/[escript]/branches/ripleygmg_from_3668/ripley/src/Rectangle.cpp
ViewVC logotype

Contents of /branches/ripleygmg_from_3668/ripley/src/Rectangle.cpp

Parent Directory Parent Directory | Revision Log Revision Log

Revision 3759 - (show annotations)
Fri Jan 6 06:54:51 2012 UTC (7 years, 11 months ago) by caltinay
File size: 105801 byte(s) 
Implemented ownSample for face elements.
1
2 /*******************************************************
3 *
4 * Copyright (c) 2003-2011 by University of Queensland
5 * Earth Systems Science Computational Center (ESSCC)
6 * http://www.uq.edu.au/esscc
7 *
8 * Primary Business: Queensland, Australia
9 * Licensed under the Open Software License version 3.0
10 * http://www.opensource.org/licenses/osl-3.0.php
11 *
12 *******************************************************/
13
14 #include <ripley/Rectangle.h>
15 extern "C" {
16 #include "paso/SystemMatrix.h"
17 }
18
19 #if USE_SILO
20 #include <silo.h>
21 #ifdef ESYS_MPI
22 #include <pmpio.h>
23 #endif
24 #endif
25
26 #include <iomanip>
27
28 using namespace std;
29
30 namespace ripley {
31
32 Rectangle::Rectangle(int n0, int n1, double l0, double l1, int d0, int d1) :
33 RipleyDomain(2),
34 m_gNE0(n0),
35 m_gNE1(n1),
36 m_l0(l0),
37 m_l1(l1),
38 m_NX(d0),
39 m_NY(d1)
40 {
41 // ensure number of subdivisions is valid and nodes can be distributed
42 // among number of ranks
43 if (m_NX*m_NY != m_mpiInfo->size)
44 throw RipleyException("Invalid number of spatial subdivisions");
45
46 if ((n0+1)%m_NX > 0 || (n1+1)%m_NY > 0)
47 throw RipleyException("Number of elements+1 must be separable into number of ranks in each dimension");
48
49 if ((m_NX > 1 && (n0+1)/m_NX<2) || (m_NY > 1 && (n1+1)/m_NY<2))
50 throw RipleyException("Too few elements for the number of ranks");
51
52 // local number of elements (including overlap)
53 m_NE0 = (m_NX>1 ? (n0+1)/m_NX : n0);
54 if (m_mpiInfo->rank%m_NX>0 && m_mpiInfo->rank%m_NX<m_NX-1)
55 m_NE0++;
56 m_NE1 = (m_NY>1 ? (n1+1)/m_NY : n1);
57 if (m_mpiInfo->rank/m_NX>0 && m_mpiInfo->rank/m_NX<m_NY-1)
58 m_NE1++;
59
60 // local number of nodes
61 m_N0 = m_NE0+1;
62 m_N1 = m_NE1+1;
63
64 // bottom-left node is at (offset0,offset1) in global mesh
65 m_offset0 = (n0+1)/m_NX*(m_mpiInfo->rank%m_NX);
66 if (m_offset0 > 0)
67 m_offset0--;
68 m_offset1 = (n1+1)/m_NY*(m_mpiInfo->rank/m_NX);
69 if (m_offset1 > 0)
70 m_offset1--;
71
72 populateSampleIds();
73 createPattern();
74 }
75
76 Rectangle::~Rectangle()
77 {
78 }
79
80 string Rectangle::getDescription() const
81 {
82 return "ripley::Rectangle";
83 }
84
85 bool Rectangle::operator==(const AbstractDomain& other) const
86 {
87 const Rectangle* o=dynamic_cast<const Rectangle*>(&other);
88 if (o) {
89 return (RipleyDomain::operator==(other) &&
90 m_gNE0==o->m_gNE0 && m_gNE1==o->m_gNE1
91 && m_l0==o->m_l0 && m_l1==o->m_l1
92 && m_NX==o->m_NX && m_NY==o->m_NY);
93 }
94
95 return false;
96 }
97
98 void Rectangle::dump(const string& fileName) const
99 {
100 #if USE_SILO
101 string fn(fileName);
102 if (fileName.length() < 6 || fileName.compare(fileName.length()-5, 5, ".silo") != 0) {
103 fn+=".silo";
104 }
105
106 const int NUM_SILO_FILES = 1;
107 const char* blockDirFmt = "/block%04d";
108 int driver=DB_HDF5;
109 string siloPath;
110 DBfile* dbfile = NULL;
111
112 #ifdef ESYS_MPI
113 PMPIO_baton_t* baton = NULL;
114 #endif
115
116 if (m_mpiInfo->size > 1) {
117 #ifdef ESYS_MPI
118 baton = PMPIO_Init(NUM_SILO_FILES, PMPIO_WRITE, m_mpiInfo->comm,
119 0x1337, PMPIO_DefaultCreate, PMPIO_DefaultOpen,
120 PMPIO_DefaultClose, (void*)&driver);
121 // try the fallback driver in case of error
122 if (!baton && driver != DB_PDB) {
123 driver = DB_PDB;
124 baton = PMPIO_Init(NUM_SILO_FILES, PMPIO_WRITE, m_mpiInfo->comm,
125 0x1338, PMPIO_DefaultCreate, PMPIO_DefaultOpen,
126 PMPIO_DefaultClose, (void*)&driver);
127 }
128 if (baton) {
129 char str[64];
130 snprintf(str, 64, blockDirFmt, PMPIO_RankInGroup(baton, m_mpiInfo->rank));
131 siloPath = str;
132 dbfile = (DBfile*) PMPIO_WaitForBaton(baton, fn.c_str(), siloPath.c_str());
133 }
134 #endif
135 } else {
136 dbfile = DBCreate(fn.c_str(), DB_CLOBBER, DB_LOCAL,
137 getDescription().c_str(), driver);
138 // try the fallback driver in case of error
139 if (!dbfile && driver != DB_PDB) {
140 driver = DB_PDB;
141 dbfile = DBCreate(fn.c_str(), DB_CLOBBER, DB_LOCAL,
142 getDescription().c_str(), driver);
143 }
144 }
145
146 if (!dbfile)
147 throw RipleyException("dump: Could not create Silo file");
148
149 /*
150 if (driver==DB_HDF5) {
151 // gzip level 1 already provides good compression with minimal
152 // performance penalty. Some tests showed that gzip levels >3 performed
153 // rather badly on escript data both in terms of time and space
154 DBSetCompression("ERRMODE=FALLBACK METHOD=GZIP LEVEL=1");
155 }
156 */
157
158 boost::scoped_ptr<double> x(new double[m_N0]);
159 boost::scoped_ptr<double> y(new double[m_N1]);
160 double* coords[2] = { x.get(), y.get() };
161 pair<double,double> xdx = getFirstCoordAndSpacing(0);
162 pair<double,double> ydy = getFirstCoordAndSpacing(1);
163 #pragma omp parallel
164 {
165 #pragma omp for nowait
166 for (dim_t i0 = 0; i0 < m_N0; i0++) {
167 coords[0][i0]=xdx.first+i0*xdx.second;
168 }
169 #pragma omp for nowait
170 for (dim_t i1 = 0; i1 < m_N1; i1++) {
171 coords[1][i1]=ydy.first+i1*ydy.second;
172 }
173 }
174 IndexVector dims = getNumNodesPerDim();
175
176 // write mesh
177 DBPutQuadmesh(dbfile, "mesh", NULL, coords, &dims[0], 2, DB_DOUBLE,
178 DB_COLLINEAR, NULL);
179
180 // write node ids
181 DBPutQuadvar1(dbfile, "nodeId", "mesh", (void*)&m_nodeId[0], &dims[0], 2,
182 NULL, 0, DB_INT, DB_NODECENT, NULL);
183
184 // write element ids
185 dims = getNumElementsPerDim();
186 DBPutQuadvar1(dbfile, "elementId", "mesh", (void*)&m_elementId[0],
187 &dims[0], 2, NULL, 0, DB_INT, DB_ZONECENT, NULL);
188
189 // rank 0 writes multimesh and multivar
190 if (m_mpiInfo->rank == 0) {
191 vector<string> tempstrings;
192 vector<char*> names;
193 for (dim_t i=0; i<m_mpiInfo->size; i++) {
194 stringstream path;
195 path << "/block" << setw(4) << setfill('0') << right << i << "/mesh";
196 tempstrings.push_back(path.str());
197 names.push_back((char*)tempstrings.back().c_str());
198 }
199 vector<int> types(m_mpiInfo->size, DB_QUAD_RECT);
200 DBSetDir(dbfile, "/");
201 DBPutMultimesh(dbfile, "multimesh", m_mpiInfo->size, &names[0],
202 &types[0], NULL);
203 tempstrings.clear();
204 names.clear();
205 for (dim_t i=0; i<m_mpiInfo->size; i++) {
206 stringstream path;
207 path << "/block" << setw(4) << setfill('0') << right << i << "/nodeId";
208 tempstrings.push_back(path.str());
209 names.push_back((char*)tempstrings.back().c_str());
210 }
211 types.assign(m_mpiInfo->size, DB_QUADVAR);
212 DBPutMultivar(dbfile, "nodeId", m_mpiInfo->size, &names[0],
213 &types[0], NULL);
214 tempstrings.clear();
215 names.clear();
216 for (dim_t i=0; i<m_mpiInfo->size; i++) {
217 stringstream path;
218 path << "/block" << setw(4) << setfill('0') << right << i << "/elementId";
219 tempstrings.push_back(path.str());
220 names.push_back((char*)tempstrings.back().c_str());
221 }
222 DBPutMultivar(dbfile, "elementId", m_mpiInfo->size, &names[0],
223 &types[0], NULL);
224 }
225
226 if (m_mpiInfo->size > 1) {
227 #ifdef ESYS_MPI
228 PMPIO_HandOffBaton(baton, dbfile);
229 PMPIO_Finish(baton);
230 #endif
231 } else {
232 DBClose(dbfile);
233 }
234
235 #else // USE_SILO
236 throw RipleyException("dump(): no Silo support");
237 #endif
238 }
239
240 const int* Rectangle::borrowSampleReferenceIDs(int fsType) const
241 {
242 switch (fsType) {
243 case Nodes:
244 case ReducedNodes: //FIXME: reduced
245 return &m_nodeId[0];
246 case DegreesOfFreedom:
247 case ReducedDegreesOfFreedom: //FIXME: reduced
248 return &m_dofId[0];
249 case Elements:
250 case ReducedElements:
251 return &m_elementId[0];
252 case FaceElements:
253 case ReducedFaceElements:
254 return &m_faceId[0];
255 default:
256 break;
257 }
258
259 stringstream msg;
260 msg << "borrowSampleReferenceIDs() not implemented for function space type "
261 << functionSpaceTypeAsString(fsType);
262 throw RipleyException(msg.str());
263 }
264
265 bool Rectangle::ownSample(int fsType, index_t id) const
266 {
267 if (getMPISize()==1)
268 return true;
269
270 switch (fsType) {
271 case Nodes:
272 case ReducedNodes: //FIXME: reduced
273 return (m_dofMap[id] < getNumDOF());
274 case DegreesOfFreedom:
275 case ReducedDegreesOfFreedom:
276 return true;
277 case Elements:
278 case ReducedElements:
279 // check ownership of element's bottom left node
280 return (m_dofMap[id%m_NE0+m_N0*(id/m_NE0)] < getNumDOF());
281 case FaceElements:
282 case ReducedFaceElements:
283 {
284 // check ownership of face element's first node
285 const IndexVector faces = getNumFacesPerBoundary();
286 dim_t n=0;
287 for (size_t i=0; i<faces.size(); i++) {
288 n+=faces[i];
289 if (id<n) {
290 index_t k;
291 if (i==1)
292 k=m_N0-1;
293 else if (i==3)
294 k=m_N0*(m_N1-1);
