standard form

© 2020 e-Learning Centre, UCSC 60

 Floating Point Example 2

• To STORE the number ...

1
−3 = −11.01 = −𝟏. 𝟏𝟎𝟏 × 𝟐𝟏
4
in FLOATING POINT NOTATION ...

1. STORE the SIGN BIT

Next, Mantissa…
© 2020 e-Learning Centre, UCSC 61
 Floating Point Example 2

• To STORE the number ...

1
−3 = −11.01 = −𝟏. 𝟏𝟎𝟏 × 𝟐𝟏
4
in FLOATING POINT NOTATION ...

2. STORE the MANTISSA BITS

1 1 0 1 0

Next, Exponent…
© 2020 e-Learning Centre, UCSC 62
 Floating Point Example 2

• To STORE the number ...

1
−3 = −11.01 = −𝟏. 𝟏𝟎𝟏 × 𝟐𝟏
4
in FLOATING POINT NOTATION ...

2. STORE the EXPONENT BITS

1 1 0 0 1 0 1 0

© 2020 e-Learning Centre, UCSC 63

 Activities

• Try to represent following decimal floating-point

numbers in binary representation given below.
• 14.5
• 11/64
• -14.5
5
• −1
8

Sign Bit Exponent Mantissa

© 2020 e-Learning Centre, UCSC 64

 Floating Point Binary to Decimal Conversion

1. Convert EXPONENT (EXCESS 4)

2. Apply EXPONENT to MANTISSA
3. Convert BINARY Fraction
4. Apply SIGN

© 2020 e-Learning Centre, UCSC 65

 Floating Point Binary to Decimal Conversion

• Example – 10111010 [Assume Exponent = 3 bits,

Mantissa = 4 bits and Exponent in Excess-3 ]
• Sign = 1, Exponent = 011, Mantissa = 1010
• STEP 1: Convert Exponent
• Excess-3 Value = 3
• Real Value = 3 - 3 = 0
• STEP 2: Apply Exponent to Mantissa
• 1.1010 x 20
• STEP 3: Convert Binary Fraction
1 1 5
• 1+ + =1
2 8 8
• STEP 4: Apply sign
5
• −1
8

© 2020 e-Learning Centre, UCSC 66

 IEEE Floating Point Standard

• IEEE has introduced standards for FP representations

• The standard is formally known as IEEE-754
• IEEE-754 Single Precision Floating Point

1 bit 8 bits 23 bits

Sign Exponent (Bias 127) Mantissa

• IEEE-754 Double Precision Floating Point

1 bit 11 bits 52 bits
Sign Exponent (Bias 1023) Mantissa

© 2020 e-Learning Centre, UCSC 67

 Round Off Error

5
• Try STORE +2 in 8-bit Floating Point Representation
16

𝟓
+𝟐 = 𝟏𝟎. 𝟎𝟏𝟎𝟏 In Binary
𝟏𝟔

© 2020 e-Learning Centre, UCSC 68

 Round Off Error

5
• Try STORE +2
16
5
+2 = 10.0101 = 𝟏. 𝟎𝟎𝟏𝟎𝟏 × 𝟐𝟏
16

1. STORE the SIGN BIT

© 2020 e-Learning Centre, UCSC 69

 Round Off Error

5
• Try STORE +2
16
5
+2 = 10.0101 = 𝟏. 𝟎𝟎𝟏𝟎𝟏 × 𝟐𝟏
16

2. STORE the MANTISSA

0 0 0 1 0 1

© 2020 e-Learning Centre, UCSC 70

 Round Off Error

5
• Try STORE +2
16
5
+2 = 10.0101 = 𝟏. 𝟎𝟎𝟏𝟎𝟏 × 𝟐𝟏
16

2. STORE the MANTISSA

0 0 0 1 0 1

The last bit of Mantissa cannot be stored. DISCARD that bit.

