CS2100 Computer Organisation

http://www.comp.nus.edu.sg/~cs2100/ Number Systems and Codes

(AY2012/3) Semester 2

NUMBER SYSTEMS & CODES

Information Representations Number Systems Read up DLD for details! Base Conversion Negative Numbers Excess Representation Floating-Point Numbers Decimal codes: BCD, Excess-3, 2421, 84-2-1 Gray Code Alphanumeric Code Error Detection and Correction

INFORMATION REPRESENTATION (1/3)

Numbers are important to computers

Represent information precisely Can be processed Represent yes or no: use 0 and 1 Represent the 4 seasons: 0, 1, 2 and 3 Matriculation number: 8 alphanumeric characters (eg: U071234X)

Examples

Sometimes, other characters are used

INFORMATION REPRESENTATION (2/3)

Bit (Binary digit)

0 and 1 Represent false and true in logic Represent the low and high states in electronic devices Byte: 8 bits Nibble: 4 bits (seldom used) Word: Multiples of byte (eg: 1 byte, 2 bytes, 4 bytes, 8 bytes, etc.), depending on the architecture of the computer system

Other units

INFORMATION REPRESENTATION (3/3)

N bits can represent up to 2N values.

Examples:
2 bits represent up to 4 values (00, 01, 10, 11)
3 bits rep. up to 8 values (000, 001, 010, , 110, 111) 4 bits rep. up to 16 values (0000, 0001, 0010, ., 1111)

To represent M values, log2M bits are required.

Examples:
32 values requires 5 bits 64 values requires 6 bits 1024 values requires 10 bits 40 values how many bits? 100 values how many bits?

DECIMAL (BASE 10) SYSTEM (1/2)

A weighted-positional number system

Base or radix is 10 (the base or radix of a number system is the total number of symbols/digits allowed in the system) Symbols/digits = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 } Position is important, as the value of each symbol/digit is dependent on its type and its position in the number Example, the 9 in the two numbers below has different values:
(7594)10 = (7 103) + (5 102) + (9 101) + (4 100)

(912)10 = (9 102) + (1 101) + (2 100)

In general,

(anan-1 a0 . f1f2 fm)10 = (an x 10n) + (an-1x10n-1) + + (a0 x 100) + (f1 x 10-1) + (f2 x 10-2) + + (fm x 10m)

DECIMAL (BASE 10) SYSTEM (2/2)

Weighing factors (or weights) are in powers of 10: 103 102 101 100 . 10-1 10-2 10-3 To evaluate the decimal number 593.68, the digit in each position is multiplied by the corresponding weight: 5 102 + 9 101 + 3 100 + 6 10-1 + 8 10-2 = (593.68)10

OTHER NUMBER SYSTEMS (1/2) Binary (base 2)

Weights in powers of 2 Binary digits (bits): 0, 1 Weights in powers of 8 Octal digits: 0, 1, 2, 3, 4, 5, 6, 7. Weights in powers of 16 Hexadecimal digits: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F. Weights in powers of R

Octal (base 8)

Hexadecimal (base 16)

Base/radix R:

OTHER NUMBER SYSTEMS (2/2) In some programming languages/software, special

notations are used to represent numbers in certain bases

In programming language C
Prefix 0 for octal. Eg: 032 represents the octal number (32)8 Prefix 0x for hexadecimal. Eg: 0x32 represents the hexadecimal number (32)16

In PCSpim (a MIPS simulator)

Prefix 0x for hexadecimal. Eg: 0x100 represents the hexadecimal number (100)16

In Verilog, the following values are the same

8b11110000: an 8-bit binary value 11110000 8hF0: an 8-bit binary value represented in hexadecimal F0 8d240: an 8-bit binary value represented in decimal 240

BASE-R TO DECIMAL CONVERSION

Easy!

1101.1012 = 123 + 122 + 120 + 12-1 + 12-3 572.68 2A.816 = =

341.245 =

QUICK REVIEW QUESTIONS (1)

DLD page 42 Questions 2-1 to 2-4.

