Datasheet

Floating-point Support
ARM DUI 0067D Copyright © 1999-2001 ARM Limited. All rights reserved. 5-25
5.4.1 Inverse hyperbolic functions (acosh, asinh, atanh)
double acosh(double x);
double asinh(double x);
double atanh(double x);
These functions are the inverses of the ANSI-required cosh , sinh and tanh :
• Because
cosh is a symmetric function (that is, it returns the same value when
applied to x or
–x ), acosh always has a choice of two return values, one positive
and one negative. It chooses the positive result.
•
acosh returns an EDOM error if called with an argument less than 1.0.
•
atanh returns an EDOM error if called with an argument whose absolute value
exceeds 1.0.
5.4.2 Cube root (cbrt)
double cbrt(double x);
This function returns the cube root of its argument.
5.4.3 Copy sign (copysign)
double copysign(double x, double y);
This function replaces the sign bit of x with the sign bit of y , and returns the result. It
causes no errors or exceptions, even when applied to NaNs and infinities.
5.4.4 Error functions (erf, erfc)
double erf(double x);
double erfc(double x);
These functions compute the standard statistical error function, related to the Normal
distribution:
•
erf computes the ordinary error function of x.
•
erfc computes one minus erf(x) . It is better to use erfc(x) than 1-erf(x) when x
is large, because the answer is more accurate.