295 else
296 k=0;
297 // determine whether to move right or up
298 const index_t delta=(i/2==0 ? m_N0 : 1);
299 return (m_dofMap[k+(id-n+faces[i])*delta] < getNumDOF());
300 }
301 }
302 return false;
303 }
304 default:
305 break;
306 }
307
308 stringstream msg;
309 msg << "ownSample() not implemented for "
310 << functionSpaceTypeAsString(fsType);
311 throw RipleyException(msg.str());
312 }
313
314 void Rectangle::setToGradient(escript::Data& out, const escript::Data& cIn) const
315 {
316 escript::Data& in = *const_cast<escript::Data*>(&cIn);
317 const dim_t numComp = in.getDataPointSize();
318 const double h0 = m_l0/m_gNE0;
319 const double h1 = m_l1/m_gNE1;
320 const double cx0 = -1./h0;
321 const double cx1 = -.78867513459481288225/h0;
322 const double cx2 = -.5/h0;
323 const double cx3 = -.21132486540518711775/h0;
324 const double cx4 = .21132486540518711775/h0;
325 const double cx5 = .5/h0;
326 const double cx6 = .78867513459481288225/h0;
327 const double cx7 = 1./h0;
328 const double cy0 = -1./h1;
329 const double cy1 = -.78867513459481288225/h1;
330 const double cy2 = -.5/h1;
331 const double cy3 = -.21132486540518711775/h1;
332 const double cy4 = .21132486540518711775/h1;
333 const double cy5 = .5/h1;
334 const double cy6 = .78867513459481288225/h1;
335 const double cy7 = 1./h1;
336
337 if (out.getFunctionSpace().getTypeCode() == Elements) {
338 /*** GENERATOR SNIP_GRAD_ELEMENTS TOP */
339 #pragma omp parallel for
340 for (index_t k1=0; k1 < m_NE1; ++k1) {
341 for (index_t k0=0; k0 < m_NE0; ++k0) {
342 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,k1, m_N0));
343 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,k1+1, m_N0));
344 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,k1+1, m_N0));
345 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,k1, m_N0));
346 double* o = out.getSampleDataRW(INDEX2(k0,k1,m_NE0));
347 for (index_t i=0; i < numComp; ++i) {
348 o[INDEX3(i,0,0,numComp,2)] = f_00[i]*cx1 + f_01[i]*cx3 + f_10[i]*cx6 + f_11[i]*cx4;
349 o[INDEX3(i,1,0,numComp,2)] = f_00[i]*cy1 + f_01[i]*cy6 + f_10[i]*cy3 + f_11[i]*cy4;
350 o[INDEX3(i,0,1,numComp,2)] = f_00[i]*cx1 + f_01[i]*cx3 + f_10[i]*cx6 + f_11[i]*cx4;
351 o[INDEX3(i,1,1,numComp,2)] = f_00[i]*cy3 + f_01[i]*cy4 + f_10[i]*cy1 + f_11[i]*cy6;
352 o[INDEX3(i,0,2,numComp,2)] = f_00[i]*cx3 + f_01[i]*cx1 + f_10[i]*cx4 + f_11[i]*cx6;
353 o[INDEX3(i,1,2,numComp,2)] = f_00[i]*cy1 + f_01[i]*cy6 + f_10[i]*cy3 + f_11[i]*cy4;
354 o[INDEX3(i,0,3,numComp,2)] = f_00[i]*cx3 + f_01[i]*cx1 + f_10[i]*cx4 + f_11[i]*cx6;
355 o[INDEX3(i,1,3,numComp,2)] = f_00[i]*cy3 + f_01[i]*cy4 + f_10[i]*cy1 + f_11[i]*cy6;
356 } /* end of component loop i */
357 } /* end of k0 loop */
358 } /* end of k1 loop */
359 /* GENERATOR SNIP_GRAD_ELEMENTS BOTTOM */
360 } else if (out.getFunctionSpace().getTypeCode() == ReducedElements) {
361 /*** GENERATOR SNIP_GRAD_REDUCED_ELEMENTS TOP */
362 #pragma omp parallel for
363 for (index_t k1=0; k1 < m_NE1; ++k1) {
364 for (index_t k0=0; k0 < m_NE0; ++k0) {
365 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,k1, m_N0));
366 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,k1+1, m_N0));
367 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,k1+1, m_N0));
368 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,k1, m_N0));
369 double* o = out.getSampleDataRW(INDEX2(k0,k1,m_NE0));
370 for (index_t i=0; i < numComp; ++i) {
371 o[INDEX3(i,0,0,numComp,2)] = cx5*(f_10[i] + f_11[i]) + cx2*(f_00[i] + f_01[i]);
372 o[INDEX3(i,1,0,numComp,2)] = cy2*(f_00[i] + f_10[i]) + cy5*(f_01[i] + f_11[i]);
373 } /* end of component loop i */
374 } /* end of k0 loop */
375 } /* end of k1 loop */
376 /* GENERATOR SNIP_GRAD_REDUCED_ELEMENTS BOTTOM */
377 } else if (out.getFunctionSpace().getTypeCode() == FaceElements) {
378 #pragma omp parallel
379 {
380 /*** GENERATOR SNIP_GRAD_FACES TOP */
381 if (m_faceOffset[0] > -1) {
382 #pragma omp for nowait
383 for (index_t k1=0; k1 < m_NE1; ++k1) {
384 const register double* f_10 = in.getSampleDataRO(INDEX2(1,k1, m_N0));
385 const register double* f_11 = in.getSampleDataRO(INDEX2(1,k1+1, m_N0));
386 const register double* f_01 = in.getSampleDataRO(INDEX2(0,k1+1, m_N0));
387 const register double* f_00 = in.getSampleDataRO(INDEX2(0,k1, m_N0));
388 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
389 for (index_t i=0; i < numComp; ++i) {
390 o[INDEX3(i,0,0,numComp,2)] = f_00[i]*cx1 + f_01[i]*cx3 + f_10[i]*cx6 + f_11[i]*cx4;
391 o[INDEX3(i,1,0,numComp,2)] = f_00[i]*cy0 + f_01[i]*cy7;
392 o[INDEX3(i,0,1,numComp,2)] = f_00[i]*cx3 + f_01[i]*cx1 + f_10[i]*cx4 + f_11[i]*cx6;
393 o[INDEX3(i,1,1,numComp,2)] = f_00[i]*cy0 + f_01[i]*cy7;
394 } /* end of component loop i */
395 } /* end of k1 loop */
396 } /* end of face 0 */
397 if (m_faceOffset[1] > -1) {
398 #pragma omp for nowait
399 for (index_t k1=0; k1 < m_NE1; ++k1) {
400 const register double* f_10 = in.getSampleDataRO(INDEX2(m_N0-1,k1, m_N0));
401 const register double* f_11 = in.getSampleDataRO(INDEX2(m_N0-1,k1+1, m_N0));
402 const register double* f_01 = in.getSampleDataRO(INDEX2(m_N0-2,k1+1, m_N0));
403 const register double* f_00 = in.getSampleDataRO(INDEX2(m_N0-2,k1, m_N0));
404 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
405 for (index_t i=0; i < numComp; ++i) {
406 o[INDEX3(i,0,0,numComp,2)] = f_00[i]*cx1 + f_01[i]*cx3 + f_10[i]*cx6 + f_11[i]*cx4;
407 o[INDEX3(i,1,0,numComp,2)] = f_10[i]*cy0 + f_11[i]*cy7;
408 o[INDEX3(i,0,1,numComp,2)] = f_00[i]*cx3 + f_01[i]*cx1 + f_10[i]*cx4 + f_11[i]*cx6;
409 o[INDEX3(i,1,1,numComp,2)] = f_10[i]*cy0 + f_11[i]*cy7;
410 } /* end of component loop i */
411 } /* end of k1 loop */
412 } /* end of face 1 */
413 if (m_faceOffset[2] > -1) {
414 #pragma omp for nowait
415 for (index_t k0=0; k0 < m_NE0; ++k0) {
416 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,0, m_N0));
417 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,0, m_N0));
418 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,1, m_N0));
419 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,1, m_N0));
420 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
421 for (index_t i=0; i < numComp; ++i) {
422 o[INDEX3(i,0,0,numComp,2)] = f_00[i]*cx0 + f_10[i]*cx7;
423 o[INDEX3(i,1,0,numComp,2)] = f_00[i]*cy1 + f_01[i]*cy6 + f_10[i]*cy3 + f_11[i]*cy4;
424 o[INDEX3(i,0,1,numComp,2)] = f_00[i]*cx0 + f_10[i]*cx7;
425 o[INDEX3(i,1,1,numComp,2)] = f_00[i]*cy3 + f_01[i]*cy4 + f_10[i]*cy1 + f_11[i]*cy6;
426 } /* end of component loop i */
427 } /* end of k0 loop */
428 } /* end of face 2 */
429 if (m_faceOffset[3] > -1) {
430 #pragma omp for nowait
431 for (index_t k0=0; k0 < m_NE0; ++k0) {
432 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,m_N1-1, m_N0));
433 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,m_N1-1, m_N0));
434 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,m_N1-2, m_N0));
435 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,m_N1-2, m_N0));
436 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
437 for (index_t i=0; i < numComp; ++i) {
438 o[INDEX3(i,0,0,numComp,2)] = f_01[i]*cx0 + f_11[i]*cx7;
439 o[INDEX3(i,1,0,numComp,2)] = f_00[i]*cy1 + f_01[i]*cy6 + f_10[i]*cy3 + f_11[i]*cy4;
440 o[INDEX3(i,0,1,numComp,2)] = f_01[i]*cx0 + f_11[i]*cx7;
441 o[INDEX3(i,1,1,numComp,2)] = f_00[i]*cy3 + f_01[i]*cy4 + f_10[i]*cy1 + f_11[i]*cy6;
442 } /* end of component loop i */
443 } /* end of k0 loop */
444 } /* end of face 3 */
445 /* GENERATOR SNIP_GRAD_FACES BOTTOM */
446 } // end of parallel section
447 } else if (out.getFunctionSpace().getTypeCode() == ReducedFaceElements) {
448 #pragma omp parallel
449 {
450 /*** GENERATOR SNIP_GRAD_REDUCED_FACES TOP */
451 if (m_faceOffset[0] > -1) {
452 #pragma omp for nowait
453 for (index_t k1=0; k1 < m_NE1; ++k1) {
454 const register double* f_10 = in.getSampleDataRO(INDEX2(1,k1, m_N0));
455 const register double* f_11 = in.getSampleDataRO(INDEX2(1,k1+1, m_N0));
456 const register double* f_01 = in.getSampleDataRO(INDEX2(0,k1+1, m_N0));
457 const register double* f_00 = in.getSampleDataRO(INDEX2(0,k1, m_N0));
458 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
459 for (index_t i=0; i < numComp; ++i) {
460 o[INDEX3(i,0,0,numComp,2)] = cx5*(f_10[i] + f_11[i]) + cx2*(f_00[i] + f_01[i]);
461 o[INDEX3(i,1,0,numComp,2)] = f_00[i]*cy0 + f_01[i]*cy7;
462 } /* end of component loop i */
463 } /* end of k1 loop */
464 } /* end of face 0 */
465 if (m_faceOffset[1] > -1) {
466 #pragma omp for nowait
467 for (index_t k1=0; k1 < m_NE1; ++k1) {
468 const register double* f_10 = in.getSampleDataRO(INDEX2(m_N0-1,k1, m_N0));
469 const register double* f_11 = in.getSampleDataRO(INDEX2(m_N0-1,k1+1, m_N0));
470 const register double* f_01 = in.getSampleDataRO(INDEX2(m_N0-2,k1+1, m_N0));
471 const register double* f_00 = in.getSampleDataRO(INDEX2(m_N0-2,k1, m_N0));
472 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
473 for (index_t i=0; i < numComp; ++i) {
474 o[INDEX3(i,0,0,numComp,2)] = cx5*(f_10[i] + f_11[i]) + cx2*(f_00[i] + f_01[i]);
