3 <meta http-equiv="Content-Type" content="text/html; charset=US-ASCII">
4 <title>Comparisons to Other Open Source Libraries</title>
5 <link rel="stylesheet" href="../math.css" type="text/css">
6 <meta name="generator" content="DocBook XSL Stylesheets V1.78.1">
7 <link rel="home" href="../index.html" title="Math Toolkit 2.1.0">
8 <link rel="up" href="../perf.html" title="Chapter 15. Performance">
9 <link rel="prev" href="tuning.html" title="Performance Tuning Macros">
10 <link rel="next" href="perf_test_app.html" title="The Performance Test Application">
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="../../../../../libs/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>
22 <div class="spirit-nav">
23 <a accesskey="p" href="tuning.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../perf.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="perf_test_app.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a>
26 <div class="titlepage"><div><div><h2 class="title" style="clear: both">
27 <a name="math_toolkit.comparisons"></a><a class="link" href="comparisons.html" title="Comparisons to Other Open Source Libraries">Comparisons to Other Open Source
29 </h2></div></div></div>
31 We've run our performance tests both for our own code, and against other open
32 source implementations of the same functions. The results are presented below
33 to give you a rough idea of how they all compare.
35 <div class="caution"><table border="0" summary="Caution">
37 <td rowspan="2" align="center" valign="top" width="25"><img alt="[Caution]" src="../../../../../doc/src/images/caution.png"></td>
38 <th align="left">Caution</th>
40 <tr><td align="left" valign="top"><p>
41 You should exercise extreme caution when interpreting these results, relative
42 performance may vary by platform, the tests use data that gives good code
43 coverage of <span class="emphasis"><em>our</em></span> code, but which may skew the results
44 towards the corner cases. Finally, remember that different libraries make
45 different choices with regard to performance verses numerical stability.
49 <a name="math_toolkit.comparisons.h0"></a>
50 <span class="phrase"><a name="math_toolkit.comparisons.comparison_to_gsl_1_13_and_cephe"></a></span><a class="link" href="comparisons.html#math_toolkit.comparisons.comparison_to_gsl_1_13_and_cephe">Comparison
51 to GSL-1.13 and Cephes</a>
54 All the results were measured on a 2.0GHz Intel T5800 Core 2 Duo, 4Gb RAM,
55 Windows Vista machine, with the test program compiled with Microsoft Visual
56 C++ 2009 using the /Ox option.
58 <div class="informaltable"><table class="table">
91 <a class="link" href="powers/cbrt.html" title="cbrt">cbrt</a>
97 <p><span class="bold"><strong>1.00</strong></span></p>
111 <p><span class="bold"><strong>1.00</strong></span></p>
121 <a class="link" href="powers/log1p.html" title="log1p">log1p</a>
127 <p><span class="bold"><strong>1.00</strong></span></p>
136 <p><span class="bold"><strong>1.00</strong></span></p>
145 <p><span class="bold"><strong>1.00</strong></span></p>
155 <a class="link" href="powers/expm1.html" title="expm1">expm1</a>
161 <p><span class="bold"><strong>1.00</strong></span></p>
170 <p><span class="bold"><strong>1.00</strong></span></p>
179 <p><span class="bold"><strong>1.00</strong></span></p>
189 <a class="link" href="sf_gamma/tgamma.html" title="Gamma">tgamma</a>
213 <p><span class="bold"><strong>1.00</strong></span></p>
223 <a class="link" href="sf_gamma/lgamma.html" title="Log Gamma">lgamma</a>
247 <p><span class="bold"><strong>1.00</strong></span></p>
257 <a class="link" href="sf_erf/error_function.html" title="Error Functions">erf</a> and
258 <a class="link" href="sf_erf/error_function.html" title="Error Functions">erfc</a>
264 <p><span class="bold"><strong>1.00</strong></span></p>
273 <p><span class="bold"><strong>1.00</strong></span></p>
282 <p><span class="bold"><strong>1.00</strong></span></p>
292 <a class="link" href="sf_gamma/igamma.html" title="Incomplete Gamma Functions">gamma_p</a> and
293 <a class="link" href="sf_gamma/igamma.html" title="Incomplete Gamma Functions">gamma_q</a>
