Defines the template class numeric_limits and two enumerations concerning floating-point representations and rounding.
Explicit specializations of the numeric_limits class describe many properties of the fundamental types, including the character, integer, and floating-point types and bool that are implementation defined rather than fixed by the rules of the C++ language. Properties described in <limits> include accuracy, minimum and maximum sized representations, rounding, and signaling type errors.
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 describes the various methods that an implementation can choose for rounding a floating-point value to an integer value.