acosh, acoshf, acoshl
From cppreference.com
Defined in header <math.h>
|
||
float acoshf( float arg ); |
(1) | (since C99) |
double acosh( double arg ); |
(2) | (since C99) |
long double acoshl( long double arg ); |
(3) | (since C99) |
Defined in header <tgmath.h>
|
||
#define acosh( arg ) |
(4) | (since C99) |
1-3) Computes the inverse hyperbolic cosine of
arg
.4) Type-generic macro: If the argument has type long double,
acoshl
is called. Otherwise, if the argument has integer type or the type double, acosh
is called. Otherwise, acoshf
is called. If the argument is complex, then the macro invokes the corresponding complex function (cacoshf, cacosh, cacoshl).Parameters
arg | - | floating point value representing the area of a hyperbolic sector |
Return value
If no errors occur, the inverse hyperbolic cosine of arg
(cosh-1
(arg), or arcosh(arg)) on the interval [0, +], is returned.
If a domain error occurs, an implementation-defined value is returned (NaN where supported).
Error handling
Errors are reported as specified in math_errhandling.
If the argument is less than 1, a domain error occurs.
If the implementation supports IEEE floating-point arithmetic (IEC 60559),
- if the argument is less than 1, FE_INVALID is raised an NaN is returned
- if the argument is 1, +0 is returned
- if the argument is +, + is returned
- if the argument is NaN, NaN is returned
Notes
Although the C standard names this function "arc hyperbolic cosine", the inverse functions of the hyperbolic functions are the area functions. Their argument is the area of a hyperbolic sector, not an arc. The correct name is "inverse hyperbolic cosine" (used by POSIX) or "area hyperbolic cosine".
Example
Run this code
#include <stdio.h> #include <math.h> #include <float.h> #include <errno.h> #include <fenv.h> #pragma STDC FENV_ACCESS ON int main(void) { printf("acosh(1) =%f\nacosh(10) =%f\n", acosh(1), acosh(10)); printf("acosh(DBL_MAX) =%f\nacosh(Inf) =%f\n", acosh(DBL_MAX), acosh(INFINITY)); //error handling errno = 0; feclearexcept(FE_ALL_EXCEPT); printf("acosh(0.5) =%f\n", acosh(0.5)); if(errno == EDOM) perror(" errno == EDOM"); if(fetestexcept(FE_INVALID)) puts(" FE_INVALID raised"); }
Possible output:
acosh(1) = 0.000000 acosh(10) = 2.993223 acosh(DBL_MAX) = 710.475860 acosh(Inf) = inf acosh(0.5) = -nan errno == EDOM: Numerical argument out of domain FE_INVALID raised
References
- C11 standard (ISO/IEC 9899:2011):
- 7.12.5.1 The acosh functions (p: 240)
- 7.25 Type-generic math <tgmath.h> (p: 373-375)
- F.10.2.1 The acosh functions (p: 520)
- C99 standard (ISO/IEC 9899:1999):
- 7.12.5.1 The acosh functions (p: 221)
- 7.22 Type-generic math <tgmath.h> (p: 335-337)
- F.9.2.1 The acosh functions (p: 457)
See also
(C99)(C99)(C99) |
computes inverse hyperbolic sine (arsinh(x)) (function) |
(C99)(C99)(C99) |
computes inverse hyperbolic tangent (artanh(x)) (function) |
(C99)(C99) |
computes hyperbolic cosine (ch(x)) (function) |
(C99)(C99)(C99) |
computes the complex arc hyperbolic cosine (function) |