Boost-Commit :

Date view	Thread view	Subject view	Author view

Subject: [Boost-commit] svn:boost r83536 - in sandbox/precision/libs/precision/doc: . html html/precision html/precision/precision
From: pbristow_at_[hidden]
Date: 2013-03-24 05:22:33

Next message: john_at_[hidden]: "[Boost-commit] svn:boost r83537 - in trunk/boost/math: constants special_functions/detail tools"
Previous message: e_float_at_[hidden]: "[Boost-commit] svn:boost r83535 - sandbox/precision/libs/precision/doc"

Author: pbristow
Date: 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
New Revision: 83536
URL: http://svn.boost.org/trac/boost/changeset/83536

Log:
Updated html and pdf from Chris's qbk - no changes to qbk
Binary files modified:
   sandbox/precision/libs/precision/doc/precision.pdf
Text files modified:
   sandbox/precision/libs/precision/doc/html/index.html | 7 +-
   sandbox/precision/libs/precision/doc/html/precision/abstract.html | 44 ++++++++++----
   sandbox/precision/libs/precision/doc/html/precision/background.html | 46 +++++++++------
   sandbox/precision/libs/precision/doc/html/precision/introduction.html | 36 ++++++-----
   sandbox/precision/libs/precision/doc/html/precision/precision.html | 21 ++++---
   sandbox/precision/libs/precision/doc/html/precision/precision/integerfixedtypes.html | 2
   sandbox/precision/libs/precision/doc/html/precision/precision/newfloattypes.html | 42 ++++++++-----
   sandbox/precision/libs/precision/doc/html/precision/references.html | 2
   sandbox/precision/libs/precision/doc/html/precision/suffixes.html | 120 +++++++++++++++++++++------------------
   sandbox/precision/libs/precision/doc/html/precision/version_id.html | 2
   10 files changed, 188 insertions(+), 134 deletions(-)

Modified: sandbox/precision/libs/precision/doc/html/index.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/index.html (original)
+++ sandbox/precision/libs/precision/doc/html/index.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -50,6 +50,7 @@
<dt>Abstract</dt>
<dt>Background</dt>
<dt>Introduction</dt>
+<dt>The proposed types and potential extensions</dt>
<dt><a href="precision/suffixes.html">How to specify constants with quad
and half precision?</a></dt>
<dt>Specifying Precision</dt>
@@ -59,8 +60,8 @@
<dt>Proposed new section</dt>
</dl></dd>
<dt>Interaction with complex</dt>
-<dt><a href="precision/microfpu.html">Improvemed safety for microcontrollers
- with an FPU</a></dt>
+<dt><a href="precision/microfpu.html">Improved efficiency and robustness
+ for microcontrollers</a></dt>
<dt>References</dt>
<dt>Version Info</dt>
</dl>
@@ -88,7 +89,7 @@
</table></div>
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
-<td align="left">Last revised: March 23, 2013 at 12:20:46 GMT</td>
+<td align="left">Last revised: March 24, 2013 at 09:21:29 GMT</td>
<td align="right"><div class="copyright-footer"></div></td>
</tr></table>
<hr>

Modified: sandbox/precision/libs/precision/doc/html/precision/abstract.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/abstract.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/abstract.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -27,32 +27,52 @@
<a name="precision.abstract"></a><a class="link" href="abstract.html" title="Abstract">Abstract</a>
</h2></div></div></div>

- It is proposed to add several optional typedefs with specified widths for floating-point
- types including <code class="computeroutput">float32_t,
- float64_t _float128_t</code>
- (similar to <code class="computeroutput">int64_t</code> for integer
- types).
+ It is proposed to add to the C++ standard several optional typedefs for floating-point
+ types with specified widths. In particular, the optional types include <code class="computeroutput">float16_t</code>, <code class="computeroutput">float32_t</code>,
+ <code class="computeroutput">float64_t</code>, <code class="computeroutput">float128_t</code>,
+ and their corresponding fast and least types. The optional floating-point types
+ are to conform with the corresponding types <code class="computeroutput">binary16</code>,
+ <code class="computeroutput">binary32</code>, <code class="computeroutput">binary64</code>,
+ and <code class="computeroutput">binary128</code> described in
+ <a href="http://en.wikipedia.org/wiki/IEEE_floating_point" target="_top">IEEE_ floating-point
+ format</a>.
 