475 o[INDEX3(i,1,0,numComp,2)] = f_10[i]*cy0 + f_11[i]*cy7;
476 } /* end of component loop i */
477 } /* end of k1 loop */
478 } /* end of face 1 */
479 if (m_faceOffset[2] > -1) {
480 #pragma omp for nowait
481 for (index_t k0=0; k0 < m_NE0; ++k0) {
482 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,0, m_N0));
483 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,0, m_N0));
484 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,1, m_N0));
485 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,1, m_N0));
486 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
487 for (index_t i=0; i < numComp; ++i) {
488 o[INDEX3(i,0,0,numComp,2)] = f_00[i]*cx0 + f_10[i]*cx7;
489 o[INDEX3(i,1,0,numComp,2)] = cy2*(f_00[i] + f_10[i]) + cy5*(f_01[i] + f_11[i]);
490 } /* end of component loop i */
491 } /* end of k0 loop */
492 } /* end of face 2 */
493 if (m_faceOffset[3] > -1) {
494 #pragma omp for nowait
495 for (index_t k0=0; k0 < m_NE0; ++k0) {
496 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,m_N1-1, m_N0));
497 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,m_N1-1, m_N0));
498 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,m_N1-2, m_N0));
499 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,m_N1-2, m_N0));
500 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
501 for (index_t i=0; i < numComp; ++i) {
502 o[INDEX3(i,0,0,numComp,2)] = f_01[i]*cx0 + f_11[i]*cx7;
503 o[INDEX3(i,1,0,numComp,2)] = cy5*(f_01[i] + f_11[i]) + cy2*(f_00[i] + f_10[i]);
504 } /* end of component loop i */
505 } /* end of k0 loop */
506 } /* end of face 3 */
507 /* GENERATOR SNIP_GRAD_REDUCED_FACES BOTTOM */
508 } // end of parallel section
509 } else {
510 stringstream msg;
511 msg << "setToGradient() not implemented for "
512 << functionSpaceTypeAsString(out.getFunctionSpace().getTypeCode());
513 throw RipleyException(msg.str());
514 }
515 }
516
517 void Rectangle::setToIntegrals(vector<double>& integrals, const escript::Data& arg) const
518 {
519 escript::Data& in = *const_cast<escript::Data*>(&arg);
520 const dim_t numComp = in.getDataPointSize();
521 const double h0 = m_l0/m_gNE0;
522 const double h1 = m_l1/m_gNE1;
523 if (arg.getFunctionSpace().getTypeCode() == Elements) {
524 const double w_0 = h0*h1/4.;
525 #pragma omp parallel
526 {
527 vector<double> int_local(numComp, 0);
528 #pragma omp for nowait
529 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
530 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
531 const double* f = in.getSampleDataRO(INDEX2(k0, k1, m_NE0));
532 for (index_t i=0; i < numComp; ++i) {
533 const register double f_0 = f[INDEX2(i,0,numComp)];
534 const register double f_1 = f[INDEX2(i,1,numComp)];
535 const register double f_2 = f[INDEX2(i,2,numComp)];
536 const register double f_3 = f[INDEX2(i,3,numComp)];
537 int_local[i]+=(f_0+f_1+f_2+f_3)*w_0;
538 } /* end of component loop i */
539 } /* end of k0 loop */
540 } /* end of k1 loop */
541
542 #pragma omp critical
543 for (index_t i=0; i<numComp; i++)
544 integrals[i]+=int_local[i];
545 } // end of parallel section
546 } else if (arg.getFunctionSpace().getTypeCode() == ReducedElements) {
547 const double w_0 = h0*h1;
548 #pragma omp parallel
549 {
550 vector<double> int_local(numComp, 0);
551 #pragma omp for nowait
552 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
553 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
554 const double* f = in.getSampleDataRO(INDEX2(k0, k1, m_NE0));
555 for (index_t i=0; i < numComp; ++i) {
556 int_local[i]+=f[i]*w_0;
557 } /* end of component loop i */
558 } /* end of k0 loop */
559 } /* end of k1 loop */
560
561 #pragma omp critical
562 for (index_t i=0; i<numComp; i++)
563 integrals[i]+=int_local[i];
564 } // end of parallel section
565 } else if (arg.getFunctionSpace().getTypeCode() == FaceElements) {
566 const double w_0 = h0/2.;
567 const double w_1 = h1/2.;
568 #pragma omp parallel
569 {
570 vector<double> int_local(numComp, 0);
571 if (m_faceOffset[0] > -1) {
572 #pragma omp for nowait
573 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
574 const double* f = in.getSampleDataRO(m_faceOffset[0]+k1);
575 for (index_t i=0; i < numComp; ++i) {
576 const register double f_0 = f[INDEX2(i,0,numComp)];
577 const register double f_1 = f[INDEX2(i,1,numComp)];
578 int_local[i]+=(f_0+f_1)*w_1;
579 } /* end of component loop i */
580 } /* end of k1 loop */
581 }
582
583 if (m_faceOffset[1] > -1) {
584 #pragma omp for nowait
585 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
586 const double* f = in.getSampleDataRO(m_faceOffset[1]+k1);
587 for (index_t i=0; i < numComp; ++i) {
588 const register double f_0 = f[INDEX2(i,0,numComp)];
589 const register double f_1 = f[INDEX2(i,1,numComp)];
590 int_local[i]+=(f_0+f_1)*w_1;
591 } /* end of component loop i */
592 } /* end of k1 loop */
593 }
594
595 if (m_faceOffset[2] > -1) {
596 #pragma omp for nowait
597 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
598 const double* f = in.getSampleDataRO(m_faceOffset[2]+k0);
599 for (index_t i=0; i < numComp; ++i) {
600 const register double f_0 = f[INDEX2(i,0,numComp)];
601 const register double f_1 = f[INDEX2(i,1,numComp)];
602 int_local[i]+=(f_0+f_1)*w_0;
603 } /* end of component loop i */
604 } /* end of k0 loop */
605 }
606
607 if (m_faceOffset[3] > -1) {
608 #pragma omp for nowait
609 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
610 const double* f = in.getSampleDataRO(m_faceOffset[3]+k0);
611 for (index_t i=0; i < numComp; ++i) {
612 const register double f_0 = f[INDEX2(i,0,numComp)];
613 const register double f_1 = f[INDEX2(i,1,numComp)];
614 int_local[i]+=(f_0+f_1)*w_0;
615 } /* end of component loop i */
616 } /* end of k0 loop */
617 }
618
619 #pragma omp critical
620 for (index_t i=0; i<numComp; i++)
621 integrals[i]+=int_local[i];
622 } // end of parallel section
623 } else if (arg.getFunctionSpace().getTypeCode() == ReducedFaceElements) {
624 #pragma omp parallel
625 {
626 vector<double> int_local(numComp, 0);
627 if (m_faceOffset[0] > -1) {
628 #pragma omp for nowait
629 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
630 const double* f = in.getSampleDataRO(m_faceOffset[0]+k1);
631 for (index_t i=0; i < numComp; ++i) {
632 int_local[i]+=f[i]*h1;
633 } /* end of component loop i */
634 } /* end of k1 loop */
635 }
636
637 if (m_faceOffset[1] > -1) {
638 #pragma omp for nowait
639 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
640 const double* f = in.getSampleDataRO(m_faceOffset[1]+k1);
641 for (index_t i=0; i < numComp; ++i) {
642 int_local[i]+=f[i]*h1;
643 } /* end of component loop i */
644 } /* end of k1 loop */
645 }
646
647 if (m_faceOffset[2] > -1) {
648 #pragma omp for nowait
649 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
650 const double* f = in.getSampleDataRO(m_faceOffset[2]+k0);
651 for (index_t i=0; i < numComp; ++i) {
652 int_local[i]+=f[i]*h0;
653 } /* end of component loop i */
654 } /* end of k0 loop */
655 }
656
657 if (m_faceOffset[3] > -1) {
658 #pragma omp for nowait
659 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
660 const double* f = in.getSampleDataRO(m_faceOffset[3]+k0);
661 for (index_t i=0; i < numComp; ++i) {
662 int_local[i]+=f[i]*h0;
663 } /* end of component loop i */
664 } /* end of k0 loop */
665 }
666
667 #pragma omp critical
668 for (index_t i=0; i<numComp; i++)
669 integrals[i]+=int_local[i];
670 } // end of parallel section
671 } else {
672 stringstream msg;
673 msg << "setToIntegrals() not implemented for "
674 << functionSpaceTypeAsString(arg.getFunctionSpace().getTypeCode());
675 throw RipleyException(msg.str());
676 }
677 }
678
679 void Rectangle::setToNormal(escript::Data& out) const
680 {
681 if (out.getFunctionSpace().getTypeCode() == FaceElements) {
682 #pragma omp parallel
683 {
684 if (m_faceOffset[0] > -1) {
685 #pragma omp for nowait
686 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
687 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
688 // set vector at two quadrature points
689 *o++ = -1.;
690 *o++ = 0.;
691 *o++ = -1.;
692 *o = 0.;
693 }
694 }
695
696 if (m_faceOffset[1] > -1) {
697 #pragma omp for nowait
698 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
699 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
700 // set vector at two quadrature points
701 *o++ = 1.;
702 *o++ = 0.;
703 *o++ = 1.;
704 *o = 0.;
705 }
706 }
707
708 if (m_faceOffset[2] > -1) {
709 #pragma omp for nowait
710 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
711 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
712 // set vector at two quadrature points
713 *o++ = 0.;
714 *o++ = -1.;
715 *o++ = 0.;
716 *o = -1.;
717 }
718 }
719
720 if (m_faceOffset[3] > -1) {
721 #pragma omp for nowait
722 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
723 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
724 // set vector at two quadrature points
725 *o++ = 0.;
