Legacy Document

fdim

You can use the fdim function to determine the positive difference between two real numbers.

double_t fdim (double_t x, double_t y);

x

Any floating-point number.

y

Any floating-point number.

DESCRIPTION

The fdim function returns the positive difference between its two arguments.

$fdim(x,y)\; =\; x+-y$ if x > y
$fdim(x,y)\; =\; +0$ if x y

EXCEPTIONS

When x and y are finite and nonzero and x > y, either the result of $fdim(x,y)$ is exact or it raises one of the following exceptions:

• inexact (if the result of x - y must be rounded)
• overflow (if the result of x - y is outside the range of the data type)
• underflow (if the result of x - y is inexact and must be represented as a denormalized number or 0)

SPECIAL CASES

Table 10-1 shows the results when one of the arguments to the fdim function is a zero, a NaN, or an Infinity. In this table, x and y are finite, nonzero floating-point numbers.

Special cases for the fdim function

Operation	Result	Exceptions raised
$fdim(+0,y)$	+0	None
$fdim(x,+0)$	x	None
$fdim(-0,y)$	+0	None
$fdim(x,-0)$	x	None
$fdim(NaN,y)$	NaN[19]	None[20]
$fdim(x,NaN)$	NaN	None
$fdim(+$ ,y)	+	None
$fdim(x,+$ )	+0	None
$fdim(-$ ,y)	+0	None
$fdim(x,-$ )	+	None

EXAMPLES

z = fdim(+INFINITY, 300);  /* z = +   - 300 = +INFINITY because 
                              +   > 300 */
z = fdim(300, +INFINITY);  /* z = +0 because 300  +   */

---

[19] If both arguments are NaN, the first NaN is returned.

---

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