- These will be defined in the global and <code class="computeroutput">std</code>
- namespaces.
+ The optional floating-point types with specified widths are to be contained
+ in a new standard library header <code class="computeroutput"><cstdfloat></code>.
+ Any of the optional floating-point types with specified widths included in
+ the implementation must have full support for the functions in <code class="computeroutput"><cmath></code> and seamlessly interoperate with <code class="computeroutput"><limits></code> and <code class="computeroutput"><complex></code>.
+ The proposed new floating-point types with specified widths will be defined
+ in the global and <code class="computeroutput">std</code> namespaces.
 

- And also to provide additional suffix(es) to specify extended precision constants
- to suit precisions lower than that of <code class="computeroutput">float</code>
- higher than that of <code class="computeroutput">long double</code>.
+ It is also proposed to provide additional suffix(es) to specify constants to
+ suit precision lower than that of <code class="computeroutput">float</code>
+ and precision higher than that of <code class="computeroutput">long
+ double</code>.
 

- The objectives are to:
+ Floating-point types with specified widths are expected to significantly improve
+ clarity of code and portability of floating-point calculations. Analogous improvements
+ for integer calculations were recently achieved via standardization of integer
+ types having specified width such as <code class="computeroutput">int8_t</code>,
+ <code class="computeroutput">int16_t</code>, <code class="computeroutput">int32_t</code>,
+ and <code class="computeroutput">int64_t</code>.
+ 
+
+ The main objectives of this proposal are to:
 
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
- Make it easier to use higher-precision.
+ Extend the range of floating-point precision.
 </li>
<li class="listitem">
 Reduce errors in precision.
 </li>
<li class="listitem">
+ Improve clarity of coding.
+ </li>
+<li class="listitem">
 Improve portability, reliability and safety.
 </li>
</ul></div>

Modified: sandbox/precision/libs/precision/doc/html/precision/background.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/background.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/background.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -27,9 +27,12 @@
<a name="precision.background"></a><a class="link" href="background.html" title="Background">Background</a>
</h2></div></div></div>

- C++11 supports floating-point calculations with its built-in types <code class="computeroutput">float</code>, <code class="computeroutput">double</code>,
- and <code class="computeroutput">long double</code>
- as well as imlementations of numerous elementary and transcendental functions.
+ Support for mathematical facilities and specialized number types in C++ is
+ progressing rapidly. Currently, C++11 supports floating-point calculations
+ with its built-in types <code class="computeroutput">float</code>,
+ <code class="computeroutput">double</code>, and <code class="computeroutput">long
+ double</code> as well as imlementations of
+ numerous elementary and transcendental functions.
 

 A variety of higher transcendental functions of pure and applied mathematics
@@ -48,22 +51,26 @@
 above and has become quite widely used.
 

- With the acceptance and release of Boost.Multiprecision
- that provides much higher precision than built-in <code class="computeroutput">long
- double</code> with cpp_dec_float
- employing a variety of backends including the well-established <a href="http://gmplib.org/" target="_top">GNU
+ There is also progress in C++ in the area of multiprecision, including support
+ of multiprecision floating-point numbers. In particular, the acceptance and
+ release of Boost.Multiprecision
+ provides much higher precision than built-in <code class="computeroutput">long
+ double</code> with its cpp_dec_float
+ data type. Boost.Multiprecision
+ has a flexible front-end that employs a variety of backends to implement multiprecision
+ floating-point types including the well-established <a href="http://gmplib.org/" target="_top">GNU
 Multiple Precision Arithmetic Library</a> and <a href="http://www.mpfr.org/" target="_top">GNU
- MPFR library</a> libraries as well as a full open-license backend developed
- from the <a href="http://calgo.acm.org/910.zip" target="_top">e_float (TOMS Algorithm
- 910)</a> library by Christopher Kormanyos and John Maddock.
+ MPFR library</a> libraries as well as a full open-license backend that
+ originates from the <a href="http://calgo.acm.org/910.zip" target="_top">e_float (TOMS
+ Algorithm 910)</a> library by Christopher Kormanyos and John Maddock.
 