299 <p><span class="bold"><strong>1.00</strong></span></p>
327 <a class="link" href="sf_gamma/igamma_inv.html" title="Incomplete Gamma Function Inverses">gamma_p_inv</a>
328 and <a class="link" href="sf_gamma/igamma_inv.html" title="Incomplete Gamma Function Inverses">gamma_q_inv</a>
334 <p><span class="bold"><strong>1.00</strong></span></p>
347 +INF <a href="#ftn.math_toolkit.comparisons.f0" class="footnote" name="math_toolkit.comparisons.f0"><sup class="footnote">[1]</sup></a>
354 <a class="link" href="sf_beta/ibeta_function.html" title="Incomplete Beta Functions">ibeta</a>
355 and <a class="link" href="sf_beta/ibeta_function.html" title="Incomplete Beta Functions">ibetac</a>
361 <p><span class="bold"><strong>1.00</strong></span></p>
389 <a class="link" href="sf_beta/ibeta_inv_function.html" title="The Incomplete Beta Function Inverses">ibeta_inv</a>
390 and <a class="link" href="sf_beta/ibeta_inv_function.html" title="The Incomplete Beta Function Inverses">ibetac_inv</a>
396 <p><span class="bold"><strong>1.00</strong></span></p>
420 <a class="link" href="bessel/bessel_first.html" title="Bessel Functions of the First and Second Kinds">cyl_bessel_j</a>
426 <p>17.89<a href="#ftn.math_toolkit.comparisons.f1" class="footnote" name="math_toolkit.comparisons.f1"><sup class="footnote">[2]</sup></a></p>
435 <p><span class="bold"><strong>1.00</strong></span></p>
444 <p><span class="bold"><strong>1.00</strong></span></p>
454 <a class="link" href="bessel/mbessel.html" title="Modified Bessel Functions of the First and Second Kinds">cyl_bessel_i</a>
460 <p><span class="bold"><strong>1.00</strong></span></p>
469 <p><span class="bold"><strong>1.00</strong></span></p>
478 <p><span class="bold"><strong>1.00</strong></span></p>
488 <a class="link" href="bessel/mbessel.html" title="Modified Bessel Functions of the First and Second Kinds">cyl_bessel_k</a>
494 <p><span class="bold"><strong>1.00</strong></span></p>
503 <p><span class="bold"><strong>1.00</strong></span></p>
518 <a class="link" href="bessel/bessel_first.html" title="Bessel Functions of the First and Second Kinds">cyl_neumann</a>
524 <p><span class="bold"><strong>1.00</strong></span></p>
533 <p><span class="bold"><strong>1.00</strong></span></p>
542 <p><span class="bold"><strong>1.00</strong></span></p>
550 <tbody class="footnotes"><tr><td colspan="4">
551 <div id="ftn.math_toolkit.comparisons.f0" class="footnote"><p><a href="#math_toolkit.comparisons.f0" class="para"><sup class="para">[1] </sup></a>
552 Cephes gets stuck in an infinite loop while trying to execute our
555 <div id="ftn.math_toolkit.comparisons.f1" class="footnote"><p><a href="#math_toolkit.comparisons.f1" class="para"><sup class="para">[2] </sup></a>
556 The performance here is dominated by a few cases where the parameters
557 grow very large: faster asymptotic expansions are available, but
558 are of limited (or even frankly terrible) precision. The same issue
559 effects all of our Bessel function implementations, but doesn't
560 necessarily show in the current performance data. More investigation
566 <a name="math_toolkit.comparisons.h1"></a>
567 <span class="phrase"><a name="math_toolkit.comparisons.comparison_to_the_r_and_dcdflib_"></a></span><a class="link" href="comparisons.html#math_toolkit.comparisons.comparison_to_the_r_and_dcdflib_">Comparison
568 to the R and DCDFLIB Statistical Libraries on Windows</a>
571 All the results were measured on a 2.0GHz Intel T5800 Core 2 Duo, 4Gb RAM,
572 Windows Vista machine, with the test program compiled with Microsoft Visual
573 C++ 2009, and R-2.9.2 compiled in "standalone mode" with MinGW-4.3
574 (R-2.9.2 appears not to be buildable with Visual C++).
577 <a name="math_toolkit.comparisons.a_comparison_to_the_r_statistica"></a><p class="title"><b>Table 15.5. A Comparison to the R Statistical Library on Windows XP</b></p>
578 <div class="table-contents"><table class="table" summary="A Comparison to the R Statistical Library on Windows XP">
611 <a class="link" href="dist_ref/dists/beta_dist.html" title="Beta Distribution">Beta Distribution</a>
