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
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
|
#define DOCTEST_CONFIG_IMPLEMENT_WITH_MAIN
#include <doctest.h>
#include <taskflow/taskflow.hpp>
// --------------------------------------------------------
// Graph generation
// --------------------------------------------------------
struct Node {
std::string name;
size_t idx {0};
size_t level {0};
bool visited {false};
std::atomic<size_t> dependents {0};
std::vector<Node*> successors;
void precede(Node& n) {
successors.emplace_back(&n);
n.dependents ++;
}
};
std::unique_ptr<Node[]> make_dag(size_t num_nodes, size_t max_degree) {
std::unique_ptr<Node[]> nodes(new Node[num_nodes]);
// Make sure nodes are in clean state
for(size_t i=0; i<num_nodes; i++) {
nodes[i].idx = i;
nodes[i].name = std::to_string(i);
REQUIRE(!nodes[i].visited);
REQUIRE(nodes[i].successors.empty());
REQUIRE(nodes[i].dependents == 0);
}
// Create a DAG
for(size_t i=0; i<num_nodes; i++) {
size_t degree {0};
for(size_t j=i+1; j<num_nodes && degree < max_degree; j++) {
if(j%2 == 1) {
nodes[i].precede(nodes[j]);
degree ++;
}
}
}
return nodes;
}
std::unique_ptr<Node[]> make_chain(size_t num_nodes) {
std::unique_ptr<Node[]> nodes(new Node[num_nodes]);
// Make sure nodes are in clean state
for(size_t i=0; i<num_nodes; i++) {
nodes[i].idx = i;
nodes[i].name = std::to_string(i);
REQUIRE(!nodes[i].visited);
REQUIRE(nodes[i].successors.empty());
REQUIRE(nodes[i].dependents == 0);
}
// Create a DAG
for(size_t i=1; i<num_nodes; i++) {
nodes[i-1].precede(nodes[i]);
}
return nodes;
}
// --------------------------------------------------------
// Testcase: StaticTraversal
// --------------------------------------------------------
void static_traversal(unsigned W) {
size_t max_degree = 4;
size_t num_nodes = 1000;
auto nodes = make_dag(num_nodes, max_degree);
tf::Taskflow tf;
tf::Executor executor(W);
std::atomic<size_t> level(0);
std::vector<tf::Task> tasks;
for(size_t i=0; i<num_nodes; ++i) {
auto task = tf.emplace([&level, v=&(nodes[i])](){
v->level = ++level;
v->visited = true;
for(size_t j=0; j<v->successors.size(); ++j) {
v->successors[j]->dependents.fetch_sub(1);
}
}).name(nodes[i].name);
tasks.push_back(task);
}
for(size_t i=0; i<num_nodes; ++i) {
for(size_t j=0; j<nodes[i].successors.size(); ++j) {
tasks[i].precede(tasks[nodes[i].successors[j]->idx]);
}
}
executor.run(tf).wait(); // block until finished
for(size_t i=0; i<num_nodes; i++) {
REQUIRE(nodes[i].visited);
REQUIRE(nodes[i].dependents == 0);
for(size_t j=0; j<nodes[i].successors.size(); ++j) {
REQUIRE(nodes[i].level < nodes[i].successors[j]->level);
}
}
}
TEST_CASE("StaticTraversal.1thread" * doctest::timeout(300)) {
static_traversal(1);
}
TEST_CASE("StaticTraversal.2threads" * doctest::timeout(300)) {
static_traversal(2);
}
TEST_CASE("StaticTraversal.3threads" * doctest::timeout(300)) {
static_traversal(3);
}
TEST_CASE("StaticTraversal.4threads" * doctest::timeout(300)) {
static_traversal(4);
}
// --------------------------------------------------------
// Testcase: DynamicTraversal
// --------------------------------------------------------
void dynamic_traversal(unsigned W) {
std::atomic<size_t> level;
std::function<void(Node*, tf::Subflow&)> traverse;
traverse = [&] (Node* n, tf::Subflow& subflow) {
REQUIRE(!n->visited);
n->visited = true;
size_t S = n->successors.size();
