Part I: Representation of the information

Basic Elements
The computer can process various types of information:
Numerical values, Texts, Images, Sound, …
BUT
Chapter 2: All this information is stored in digital form
Representation of information &
System number
Information

2023-2024 Instructions Data

Computer Architecture 1 Course, 1st year Computer Science Engineer 1 Computer Architecture 1 Course, 1st year Computer Science Engineer 2

Coding of information Some sizes

Whatever its nature (number, text, image, sound, 210 bits = 1024 bits = 1 Kb (1 Kilo bits) / 210 Ø = 1024 Ø =
1 KØ (1 Kilo Ø)
or video), digital information processed by a
210 Kb = 1024 Kb = 1 Mb (1 Mega bits) / 210 KØ = 1024
computer is always represented in binary form (a
KØ = 1 MØ (1 Mega Ø)
sequence of 0 and 1). For example: 01111011,
210 Mb = 1024 Mb = 1 Gb (1 Giga bits) / 210 MØ = 1024
11000000….. MØ = 1 GØ (1 Giga Ø)
 The smallest unit of information transmitted by a
210 Gb = 1024 Gb = 1 Tb (1 Tera bits)/ 210 GØ = 1024 GØ
computer is called Bit (BInary digiT) (which can take
= 1 TØ (1 Tera Ø)
two values: 0 or 1)
 A unit of information made up of 8 bits is called a
“byte”
Computer Architecture 1 Course, 1st year Computer Science Engineer 3 Computer Architecture 1 Course, 1st year Computer Science Engineer 4

Definition of Information Coding Steps in coding information

The coding of information consists of establishing a The coding of information passes by three steps:
correspondence between the (usual) external
representation of the information (text, number, image,
etc.), and its internal representation in the machine, which 1. Representation of information by a series of numbers
(Digitalization)
is always a series of bits.
2. Encoding each number in binary form
Example: the number 22
3. Represent each binary element by a physical state
(electrical signal)
• Its external representation = 22
• Its internal representation (in binary) = 00010110
How to represent numbers (integers, real, etc.) and
characters (letters, mathematical symbols, etc.) in the
machine?
6
Computer Architecture 1 Course, 1st year Computer Science Engineer 5 Computer Architecture 1 Course, 1st year Computer Science Engineer

1
 1. Representation of numbers: Representation of numbers in a base b
Number systems
 Number systems describe how numbers are represented. • A number (XXX)b indicates the representation of a number
 A num ber s ys tem is defined by:
XXX in the base b.
 A l p h a b e t ( A ) : A s e t o f s y m b o l s ( n u m b e r s ) : A = { a 1, • The usual bases that we know and use every day are:
a 2, … . , a n } - base 10 (decimal system) to represent different
 Rules for wr it ing num bers: J uxtapos ition of quantities, different figures and numbers, and
s ym bols
- base 60 to represent time.
 a 1a 3: i s a w o r d
• In a num ber s ys tem, the num ber of distinc t s ym bols is
c alled the bas e of the num ber s ystem (the c ardinal of the How to represent a number in a base b?
s et A) .
If b ≤ 10, we simply use the numbers 0 to b-1
• In c om puting , the m ost us ed bas es are binar y, octal, and
hexadecimal . Example: base 8 (octal system): any number will be the
combination of digits belonging to the set {0,…, 7}
7 8
Computer Architecture 1 Course, 1st year Computer Science Engineer

Representation of numbers in a base b Decimal system

(continued) • It is the number system that we frequently use in our daily activities.
• Based on 10 symbols {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} === base 10
If b > 10, we simply use the numbers 0 to 9 then the letters in • It is a positional system: Each position has a weight.
alphabetical order.
Example:
The number 5368 is written as following :
Example:
5368 = 8*100 +6*101 +3*102 +5*103
Base 16 (hexadecimal system): any number will be the
combination of symbols belonging to {0,…, 9, A, B, C, D, E, F}
Low weight Heavy weight
such that: (A=10 , ….., F=15).