726 *o++ = 1.;
727 *o++ = 0.;
728 *o = 1.;
729 }
730 }
731 } // end of parallel section
732 } else if (out.getFunctionSpace().getTypeCode() == ReducedFaceElements) {
733 #pragma omp parallel
734 {
735 if (m_faceOffset[0] > -1) {
736 #pragma omp for nowait
737 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
738 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
739 *o++ = -1.;
740 *o = 0.;
741 }
742 }
743
744 if (m_faceOffset[1] > -1) {
745 #pragma omp for nowait
746 for (index_t k1 = 0; k1 < m_NE1; ++k1) {
747 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
748 *o++ = 1.;
749 *o = 0.;
750 }
751 }
752
753 if (m_faceOffset[2] > -1) {
754 #pragma omp for nowait
755 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
756 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
757 *o++ = 0.;
758 *o = -1.;
759 }
760 }
761
762 if (m_faceOffset[3] > -1) {
763 #pragma omp for nowait
764 for (index_t k0 = 0; k0 < m_NE0; ++k0) {
765 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
766 *o++ = 0.;
767 *o = 1.;
768 }
769 }
770 } // end of parallel section
771
772 } else {
773 stringstream msg;
774 msg << "setToNormal() not implemented for "
775 << functionSpaceTypeAsString(out.getFunctionSpace().getTypeCode());
776 throw RipleyException(msg.str());
777 }
778 }
779
780 Paso_SystemMatrixPattern* Rectangle::getPattern(bool reducedRowOrder,
781 bool reducedColOrder) const
782 {
783 if (reducedRowOrder || reducedColOrder)
784 throw RipleyException("getPattern() not implemented for reduced order");
785
786 return m_pattern;
787 }
788
789 void Rectangle::Print_Mesh_Info(const bool full) const
790 {
791 RipleyDomain::Print_Mesh_Info(full);
792 if (full) {
793 cout << " Id Coordinates" << endl;
794 cout.precision(15);
795 cout.setf(ios::scientific, ios::floatfield);
796 pair<double,double> xdx = getFirstCoordAndSpacing(0);
797 pair<double,double> ydy = getFirstCoordAndSpacing(1);
798 for (index_t i=0; i < getNumNodes(); i++) {
799 cout << " " << setw(5) << m_nodeId[i]
800 << " " << xdx.first+(i%m_N0)*xdx.second
801 << " " << ydy.first+(i/m_N0)*ydy.second << endl;
802 }
803 }
804 }
805
806 IndexVector Rectangle::getNumNodesPerDim() const
807 {
808 IndexVector ret;
809 ret.push_back(m_N0);
810 ret.push_back(m_N1);
811 return ret;
812 }
813
814 IndexVector Rectangle::getNumElementsPerDim() const
815 {
816 IndexVector ret;
817 ret.push_back(m_NE0);
818 ret.push_back(m_NE1);
819 return ret;
820 }
821
822 IndexVector Rectangle::getNumFacesPerBoundary() const
823 {
824 IndexVector ret(4, 0);
825 //left
826 if (m_offset0==0)
827 ret[0]=m_NE1;
828 //right
829 if (m_mpiInfo->rank%m_NX==m_NX-1)
830 ret[1]=m_NE1;
831 //bottom
832 if (m_offset1==0)
833 ret[2]=m_NE0;
834 //top
835 if (m_mpiInfo->rank/m_NX==m_NY-1)
836 ret[3]=m_NE0;
837 return ret;
838 }
839
840 pair<double,double> Rectangle::getFirstCoordAndSpacing(dim_t dim) const
841 {
842 if (dim==0) {
843 return pair<double,double>((m_l0*m_offset0)/m_gNE0, m_l0/m_gNE0);
844 } else if (dim==1) {
845 return pair<double,double>((m_l1*m_offset1)/m_gNE1, m_l1/m_gNE1);
846 }
847 throw RipleyException("getFirstCoordAndSpacing(): invalid argument");
848 }
849
850 //protected
851 dim_t Rectangle::getNumDOF() const
852 {
853 return (m_gNE0+1)/m_NX*(m_gNE1+1)/m_NY;
854 }
855
856 //protected
857 dim_t Rectangle::getNumFaceElements() const
858 {
859 const IndexVector faces = getNumFacesPerBoundary();
860 dim_t n=0;
861 for (size_t i=0; i<faces.size(); i++)
862 n+=faces[i];
863 return n;
864 }
865
866 //protected
867 void Rectangle::assembleCoordinates(escript::Data& arg) const
868 {
869 escriptDataC x = arg.getDataC();
870 int numDim = m_numDim;
871 if (!isDataPointShapeEqual(&x, 1, &numDim))
872 throw RipleyException("setToX: Invalid Data object shape");
873 if (!numSamplesEqual(&x, 1, getNumNodes()))
874 throw RipleyException("setToX: Illegal number of samples in Data object");
875
876 pair<double,double> xdx = getFirstCoordAndSpacing(0);
877 pair<double,double> ydy = getFirstCoordAndSpacing(1);
878 arg.requireWrite();
879 #pragma omp parallel for
880 for (dim_t i1 = 0; i1 < m_N1; i1++) {
881 for (dim_t i0 = 0; i0 < m_N0; i0++) {
882 double* point = arg.getSampleDataRW(i0+m_N0*i1);
883 point[0] = xdx.first+i0*xdx.second;
884 point[1] = ydy.first+i1*ydy.second;
885 }
886 }
887 }
888
889 //protected
890 dim_t Rectangle::insertNeighbourNodes(IndexVector& index, index_t node) const
891 {
892 const dim_t nDOF0 = (m_gNE0+1)/m_NX;
893 const dim_t nDOF1 = (m_gNE1+1)/m_NY;
894 const int x=node%nDOF0;
895 const int y=node/nDOF0;
896 dim_t num=0;
897 // loop through potential neighbours and add to index if positions are
898 // within bounds
899 for (int i1=-1; i1<2; i1++) {
900 for (int i0=-1; i0<2; i0++) {
901 // skip node itself
902 if (i0==0 && i1==0)
903 continue;
904 // location of neighbour node
905 const int nx=x+i0;
906 const int ny=y+i1;
907 if (nx>=0 && ny>=0 && nx<nDOF0 && ny<nDOF1) {
908 index.push_back(ny*nDOF0+nx);
909 num++;
910 }
911 }
912 }
913
914 return num;
915 }
916
917 //protected
918 void Rectangle::nodesToDOF(escript::Data& out, escript::Data& in) const
919 {
920 const dim_t numComp = in.getDataPointSize();
921 out.requireWrite();
922
923 const index_t left = (m_offset0==0 ? 0 : 1);
924 const index_t bottom = (m_offset1==0 ? 0 : 1);
925 const dim_t nDOF0 = (m_gNE0+1)/m_NX;
926 const dim_t nDOF1 = (m_gNE1+1)/m_NY;
927 #pragma omp parallel for
928 for (index_t i=0; i<nDOF1; i++) {
929 for (index_t j=0; j<nDOF0; j++) {
930 const index_t n=j+left+(i+bottom)*m_N0;
931 const double* src=in.getSampleDataRO(n);
932 copy(src, src+numComp, out.getSampleDataRW(j+i*nDOF0));
933 }
934 }
935 }
936
937 //protected
938 void Rectangle::dofToNodes(escript::Data& out, escript::Data& in) const
939 {
940 const dim_t numComp = in.getDataPointSize();
941 Paso_Coupler* coupler = Paso_Coupler_alloc(m_connector, numComp);
942 in.requireWrite();
943 Paso_Coupler_startCollect(coupler, in.getSampleDataRW(0));
944
945 const dim_t numDOF = getNumDOF();
946 out.requireWrite();
947 const double* buffer = Paso_Coupler_finishCollect(coupler);
948
949 #pragma omp parallel for
950 for (index_t i=0; i<getNumNodes(); i++) {
951 const double* src=(m_dofMap[i]<numDOF ?
952 in.getSampleDataRO(m_dofMap[i])
953 : &buffer[(m_dofMap[i]-numDOF)*numComp]);
954 copy(src, src+numComp, out.getSampleDataRW(i));
955 }
956 }
957
958 //private
959 void Rectangle::populateSampleIds()
960 {
961 // identifiers are ordered from left to right, bottom to top globablly.
962
963 // build node distribution vector first.
964 // rank i owns m_nodeDistribution[i+1]-nodeDistribution[i] nodes
965 m_nodeDistribution.assign(m_mpiInfo->size+1, 0);
966 const dim_t numDOF=getNumDOF();
967 for (dim_t k=1; k<m_mpiInfo->size; k++) {
968 m_nodeDistribution[k]=k*numDOF;
969 }
970 m_nodeDistribution[m_mpiInfo->size]=getNumDataPointsGlobal();
971 m_nodeId.resize(getNumNodes());
972 m_dofId.resize(numDOF);
973 m_elementId.resize(getNumElements());
974 m_faceId.resize(getNumFaceElements());
975
976 #pragma omp parallel
977 {
978 // nodes
979 #pragma omp for nowait
980 for (dim_t i1=0; i1<m_N1; i1++) {
981 for (dim_t i0=0; i0<m_N0; i0++) {
982 m_nodeId[i0+i1*m_N0] = (m_offset1+i1)*(m_gNE0+1)+m_offset0+i0;
983 }
984 }
985
986 // degrees of freedom
987 #pragma omp for nowait
988 for (dim_t k=0; k<numDOF; k++)
989 m_dofId[k] = m_nodeDistribution[m_mpiInfo->rank]+k;
990
991 // elements
992 #pragma omp for nowait
993 for (dim_t i1=0; i1<m_NE1; i1++) {
994 for (dim_t i0=0; i0<m_NE0; i0++) {
995 m_elementId[i0+i1*m_NE0]=(m_offset1+i1)*m_gNE0+m_offset0+i0;
996 }
997 }
998
999 // face elements
1000 #pragma omp for
1001 for (dim_t k=0; k<getNumFaceElements(); k++)
1002 m_faceId[k]=k;
1003 } // end parallel section
1004
1005 m_nodeTags.assign(getNumNodes(), 0);
1006 updateTagsInUse(Nodes);
1007
1008 m_elementTags.assign(getNumElements(), 0);
1009 updateTagsInUse(Elements);
1010
1011 // generate face offset vector and set face tags
1012 const IndexVector facesPerEdge = getNumFacesPerBoundary();
1013 const index_t LEFT=1, RIGHT=2, BOTTOM=10, TOP=20;
1014 const index_t faceTag[] = { LEFT, RIGHT, BOTTOM, TOP };
1015 m_faceOffset.assign(facesPerEdge.size(), -1);
1016 m_faceTags.clear();
1017 index_t offset=0;
1018 for (size_t i=0; i<facesPerEdge.size(); i++) {
1019 if (facesPerEdge[i]>0) {
1020 m_faceOffset[i]=offset;
1021 offset+=facesPerEdge[i];
1022 m_faceTags.insert(m_faceTags.end(), facesPerEdge[i], faceTag[i]);
1023 }
1024 }
1025 setTagMap("left", LEFT);
1026 setTagMap("right", RIGHT);
1027 setTagMap("bottom", BOTTOM);
1028 setTagMap("top", TOP);
1029 updateTagsInUse(FaceElements);
1030 }
1031
1032 //private
1033 void Rectangle::createPattern()
1034 {
1035 const dim_t nDOF0 = (m_gNE0+1)/m_NX;
1036 const dim_t nDOF1 = (m_gNE1+1)/m_NY;
1037 const index_t left = (m_offset0==0 ? 0 : 1);
1038 const index_t bottom = (m_offset1==0 ? 0 : 1);
1039
1040 // populate node->DOF mapping with own degrees of freedom.