627 <p><span class="bold"><strong>1.00</strong></span></p>
646 <a class="link" href="dist_ref/dists/beta_dist.html" title="Beta Distribution">Beta Distribution</a>
653 <p><span class="bold"><strong>1.00</strong></span></p>
662 <p>67.66<a href="#ftn.math_toolkit.comparisons.f2" class="footnote" name="math_toolkit.comparisons.f2"><sup class="footnote">[1]</sup></a></p>
681 <a class="link" href="dist_ref/dists/binomial_dist.html" title="Binomial Distribution">Binomial
706 <p><span class="bold"><strong>1.00</strong></span></p>
716 <a class="link" href="dist_ref/dists/binomial_dist.html" title="Binomial Distribution">Binomial
717 Distribution</a> Quantile
723 <p><span class="bold"><strong>1.00</strong></span></p>
751 <a class="link" href="dist_ref/dists/cauchy_dist.html" title="Cauchy-Lorentz Distribution">Cauchy Distribution</a>
758 <p><span class="bold"><strong>1.00</strong></span></p>
782 <a class="link" href="dist_ref/dists/cauchy_dist.html" title="Cauchy-Lorentz Distribution">Cauchy Distribution</a>
789 <p><span class="bold"><strong>1.00</strong></span></p>
798 <p><span class="bold"><strong>1.00</strong></span></p>
813 <a class="link" href="dist_ref/dists/chi_squared_dist.html" title="Chi Squared Distribution">Chi
814 Squared Distribution</a> CDF
838 <p><span class="bold"><strong>1.00</strong></span></p>
848 <a class="link" href="dist_ref/dists/chi_squared_dist.html" title="Chi Squared Distribution">Chi
849 Squared Distribution</a> Quantile
855 <p><span class="bold"><strong>1.00</strong></span></p>
883 <a class="link" href="dist_ref/dists/exp_dist.html" title="Exponential Distribution">Exponential
890 <p><span class="bold"><strong>1.00</strong></span></p>
914 <a class="link" href="dist_ref/dists/exp_dist.html" title="Exponential Distribution">Exponential
915 Distribution</a> Quantile
930 <p><span class="bold"><strong>1.00</strong></span></p>
945 <a class="link" href="dist_ref/dists/f_dist.html" title="F Distribution">Fisher F Distribution</a>
952 <p><span class="bold"><strong>1.00</strong></span></p>
980 <a class="link" href="dist_ref/dists/f_dist.html" title="F Distribution">Fisher F Distribution</a>
987 <p><span class="bold"><strong>1.00</strong></span></p>
1007 <p>(2.205e-005s)</p>
1015 <a class="link" href="dist_ref/dists/gamma_dist.html" title="Gamma (and Erlang) Distribution">Gamma Distribution</a>
1024 <p>(6.240e-007s)</p>
1033 <p>(1.279e-006s)</p>
1040 <p><span class="bold"><strong>1.00</strong></span></p>
1042 <p>(4.111e-007s)</p>
1050 <a class="link" href="dist_ref/dists/gamma_dist.html" title="Gamma (and Erlang) Distribution">Gamma Distribution</a>
1059 <p>(2.179e-006s)</p>
1068 <p>(1.102e-005s)</p>
1075 <p><span class="bold"><strong>1.00</strong></span></p>
1077 <p>(1.764e-006s)</p>
1085 <a class="link" href="dist_ref/dists/hypergeometric_dist.html" title="Hypergeometric Distribution">hypergeometric
1086 Distribution</a> CDF
1092 <p>3.60<a href="#ftn.math_toolkit.comparisons.f3" class="footnote" name="math_toolkit.comparisons.f3"><sup class="footnote">[2]</sup></a></p>
1094 <p>(5.987e-007s)</p>
1101 <p><span class="bold"><strong>1.00</strong></span></p>
1103 <p>(1.665e-007s)</p>
1116 <a class="link" href="dist_ref/dists/hypergeometric_dist.html" title="Hypergeometric Distribution">hypergeometric
1117 Distribution</a> Quantile
1123 <p><span class="bold"><strong>1.00</strong></span></p>
1125 <p>(5.684e-007s)</p>
1134 <p>(2.004e-006s)</p>
1147 <a class="link" href="dist_ref/dists/logistic_dist.html" title="Logistic Distribution">Logistic
1148 Distribution</a> CDF
1154 <p><span class="bold"><strong>1.00</strong></span></p>
1156 <p>(1.714e-007s)</p>
1165 <p>(8.984e-007s)</p>
1178 <a class="link" href="dist_ref/dists/logistic_dist.html" title="Logistic Distribution">Logistic
1179 Distribution</a> Quantile
1187 <p>(2.084e-007s)</p>
1194 <p><span class="bold"><strong>1.00</strong></span></p>
1196 <p>(2.043e-007s)</p>
1209 <a class="link" href="dist_ref/dists/lognormal_dist.html" title="Log Normal Distribution">Log-normal
1210 Distribution</a> CDF
1216 <p><span class="bold"><strong>1.00</strong></span></p>
