Chapter Three

Number Systems and Data

Representation

1
3.1 Data Types
 Binary information is stored in memory or processor registers
 Registers contain either data or control information
 Data are numbers and other binary-coded information
 Control information is a bit or a group of bits used to specify the
sequence of command signals
 Data types found in the registers of digital computers
 Numbers used in arithmetic computations
 Letters of the alphabet used in data processing
 Other discrete symbols used for specific purpose
» Number Letter) gray code, error detection code, …
 Number Systems
 Base or Radix r system : uses distinct symbols for r digits
 Most common number system :Decimal, Binary, Octal, Hexadecimal
 Positional-value(weight) System : r2 r 1r0.r-1 r-2 r-3
» Multiply each digit by an integer power of r and then form he sum of all
weighted digits

2
 Decimal System/Base-10 System
 Composed of 10 symbols or numerals(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0)
 Binary System/Base-2 System
 Composed of 10 symbols or numerals(0, 1)
 Bit = Binary digit
 Hexadecimal System/Base-16 System : Tab. 3-2
 Composed of 16 symbols or numerals(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F)
 Binary-to-Decimal Conversions
1011.1012 = (1 x 23) + (0 x 22)+ (1 x 21) + (1 x 2o) + (1 x 2-1) + (0 x 2-2) + (1 x 2-3)
= 810+ 0 + 210 + 110 + 0.510 + 0 + 0.12510
= 11.62510
 Decimal-to-Binary Conversions Decimal point conversion
Repeated division 0.375 x 2 = 0.750 integer 0 MSB
37 / 2 = 18 remainder 1 (binary number will end with 1) : LSB 0.750 x 2 = 1.500 integer 1 .
18 / 2 = 9 remainder 0 0.500 x 2 = 1.000 integer 1 LSB
Read the result downward .37510 = .0112
9 / 2 = 4 remainder 1
4 / 2 = 2 remainder 0
2 / 2 = 1 remainder 0
1 / 2 = 0 remainder 1 (binary number will start with 1) : MSB
Read the result upward to give an answer of 37 10 = 1001012

3
 Hex-to-Decimal Conversion Table 3-2
Hex Binary Decimal
2 1 o 0 0000 0
2AF16 = (2 x 16 ) + (10 x 16 ) + (15 x 16 ) 1 0001 1
= 51210 + 16010 + 1510 2 0010 2
3 0011 3
= 68710 4 0100 4
5 0101 5
 Decimal-to-Hex Conversion 6
7
0110
0111
6
7
42310 / 16 = 26 remainder 7 (Hex number will end with 7) : LSB 8 1000 8
9 1001 9
2610 / 16 = 1 remainder 10 A 1010 10
B 1011 11
110 / 16 = 0 remainder 1 (Hex number will start with 1) : MSB C 1100 12
D 1101 13
Read the result upward to give an answer of 42310 = 1A716 E 1110 14
F 1111 15

 Hex-to-Binary Conversion  Binary-to-Hex Conversion

9F216 = 9 F 2 1 1 1 0 1 0 0 1 1 02 = 0 0 1 1 1 0 1 0 0 1 1 0

= 1001 1111 0010 3 A 6

= 1001111100102 = 3A616

4
 Binary-Coded-Decimal Code
 Each digit of a decimal number is represented by its binary equivalent
8 7 4 (Decimal)

1000 0111 0100 (BCD)

 Only the four bit binary numbers from 0000 through 1001 are used
 Comparison of BCD and Binary
13710 = 100010012 (Binary) - require only 8 bits
13710 = 0001 0011 0111BCD (BCD) - require 12 bits
 Alphanumeric Representation
 Alphanumeric character set
» 10 decimal digits, 26 letters, special character($, +, =,….)
» A complete list of ASCII : (Find in Google: or https://www.ascii-code.com/)
 ASCII(American Standard Code for Information Interchange)
» Standard alphanumeric binary code uses seven bits to code 128
characters, but extended ASCII use 8-bit and code 256 characters.

