COMPUTER ORGANIZATION

AND ASSEMBLY LANGUAGE

Lecture 7 & 8
Floating Point Representation
Binary Coded Decimal
Course Instructor: Engr. Aisha Danish
Real Numbers
 Numbers with fractions
 Could be done in pure binary
 1001.1010 = 24 + 20 +2-1 + 2-3 =9.625
 Where is the binary point?
 Fixed?
 Very limited
 Moving?
 How do you show where it is?
 Floating point
 In computing, floating point describes a method of
representing an approximation of a real number in a way that
can support a wide range of values
 The numbers are, in general, represented approximately to a
fixed number of significant digits (the significand) and scaled
using an exponent
 The term floating point refers to the fact that a number's 
radix point (decimal point, or, more commonly in computers,
binary point) can "float"; that is, it can be placed anywhere
relative to the significant digits of the number
 Floating point can be thought of as a computer realization of 
scientific notation
 How floating-point numbers work?

The idea is to compose a number of two main parts:

 A significand that contains the number’s digits.

 Negative significands represent negative numbers.

 An exponent that says where the decimal (or binary) point is

placed relative to the beginning of the significand.

 Negative exponents represent numbers that are very small (i.e.

close to zero).
Such a format satisfies all the requirements:
 It can represent numbers at widely different magnitudes (limited

by the length of the exponent)

 It provides the same relative accuracy at all magnitudes (limited

by the length of the significand)

 Floating Point

 A typical 32-bit floating-point format

 The leftmost bit stores the sign of the number
 The exponent value is stored in the next 8 bits
 The representation used is known as a biased representation
 A fixed value, called the bias, is subtracted from the field to get the true exponent
value
 Typically, the bias equals 2^(k-1)-1where k is the number of bits in the binary
exponent. In this case, the 8-bit field yields the numbers 0 through 255
 With a bias of 127(2^7-1), the true exponent values are in the range -127 to +128
 In this example, the base is assumed to be 2
Floating Point Examples
Signs for Floating Point
 Mantissa/Significand is stored in 2s compliment
 Exponent is in excess or biased notation
 e.g. Excess (bias) 128 means
 8 bit exponent field
 Pure value range 0-255
 Subtract 127 to get correct value
 Range -127 to +128
 Normalization
 A normalized number is one in which the most significant digit
of the significand is nonzero
 For base 2 representation, a normalized number is therefore one
in which the most significant bit of the significand is one.
 The typical convention is that there is one bit to the left of the
radix point
 i.e. exponent is adjusted so that leading bit (MSB) of mantissa is
1
 Since it is always 1 there is no need to store it
 (Scientific notation where numbers are normalized to give a
single digit before the decimal point e.g. 3.123 x 10 3)
IEEE 754
 Developed to facilitate the portability of programs
from one processor to another and to encourage the
development of sophisticated, numerically oriented
programs
 The standard has been widely adopted and is used
on virtually all contemporary processors and
arithmetic coprocessors
 32 and 64 bit standards
 8 and 11 bit exponent respectively
IEEE 754 Formats
Required Reading
 Stallings Chapter 9
 IEEE 754 on IEEE Web site
 The 8421 BCD Code

 BCD stands for Binary-Coded Decimal.

 A BCD number is a four-bit binary group that

represents one of the ten decimal digits 0 through 9.
Example:
Decimal number 4926 4 9 2 6

8421 BCD coded number 0100 1001 0010 0110

 QUIZ
Convert the BCD coded number 1000 0111 0001
into decimal.

BCD Coded Number 1000 0111 0001

Decimal Number 8 7 1
 QUIZ

Convert the decimal number 350 to its BCD

equivalent.

Decimal Number 3 5 0

BCD Coded Number 0011 0101 0000

Binary Coded Decimal (BCD)
 Would it be easy for you if you can replace a
decimal number with an individual binary code?
 Such as 00011001 = 1910
 The 8421 code is a type of BCD to do that.
 BCD code provides an excellent interface to binary
systems:
 Keypad inputs
 Digital readouts
 Binary Coded Decimal
Decimal
Digit 0 1 2 3 4 5 6 7 8 9

BCD 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001

Note: 1010, 1011, 1100, 1101, 1110, and 1111 are INVALID CODE!

ex1: dec-to-BCD ex2: BCD-to-dec

(a) 35 (a) 10000110
(b) 98 (b) 001101010001
(c) 170 (c) 1001010001110000
(d) 2469
 BCD Addition
 BCD is a numerical code and can be used in arithmetic
operations. Here is how to add two BCD numbers:
 Add the two BCD numbers, using the rules for basic binary
addition.
 If a 4-bit sum is equal to or less than 9, it is a valid BCD
number.
 If a 4-bit sum > 9, or if a carry out of the 4-bit group is
generated it is an invalid result. Add 6 (0110) to a 4-bit sum in
order to skip the six the invalid states and return the code to
8421. If a carry results when 6 is added, simply add the carry
to the next 4-bit group.
BCD Addition
 Add the following numbers
(a) 0011+0100
(b) 00100011 + 00010101
(c) 10000110 + 00010011
(d) 010001010000 + 010000010111
(e) 1001 + 0100
(f) 1001 + 1001
(g) 00010110 + 00010101
(h) 01100111 + 01010011