DECIMAL TO BINARY CONVERSION

Method 1

Sum-of-Weights Method Repeated Division-by-2 Method (for whole numbers) Repeated Multiplication-by-2 Method (for fractions)

Method 2

SUM-OF-WEIGHTS METHOD

Determine the set of binary weights whose sum is equal to the decimal number

(9)10 = 8 + 1 = 23 + 20 = (1001)2 (18)10 = 16 + 2 = 24 + 21 = (10010)2 (58)10 = 32 + 16 + 8 + 2 = 25 + 24 + 23 + 21 = (111010)2 (0.625)10 = 0.5 + 0.125 = 2-1 + 2-3 = (0.101)2

REPEATED DIVISION-BY-2
To convert a whole number to binary, use successive
division by 2 until the quotient is 0. The remainders form the answer, with the first remainder as the least significant bit (LSB) and the last as the most significant bit (MSB). (43)10 = (101011)2
2 2 2 2 2 2 43 21 10 5 2 1 0 rem 1 LSB rem 1 rem 0 rem 1 rem 0 rem 1 MSB

REPEATED MULTIPLICATION-BY-2
To convert decimal fractions to binary, repeated
multiplication by 2 is used, until the fractional product is 0 (or until the desired number of decimal places). The carried digits, or carries, produce the answer, with the first carry as the MSB, and the last as the LSB. (0.3125)10 = (.0101)2
0.3125 2=0.625 0.625 2=1.25 0.25 2=0.50 0.5 2=1.00 Carry 0 1 0 1 MSB

LSB

CONVERSION BETWEEN DECIMAL AND OTHER BASES

Base-R to decimal: multiply digits with their
corresponding weights.

Decimal to binary (base 2)

Whole numbers repeated division-by-2 Fractions: repeated multiplication-by-2

Decimal to base-R
Whole numbers: repeated division-by-R Fractions: repeated multiplication-by-R

QUICK REVIEW QUESTIONS (2)

DLD page 42 Questions 2-5 to 2-8.

CONVERSION BETWEEN BASES

In general, conversion between bases can be done via decimal:

Base-2 Base-3 Base-4 Base-R Base-2 Base-3 Decimal Base-4 . Base-R

Shortcuts for conversion between bases 2, 4, 8, 16 (see next slide)

BINARY TO OCTAL/HEXADECIMAL CONVERSION

Binary Octal: partition in groups of 3

(10 111 011 001 . 101 110)2 = (2731.56)8 = (101 1101 1001 . 1011 1000)2 = (5D9.B8)16 =

Octal Binary: reverse

Binary Hexadecimal: partition in groups of 4

Hexadecimal Binary: reverse

QUICK REVIEW QUESTIONS (3)

DLD page 42 Questions 2-9 to 2-10.

READING ASSIGNMENT

Binary arithmetic operations

Read up DLD section 2.6, pg 20 21.

PEEKING AHEAD (1/2)

Function simplification (eg: Quine-McCluskey) In computer-speak, units are in powers of 2

Memory addressing (see next slide)

PEEKING AHEAD (2/2)

Memory addressing

Assume 210 bytes in memory, and each word contains 4 bytes.

Addresses
binary decimal 00101101 0000000000 0 01010101 0000000001 1 10111100 0000000010 2 01111001 0000000011 3 11001100 0000000100 4 10000101 0000000101 5 11010111 0000000110 6 00011000 0000000111 7 01101101 0000001000 8 10011011 0000001001 9 11010101 0000001010 10 01000001 0000001011 11 . . . . 1111111111 1023

Memory

NEGATIVE NUMBERS

Unsigned numbers: only non-negative values. Signed numbers: include all values (positive and negative) There are 3 common representations for signed binary numbers:

Sign-and-Magnitude 1s Complement 2s Complement

SIGN-AND-MAGNITUDE (1/3)

The sign is represented by a sign bit

0 for + 1 for -

Eg: a 1-bit sign and 7-bit magnitude format.

sign

magnitude

00110100 +1101002 = ? 10010011 -100112 = ?

