Floating-Point Intrinsic Functions

Article
06/30/2006

The floating-point intrinsic functions can be separated into a group of general-purpose functions and three groups of related transcendental functions. Except for _isunordered, these functions have always been part of the Microsoft C Run-Time Library for Windows CE, and the intrinsic versions perform the same operation as the run-time versions except as noted. For more information, see Microsoft C Run-Time Library for Windows CE .NET.

All floating-point operations should adhere to the IEEE Standard for Binary Floating-Point Arithmetic, or IEEE-754. The IEEE standard reserves some exponent fields for special values. These special values include zero and the value NaN, Not a Number. NaN represents a value that is not a real number, such as the indeterminate result of dividing a real number by zero.

The floating-point intrinsic functions are available in both double-precision and single-precision versions, and you can implement them as single instructions on some platforms. The following table shows the general-purpose floating-point intrinsic functions.

Operation	Description
fabs fabsf	Calculates the absolute value of a floating-point number.
floor floorf	Calculates the floor (greatest integer less than or equal to) value of a number.
ceil ceilf	Calculates the ceiling (smallest integer greater than or equal to) value of a number. This is the result of rounding up.
fmod fmodf	Calculates the modulus, or floating-point remainder.
sqrt sqrtf	Calculates the square root of a floating-point number.
_isnan _isnanf	Determines a Boolean value, true when argument is not a number (NaN). Returns false when argument is zero, subnormal, or infinite.
_isunordered _isunorderedf	Determines a Boolean value, returns true when arguments are unordered. This function is not part of the Microsoft C Run-Time Library for Windows CE .NET.

In addition to general-purpose floating-point intrinsic functions, Windows CE .NET provides support for transcendental floating-point functions. These intrinsic functions implement logarithmic, trigonometric, and hyperbolic functions. The transcendental intrinsic functions take double-precision arguments and return a double-precision result. These functions have always been available as intrinsic on all platforms and are provided for backwards compatibility. In most cases they are implemented by calling the Microsoft C Run-Time Library.

The following table shows the logarithmic functions.

Intrinsic function	Description
exp	Calculates the exponential function value of a floating-point number.
log	Calculates the natural logarithm (base e) of a floating-point number.
log10	Calculates the common logarithm (base 10) of a floating-point number.
pow	Calculates the value of x raised to the power of y, x^y.

The following table shows the trigonometric functions.

Intrinsic function	Description
cos	Calculates the cosine of a floating-point number from a radian argument.
sin	Calculates the sine of a floating-point number from a radian argument.
tan	Calculates the tangent of a floating-point number from a radian argument.
acos	Calculates the arccosine of a floating-point number, in radians.
asin	Calculates the arcsine of a floating-point number, in radians.
atan	Calculates the arctangent of a floating-point number, in radians.
atan2	Calculates the arctangent of a floating-point number y/x as a rational number.

The following table shows the hyperbolic functions.

Intrinsic function	Description
sinh	Calculates the hyperbolic cosine of a floating-point number.
cosh	Calculates the hyperbolic sine of a floating-point number.
tanh	Calculates the hyperbolic tangent of a floating-point number.

Share via

Floating-Point Intrinsic Functions

See Also

Additional resources