1<html> 2<head> 3<meta http-equiv="Content-Type" content="text/html; charset=UTF-8"> 4<title>Benchmarks</title> 5<link rel="stylesheet" href="../../../../../doc/src/boostbook.css" type="text/css"> 6<meta name="generator" content="DocBook XSL Stylesheets V1.79.1"> 7<link rel="home" href="../index.html" title="Chapter 1. Boost.Histogram"> 8<link rel="up" href="../index.html" title="Chapter 1. Boost.Histogram"> 9<link rel="prev" href="guide.html" title="User guide"> 10<link rel="next" href="concepts.html" title="Concepts"> 11</head> 12<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF"> 13<table cellpadding="2" width="100%"><tr> 14<td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../../../boost.png"></td> 15<td align="center"><a href="../../../../../index.html">Home</a></td> 16<td align="center"><a href="../../../../libraries.htm">Libraries</a></td> 17<td align="center"><a href="http://www.boost.org/users/people.html">People</a></td> 18<td align="center"><a href="http://www.boost.org/users/faq.html">FAQ</a></td> 19<td align="center"><a href="../../../../../more/index.htm">More</a></td> 20</tr></table> 21<hr> 22<div class="spirit-nav"> 23<a accesskey="p" href="guide.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="concepts.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a> 24</div> 25<div class="section"> 26<div class="titlepage"><div><div><h2 class="title" style="clear: both"> 27<a name="histogram.benchmarks"></a><a class="link" href="benchmarks.html" title="Benchmarks">Benchmarks</a> 28</h2></div></div></div> 29<div class="toc"><dl class="toc"> 30<dt><span class="section"><a href="benchmarks.html#histogram.benchmarks.fill_performance">Fill performance</a></span></dt> 31<dt><span class="section"><a href="benchmarks.html#histogram.benchmarks.iteration_performance">Iteration 32 performance</a></span></dt> 33</dl></div> 34<p> 35 The library is designed to be fast. When configured correctly, it is one of 36 the fastest libraries on the market. If you find a library that is faster than 37 Boost.Histogram, please submit an issue on Github. We care about performance. 38 </p> 39<p> 40 That being said, the time spend in filling the histogram is usually not the 41 bottleneck of an application. Only in processing of really large data sets 42 the performance of the histogram can be important. 43 </p> 44<p> 45 All benchmarks are compiled on a laptop with a 2,9 GHz Intel Core i5 processor 46 with Apple LLVM (clang-1001.0.46.4) and the flags <code class="computeroutput"><span class="special">-</span><span class="identifier">DNDEBUG</span> <span class="special">-</span><span class="identifier">O3</span> <span class="special">-</span><span class="identifier">funsafe</span><span class="special">-</span><span class="identifier">math</span><span class="special">-</span><span class="identifier">optimizations</span></code>. Adding <code class="computeroutput"><span class="special">-</span><span class="identifier">fno</span><span class="special">-</span><span class="identifier">exceptions</span> 47 <span class="special">-</span><span class="identifier">fno</span><span class="special">-</span><span class="identifier">rtti</span></code> would 48 increase the Boost.Histogram performance by another (10-20) %, but this is 49 not done here out of fairness, since the ROOT histograms do not compile with 50 these options. 51 </p> 52<div class="section"> 53<div class="titlepage"><div><div><h3 class="title"> 54<a name="histogram.benchmarks.fill_performance"></a><a class="link" href="benchmarks.html#histogram.benchmarks.fill_performance" title="Fill performance">Fill performance</a> 55</h3></div></div></div> 56<p> 57 The fill performance of different configurations of Boost.Histogram are compared 58 with histogram classes and functions from other libraries. Random numbers 59 from a uniform and a normal distribution are filled into histograms with 60 1, 2, 3, and 6 axes. 100 bins per axis are used for 1, 2, 3 axes. 10 bins 61 per axis for the case with 6 axes. The histogram are filled with the call 62 operator <code class="computeroutput"><span class="keyword">operator</span><span class="special">()</span></code> 63 and the more efficient <code class="computeroutput"><span class="identifier">fill</span></code>-method, 64 which accepts large chunks of values at once. The GSL offers only 1D and 65 2D histograms, so there are no entries for the higher dimensional benchmarks. 66 Raw timing results are converted to average number of CPU cycles used per 67 input value. 