A number of n digits (symbols) is a sequence (ai), 0 ≤ i ≤ n-1: an- 5368, 135=5*103+3*102 + 6*101 + 8*100 + 1*10-1+3*10-2+5*10-3
1 ….. a1 a0 such that: a0 is the least significant term and an-1 is
the most significant term. Decimal part
Integer part

9 10
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

Binary system Transcoding: Bases change

Tr ans c oding ( or bas es c onv er s ion) is the oper ation, whic h
• Base (b)=2 allows to g o fr om the r epr es entation of a num ber in one bas e
to its r epr es entation in another bas e.
• The system used in computers
• It uses two digits {0,1} Decimal Base b
Number N= integer part, decimal part (example: 15, 23)
• Example: (10111101)2

• The polynomi al form : Integ er p art: T he successive division method

Divide the number by B N B
(10111101)2 =20*1+21*0+22*1+23*1+24*1+25*1+26*0+27*1 T hen t he quot i ent by B
R1 q1 B
= (190)10
and so on until obtaining a zero quotient
R 2 q2 B
take the remainders of successive divisions on R 3 qn B
the base X in the opposite direction.
Rn 0
11 (N) 1 0 = (R n . . R 3 R 2 R 1 ) B 12
Computer Architecture 1 Course, 1st year Computer Science Engineer

2
 Decimal base to base b (0,23)10=(?)2
• Integer part:
Successive multiplications until having a zero result or obtaining a • Now let's move on to the decimal part :
given precision (0,23)10=(?)2
Example:
0,23x2= 0,46 integer number is 0
(115,23)10= (?)2
0,46x2= 0,92 integer number is 0
With a precision of 6 places after the decimal point. 0,92x2= 1,84 integer number is 1
We treat each part separately 0,84x2= 1,68 integer number is 1
Integral part: 0,68x2= 1,36 integer number is 1
115÷2=57 remainder 1
57÷2= 28 remainder 1
0,36x2= 0,72 integer number is 0
28÷2= 14 remaider 0
14÷2= 7 remainder 0 • Hence: (0,23)10=(0,001110)2
7÷2= 3 remainder 1
3÷2=1 remainder 1 • Final result: (115,23)10= (1110011, 001110)2
1÷2=0 remainder 1 (quotient=0 stop)
Hence: (115)10= (1110011)2 13 14
Computer Architecture 1 Course, 1st year Computer Science Engineer

Decimal Base to the Octal and Conversion from base b to the

Hexadecimal base decimal base
• Use polynomial expansion
Decimal  Octal Decimal  hexadecimal
X = (an.. a2a1a 0) b

=b 0 a 0 +b 1 a 1+… b n a n = (∑aibi)10
25 8 25 16 Examples:
1 3 8 9 1 16
• (11011101,1)2 = 2-1*1+ 20*1+21*0+22*1+23*1+24*1+25*0+26*1+27*1= (221,5)10
3 0 1 0 • (175,26)8= 8-1*2+8-2*6 + 80*5+81*7+82*1
• (14)16=160*4+161*1=(20)10
• (1011)2 = (1 × 23 + 0×22 + 1×21 + 1×20)10 = (1×8 + 0×4 + 1×2 + 1×1)10 = (11)10

• (16257)8 = 1 × 84 + 6×83 + 2×82 + 5×81 + 7×80 = 1×4096 + 6×512 + 2×64 + 5×8 + 7

(25)10= (31)8 (25)10= (19)16 = 4096 + 3072 + 128+ 40 + 7 = 7343

 (F53)16 = 15 × 162 + 5 × 161 + 3 × 160 = 15 × 256 + 5 × 16 + 3 = 3840 + 80 + 3 = 3923
15 16
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

Conversion from binary base to the Conversion from the octal base to the
octal base binary base
 Making 3-bits groups starting from the least significant one.  Replace each symbol in the octal base with its 3-bit binary
value
 Replace each group with the corresponding octal value. Octale Binaire
Example: ( 213)8
 3 binary digits ⇒ one octal digit

2 1 3
( 10111101)2
010 001 101
2 7 5

(275)8
17 18
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

3
 Arithmetic operations (the
Application exercises
addition)
In binary In Octal In Hexadecimal Perform the following operations and transform
the result to decimal each time:
The ad d ition

+ 0 1 retain
11
• (1111,101)2+(10,1)2=(?)2
0 0 1 1 111 11
11011
375 1BA • (45)8+(75)8=(?)8
+
1 1 10
+ + • (AB4)16+(253)16=(?)16
33 F6
10110
The multiplication
(110001)2 (430)8 (2B0)16
* 0 1

