CPNM Lecture 8 - Number Systems and Number

Representations

Mridul Sankar Barik

Jadavpur University

November 22, 2021

Number System

I A number system is merely a convention for representing

quantities
I A base is the number used as the reference for constructing
the system and also defines number of distinct symbols
(digits)
I Different quantities are represented as combinations of the
basic digits, with the position or place value specifying the
magnitude
I Positional notation: dn dn−1 . . . d0 .d−1 d−2 . . . d−m is equal to
dn × b n + dn−1 × b n−1 + . . . + d0 × b 0 + d−1 × b −1 + d−2 ×
b −2 + . . . + d−m × b −m
I A radix point is the symbol used in numerical representations
to separate the integer part of a number from its fractional
part
Binary I
I Base: 2
I Distinct Symbols: 0, 1
I Any binary number an an−1 . . . a0 .a−1 a−2 . . . a−m is equal to
an × 2n + an−1 × 2n−1 + . . . + a0 × 20 + a−1 × 2−1 + a−2 ×
2−2 + . . . + a−m × 2−m in decimal
I Example: (1011)2 = (1 × 23 + 0 × 22 + 1 × 21 + 1 × 20 )2 =
(11)10
I Example: (101011)2 =
(1 × 25 + 0 × 24 + 1 × 23 + 0 × 22 + 1 × 21 + 1 × 20 )2 = (43)10
I Example:
101.011 = 1×22 +0×21 +1×20 +0×2−1 +1×2−2 +1×2−2 =
4 + 0 + 1 + 0 + 41 + 18 = 5 + 0.25 + 0.125 = 5.375
Binary II

I Decimal to Binary
I For the integral part successively divide by 2 until the quotient
is 0; take the remainders in reverse order
I For the fraction part successively multiply by 2 until the
fraction part becomes 0; take the integral part of the products
in order
I Example: Convert 10.625 to binary
Quotient Remainder
10/2 5 0
5/2 2 1
2/2 1 0
1/2 0 1

Table 1: Integral Part

I (10)10 = (1010)2
Binary III
Integral Part
0.625 * 2 1.250 1
0.250 * 2 0.500 0
0.500 * 2 1.000 1

Table 2: Fraction Part

I (0.625)10 =(0.101)2
I (10.625)10 =(1010.101)2
I Binary Addition:
I 0+0=0
I 0+1=1
I 1+0=1
I 1 + 1 = 0 and carry 1
I Example
1 0 1 0
+ 1 1 0 1
1 0 1 1 1
Binary IV
I Example
1 1 1 0 1 1
+ 1 0 1 0 0 1
1 1 0 0 1 0 0
I Binary Subtraction:
I 0−0=0
I 0 − 1 = 1 and borrow 1
I 1−0=1
I 1−1=0
I Example
1 0 1 0
- 0 1 0 1
0 1 0 1
Octal I

I Base: 8
I Distinct Symbols: 0, 1, 2, 3, 4, 5, 6, 7
I Example: (26)8 = (2 × 81 + 6 × 80 )10 = (22)10
I Example: (723)8 = (7 × 82 + 2 × 81 + 3 × 80 )10 =
(448 + 16 + 3)10 = (467)10
I Binary to Octal: group bits from least significant position in
group size of three
I Example: (100101110101)2 = (100 101 110 101)2 = (4565)8
I Add extra zeros in most significant position if number of bits is
not a multiple of three
I Example: (1010110010)2 = (1 010 110 010)2 =
(001 010 110 010)2 = (1262)8
I Example: (1001111010.1010010)2 =
(001 001 111 010.101 001 000)2 = (1172.510)8
Hexadecimal