SIGN-AND-MAGNITUDE (2/3)

Largest value: 01111111 = +12710 Smallest value: 11111111 = -12710 Zeros: 00000000 = +010 10000000 = -010 Range: -12710 to +12710 Question:

For an n-bit sign-and-magnitude representation, what is the range of values that can be represented?

SIGN-AND-MAGNITUDE (3/3)

To negate a number, just invert the sign bit. Examples:

How to negate 00100001sm (decimal 33)? Answer: 10100001sm (decimal -33) How to negate 10000101sm (decimal -5)? Answer: 00000101sm (decimal +5)

1s COMPLEMENT (1/3)
Given a number x which can be expressed as an n-bit binary number, its negated value can be obtained in 1scomplement representation using: -x = 2n x 1 Example: With an 8-bit number 00001100 (or 1210), its negated value expressed in 1s-complement is: -000011002 = 28 12 1 (calculation in decimal) = 243 = 111100111s (This means that -1210 is written as 11110011 in 1scomplement representation.)

1s COMPLEMENT (2/3)

Essential technique to negate a value: invert all the bits. Largest value: 01111111 = +12710 Smallest value: 10000000 = -12710 Zeros: 00000000 = +010 11111111 = -010 Range: -12710 to +12710 The most significant (left-most) bit still represents the sign: 0 for positive; 1 for negative.

1s COMPLEMENT (3/3)

Examples (assuming 8-bit numbers): (14)10 = (00001110)2 = (00001110)1s -(14)10 = -(00001110)2 = (11110001)1s -(80)10 = -( ? )2 = ( ? )1s

2s COMPLEMENT (1/3)
Given a number x which can be expressed as an n-bit binary number, its negated value can be obtained in 2scomplement representation using: -x = 2n x Example: With an 8-bit number 00001100 (or 1210), its negated value expressed in 2s-complement is: -000011002 = 28 12 (calculation in decimal) = 244 = 111101002s (This means that -1210 is written as 11110100 in 2scomplement representation.)

2s COMPLEMENT (2/3)

Essential technique to negate a value: invert all the bits, then add 1. Largest value: 01111111 = +12710 Smallest value: 10000000 = -12810 Zero: 00000000 = +010 Range: -12810 to +12710 The most significant (left-most) bit still represents the sign: 0 for positive; 1 for negative.

2s COMPLEMENT (3/3)

Examples (assuming 8-bit numbers): (14)10 = (00001110)2 = (00001110)2s -(14)10 = -(00001110)2 = (11110010)2s -(80)10 = -( ? )2 = ( ? )2s

Compare with slide 30.

1s complement: (14)10 = (00001110)2 = (00001110)1s -(14)10 = -(00001110)2 = (11110001)1s

READING ASSIGNMENT

Download from the course website and read the Supplement Notes on Lecture 2: Number Systems. Work out the exercises in there and discuss them in the IVLE forum if you have doubts.

COMPARISONS
4-bit system
Positive values
Value +7 +6 +5 +4 +3 +2 +1 +0 Sign-andMagnitude 0111 0110 0101 0100 0011 0010 0001 0000 1s Comp. 0111 0110 0101 0100 0011 0010 0001 0000

Important!

Negative values
2s Comp. 0111 0110 0101 0100 0011 0010 0001 0000
Value -0 -1 -2 -3 -4 -5 -6 -7 -8 Sign-andMagnitude 1000 1001 1010 1011 1100 1101 1110 1111 1s Comp. 1111 1110 1101 1100 1011 1010 1001 1000 2s Comp. 1111 1110 1101 1100 1011 1010 1001 1000

COMPLEMENT ON FRACTIONS

We can extend the idea of complement on fractions. Examples:

Negate 0101.01 in 1s-complement Answer: 1010.10 Negate 111000.101 in 1s-complement Answer: 000111.010 Negate 0101.01 in 2s-complement Answer: 1010.11

2s COMPLEMENT ADDITION/SUBTRACTION (1/3)

Algorithm for addition, A + B:

1. 2. 3. Perform binary addition on the two numbers. Ignore the carry out of the MSB. Check for overflow. Overflow occurs if the carry in and carry out of the MSB are different, or if result is opposite sign of A and B.

