CBRT(3,L)                   AIX Technical Reference                   CBRT(3,L)



-------------------------------------------------------------------------------
cbrt, exp, expm1, log, log10, log1p, pow, sqrt



PURPOSE

Computes exponential, logarithm, power, square root and cube root functions.

LIBRARY

Math Library (libm.a)

SYNTAX

#include <math.h>





                double cbrt (x)             double log10 (x)
                double x;                   double x;

                double exp (x)              double log1p (x)
                double x;                   double x;

                double expm1 (x)            double pow (x, y)
                double x;                   double x, y;

                double log (x)              double sqrt (x)
                double x;                   double x;



DESCRIPTION

The cbrt subroutine returns the cube root of x.

The exp subroutine to returns e(x).

The expm1 subroutine returns log (1+x).

The log subroutine returns the natural logarithm of x,
ln x
.  The value of x must be positive.

The log10 subroutine returns the logarithm base 10 of x,
log sub 10 % x
.  The value of x must be positive.

The log1p subroutine returns log (1+x).



Processed November 7, 1990         CBRT(3,L)                                  1






CBRT(3,L)                   AIX Technical Reference                   CBRT(3,L)




The pow subroutine returns x(y).  The values of x and y may not both be 0.  If
x is negative or 0, then y must be an integer.

The sqrt subroutine returns the square root of x.  The value of x cannot be
negative.

Note:  The expm1 and log1p subroutines are useful to guarantee that financial
       calculations of ((1+x(n)-1/ x (expm1(n* log1p (x))/x are accurate when x
       is very small (for example, small daily interest rates).  They also make
       it easier to write accurate inverse hyperbolic functions.

ERROR CONDITIONS

The exp, log, log10, and sqrt subroutines can perform either of the following
types of error handling.  The pow subroutine always handles errors according to
the second method.  Both types of error handling allow you to define special
actions to be taken when an error occurs.

  1. For cbrt, exp, log, log10, and sqrt, exception handling will be performed
    according to ANSI/IEEE standard 754 for binary floating-point arithmetic by
    default.

    If a hardware floating-point processor is installed in your system, then
    using this option can provide greater performance in addition to IEEE
    exception handling.  This mode instructs the C compiler to generate code
    that avoids the overhead of the math library subroutines by generating math
    coprocessor code in-line.

  2. The second method is matherr error handling, as described on page
    matherr-1.  The default error-handling procedures for these subroutines
    are as follows:

    cbrt   If x is negative, cbrt returns the value 0, sets errno to EDOM and
           writes an error message to the standard error output.

    exp    If the correct value overflows, exp returns HUGE and sets errno to
           ERANGE.

    expm1  If the correct value overflows, expm1 returns HUGE but does not
           modify errno.

    log     If x is negative or 0, log returns the value -HUGE, sets errno to
           EDOM if x is negative and to ERANGE if x is 0, and writes an error
           message to the standard error output.

    log10  If x is negative or 0, log10 returns the value -HUGE, sets errno to
           EDOM if x is negative and to ERANGE if x is 0, and writes an error
           message to the standard error output.






Processed November 7, 1990         CBRT(3,L)                                  2






CBRT(3,L)                   AIX Technical Reference                   CBRT(3,L)



    log1p  If x is less than -1, then log1p returns the value QNaN.  If x
           equals -1, then log1p returns the value -HUGE.  In neither case is
           errno set.

    pow    If x is negative or 0 and y is not an integer, or if x and y are
           both 0, pow returns the value 0, sets errno to EDOM, and writes an
           error message to the standard error output.  If the correct value
           overflows, pow returns HUGE and sets errno to ERANGE.

    sqrt   If x is negative, sqrt returns the value 0, sets errno to EDOM, and
           writes an error message to the standard error output.

This second method of error handling is invoked by including the -z option on
the cc command line.

RELATED INFORMATION

In this book:  "hypot, cabs,"  "matherr," and "sinh, cosh, tanh."





































Processed November 7, 1990         CBRT(3,L)                                  3