68 </p> 69<p> 70 There is one bar for each benchmark and the upper end has a hatched part. 71 The full bar is the result when the histograms are filled with random normally 72 distributed data that falls outside of the axis domain in about 10 % of the 73 cases. This makes the branch predictors in the CPU fail every now and then, 74 which degrades performance. The bar without the hatched part is the result 75 when the histograms are filled with uniform random numbers which are always 76 inside the axis range. 77 </p> 78<p> 79 <span class="inlinemediaobject"><object type="image/svg+xml" data="../../fill_performance.svg" width="630" height="540"></object></span> 80 </p> 81<div class="variablelist"> 82<p class="title"><b></b></p> 83<dl class="variablelist"> 84<dt><span class="term">ROOT 6</span></dt> 85<dd><p> 86 <a href="https://root.cern.ch" target="_top">ROOT classes</a> (<code class="computeroutput"><span class="identifier">TH1I</span></code> for 1D, <code class="computeroutput"><span class="identifier">TH2I</span></code> 87 for 2D, <code class="computeroutput"><span class="identifier">TH3I</span></code> for 3D 88 and <code class="computeroutput"><span class="identifier">THnI</span></code> for 6D) 89 </p></dd> 90<dt><span class="term">GSL</span></dt> 91<dd><p> 92 <a href="https://www.gnu.org/software/gsl/doc/html/histogram.html" target="_top">GSL 93 histograms</a> for 1D and 2D 94 </p></dd> 95<dt><span class="term">boost-sta</span></dt> 96<dd><p> 97 Histogram with <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">tuple</span><span class="special"><</span><span class="identifier">axis</span><span class="special">::</span><span class="identifier">regular</span><span class="special"><>></span></code> and <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special"><</span><span class="keyword">int</span><span class="special">></span></code> storage 98 </p></dd> 99<dt><span class="term">boost-dyn</span></dt> 100<dd><p> 101 Histogram with <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special"><</span><span class="identifier">axis</span><span class="special">::</span><span class="identifier">variant</span><span class="special"><</span><span class="identifier">axis</span><span class="special">::</span><span class="identifier">regular</span><span class="special"><>>></span></code> and <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special"><</span><span class="keyword">int</span><span class="special">></span></code> storage 102 </p></dd> 103</dl> 104</div> 105<p> 106 Boost.Histogram is faster than other libraries. Simultaneously, it is much 107 more flexible, since the axis and storage types can be customized. When 108 <code class="computeroutput"><span class="keyword">operator</span><span class="special">()</span></code> 109 is used, a histogram with compile-time configured axes (boost-sta-...) is 110 always faster than the equivalent alternatives from other libraries. The 111 histogram with run-time configured axes (boost-dyn-...) is comparable or 112 slower than other libraries, but offers a run-time flexibility that the alternatives 113 do not. If the <code class="computeroutput"><span class="identifier">fill</span></code> method 114 is used, filling either type of histogram is much faster (up to a factor 115 6) than filling histograms in other libraries, and the performance difference 116 between compile-time and run-time configured axes is mostly vanishes. 117 </p> 118</div> 119<div class="section"> 120<div class="titlepage"><div><div><h3 class="title"> 121<a name="histogram.benchmarks.iteration_performance"></a><a class="link" href="benchmarks.html#histogram.benchmarks.iteration_performance" title="Iteration performance">Iteration 122 performance</a> 123</h3></div></div></div> 124<p> 125 Boost.Histogram provides the <code class="computeroutput"><a class="link" href="../boost/histogram/indexed.html" title="Function template indexed">boost::histogram::indexed</a></code> 126 range generator for convenient iteration over the histogram cells. Using 127 the range generator is very convenient and it is faster than by writing nested 128 for-loops. 129 </p> 130<pre class="programlisting"><span class="comment">// nested for loops over 2d histogram</span> 131<span class="keyword">for</span> <span class="special">(</span><span class="keyword">int</span> <span class="identifier">i</span> <span class="special">=</span> <span class="number">0</span><span class="special">;</span> <span class="identifier">i</span> <span class="special"><</span> <span class="identifier">h</span><span class="special">.</span><span class="identifier">axis</span><span class="special">(</span><span class="number">0</span><span class="special">).