Algorithm for subtraction, A B: A B = A + (-B)

1. 2. Take 2s-complement of B. Add the 2s-complement of B to A.

OVERFLOW

Signed numbers are of a fixed range. If the result of addition/subtraction goes beyond this range, an overflow occurs. Overflow can be easily detected:

positive add positive negative negative add negative positive Range of value: -810 to 710 01012s + 01102s = 10112s 510 + 610 = -510 ?! (overflow!) 10012s + 11012s = 101102s (discard end-carry) = 01102s -710 + -310 = 610 ?! (overflow!)

Example: 4-bit 2s-complement system

2s COMPLEMENT ADDITION/SUBTRACTION (2/3)

Examples: 4-bit system

+3 + +4 ---+7 ---+6 + -3 ---+3 ---0011 + 0100 ------0111 ------0110 + 1101 ------10011 -------2 + -6 ----8 ---+4 + -7 ----3 ---1110 + 1010 ------11000 ------0100 + 1001 ------1101 -------

Which of the above is/are overflow(s)?

2s COMPLEMENT ADDITION/SUBTRACTION (3/3)

Examples: 4-bit system

-3 + -6 ----9 ---1101 + 1010 ------10111 ------+5 + +6 ---+11 ---0101 + 0110 ------1011 -------

Which of the above is/are overflow(s)?

1s COMPLEMENT ADDITION/SUBTRACTION (1/2)

Algorithm for addition, A + B:

1. 2. 3. Perform binary addition on the two numbers. If there is a carry out of the MSB, add 1 to the result. Check for overflow. Overflow occurs if result is opposite sign of A and B.

Algorithm for subtraction, A B: A B = A + (-B)

1. 2. Take 1s-complement of B. Add the 1s-complement of B to A.

1s COMPLEMENT ADDITION/SUBTRACTION (2/2)

Examples: 4-bit system

+3 + +4 ---+7 ----2 + -5 ----7 ---0011 + 0100 ------0111 ------1101 + 1010 -----10111 + 1 -----1000

Any overflow?
+5 + -5 ----0 ----3 + -7 ----10 ---0101 + 1010 ------1111 ------1100 + 1000 ------10100 + 1 ------0101

QUICK REVIEW QUESTIONS (4)

DLD pages 42 - 43 Questions 2-13 to 2-18.

EXCESS REPRESENTATION (1/2) Besides sign-and-magnitude and

complement schemes, the excess representation is another scheme. It allows the range of values to be distributed evenly between the positive and negative values, by a simple translation (addition/subtraction). Example: Excess-4 representation on 3-bit numbers. See table on the right.

Excess-4 Representation 000 001 010 011 100 101 110

Value -4 -3 -2 -1 0 1 2 3

Questions: What if we use Excess-2 on 3-bit numbers? Excess-7?

111

EXCESS REPRESENTATION (2/2) Example: For 4-bit numbers, we may use excess-7 or

excess-8. Excess-8 is shown below. Fill in the values.

Excess-8 Representation 0000 0001 0010 0011 0100 0101 0110 0111

Value -8

Excess-8 Representation 1000 1001 1010 1011 1100 1101 1110 1111

Value

FIXED POINT NUMBERS (1/2)

In fixed point representation, the binary point is assumed to be at a fixed location.

For example, if the binary point is at the end of an 8-bit representation as shown below, it can represent integers from -128 to +127.

binary point

FIXED POINT NUMBERS (2/2)

In general, the binary point may be assumed to be at any pre-fixed location.

