The special functions from:
The Naval Surface Warfare Center (NSWC) Mathematics Library
-----------------------------------------------------------
This Fortran 90 version has been compiled by Alan Miller.
Accuracy
--------
The accuracy of the single precision versions of these routines is
typically
about 13-14 decimal digits, which is `nearly' double precision accuracy
for
many computer/compiler combinations, while the double precision versions
are often accurate to 25-27 decimals. These accuracies assume that the
machine being used have sufficiently large word sizes.
The common 4-byte (single precision) and 8-byte (double precision)
representations of floating-point numbers give nearly 7 and 16 decimal
digit representations respectively. The single precision functions in
this library could be used in double precision where 13-14 decimal digit
accuracy is adequate, as they will probably be fadter than functions from
other sources giving full 16 decimal digit accuracy. Similarly, the
double precision functions could be used in quadruple precision when
25-27 decimals is adequate. The first letter of the function name is
usually D if it is in double precision.
Functions
---------
N.B. There are many functions and routines which are intended to be used
only by other functions and not called directly by the user.
CPABS Calculates SQRT(x^2 + y^2)
DCPABS
CREC Reciprocal of a complex number
DCREC
CDIV Division of two complex numbers
CDIVID Division of double precision complex numbers, each stored as a pair
of double precision numbers
CAPO Conversion from polar to cartesian coordinates
SIN0 Sine of x*pi/2
COS0 Cosine of x*pi/2
SIN1 Sine of x*pi
DSIN1
COS1 Cosine of x*pi
SNHCSH Calculates several hyperbolic functions:
SINHM(X) = SINH(X) - X
COSHM(X) = COSH(X) - 1