1218 <p>(3.579e-007s)</p>
1227 <p>(5.332e-007s)</p>
1240 <a class="link" href="dist_ref/dists/lognormal_dist.html" title="Log Normal Distribution">Log-normal
1241 Distribution</a> Quantile
1247 <p><span class="bold"><strong>1.00</strong></span></p>
1249 <p>(9.622e-007s)</p>
1258 <p>(1.507e-006s)</p>
1271 <a class="link" href="dist_ref/dists/negative_binomial_dist.html" title="Negative Binomial Distribution">Negative
1272 Binomial Distribution</a> CDF
1278 <p><span class="bold"><strong>1.00</strong></span></p>
1280 <p>(6.227e-007s)</p>
1289 <p>(1.403e-006s)</p>
1298 <p>(1.378e-006s)</p>
1306 <a class="link" href="dist_ref/dists/negative_binomial_dist.html" title="Negative Binomial Distribution">Negative
1307 Binomial Distribution</a> Quantile
1313 <p><span class="bold"><strong>1.00</strong></span></p>
1315 <p>(8.594e-006s)</p>
1322 <p>43.43<a href="#ftn.math_toolkit.comparisons.f4" class="footnote" name="math_toolkit.comparisons.f4"><sup class="footnote">[3]</sup></a></p>
1324 <p>(3.732e-004s)</p>
1333 <p>(2.994e-005s)</p>
1341 <a class="link" href="dist_ref/dists/nc_chi_squared_dist.html" title="Noncentral Chi-Squared Distribution">Noncentral
1342 Chi Squared Distribution</a> CDF
1350 <p>(3.926e-006s)</p>
1359 <p>(1.450e-004s)</p>
1366 <p><span class="bold"><strong>1.00</strong></span></p>
1368 <p>(1.814e-006s)</p>
1376 <a class="link" href="dist_ref/dists/nc_chi_squared_dist.html" title="Noncentral Chi-Squared Distribution">Noncentral
1377 Chi Squared Distribution</a> Quantile
1385 <p>(3.393e-004s)</p>
1392 <p>393.90<a href="#ftn.math_toolkit.comparisons.f5" class="footnote" name="math_toolkit.comparisons.f5"><sup class="footnote">[4]</sup></a></p>
1394 <p>(2.673e-002s)</p>
1401 <p><span class="bold"><strong>1.00</strong></span></p>
1403 <p>(6.786e-005s)</p>
1411 <a class="link" href="dist_ref/dists/nc_f_dist.html" title="Noncentral F Distribution">Noncentral
1412 F Distribution</a> CDF
1420 <p>(1.128e-005s)</p>
1427 <p><span class="bold"><strong>1.00</strong></span></p>
1429 <p>(7.087e-006s)</p>
1436 <p><span class="bold"><strong>1.00</strong></span></p>
1438 <p>(4.274e-006s)</p>
1446 <a class="link" href="dist_ref/dists/nc_f_dist.html" title="Noncentral F Distribution">Noncentral
1447 F Distribution</a> Quantile
1453 <p><span class="bold"><strong>1.00</strong></span></p>
1455 <p>(4.750e-004s)</p>
1464 <p>(7.681e-004s)</p>
1471 <p><span class="bold"><strong>1.00</strong></span></p>
1473 <p>(4.274e-006s)</p>
1481 <a class="link" href="dist_ref/dists/nc_t_dist.html" title="Noncentral T Distribution">noncentral
1482 T distribution</a> CDF
1490 <p>(1.852e-005s)</p>
1497 <p><span class="bold"><strong>1.00</strong></span></p>
1499 <p>(5.436e-006s)</p>
1512 <a class="link" href="dist_ref/dists/nc_t_dist.html" title="Noncentral T Distribution">noncentral
1513 T distribution</a> Quantile
1521 <p>(5.768e-004s)</p>
1528 <p><span class="bold"><strong>1.00</strong></span><a href="#ftn.math_toolkit.comparisons.f6" class="footnote" name="math_toolkit.comparisons.f6"><sup class="footnote">[5]</sup></a></p>
1530 <p>(4.411e-004s)</p>
1543 <a class="link" href="dist_ref/dists/normal_dist.html" title="Normal (Gaussian) Distribution">Normal Distribution</a>
1550 <p><span class="bold"><strong>1.00</strong></span></p>
1552 <p>(8.373e-008s)</p>
1561 <p>(1.409e-007s)</p>
1570 <p>(5.029e-007s)</p>
1578 <a class="link" href="dist_ref/dists/normal_dist.html" title="Normal (Gaussian) Distribution">Normal Distribution</a>
1587 <p>(1.521e-007s)</p>
1594 <p><span class="bold"><strong>1.00</strong></span></p>
1596 <p>(1.182e-007s)</p>
1605 <p>(1.283e-006s)</p>
1613 <a class="link" href="dist_ref/dists/poisson_dist.html" title="Poisson Distribution">Poisson
1614 Distribution</a> CDF
1622 <p>(5.193e-007s)</p>
1631 <p>(1.314e-006s)</p>
1638 <p><span class="bold"><strong>1.00</strong></span></p>
1640 <p>(4.410e-007s)</p>
1648 <a class="link" href="dist_ref/dists/poisson_dist.html" title="Poisson Distribution">Poisson
1655 <p><span class="bold"><strong>1.00</strong></span></p>
1657 <p>(1.203e-006s)</p>
1666 <p>(2.642e-006s)</p>
