Legacy Document

rint

You can use the round-to-integer operation (rint function) to round a number to the nearest integer in the current rounding direction.

double_t rint(double_t x);

x

Any floating-point number.

DESCRIPTION

The rint function rounds its argument to an integer in the current rounding direction. The available rounding directions are upward, downward, to nearest (default), and toward zero. With the default rounding direction, if the argument is equally near two integers, the even integer is used as the result.

In each floating-point data type, all values of sufficiently great magnitude are integers. For example, in single format, all numbers whose magnitudes are at least $223$ are integers. This means that + and $-$ are already integers and return exact results.

The rint function performs the round-to-integer arithmetic operation described in the IEEE standard. For other C functions that perform rounding to integer, see Chapter 9, "Conversion Functions."

EXCEPTIONS

When x is finite and nonzero, either the result of $rint(x)$ is exact or it raises the following exception

• inexact (if x is not an integer)

SPECIAL CASES

Table 6-9 shows the results when the argument to the round-to-integer operation is a zero, a NaN, or an Infinity.

Table 6-9 Special cases for floating-point round-to-integer

Operation	Result	Exceptions raised
$rint(+0)$	+0	None
$rint(-0)$	$-0$	None
$rint(NaN)$	NaN	None[11]
$rint(+$ )	+	None
$rint(-$ )	$-$	None

EXAMPLES

Table 6-10 shows some example results of rint, given different rounding directions.

Examples of rint

Example	Current rounding direction
	To nearest	Toward 0	Downward	Upward
rint(1.5)	2	1	1	2
rint(2.5)	2	2	2	3
rint(-2.2)	-2	-2	-3	-2

---

[11] If the NaN is a signaling NaN, the invalid exception is raised.