Implementation of Floating-Point Arithmetic Circuits

The representation of real numbers in binary systems caused floating-point numerical representations to be established. Besides precisely representing a numerical value, calculating the result of arithmetic operations between two floating point values, such as adding, multiplying, dividing and subtracting, without error and with immediate response time is a crucial point. Currently the IEEE 754 standard is the representation standard used in computer systems, and this work presents a simplified implementation of floating point representation, in which the number of bits used is parametrizable. The implemented circuits make use only of combinational logic, in order to optimize the response time of the operations, and the amount of logic elements used, according to the desired application.

Autorentext

Victor Augusto Machado Santos, ingénieur électricien. Programme de premier cycle en génie électrique, robotique et automatisation industrielle, université fédérale de Juiz de Fora, Brésil.