0 0 0

1 0 1
19 20
Computer Architecture 1 Course, 1st year Computer Science Engineer

Coding of natural (unsigned) integers

2. Internal data representation -pure binary code-
 A na t ur a l num b e r is a p o s it iv e o r z e r o int e g e r.
 To e nc o d e na t ur a l num b e r s , we us e p ur e b in a r y c o d e .
Unsigned Integers  T he na t ur a l num b e r is r e p r e s e nt e d in b a s e 2 o n n b it s .
 T he r a ng e o f num b e r s o n n b it s is : [ 0 , 2 n - 1 ]
 W it h n b it s , we c a n r e p r e s e nt 2 n num b e r s .

A number N Signed Integers

Example:
• On a byte, (17) 10 is coded in pure binary: 00010001
Real (fixed point, floating-
point)  The number of bits to use depends on the range of
numbers we want to use
 Using 1 byte (8 bits): we can code 2 8 values: [0 ; 255]
 Using 2 bytes (16 bits): we can encode 2 16 values: [0 ; 2 16-1 ]
 Using (n bits): we can code 2 n values: [0 ; 2 n-1 ]
21 22
Computer Architecture 1 Course, 1st year Computer Science Engineer

Coding of signed integers Signed magnitude representation

• These are num bers with a + or - sign. The most significant bit is used to represent the sign of the number
• 1: for a negative number
Example: -24, +354,….. • 0: for a positive number

• There are at least three techniques allowing the • The other (n-1) bits encode the magnitude (the absolute
representation of signed integers: value) of the number

• W ith n bits, we encode all the numbers between -(2 n-1 -1) and
1. Signed m agnitude representation (2 n-1 -1)

2. One's com plement representation (C1). • Example :

• On 8 bits, we can encode the numbers -13 and +17 in signed
3. Two’s com plement representation (C2) magnitude as follows:
-13 is coded by: 1 0 0 0 1 1 0 1
+17 is coded by: 0 0 0 1 0 0 0 1

23 24
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

4
 Advantages and disadvantages of
One's complement (C1)
signed magnitude
• Advantages: Simple • The first bit is reserved for the sign.
• Disadvantages (limites): • If the number is positive then the number keeps its format.
• If the number is negative then each bit (of the remaining bits) is
 Two representations of zero : inverted (0 becomes 1 and 1 becomes 0) (by completing on the left
with 0s to obtain an n-bit code).
On 8 bits : +0 = 00000000
• The number of possible combinations on n bits is 2n
-0 = 10000000 • With n bits, we encode all the numbers between -(2n-1-1) and (2n-1-1)
 Multiplication and addition are less obvious • Two combinations for 0
• Examples:
For example, we add -3 and -1 on 4 bits  -5 on 8 bits
5= (00000101)2
-5= (11111010)C1
 +7 on 8 bits
25 +7= (00000111)2= (00000111)C1 26
Computer Architecture 1 Course, 1st year Computer Science Engineer

Addition and subtraction in C1 Two's complement (C2)

It is based on the following principle: The represent ation of a number X in t wo's com plem ent on n bit s i s
• If no carry is generated by the sign bit then the result is correct and it is done as f ol l ows:
represented in C1.
• i f ( X > = 0 ) ( n u m b e r f r o m 0 t o ( 2 n-1 - 1 ) ) t h e n X i s c o d e d i n t h e s a m e
• Otherwise, it will be removed and added to the result of the operation, this w a y a s i n p u r e b i n a r y,
is represented in C1.
• i f ( X < 0 ) ( n u m b e r f r o m - ( 2 n-1- 1 ) t o 0 ) t h e n :
 Example 1:
 C od e | X| i n bi na ry by c om pl et i ng o n t h e l ef t wi t h 0 t o o bt ai n an n-
-14+5 on 5 bits
bit code
-14+5=(-1110+0101)2=(11110)SM+(00101)SM= (10001)C1+(00101)C1
 Inv ert all bits of the binary representation (one's complement) ;
=(10110)C1 no retain, so the result is correct and it is written in C1.
 Add 1 t o t he resul t (t wo' s com pl em ent or C2)
We must transform it into a binary number then decimal:
 The num ber of possi bl e com bi nati on s on n bi ts i s 2n
(10110)C1=(11001)SM= (-1001)2=-9
 W i t h n b i t s , w e e n c o d e a l l t h e n u m b e r s b e t w e e n -(2n-1) and (2n-1-1)
 Exemple 2:
14-6 on 5 bits Examples:
-5 on 8 bits
14-6=(1110-0110)2=(01110+10110)SVA=(01110+11001)C1=(00111)C1
-5= (00000101)2 = (11111010)C1 = (11111011)C2
with retained 1, the latter is added to the result obtained and we obtain
 +7 on 8 bits
(01000)C1= 8 since the number is positive
27 +7= (00000111)2= (00000111)C1 = (00000111)C2 28