1675 <p>(9.457e-006s)</p>
1683 <a class="link" href="dist_ref/dists/students_t_dist.html" title="Students t Distribution">Students
1684 t Distribution</a> CDF
1690 <p><span class="bold"><strong>1.00</strong></span></p>
1692 <p>(8.655e-007s)</p>
1701 <p>(9.166e-007s)</p>
1710 <p>(8.999e-007s)</p>
1718 <a class="link" href="dist_ref/dists/students_t_dist.html" title="Students t Distribution">Students
1719 t Distribution</a> Quantile
1725 <p><span class="bold"><strong>1.00</strong></span></p>
1727 <p>(2.294e-006s)</p>
1736 <p>(3.131e-006s)</p>
1745 <p>(1.106e-005s)</p>
1753 <a class="link" href="dist_ref/dists/weibull_dist.html" title="Weibull Distribution">Weibull
1754 Distribution</a> CDF
1760 <p><span class="bold"><strong>1.00</strong></span></p>
1762 <p>(1.865e-007s)</p>
1771 <p>(4.341e-007s)</p>
1784 <a class="link" href="dist_ref/dists/weibull_dist.html" title="Weibull Distribution">Weibull
1785 Distribution</a> Quantile
1791 <p><span class="bold"><strong>1.00</strong></span></p>
1793 <p>(3.608e-007s)</p>
1802 <p>(4.410e-007s)</p>
1813 <tbody class="footnotes"><tr><td colspan="4">
1814 <div id="ftn.math_toolkit.comparisons.f2" class="footnote"><p><a href="#math_toolkit.comparisons.f2" class="para"><sup class="para">[1] </sup></a>
1815 There are a small number of our test cases where the R library
1816 fails to converge on a result: these tend to dominate the performance
1819 <div id="ftn.math_toolkit.comparisons.f3" class="footnote"><p><a href="#math_toolkit.comparisons.f3" class="para"><sup class="para">[2] </sup></a>
1820 This result is somewhat misleading: for small values of the parameters
1821 there is virtually no difference between the two libraries, but
1822 for large values the Boost implementation is <span class="emphasis"><em>much</em></span>
1823 slower, albeit with much improved precision.
1825 <div id="ftn.math_toolkit.comparisons.f4" class="footnote"><p><a href="#math_toolkit.comparisons.f4" class="para"><sup class="para">[3] </sup></a>
1826 The R library appears to use a linear-search strategy, that can
1827 perform very badly in a small number of pathological cases, but
1828 may or may not be more efficient in "typical" cases
1830 <div id="ftn.math_toolkit.comparisons.f5" class="footnote"><p><a href="#math_toolkit.comparisons.f5" class="para"><sup class="para">[4] </sup></a>
1831 There are a small number of our test cases where the R library
1832 fails to converge on a result: these tend to dominate the performance
1835 <div id="ftn.math_toolkit.comparisons.f6" class="footnote"><p><a href="#math_toolkit.comparisons.f6" class="para"><sup class="para">[5] </sup></a>
1836 There are a small number of our test cases where the R library
1837 fails to converge on a result: these tend to dominate the performance
1843 <br class="table-break"><h4>
1844 <a name="math_toolkit.comparisons.h2"></a>
1845 <span class="phrase"><a name="math_toolkit.comparisons.comparison_to_the_r_statistical_"></a></span><a class="link" href="comparisons.html#math_toolkit.comparisons.comparison_to_the_r_statistical_">Comparison
1846 to the R Statistical Library on Linux</a>
1849 All the results were measured on a 2.0GHz Intel T5800 Core 2 Duo, 4Gb RAM,
1850 Ubuntu Linux 9 machine, with the test program and R-2.9.2 compiled with GNU
1851 G++ 4.3.3 using -O3 -DNDEBUG=1.
1854 <a name="math_toolkit.comparisons.a_comparison_to_the_r_statistic0"></a><p class="title"><b>Table 15.6. A Comparison to the R Statistical Library on Linux</b></p>
1855 <div class="table-contents"><table class="table" summary="A Comparison to the R Statistical Library on Linux">
1865 Statistical Function
1888 <a class="link" href="dist_ref/dists/beta_dist.html" title="Beta Distribution">Beta Distribution</a>
1897 <p>(3.189e-006s)</p>
1904 <p><span class="bold"><strong>1.00</strong></span></p>
1906 <p>(1.526e-006s)</p>
1915 <p>(1.822e-006s)</p>
1923 <a class="link" href="dist_ref/dists/beta_dist.html" title="Beta Distribution">Beta Distribution</a>
1930 <p><span class="bold"><strong>1.00</strong></span></p>
1932 <p>(1.185e-005s)</p>