1041 // The rest is assigned in the loop further down
1042 m_dofMap.assign(getNumNodes(), 0);
1043 #pragma omp parallel for
1044 for (index_t i=bottom; i<m_N1; i++) {
1045 for (index_t j=left; j<m_N0; j++) {
1046 m_dofMap[i*m_N0+j]=(i-bottom)*nDOF0+j-left;
1047 }
1048 }
1049
1050 // build list of shared components and neighbours by looping through
1051 // all potential neighbouring ranks and checking if positions are
1052 // within bounds
1053 const dim_t numDOF=nDOF0*nDOF1;
1054 vector<IndexVector> colIndices(numDOF); // for the couple blocks
1055 RankVector neighbour;
1056 IndexVector offsetInShared(1,0);
1057 IndexVector sendShared, recvShared;
1058 int numShared=0;
1059 const int x=m_mpiInfo->rank%m_NX;
1060 const int y=m_mpiInfo->rank/m_NX;
1061 for (int i1=-1; i1<2; i1++) {
1062 for (int i0=-1; i0<2; i0++) {
1063 // skip this rank
1064 if (i0==0 && i1==0)
1065 continue;
1066 // location of neighbour rank
1067 const int nx=x+i0;
1068 const int ny=y+i1;
1069 if (nx>=0 && ny>=0 && nx<m_NX && ny<m_NY) {
1070 neighbour.push_back(ny*m_NX+nx);
1071 if (i0==0) {
1072 // sharing top or bottom edge
1073 const int firstDOF=(i1==-1 ? 0 : numDOF-nDOF0);
1074 const int firstNode=(i1==-1 ? left : m_N0*(m_N1-1)+left);
1075 offsetInShared.push_back(offsetInShared.back()+nDOF0);
1076 for (dim_t i=0; i<nDOF0; i++, numShared++) {
1077 sendShared.push_back(firstDOF+i);
1078 recvShared.push_back(numDOF+numShared);
1079 if (i>0)
1080 colIndices[firstDOF+i-1].push_back(numShared);
1081 colIndices[firstDOF+i].push_back(numShared);
1082 if (i<nDOF0-1)
1083 colIndices[firstDOF+i+1].push_back(numShared);
1084 m_dofMap[firstNode+i]=numDOF+numShared;
1085 }
1086 } else if (i1==0) {
1087 // sharing left or right edge
1088 const int firstDOF=(i0==-1 ? 0 : nDOF0-1);
1089 const int firstNode=(i0==-1 ? bottom*m_N0 : (bottom+1)*m_N0-1);
1090 offsetInShared.push_back(offsetInShared.back()+nDOF1);
1091 for (dim_t i=0; i<nDOF1; i++, numShared++) {
1092 sendShared.push_back(firstDOF+i*nDOF0);
1093 recvShared.push_back(numDOF+numShared);
1094 if (i>0)
1095 colIndices[firstDOF+(i-1)*nDOF0].push_back(numShared);
1096 colIndices[firstDOF+i*nDOF0].push_back(numShared);
1097 if (i<nDOF1-1)
1098 colIndices[firstDOF+(i+1)*nDOF0].push_back(numShared);
1099 m_dofMap[firstNode+i*m_N0]=numDOF+numShared;
1100 }
1101 } else {
1102 // sharing a node
1103 const int dof=(i0+1)/2*(nDOF0-1)+(i1+1)/2*(numDOF-nDOF0);
1104 const int node=(i0+1)/2*(m_N0-1)+(i1+1)/2*m_N0*(m_N1-1);
1105 offsetInShared.push_back(offsetInShared.back()+1);
1106 sendShared.push_back(dof);
1107 recvShared.push_back(numDOF+numShared);
1108 colIndices[dof].push_back(numShared);
1109 m_dofMap[node]=numDOF+numShared;
1110 ++numShared;
1111 }
1112 }
1113 }
1114 }
1115
1116 // create connector
1117 Paso_SharedComponents *snd_shcomp = Paso_SharedComponents_alloc(
1118 numDOF, neighbour.size(), &neighbour[0], &sendShared[0],
1119 &offsetInShared[0], 1, 0, m_mpiInfo);
1120 Paso_SharedComponents *rcv_shcomp = Paso_SharedComponents_alloc(
1121 numDOF, neighbour.size(), &neighbour[0], &recvShared[0],
1122 &offsetInShared[0], 1, 0, m_mpiInfo);
1123 m_connector = Paso_Connector_alloc(snd_shcomp, rcv_shcomp);
1124 Paso_SharedComponents_free(snd_shcomp);
1125 Paso_SharedComponents_free(rcv_shcomp);
1126
1127 // create main and couple blocks
1128 Paso_Pattern *mainPattern = createMainPattern();
1129 Paso_Pattern *colPattern, *rowPattern;
1130 createCouplePatterns(colIndices, numShared, &colPattern, &rowPattern);
1131
1132 // allocate paso distribution
1133 Paso_Distribution* distribution = Paso_Distribution_alloc(m_mpiInfo,
1134 const_cast<index_t*>(&m_nodeDistribution[0]), 1, 0);
1135
1136 // finally create the system matrix
1137 m_pattern = Paso_SystemMatrixPattern_alloc(MATRIX_FORMAT_DEFAULT,
1138 distribution, distribution, mainPattern, colPattern, rowPattern,
1139 m_connector, m_connector);
1140
1141 Paso_Distribution_free(distribution);
1142
1143 // useful debug output
1144 /*
1145 cout << "--- rcv_shcomp ---" << endl;
1146 cout << "numDOF=" << numDOF << ", numNeighbors=" << neighbour.size() << endl;
1147 for (size_t i=0; i<neighbour.size(); i++) {
1148 cout << "neighbor[" << i << "]=" << neighbour[i]
1149 << " offsetInShared[" << i+1 << "]=" << offsetInShared[i+1] << endl;
1150 }
1151 for (size_t i=0; i<recvShared.size(); i++) {
1152 cout << "shared[" << i << "]=" << recvShared[i] << endl;
1153 }
1154 cout << "--- snd_shcomp ---" << endl;
1155 for (size_t i=0; i<sendShared.size(); i++) {
1156 cout << "shared[" << i << "]=" << sendShared[i] << endl;
1157 }
1158 cout << "--- dofMap ---" << endl;
1159 for (size_t i=0; i<m_dofMap.size(); i++) {
1160 cout << "m_dofMap[" << i << "]=" << m_dofMap[i] << endl;
1161 }
1162 cout << "--- colIndices ---" << endl;
1163 for (size_t i=0; i<colIndices.size(); i++) {
1164 cout << "colIndices[" << i << "].size()=" << colIndices[i].size() << endl;
1165 }
1166 */
1167
1168 /*
1169 cout << "--- main_pattern ---" << endl;
1170 cout << "M=" << mainPattern->numOutput << ", N=" << mainPattern->numInput << endl;
1171 for (size_t i=0; i<mainPattern->numOutput+1; i++) {
1172 cout << "ptr[" << i << "]=" << mainPattern->ptr[i] << endl;
1173 }
1174 for (size_t i=0; i<mainPattern->ptr[mainPattern->numOutput]; i++) {
1175 cout << "index[" << i << "]=" << mainPattern->index[i] << endl;
1176 }
1177 */
1178
1179 /*
1180 cout << "--- colCouple_pattern ---" << endl;
1181 cout << "M=" << colPattern->numOutput << ", N=" << colPattern->numInput << endl;
1182 for (size_t i=0; i<colPattern->numOutput+1; i++) {
1183 cout << "ptr[" << i << "]=" << colPattern->ptr[i] << endl;
1184 }
1185 for (size_t i=0; i<colPattern->ptr[colPattern->numOutput]; i++) {
1186 cout << "index[" << i << "]=" << colPattern->index[i] << endl;
1187 }
1188 */
1189
1190 /*
1191 cout << "--- rowCouple_pattern ---" << endl;
1192 cout << "M=" << rowPattern->numOutput << ", N=" << rowPattern->numInput << endl;
1193 for (size_t i=0; i<rowPattern->numOutput+1; i++) {
1194 cout << "ptr[" << i << "]=" << rowPattern->ptr[i] << endl;
1195 }
1196 for (size_t i=0; i<rowPattern->ptr[rowPattern->numOutput]; i++) {
1197 cout << "index[" << i << "]=" << rowPattern->index[i] << endl;
1198 }
1199 */
1200
1201 Paso_Pattern_free(mainPattern);
1202 Paso_Pattern_free(colPattern);
1203 Paso_Pattern_free(rowPattern);
1204 }
1205
1206 //protected
1207 void Rectangle::interpolateNodesOnElements(escript::Data& out,
1208 escript::Data& in, bool reduced) const
1209 {
1210 const dim_t numComp = in.getDataPointSize();
1211 if (reduced) {
1212 /*** GENERATOR SNIP_INTERPOLATE_REDUCED_ELEMENTS TOP */
1213 const double c0 = .25;
1214 #pragma omp parallel for
1215 for (index_t k1=0; k1 < m_NE1; ++k1) {
1216 for (index_t k0=0; k0 < m_NE0; ++k0) {
1217 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,k1+1, m_N0));
1218 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,k1, m_N0));
1219 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,k1, m_N0));
1220 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,k1+1, m_N0));
1221 double* o = out.getSampleDataRW(INDEX2(k0,k1,m_NE0));
1222 for (index_t i=0; i < numComp; ++i) {
1223 o[INDEX2(i,numComp,0)] = c0*(f_00[i] + f_01[i] + f_10[i] + f_11[i]);
1224 } /* end of component loop i */
1225 } /* end of k0 loop */
1226 } /* end of k1 loop */
1227 /* GENERATOR SNIP_INTERPOLATE_REDUCED_ELEMENTS BOTTOM */
1228 } else {
1229 /*** GENERATOR SNIP_INTERPOLATE_ELEMENTS TOP */
1230 const double c0 = .16666666666666666667;
1231 const double c1 = .044658198738520451079;
1232 const double c2 = .62200846792814621559;
1233 #pragma omp parallel for
1234 for (index_t k1=0; k1 < m_NE1; ++k1) {
1235 for (index_t k0=0; k0 < m_NE0; ++k0) {
1236 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,k1, m_N0));
1237 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,k1+1, m_N0));
1238 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,k1+1, m_N0));
1239 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,k1, m_N0));
1240 double* o = out.getSampleDataRW(INDEX2(k0,k1,m_NE0));
1241 for (index_t i=0; i < numComp; ++i) {
1242 o[INDEX2(i,numComp,0)] = f_00[i]*c2 + f_11[i]*c1 + c0*(f_01[i] + f_10[i]);
1243 o[INDEX2(i,numComp,1)] = f_01[i]*c1 + f_10[i]*c2 + c0*(f_00[i] + f_11[i]);
