2
0
mirror of https://github.com/boostorg/math.git synced 2026-01-26 06:42:12 +00:00
Files
math/test/test_gamma.cpp
John Maddock 4ecb968fc7 Begun to refactor and tighten up the tests so that results are actually regression tested as well as printed to screen.
Added code for extracting the sign of a number.
Fixed a few bugs and tidied up the code a little in places.
Changed test_result to properly encapsulate it's contents.


[SVN r3106]
2006-07-25 09:17:33 +00:00

336 lines
13 KiB
C++

// (C) Copyright John Maddock 2006.
// Use, modification and distribution are subject to 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)
#include <boost/math/concepts/real_concept.hpp>
#include <boost/math/special_functions/gamma.hpp>
#include <boost/test/included/test_exec_monitor.hpp>
#include <boost/test/floating_point_comparison.hpp>
#include <boost/math/tools/stats.hpp>
#include <boost/math/tools/test.hpp>
#include <boost/math/constants/constants.hpp>
#include <boost/type_traits/is_floating_point.hpp>
#include <boost/array.hpp>
#include <boost/lambda/lambda.hpp>
#include <boost/lambda/bind.hpp>
#include "test_gamma_hooks.hpp"
#include "handle_test_result.hpp"
//
// DESCRIPTION:
// ~~~~~~~~~~~~
//
// This file tests the functions tgamma and lgamma, and the
// function tgammap1m1. There are two sets of tests, spot
// tests which compare our results with selected values computed
// using the online special function calculator at
// functions.wolfram.com, while the bulk of the accuracy tests
// use values generated with NTL::RR at 1000-bit precision
// and our generic versions of these functions.
//
// Note that when this file is first run on a new platform many of
// these tests will fail: the default accuracy is 1 epsilon which
// is too tight for most platforms. In this situation you will
// need to cast a human eye over the error rates reported and make
// a judgement as to whether they are acceptable. Either way please
// report the results to the Boost mailing list. Acceptable rates of
// error are marked up below as a series of regular expressions that
// identify the compiler/stdlib/platform/data-type/test-data/test-function
// along with the maximum expected peek and RMS mean errors for that
// test.
//
void expected_results()
{
//
// Define the max and mean errors expected for
// various compilers and platforms.
//
const char* largest_type;
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
if(boost::math::tools::digits<double>() == boost::math::tools::digits<long double>())
{
largest_type = "(long\\s+)?double";
}
else
{
largest_type = "long double";
}
#else
largest_type = "(long\\s+)?double";
#endif
//
// Catch all cases come last:
//
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"factorials", // test data group
"boost::math::tgamma", 4, 1); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"factorials", // test data group
"boost::math::lgamma", 9, 1); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"near (0|-55)", // test data group
"boost::math::(t|l)gamma", 5, 3); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"near (1|2|-10)", // test data group
"boost::math::tgamma", 5, 3); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"near (1|2|-10)", // test data group
"boost::math::lgamma", 14, 7); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
largest_type, // test type(s)
"tgammap1m1.*", // test data group
"boost::math::tgammap1m1", 30, 9); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
"real_concept", // test type(s)
"factorials", // test data group
"boost::math::tgamma", 70, 25); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
"real_concept", // test type(s)
"factorials", // test data group
"boost::math::lgamma", 30, 4); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
"real_concept", // test type(s)
"near.*", // test data group
"boost::math::tgamma", 40, 20); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
"real_concept", // test type(s)
"near.*", // test data group
"boost::math::lgamma", 10000000, 10000000); // test function
add_expected_result(
".*", // compiler
".*", // stdlib
".*", // platform
"real_concept", // test type(s)
"tgammap1m1.*", // test data group
"boost::math::tgammap1m1", 20, 5); // test function
//
// Finish off by printing out the compiler/stdlib/platform names,
// we do this to make it easier to mark up expected error rates.
//
std::cout << "Tests run with " << BOOST_COMPILER << ", "
<< BOOST_STDLIB << ", " << BOOST_PLATFORM << std::endl;
}
template <class T>
void do_test_gamma(const T& data, const char* type_name, const char* test_name)
{
typedef typename T::value_type row_type;
typedef typename row_type::value_type value_type;
typedef value_type (*pg)(value_type);
pg funcp = boost::math::tgamma;
boost::math::tools::test_result<value_type> result;
std::cout << "Testing " << test_name << " with type " << type_name
<< "\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n";
//
// test tgamma against data:
//
result = boost::math::tools::test(
data,
boost::lambda::bind(funcp, boost::lambda::ret<value_type>(boost::lambda::_1[0])),
boost::lambda::ret<value_type>(boost::lambda::_1[1]));
handle_test_result(result, data[result.worst()], result.worst(), type_name, "boost::math::tgamma", test_name);
#ifdef TEST_OTHER
if(::boost::is_floating_point<value_type>::value){
funcp = other::tgamma;
result = boost::math::tools::test(
data,
boost::lambda::bind(funcp, boost::lambda::ret<value_type>(boost::lambda::_1[0])),
boost::lambda::ret<value_type>(boost::lambda::_1[1]));
print_test_result(result, data[result.worst()], result.worst(), type_name, "other::tgamma");
