208 lines
6.5 KiB
Plaintext
208 lines
6.5 KiB
Plaintext
[section:next_float Floating-Point Representation Distance (ULP),
|
|
and Finding Adjacent Floating-Point Values]
|
|
|
|
[@http://en.wikipedia.org/wiki/Unit_in_the_last_place Unit of Least Precision or Unit in the Last Place]
|
|
is the gap between two different, but as close as possible, floating-point numbers.
|
|
|
|
Most decimal values, for example 0.1, cannot be exactly represented as floating-point values,
|
|
but will be stored as the
|
|
[@http://en.wikipedia.org/wiki/Floating_point#Representable_numbers.2C_conversion_and_rounding
|
|
closest representable floating-point].
|
|
|
|
Functions are provided for finding adjacent greater and lesser floating-point values,
|
|
and estimating the number of gaps between any two floating-point values.
|
|
|
|
The floating-point type FPT must have has a fixed number of bits in the representation.
|
|
The number of bits may set at runtime, but must be the same for all numbers.
|
|
For example, __NTL_quad_float type (fixed 128-bit representation)
|
|
or __NTL_RR type (arbitrary but fixed decimal digits, default 150)
|
|
but *not* a type that extends the representation to provide an exact representation
|
|
for any number, for example [@http://keithbriggs.info/xrc.html XRC eXact Real in C].
|
|
|
|
[section:nextafter Finding the Next Representable Value in a Specific Direction (nextafter)]
|
|
|
|
[h4 Synopsis]
|
|
|
|
``
|
|
#include <boost/math/special_functions/next.hpp>
|
|
``
|
|
|
|
namespace boost{ namespace math{
|
|
|
|
template <class FPT>
|
|
FPT nextafter(FPT val, FPT direction);
|
|
|
|
}} // namespaces
|
|
|
|
[h4 Description - nextafter]
|
|
|
|
This is an implementation of the `nextafter` function included in the C99 standard.
|
|
(It is also effectively an implementation of the C99 'nexttoward' legacy function
|
|
which differs only having a long double direction,
|
|
and can generally serve in its place if required).
|
|
|
|
[note The C99 functions must use suffixes f and l to distinguish float and long double versions.
|
|
C++ uses the template mechanism instead.]
|
|
|
|
Returns the next representable value after /x/ in the direction of /y/. If
|
|
`x == y` then returns /x/. If /x/ is non-finite then returns the result of
|
|
a __domain_error. If there is no such value in the direction of /y/ then
|
|
returns an __overflow_error.
|
|
|
|
[warning The template parameter FTP must be a floating-point type.
|
|
An integer type, for example, will produce an unhelpful error message.]
|
|
|
|
[tip Nearly always, you just want the next or prior representable value,
|
|
so instead use `float_next` or `float_prior` below.]
|
|
|
|
[h4 Examples - nextafter]
|
|
|
|
The two representations using a 32-bit float either side of unity are:
|
|
``
|
|
The nearest (exact) representation of 1.F is 1.00000000
|
|
nextafter(1.F, 999) is 1.00000012
|
|
nextafter(1/f, -999) is 0.99999994
|
|
|
|
The nearest (not exact) representation of 0.1F is 0.100000001
|
|
nextafter(0.1F, 10) is 0.100000009
|
|
nextafter(0.1F, 10) is 0.099999994
|
|
``
|
|
|
|
[endsect]
|
|
|
|
[section:float_next Finding the Next Greater Representable Value (float_next)]
|
|
|
|
[h4 Synopsis]
|
|
|
|
``
|
|
#include <boost/math/special_functions/next.hpp>
|
|
``
|
|
|
|
namespace boost{ namespace math{
|
|
|
|
template <class FPT>
|
|
FPT float_next(FPT val);
|
|
|
|
}} // namespaces
|
|
|
|
[h4 Description - float_next]
|
|
|
|
Returns the next representable value which is greater than /x/.
|
|
If /x/ is non-finite then returns the result of
|
|
a __domain_error. If there is no such value greater than /x/ then
|
|
returns an __overflow_error.