5
 3.2 Complements
 Complements are used in digital computers for simplifying the
subtraction operation and for logical manipulation
 There are two types of complements for base r system
 1) r’s complement 2) (r-1)’s complement
» Binary number : 2’s or 1’s complement
» Decimal number : 10’s or 9’s complement
N : given number
 (r-1)’s Complement r : base
 (r-1)’s Complement of N = (rn-1)-N n : digit number
» 9’s complement of N=546700
(106-1)-546700= (1000000-1)-546700= 999999-546700
= 453299 546700(N) + 453299(9’s com)
» 1’s complement of N=101101 =999999
(26-1)-101101= (1000000-1)-101101= 111111-101101
= 010010 101101(N) + 010010(1’s com)
=111111
 r’s Complement
 r’s Complement of N = rn-N * r’s Complement
(r-1)’s Complement +1 =(rn-1)-N+1= rn-N
» 10’s complement of 2389= 7610+1= 7611
» 2’s complement of 1101100= 0010011+1= 0010100
6
  Subtraction of Unsigned Numbers (M-N), N0
 1) M + (rn-N)
 2) M  N : Discard end carry, Result = M-N
 3) M  N : No end carry, Result = - r’s complement of (N-M)

» Decimal Example)
M  N 72532(M) - 13250(N) = 59282 MN 13250(M) - 72532(N) = -59282
72532 13250
Discard + 86750 (10’s complement of 13250) + 27468 (10’s complement of 72532)
End Carry 1 59282 0 40718
No End Carry
Result = 59282 Result = -(10’s complement of 40718)
= -(59281+1) = -59282
» Binary Example)
X  Y 1010100(X) - 1000011(Y) = 0010001 XY 1000011(X) - 1010100(Y) = -0010001
1010100 1000011
+ 0111101 (2’s complement of 1000011) + 0101100 (2’s complement of 1010100)
1 0010001 0 1101111
Result = 0010001 Result = -(2’s complement of 1101111)
= -(0010000+1) = -0010001

7
 *Numeric Data
1) Fixed Point
 3.3 Fixed-Point Representation 2) Floating Point
 Computers must represent everything with 1’s and 0’s, including the
sign of a number and fixed/floating point number
* 32.25
 Binary/Decimal Point 1) 0.25, 2) 32.0, 3) 32.25
 The position of the binary/decimal point is needed to represent fractions,
integers, or mixed integer-fraction number
 Two ways of specifying the position of the binary point in a register
 1) Fixed Point : the binary point is always fixed in one position
» A binary point in the extreme left of the register(Fraction : 0.xxxxx)
» A binary point in the extreme right of the register(Integer : xxxxx.0)
 The binary point is not actually present, but the number stored in the register is treated
as a fraction or as an integer
 2) Floating Point : the second register is used to designate the position of the
binary point in the first register
* MSB for Sign
 Integer Representation +14 -14 “0” is plus +
“1” is minus -
 Signed-magnitude representation 0 0001110 1 0001110
Most
Common  Signed-1’s complement representation 0 0001110 1 1110001
 Signed-2’s complement representation
0 0001110 1 1110010
8
  Arithmetic Addition (-12) + (-13) = -25
 Addition Rules of Ordinary Arithmetic
(+12) + (+13) = +25
» The signs are same : sign= common sign, result= add (+25) + (-37)
» The signs are different : sign= larger sign, result= larger-smaller = 37 - 25 = -12

 Addition Rules of the signed 2’s complement *Addition Example)