I Base: 16
I Distinct Symbols: 0, 1, 2, 3, 4, 5 ,6, 7, 8, 9, a, b, c, d, e, f
I Example: (2a)16 = (2 × 161 + 10 × 160 )10 = (42)10
I Example: (3e8)16 = (3 × 162 + 14 × 161 + 8 × 160 )10 =
(768 + 224 + 8)10 = (1000)10
I Binary to Hexadecimal: group bits from least significant
position in group size of four
I Example: (1100101011001011)2 = (1100 1010 1100 1011)2 =
(cacb)16
I Add extra zeros in most significant position if number of bits is
not a multiple of four
I Example: (1001010110010)2 = (1 0010 1011 0010)2 =
(0001 0010 1011 0010)2 = (12b2)16
Numbers with Different Bases
Decimal Binary Octal Hexadecimal
00 0000 00 0
01 0001 01 1
02 0010 02 2
03 0011 03 3
04 0100 04 4
05 0101 05 5
06 0110 06 6
07 0111 07 7
08 1000 10 8
09 1001 11 9
10 1010 12 a
11 1011 13 b
12 1100 14 c
13 1101 15 d
14 1110 16 e
15 1111 17 f
(r-1)’s Complement I
I To find (r-1)’s complement, subtract each digit of a given
number from the largest digit in that number system
I Example: Find 9’s complement of (7548)10
9 9 9 9
-7 -5 -4 -8
2 4 5 1
I Example: Find 7’s complement of (7543)8
7 7 7 7
-7 -5 -4 -3
0 2 3 4
I Example: Find 15’s complement of 15a7
f f f f
-1 -5 -a -7
e a 5 8
(r-1)’s Complement II

I Example: Find 1’s complement of 1101

1 1 1 1
-1 -1 -0 -1
0 0 1 0
r’s Complement
I To find r’s complement of a given number add 1 to (r-1)’s
complement of that number
I Example: Find 10’s complement of 2394
9 9 9 9
-2 -3 -9 -4
7 6 0 5
+ 1
7 6 0 6
I Example: Find 2’s complement of 1101
1 1 1 1
-1 -1 -0 -1
0 0 1 0
+ 1
0 0 1 1
Subtraction using r’s Complement

I Subtraction of two positive numbers (M-N) both of base r

may be done as follows
I Add the minuend to the r’s complement of subtrahend N
I Inspect the result obtained in step1 for an end carry
I If an end carry occurs discard it
I If an end carry does not occur, take the r’s complement of the
number obtained in step1 and place a negative sign in front
Subtraction using r’s Complement - Example

I Subtract 2957 - 1396

I Find 10’s complement of 1396
9 9 9 9
-1 -3 -9 -6
8 6 0 3
+ 1
8 6 0 4
I Add 8604 to 2956
2 9 5 7
8 6 0 4
1 1 5 6 1
I Discard end carry and the result is 1561
Subtraction using r’s Complement - Example I
I Subtract 1396 - 2957
I Find 10’s complement of 2957
9 9 9 9
-2 -9 -5 -7
7 0 4 2
+ 1
7 0 4 3
I Add 8604 to 2956
1 3 9 6
7 0 4 3
8 4 3 9
I No end carry, so result is negative and is in 10’s complement
form
I Find 10’s complement of 8439
Subtraction using r’s Complement - Example II

9 9 9 9
-8 -4 -3 -9
1 5 6 0
+ 1
1 5 6 1

I So, the result is -1561

Integer Representation I
I Assume that, 3 bits are used for storing integers
I If the integers are unsigned, then there are 8 distinct integer
values that can be stored. Range = 0 to 2n − 1
Integer Value Binary
0 000
1 001
2 010
3 011
4 100
5 101
6 110
7 111

I If the integers are signed, then most significant bit is used as

the sign bit
I Using Sign Magnitude method for representing signed
numbers. Range = +(2n−1 − 1) to −(2n−1 − 1); Drawback -
Two representations of zero
Integer Value Binary
(Signed Magnitude)
+0 000
+1 001
+2 010
+3 011
-0 100
-1 101
-2 110
-3 111
Integer Representation II
I Using 1’s Complement Method for representing signed
numbers. Range = +(2n−1 − 1) to −(2n−1 − 1); Drawback -
Two representations of zero
Integer Value Binary
(1’s Complement)
+0 000
+1 001
+2 010
+3 011
-3 100
-2 101
-1 110
-0 111

I Using 2’s Complement Method for representing signed

numbers. Range = +(2n−1 − 1) to −(2n−1 ); Single
representation of zero
Integer Value Binary
(2’s Complement)
0 000
+1 001
+2 010
+3 011
-4 100
-3 101
-2 110
-1 111
Floating Point Representation