1939 <p>30.51<a href="#ftn.math_toolkit.comparisons.f7" class="footnote" name="math_toolkit.comparisons.f7"><sup class="footnote">[1]</sup></a></p>
1941 <p>(3.616e-004s)</p>
1950 <p>(2.989e-005s)</p>
1958 <a class="link" href="dist_ref/dists/binomial_dist.html" title="Binomial Distribution">Binomial
1959 Distribution</a> CDF
1967 <p>(9.175e-007s)</p>
1976 <p>(7.476e-007s)</p>
1983 <p><span class="bold"><strong>1.00</strong></span></p>
1985 <p>(2.081e-007s)</p>
1993 <a class="link" href="dist_ref/dists/binomial_dist.html" title="Binomial Distribution">Binomial
1994 Distribution</a> Quantile
2002 <p>(6.925e-006s)</p>
2009 <p><span class="bold"><strong>1.00</strong></span></p>
2011 <p>(4.407e-006s)</p>
2020 <p>(3.274e-005s)</p>
2028 <a class="link" href="dist_ref/dists/cauchy_dist.html" title="Cauchy-Lorentz Distribution">Cauchy Distribution</a>
2035 <p><span class="bold"><strong>1.00</strong></span></p>
2037 <p>(1.594e-007s)</p>
2046 <p>(1.654e-007s)</p>
2059 <a class="link" href="dist_ref/dists/cauchy_dist.html" title="Cauchy-Lorentz Distribution">Cauchy Distribution</a>
2068 <p>(1.752e-007s)</p>
2075 <p><span class="bold"><strong>1.00</strong></span></p>
2077 <p>(1.448e-007s)</p>
2090 <a class="link" href="dist_ref/dists/chi_squared_dist.html" title="Chi Squared Distribution">Chi
2091 Squared Distribution</a> CDF
2099 <p>(1.376e-006s)</p>
2108 <p>(1.243e-006s)</p>
2115 <p><span class="bold"><strong>1.00</strong></span></p>
2117 <p>(5.270e-007s)</p>
2125 <a class="link" href="dist_ref/dists/chi_squared_dist.html" title="Chi Squared Distribution">Chi
2126 Squared Distribution</a> Quantile
2132 <p><span class="bold"><strong>1.00</strong></span></p>
2134 <p>(4.252e-006s)</p>
2143 <p>(5.700e-006s)</p>
2152 <p>(1.477e-005s)</p>
2160 <a class="link" href="dist_ref/dists/exp_dist.html" title="Exponential Distribution">Exponential
2161 Distribution</a> CDF
2167 <p><span class="bold"><strong>1.00</strong></span></p>
2169 <p>(1.342e-007s)</p>
2178 <p>(1.677e-007s)</p>
2191 <a class="link" href="dist_ref/dists/exp_dist.html" title="Exponential Distribution">Exponential
2192 Distribution</a> Quantile
2198 <p><span class="bold"><strong>1.00</strong></span></p>
2200 <p>(8.827e-008s)</p>
2209 <p>(9.470e-008s)</p>
2222 <a class="link" href="dist_ref/dists/f_dist.html" title="F Distribution">Fisher F Distribution</a>
2231 <p>(2.324e-006s)</p>
2240 <p>(1.711e-006s)</p>
2247 <p><span class="bold"><strong>1.00</strong></span></p>
2249 <p>(1.437e-006s)</p>
2257 <a class="link" href="dist_ref/dists/f_dist.html" title="F Distribution">Fisher F Distribution</a>
2266 <p>(1.577e-005s)</p>
2273 <p><span class="bold"><strong>1.00</strong></span></p>
2275 <p>(1.033e-005s)</p>
2284 <p>(2.719e-005s)</p>
2292 <a class="link" href="dist_ref/dists/gamma_dist.html" title="Gamma (and Erlang) Distribution">Gamma Distribution</a>
2301 <p>(1.582e-006s)</p>
2310 <p>(1.309e-006s)</p>
2317 <p><span class="bold"><strong>1.00</strong></span></p>
2319 <p>(4.980e-007s)</p>
2327 <a class="link" href="dist_ref/dists/gamma_dist.html" title="Gamma (and Erlang) Distribution">Gamma Distribution</a>
2336 <p>(4.770e-006s)</p>
2345 <p>(1.513e-005s)</p>
2352 <p><span class="bold"><strong>1.00</strong></span></p>
2354 <p>(2.179e-006s)</p>
2362 <a class="link" href="dist_ref/dists/hypergeometric_dist.html" title="Hypergeometric Distribution">hypergeometric
2363 Distribution</a> CDF
2369 <p>2.20<a href="#ftn.math_toolkit.comparisons.f8" class="footnote" name="math_toolkit.comparisons.f8"><sup class="footnote">[2]</sup></a></p>
2371 <p>(3.522e-007s)</p>
2378 <p><span class="bold"><strong>1.00</strong></span></p>
2380 <p>(1.601e-007s)</p>
2393 <a class="link" href="dist_ref/dists/hypergeometric_dist.html" title="Hypergeometric Distribution">hypergeometric
2394 Distribution</a> Quantile
2400 <p><span class="bold"><strong>1.00</strong></span></p>
2402 <p>(8.279e-007s)</p>
2411 <p>(2.125e-006s)</p>
2424 <a class="link" href="dist_ref/dists/logistic_dist.html" title="Logistic Distribution">Logistic
2425 Distribution</a> CDF
2431 <p><span class="bold"><strong>1.00</strong></span></p>
2433 <p>(9.398e-008s)</p>
