1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 03:33:20 +01:00
llvm-mirror/unittests/FuzzMutate/ReservoirSamplerTest.cpp
Justin Bogner c5917e5476 Re-apply "Introduce FuzzMutate library"
Same as r311392 with some fixes for library dependencies. Thanks to
Chapuni for helping work those out!

Original commit message:

This introduces the FuzzMutate library, which provides structured
fuzzing for LLVM IR, as described in my EuroLLVM 2017 talk. Most of
the basic mutators to inject and delete IR are provided, with support
for most basic operations.

llvm-svn: 311402
2017-08-21 22:57:06 +00:00

70 lines
2.1 KiB
C++

//===- ReservoirSampler.cpp - Tests for the ReservoirSampler --------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "llvm/FuzzMutate/Random.h"
#include "gtest/gtest.h"
#include <random>
using namespace llvm;
TEST(ReservoirSamplerTest, OneItem) {
std::mt19937 Rand;
auto Sampler = makeSampler(Rand, 7, 1);
ASSERT_FALSE(Sampler.isEmpty());
ASSERT_EQ(7, Sampler.getSelection());
}
TEST(ReservoirSamplerTest, NoWeight) {
std::mt19937 Rand;
auto Sampler = makeSampler(Rand, 7, 0);
ASSERT_TRUE(Sampler.isEmpty());
}
TEST(ReservoirSamplerTest, Uniform) {
std::mt19937 Rand;
// Run three chi-squared tests to check that the distribution is reasonably
// uniform.
std::vector<int> Items = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
int Failures = 0;
for (int Run = 0; Run < 3; ++Run) {
std::vector<int> Counts(Items.size(), 0);
// We need $np_s > 5$ at minimum, but we're better off going a couple of
// orders of magnitude larger.
int N = Items.size() * 5 * 100;
for (int I = 0; I < N; ++I) {
auto Sampler = makeSampler(Rand, Items);
Counts[Sampler.getSelection()] += 1;
}
// Knuth. TAOCP Vol. 2, 3.3.1 (8):
// $V = \frac{1}{n} \sum_{s=1}^{k} \left(\frac{Y_s^2}{p_s}\right) - n$
double Ps = 1.0 / Items.size();
double Sum = 0.0;
for (int Ys : Counts)
Sum += Ys * Ys / Ps;
double V = (Sum / N) - N;
assert(Items.size() == 10 && "Our chi-squared values assume 10 items");
// Since we have 10 items, there are 9 degrees of freedom and the table of
// chi-squared values is as follows:
//
// | p=1% | 5% | 25% | 50% | 75% | 95% | 99% |
// v=9 | 2.088 | 3.325 | 5.899 | 8.343 | 11.39 | 16.92 | 21.67 |
//
// Check that we're in the likely range of results.
//if (V < 2.088 || V > 21.67)
if (V < 2.088 || V > 21.67)
++Failures;
}
EXPECT_LT(Failures, 3) << "Non-uniform distribution?";
}