next up previous
Next: Single Precision Up: Introductory Material Previous: Introduction to Fortran

Numerical Representation

A bit is a binary number in which 0 is typically used to represent off (or false or positive) and 1 is typically used to represent on (or true or negative). Some sample binary (i.e., base 2) numbers are 0, 1, 10, 11, 100, 101, 110, 111, which represent 0, 1, 2, 3, 4, 5, 6, 7 in base 10.

A byte of data typically consists of 8 bits. A word of data typically refers to 4 bytes (i.e., 32 bits). However, as technology continues to advance, more and more machines are constructed with the basic word having 8 bytes (i.e., 64 bits) of data. A single precision number is typically represented by a word and a double precision number is typically represented by two words.

For the sake of consistency and portability between machines, IEEE arithmetic defines the following.





Russ Miller
Thu Sep 14 13:56:19 EDT 1995