Representation of some numbers on 4 bits

C2 MS Decimal
Addition and subtraction in C2
0000 0000 0 It is based on the following principle:
0001 0001 +1 • If there is a carry generated by the sign bit, it is ignored and the result is in C2
0010 0010 +2 • Otherwise the result is correct and in C2
0011 0011 +3 Example 1:
0100 0100 +4 -14+5 on 5 bits
0101 0101 +5 -14+5=(-1110+0101)2=(11110)SM+(00101)SM= (10001)C1+(00101)C1
0110 0110 +6 = (10010)C2+(00101)C2 =(10111)C2 no carry, so the result is correct and it is
0111 0111 +7 written in C2
1000 -8 We must transform it into a binary number then decimal
1001 1111 -7 (10111)C2= (10110)C1 =(11001)SM= (-1001)2=-9
1010 1110 -6 Example 2:
1011 1101 -5 14-6 on 5 bits
1100 1100 -4 14-6=(1110-0110)2=(01110+10110)SAV=(01110+11001)C1=(01110+11010)C2=(01000)C2
1101 1011 -3 with a carry 1, the latter is ignored and we obtain
1110 1010 -2 (01000)C2= (00110)C1=(00110)SAV=8 since the number is positive
1111 1001 -1
29 30
Computer Architecture 1 Course, 1st year Computer Science Engineer

5
 Representation of real numbers Fixed point
• A real num ber = the integer part + the decim al
• Example: +15,23, -234,01…… part .
• Two questions arise:
• T he integer part is coded o n “p” bits by
1. How to represent the comma in a machine? perform ing successiv e div isions b y 2 .
The designers did not take the comma into account, but
they offered a place in the representation of numbers. • T he decim al part is encoded on “q” bits b y
perform ing successive m ultiplications b y 2 until
2. How to tell the machine the position of the decimal the decim al part is zero o r the num ber of bits q is
point? reached .
0,625*2=1,25
 Fixed point Example : 12,625=(?)2 fixed point format 0,25*2=0,5
 Floating point Integer part : 12 = (00001100)2 0,5*2=1,0
Decim al part: 0,625 = (?)2
(12,625)10=(001100,101)2
31 32
Computer Architecture 1 Course, 1st year Computer Science Engineer

The IEEE 754 standard

Floating point (IEEE 754 standard)
(single precision)
• In c om puti ng, the IE EE 754 s tandard has bec om e
es tabl i s hed for the c odi ng of fl oati ng num bers .
• In the IE E E 754 s tandard, a floati ng poi nt num ber i s • Single precision on 32 bits
al ways repres ented by a tri pl e
 1 bit of the sign
(S; E ;M)
 S: the sign is coded on 1 most significant bit  8 bits for the exponent
(1: negative and 0 positive)  23 bits for the mantissa
 E: the exponent
 M: the mantissa The sign The exponent The Mantissa

The sign The exponent The Mantissa 1 bit 8 bits 23 bits

1 bit P bits q bits

33 34
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

IEEE 754 standard (double

T h e IE E E 754 sta n d ard
precision)
IEEE 754 Coding Steps:
• Double precision on 64 bits :
[Link] representation of the number X in
floating point format : X= ± 1,M . 2dec
1 bit of the sign
11 bits for the exponent • Example:
 52 bits for the mantissa
(24,5)10=(11000,1)2 (fixe point)
The sign The exponent The Mantissa
= + 1,10001 * 2 4 (floating point)
1 bit 11 bits 52 bits