 Since Boost.Multiprecision
 and Boost.Math
- work seamlessly, allowing a <code class="computeroutput">float_type
- typedef</code> to be switched from a built-in
- type to hundreds of decimal digits; then all the special functions and distributions
- can be used at any chosen precision.
+ work seamlessly, a <code class="computeroutput">float_type typedef</code> can be used to switch from a built-in
+ type to hundreds of decimal digits. This allows all the special functions and
+ distributions in Boost.Math
+ to be used at any chosen precision.
 

 Other users and domains are finding the need and utility of decimal
@@ -72,14 +79,15 @@
 

 Of course, moving away from hardware supported types to software using C++
- templates carries a small price at compile-time, and a much bigger price at
- runtime.
+ templates carries a small price at compile-time, and potentially a much bigger
+ price at runtime. Nonetheless, the new numerical types have wide ranges of
+ application required in numerous programming domains.
 

 All these development have made C++ much more attractive to the scientific
- and engeering community, especially those needing higher (or lower) precision
- for some (if not all) of the calculations, previously the domain covered by
- computer algebra systems for which the precision can be arbitrary.
+ and engeering community, especially those needing mathematical functions and
+ higher (or lower) precision for some of their calculations. Previously these
+ domains were predominantly covered by computer algebra systems.
 
<div class="footnotes">
 <hr style="width:100; align:left;">

Modified: sandbox/precision/libs/precision/doc/html/precision/introduction.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/introduction.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/introduction.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -7,7 +7,7 @@
<link rel="home" href="../index.html" title="Specific-Width Floating-Point Types">
<link rel="up" href="../index.html" title="Specific-Width Floating-Point Types">
<link rel="prev" href="background.html" title="Background">
-<link rel="next" href="suffixes.html" title="How to specify constants with quad and half precision?">
+<link rel="next" href="thetypes.html" title="The proposed types and potential extensions">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table cellpadding="2" width="100%"><tr>
@@ -20,7 +20,7 @@
</tr></table>
<hr>
<div class="spirit-nav">
-<a accesskey="p" href="background.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="suffixes.html"><img src="../images/next.png" alt="Next"></a>
+<a accesskey="p" href="background.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="thetypes.html"><img src="../images/next.png" alt="Next"></a>
</div>
<div class="section precision_introduction">
<div class="titlepage"><div><div><h2 class="title" style="clear: both">
@@ -37,38 +37,40 @@

 It is, however, emphasized that floating-point adherence to <a href="http://en.wikipedia.org/wiki/IEEE_floating_point" target="_top">IEEE_
 floating-point format</a> is not mandated by the current C++ language standard.
- Nor does the standard enforce specific widths or precisions of the built-in
- types <code class="computeroutput">float</code>, <code class="computeroutput">double</code>,
+ Nor does the standard specify the widths or precisions of its built-in types
+ <code class="computeroutput">float</code>, <code class="computeroutput">double</code>,
 and <code class="computeroutput">long double</code>.
- This can lead to portability problems and reduce reliability and safety.
+ This can lead to portability problems, introduce poor efficiency on cost-sensitive
+ microcontroller architectures, and reduce reliability and safety.
 

- This situation reveals a need for more standard ways to specify precision.
- In addition, it is desirable to extend the precision of existing types to lower
- and higher precisions. The extension to lower precision is expected to simplify
+ This situation reveals a need for a standard way to specify precision. It is
+ also desirable to extend the precision of existing types to both lower and
+ higher precisions. The extension to lower precision is expected to simplify
 and improve efficiency of floating-point implementations on cost-sensitive
 architectures such as small microcontrollers. The extension to higher precision
 is useful for large-scale high-performance numerical calculations and should
- ease the progression to extended precision by providing precision-steps with
- finer granularity.
+ ease the transition to multiprecision by providing built-in types with progressing
+ precision of finer granularity.
 