» Add the two numbers including their sign bits + 6 00000110 - 6 11111010
» Discard any carry out of the sign bit position + 13 00001101 + 13 00001101
+ 19 00010011 + 7 00000111
 Arithmetic Subtraction
 Subtraction is changed to an Addition + 6 00000110 - 6 11111010
- 13 11110011 - 13 11110011
» (± A) - (+ B) = (± A) + (- B)
- 7 11111001 - 19 11101101
» (± A) - ( - B) = (± A) + (+ B)
* Subtraction Example) (- 6) - ( - 13) = +7
11111010 - 11110011 = 11111010 + 2’s comp of 11110011
Discard = 11111010 + 00001101
End Carry = 1 00000111 = +7

 Overflow
 Two numbers of n digits each are added and the sum occupies n+1 digits
 n + 1 bit cannot be accommodated in a register with a standard length of n
bits(many computer detect the occurrence of an overflow, and a
corresponding F/F is set) 9
  Overflow
 An overflow may occur if the two numbers added are both positive or both
negative
» When two unsigned numbers are added
 an overflow is detected from the end carry out of the MSB position
» When two signed numbers are added * Overflow Example)
 the MSB always represents the sign
out in out in
carries 0 1 carries 1 0
- the sign bit is treated as part of the number
+ 70 0 1000110 - 70 1 0111010
- the end carry does not indicate an overflow
+ 80 0 1010000 - 80 1 0110000
 Overflow Detection + 150 1 0010110 - 150 0 1101010
 Detected by observing the carry into the sign bit position and the carry out
of the sign bit position
 If these two carries are not equal, an overflow *Decimal Exam) (+375) + (-240)
condition is produced 375 + (10’s comp of 240)= 375 + 760

 Decimal Fixed-Point Representation 0 375 (0000 0011 0111 0101)

 A 4 bit decimal code requires four F/Fs +9 760 (1001 0111 0110 0000)
* Advantage * 0 135 (0000 0001 0011 0101)
for each decimal digit
Computer I/O
data are generated  The representation of 4385 in BCD requires 16 F/Fs (0100 0011 1000 0101)
by people who use
 The representation in decimal is wasting a considerable amount of storage
the decimal
system space and the circuits required to perform decimal arithmetic are more
complex 10
 3.4 Floating-Point Representation
 The floating-point representation of a number has two parts
 1) Mantissa : signed, fixed-point number
 2) Exponent : position of binary(decimal) point
* Decimal + 6132.789
 Scientific notation : m x re (+0.6132789 x 10+4) Fraction Exponent
+0.6132789 +4
 m : mantissa, r : radix, e : exponent
 Example : m x 2e = +(.1001110)2 x 2+4 Fraction Exponent
01001110 000100
 Normalization
 Most significant digit of mantissa is nonzero
 N.B In this representation decimal numbers are represented with a
fixed length format. In order not to waste bits, the representation will
normalize all the numbers.
 For example, 0.000123 wastes three zeroes on the left before non -
zero digits. There zeroes have no meaning except to indicate the
position of the Decimal point.
 Normalizing this number result in .123x10-3 .123 is the normalized
mantissa; -3 is the exponent. We have normalized this by eliminating all
the meaningless zeroes to the left of the first non-zero digit and by
adjusting the exponent. 11
 IEEE Floating Point
 IEEE Standard 754
 Supported by all major CPUs
 The IEEE standards committee consisted mostly of hardware
people, plus a few academics led by W. Kahan at Berkeley.
 Main goals:
 Consistent representation of floating point numbers by all
machines .
 Correctly rounded floating point operations.
 Consistent treatment of exceptional situations such as division
by zero.
 Example 1: 22.1 is normalized as .221x102.
 The general form of floating point representation is Mx10E where M
is the mantissa, and E is the exponent.
 It can be seen that a normalized number is characterized by a
mantissa less than 1 and greater than or equal to 1, all cases when
the number is not zero.
12
 Example 2: 111.01 is normalized as .11101x23.
 The mantissa is 11101. The exponent is 3.
 The general structure of floating point is:

 In representing a number in floating point we use 1 bit for sing, some

bits for exponent and the remaining bit for mantissa.
 In floating point representation the exponent is represented by a
biased exponent (Characteristic).
 Biased exponent = true exponent + excess 2n-1,
where n is the number of bits representing the
exponent.

