The enumeration describes the various methods that an implementation can choose for representing a denormalized floating-point value — one too small to represent as a normalized value:
The enumeration returns:
denorm_indeterminate if the presence or absence of denormalized forms cannot be determined at translation time.
denorm_absent if denormalized forms are absent.
denorm_present if denormalized forms are present.
See numeric_limits::has_denorm for an example in which the values of this enumeration may be accessed.