-
Notifications
You must be signed in to change notification settings - Fork 1
/
Boruvka-MST-parallel.cpp
389 lines (285 loc) · 13 KB
/
Boruvka-MST-parallel.cpp
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
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
#include<bits/stdc++.h>
#define INF INT_MAX
#define __CL_ENABLE_EXCEPTIONS
#define CL_HPP_TARGET_OPENCL_VERSION 220
#define CL_DEVICE_TYPE_DEFAULT CL_DEVICE_TYPE_GPU
#include<CL/cl.hpp>
// a single node will store an edge
struct node
{
int v1,v2,weight;
};
//this struct will store parent and rank
// of the vertex which will be needed
//during union-find
struct vertex_record
{
int parent,rank;
};
//thread-safe find_set function of DSU
int get_parent(int x,struct vertex_record *parent)
{
// //printf("x=%d\n",x);
while (x!=parent[x].parent)
{
//printf("x=%d,parent=%d\n",x,parent[x].parent);
int t=parent[x].parent;
//printf("t=%d\n",t);
// __global
__sync_bool_compare_and_swap(&parent[x].parent,t,parent[t].parent);
x=parent[t].parent;
}
return x;
}
//function for reading the graph from file
void take_graph_input_weighted(std::string filename,
std::vector<struct node>&edges,
int &vertex_count,
int &edge_count,
int &starting_vertex )
{
std::ifstream infile(filename);
int v1,v2,w;
std::set<int>vertex_set;
while(infile>>v1>>v2>>w)
{
edge_count++;
edges.push_back({v1,v2,w}); //vector of nodes that will store input edges
vertex_set.insert(v1),vertex_set.insert(v2); // a temporary set to get the vertex count
}
vertex_count=vertex_set.size();
starting_vertex=*(vertex_set.begin());
int add=0;
if(starting_vertex==0) //in case the smallest vertex id starts from 0,
// we will start from 1.
add=1;
for(auto &x:edges)
{
x.v1=x.v1+add,x.v2=x.v2+add;
}
}
//doing the necessary initialization for kernels
void device_setup_OpenCL(std::string kernel_filename,
std::string kernel_funcname1,
std::string kernel_funcname2,
std::string kernel_funcname3,
std::string kernel_funcname4,
std::vector<cl::Platform>&platforms,
cl::Context &context,
std::vector<cl::Device>&devices,
cl::CommandQueue &queue,
cl::Program &program,
cl::Kernel &kernel_find_min_edge_weight,
cl::Kernel &kernel_find_min_edge_index,
cl::Kernel &kernel_combine_components,
cl::Kernel &kernel_set_max_edge_weight
)
{
// std::cout<<"Entering device setup\n";
cl::Platform::get(&platforms);
// std::cout<<"Platforms queried\n";
cl_context_properties cps[3]={
CL_CONTEXT_PLATFORM,
(cl_context_properties)(platforms[0])(),
0
};
// std::cout<<"context properties obtained\n";
context=cl::Context(CL_DEVICE_TYPE_GPU,cps);
// std::cout<<"Success2\n";
devices=context.getInfo<CL_CONTEXT_DEVICES>();
queue=cl::CommandQueue(context,devices[0],CL_QUEUE_PROFILING_ENABLE);
std::ifstream sourceFile(kernel_filename);
std::string sourceCode(
std::istreambuf_iterator<char>(sourceFile),
(std::istreambuf_iterator<char>()));
cl::Program::Sources sources(1,
std::make_pair(sourceCode.c_str(), sourceCode.length()+1));
// Make program of the source code in the context
program = cl::Program(context, sources);
// Build program for these specific devices
program.build(devices);
kernel_find_min_edge_weight=cl::Kernel(program,kernel_funcname1.c_str());
kernel_find_min_edge_index=cl::Kernel(program,kernel_funcname2.c_str());
kernel_combine_components=cl::Kernel(program,kernel_funcname3.c_str());
kernel_set_max_edge_weight=cl::Kernel(program,kernel_funcname4.c_str());
}
//setting up buffers that will be passed to kernels
void buffer_setup(cl::Context context,
cl::CommandQueue queue,
cl::Buffer buffer_arr[],
int *current_edges,
std::vector<struct node>&edges,
struct vertex_record* parent,
struct vertex_record** parent_ptr,
int* min_weight,
int* min_weight_edge_index_component_buffer,
int vertex_count,
int edge_count)
{
buffer_arr[0]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(int)*(edge_count));
buffer_arr[1]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(struct node)*(edge_count+1));
buffer_arr[2]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(vertex_record)*(vertex_count+1));
buffer_arr[3]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(int)*(vertex_count+1));
buffer_arr[4]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(int)*(vertex_count+1));
buffer_arr[5]=cl::Buffer(context,CL_MEM_READ_WRITE,
sizeof(vertex_record*)*(vertex_count+1));
queue.enqueueWriteBuffer(buffer_arr[0],
CL_TRUE,0,sizeof(int)*(edge_count),(void*)current_edges);
queue.enqueueWriteBuffer(buffer_arr[1],
CL_TRUE,0,sizeof(struct node)*(edge_count+1),(void*)edges.data());
queue.enqueueWriteBuffer(buffer_arr[2],
CL_TRUE,0,sizeof(vertex_record)*(vertex_count+1),(void*)parent);
queue.enqueueWriteBuffer(buffer_arr[3],
CL_TRUE,0,sizeof(int)*(vertex_count+1),(void*)min_weight);
queue.enqueueWriteBuffer(buffer_arr[4],
CL_TRUE,0,sizeof(int)*(vertex_count+1),
(void*)min_weight_edge_index_component_buffer);
queue.enqueueWriteBuffer(buffer_arr[5],
CL_TRUE,0,sizeof(vertex_record*)*(vertex_count+1),
(void*)parent_ptr);
}
//driver function for getting component count
int count_components(cl::CommandQueue queue,
cl::Buffer buffer_arr[],
struct vertex_record* parent,
cl::Kernel kernel_combine_components,
cl::Kernel kernel_set_max_edge_weight,
int *current_edges,
int edge_count,
int vertex_count)
{
kernel_combine_components.setArg(0,buffer_arr[0]);
kernel_combine_components.setArg(1,buffer_arr[1]);
kernel_combine_components.setArg(2,buffer_arr[2]);
cl::NDRange global=cl::NDRange(edge_count);
cl::NDRange local=cl::NDRange(1);
queue.enqueueNDRangeKernel
(kernel_combine_components,cl::NullRange,global,local);
std::set<int>distinct_components;
queue.enqueueReadBuffer(buffer_arr[2],
CL_TRUE,0,(vertex_count+1)*sizeof(vertex_record),
parent); //the parent array has the parent and rank information
// for each vertex
for(int i=1;i<=vertex_count;i++)
{
distinct_components.insert(get_parent(i,parent)); //distinct_components will store how many different parents
// are there
}
for(auto u:distinct_components)
std::cout<<u<<",";
std::cout<<"\n";
return distinct_components.size(); //returning the total number of connected components in the graph,
// for a connected graph this should be 1.
//return 0;
}
int main()
{
std::string filename="USA-road-d-for-Boruvka.txt";
int vertex_count=0,edge_count=0,starting_vertex=-1;
std::vector<struct node>edges(1);
take_graph_input_weighted(filename,
edges,
vertex_count,
edge_count,
starting_vertex
);
if(starting_vertex==-1)
{
std::cout<<"graph not read properly from file\n";
exit(0);
}
else
{
std::cout<<"graph read complete\n";
// struct vertex_record* parent=(vertex_record*)malloc
// (sizeof(vertex_record)*(vertex_count+1));
// struct vertex_record **parent=new vertex_record*[vertex_count+1];
struct vertex_record **parent_ptr=(vertex_record**)malloc(sizeof(vertex_record*)
*(vertex_count+1)); //ignore this array of pointers
struct vertex_record *parent=(vertex_record*)malloc(sizeof(vertex_record)
*(vertex_count+1));
int *min_weight=(int*)malloc(sizeof(int)*(vertex_count+1)); //ignore for now
int *min_weight_edge_index_component=(int*)malloc(sizeof(int)*(vertex_count+1)); //ignore for now
for(int i=0;i<=vertex_count;i++)
{
parent[i].parent=i;
parent[i].rank=0;
}
try
{
std::vector<cl::Platform> platforms;
cl::Context context;
std::vector<cl::Device> devices;
cl::CommandQueue queue;
cl::Program program;
cl::Kernel kernel_find_min_edge_weight,kernel_find_min_edge_index,
kernel_combine_components, //this kernel has been used in the driver function
//above
kernel_set_max_edge_weight;
std::string kernel_filename="Boruvka-adopted-version-alternate.cl";
std::string funcname1="find_min_edge_weight";
std::string funcname2="find_min_edge_index";
std::string funcname3="combine_components"; //this kernel function will take an edge
//and perform union find between its vertices.
std::string funcname4="set_max_edge_weight";
device_setup_OpenCL(kernel_filename,
funcname1,
funcname2,
funcname3,
funcname4,
platforms,
context,
devices,
queue,
program,
kernel_find_min_edge_weight,
kernel_find_min_edge_index,
kernel_combine_components,
kernel_set_max_edge_weight);
std::cout<<"Device setup complete\n";
cl::Buffer current_edges_buffer,edges_buffer,parent_buffer,
min_weight_buffer,min_weight_edge_index_component_buffer,parent_ptr_buffer;
cl::Buffer buffer_arr[]={ current_edges_buffer,edges_buffer,
parent_buffer,min_weight_buffer,min_weight_edge_index_component_buffer,
parent_ptr_buffer};
int *current_edges=(int*)malloc(sizeof(int)*(edge_count)); //this will store the edge index on
//which each thread will operate
for(int i=1;i<=edge_count;i++)
{
current_edges[i-1]=i;
}
buffer_setup(context,
queue,
buffer_arr,
current_edges,
edges,
parent,
parent_ptr,
min_weight,
min_weight_edge_index_component,
vertex_count,
edge_count);
std::cout<<"Buffer setup success\n";
int component_count=0;
component_count=count_components(queue,
buffer_arr,
parent,
kernel_combine_components,
kernel_set_max_edge_weight,
current_edges,
edge_count,
vertex_count);
std::cout<<component_count<<"\n";
}
catch(cl::Error error)
{
std::cout << error.what() << "(" << error.err() << ")" << std::endl;
exit(0);
}
}
}