13
 Example 1.
 Represent –234.375 in floating point using 7 bit for exponent and 16 bit
for mantissa.
 First we have to change to normalized binary
 i.e 234 = 11100010
 0.375= 0.011 , Then
 234.375 = 11100010.011 = 0.11100010011x2 8
 True exponent = 8
 Excess 2 n-1 = 2 7-1= 26= 64, Finally;
 Biased exponent = 8+26 = 8+64 = 72
= 100 1000 2
 Therefore –234.375 is represented as:

14
 Example 2.
 Represent 34.25 in floating point using 7 bit for
exponent and 24 bits for mantissa.
 34.25 = 1000 10.0 12
 The normalized form of 34.25 = .10001001x 26
 True exponent = 62 n-1 = 2 7-1= 6+26
=6+64=70
 70 = 10001102
 Therefore, 34.25 is represented as

0 1000110 100010010000…..0

15
 3.5 Error Detection Codes
 Binary information transmitted through some form of communication
medium is subject to external noise

~
Transmitter Receiver
 Parity Bit
 An extra bit included with a binary message to make the total number of 1’s
either odd or even
 Even-parity method
 The value of the parity bit is chosen so that the total number of 1s (including the
parity bit) is an even number
1 1 0 0 0 0 1 1
Added parity bit
 Odd-parity method
 Exactly the same way except that the total number of 1s is an odd number

1 1 0 0 0 0 0 1
Added parity bit
16
 Even/odd parity (1)

 Computers can sometimes make errors when

they transmit data.
 Even/odd parity:
 is basic method for detecting if an odd number of bits
has been switched by accident.
 Odd parity:
 The number of 1-bit must add up to an odd number
 Even parity:
 The number of 1-bit must add up to an even number

17
 Even/odd parity (2)

 The computer knows which parity it is using

 If it uses an even parity:
 If the number of of 1-bit add up to an odd number
then it knows there was an error:
 If it uses an odd:
 If the number of of 1-bit add up to an even number
then it knows there was an error:
 However, If an even number of 1-bit is flipped the parity
will still be the same. But an error occurs
 The even/parity can’t this detect this error:

18
 Even/odd parity (3)

 It is useful when an odd number of 1-bits is flipped.

 Suppose we have an 7-bit binary word (7-digits).

 If you need to change the parity you need to add 1
(parity bit) to the binary word.
 You now have 8 digit word.
 However, the computer knows that the added bit is a
parity bit and therefore ignore it.

19
 Example (1)

 Suppose you receive a binary bit word “0101” and you know you are
using an odd parity.
 Is the binary word errored?
 The answer is yes:
 There are 2 1-bit, which is an even number
 We are using an odd parity
 So there must have an error.

20
 Parity Bit
 A single bit is appended to each data chunk
 makes the number of 1 bits even/odd
 Example: even parity
 1000000(1)
 1111101(0)
 1001001(1)
 Example: odd parity
 1000000(0)
 1111101(1)
 1001001(0)

21
 Parity Checking

 Assume we are using even parity with 7-bit ASCII.

 The letter V in 7-bit ASCII is encoded as 0110101.
 How will the letter V be transmitted?
 Because there are four 1s (an even number), parity is set to zero.
 This would be transmitted as: 01101010.
 If we are using an odd parity:
 The letter V will be transmitted as 01101011

22
 Exercise
 Suppose you are using an odd parity. What should the binary word
“1010” look like after you add the parity bit?
 Answer:
 There is an even number of 1-bits.
 So we need to add another 1-bit
 Our new word will look like “10101”.
 Suppose you are using an even parity. What should the binary word
“1010” look like after you add a parity bit?
 Answer:
 There is an even number of 1’s.
 So we need to add another 0
 Our new word will look like “10100”.

23