© 2020 e-Learning Centre, UCSC 71

 Round Off Error

5
• Try STORE +2
16
5
+2 = 10.0101 = 𝟏. 𝟎𝟎𝟏𝟎𝟏 × 𝟐𝟏
16

3. STORE the EXPONENT

0 1 0 0 0 0 1 0

© 2020 e-Learning Centre, UCSC 72

 Round Off Error

5
• Try STORE +2
16
5
+2 = 10.0101 = 𝟏. 𝟎𝟎𝟏𝟎𝟏 × 𝟐𝟏
16

3. STORE the EXPONENT

0 1 0 0 0 0 1 0

Conversion this number back to decimal will not give the

𝟓
original number (+𝟐 )
𝟏𝟔
© 2020 e-Learning Centre, UCSC 73
 Round Off Error

5
• Original Value = 2
16
• Stored Value = 01000010 = 𝟏. 𝟎𝟎𝟏𝟎 × 𝟐𝟏
= 10.01
1
=2
4
• Therefore, the Rounding Off Error in this case is
5 1
=2 −2
16 4

1
=
16

© 2020 e-Learning Centre, UCSC 74

 Range of Floating Point Representations

In a floating point representation model,

• What is the Largest Positive number ?
• What is the Smallest Positive number ?
• What is the Largest Negative number ?
• What is the Smallest Negative number ?

© 2020 e-Learning Centre, UCSC 75

 Range: The Largest Positive Number

• What is the largest positive number ?

• Sign bit should be 0
• Mantissa should be all 1s
• Exponent should be maximum (execess-k)

0 1 1 1 1 1 1 1

= 1.1111 x 24

= 31

© 2020 e-Learning Centre, UCSC 76

 Range: The Smallest Positive Number

• What is the largest positive number ?

• Sign bit should be 0
• Mantissa should be all 0s
• Exponent should be minimum (execess-k)

0 0 0 0 0 0 0 0

= 1.0000 x 2-3 = 1/8

= 0.125

© 2020 e-Learning Centre, UCSC 77

 Range: The Smallest Negative Number

• What is the largest positive number ?

• Sign bit should be 1
• Mantissa should be all 0s
• Exponent should be minimum (execess-k)

1 0 0 0 0 0 0 0

= -1.0000 x 2-3 = -1/8

= - 0.125

© 2020 e-Learning Centre, UCSC 78

 Range: The Largest Negative Number

• What is the largest positive number ?

• Sign bit should be 1
• Mantissa should be all 1s
• Exponent should be maximum (execess-k)

1 1 1 1 1 1 1 1

= -1.1111 x 24

= - 31

© 2020 e-Learning Centre, UCSC 79

 Range

• The range of 8 bit floating point representation that

we used can be illustrated as below.

Range Range
- 31 - 0.125 0.125 + 31

© 2020 e-Learning Centre, UCSC 80

 Challenge..!

• Try to find the ranges for IEEE-754 single and double

precision representations. (Fill the tables)

• IEEE-754 Single Precision

Negative Min Negative Max Positive Min Positive Max

• IEEE-754 Double Precision

Negative Min Negative Max Positive Min Positive Max

© 2020 e-Learning Centre, UCSC 81

 Special Cases: Zero

• Zero cannot be represented with this format of FP as

we assume there is always 1 immediate left of the
radix point.
• Therefore, IEEE-754 defines all 0s in mantissa and
exponent denote zero.
• Both cases are regarded as standard for 0
0 0000 0000 000 0000 0000 0000 0000 0000

1 0000 0000 000 0000 0000 0000 0000 0000

© 2020 e-Learning Centre, UCSC 82

 Special Cases: Infinity and NaN

• Infinity is denoted by all 1s in exponent with all 0s in

mantissa.
+ Infinity 0 1111 1111 000 0000 0000 0000 0000 0000

- Infinity 1 1111 1111 000 0000 0000 0000 0000 0000

• Not a Number (NaN) is denoted by all 1s in exponent

with any non-zero value in mantissa. Sign bit has no
meaning in this situation.

1 1111 1111 Any Non-zero Value

© 2020 e-Learning Centre, UCSC 83

 Special Cases: Denormalized Numbers

• IEEE-754 assumes that all FP numbers are normalized

(there is always single digit at the immediate left to
radix point and that digit is 1. Ex. 1.0101 x 2-3)
• However, IEEE-754 allows denormalized representation
• When exponent is all 0s and mantissa is non-zero, it is
assumed the number is in denormalized form.
• Ex. 1 0000 0000 Any Non-zero Value

Here, exponent is in excess-127 form. Hence the true value of

exponent is -127

© 2020 e-Learning Centre, UCSC 84

 Range for IEEE-754

• Max: 1.111 1111 1111 1111 1111 1111 x 2127

• Even though 2128 can be represented, it will make
all 1s in exponent, which is the NaN
• Min: 0. 000 0000 0000 0000 0001 x 2-127
• Since the exponent is -127 which is converted to all
0s with Excess-127 notation, the value represented
will be regarded as denormalized number. Hence,
there is no 1 left to the radix point.

