ABSTRACT
Most of the modern computers follow Institute of Electrical and Electronics Engineers (IEEE) standards for storing and using floating point numbers. Denormal numbers or denormalized numbers are those that are less than the smallest normalized floating point number available in the system. They actually fill the gap between zero and the smallest floating point number. As these numbers are smaller than the smallest floating point number, these numbers are also called subnormal numbers. There are two zeros in this representation: positive zero (+0) and negative zero (-0). For positive zero, all bits are switched off. For negative zero, sign bit is on and all other bits are zero. IEEE exception handling in Fortran is performed through three modules: ieee_exceptions, ieee_arithmetic and ieee_features. The module ieee_exceptions is used for exceptions, the module ieee_arithmetic is used for IEEE arithmetic and the module ieee_features is used to control the features supported by the compiler.