1244 o[INDEX2(i,numComp,2)] = f_01[i]*c2 + f_10[i]*c1 + c0*(f_00[i] + f_11[i]);
1245 o[INDEX2(i,numComp,3)] = f_00[i]*c1 + f_11[i]*c2 + c0*(f_01[i] + f_10[i]);
1246 } /* end of component loop i */
1247 } /* end of k0 loop */
1248 } /* end of k1 loop */
1249 /* GENERATOR SNIP_INTERPOLATE_ELEMENTS BOTTOM */
1250 }
1251 }
1252
1253 //protected
1254 void Rectangle::interpolateNodesOnFaces(escript::Data& out, escript::Data& in,
1255 bool reduced) const
1256 {
1257 const dim_t numComp = in.getDataPointSize();
1258 if (reduced) {
1259 const double c0 = .5;
1260 #pragma omp parallel
1261 {
1262 /*** GENERATOR SNIP_INTERPOLATE_REDUCED_FACES TOP */
1263 if (m_faceOffset[0] > -1) {
1264 #pragma omp for nowait
1265 for (index_t k1=0; k1 < m_NE1; ++k1) {
1266 const register double* f_00 = in.getSampleDataRO(INDEX2(0,k1, m_N0));
1267 const register double* f_01 = in.getSampleDataRO(INDEX2(0,k1+1, m_N0));
1268 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
1269 for (index_t i=0; i < numComp; ++i) {
1270 o[INDEX2(i,numComp,0)] = c0*(f_00[i] + f_01[i]);
1271 } /* end of component loop i */
1272 } /* end of k1 loop */
1273 } /* end of face 0 */
1274 if (m_faceOffset[1] > -1) {
1275 #pragma omp for nowait
1276 for (index_t k1=0; k1 < m_NE1; ++k1) {
1277 const register double* f_10 = in.getSampleDataRO(INDEX2(m_N0-1,k1, m_N0));
1278 const register double* f_11 = in.getSampleDataRO(INDEX2(m_N0-1,k1+1, m_N0));
1279 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
1280 for (index_t i=0; i < numComp; ++i) {
1281 o[INDEX2(i,numComp,0)] = c0*(f_10[i] + f_11[i]);
1282 } /* end of component loop i */
1283 } /* end of k1 loop */
1284 } /* end of face 1 */
1285 if (m_faceOffset[2] > -1) {
1286 #pragma omp for nowait
1287 for (index_t k0=0; k0 < m_NE0; ++k0) {
1288 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,0, m_N0));
1289 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,0, m_N0));
1290 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
1291 for (index_t i=0; i < numComp; ++i) {
1292 o[INDEX2(i,numComp,0)] = c0*(f_00[i] + f_10[i]);
1293 } /* end of component loop i */
1294 } /* end of k0 loop */
1295 } /* end of face 2 */
1296 if (m_faceOffset[3] > -1) {
1297 #pragma omp for nowait
1298 for (index_t k0=0; k0 < m_NE0; ++k0) {
1299 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,m_N1-1, m_N0));
1300 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,m_N1-1, m_N0));
1301 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
1302 for (index_t i=0; i < numComp; ++i) {
1303 o[INDEX2(i,numComp,0)] = c0*(f_01[i] + f_11[i]);
1304 } /* end of component loop i */
1305 } /* end of k0 loop */
1306 } /* end of face 3 */
1307 /* GENERATOR SNIP_INTERPOLATE_REDUCED_FACES BOTTOM */
1308 } // end of parallel section
1309 } else {
1310 const double c0 = 0.21132486540518711775;
1311 const double c1 = 0.78867513459481288225;
1312 #pragma omp parallel
1313 {
1314 /*** GENERATOR SNIP_INTERPOLATE_FACES TOP */
1315 if (m_faceOffset[0] > -1) {
1316 #pragma omp for nowait
1317 for (index_t k1=0; k1 < m_NE1; ++k1) {
1318 const register double* f_01 = in.getSampleDataRO(INDEX2(0,k1+1, m_N0));
1319 const register double* f_00 = in.getSampleDataRO(INDEX2(0,k1, m_N0));
1320 double* o = out.getSampleDataRW(m_faceOffset[0]+k1);
1321 for (index_t i=0; i < numComp; ++i) {
1322 o[INDEX2(i,numComp,0)] = f_00[i]*c1 + f_01[i]*c0;
1323 o[INDEX2(i,numComp,1)] = f_00[i]*c0 + f_01[i]*c1;
1324 } /* end of component loop i */
1325 } /* end of k1 loop */
1326 } /* end of face 0 */
1327 if (m_faceOffset[1] > -1) {
1328 #pragma omp for nowait
1329 for (index_t k1=0; k1 < m_NE1; ++k1) {
1330 const register double* f_10 = in.getSampleDataRO(INDEX2(m_N0-1,k1, m_N0));
1331 const register double* f_11 = in.getSampleDataRO(INDEX2(m_N0-1,k1+1, m_N0));
1332 double* o = out.getSampleDataRW(m_faceOffset[1]+k1);
1333 for (index_t i=0; i < numComp; ++i) {
1334 o[INDEX2(i,numComp,0)] = f_10[i]*c1 + f_11[i]*c0;
1335 o[INDEX2(i,numComp,1)] = f_10[i]*c0 + f_11[i]*c1;
1336 } /* end of component loop i */
1337 } /* end of k1 loop */
1338 } /* end of face 1 */
1339 if (m_faceOffset[2] > -1) {
1340 #pragma omp for nowait
1341 for (index_t k0=0; k0 < m_NE0; ++k0) {
1342 const register double* f_10 = in.getSampleDataRO(INDEX2(k0+1,0, m_N0));
1343 const register double* f_00 = in.getSampleDataRO(INDEX2(k0,0, m_N0));
1344 double* o = out.getSampleDataRW(m_faceOffset[2]+k0);
1345 for (index_t i=0; i < numComp; ++i) {
1346 o[INDEX2(i,numComp,0)] = f_00[i]*c1 + f_10[i]*c0;
1347 o[INDEX2(i,numComp,1)] = f_00[i]*c0 + f_10[i]*c1;
1348 } /* end of component loop i */
1349 } /* end of k0 loop */
1350 } /* end of face 2 */
1351 if (m_faceOffset[3] > -1) {
1352 #pragma omp for nowait
1353 for (index_t k0=0; k0 < m_NE0; ++k0) {
1354 const register double* f_11 = in.getSampleDataRO(INDEX2(k0+1,m_N1-1, m_N0));
1355 const register double* f_01 = in.getSampleDataRO(INDEX2(k0,m_N1-1, m_N0));
1356 double* o = out.getSampleDataRW(m_faceOffset[3]+k0);
1357 for (index_t i=0; i < numComp; ++i) {
1358 o[INDEX2(i,numComp,0)] = f_01[i]*c1 + f_11[i]*c0;
1359 o[INDEX2(i,numComp,1)] = f_01[i]*c0 + f_11[i]*c1;
1360 } /* end of component loop i */
1361 } /* end of k0 loop */
1362 } /* end of face 3 */
1363 /* GENERATOR SNIP_INTERPOLATE_FACES BOTTOM */
1364 } // end of parallel section
1365 }
1366 }
1367
1368 //protected
1369 void Rectangle::assemblePDESingle(Paso_SystemMatrix* mat,
1370 escript::Data& rhs, const escript::Data& A, const escript::Data& B,
1371 const escript::Data& C, const escript::Data& D,
1372 const escript::Data& X, const escript::Data& Y,
1373 const escript::Data& d, const escript::Data& y) const
1374 {
1375 const double h0 = m_l0/m_gNE0;
1376 const double h1 = m_l1/m_gNE1;
1377 /*** GENERATOR SNIP_PDE_SINGLE_PRE TOP */
1378 const double w0 = -0.1555021169820365539*h1/h0;
1379 const double w1 = 0.041666666666666666667;
1380 const double w10 = -0.041666666666666666667*h0/h1;
1381 const double w11 = 0.1555021169820365539*h1/h0;
1382 const double w12 = 0.1555021169820365539*h0/h1;
1383 const double w13 = 0.01116454968463011277*h0/h1;
1384 const double w14 = 0.01116454968463011277*h1/h0;
1385 const double w15 = 0.041666666666666666667*h1/h0;
1386 const double w16 = -0.01116454968463011277*h0/h1;
1387 const double w17 = -0.1555021169820365539*h0/h1;
1388 const double w18 = -0.33333333333333333333*h1/h0;
1389 const double w19 = 0.25000000000000000000;
1390 const double w2 = -0.15550211698203655390;
1391 const double w20 = -0.25000000000000000000;
1392 const double w21 = 0.16666666666666666667*h0/h1;
1393 const double w22 = -0.16666666666666666667*h1/h0;
1394 const double w23 = -0.16666666666666666667*h0/h1;
1395 const double w24 = 0.33333333333333333333*h1/h0;
1396 const double w25 = 0.33333333333333333333*h0/h1;
1397 const double w26 = 0.16666666666666666667*h1/h0;
1398 const double w27 = -0.33333333333333333333*h0/h1;
1399 const double w28 = -0.032861463941450536761*h1;
1400 const double w29 = -0.032861463941450536761*h0;
1401 const double w3 = 0.041666666666666666667*h0/h1;
1402 const double w30 = -0.12264065304058601714*h1;
1403 const double w31 = -0.0023593469594139828636*h1;
1404 const double w32 = -0.008805202725216129906*h0;
1405 const double w33 = -0.008805202725216129906*h1;
1406 const double w34 = 0.032861463941450536761*h1;
1407 const double w35 = 0.008805202725216129906*h1;
1408 const double w36 = 0.008805202725216129906*h0;
1409 const double w37 = 0.0023593469594139828636*h1;
1410 const double w38 = 0.12264065304058601714*h1;
1411 const double w39 = 0.032861463941450536761*h0;
1412 const double w4 = 0.15550211698203655390;
1413 const double w40 = -0.12264065304058601714*h0;
1414 const double w41 = -0.0023593469594139828636*h0;
1415 const double w42 = 0.0023593469594139828636*h0;
1416 const double w43 = 0.12264065304058601714*h0;
1417 const double w44 = -0.16666666666666666667*h1;
1418 const double w45 = -0.083333333333333333333*h0;
1419 const double w46 = 0.083333333333333333333*h1;
1420 const double w47 = 0.16666666666666666667*h1;
1421 const double w48 = 0.083333333333333333333*h0;
1422 const double w49 = -0.16666666666666666667*h0;
1423 const double w5 = -0.041666666666666666667;
