NAME
math — math functions, constants, and types
DESCRIPTION
This file contains declarations of all the functions in the
Math Library (described in Section (3M)).
For
Itanium®-based
systems, included in a compilation under the
-fpwidetypes
option, this file defines the types
- extended
The 80-bit double-extended type on Itanium-based systems.
- quad
An IEEE 754-compliant, 128-bit floating-point type.
On HPUX,
quad
is a synonym for
long double.
It defines the types
- float_t
A floating type, at least as wide as
float.
On PA-RISC systems,
float_t
is
float.
On Itanium-based systems, if
FLT_EVAL_METHOD
equals 0,
float_t
is
float;
if
FLT_EVAL_METHOD
equals 1,
float_t
is
double;
if
FLT_EVAL_METHOD
equals -2,
float_t
is
extended.
- double_t
A floating type, at least as wide as
double.
On PA-RISC systems,
double_t
is
double.
For Itanium-based systems, if
FLT_EVAL_METHOD
equals 0 or 1,
double_t
is
double;
if
FLT_EVAL_METHOD
equals -2,
double_t
is
extended.
For Itanium-based systems, the value of
FLT_EVAL_METHOD
is 0, 1, or -2, according to the use of the compiler option
-fpeval=float(default),
-fpeval=double,
or
-fpeval=extended.
It defines the following constants,
which, for Itanium-based systems, may be used to
initialize statics and aggregates:
- NAN
A quiet NaN (Not-a-Number) value (type
float).
- INFINITY
A positive infinity value (type
float).
- HUGE_VAL
The maximum value (type
double)
representable in the
double
type (IEEE positive infinity).
- HUGE_VALF
The maximum value (type
float)
representable in the
float
type (IEEE positive infinity).
- HUGE_VALL
For Itanium-based systems only, the maximum value (type
long double)
representable in the
long double
type (IEEE positive infinity).
- HUGE_VALW
For Itanium-based systems only, the maximum value (type
extended)
representable in the
extended
type (IEEE positive infinity).
To use
HUGE_VALW
compile with the
-fpwidetypes
option.
- HUGE_VALQ
For Itanium-based systems only, equivalent to
HUGE_VALL.
To use
HUGE_VALQ
compile with the
-fpwidetypes
option.
It defines the following integer constants which give
special-case values returned by the
ilogb()
functions:
- FP_ILOGB0
Returned by
ilogb()
if its argument is zero.
- FP_ILOGBNAN
Returned by
ilogb()
if its argument is a NaN.
For Itanium-based systems,
it defines the following integer constants which identify the
error handling method supported by
<math.h>
functions:
- MATH_ERRNO
Indicates support for the ISO/IEC C99
errno
specification.
- MATH_ERREXCEPT
Indicates support for the ISO/IEC C99
exception flag
specification.
- math_errhandling
Defined to be
MATH_ERRNO
if the
+Olibmerrno
compiler option (not the default) is used, or defined to be
MATH_ERREXCEPT
otherwise.
It defines the following mathematical constants (type double)
for user convenience:
- M_E
The base of natural logarithms
(e).
- M_LOG2E
The base-2 logarithm of
e.
- M_LOG10E
The base-10 logarithm of
e.
- M_LN2
The natural logarithm of 2.
- M_LN10
The natural logarithm of 10.
- M_PI
The ratio of the circumference of a circle to its diameter.
(There are also several fractions of
Pi,
its reciprocal, and its square root:
M_PI_2,
M_PI_4,
M_1_PI,
M_2_PI,
and
M_2_SQRTPI).
- M_SQRT2
The positive square root of 2.
- M_SQRT1_2
The positive square root of 1/2.
For the definitions of various machine-dependent constants,
see the description of the
<values.h>
header file.
To use one of the types
extended
or
quad,
or one of the macros
HUGE_VALW
or
HUGE_VALQ,
compile with the
-fpwidetypes
option.
STANDARDS CONFORMANCE
<math.h>: SVID3, XPG4.2, ANSI C, ISO/IEC C99
