math.ilogb Method (Single)

mathilogb Method (Single)

Gets the unbiased exponent of the specified floating-point number.

Namespace: C
Assembly: C.math (in C.math.dll) Version: 1.0.0.0 (1.0.0.0)

Syntax

public static int ilogb(
	float number
)

Public Shared Function ilogb ( 
	number As Single
) As Integer

public:
static int ilogb(
	float number
)

static member ilogb : 
        number : float32 -> int

Parameters

number: Type: SystemSingle
A floating-point number.

Return Value

Type: Int32
The unbiased exponent of the specified floating-point number, or a special value if number is not normal or subnormal.

Remarks

The unbiased exponent is the integral part of the logarithm base 2 of number. The unbiased exponent is such that:

number = significand(number) * 2^ilogb(number).

The return unbiased exponent is valid for all normal and subnormal numbers. Special values are treated as follows.

If number is ±0, FP_ILOGB0 is returned.
If number is infinite, INT_MAX is returned.
If number is NaN, FP_ILOGBNAN is returned.

If number is not zero, infinite, or NaN, the value returned is exactly equivalent to (Int32)logb(number).

The value of the exponent returned by ilogb(Single) is always 1 less than the exponent retuned by frexp(Single, Int32) because of the different normalization requirements: for ilogb(Single), the normalized significand is in the interval [1, 2), but for frexp(Single, Int32), the normalized significand is in the interval [0.5, 1).

See ilogb in the C standard documentation.

Examples

C++

Copy

Assert.IsTrue(math.ilogb(1F) == 0);
Assert.IsTrue(math.ilogb((float)System.Math.E) == 1);
Assert.IsTrue(math.ilogb(1024F) == 10);
Assert.IsTrue(math.ilogb(-2000F) == 10);

Assert.IsTrue(math.ilogb(2F) == 1);
Assert.IsTrue(math.ilogb((float)Math.Pow(2F, 56F)) == 56);
Assert.IsTrue(math.ilogb(1.1F * (float)Math.Pow(2F, -149F)) == -149);
Assert.IsTrue(math.ilogb((float)Math.Pow(2F, -150F)) == math.FP_ILOGB0);
Assert.IsTrue(math.ilogb((float)Math.Pow(2F, 128F)) == math.INT_MAX);
Assert.IsTrue(math.ilogb((float)Math.Pow(2D, 127F)) == 127);
Assert.IsTrue(math.ilogb(2F * (float)Math.Pow(2F, 102F)) == 103);

Assert.IsTrue(math.ilogb(math.FLT_DENORM_MIN) == math.FLT_EXP_MIN - math.FLT_MANT_BITS);
Assert.IsTrue(math.ilogb(math.FLT_DENORM_MAX) == math.FLT_EXP_MIN - 1);
Assert.IsTrue(math.ilogb(math.FLT_MIN) == math.FLT_EXP_MIN);
Assert.IsTrue(math.ilogb(math.FLT_MAX) == math.FLT_EXP_MAX);

Assert.IsTrue(math.ilogb(System.Single.PositiveInfinity) == math.INT_MAX);
Assert.IsTrue(math.ilogb(System.Single.NegativeInfinity) == math.INT_MAX);
Assert.IsTrue(math.ilogb(0F) == math.FP_ILOGB0);
Assert.IsTrue(math.ilogb(-0F) == math.FP_ILOGB0);
Assert.IsTrue(math.ilogb(System.Single.NaN) == math.FP_ILOGBNAN);
Assert.IsTrue(math.ilogb(-System.Single.NaN) == math.FP_ILOGBNAN);

Assert.IsTrue(math.ilogb(1F) = 0);
Assert.IsTrue(math.ilogb(CSng(System.Math.E)) = 1);
Assert.IsTrue(math.ilogb(1024F) = 10);
Assert.IsTrue(math.ilogb(-2000F) = 10);

Assert.IsTrue(math.ilogb(2F) = 1);
Assert.IsTrue(math.ilogb(CSng(Math.Pow(2F, 56F))) = 56);
Assert.IsTrue(math.ilogb(1.1F * CSng(Math.Pow(2F, -149F))) = -149);
Assert.IsTrue(math.ilogb(CSng(Math.Pow(2F, -150F))) = math.FP_ILOGB0);
Assert.IsTrue(math.ilogb(CSng(Math.Pow(2F, 128F))) = math.INT_MAX);
Assert.IsTrue(math.ilogb(CSng(Math.Pow(2D, 127F))) = 127);
Assert.IsTrue(math.ilogb(2F * CSng(Math.Pow(2F, 102F))) = 103);

Assert.IsTrue(math.ilogb(math.FLT_DENORM_MIN) = math.FLT_EXP_MIN - math.FLT_MANT_BITS);
Assert.IsTrue(math.ilogb(math.FLT_DENORM_MAX) = math.FLT_EXP_MIN - 1);
Assert.IsTrue(math.ilogb(math.FLT_MIN) = math.FLT_EXP_MIN);
Assert.IsTrue(math.ilogb(math.FLT_MAX) = math.FLT_EXP_MAX);

Assert.IsTrue(math.ilogb(System.Single.PositiveInfinity) = math.INT_MAX);
Assert.IsTrue(math.ilogb(System.Single.NegativeInfinity) = math.INT_MAX);
Assert.IsTrue(math.ilogb(0F) = math.FP_ILOGB0);
Assert.IsTrue(math.ilogb(-0F) = math.FP_ILOGB0);
Assert.IsTrue(math.ilogb(System.Single.NaN) = math.FP_ILOGBNAN);
Assert.IsTrue(math.ilogb(-System.Single.NaN) = math.FP_ILOGBNAN);

Reference

math Class

ilogb Overload

C Namespace