1424 const double w50 = 0.16666666666666666667*h0;
1425 const double w51 = -0.083333333333333333333*h1;
1426 const double w52 = 0.025917019497006092316*h0*h1;
1427 const double w53 = 0.0018607582807716854616*h0*h1;
1428 const double w54 = 0.0069444444444444444444*h0*h1;
1429 const double w55 = 0.09672363354357992482*h0*h1;
1430 const double w56 = 0.00049858867864229740201*h0*h1;
1431 const double w57 = 0.055555555555555555556*h0*h1;
1432 const double w58 = 0.027777777777777777778*h0*h1;
1433 const double w59 = 0.11111111111111111111*h0*h1;
1434 const double w6 = -0.01116454968463011277*h1/h0;
1435 const double w60 = -0.19716878364870322056*h1;
1436 const double w61 = -0.19716878364870322056*h0;
1437 const double w62 = -0.052831216351296779436*h0;
1438 const double w63 = -0.052831216351296779436*h1;
1439 const double w64 = 0.19716878364870322056*h1;
1440 const double w65 = 0.052831216351296779436*h1;
1441 const double w66 = 0.19716878364870322056*h0;
1442 const double w67 = 0.052831216351296779436*h0;
1443 const double w68 = -0.5*h1;
1444 const double w69 = -0.5*h0;
1445 const double w7 = 0.011164549684630112770;
1446 const double w70 = 0.5*h1;
1447 const double w71 = 0.5*h0;
1448 const double w72 = 0.1555021169820365539*h0*h1;
1449 const double w73 = 0.041666666666666666667*h0*h1;
1450 const double w74 = 0.01116454968463011277*h0*h1;
1451 const double w75 = 0.25*h0*h1;
1452 const double w8 = -0.011164549684630112770;
1453 const double w9 = -0.041666666666666666667*h1/h0;
1454 /* GENERATOR SNIP_PDE_SINGLE_PRE BOTTOM */
1455
1456 rhs.requireWrite();
1457 #pragma omp parallel
1458 {
1459 for (index_t k1_0=0; k1_0<2; k1_0++) { // colouring
1460 #pragma omp for
1461 for (index_t k1=k1_0; k1<m_NE1; k1+=2) {
1462 for (index_t k0=0; k0<m_NE0; ++k0) {
1463 bool add_EM_S=false;
1464 bool add_EM_F=false;
1465 vector<double> EM_S(4*4, 0);
1466 vector<double> EM_F(4, 0);
1467 const index_t e = k0 + m_NE0*k1;
1468 /*** GENERATOR SNIP_PDE_SINGLE TOP */
1469 ///////////////
1470 // process A //
1471 ///////////////
1472 if (!A.isEmpty()) {
1473 add_EM_S=true;
1474 const double* A_p=const_cast<escript::Data*>(&A)->getSampleDataRO(e);
1475 if (A.actsExpanded()) {
1476 const register double A_00_0 = A_p[INDEX3(0,0,0,2,2)];
1477 const register double A_01_0 = A_p[INDEX3(0,1,0,2,2)];
1478 const register double A_10_0 = A_p[INDEX3(1,0,0,2,2)];
1479 const register double A_11_0 = A_p[INDEX3(1,1,0,2,2)];
1480 const register double A_00_1 = A_p[INDEX3(0,0,1,2,2)];
1481 const register double A_01_1 = A_p[INDEX3(0,1,1,2,2)];
1482 const register double A_10_1 = A_p[INDEX3(1,0,1,2,2)];
1483 const register double A_11_1 = A_p[INDEX3(1,1,1,2,2)];
1484 const register double A_00_2 = A_p[INDEX3(0,0,2,2,2)];
1485 const register double A_01_2 = A_p[INDEX3(0,1,2,2,2)];
1486 const register double A_10_2 = A_p[INDEX3(1,0,2,2,2)];
1487 const register double A_11_2 = A_p[INDEX3(1,1,2,2,2)];
1488 const register double A_00_3 = A_p[INDEX3(0,0,3,2,2)];
1489 const register double A_01_3 = A_p[INDEX3(0,1,3,2,2)];
1490 const register double A_10_3 = A_p[INDEX3(1,0,3,2,2)];
1491 const register double A_11_3 = A_p[INDEX3(1,1,3,2,2)];
1492 const register double tmp4_0 = A_10_1 + A_10_2;
1493 const register double tmp12_0 = A_11_0 + A_11_2;
1494 const register double tmp2_0 = A_11_0 + A_11_1 + A_11_2 + A_11_3;
1495 const register double tmp10_0 = A_01_3 + A_10_3;
1496 const register double tmp14_0 = A_01_0 + A_01_3 + A_10_0 + A_10_3;
1497 const register double tmp0_0 = A_01_0 + A_01_3;
1498 const register double tmp13_0 = A_01_2 + A_10_1;
1499 const register double tmp3_0 = A_00_2 + A_00_3;
1500 const register double tmp11_0 = A_11_1 + A_11_3;
1501 const register double tmp18_0 = A_01_1 + A_10_1;
1502 const register double tmp1_0 = A_00_0 + A_00_1;
1503 const register double tmp15_0 = A_01_1 + A_10_2;
1504 const register double tmp5_0 = A_00_0 + A_00_1 + A_00_2 + A_00_3;
1505 const register double tmp16_0 = A_10_0 + A_10_3;
1506 const register double tmp6_0 = A_01_3 + A_10_0;
1507 const register double tmp17_0 = A_01_1 + A_01_2;
1508 const register double tmp9_0 = A_01_0 + A_10_0;
1509 const register double tmp7_0 = A_01_0 + A_10_3;
1510 const register double tmp8_0 = A_01_1 + A_01_2 + A_10_1 + A_10_2;
1511 const register double tmp19_0 = A_01_2 + A_10_2;
1512 const register double tmp14_1 = A_10_0*w8;
1513 const register double tmp23_1 = tmp3_0*w14;
1514 const register double tmp35_1 = A_01_0*w8;
1515 const register double tmp54_1 = tmp13_0*w8;
1516 const register double tmp20_1 = tmp9_0*w4;
1517 const register double tmp25_1 = tmp12_0*w12;
1518 const register double tmp2_1 = A_01_1*w4;
1519 const register double tmp44_1 = tmp7_0*w7;
1520 const register double tmp26_1 = tmp10_0*w4;
1521 const register double tmp52_1 = tmp18_0*w8;
1522 const register double tmp48_1 = A_10_1*w7;
1523 const register double tmp46_1 = A_01_3*w8;
1524 const register double tmp50_1 = A_01_0*w2;
1525 const register double tmp8_1 = tmp4_0*w5;
1526 const register double tmp56_1 = tmp19_0*w8;
1527 const register double tmp9_1 = tmp2_0*w10;
1528 const register double tmp19_1 = A_10_3*w2;
1529 const register double tmp47_1 = A_10_2*w4;
1530 const register double tmp16_1 = tmp3_0*w0;
1531 const register double tmp18_1 = tmp1_0*w6;
1532 const register double tmp31_1 = tmp11_0*w12;
1533 const register double tmp55_1 = tmp15_0*w2;
1534 const register double tmp39_1 = A_10_2*w7;
1535 const register double tmp11_1 = tmp6_0*w7;
1536 const register double tmp40_1 = tmp11_0*w17;
1537 const register double tmp34_1 = tmp15_0*w8;
1538 const register double tmp33_1 = tmp14_0*w5;
1539 const register double tmp24_1 = tmp11_0*w13;
1540 const register double tmp3_1 = tmp1_0*w0;
1541 const register double tmp5_1 = tmp2_0*w3;
1542 const register double tmp43_1 = tmp17_0*w5;
1543 const register double tmp15_1 = A_01_2*w4;
1544 const register double tmp53_1 = tmp19_0*w2;
1545 const register double tmp27_1 = tmp3_0*w11;
1546 const register double tmp32_1 = tmp13_0*w2;
1547 const register double tmp10_1 = tmp5_0*w9;
1548 const register double tmp37_1 = A_10_1*w4;
1549 const register double tmp38_1 = tmp5_0*w15;
1550 const register double tmp17_1 = A_01_1*w7;
1551 const register double tmp12_1 = tmp7_0*w4;
1552 const register double tmp22_1 = tmp10_0*w7;
1553 const register double tmp57_1 = tmp18_0*w2;
1554 const register double tmp28_1 = tmp9_0*w7;
1555 const register double tmp29_1 = tmp1_0*w14;
1556 const register double tmp51_1 = tmp11_0*w16;
1557 const register double tmp42_1 = tmp12_0*w16;
1558 const register double tmp49_1 = tmp12_0*w17;
1559 const register double tmp21_1 = tmp1_0*w11;
1560 const register double tmp1_1 = tmp0_0*w1;
1561 const register double tmp45_1 = tmp6_0*w4;
1562 const register double tmp7_1 = A_10_0*w2;
1563 const register double tmp6_1 = tmp3_0*w6;
1564 const register double tmp13_1 = tmp8_0*w1;
1565 const register double tmp36_1 = tmp16_0*w1;
1566 const register double tmp41_1 = A_01_3*w2;
1567 const register double tmp30_1 = tmp12_0*w13;
1568 const register double tmp4_1 = A_01_2*w7;
1569 const register double tmp0_1 = A_10_3*w8;
1570 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1 + tmp4_1 + tmp5_1 + tmp6_1 + tmp7_1 + tmp8_1;
1571 EM_S[INDEX2(1,2,4)]+=tmp10_1 + tmp11_1 + tmp12_1 + tmp13_1 + tmp9_1;
1572 EM_S[INDEX2(3,2,4)]+=tmp14_1 + tmp15_1 + tmp16_1 + tmp17_1 + tmp18_1 + tmp19_1 + tmp1_1 + tmp5_1 + tmp8_1;
1573 EM_S[INDEX2(0,0,4)]+=tmp13_1 + tmp20_1 + tmp21_1 + tmp22_1 + tmp23_1 + tmp24_1 + tmp25_1;
1574 EM_S[INDEX2(3,3,4)]+=tmp13_1 + tmp26_1 + tmp27_1 + tmp28_1 + tmp29_1 + tmp30_1 + tmp31_1;
1575 EM_S[INDEX2(3,0,4)]+=tmp10_1 + tmp32_1 + tmp33_1 + tmp34_1 + tmp9_1;
1576 EM_S[INDEX2(3,1,4)]+=tmp35_1 + tmp36_1 + tmp37_1 + tmp38_1 + tmp39_1 + tmp40_1 + tmp41_1 + tmp42_1 + tmp43_1;
1577 EM_S[INDEX2(2,1,4)]+=tmp10_1 + tmp13_1 + tmp44_1 + tmp45_1 + tmp9_1;
1578 EM_S[INDEX2(0,2,4)]+=tmp36_1 + tmp38_1 + tmp43_1 + tmp46_1 + tmp47_1 + tmp48_1 + tmp49_1 + tmp50_1 + tmp51_1;
1579 EM_S[INDEX2(2,0,4)]+=tmp0_1 + tmp15_1 + tmp17_1 + tmp1_1 + tmp38_1 + tmp49_1 + tmp51_1 + tmp7_1 + tmp8_1;
1580 EM_S[INDEX2(1,3,4)]+=tmp14_1 + tmp19_1 + tmp1_1 + tmp2_1 + tmp38_1 + tmp40_1 + tmp42_1 + tmp4_1 + tmp8_1;