Example: Two fractional bits are assumed as shown below.

fraction part assumed binary point

integer part

If 2s complement is used, we can represent values like:

011010.112s = 26.7510 111110.112s = -000001.012 = -1.2510

FLOATING POINT NUMBERS (1/4)

Fixed point numbers have limited range. Floating point numbers allow us to represent very large or very small numbers. Examples:
0.23 1023 (very large positive number) 0.5 10-37 (very small positive number) -0.2397 10-18 (very small negative number)

FLOATING POINT NUMBERS (2/4)

3 parts: sign, mantissa and exponent The base (radix) is assumed to be 2. Sign bit: 0 for positive, 1 for negative.
sign mantissa exponent

Mantissa is usually in normalised form (the integer part is zero and the fraction part must not begin with zero)
0.01101 24 normalised 101011.0110 2-4 normalised

Trade-off:

More bits in mantissa better precision More bits in exponent larger range of values

FLOATING POINT NUMBERS (3/4)

Exponent is usually expressed in complement or excess format. Example: Express -6.510 in base-2 normalised form -6.510 = -110.12 = -0.11012 23 Assuming that the floating-point representation contains 1-bit, 5-bit normalised mantissa, and 4-bit exponent. The above example will be stored as if the exponent is in 1s or 2s complement. 1 11010 0011

FLOATING POINT NUMBERS (4/4)

Example: Express 0.187510 in base-2 normalised form 0.187510 = 0.00112 = 0.11 2-2 Assume this floating-point representation:1-bit sign, 5-bit normalised mantissa, and 4-bit exponent. The above example will be represented as 0 0 0 11000 11000 11000 1101 1110 0110
If exponent is in 1s complement. If exponent is in 2s complement. If exponent is in excess-8.

QUICK REVIEW QUESTIONS (5)

DLD page 43 Questions 2-19 to 2-20.

READING ASSIGNMENT

Arithmetic operations on floating point numbers

DLD page 31 DLD pages 32 - 33 IEEE standard 754 floating point numbers: http://steve.hollasch.net/cgindex/coding/ieeefloat.html

IEEE floating point representation

DECIMAL CODES

Decimal numbers are favoured by humans. Binary numbers are natural to computers. Hence, conversion is required. If little calculation is required, we can use some coding schemes to store decimal numbers, for data transmission purposes. Examples: BCD (or 8421), Excess-3, 84-2-1, 2421, etc. Each decimal digit is represented as a 4-bit code. The number of digits in a code is also called the length of the code.

BINARY CODE DECIMAL (BCD) (1/2) Decimal

Some codes are unused, like 1010BCD, 1011BCD, 1111BCD. These codes are considered as errors. Easy to convert, but arithmetic operations are more complicated. Suitable for interfaces such as keypad inputs.
digit 0 1 2 3 4 5 6 7 8 9

BCD

0000 0001 0010 0011 0100 0101 0110 0111 1000 1001

BINARY CODE DECIMAL (BCD) (2/2)

Examples of conversion between BCD values and decimal values:

(234)10 = (0010 0011 0100)BCD (7093)10 = (0111 0000 1001 0011)BCD (1000 0110)BCD = (86)10 (1001 0100 0111 0010)BCD = (9472)10 Example: (234)10 = (11101010)2

Note that BCD is not equivalent to binary

OTHER DECIMAL CODES

Decimal Digit 0 1 2 3 4 5 6 7 8 9 BCD 8421 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 Excess-3 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 84-2-1 0000 0111 0110 0101 0100 1011 1010 1001 1000 1111 2*421 0000 0001 0010 0011 0100 1011 1100 1101 1110 1111 Biquinary 5043210 0100001 0100010 0100100 0101000 0110000 1000001 1000010 1000100 1001000 1010000

Self-complementing code: codes for complementary digits are also complementary to each other. Error-detecting code: biquinary code (bi=two, quinary=five).

SELF-COMPLEMENTING The codes representing the pair of complementary digits CODES

are also complementary to each other. Example: Excess-3 code 0: 0011

1: 2: 3: 4: 5: 6: 7: 8: 9: 0100 0101 0110 0111 1000 1001 1010 1011 1100

Question: What are the other self-complementing codes?

