freebsd-nq/lib/msun/man/exp.3
2004-07-02 23:52:20 +00:00

328 lines
6.9 KiB
Groff

.\" Copyright (c) 1985, 1991 Regents of the University of California.
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\" 3. All advertising materials mentioning features or use of this software
.\" must display the following acknowledgement:
.\" This product includes software developed by the University of
.\" California, Berkeley and its contributors.
.\" 4. Neither the name of the University nor the names of its contributors
.\" may be used to endorse or promote products derived from this software
.\" without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" from: @(#)exp.3 6.12 (Berkeley) 7/31/91
.\" $FreeBSD$
.\"
.Dd July 31, 1991
.Dt EXP 3
.Os
.Sh NAME
.Nm exp ,
.Nm expf ,
.Nm exp10 ,
.Nm exp10f ,
.Nm expm1 ,
.Nm expm1f ,
.Nm log ,
.Nm logf ,
.Nm log10 ,
.Nm log10f ,
.Nm log1p ,
.Nm log1pf ,
.Nm pow ,
.Nm powf
.Nd exponential, logarithm, power functions
.Sh LIBRARY
.Lb libm
.Sh SYNOPSIS
.In math.h
.Ft double
.Fn exp "double x"
.Ft float
.Fn expf "float x"
.Ft double
.Fn expm1 "double x"
.Ft float
.Fn expm1f "float x"
.Ft double
.Fn log "double x"
.Ft float
.Fn logf "float x"
.Ft double
.Fn log10 "double x"
.Ft float
.Fn log10f "float x"
.Ft double
.Fn log1p "double x"
.Ft float
.Fn log1pf "float x"
.Ft double
.Fn pow "double x" "double y"
.Ft float
.Fn powf "float x" "float y"
.Sh DESCRIPTION
The
.Fn exp
and the
.Fn expf
functions compute the exponential value of the given argument
.Fa x .
.Pp
The
.Fn expm1
and the
.Fn expm1f
functions compute the value exp(x)\-1 accurately even for tiny argument
.Fa x .
.Pp
The
.Fn log
and the
.Fn logf
functions compute the value of the natural logarithm of argument
.Fa x .
.Pp
The
.Fn log10
and the
.Fn log10f
functions compute the value of the logarithm of argument
.Fa x
to base 10.
.Pp
The
.Fn log1p
and the
.Fn log1pf
functions compute
the value of log(1+x) accurately even for tiny argument
.Fa x .
.Pp
The
.Fn pow
and the
.Fn powf
functions compute the value
of
.Ar x
to the exponent
.Ar y .
.Sh ERROR (due to Roundoff etc.)
.Fn exp x ,
.Fn log x ,
.Fn expm1 x
and
.Fn log1p x
are accurate to within
an
.Em ulp ,
and
.Fn log10 x
to within about 2
.Em ulps ;
an
.Em ulp
is one
.Em Unit
in the
.Em Last
.Em Place .
The error in
.Fn pow x y
is below about 2
.Em ulps
when its
magnitude is moderate, but increases as
.Fn pow x y
approaches
the over/underflow thresholds until almost as many bits could be
lost as are occupied by the floating\-point format's exponent
field; that is 8 bits for
.Tn "VAX D"
and 11 bits for IEEE 754 Double.
No such drastic loss has been exposed by testing; the worst
errors observed have been below 20
.Em ulps
for
.Tn "VAX D" ,
300
.Em ulps
for
.Tn IEEE
754 Double.
Moderate values of
.Fn pow
are accurate enough that
.Fn pow integer integer
is exact until it is bigger than 2**56 on a
.Tn VAX ,
2**53 for
.Tn IEEE
754.
.Sh RETURN VALUES
These functions will return the appropriate computation unless an error
occurs or an argument is out of range.
The functions
.Fn exp ,
.Fn expm1 ,
.Fn pow
detect if the computed value will overflow,
set the global variable
.Va errno
to
.Er ERANGE
and cause a reserved operand fault on a
.Tn VAX
or
.Tn Tahoe .
The functions
.Fn pow x y
checks to see if
.Fa x
< 0 and
.Fa y
is not an integer, in the event this is true,
the global variable
.Va errno
is set to
.Er EDOM
and on the
.Tn VAX
and
.Tn Tahoe
generate a reserved operand fault.
On a
.Tn VAX
and
.Tn Tahoe ,
.Va errno
is set to
.Er EDOM
and the reserved operand is returned
by log unless
.Fa x
> 0, by
.Fn log1p
unless
.Fa x
> \-1.
.Sh NOTES
The functions exp(x)\-1 and log(1+x) are called
expm1 and logp1 in
.Tn BASIC
on the Hewlett\-Packard
.Tn HP Ns \-71B
and
.Tn APPLE
Macintosh,
.Tn EXP1
and
.Tn LN1
in Pascal, exp1 and log1 in C
on
.Tn APPLE
Macintoshes, where they have been provided to make
sure financial calculations of ((1+x)**n\-1)/x, namely
expm1(n\(**log1p(x))/x, will be accurate when x is tiny.
They also provide accurate inverse hyperbolic functions.
.Pp
The function
.Fn pow x 0
returns x**0 = 1 for all x including x = 0,
.if n \
Infinity
.if t \
\(if
(not found on a
.Tn VAX ) ,
and
.Em NaN
(the reserved
operand on a
.Tn VAX ) .
Previous implementations of pow may
have defined x**0 to be undefined in some or all of these
cases.
Here are reasons for returning x**0 = 1 always:
.Bl -enum -width indent
.It
Any program that already tests whether x is zero (or
infinite or \*(Na) before computing x**0 cannot care
whether 0**0 = 1 or not.
Any program that depends
upon 0**0 to be invalid is dubious anyway since that
expression's meaning and, if invalid, its consequences
vary from one computer system to another.
.It
Some Algebra texts (e.g.\& Sigler's) define x**0 = 1 for
all x, including x = 0.
This is compatible with the convention that accepts a[0]
as the value of polynomial
.Bd -literal -offset indent
p(x) = a[0]\(**x**0 + a[1]\(**x**1 + a[2]\(**x**2 +...+ a[n]\(**x**n
.Ed
.Pp
at x = 0 rather than reject a[0]\(**0**0 as invalid.
.It
Analysts will accept 0**0 = 1 despite that x**y can
approach anything or nothing as x and y approach 0
independently.
The reason for setting 0**0 = 1 anyway is this:
.Bd -ragged -offset indent
If x(z) and y(z) are
.Em any
functions analytic (expandable
in power series) in z around z = 0, and if there
x(0) = y(0) = 0, then x(z)**y(z) \(-> 1 as z \(-> 0.
.Ed
.It
If 0**0 = 1, then
.if n \
infinity**0 = 1/0**0 = 1 too; and
.if t \
\(if**0 = 1/0**0 = 1 too; and
then \*(Na**0 = 1 too because x**0 = 1 for all finite
and infinite x, i.e., independently of x.
.El
.Sh SEE ALSO
.Xr math 3
.Sh HISTORY
A
.Fn exp ,
.Fn log
and
.Fn pow
functions
appeared in
.At v6 .
A
.Fn log10
function
appeared in
.At v7 .
The
.Fn log1p
and
.Fn expm1
functions appeared in
.Bx 4.3 .