parallel-cpp/reduce/test.cpp at main · neudachina/parallel-cpp · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
#include "reduce.h"

#include "timeout_guard.h"
#include <gtest/gtest.h>

#include <random>


static std::vector<int> gen_vector(size_t size) {
  static std::mt19937 gen;
  std::uniform_int_distribution<int> dist(1);
  std::vector<int> v(size);
  for (int& x : v) {
    x = dist(gen);
  }
  return v;
}

TEST(Reduce, OneThreadSum) {
  const size_t n_threads = 1;
  const size_t size = 1024;
  std::vector<int> v = gen_vector(size);

  auto sum = [] (int a, int b) {
      return a + b;
  };

  int result = std::reduce(v.begin(), v.end(), 0, sum);

  TimeoutGuard guard(50ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 0, sum, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, OneThreadProd) {
  const size_t n_threads = 1;
  const size_t size = 1024;
  std::vector<int> v = gen_vector(size);

  auto prod = [] (int a, int b) {
      return a * b;
  };

  int result = std::reduce(v.begin(), v.end(), 1, prod);

  TimeoutGuard guard(50ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 1, prod, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, TwoThreadSum) {
  const size_t n_threads = 2;
  const size_t size = 1ull << 22;
  std::vector<int> v = gen_vector(size);

  auto sum = [] (int a, int b) {
      return a + b;
  };

  int result = std::reduce(v.begin(), v.end(), 0, sum);
  TimeoutGuard guard(250ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 0, sum, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, TwoThreadsProd) {
  const size_t n_threads = 2;
  const size_t size = 1ull << 22;
  std::vector<int> v = gen_vector(size);

  auto prod = [] (int a, int b) {
      return a * b;
  };

  int result = std::reduce(v.begin(), v.end(), 1, prod);
  TimeoutGuard guard(250ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 1, prod, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, TenThreadSum) {
  const size_t n_threads = 10;
  const size_t size = 1ull << 22;
  std::vector<int> v = gen_vector(size);

  auto sum = [] (int a, int b) {
      return a + b;
  };

  int result = std::reduce(v.begin(), v.end(), 0, sum);
  TimeoutGuard guard(250ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 0, sum, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, TenThreadsProd) {
  const size_t n_threads = 10;
  const size_t size = 1ull << 22;
  std::vector<int> v = gen_vector(size);

  auto prod = [] (int a, int b) {
      return a * b;
  };

  int result = std::reduce(v.begin(), v.end(), 1, prod);
  TimeoutGuard guard(250ms);
  int parallel_result = parallel_reduce(v.begin(), v.end(), 1, prod, n_threads);
  ASSERT_EQ(result, parallel_result);
}

TEST(Reduce, TwoThreadsFaster) {
  const size_t size = 1ull << 22;
  std::vector<int> v = gen_vector(size);

  auto sum = [] (int a, int b) {
      return a + b;
  };

  TimeoutGuard guard(500ms);
  auto start1 = std::chrono::high_resolution_clock::now();
  int parallel_result1 = parallel_reduce(v.begin(), v.end(), 0, sum, 1);
  auto end1 = std::chrono::high_resolution_clock::now();
  auto dur1 = end1 - start1;

  auto start2 = std::chrono::high_resolution_clock::now();
  int parallel_result2 = parallel_reduce(v.begin(), v.end(), 0, sum, 2);
  auto end2 = std::chrono::high_resolution_clock::now();
  auto dur2 = end2 - start2;

  ASSERT_EQ(parallel_result1, parallel_result2);
  ASSERT_GE(dur1.count(), 5 * dur2.count() / 4);
}