|
|
|
|
Has the same effect as
|
|
|
|
nextafter(val, (std::numeric_limits<FPT>::max)());
|
|
|
|
[endsect] [/section:float_next Finding the Next Greater Representable Value (float_prior)]
|
|
|
|
[section:float_prior Finding the Next Smaller Representable Value (float_prior)]
|
|
|
|
[h4 Synopsis]
|
|
|
|
``
|
|
#include <boost/math/special_functions/next.hpp>
|
|
``
|
|
|
|
namespace boost{ namespace math{
|
|
|
|
template <class FPT>
|
|
FPT float_prior(FPT val);
|
|
|
|
}} // namespaces
|
|
|
|
|
|
[h4 Description - float_prior]
|
|
|
|
Returns the next representable value which is less than /x/.
|
|
If /x/ is non-finite then returns the result of
|
|
a __domain_error. If there is no such value less than /x/ then
|
|
returns an __overflow_error.
|
|
|
|
Has the same effect as
|
|
|
|
nextafter(val, -(std::numeric_limits<FPT>::max)()); // Note most negative value -max.
|
|
|
|
[endsect] [/section:float_prior Finding the Next Smaller Representable Value (float_prior)]
|
|
|
|
[section:float_distance Calculating the Representation Distance
|
|
Between Two Floating Point Values (ULP) float_distance]
|
|
|
|
Function float_distance finds the number of gaps/bits/ULP between any two floating-point values.
|
|
If the significands of floating-point numbers are viewed as integers,
|
|
then their difference is the number of ULP/gaps/bits different.
|
|
|
|
[h4 Synopsis]
|
|
|
|
``
|
|
#include <boost/math/special_functions/next.hpp>
|
|
``
|
|
|
|
namespace boost{ namespace math{
|
|
|
|
template <class FPT>
|
|
FPT float_distance(FPT a, FPT b);
|
|
|
|
}} // namespaces
|
|
|
|
[h4 Description - float_distance]
|
|
|
|
Returns the distance between /a/ and /b/: the result is always
|
|
a signed integer value (stored in floating-point type FPT)
|
|
representing the number of distinct representations between /a/ and /b/.
|
|
|
|
Note that
|
|
|
|
* `float_distance(a, a)` always returns 0.
|
|
* `float_distance(float_next(a), a)` always returns -1.
|
|
* `float_distance(float_prior(a), a)` always returns 1.
|
|
|
|
The function `float_distance` is equivalent to calculating the number
|
|
of ULP (Units in the Last Place) between /a/ and /b/ except that it
|
|
returns a signed value indicating whether `a > b` or not.
|
|
|
|
If the distance is too great then it may not be able
|
|
to be represented as an exact integer by type FPT,
|
|
but in practice this is unlikely to be a issue.
|
|
|
|
[endsect] [/section:float_distance Calculating the Representation Distance
|
|
Between Two Floating Point Values (ULP) float_distance]
|
|
|
|
[section:float_advance Advancing a Floating Point Value by a Specific
|
|
Representation Distance (ULP) float_advance]
|
|
|
|
Function float_advance advances a floating point number by a specified number
|
|
of ULP.
|
|
|
|
[h4 Synopsis]
|
|
|
|
``
|
|
#include <boost/math/special_functions/next.hpp>
|
|
``
|
|
|
|
namespace boost{ namespace math{
|
|
|
|
template <class FPT>
|
|
FPT float_advance(FPT val, int distance);
|
|
|
|
}} // namespaces
|
|
|
|
[h4 Description - float_advance]
|
|
|
|
Returns a floating point number /r/ such that `float_distance(val, r) == distance`.
|
|
|
|
[endsect] [/section:float_advance]
|
|
|
|
[endsect] [/ section:next_float Floating-Point Representation Distance (ULP),
|
|
and Finding Adjacent Floating-Point Values]
|
|
|
|
[/
|
|
Copyright 2008 John Maddock and Paul A. Bristow.
|
|
Distributed under the Boost Software License, Version 1.0.
|
|
(See accompanying file LICENSE_1_0.txt or copy at
|
|
http://www.boost.org/LICENSE_1_0.txt).
|
|
]
|
|
|