</span><span class="identifier">size</span><span class="special">();</span> <span class="special">++</span><span class="identifier">i</span><span class="special">)</span> <span class="special">{</span> 132 <span class="keyword">for</span> <span class="special">(</span><span class="keyword">int</span> <span class="identifier">j</span> <span class="special">=</span> <span class="number">0</span><span class="special">;</span> <span class="identifier">j</span> <span class="special"><</span> <span class="identifier">h</span><span class="special">.</span><span class="identifier">axis</span><span class="special">(</span><span class="number">1</span><span class="special">).</span><span class="identifier">size</span><span class="special">();</span> <span class="special">++</span><span class="identifier">j</span><span class="special">)</span> <span class="special">{</span> 133 <span class="identifier">std</span><span class="special">::</span><span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">i</span> <span class="special"><<</span> <span class="string">" "</span> <span class="special"><<</span> <span class="identifier">j</span> <span class="special"><<</span> <span class="string">" "</span> <span class="special"><<</span> <span class="identifier">h</span><span class="special">.</span><span class="identifier">at</span><span class="special">(</span><span class="identifier">i</span><span class="special">,</span> <span class="identifier">j</span><span class="special">)</span> <span class="special"><<</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">endl</span><span class="special">;</span> 134 <span class="special">}</span> 135<span class="special">}</span> 136 137<span class="comment">// same, with indexed range generator</span> 138<span class="keyword">for</span> <span class="special">(</span><span class="keyword">auto</span><span class="special">&&</span> <span class="identifier">x</span> <span class="special">:</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">histogram</span><span class="special">::</span><span class="identifier">indexed</span><span class="special">(</span><span class="identifier">h</span><span class="special">))</span> <span class="special">{</span> 139 <span class="identifier">std</span><span class="special">::</span><span class="identifier">cout</span> <span class="special"><<</span> <span class="identifier">x</span><span class="special">.</span><span class="identifier">index</span><span class="special">(</span><span class="number">0</span><span class="special">)</span> <span class="special"><<</span> <span class="string">" "</span> <span class="special"><<</span> <span class="identifier">x</span><span class="special">.</span><span class="identifier">index</span><span class="special">(</span><span class="number">1</span><span class="special">)</span> <span class="special"><<</span> <span class="string">" "</span> <span class="special"><<</span> <span class="special">*</span><span class="identifier">x</span> <span class="special"><<</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">endl</span><span class="special">;</span> 140<span class="special">}</span> 141</pre> 142<p> 143 The access time per bin is compared for these two iteration strategies for 144 histograms that hold the axes in a <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">tuple</span></code> 145 (tuple), in a <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span></code> (vector), and in a <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special"><</span><span class="identifier">boost</span><span class="special">::</span><span class="identifier">histogram</span><span class="special">::</span><span class="identifier">axis</span><span class="special">::</span><span class="identifier">variant</span><span class="special">></span></code> (vector of variants). The access time 146 per bin is measured for axis with 4 to 128 bins per axis. 147 </p> 148<p> 149 <span class="inlinemediaobject"><object type="image/svg+xml" data="../../iteration_performance.svg" width="900" height="450"></object></span> 150 </p> 151</div> 152</div> 153<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr> 154<td align="left"></td> 155<td align="right"><div class="copyright-footer">Copyright © 2016-2019 Hans 156 Dembinski<p> 157 Distributed under the Boost Software License, Version 1.0. (See accompanying 158 file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) 159 </p> 160</div></td> 161</tr></table> 162<hr> 163<div class="spirit-nav"> 164<a accesskey="p" href="guide.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../index.html"><img src="../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="concepts.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a> 165</div> 166</body> 167</html> 168