35 36
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

6
 IEEE 754 Coding Steps Application exercise
[Link] of the exponent E (biased/shifted or
normalized) • Question:
Convert the decimal number (12,25)10 in the floating
Exponent (E biased) = dec + 2p-1 – 1 point format according to the IEEE 754 single
precision standard
• Solution:
 Single precision (32 bits, p=8) : E = dec +127 Converting the number 12.25 to binary
(12,25)10=(1100,01)2
 Double precision (64 bits, p=11) : E= dec +1023 =1,10001*23
Hence
Shift (power)

• 12,25=1,10001*23
37 Th e mantissa (M) 38
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

Application exercise Converting IEEE 754 to Decimal

(continued)
• Estimation of the elements of the number Conv erting a num ber X from IEEE 754 to decim al
m eans decom posing this num ber into its elem ents :
T he s ig n (S) =0 (pos itiv e num ber)
S,E ; then estim ating its representation in floating
The 8-bit exponent : E=dec+127=3+127=130=> (10000010)2 point form at (X= ± 1,M . 2dec )
 T he mantissa M on 23 bits => 10001000000000000000000 Example:
• X=(010000010110100000000000000000000)
0 10000010 10001000000000000000000 • X= 0 10000010 110100000000000000000000
M
S E
1 bit
8 bits 23 bits

So, the number (12,25)10 in floating point according to the IEEE 754 S=0 => positive number
single precision standard is : E=(10000010) 2=130 ; E=
dec+ 127 => dec=130-127=3
01000001010001000000000000000000
X= + 1,M * 23 = 1,11010000000000000000000 * 23 (dec=3)

39 X=+ (1110,10) 2 = (14,5)10 40

or more readably in hexadecimal representation (41040000)16 Computer Architecture 1 Course, 1st year Computer Science Engineer

Binary Coded Decimal (BCD) 3. Characters encoding

 T he BCD code is the 8,4,2,1 code .
 8 , 4, 2, and 1 are w eights
 The characters:
Alphabetical (A-Z , a-z),
 This code is the sim plest, m ost intuitive binary
code for decim al digits a n d uses the sam e pow ers o f Digital (0,1,2,3,4,5,6,7,8,9),
2 a s a binary num ber, b u t onl y encodes the first ten Punctuation( ;. ? ! … . )
v alues from 0 to 9 .
Specials (&, $, %,,…)
Exam ple : 12 (1100 in pure binary, 0001 0010 in DCB)

Do NOT mix up conversion of a decimal number to a binary number Character coding is done using a table of
with coding a decimal number with a BINARY CODE. correspondence between characters and binary
1310 = 11012 (This is conversion) num bers.
13  0001|0011 (This is coding)
41 42
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

7
 Character encoding: ASCII code The ASCII code table (1)

 ASCII (American Standard Code for Information Interchange)

is a computer standard for character encoding that emerged
in the 1960s.
 The basic ASCII code represented 7-bit characters (128
possible characters, from 0 to 127).
o Codes from 48 to 57: numbers in order (0,1,…,9)
o Codes from 65 to 90: capital letters (A….Z)
o Codes from 97 to 122: lowercase letters (a….z).

 This code was developed for the English language, so it does not
contain accented characters or language-specific characters.

43 44
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer

Extended ASCII code The ASCII code table (2)

 The ASCII code has been extended to 8 bits to be able to
encode more characters (0 to 255) => extended ASCII code.
 Allows us to code accented characters: à, é, è,…etc.

45 46

Examples of ASCII encoding Unicode code

 Developed in 1991
Binary Hexadecimal Decimal  It uses 16 bits to represent 65,536 characters (0 to
H = 01001000 = 48 = 72
e = 01100101 = 65 = 101 65,535)
l = 01101100 = 6C = 108  Unicode defines tens of thousands of codes, but the
l = 01101100 = 6C = 108
o = 01101111 = 6F = 111 first 128 remain compatible with ASCII.
, = 00101100 = 2C = 44  It codes most alphabets: Arabic, Chinese, Turkish, etc.
Espace = 00100000 = 20 = 32
w = 01110111 = 77 = 119  We refer to a character by its number written in
o = 01100111 = 67 = 103
r = 01110010 = 72 = 114
hexadecimal preceded by “U+”.
l = 01101100 = 6C = 108  For example, the Latin letter “a” corresponds to
d = 01100100 = 64 = 100
U+0061 (in hexadecimal)
 [Link]
47 48
Computer Architecture 1 Course, 1st year Computer Science Engineer Computer Architecture 1 Course, 1st year Computer Science Engineer