 All of these improvements should improve portability, reliability, and safety
 of floating-point calculations in C++ by ensuring that the actual precision
 of a floating-point type can be exactly determined both at compile-time as
 well as during the run of a calculation. Strong interest in floating-point
- types with specified widths has, for example, been expressed on the <a href="http://lists.boost.org/Archives/boost/2013/03/201786.php" target="_top">Boost list discussion
- of precise floating-point types</a>.
+ types with specified widths has, for example, recently been expressed on the
+ <a href="http://lists.boost.org/Archives/boost/2013/03/201786.php" target="_top">Boost
+ list discussion of precise floating-point types</a>.
 

- Recent specification of integer types with specified widths in C99, C11, C++11,
- and <a href="http://open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3376.pdf" target="_top">C++
+ Recent specification of integer types having specified widths in C99, C11,
+ C++11, and <a href="http://open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3376.pdf" target="_top">C++
 draft specification</a> has drastically improved integer algorithm portability
 and range.
 

- One example of how specific-size integer types have proven to be essential
- is described by Robert Ramey <a href="http://lists.boost.org/Archives/boost/2002/11/40432.php" target="_top">Usefulness
+ One example of how integer types having specified widths have proven to be
+ essential is described by Robert Ramey <a href="http://lists.boost.org/Archives/boost/2002/11/40432.php" target="_top">Usefulness
 of fixed integer sizes in portability (for Boost serialization library).</a>
 

@@ -120,7 +122,7 @@
</tr></table>
<hr>
<div class="spirit-nav">
-<a accesskey="p" href="background.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="suffixes.html"><img src="../images/next.png" alt="Next"></a>
+<a accesskey="p" href="background.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="thetypes.html"><img src="../images/next.png" alt="Next"></a>
</div>
</body>
</html>

Modified: sandbox/precision/libs/precision/doc/html/precision/precision.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/precision.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/precision.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -36,13 +36,15 @@
 
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
- <code class="computeroutput">float24_t,
+ <code class="computeroutput">float11_t,
+ float24_t,
 float53_t,
 float113_t,
 ...</code>
 </li>
<li class="listitem">
- <code class="computeroutput">float32_t,
+ <code class="computeroutput">float16_t,
+ float32_t,
 float64_t,
 float128_t,
 ...</code>
@@ -50,15 +52,16 @@
</ul></div>

 The first set above is intuitively coined from IEE754:2008.
- It is also consistent with the gist of <code class="computeroutput">std::uint32_t</code>,
- et al in so far as the number of binary digits of significand
- precision is contained within the name of the data type.
+ It is also consistent with the gist of integer types with specified width such
+ as <code class="computeroutput">int64_t</code>, in so far as the
+ number of binary digits of significand precision is contained
+ within the name of the data type.
 

- On the other hand, the second set using the size of the whole type
- may probably seem more intuitive to users. The exact layout and number of significand
- and exponent bits can be confirmed as IEEE754 by checking <code class="computeroutput">std::numeric_limits<type>::is_iec559
- == true</code>.
+ On the other hand, the second set with the size of the whole type
+ contained within the name may be more intuitive to users. The exact layout
+ and number of significand and exponent bits can be confirmed as IEEE754 by
+ checking <code class="computeroutput">std::numeric_limits<type>::is_iec559 == true</code>.
 

 With the availability of Boost.Multprecision, C++ programmers can now easily

Modified: sandbox/precision/libs/precision/doc/html/precision/precision/integerfixedtypes.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/precision/integerfixedtypes.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/precision/integerfixedtypes.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -35,7 +35,7 @@
 
<pre class="programlisting">namespace std
{
- typedef signed integer type int8_t; // optional
+ typedef signed integer type int8_t; // optional
 typedef signed integer type int16_t; // optional
 typedef signed integer type int32_t; // optional
 typedef signed integer type int64_t; // optional

Modified: sandbox/precision/libs/precision/doc/html/precision/precision/newfloattypes.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/precision/newfloattypes.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/precision/newfloattypes.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -27,7 +27,8 @@
<a name="precision.precision.newfloattypes"></a><a class="link" href="newfloattypes.html" title="Proposed new section">Proposed new section</a>
</h3></div></div></div>

- Add the following text to <cstdint>
+ Add a new header <code class="computeroutput"><cstdfloat></code> to the standard library. Add the following
+ text to <cstdint>
 
