Click or drag to resize
mathfrexp Method (Double, Int32)
Decomposes the given floating-point number into a normalized fraction and an integral power of two.

Namespace:  C
Assembly:  C.math (in C.math.dll) Version: 1.0.0.0 (1.0.0.0)
Syntax
Copy
public static double frexp(
	double number,
	ref int exponent
)

Parameters

number
Type: SystemDouble
A floating-point number.
exponent
Type: SystemInt32
Reference to an Int32 value to store the exponent to.

Return Value

Type: Double
A fraction in the range [0.5, 1) so that number = fraction * 2^exponent.
Remarks

Special values are treated as follows.

  • If number is ±0, it is returned, and 0 is returned in exponent.
  • If number is infinite, it is returned, and an undefined value is returned in exponent.
  • If number is NaN, it is returned, and an undefined value is returned in exponent.

The function frexp(Double, Int32), together with its dual, ldexp(Double, Int32), can be used to manipulate the representation of a floating-point number without direct bit manipulations.

The relation of frexp(Double, Int32) to logb(Double) and scalbn(Double, Int32) is:

exponent = (number == 0) ? 0 : (int)(1 + logb(number))
fraction = scalbn(number, -exponent)

See frexp in the C standard documentation.

Examples
Copy
Assert.IsTrue(math.frexp(12.8D, ref exponent) = 0.8D);
Assert.IsTrue(exponent = 4);

Assert.IsTrue(math.frexp(0.25D, ref exponent) == 0.5D);
Assert.IsTrue(exponent == -1);

Assert.IsTrue(math.frexp(System.Math.Pow(2D, 1023), ref exponent) == 0.5D);
Assert.IsTrue(exponent == 1024);

Assert.IsTrue(math.frexp(-System.Math.Pow(2D, -1074), ref exponent) == -0.5D);
Assert.IsTrue(exponent == -1073);
See Also