Floating point representation, IEEE standard

IEEE Standard 754 floating point is the most common representation today for real numbers on computers, including Intel-based PC’s, Macintoshes, and most Unix platforms.

IEEE floating point numbers have three basic components: the sign, the exponent, and the mantissa. The mantissa is composed of the fraction and an implicit leading digit (explained below). The exponent base(2) is implicit and need not be stored.

The following figure shows the layout for single (32-bit) and double (64-bit) precision floating-point values. The number of bits for each field are shown (bit ranges are in square brackets):

Sign

Leave a Reply

Your email address will not be published. Required fields are marked *