1<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" 2 xml:id="std.numerics" xreflabel="Numerics"> 3<?dbhtml filename="numerics.html"?> 4 5<info><title> 6 Numerics 7 <indexterm><primary>Numerics</primary></indexterm> 8</title> 9 <keywordset> 10 <keyword>ISO C++</keyword> 11 <keyword>library</keyword> 12 </keywordset> 13</info> 14 15 16 17<!-- Sect1 01 : Complex --> 18<section xml:id="std.numerics.complex" xreflabel="complex"><info><title>Complex</title></info> 19<?dbhtml filename="complex.html"?> 20 21 <para> 22 </para> 23 <section xml:id="numerics.complex.processing" xreflabel="complex Processing"><info><title>complex Processing</title></info> 24 25 <para> 26 </para> 27 <para>Using <code>complex<></code> becomes even more comple- er, sorry, 28 <emphasis>complicated</emphasis>, with the not-quite-gratuitously-incompatible 29 addition of complex types to the C language. David Tribble has 30 compiled a list of C++98 and C99 conflict points; his description of 31 C's new type versus those of C++ and how to get them playing together 32 nicely is 33<link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://david.tribble.com/text/cdiffs.htm#C99-complex">here</link>. 34 </para> 35 <para><code>complex<></code> is intended to be instantiated with a 36 floating-point type. As long as you meet that and some other basic 37 requirements, then the resulting instantiation has all of the usual 38 math operators defined, as well as definitions of <code>op<<</code> 39 and <code>op>></code> that work with iostreams: <code>op<<</code> 40 prints <code>(u,v)</code> and <code>op>></code> can read <code>u</code>, 41 <code>(u)</code>, and <code>(u,v)</code>. 42 </para> 43 <para>As an extension to C++11 and for increased compatibility with C, 44 <code><complex.h></code> includes both <code><complex></code> 45 and the C99 <code><complex.h></code> (if the C library provides 46 it). 47 </para> 48 49 </section> 50</section> 51 52<!-- Sect1 02 : Generalized Operations --> 53<section xml:id="std.numerics.generalized_ops" xreflabel="Generalized Ops"><info><title>Generalized Operations</title></info> 54<?dbhtml filename="generalized_numeric_operations.html"?> 55 56 <para> 57 </para> 58 59 <para>There are four generalized functions in the <numeric> header 60 that follow the same conventions as those in <algorithm>. Each 61 of them is overloaded: one signature for common default operations, 62 and a second for fully general operations. Their names are 63 self-explanatory to anyone who works with numerics on a regular basis: 64 </para> 65 <itemizedlist> 66 <listitem><para><code>accumulate</code></para></listitem> 67 <listitem><para><code>inner_product</code></para></listitem> 68 <listitem><para><code>partial_sum</code></para></listitem> 69 <listitem><para><code>adjacent_difference</code></para></listitem> 70 </itemizedlist> 71 <para>Here is a simple example of the two forms of <code>accumulate</code>. 72 </para> 73 <programlisting> 74 int ar[50]; 75 int someval = somefunction(); 76 77 // ...initialize members of ar to something... 78 79 int sum = std::accumulate(ar,ar+50,0); 80 int sum_stuff = std::accumulate(ar,ar+50,someval); 81 int product = std::accumulate(ar,ar+50,1,std::multiplies<int>()); 82 </programlisting> 83 <para>The first call adds all the members of the array, using zero as an 84 initial value for <code>sum</code>. The second does the same, but uses 85 <code>someval</code> as the starting value (thus, <code>sum_stuff == sum + 86 someval</code>). The final call uses the second of the two signatures, 87 and multiplies all the members of the array; here we must obviously 88 use 1 as a starting value instead of 0. 89 </para> 90 <para>The other three functions have similar dual-signature forms. 91 </para> 92 93</section> 94 95<!-- Sect1 03 : Interacting with C --> 96<section xml:id="std.numerics.c" xreflabel="Interacting with C"><info><title>Interacting with C</title></info> 97<?dbhtml filename="numerics_and_c.html"?> 98 99 100 <section xml:id="numerics.c.array" xreflabel="Numerics vs. Arrays"><info><title>Numerics vs. Arrays</title></info> 101 102 103 <para>One of the major reasons why FORTRAN can chew through numbers so well 104 is that it is defined to be free of pointer aliasing, an assumption 105 that C89 is not allowed to make, and neither is C++98. C99 adds a new 106 keyword, <code>restrict</code>, to apply to individual pointers. The 107 C++ solution is contained in the library rather than the language 108 (although many vendors can be expected to add this to their compilers 109 as an extension). 110 </para> 111 <para>That library solution is a set of two classes, five template classes, 112 and "a whole bunch" of functions. The classes are required 113 to be free of pointer aliasing, so compilers can optimize the 114 daylights out of them the same way that they have been for FORTRAN. 115 They are collectively called <code>valarray</code>, although strictly 116 speaking this is only one of the five template classes, and they are 117 designed to be familiar to people who have worked with the BLAS 118 libraries before. 119 </para> 120 121 </section> 122 123 <section xml:id="numerics.c.c99" xreflabel="C99"><info><title>C99</title></info> 124 125 126 <para>In addition to the other topics on this page, we'll note here some 127 of the C99 features that appear in libstdc++. 128 </para> 129 <para>The C99 features depend on the <code>--enable-c99</code> configure flag. 130 This flag is already on by default, but it can be disabled by the 131 user. Also, the configuration machinery will disable it if the 132 necessary support for C99 (e.g., header files) cannot be found. 133 </para> 134 <para>As of GCC 3.0, C99 support includes classification functions 135 such as <code>isnormal</code>, <code>isgreater</code>, 136 <code>isnan</code>, etc. 137 The functions used for 'long long' support such as <code>strtoll</code> 138 are supported, as is the <code>lldiv_t</code> typedef. Also supported 139 are the wide character functions using 'long long', like 140 <code>wcstoll</code>. 141 </para> 142 143 </section> 144</section> 145 146</chapter> 147