for(size_t i=0; i<S; i++) {
if(n->successors[i]->dependents.fetch_sub(1) == 1) {
n->successors[i]->level = ++level;
subflow.emplace([s=n->successors[i], &traverse](tf::Subflow &subflow2){
traverse(s, subflow2);
});
}
}
};
size_t max_degree = 4;
size_t num_nodes = 1000;
auto nodes = make_dag(num_nodes, max_degree);
std::vector<Node*> src;
for(size_t i=0; i<num_nodes; i++) {
if(nodes[i].dependents == 0) {
src.emplace_back(&(nodes[i]));
}
}
level = 0;
tf::Taskflow tf;
tf::Executor executor(W);
for(size_t i=0; i<src.size(); i++) {
tf.emplace([s=src[i], &traverse](tf::Subflow& subflow){
traverse(s, subflow);
});
}
executor.run(tf).wait(); // block until finished
for(size_t i=0; i<num_nodes; i++) {
REQUIRE(nodes[i].visited);
REQUIRE(nodes[i].dependents == 0);
for(size_t j=0; j<nodes[i].successors.size(); ++j) {
REQUIRE(nodes[i].level < nodes[i].successors[j]->level);
}
}
}
TEST_CASE("DynamicTraversal.1thread" * doctest::timeout(300)) {
dynamic_traversal(1);
}
TEST_CASE("DynamicTraversal.2threads" * doctest::timeout(300)) {
dynamic_traversal(2);
}
TEST_CASE("DynamicTraversal.3threads" * doctest::timeout(300)) {
dynamic_traversal(3);
}
TEST_CASE("DynamicTraversal.4threads" * doctest::timeout(300)) {
dynamic_traversal(4);
}
// --------------------------------------------------------
// Testcase: ParallelTraversal
// --------------------------------------------------------
/*void parallel_traversal(unsigned num_threads) {
tf::Executor executor(num_threads);
std::vector<std::thread> threads;
for(unsigned t=0; t<num_threads; ++t) {
threads.emplace_back([&](){
std::atomic<size_t> level {0};
size_t max_degree = 4;
size_t num_nodes = 1000;
auto nodes = make_dag(num_nodes, max_degree);
std::vector<Node*> src;
for(size_t i=0; i<num_nodes; i++) {
if(nodes[i].dependents == 0) {
src.emplace_back(&(nodes[i]));
}
}
std::function<void(Node*, tf::Subflow&)> traverse;
traverse = [&] (Node* n, tf::Subflow& subflow) {
REQUIRE(!n->visited);
n->visited = true;
size_t S = n->successors.size();
for(size_t i=0; i<S; i++) {
if(n->successors[i]->dependents.fetch_sub(1) == 1) {
n->successors[i]->level = ++level;
subflow.emplace([s=n->successors[i], &traverse](tf::Subflow &subflow){
traverse(s, subflow);
});
}
}
};
tf::Taskflow tf;
for(size_t i=0; i<src.size(); i++) {
tf.emplace([s=src[i], &traverse](tf::Subflow& subflow){
traverse(s, subflow);
});
}
executor.run(tf).wait(); // block until finished
for(size_t i=0; i<num_nodes; i++) {
REQUIRE(nodes[i].visited);
REQUIRE(nodes[i].dependents == 0);
for(size_t j=0; j<nodes[i].successors.size(); ++j) {
REQUIRE(nodes[i].level < nodes[i].successors[j]->level);
}
}
});
}
for(auto& thread : threads) thread.join();
}
TEST_CASE("ParallelTraversal.1" * doctest::timeout(300)) {
parallel_traversal(1);
}
TEST_CASE("ParallelTraversal.2" * doctest::timeout(300)) {
parallel_traversal(2);
}
TEST_CASE("ParallelTraversal.3" * doctest::timeout(300)) {
parallel_traversal(3);
}
TEST_CASE("ParallelTraversal.4" * doctest::timeout(300)) {
parallel_traversal(4);
}
TEST_CASE("ParallelTraversal.5" * doctest::timeout(300)) {
parallel_traversal(5);
}
TEST_CASE("ParallelTraversal.6" * doctest::timeout(300)) {
parallel_traversal(6);
}
TEST_CASE("ParallelTraversal.7" * doctest::timeout(300)) {
parallel_traversal(7);
}
TEST_CASE("ParallelTraversal.8" * doctest::timeout(300)) {
parallel_traversal(8);
}
*/
|