2442 <p>(2.588e-007s)</p>
2455 <a class="link" href="dist_ref/dists/logistic_dist.html" title="Logistic Distribution">Logistic
2456 Distribution</a> Quantile
2462 <p><span class="bold"><strong>1.00</strong></span></p>
2464 <p>(9.893e-008s)</p>
2473 <p>(1.285e-007s)</p>
2486 <a class="link" href="dist_ref/dists/lognormal_dist.html" title="Log Normal Distribution">Log-normal
2487 Distribution</a> CDF
2493 <p><span class="bold"><strong>1.00</strong></span></p>
2495 <p>(1.831e-007s)</p>
2504 <p>(2.539e-007s)</p>
2517 <a class="link" href="dist_ref/dists/lognormal_dist.html" title="Log Normal Distribution">Log-normal
2518 Distribution</a> Quantile
2526 <p>(5.551e-007s)</p>
2533 <p><span class="bold"><strong>1.00</strong></span></p>
2535 <p>(5.037e-007s)</p>
2548 <a class="link" href="dist_ref/dists/negative_binomial_dist.html" title="Negative Binomial Distribution">Negative
2549 Binomial Distribution</a> CDF
2557 <p>(1.563e-006s)</p>
2564 <p><span class="bold"><strong>1.00</strong></span></p>
2566 <p>(1.444e-006s)</p>
2573 <p><span class="bold"><strong>1.00</strong></span></p>
2575 <p>(1.444e-006s)</p>
2583 <a class="link" href="dist_ref/dists/negative_binomial_dist.html" title="Negative Binomial Distribution">Negative
2584 Binomial Distribution</a> Quantile
2590 <p><span class="bold"><strong>1.00</strong></span></p>
2592 <p>(1.700e-005s)</p>
2599 <p>25.92<a href="#ftn.math_toolkit.comparisons.f9" class="footnote" name="math_toolkit.comparisons.f9"><sup class="footnote">[3]</sup></a></p>
2601 <p>(4.407e-004s)</p>
2610 <p>(3.274e-005s)</p>
2618 <a class="link" href="dist_ref/dists/nc_chi_squared_dist.html" title="Noncentral Chi-Squared Distribution">Noncentral
2619 Chi Squared Distribution</a> CDF
2627 <p>(2.841e-005s)</p>
2636 <p>(1.405e-004s)</p>
2643 <p><span class="bold"><strong>1.00</strong></span></p>
2645 <p>(5.617e-006s)</p>
2653 <a class="link" href="dist_ref/dists/nc_chi_squared_dist.html" title="Noncentral Chi-Squared Distribution">Noncentral
2654 Chi Squared Distribution</a> Quantile
2662 <p>(1.879e-003s)</p>
2669 <p>144.91<a href="#ftn.math_toolkit.comparisons.f10" class="footnote" name="math_toolkit.comparisons.f10"><sup class="footnote">[4]</sup></a></p>
2671 <p>(3.214e-002s)</p>
2678 <p><span class="bold"><strong>1.00</strong></span></p>
2680 <p>(2.218e-004s)</p>
2688 <a class="link" href="dist_ref/dists/nc_f_dist.html" title="Noncentral F Distribution">Noncentral
2689 F Distribution</a> CDF
2697 <p>(5.868e-005s)</p>
2706 <p>(8.058e-006s)</p>
2713 <p><span class="bold"><strong>1.00</strong></span></p>
2715 <p>(5.682e-006s)</p>
2723 <a class="link" href="dist_ref/dists/nc_f_dist.html" title="Noncentral F Distribution">Noncentral
2724 F Distribution</a> Quantile
2732 <p>(7.869e-004s)</p>
2741 <p>(9.256e-004s)</p>
2748 <p><span class="bold"><strong>1.00</strong></span></p>
2750 <p>(1.396e-004s)</p>
2758 <a class="link" href="dist_ref/dists/nc_t_dist.html" title="Noncentral T Distribution">noncentral
2759 T distribution</a> CDF
2767 <p>(3.357e-005s)</p>
2774 <p><span class="bold"><strong>1.00</strong></span></p>
2776 <p>(6.844e-006s)</p>
2789 <a class="link" href="dist_ref/dists/nc_t_dist.html" title="Noncentral T Distribution">noncentral
2790 T distribution</a> Quantile
2798 <p>(9.265e-004s)</p>
2805 <p><span class="bold"><strong>1.00</strong></span><a href="#ftn.math_toolkit.comparisons.f11" class="footnote" name="math_toolkit.comparisons.f11"><sup class="footnote">[5]</sup></a></p>
2807 <p>(5.916e-004s)</p>
2820 <a class="link" href="dist_ref/dists/normal_dist.html" title="Normal (Gaussian) Distribution">Normal Distribution</a>
2827 <p><span class="bold"><strong>1.00</strong></span></p>
2829 <p>(1.074e-007s)</p>
2838 <p>(1.245e-007s)</p>
2847 <p>(5.762e-007s)</p>
2855 <a class="link" href="dist_ref/dists/normal_dist.html" title="Normal (Gaussian) Distribution">Normal Distribution</a>
2864 <p>(1.902e-007s)</p>
2871 <p><span class="bold"><strong>1.00</strong></span></p>
2873 <p>(1.490e-007s)</p>
2882 <p>(1.542e-006s)</p>
2890 <a class="link" href="dist_ref/dists/poisson_dist.html" title="Poisson Distribution">Poisson