GRAY CODE (1/3)

Unweighted (not an arithmetic code) Only a single bit change from one code value to the next. Not restricted to decimal digits: n bits 2n values. Good for error detection. Example: 4-bit standard Gray code
Decimal 0 1 2 3 4 5 6 7 Binary 0000 0001 0010 0011 0100 0101 0110 0111 Gray Code 0000 0001 0011 0010 0110 0111 0101 0100 Decimal 8 9 10 11 12 13 14 15 Binary 1000 1001 1010 1011 1100 1101 1110 1111 Gray code 1100 1101 1111 1110 1010 1011 1001 1000

GRAY CODE (2/3)

Generating a 4-bit standard Gray code sequence.

0 0 0 0 0 0 0 0

0 0 0 0 0 1 0 1 0 1 0 1

0 0 1 0 1 0 1 1 0 0

0 1 1 0 0 1 1 0

1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1

1 1 1 1 0 0 0 0

0 0 1 1 1 1 0 0

0 1 1 0 0 1 1 0

Questions: How to generate 5-bit standard Gray code sequence? 6-bit standard Gray code sequence?

GRAY CODE (3/3)

1 1
1

0 0

0 1
1 1 0

0
0

sensors

0 0 0 0 0 0 0 1 0 1 1 0 0 0 1 0 0 1

0 0 1 1 1 0 0 0 0 0 0 0 0 1 0 0 1 0

0 0 1 1 1 1

Binary coded: 111 110 000

1 11 1 1 0 0 0 0 0 1 0 0 0 0 1 0 0

0 10 0 1 0 1 1 1 1 1 0

mis-aligned sensors

0
1

1
1

0
0
0 00 0 1 1 1 1 0 1 1 0 1 0 0 1 1 0

1 1

1 1 1 1 1

1 11 0 1 0 1 1 1 1 1 0

mis-aligned sensors

0 11 01 0 1 1 1 1 0 1 1 1 0 0 1 0

Gray coded: 111 101

0 0 1 0 1 1 1 1 0 0 1 0 0 0 1 1 0 1

READING ASSIGNMENT

Conversion between standard Gray code and binary

DLD page 36.

QUICK REVIEW QUESTIONS (6)

DLD pages 43 - 44 Questions 2-22 to 2-25.

ALPHANUMERIC CODES (1/3)

Computers also handle textual data. Character set frequently used:

alphabets: digits: A Z, a z 0 9 NULL, BELL, CR, etc.

special symbols: $, ., @, *, etc. non-printable:

Examples

ASCII (8 bits), Unicode

ALPHANUMERIC CODES (2/3)

ASCII

American Standard Code for Information Interchange 7 bits, plus a parity bit for error detection Odd or even parity

Character 0 1 ... 9 : A B ... Z [ \

ASCII Code 0110000 0110001 ... 0111001 0111010 1000001 1000010 ... 1011010 1011011 1011100

ALPHANUMERIC CODES (3/3)

ASCII table
LSBs 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 000 NUL SOH STX ETX EOT ENQ ACK BEL BS HT LF VT FF CR O SI 001 DLE DC1 DC2 DC3 DC4 NAK SYN ETB CAN EM SUB ESC FS GS RS US 010 SP ! # $ % & ( ) * + , . / MSBs 011 100 0 @ 1 A 2 B 3 C 4 D 5 E 6 F 7 G 8 H 9 I : J ; K < L = M > N ? O 101 P Q R S T U V W X Y Z [ \ ] ^ _

A: 1000001
110 ` a b c d e f g h i j k l m n o 111 p q r s t u v w x y z { | } ~ DEL

ERROR DETECTION (1/4)

Errors can occur during data transmission. They should be detected, so that re-transmission can be requested. With binary numbers, usually single-bit errors occur.

Example: 0010 erroneously transmitted as 0011 or 0000 or 0110 or 1010.

Decimal digit 0 1 2 3 4 5 6 7 8 9 Biquinary 5043210 0100001 0100010 0100100 0101000 0110000 1000001 1000010 1000100 1001000 1010000

Biquinary code has length 7; it uses 3 additional bits for error-detection.