<div class="note"><table border="0" summary="Note">
<tr>
@@ -38,20 +39,20 @@
 It is not obvious where these <code class="computeroutput">typedef</code>s
 should reside. The obvious place is <code class="computeroutput"><cstdint></code>
 but <code class="computeroutput">int</code> implies integer types.
- (or a new <cstdfloat>?)
+ Here, we prefer the new header <code class="computeroutput"><cstdfloat></code>.
 </td></tr>
</table></div>

- 18.4? Arithmetic types [cstdfloat] (or cstdarith] 18.4.2? Header <cstdfloat>
+ 18.4? Floating-Point Types Having Specified Width 18.4.2? Header <cstdfloat>
 synopsis [cstdfloat.syn]
 
<pre class="programlisting">namespace std {
- typedef signed floating-point type float_16_t; // optional.
- typedef signed floating-point type float_32_t; // optional.
- typedef signed floating-point type float_64_t; // optional.
- typedef signed floating-point type float_80_t; // optional.
- typedef signed floating-point type float_128_t; // optional.
- typedef signed floating-point type float_256_t; // optional.
+ typedef signed floating-point type float16_t; // optional.
+ typedef signed floating-point type float32_t; // optional.
+ typedef signed floating-point type float64_t; // optional.
+ typedef signed floating-point type float80_t; // optional.
+ typedef signed floating-point type float128_t; // optional.
+ typedef signed floating-point type float256_t; // optional.
 typedef signed floating-point type floatmax_t; // optional.

 typedef signed floating-point type float_least16_t; // optional.
@@ -73,15 +74,24 @@
 It is not proposed to make any change to <code class="computeroutput">std::numeric_limits</code>.
 

- It is obviously highly desirable that <code class="computeroutput">std::numeric_limits</code>
- is specialized for all floating-point types. And experience with Boost.Math
- and Boost.Multiprecision
- is that the normal set of trig and others useful functions is also essential
- to make the type useful in real-life.
+ It is, nonethelsee, mandatory that <code class="computeroutput">std::numeric_limits</code>
+ is specialized for all floating-point types. Thus their must be explicit
+ template specializations for all of the optional floating-point types with
+ specified precision that are included in the implementation. This will ensure
+ that programs can use the established <code class="computeroutput">std::numeric_limits<>::is_iec559</code>
+ member to determine if a floating-point type conforms with <a href="http://en.wikipedia.org/wiki/IEEE_floating_point" target="_top">IEEE_
+ floating-point format</a>.
 

- Programs can then use this to determine if a floating-point type is IEEE
- 754 using <code class="computeroutput">std::numeric_limits<>::is_iec559</code>.
+ Experience with Boost.Math
+ and Boost.Multiprecision
+ has shown that the normal set of elementary and transcendental functions
+ (and possibly additional higher transcendental functions) is also essential
+ to make the type useful in real-life computational regimes. Therefore, the
+ implementation must provide support for the mathematical functions in the
+ <code class="computeroutput">std</code> namespace (TBD: Chris
+ list functions) for each of the floating-point types with specified precision
+ included in the implementation.
 
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>

Modified: sandbox/precision/libs/precision/doc/html/precision/references.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/references.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/references.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -6,7 +6,7 @@
<meta name="generator" content="DocBook XSL Stylesheets V1.77.1">
<link rel="home" href="../index.html" title="Specific-Width Floating-Point Types">
<link rel="up" href="../index.html" title="Specific-Width Floating-Point Types">
-<link rel="prev" href="microfpu.html" title="Improvemed safety for microcontrollers with an FPU">
+<link rel="prev" href="microfpu.html" title="Improved efficiency and robustness for microcontrollers">
<link rel="next" href="version_id.html" title="Version Info">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">