2891 Distribution</a> CDF
2899 <p>(1.198e-006s)</p>
2908 <p>(1.110e-006s)</p>
2915 <p><span class="bold"><strong>1.00</strong></span></p>
2917 <p>(4.937e-007s)</p>
2925 <a class="link" href="dist_ref/dists/poisson_dist.html" title="Poisson Distribution">Poisson
2934 <p>(3.032e-006s)</p>
2941 <p><span class="bold"><strong>1.00</strong></span></p>
2943 <p>(2.724e-006s)</p>
2952 <p>(1.110e-005s)</p>
2960 <a class="link" href="dist_ref/dists/students_t_dist.html" title="Students t Distribution">Students
2961 t Distribution</a> CDF
2969 <p>(2.020e-006s)</p>
2976 <p><span class="bold"><strong>1.00</strong></span></p>
2978 <p>(9.321e-007s)</p>
2987 <p>(1.021e-006s)</p>
2995 <a class="link" href="dist_ref/dists/students_t_dist.html" title="Students t Distribution">Students
2996 t Distribution</a> Quantile
3004 <p>(3.972e-006s)</p>
3011 <p><span class="bold"><strong>1.00</strong></span></p>
3013 <p>(3.364e-006s)</p>
3022 <p>(1.308e-005s)</p>
3030 <a class="link" href="dist_ref/dists/weibull_dist.html" title="Weibull Distribution">Weibull
3031 Distribution</a> CDF
3037 <p><span class="bold"><strong>1.00</strong></span></p>
3039 <p>(3.662e-007s)</p>
3048 <p>(3.808e-007s)</p>
3061 <a class="link" href="dist_ref/dists/weibull_dist.html" title="Weibull Distribution">Weibull
3062 Distribution</a> Quantile
3068 <p><span class="bold"><strong>1.00</strong></span></p>
3070 <p>(4.112e-007s)</p>
3079 <p>(4.317e-007s)</p>
3090 <tbody class="footnotes"><tr><td colspan="4">
3091 <div id="ftn.math_toolkit.comparisons.f7" class="footnote"><p><a href="#math_toolkit.comparisons.f7" class="para"><sup class="para">[1] </sup></a>
3092 There are a small number of our test cases where the R library
3093 fails to converge on a result: these tend to dominate the performance
3096 <div id="ftn.math_toolkit.comparisons.f8" class="footnote"><p><a href="#math_toolkit.comparisons.f8" class="para"><sup class="para">[2] </sup></a>
3097 This result is somewhat misleading: for small values of the parameters
3098 there is virtually no difference between the two libraries, but
3099 for large values the Boost implementation is <span class="emphasis"><em>much</em></span>
3100 slower, albeit with much improved precision.
3102 <div id="ftn.math_toolkit.comparisons.f9" class="footnote"><p><a href="#math_toolkit.comparisons.f9" class="para"><sup class="para">[3] </sup></a>
3103 The R library appears to use a linear-search strategy, that can
3104 perform very badly in a small number of pathological cases, but
3105 may or may not be more efficient in "typical" cases
3107 <div id="ftn.math_toolkit.comparisons.f10" class="footnote"><p><a href="#math_toolkit.comparisons.f10" class="para"><sup class="para">[4] </sup></a>
3108 There are a small number of our test cases where the R library
3109 fails to converge on a result: these tend to dominate the performance
3112 <div id="ftn.math_toolkit.comparisons.f11" class="footnote"><p><a href="#math_toolkit.comparisons.f11" class="para"><sup class="para">[5] </sup></a>
3113 There are a small number of our test cases where the R library
3114 fails to converge on a result: these tend to dominate the performance
3120 <br class="table-break">
3122 <table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
3123 <td align="left"></td>
3124 <td align="right"><div class="copyright-footer">Copyright © 2006-2010, 2012-2014 Nikhar Agrawal,
3125 Anton Bikineev, Paul A. Bristow, Marco Guazzone, Christopher Kormanyos, Hubert
3126 Holin, Bruno Lalande, John Maddock, Johan Råde, Gautam Sewani, Benjamin Sobotta,
3127 Thijs van den Berg, Daryle Walker and Xiaogang Zhang<p>
3128 Distributed under the Boost Software License, Version 1.0. (See accompanying
3129 file LICENSE_1_0.txt or copy at <a href="http://www.boost.org/LICENSE_1_0.txt" target="_top">http://www.boost.org/LICENSE_1_0.txt</a>)
3134 <div class="spirit-nav">
3135 <a accesskey="p" href="tuning.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../perf.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="perf_test_app.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a>