ABSTRACT
Chapter 4 deals with numbers representation in floating-point format for expressing the real numbers. Non-normalized numbers are introduced. The main topic is a worldwide accepted and applied IEEE 754 standard as a hardware independent. Among others, the following issues are highlighted: single and double precision, special values and exceptions. The changes imposed by IEEE 854 update and related to the need of support shorter than single and longer than double precision new formats are announced. Additionally, some key-value features of a FPU floating point unit as specialized arithmetic coprocessor were pointed out. The universal method of conversion to another radix is provided.