ERROR DETECTION (2/4)

Parity bit

Even parity: additional bit added to make total number of 1s even. Odd parity: additional bit added to make total number of 1s odd.

Example of odd parity on ASCII values.

Character 0 1 ... 9 : A B ... Z [ \

ASCII Code 0110000 1 0110001 0 ... 0111001 1 0111010 1 1000001 1 1000010 1 ... 1011010 1 1011011 0 1011100 1

Parity bits

ERROR DETECTION (3/4)

Parity bit can detect odd number of errors but not even number of errors.

Example: Assume odd parity,

10011 10001 (detected) 10011 10101 (not detected)
0110 1 0001 0 1011 0 1111 1 1001 1 0101 0

Parity bits can also be applied to a block of data.

Column-wise parity

Row-wise parity

ERROR DETECTION (4/4)

Sometimes, it is not enough to do error detection. We may want to correct the errors. Error correction is expensive. In practice, we may use only single-bit error correction. Popular technique: Hamming code

ERROR CORRECTION (1/7)

Given this 3-bit code C1

{ 000, 110, 011, 101 }

With 4 code words, we actually need only 2 bits.

We call this k, the number of original message bits.

To add error detection/correction ability, we use more bits than necessary.

In this case, the length of each codeword is 3

We define code efficiency (or rate) by

k / length of codeword

Hence, efficiency of C1 is 2/3.

ERROR CORRECTION (2/7)

Given this 3-bit code C1

{ 000, 110, 011, 101 }

Can C1 detect a single bit error? Can C1 correct a single bit error?

Sometimes, we use 1 error for single bit error, 2 errors for 2 bits error, etc.

ERROR CORRECTION (3/7)

The distance d between any two code words in a code is the sum of the number of differences between the codewords.

Example: d(000, 110) = 2; d(0110,1011) = 3.

The Hamming distance of a code is the minimum distance between any two code words in the code.

Example: The Hamming distance of C1 is 2.

A code with Hamming distance of 2 can detect 1 error.

ERROR CORRECTION (4/7)

Given this 6-bit code C2

{ 000000, 111000, 001110, 110011 }

What is its efficiency? What is its Hamming distance? Can it correct 1 error?

Can it correct 2 errors?

ERROR CORRECTION (5/7)Self-study

Hamming code: a popular error-correction code Procedure

Parity bits are at positions that are powers of 2 (i.e. 1, 2, 4, 8, 16, ) All other positions are data bits Each parity bit checks some of the data bits
Position 1: Check 1 bit, skip 1 bit (1, 3, 5, 7, 9, 11, ) Position 2: Check 2 bits, skip 2 bits (2, 3, 6, 7, 10, 11, ) Position 4: Check 4 bits, skip 4 bits (4-7, 12-15, 20-23, ) Position 8: Check 8 bits, skip 8 bits (8-15, 24-31, 40-47, )

Set the parity bit accordingly so that total number of 1s in the positions it checks is even.

ERROR CORRECTION (6/7)Self-study

Example: Data 10011010 Insert positions for parity bits: __1_001_1010

Position 1: ? _ 1 _ 0 0 1 _ 1 0 1 0 so ? must be 0 Position 2: 0 ? 1 _ 0 0 1 _ 1 0 1 0 so ? must be 1 Position 4: 0 1 1 ? 0 0 1 _ 1 0 1 0 so ? must be 1 Position 8: 0 1 1 1 0 0 1 ? 1 0 1 0 so ? must be 0

Answer: 0 1 1 1 0 0 1 0 1 0 1 0

ERROR CORRECTION (7/7)Self-study

Suppose 1 error occurred and the received data is: 011100101110 How to determine which bit is in error? Check which parity bits are in error.

Answer: parity bits 2 and 8. Answer: 2 + 8 = 10. Hence data bit 10 is in error.

Add the positions of these erroneous parity bits

Corrected data: 0 1 1 1 0 0 1 0 1 0 1 0

END