Modified: sandbox/precision/libs/precision/doc/html/precision/suffixes.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/suffixes.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/suffixes.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -6,7 +6,7 @@
<meta name="generator" content="DocBook XSL Stylesheets V1.77.1">
<link rel="home" href="../index.html" title="Specific-Width Floating-Point Types">
<link rel="up" href="../index.html" title="Specific-Width Floating-Point Types">
-<link rel="prev" href="introduction.html" title="Introduction">
+<link rel="prev" href="thetypes.html" title="The proposed types and potential extensions">
<link rel="next" href="precision.html" title="Specifying Precision">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
@@ -20,7 +20,7 @@
</tr></table>
<hr>
<div class="spirit-nav">
-<a accesskey="p" href="introduction.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="precision.html"><img src="../images/next.png" alt="Next"></a>
+<a accesskey="p" href="thetypes.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="precision.html"><img src="../images/next.png" alt="Next"></a>
</div>
<div class="section precision_suffixes">
<div class="titlepage"><div><div><h2 class="title" style="clear: both">
@@ -28,61 +28,71 @@
 and half precision?</a>
</h2></div></div></div>

+ The standard specifies that the type of a floating-point literal is double
+ unless explicitly specified by a suffix. The standard continues by specifying
+ that the suffixes <code class="computeroutput">f</code> and <code class="computeroutput">F</code> specify <code class="computeroutput">float</code>,
+ and the suffixes <code class="computeroutput">l</code> and <code class="computeroutput">L</code> specify <code class="computeroutput">long
+ double</code>.
+ 
+
 Recent discussion on extended precision floating-point types in C++ has also
 raised the issue of how to specify constant values with a precision greater
 than <code class="computeroutput">long double</code>,
- now signified by the suffix <code class="computeroutput">L</code>.
- 
-
- One obvious way is to add <code class="computeroutput">Q</code>
- or <code class="computeroutput">q</code> suffixes to signify that
- a constant has at least 128-bits (about 40 decimal digits) of precision.
- 
-
- There may also be a need for 256-bit (about 80 decimal digits) precision, and
- perhaps 512-bits (about 155 decimal digits) precision.
- 
-
- At present, the only way to provide constant values is to use a string to extended-precision
- type conversion.
- 
-
- This <code class="computeroutput">from_string</code> method is
- used for Boost.Math,
- Boost.Multiprecision
- and GCC libquadmath,
- for example.
- 
-
- It would also be useful to have a method of interrogating the size of types,
- similar to that provided by <a href="http://gcc.gnu.org/onlinedocs/cpp/Common-Predefined-Macros.html" target="_top">GCC
- 3.7.2 Common Predefined Macros</a>, for example, <code class="computeroutput">__SIZEOF_LONG_DOUBLE__</code>
- (but is not defined for <code class="computeroutput">__float128</code>
- nor <code class="computeroutput">__float80</code>)
- 
-
- We refer to floating-point types with fixed precision such as 24, 53, 113 or
- more binary significand digits, (and possibly even extending beyond these to
- potential multiprecision types).
- 
-
- These are defined in <a href="http://dx.doi.org/10.1109/IEEESTD.2008.4610935" target="_top">IEEE
- Standard for Floating-point Arithmetic, IEEE Std 754-2008</a>.
- 
-
- There are detailed descriptions at <a href="http://en.wikipedia.org/wiki/IEEE_floating_point" target="_top">IEEE_
- floating-point format</a>, with more detailed descrptions of each type
- at <a href="http://en.wikipedia.org/wiki/Half_precision_floating-point_format" target="_top">IEEE
- half-precision floating-point format</a>, <a href="http://en.wikipedia.org/wiki/Single_precision_floating-point_format" target="_top">IEEE
- single-precision floating-point format</a>, <a href="http://en.wikipedia.org/wiki/Double_precision_floating-point_format" target="_top">IEEE
- double-precision floating-point format</a>, <a href="http://en.wikipedia.org/wiki/Quadruple-precision_floating-point_format" target="_top">Quadruple-precision
- floating-point format</a>, and <a href="http://en.wikipedia.org/wiki/Extended_precision#IEEE_754_extended_precision_formats" target="_top">IEEE
- 754 extended precision formats and x86 80-bit Extended Precision Format</a>
- and these correspond to the proposed types below <code class="computeroutput">float16_t</code>
- ....
- 
-
- TBD by Chris: Suffix for half-precision.
+ now signified by the suffix <code class="computeroutput">L</code>
+ or <code class="computeroutput">l</code>. One possible way is to
+ add <code class="computeroutput">Q</code> or <code class="computeroutput">q</code>
+ suffixes to signify that floating-point literal has quadruple precision.
+ 
+
+ Code using the <code class="computeroutput">Q</code> suffix scheme
+ is shown in the sample below.
+ 
+<pre class="programlisting">#include <cstdfloat>
+
+constexpr std::float128_t pi = 3.1415926535897932384626433832795028841972Q;
+</pre>
+
+ For half-precision floating-point literals, the suffix <code class="computeroutput">H</code>
+ or <code class="computeroutput">h</code> could be used. One potential
+ suffix for octuple-precision floating-point literals is <code class="computeroutput">O</code>
+ or <code class="computeroutput">o</code>.
+ 
+
+ For example,
+ 
+<pre class="programlisting">#include <cstdfloat>
+
+constexpr std::float16_t euler_gamma = 0.577216H;
+</pre>
+
+ Higher precisions also require construction from floating-point literals. As
+ the list of available suffixes dwindles, however, available suffixes might
+ run out and the myriad of suffixes may become confusing. Floating-point literals
+ for precisions higher than quadruple precision, then, might be better served
+ with construction from string literals.
+ 
+
+ An alternative suffix scheme could use hybrid suffixes composed of, say, the
+ letter <code class="computeroutput">F</code> or <code class="computeroutput">f</code>
+ which stands for floating-point, to which the specified width of the type is
+ appended, for example <code class="computeroutput">F16</code>,
+ <code class="computeroutput">F32</code>, <code class="computeroutput">F64</code>,
+ <code class="computeroutput">F128</code>, etc. Code using the
+ <code class="computeroutput">F128</code> suffix scheme is shown
+ in the sample below.
+ 
+<pre class="programlisting">#include <cstdfloat>
+
+constexpr std::float128_t pi = 3.1415926535897932384626433832795028841972F128;
+
+constexpr std::float16_t euler_gamma = 0.577216F16;
+</pre>
+
+ This suffix scheme is unequivocal and it can be easily extended to unlimited
+ precision. On the other hand, it may be difficult for programmers to separate
+ the character part of the suffix from its numerical part when analyzing source
+ code. For example, it is particularly difficult to resolve the suffix in <code class="computeroutput">euler_gamma</code> above, since the <code class="computeroutput">16F16</code> obscures
+ the <code class="computeroutput">F</code>.
 