© 2020 e-Learning Centre, UCSC 85

 Range for IEEE-754

• With the sign(+ or -), these Min and Max help to

define the -Min, -Max and +Min, +Max respectively.

Negative Positive
Overflow Overflow
0

Range Range
- Max - Min + Min + Max

Negative Positive
Underflow Underflow

© 2020 e-Learning Centre, UCSC 86

 Problems with Floating Point Numbers

• Out of range: Overflow and Underflow

• Within the range: Rounding off error
• Floating point arthmetic is not always associative
•Ex. There can be instances where three FP numbers
𝑎, 𝑏 and 𝑐, that
• 𝑎 + 𝑏 + 𝑐 ≠ 𝑎 + (𝑏 + 𝑐)
• There can be instance where the FP arithmetic is not
distributive.
• Ex. a × 𝑏 + 𝑐 ≠ 𝑎𝑏 + 𝑎𝑐

© 2020 e-Learning Centre, UCSC 87

 Character Codes

• Character encoding is the process of assigning

numbers to graphical characters that gives the
meaning to its representation.
• Character codes have evolved from basic coding
systems such Binary Coded Decimal to Unicode
• Some of the Coding Systems
• EBCDIC
• ASCII
• Unicode

© 2020 e-Learning Centre, UCSC 88

 EBCDIC

• EBCDIC stands for Extended Binary Coded Decimal

Interchange Code.
• 8-bit Binary Coded Decimal representation.
• Characters are represented by appending digit bits to
zone bits (see the figure in the next slide).
• Character A = 1100 0001
• Character a = 1000 0001

© 2020 e-Learning Centre, UCSC 89

EBCDIC Code

© 2020 e-Learning Centre, UCSC 90

EBCDIC Code Abbreviations

© 2020 e-Learning Centre, UCSC 91

 ASCII

• American Standard Code for Information Interchange

(ASCII)
• Use bit patterns of length seven to represent
•Letters of English alphabet: a - z and A - Z
•Digits: 0 – 9
•Punctuation symbols: (, ), [, ], {, }, ’, ”, !, /, \
•Arithmetic Operation symbols: +, -, *, <, >, =
•Special symbols: (space), %, $, #, &, @, ^
• 27 = 128 characters can be represented by ASCII

© 2020 e-Learning Centre, UCSC 92

 Character Representation: ASCII Table

Symbol ASCII Symbol ASCII Symbol ASCII Symbol ASCII

(space) 00100000 A 01000001 a 01100001 0 00110000

! 00100001 B 01000010 b 01100010 1 00110001

“ 00100010 C 01000011 c 01100011 2 00110010

# 00100011 D 01000100 d 01100100 3 00110011

$ 00100100 E 01000101 e 01100101 4 00110100

% 00100101 F 01000110 f 01100110 5 00110101

& 00100110 G 01000111 g 01100111 6 00110110

……. …….. ……. …….

© 2020 e-Learning Centre, UCSC 93
Character Representation: ASCII Table

© 2020 e-Learning Centre, UCSC 94

 Character Representation: Unicode

• EBCDIC and ASCII are built around the Latin alphabet

• Are restricted in their ability for representing non-
Latin alphabet
• Countries developed their own codes for native
languages

• Unicode: 16-bit system that can encode the characters

of most languages

• 16 bits = 216 = 65,636 characters

© 2020 e-Learning Centre, UCSC 95

 Character Representation: Unicode

• The Java programming language and some operating

systems now use Unicode as their default character
code

• Unicode code space is divided into six parts

• The first part is for Western alphabet codes,
including English, Greek, and Russian

• Downward compatible with ASCII and Latin-1

character sets

© 2020 e-Learning Centre, UCSC 96

 Character Representation: Example

• English section of Unicode Table

• ACSII equivalent of A is 4116
• Unicode is equivalent of A: 00 4116

Full chart list:

http://www.unicode.org/charts/

© 2020 e-Learning Centre, UCSC 97

 Sinhala Unicode

E.g. Unicode for Sinhala language’s character අ is 0D85

Sinhala Unicode Chart: https://unicode.org/charts/PDF/U0D80.pdf

© 2020 e-Learning Centre, UCSC 98

 Tamil Unicode

E.g. Unicode for Tamil language’s character அ is 0B85

Sinhala Unicode Chart: https://unicode.org/charts/PDF/U0B80.pdf

© 2020 e-Learning Centre, UCSC 99

Thank you..!

10
© 2020 e-Learning Centre, UCSC
0