| 1 | //===----------------------------------------------------------------------===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | // REQUIRES: long_tests |
| 10 | |
| 11 | // <random> |
| 12 | |
| 13 | // template<class _IntType = int> |
| 14 | // class uniform_int_distribution |
| 15 | |
| 16 | // template<class _URNG> result_type operator()(_URNG& g); |
| 17 | |
| 18 | #include <random> |
| 19 | #include <cassert> |
| 20 | #include <climits> |
| 21 | #include <cmath> |
| 22 | #include <cstddef> |
| 23 | #include <cstdint> |
| 24 | #include <limits> |
| 25 | #include <numeric> |
| 26 | #include <vector> |
| 27 | |
| 28 | #include "test_macros.h" |
| 29 | |
| 30 | template <class T> |
| 31 | T sqr(T x) { |
| 32 | return x * x; |
| 33 | } |
| 34 | |
| 35 | template <class ResultType, class EngineType> |
| 36 | void test_statistics(ResultType a, ResultType b) { |
| 37 | ASSERT_SAME_TYPE(typename std::uniform_int_distribution<ResultType>::result_type, ResultType); |
| 38 | |
| 39 | EngineType g; |
| 40 | std::uniform_int_distribution<ResultType> dist(a, b); |
| 41 | assert(dist.a() == a); |
| 42 | assert(dist.b() == b); |
| 43 | std::vector<ResultType> u; |
| 44 | for (int i = 0; i < 10000; ++i) { |
| 45 | ResultType v = dist(g); |
| 46 | assert(a <= v && v <= b); |
| 47 | u.push_back(v); |
| 48 | } |
| 49 | |
| 50 | // Quick check: The chance of getting *no* hits in any given tenth of the range |
| 51 | // is (0.9)^10000, or "ultra-astronomically low." |
| 52 | bool bottom_tenth = false; |
| 53 | bool top_tenth = false; |
| 54 | for (std::size_t i = 0; i < u.size(); ++i) { |
| 55 | bottom_tenth = bottom_tenth || (u[i] <= (a + (b / 10) - (a / 10))); |
| 56 | top_tenth = top_tenth || (u[i] >= (b - (b / 10) + (a / 10))); |
| 57 | } |
| 58 | assert(bottom_tenth); // ...is populated |
| 59 | assert(top_tenth); // ...is populated |
| 60 | |
| 61 | // Now do some more involved statistical math. |
| 62 | double mean = std::accumulate(u.begin(), u.end(), 0.0) / u.size(); |
| 63 | double var = 0; |
| 64 | double skew = 0; |
| 65 | double kurtosis = 0; |
| 66 | for (std::size_t i = 0; i < u.size(); ++i) { |
| 67 | double dbl = (u[i] - mean); |
| 68 | double d2 = dbl * dbl; |
| 69 | var += d2; |
| 70 | skew += dbl * d2; |
| 71 | kurtosis += d2 * d2; |
| 72 | } |
| 73 | var /= u.size(); |
| 74 | double dev = std::sqrt(x: var); |
| 75 | skew /= u.size() * dev * var; |
| 76 | kurtosis /= u.size() * var * var; |
| 77 | |
| 78 | double expected_mean = double(a) + double(b)/2 - double(a)/2; |
| 79 | double expected_var = (sqr(double(b) - double(a) + 1) - 1) / 12; |
| 80 | |
| 81 | double range = double(b) - double(a) + 1.0; |
| 82 | assert(range > range / 10); // i.e., it's not infinity |
| 83 | |
| 84 | assert(std::abs(mean - expected_mean) < range / 100); |
| 85 | assert(std::abs(var - expected_var) < expected_var / 50); |
| 86 | assert(-0.1 < skew && skew < 0.1); |
| 87 | assert(1.6 < kurtosis && kurtosis < 2.0); |
| 88 | } |
| 89 | |
| 90 | template <class ResultType, class EngineType> |
| 91 | void test_statistics() { |
| 92 | test_statistics<ResultType, EngineType>(0, std::numeric_limits<ResultType>::max()); |
| 93 | } |
| 94 | |
| 95 | int main(int, char**) |
| 96 | { |
| 97 | test_statistics<int, std::minstd_rand0>(); |
| 98 | test_statistics<int, std::minstd_rand>(); |
| 99 | test_statistics<int, std::mt19937>(); |
| 100 | test_statistics<int, std::mt19937_64>(); |
| 101 | test_statistics<int, std::ranlux24_base>(); |
| 102 | test_statistics<int, std::ranlux48_base>(); |
| 103 | test_statistics<int, std::ranlux24>(); |
| 104 | test_statistics<int, std::ranlux48>(); |
| 105 | test_statistics<int, std::knuth_b>(); |
| 106 | test_statistics<int, std::minstd_rand0>(a: -6, b: 106); |
| 107 | test_statistics<int, std::minstd_rand>(a: 5, b: 100); |
| 108 | |
| 109 | test_statistics<short, std::minstd_rand0>(); |
| 110 | test_statistics<int, std::minstd_rand0>(); |
| 111 | test_statistics<long, std::minstd_rand0>(); |
| 112 | test_statistics<long long, std::minstd_rand0>(); |
| 113 | |
| 114 | test_statistics<unsigned short, std::minstd_rand0>(); |
| 115 | test_statistics<unsigned int, std::minstd_rand0>(); |
| 116 | test_statistics<unsigned long, std::minstd_rand0>(); |
| 117 | test_statistics<unsigned long long, std::minstd_rand0>(); |
| 118 | |
| 119 | test_statistics<short, std::minstd_rand0>(SHRT_MIN, SHRT_MAX); |
| 120 | |
| 121 | #if defined(_LIBCPP_VERSION) // extension |
| 122 | test_statistics<std::int8_t, std::minstd_rand0>(); |
| 123 | test_statistics<std::uint8_t, std::minstd_rand0>(); |
| 124 | |
| 125 | #if !defined(TEST_HAS_NO_INT128) |
| 126 | test_statistics<__int128_t, std::minstd_rand0>(); |
| 127 | test_statistics<__uint128_t, std::minstd_rand0>(); |
| 128 | |
| 129 | test_statistics<__int128_t, std::minstd_rand0>(-100, 900); |
| 130 | test_statistics<__int128_t, std::minstd_rand0>(0, UINT64_MAX); |
| 131 | test_statistics<__int128_t, std::minstd_rand0>(std::numeric_limits<__int128_t>::min(), std::numeric_limits<__int128_t>::max()); |
| 132 | test_statistics<__uint128_t, std::minstd_rand0>(0, UINT64_MAX); |
| 133 | #endif |
| 134 | #endif |
| 135 | |
| 136 | return 0; |
| 137 | } |
| 138 | |