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
@@ -95,7 +105,7 @@
</tr></table>
<hr>
<div class="spirit-nav">
-<a accesskey="p" href="introduction.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="precision.html"><img src="../images/next.png" alt="Next"></a>
+<a accesskey="p" href="thetypes.html"><img src="../images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../images/home.png" alt="Home"></a><a accesskey="n" href="precision.html"><img src="../images/next.png" alt="Next"></a>
</div>
</body>
</html>

Modified: sandbox/precision/libs/precision/doc/html/precision/version_id.html
==============================================================================
--- sandbox/precision/libs/precision/doc/html/precision/version_id.html (original)
+++ sandbox/precision/libs/precision/doc/html/precision/version_id.html 2013-03-24 05:22:31 EDT (Sun, 24 Mar 2013)
@@ -26,7 +26,7 @@
<a name="precision.version_id"></a><a class="link" href="version_id.html" title="Version Info">Version Info</a>
</h2></div></div></div>

- Last edit to Quickbook file precision.qbk was at 12:20:46 PM on 2013-Mar-23.
+ Last edit to Quickbook file precision.qbk was at 09:21:29 AM on 2013-Mar-24.

<div class="tip"><table border="0" summary="Tip">
<tr>

Modified: sandbox/precision/libs/precision/doc/precision.pdf
==============================================================================
Binary files. No diff available.

Next message: john_at_[hidden]: "[Boost-commit] svn:boost r83537 - in trunk/boost/math: constants special_functions/detail tools"
Previous message: e_float_at_[hidden]: "[Boost-commit] svn:boost r83535 - sandbox/precision/libs/precision/doc"

Date view	Thread view	Subject view	Author view

Boost-Commit list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk