OpenFPM_pdata  4.1.0
Project that contain the implementation of distributed structures
vector_dist_funcs.hpp
1 /*
2  * vector_dist_funcs.hpp
3  *
4  * Created on: Aug 15, 2018
5  * Author: i-bird
6  */
7 
8 #ifndef VECTOR_DIST_FUNCS_HPP_
9 #define VECTOR_DIST_FUNCS_HPP_
10 
12 struct proc_without_prp
13 {
15  template<typename T1, typename T2> inline static void proc(size_t lbl, size_t cnt, size_t id, T1 & v_prp, T2 & m_prp)
16  {
17  m_prp.get(lbl).set(cnt, v_prp.get(id));
18  }
19 };
20 
22 struct proc_without_prp_device
23 {
25  template<typename T1, typename T2> __device__ inline static void proc(size_t cnt, size_t id, T1 & v_prp, T2 & m_prp)
26  {
27  m_prp.set(cnt, v_prp.get(id));
28  }
29 };
30 
31 
40 template<typename m_opart_type>
41 inline size_t get_end_valid(long int & end, long int & end_id, m_opart_type & m_opart)
42 {
43  end_id--;
44 
45  while (end >= 0 && end_id >= 0 && (long int)m_opart.template get<0>(end) == end_id)
46  {
47  end_id--;
48  end--;
49  }
50 
51  return end_id;
52 }
53 
55 
70 template<typename proc_class, typename Top,typename Pmr, typename T1, typename T2, typename T3, typename T4>
71 inline void process_map_particle(size_t i, long int & end, long int & id_end, Top & m_opart, Pmr p_map_req, T1 & m_pos, T2 & m_prp, T3 & v_pos, T4 & v_prp, openfpm::vector<size_t> & cnt)
72 {
73  long int prc_id = m_opart.template get<2>(i);
74  size_t id = m_opart.template get<0>(i);
75 
76  if (prc_id >= 0)
77  {
78  size_t lbl = p_map_req.get(prc_id);
79 
80  m_pos.get(lbl).set(cnt.get(lbl), v_pos.get(id));
81  proc_class::proc(lbl,cnt.get(lbl),id,v_prp,m_prp);
82 
83  cnt.get(lbl)++;
84 
85  // swap the particle
86  // If a particle migrate we have an hole we cover this hole using a particle from the end of the vector
87  long int id_valid = get_end_valid(end,id_end,m_opart);
88 
89  if (id_valid > 0 && (long int)id < id_valid)
90  {
91  v_pos.set(id,v_pos.get(id_valid));
92  v_prp.set(id,v_prp.get(id_valid));
93  }
94  }
95  else
96  {
97  // swap the particle
98  long int id_valid = get_end_valid(end,id_end,m_opart);
99 
100  if (id_valid > 0 && (long int)id < id_valid)
101  {
102  v_pos.set(id,v_pos.get(id_valid));
103  v_prp.set(id,v_prp.get(id_valid));
104  }
105  }
106 }
107 
108 
110 template<typename proc_class, typename Top, typename T1, typename T2, typename T3, typename T4>
111 __device__ inline void process_map_device_particle(unsigned int i, unsigned int offset, Top & m_opart, T1 & m_pos, T2 & m_prp, T3 & v_pos, T4 & v_prp)
112 {
113  size_t id = m_opart.template get<0>(i+offset);
114 
115  m_pos.set(i, v_pos.get(id));
116  proc_class::proc(i,id,v_prp,m_prp);
117 }
118 
119 
121 template<typename Top, typename T2, typename T4, unsigned int ... prp>
122 __device__ inline void process_ghost_device_particle_prp(unsigned int i, unsigned int offset, Top & g_opart, T2 & m_prp, T4 & v_prp)
123 {
124  unsigned int id = g_opart.template get<1>(i+offset) & 0xFFFFFFFF;
125 
126  // source object type
127  typedef decltype(v_prp.get(id)) encap_src;
128  // destination object type
129  typedef decltype(m_prp.get(i)) encap_dst;
130 
131  // Copy only the selected properties
132  object_si_d<encap_src, encap_dst, OBJ_ENCAP, prp...>(v_prp.get(id), m_prp.get(i));
133 }
134 
136 template<typename T2, typename T4, unsigned int ... prp>
137 __device__ inline void process_ghost_device_particle_prp(unsigned int i, unsigned int offset, T2 & m_prp, T4 & v_prp)
138 {
139  unsigned int id = i+offset;
140 
141  // source object type
142  typedef decltype(v_prp.get(id)) encap_src;
143  // destination object type
144  typedef decltype(m_prp.get(i)) encap_dst;
145 
146  // Copy only the selected properties
147  object_si_d<encap_src, encap_dst, OBJ_ENCAP, prp...>(v_prp.get(id), m_prp.get(i));
148 }
149 
150 template<typename base_type, unsigned int prp>
151 struct compare_host_device
152 {
153  template<typename St, typename vector_type>
154  static bool compare(vector_type & v_prp,St & tol, St & near, bool silent = false)
155  {
156  bool ret = true;
157 
158  // Create a temporal
159  openfpm::vector<aggregate<base_type>> tmp;
160 
161  tmp.resize(v_prp.size());
162 
163  // move host memory to tmp
164  auto it = v_prp.getIterator();
165 
166  while (it.isNext())
167  {
168  auto p = it.get();
169 
170  tmp.template get<0>(p) = v_prp.template get<prp>(p);
171 
172  ++it;
173  }
174 
175  v_prp.template deviceToHost<prp>();
176 
177  // move host memory to tmp
178  it = v_prp.getIterator();
179 
180  while (it.isNext())
181  {
182  auto p = it.get();
183 
184  if (fabs(tmp.template get<0>(p) - v_prp.template get<prp>(p)) >= tol && (fabs(tmp.template get<0>(p)) > near && fabs(v_prp.template get<prp>(p)) ) )
185  {
186  std::cout << "Host and Device buffer differ over set tollerance: " << "Host[" << p << "]=" << tmp.template get<0>(p)
187  << " Device[" << p << "]="<< v_prp.template get<0>(p) <<
188  " differ more than: " << tol << std::endl;
189  ret = false;
190  }
191 
192  ++it;
193  }
194 
195  //restore
196  it = tmp.getIterator();
197 
198  while (it.isNext())
199  {
200  auto p = it.get();
201 
202  v_prp.template get<prp>(p) = tmp.template get<0>(p);
203 
204  ++it;
205  }
206 
207  return ret;
208  }
209 };
210 
211 template<typename base_type,unsigned int N1, unsigned int prp>
212 struct compare_host_device<Point<N1,base_type>,prp>
213 {
214  template<typename St, typename vector_type>
215  static bool compare(vector_type & v_pos,St & tol, St & near, bool silent = false)
216  {
217  bool ret = true;
218 
219  // Create a temporal
220  openfpm::vector<Point<N1,base_type>> tmp;
221 
222  tmp.resize(v_pos.size());
223 
224  // move host memory to tmp
225  auto it = v_pos.getIterator();
226 
227  while (it.isNext())
228  {
229  auto p = it.get();
230 
231  tmp.get(p) = v_pos.get(p);
232 
233  ++it;
234  }
235 
236  v_pos.template deviceToHost<prp>();
237 
238  // move host memory to tmp
239  it = v_pos.getIterator();
240 
241  while (it.isNext())
242  {
243  auto p = it.get();
244 
245  for (size_t j = 0 ; j < N1 ; j++)
246  {
247  if (fabs(tmp.template get<0>(p)[j] - v_pos.template get<0>(p)[j]) >= tol && (fabs(tmp.template get<0>(p)[j]) > near && fabs(v_pos.template get<0>(p)[j]) ) )
248  {
249  std::cout << "Host and Device buffer differ over set tollerance: " << "Host[" << p << "][" << j <<"]=" << tmp.template get<0>(p)[j]
250  << " Device[" << p << "][" << j << "]=" << v_pos.template get<0>(p)[j] <<
251  " differ more than: " << tol << std::endl;
252  ret = false;
253  }
254  }
255 
256  ++it;
257  }
258 
259  //restore
260  it = tmp.getIterator();
261 
262  while (it.isNext())
263  {
264  auto p = it.get();
265 
266  v_pos.get(p) = tmp.get(p);
267 
268  ++it;
269  }
270 
271  return ret;
272  }
273 };
274 
275 template<typename base_type,unsigned int N1, unsigned int prp>
276 struct compare_host_device<base_type[N1],prp>
277 {
278  template<typename St, typename vector_type>
279  static bool compare(vector_type & v_prp,St & tol, St & near, bool silent = false)
280  {
281  bool ret = true;
282 
283  // Create a temporal
284  openfpm::vector<aggregate<base_type[N1]>> tmp;
285 
286  tmp.resize(v_prp.size());
287 
288  // move host memory to tmp
289  auto it = v_prp.getIterator();
290 
291  while (it.isNext())
292  {
293  auto p = it.get();
294 
295  for (size_t j = 0 ; j < N1 ; j++)
296  {
297  tmp.template get<0>(p)[j] = v_prp.template get<prp>(p)[j];
298  }
299 
300  ++it;
301  }
302 
303  v_prp.template deviceToHost<prp>();
304 
305  // move host memory to tmp
306  it = v_prp.getIterator();
307 
308  while (it.isNext())
309  {
310  auto p = it.get();
311 
312  for (size_t j = 0 ; j < N1 ; j++)
313  {
314  if (fabs(tmp.template get<0>(p)[j] - v_prp.template get<prp>(p)[j]) >= tol && (fabs(tmp.template get<0>(p)[j]) > near && fabs(v_prp.template get<prp>(p)[j]) ) )
315  {
316  std::cout << "Host and Device buffer differ over set tollerance: " << "Host[" << p << "]=" << tmp.template get<0>(p)[j]
317  << " Device[" << p << "]="<< v_prp.template get<prp>(p)[j] <<
318  " differ more than: " << tol << std::endl;
319  ret = false;
320  }
321  }
322 
323  ++it;
324  }
325 
326  //restore
327  it = v_prp.getIterator();
328 
329  while (it.isNext())
330  {
331  auto p = it.get();
332 
333  for (size_t j = 0 ; j < N1 ; j++)
334  {
335  v_prp.template get<prp>(p)[j] = tmp.template get<0>(p)[j];
336  }
337 
338  ++it;
339  }
340 
341  return ret;
342  }
343 };
344 
345 template<typename base_type,unsigned int N1 , unsigned int N2, unsigned int prp>
346 struct compare_host_device<base_type[N1][N2],prp>
347 {
348  template<typename St, typename vector_type>
349  static bool compare(vector_type & v_prp,St & tol, St & near, bool silent = false)
350  {
351  bool ret = true;
352 
353  // Create a temporal
354  openfpm::vector<aggregate<base_type[N1][N2]>> tmp;
355 
356  tmp.resize(v_prp.size());
357 
358  // move host memory to tmp
359  auto it = v_prp.getIterator();
360 
361  while (it.isNext())
362  {
363  auto p = it.get();
364 
365  for (size_t j = 0 ; j < N1 ; j++)
366  {
367  for (size_t k = 0 ; k < N2 ; k++)
368  {
369  tmp.template get<0>(p)[j][k] = v_prp.template get<prp>(p)[j][k];
370  }
371  }
372 
373  ++it;
374  }
375 
376  v_prp.template deviceToHost<prp>();
377 
378  // move host memory to tmp
379  it = v_prp.getIterator();
380 
381  while (it.isNext())
382  {
383  auto p = it.get();
384 
385  for (size_t j = 0 ; j < N1 ; j++)
386  {
387  for (size_t k = 0 ; k < N2 ; k++)
388  {
389  if (fabs(tmp.template get<0>(p)[j][k] - v_prp.template get<prp>(p)[j][k]) >= tol && (fabs(tmp.template get<0>(p)[j][k]) > near && fabs(v_prp.template get<prp>(p)[j][k]) ) )
390  {
391  std::cout << "Host and Device buffer differ over set tollerance: " << "Host[" << p << "][" << j << "][" << k << "]=" << tmp.template get<0>(p)[j][k]
392  << " Device[" << p << "][" << j << "][" << k << "]=" << v_prp.template get<prp>(p)[j][k] << " differ more than: " << tol << std::endl;
393 
394  ret = false;
395  }
396  }
397  }
398 
399  ++it;
400  }
401 
402  //restore
403  it = v_prp.getIterator();
404 
405  while (it.isNext())
406  {
407  auto p = it.get();
408 
409  for (size_t j = 0 ; j < N1 ; j++)
410  {
411  for (size_t k = 0 ; k < N2 ; k++)
412  {
413  v_prp.template get<prp>(p)[j][k] = tmp.template get<0>(p)[j][k];
414  }
415  }
416 
417  ++it;
418  }
419 
420  return ret;
421  }
422 };
423 
424 #endif /* VECTOR_DIST_FUNCS_HPP_ */
proc_without_prp::proc
static void proc(size_t lbl, size_t cnt, size_t id, T1 &v_prp, T2 &m_prp)
process the particle
Definition: vector_dist_funcs.hpp:15
proc_without_prp_device
process the particle without properties
Definition: vector_dist_funcs.hpp:22
Point
This class implement the point shape in an N-dimensional space.
Definition: Point.hpp:27
compare_host_device
Definition: vector_dist_funcs.hpp:151
vector_dist::getIterator
vector_dist_iterator getIterator()
Get an iterator that traverse domain and ghost particles.
Definition: vector_dist.hpp:2037
proc_without_prp
process the particle without properties
Definition: vector_dist_funcs.hpp:12
vector_dist
Distributed vector.
Definition: vector_dist.hpp:297
vector_dist_iterator::get
vect_dist_key_dx get()
Get the actual key.
Definition: vector_dist_iterator.hpp:75
proc_without_prp_device::proc
static __device__ void proc(size_t cnt, size_t id, T1 &v_prp, T2 &m_prp)
process the particle
Definition: vector_dist_funcs.hpp:25
object_si_d
It copy the properties from one object to another.
Definition: object_si_d.hpp:189