}
#endif
//
// test lgamma against data:
//
funcp = boost::math::lgamma;
result = boost::math::tools::test(
data,
boost::lambda::bind(funcp, boost::lambda::ret<value_type>(boost::lambda::_1[0])),
boost::lambda::ret<value_type>(boost::lambda::_1[2]));
handle_test_result(result, data[result.worst()], result.worst(), type_name, "boost::math::lgamma", test_name);
#ifdef TEST_OTHER
if(::boost::is_floating_point<value_type>::value){
funcp = other::lgamma;
result = boost::math::tools::test(
data,
boost::lambda::bind(funcp, boost::lambda::ret<value_type>(boost::lambda::_1[0])),
boost::lambda::ret<value_type>(boost::lambda::_1[2]));
print_test_result(result, data[result.worst()], result.worst(), type_name, "other::lgamma");
}
#endif
std::cout << std::endl;
}
template <class T>
void do_test_gammap1m1(const T& data, const char* type_name, const char* test_name)
{
typedef typename T::value_type row_type;
typedef typename row_type::value_type value_type;
typedef value_type (*pg)(value_type);
pg funcp = boost::math::tgammap1m1;
boost::math::tools::test_result<value_type> result;
std::cout << "Testing " << test_name << " with type " << type_name
<< "\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n";
//
// test tgammap1m1 against data:
//
result = boost::math::tools::test(
data,
boost::lambda::bind(funcp, boost::lambda::ret<value_type>(boost::lambda::_1[0])),
boost::lambda::ret<value_type>(boost::lambda::_1[1]));
handle_test_result(result, data[result.worst()], result.worst(), type_name, "boost::math::tgammap1m1", test_name);
std::cout << std::endl;
}
template <class T>
void test_gamma(T, const char* name)
{
//
// The actual test data is rather verbose, so it's in a separate file
//
// The contents are as follows, each row of data contains
// three items, input value, gamma and lgamma:
//
// gamma and lgamma at integer and half integer values:
// boost::array<boost::array<T, 3>, N> factorials;
//
// gamma and lgamma for z near 0:
// boost::array<boost::array<T, 3>, N> near_0;
//
// gamma and lgamma for z near 1:
// boost::array<boost::array<T, 3>, N> near_1;
//
// gamma and lgamma for z near 2:
// boost::array<boost::array<T, 3>, N> near_2;
//
// gamma and lgamma for z near -10:
// boost::array<boost::array<T, 3>, N> near_m10;
//
// gamma and lgamma for z near -55:
// boost::array<boost::array<T, 3>, N> near_m55;
//
// The last two cases are chosen more or less at random,
// except that one is even and the other odd, and both are
// at negative poles. The data near zero also tests near
// a pole, the data near 1 and 2 are to probe lgamma as
// the result -> 0.
//
# include "test_gamma_data.ipp"
do_test_gamma(factorials, name, "factorials");
do_test_gamma(near_0, name, "near 0");
do_test_gamma(near_1, name, "near 1");
do_test_gamma(near_2, name, "near 2");
do_test_gamma(near_m10, name, "near -10");
do_test_gamma(near_m55, name, "near -55");
//
// And now tgammap1m1 which computes gamma(1+dz)-1:
//
do_test_gammap1m1(gammap1m1_data, name, "tgammap1m1(dz)");
}
template <class T>
void test_spots(T)
{
//
// basic sanity checks, tolerance is 50 epsilon expressed as a percentage:
//
T tolerance = boost::math::tools::epsilon<T>() * 5000;
BOOST_CHECK_CLOSE(::boost::math::tgamma(static_cast<T>(3.5)), static_cast<T>(3.3233509704478425511840640312646472177454052302295L), tolerance);
BOOST_CHECK_CLOSE(::boost::math::tgamma(static_cast<T>(0.125)), static_cast<T>(7.5339415987976119046992298412151336246104195881491L), tolerance);
BOOST_CHECK_CLOSE(::boost::math::tgamma(static_cast<T>(-0.125)), static_cast<T>(-8.7172188593831756100190140408231437691829605421405L), tolerance);
BOOST_CHECK_CLOSE(::boost::math::tgamma(static_cast<T>(-3.125)), static_cast<T>(1.1668538708507675587790157356605097019141636072094L), tolerance);
BOOST_CHECK_CLOSE(::boost::math::tgamma(static_cast<T>(-53249.0/1024)), static_cast<T>(-1.2646559519067605488251406578743995122462767733517e-65L), tolerance);
int sign = 1;
BOOST_CHECK_CLOSE(::boost::math::lgamma(static_cast<T>(3.5), &sign), static_cast<T>(1.2009736023470742248160218814507129957702389154682L), tolerance);
BOOST_CHECK(sign == 1);
BOOST_CHECK_CLOSE(::boost::math::lgamma(static_cast<T>(0.125), &sign), static_cast<T>(2.0194183575537963453202905211670995899482809521344L), tolerance);
BOOST_CHECK(sign == 1);
BOOST_CHECK_CLOSE(::boost::math::lgamma(static_cast<T>(-0.125), &sign), static_cast<T>(2.1653002489051702517540619481440174064962195287626L), tolerance);
BOOST_CHECK(sign == -1);
BOOST_CHECK_CLOSE(::boost::math::lgamma(static_cast<T>(-3.125), &sign), static_cast<T>(0.1543111276840418242676072830970532952413339012367L), tolerance);
BOOST_CHECK(sign == 1);
BOOST_CHECK_CLOSE(::boost::math::lgamma(static_cast<T>(-53249.0/1024), &sign), static_cast<T>(-149.43323093420259741100038126078721302600128285894L), tolerance);
BOOST_CHECK(sign == -1);
}
int test_main(int, char* [])
{
expected_results();
test_spots(0.0F);
test_spots(0.0);
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
test_spots(0.0L);
test_spots(boost::math::concepts::real_concept(0.1));
#endif
test_gamma(0.1F, "float");
test_gamma(0.1, "double");
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
test_gamma(0.1L, "long double");
test_gamma(boost::math::concepts::real_concept(0.1), "real_concept");
#else
std::cout << "<note>The long double tests have been disabled on this platform "
"either because the long double overloads of the usual math functions are "
"not available at all, or because they are too inaccurate for these tests "
"to pass.</note>" << std::cout;
#endif
return 0;
}