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
| #include <iostream>
#include <vector>
#include <thread>
#include <cmath>
#include <algorithm>
#include <numeric>
#include <chrono>
template<typename T, typename BinaryOp>
T serial_reduce(const std::vector<T>& data, T init, BinaryOp op) {
T result = init;
for (const auto& item : data) {
result = op(result, item);
}
return result;
}
template<typename T, typename BinaryOp>
T parallel_reduce(const std::vector<T>& data, T init, BinaryOp op, int num_threads) {
int n = data.size();
if (n == 0) return init;
std::vector<T> partial_results(num_threads, init);
std::vector<std::thread> threads;
int elements_per_thread = std::ceil(static_cast<double>(n) / num_threads);
for (int i = 0; i < num_threads; i++) {
threads.emplace_back([i, &data, &partial_results, &op, elements_per_thread, n]() {
int start = i * elements_per_thread;
int end = std::min((i + 1) * elements_per_thread, n);
if (start < n) {
T local_result = data[start];
for (int j = start + 1; j < end; j++) {
local_result = op(local_result, data[j]);
}
partial_results[i] = local_result;
}
});
}
for (auto& thread : threads) {
thread.join();
}
T final_result = init;
for (const auto& partial : partial_results) {
final_result = op(final_result, partial);
}
return final_result;
}
template<typename T, typename BinaryOp>
T tree_reduce(const std::vector<T>& data, int start, int end, T init, BinaryOp op) {
if (start == end) return init;
if (start + 1 == end) return data[start];
int mid = (start + end) / 2;
T left = tree_reduce(data, start, mid, init, op);
T right = tree_reduce(data, mid, end, init, op);
return op(left, right);
}
int main() {
const int size = 1000000;
std::vector<int> data(size);
for (int i = 0; i < size; i++) {
data[i] = i % 100;
}
auto start = std::chrono::high_resolution_clock::now();
int serial_sum = serial_reduce(data, 0, [](int a, int b) { return a + b; });
auto end = std::chrono::high_resolution_clock::now();
std::chrono::duration<double> serial_duration = end - start;
start = std::chrono::high_resolution_clock::now();
int parallel_sum = parallel_reduce(data, 0, [](int a, int b) { return a + b; }, 4);
end = std::chrono::high_resolution_clock::now();
std::chrono::duration<double> parallel_duration = end - start;
start = std::chrono::high_resolution_clock::now();
int tree_sum = tree_reduce(data, 0, data.size(), 0, [](int a, int b) { return a + b; });
end = std::chrono::high_resolution_clock::now();
std::chrono::duration<double> tree_duration = end - start;
std::cout << "Serial reduction result: " << serial_sum << ", time: " << serial_duration.count() << "s\n";
std::cout << "Parallel reduction result: " << parallel_sum << ", time: " << parallel_duration.count() << "s\n";
std::cout << "Tree reduction result: " << tree_sum << ", time: " << tree_duration.count() << "s\n";
return 0;
}
|
