The IEEE Standard for Floating-Point Arithmetic (IEEE 754) is a technical standard for floating-point computation which was established in 1985 by the Institute of Electrical and Electronics Engineers (IEEE). The standard addressed many problems found in the diverse floating point implementations that made them difficult to use reliably and reduced their portability. IEEE Standard 754 floating point is the most common representation today for real numbers on computers, including Intel-based PC’s, Macs, and most Unix platforms.
There are several ways to represent floating point number but IEEE 754 is the most efficient in most cases. IEEE 754 has 3 basic components:
- The Sign of Mantissa –
This is as simple as the name. 0 represents a positive number while 1 represents a negative number.
- The Biased exponent –
The exponent field needs to represent both positive and negative exponents. A bias is added to the actual exponent in order to get the stored exponent.
- The Normalised Mantissa –
The mantissa is part of a number in scientific notation or a floating-point number, consisting of its significant digits. Here we have only 2 digits, i.e. O and 1. So a normalised mantissa is one with only one 1 to the left of the decimal.
IEEE 754 numbers are divided into two based on the above three components: single precision and double precision.
|TYPES||SIGN||BIASED EXPONENT||NORMALISED MANTISA||BIAS|
|Single precision||1(31st bit)||8(30-23)||23(22-0)||127|
|Double precision||1(63rd bit)||11(62-52)||52(51-0)||1023|
85.125 85 = 1010101 0.125 = 001 85.125 = 1010101.001 =1.010101001 x 2^6 sign = 0 1. Single precision: biased exponent 127+6=133 133 = 10000101 Normalised mantisa = 010101001 we will add 0's to complete the 23 bits The IEEE 754 Single precision is: = 0 10000101 01010100100000000000000 This can be written in hexadecimal form 42AA4000 2. Double precision: biased exponent 1023+6=1029 1029 = 10000000101 Normalised mantisa = 010101001 we will add 0's to complete the 52 bits The IEEE 754 Double precision is: = 0 10000000101 0101010010000000000000000000000000000000000000000000 This can be written in hexadecimal form 4055480000000000
Special Values: IEEE has reserved some values that can ambiguity.
- Zero –
Zero is a special value denoted with an exponent and mantissa of 0. -0 and +0 are distinct values, though they both are equal.
- Denormalised –
If the exponent is all zeros, but the mantissa is not then the value is a denormalized number. This means this number does not have an assumed leading one before the binary point.
- Infinity –
The values +infinity and -infinity are denoted with an exponent of all ones and a mantissa of all zeros. The sign bit distinguishes between negative infinity and positive infinity. Operations with infinite values are well defined in IEEE.
- Not A Number (NAN) –
The value NAN is used to represent a value that is an error. This is represented when exponent field is all ones with a zero sign bit or a mantissa that it not 1 followed by zeros. This is a special value that might be used to denote a variable that doesn’t yet hold a value.
|255||not 0||Not a number (NAN)|
Similar for Double precision (just replacing 255 by 2049), Ranges of Floating point numbers:
|Single Precision||± 2-149 to (1 – 2-23)×2-126||± 2-126 to (2 – 2-23)×2127||± approximately 10-44.85 to approximately 1038.53|
|Double Precision||± 2-1074 to (1 – 2-52)×2-1022||± 2-1022 to (2 – 2-52)×21023||± approximately 10-323.3 to approximately 10308.3|
The range of positive floating point numbers can be split into normalized numbers, and denormalized numbers which use only a portion of the fractions’s precision. Since every floating-point number has a corresponding, negated value, the ranges above are symmetric around zero.
There are five distinct numerical ranges that single-precision floating-point numbers are not able to represent with the scheme presented so far:
- Negative numbers less than – (2 – 2-23) × 2127 (negative overflow)
- Negative numbers greater than – 2-149 (negative underflow)
- Positive numbers less than 2-149 (positive underflow)
- Positive numbers greater than (2 – 2-23) × 2127 (positive overflow)
Overflow generally means that values have grown too large to be represented. Underflow is a less serious problem because is just denotes a loss of precision, which is guaranteed to be closely approximated by zero.
Table of the total effective range of finite IEEE floating-point numbers is shown below:
|Single||± (2 – 2-23) × 2127||approximately ± 1038.53|
|Double||± (2 – 2-52) × 21023||approximately ± 10308.25|
Special Operations –
|n ÷ ±Infinity||0|
|±Infinity × ±Infinity||±Infinity|
|±nonZero ÷ ±0||±Infinity|
|±finite × ±Infinity||±Infinity|
|Infinity + Infinity
Infinity – -Infinity
|-Infinity – Infinity
-Infinity + – Infinity
|±0 ÷ ±0||NaN|
|±Infinity ÷ ±Infinity||NaN|
|±Infinity × 0||NaN|
|NaN == NaN||False|
- Program for conversion of 32 Bits Single Precision IEEE 754 Floating Point Representation
- Introduction of Floating Point Representation
- Floating Point Representation - Basics
- Canonical and Standard Form
- 8086 program to determine squares of numbers in an array of n numbers
- 8086 program to determine cubes of numbers in an array of n numbers
- 8085 program to count total odd numbers in series of 10 numbers
- 8085 program to count total even numbers in series of 10 numbers
- 8085 program to add two 16 bit numbers
- 8085 program to add two 8 bit numbers
- 8086 program to add two 8 bit BCD numbers
- 8085 program to add 2-BCD numbers
- 8085 program to sum of two 8 bit numbers without carry
- Maximum possible Bitwise OR of the two numbers from the range [L, R]
- Count numbers have all 1s together in binary representation
If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to firstname.lastname@example.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please Improve this article if you find anything incorrect by clicking on the "Improve Article" button below.