I A fixed point representation uses fixed number of digits after

the radix point whereas a floating point representation allows
varying number of digits
I Floating point numbers are represented in the form m × b e ,
where m is a fractional part and e an integer part
I m is called a mantissa or significand and e an exponent or
characteristic
1
I Normalization: requires that b ≤m<1
1
I Example: 34 = 0.029411765... can be stored as 0.0294 × 100
in a system that allows four decimal places in mantissa.
However, normalized representation 0.2941 × 10−1 allows us
to retain one additional significant digit
I Source of Error - mantissa holds only a finite number of
significant digits
Floating Point Numbers - Example of a Hypothetical
Machine I
[Taken from Rajaraman]
I Each location can store 6 digits and has provision to store one
or more signs
I Fixed point representation: | 2456 | 24 |
I Max. no. = 9999.99
I Min. no = 0000.01
I Normalized floating point representation:
I Mantissa: less than 1 or greater than equal to 0.1
I Exponent: power of 10 which multiplies mantissa
I Example: 44.85 × 106 = 0.4485 × 108
(+) | 4485 | (+)08 |
I Example: .004854 = 0.4854 × 10−2
(+) | 4854 | (−)02 |
I Range: 0.9999 × 1099 to 0.1000 × 10−99
Floating Point Numbers - Example of a Hypothetical
Machine II
[Taken from Rajaraman]
I Arithmetic operations with normalized floating point numbers
I Addition:
I .4546E5 + .5433E5
Exponents are equal add mantissas, sum = .9979E5
I .4546E5 + .5433E7
Operand with larger exponent is kept as it is
Operand with smaller exponent is shifted right by number of
places equal to the difference in two exponents
.4546E5 + .5433E7 = .0045E7 + .5433E7 = .5478E7
I If the two numbers are m1 × b e1 and m2 × b e2 and e1 < e2
then the first number is transformed as m1 × b −(e2 −e1 ) × b e2
before addition
I .4546E3 + .5433E7 = .0000E7 + .5433E7 = .5433E7
I .6434E3 + .4845E3 = 1.1279E3 = .1127E4
I .6434E99 + .4845E99 = 1.1279E99 = .1127E100
Exponent part cannot store more than two digits ⇒
OVERFLOW
Floating Point Numbers - Example of a Hypothetical
Machine III
[Taken from Rajaraman]
I Subtraction
I .5452E-3 - .9432E-4 = .5452E-3 - .0943E-3 = .4509E-3
I .5452E3 - .5424E3 = .0028E3 = .2800E1
I .5452E-99 - .5424E-99 = .0028E-99 = .2800E-101
Cannot store a number smaller than the smallest number ⇒
UNDERFLOW
I Multiplication:
I (i) multiply the mantissa; (ii) add the exponent; (iii) result
mantissa is normalized and the exponent adjusted
appropriately
I .5543E12 × .4111E-15 = .22787273E-3 = .2278E-3
I .1111E10 × .1234E15 = .01370974E25 = .1370E24
I .1111E51 × .4444E50 = .04937284E101 = .4937E100 ⇒
OVERFLOW
I .1234E-49 × .1111E-54 = .01370974E-103 = .1370E-104 ⇒
UNDERFLOW
Floating Point Numbers - Example of a Hypothetical
Machine IV
[Taken from Rajaraman]

I Division:
I (i) mantissa of numerator is divided by that of the
denominator; (ii) denominator exponent is subtracted from
the numerator exponent; (iii) quotient mantissa is normalized
and exponent adjusted appropriately
I .9998E1 ÷ .1000E-99 = 9.9980E100 = .9998E101 ⇒
OVERFLOW
I .9998E-5 ÷ .1000E98 = 9.9980E-103 = .9998E-102 ⇒
UNDERFLOW
I .1000E5 ÷ .9999E3 = .100010001E2 = .1000E2
Floating Point Numbers - Example I

I Assume 8 bits are used for representation. 1 Sign bit, 3 bits

for exponent and 4 bits for mantissa → b | bbb | bbbb
normalized
I 110.11 −−−−−−→ 0.11011 × 23 → 0 | 011 | 1101
normalized
I 0.01011 −−−−−−→ 0.1011 × 2−1 → 0 | 111 | 1011
normalized
I 0.001011 −−−−−−→ 0.1011 × 2−2 → 0 | 110 | 1011
I When we normalize a decimal fraction, the digit after decimal
point can be any one of 1, 2, ..., 9
I When we normalize binary fraction, the bit after the binary
point is always 1 → so, we can store an extra bit by assuming
that the first bit in the mantissa is always 1 and this bit is not
actually stored
normalized
0.01011 −−−−−−→ 1.0110 × 2−2 → 0 | 110 | 0110
normalized
0.001011 −−−−−−→ 1.0110 × 2−3 → 0 | 101 | 0110