1581 EM_S[INDEX2(2,3,4)]+=tmp16_1 + tmp18_1 + tmp35_1 + tmp36_1 + tmp41_1 + tmp43_1 + tmp47_1 + tmp48_1 + tmp5_1;
1582 EM_S[INDEX2(2,2,4)]+=tmp24_1 + tmp25_1 + tmp27_1 + tmp29_1 + tmp33_1 + tmp52_1 + tmp53_1;
1583 EM_S[INDEX2(1,0,4)]+=tmp36_1 + tmp37_1 + tmp39_1 + tmp3_1 + tmp43_1 + tmp46_1 + tmp50_1 + tmp5_1 + tmp6_1;
1584 EM_S[INDEX2(0,3,4)]+=tmp10_1 + tmp33_1 + tmp54_1 + tmp55_1 + tmp9_1;
1585 EM_S[INDEX2(1,1,4)]+=tmp21_1 + tmp23_1 + tmp30_1 + tmp31_1 + tmp33_1 + tmp56_1 + tmp57_1;
1586 } else { /* constant data */
1587 const register double A_00 = A_p[INDEX2(0,0,2)];
1588 const register double A_01 = A_p[INDEX2(0,1,2)];
1589 const register double A_10 = A_p[INDEX2(1,0,2)];
1590 const register double A_11 = A_p[INDEX2(1,1,2)];
1591 const register double tmp0_0 = A_01 + A_10;
1592 const register double tmp0_1 = A_00*w18;
1593 const register double tmp10_1 = A_01*w20;
1594 const register double tmp12_1 = A_00*w26;
1595 const register double tmp4_1 = A_00*w22;
1596 const register double tmp8_1 = A_00*w24;
1597 const register double tmp13_1 = A_10*w19;
1598 const register double tmp9_1 = tmp0_0*w20;
1599 const register double tmp3_1 = A_11*w21;
1600 const register double tmp11_1 = A_11*w27;
1601 const register double tmp1_1 = A_01*w19;
1602 const register double tmp6_1 = A_11*w23;
1603 const register double tmp7_1 = A_11*w25;
1604 const register double tmp2_1 = A_10*w20;
1605 const register double tmp5_1 = tmp0_0*w19;
1606 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1;
1607 EM_S[INDEX2(1,2,4)]+=tmp4_1 + tmp5_1 + tmp6_1;
1608 EM_S[INDEX2(3,2,4)]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1;
1609 EM_S[INDEX2(0,0,4)]+=tmp5_1 + tmp7_1 + tmp8_1;
1610 EM_S[INDEX2(3,3,4)]+=tmp5_1 + tmp7_1 + tmp8_1;
1611 EM_S[INDEX2(3,0,4)]+=tmp4_1 + tmp6_1 + tmp9_1;
1612 EM_S[INDEX2(3,1,4)]+=tmp10_1 + tmp11_1 + tmp12_1 + tmp13_1;
1613 EM_S[INDEX2(2,1,4)]+=tmp4_1 + tmp5_1 + tmp6_1;
1614 EM_S[INDEX2(0,2,4)]+=tmp10_1 + tmp11_1 + tmp12_1 + tmp13_1;
1615 EM_S[INDEX2(2,0,4)]+=tmp11_1 + tmp12_1 + tmp1_1 + tmp2_1;
1616 EM_S[INDEX2(1,3,4)]+=tmp11_1 + tmp12_1 + tmp1_1 + tmp2_1;
1617 EM_S[INDEX2(2,3,4)]+=tmp0_1 + tmp10_1 + tmp13_1 + tmp3_1;
1618 EM_S[INDEX2(2,2,4)]+=tmp7_1 + tmp8_1 + tmp9_1;
1619 EM_S[INDEX2(1,0,4)]+=tmp0_1 + tmp10_1 + tmp13_1 + tmp3_1;
1620 EM_S[INDEX2(0,3,4)]+=tmp4_1 + tmp6_1 + tmp9_1;
1621 EM_S[INDEX2(1,1,4)]+=tmp7_1 + tmp8_1 + tmp9_1;
1622 }
1623 }
1624 ///////////////
1625 // process B //
1626 ///////////////
1627 if (!B.isEmpty()) {
1628 add_EM_S=true;
1629 const double* B_p=const_cast<escript::Data*>(&B)->getSampleDataRO(e);
1630 if (B.actsExpanded()) {
1631 const register double B_0_0 = B_p[INDEX2(0,0,2)];
1632 const register double B_1_0 = B_p[INDEX2(1,0,2)];
1633 const register double B_0_1 = B_p[INDEX2(0,1,2)];
1634 const register double B_1_1 = B_p[INDEX2(1,1,2)];
1635 const register double B_0_2 = B_p[INDEX2(0,2,2)];
1636 const register double B_1_2 = B_p[INDEX2(1,2,2)];
1637 const register double B_0_3 = B_p[INDEX2(0,3,2)];
1638 const register double B_1_3 = B_p[INDEX2(1,3,2)];
1639 const register double tmp3_0 = B_0_0 + B_0_2;
1640 const register double tmp1_0 = B_1_2 + B_1_3;
1641 const register double tmp2_0 = B_0_1 + B_0_3;
1642 const register double tmp0_0 = B_1_0 + B_1_1;
1643 const register double tmp63_1 = B_1_1*w42;
1644 const register double tmp79_1 = B_1_1*w40;
1645 const register double tmp37_1 = tmp3_0*w35;
1646 const register double tmp8_1 = tmp0_0*w32;
1647 const register double tmp71_1 = B_0_1*w34;
1648 const register double tmp19_1 = B_0_3*w31;
1649 const register double tmp15_1 = B_0_3*w34;
1650 const register double tmp9_1 = tmp3_0*w34;
1651 const register double tmp35_1 = B_1_0*w36;
1652 const register double tmp66_1 = B_0_3*w28;
1653 const register double tmp28_1 = B_1_0*w42;
1654 const register double tmp22_1 = B_1_0*w40;
1655 const register double tmp16_1 = B_1_2*w29;
1656 const register double tmp6_1 = tmp2_0*w35;
1657 const register double tmp55_1 = B_1_3*w40;
1658 const register double tmp50_1 = B_1_3*w42;
1659 const register double tmp7_1 = tmp1_0*w29;
1660 const register double tmp1_1 = tmp1_0*w32;
1661 const register double tmp57_1 = B_0_3*w30;
1662 const register double tmp18_1 = B_1_1*w32;
1663 const register double tmp53_1 = B_1_0*w41;
1664 const register double tmp61_1 = B_1_3*w36;
1665 const register double tmp27_1 = B_0_3*w38;
1666 const register double tmp64_1 = B_0_2*w30;
1667 const register double tmp76_1 = B_0_1*w38;
1668 const register double tmp39_1 = tmp2_0*w34;
1669 const register double tmp62_1 = B_0_1*w31;
1670 const register double tmp56_1 = B_0_0*w31;
1671 const register double tmp49_1 = B_1_1*w36;
1672 const register double tmp2_1 = B_0_2*w31;
1673 const register double tmp23_1 = B_0_2*w33;
1674 const register double tmp38_1 = B_1_1*w43;
1675 const register double tmp74_1 = B_1_2*w41;
1676 const register double tmp43_1 = B_1_1*w41;
1677 const register double tmp58_1 = B_0_2*w28;
1678 const register double tmp67_1 = B_0_0*w33;
1679 const register double tmp33_1 = tmp0_0*w39;
1680 const register double tmp4_1 = B_0_0*w28;
1681 const register double tmp20_1 = B_0_0*w30;
1682 const register double tmp13_1 = B_0_2*w38;
1683 const register double tmp65_1 = B_1_2*w43;
1684 const register double tmp0_1 = tmp0_0*w29;
1685 const register double tmp41_1 = tmp3_0*w33;
1686 const register double tmp73_1 = B_0_2*w37;
1687 const register double tmp69_1 = B_0_0*w38;
1688 const register double tmp48_1 = B_1_2*w39;
1689 const register double tmp59_1 = B_0_1*w33;
1690 const register double tmp17_1 = B_1_3*w41;
1691 const register double tmp5_1 = B_0_3*w33;
1692 const register double tmp3_1 = B_0_1*w30;
1693 const register double tmp21_1 = B_0_1*w28;
1694 const register double tmp42_1 = B_1_0*w29;
1695 const register double tmp54_1 = B_1_2*w32;
1696 const register double tmp60_1 = B_1_0*w39;
1697 const register double tmp32_1 = tmp1_0*w36;
1698 const register double tmp10_1 = B_0_1*w37;
1699 const register double tmp14_1 = B_0_0*w35;
1700 const register double tmp29_1 = B_0_1*w35;
1701 const register double tmp26_1 = B_1_2*w36;
1702 const register double tmp30_1 = B_1_3*w43;
1703 const register double tmp70_1 = B_0_2*w35;
1704 const register double tmp34_1 = B_1_3*w39;
1705 const register double tmp51_1 = B_1_0*w43;
1706 const register double tmp31_1 = B_0_2*w34;
1707 const register double tmp45_1 = tmp3_0*w28;
1708 const register double tmp11_1 = tmp1_0*w39;
1709 const register double tmp52_1 = B_1_1*w29;
1710 const register double tmp44_1 = B_1_3*w32;
1711 const register double tmp25_1 = B_1_1*w39;
1712 const register double tmp47_1 = tmp2_0*w33;
1713 const register double tmp72_1 = B_1_3*w29;
1714 const register double tmp40_1 = tmp2_0*w28;
1715 const register double tmp46_1 = B_1_2*w40;
1716 const register double tmp36_1 = B_1_2*w42;
1717 const register double tmp24_1 = B_0_0*w37;
1718 const register double tmp77_1 = B_0_3*w35;
1719 const register double tmp68_1 = B_0_3*w37;
1720 const register double tmp78_1 = B_0_0*w34;
1721 const register double tmp12_1 = tmp0_0*w36;
1722 const register double tmp75_1 = B_1_0*w32;
1723 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1 + tmp4_1 + tmp5_1;
1724 EM_S[INDEX2(1,2,4)]+=tmp6_1 + tmp7_1 + tmp8_1 + tmp9_1;
1725 EM_S[INDEX2(3,2,4)]+=tmp10_1 + tmp11_1 + tmp12_1 + tmp13_1 + tmp14_1 + tmp15_1;
1726 EM_S[INDEX2(0,0,4)]+=tmp16_1 + tmp17_1 + tmp18_1 + tmp19_1 + tmp20_1 + tmp21_1 + tmp22_1 + tmp23_1;
1727 EM_S[INDEX2(3,3,4)]+=tmp24_1 + tmp25_1 + tmp26_1 + tmp27_1 + tmp28_1 + tmp29_1 + tmp30_1 + tmp31_1;
1728 EM_S[INDEX2(3,0,4)]+=tmp32_1 + tmp33_1 + tmp6_1 + tmp9_1;
1729 EM_S[INDEX2(3,1,4)]+=tmp34_1 + tmp35_1 + tmp36_1 + tmp37_1 + tmp38_1 + tmp39_1;
1730 EM_S[INDEX2(2,1,4)]+=tmp32_1 + tmp33_1 + tmp40_1 + tmp41_1;
1731 EM_S[INDEX2(0,2,4)]+=tmp42_1 + tmp43_1 + tmp44_1 + tmp45_1 + tmp46_1 + tmp47_1;
1732 EM_S[INDEX2(2,0,4)]+=tmp45_1 + tmp47_1 + tmp48_1 + tmp49_1 + tmp50_1 + tmp51_1;
1733 EM_S[INDEX2(1,3,4)]+=tmp37_1 + tmp39_1 + tmp52_1 + tmp53_1 + tmp54_1 + tmp55_1;
1734 EM_S[INDEX2(2,3,4)]+=tmp11_1 + tmp12_1 + tmp56_1 + tmp57_1 + tmp58_1 + tmp59_1;
1735 EM_S[INDEX2(2,2,4)]+=tmp60_1 + tmp61_1 + tmp62_1 + tmp63_1 + tmp64_1 + tmp65_1 + tmp66_1 + tmp67_1;
1736 EM_S[INDEX2(1,0,4)]+=tmp0_1 + tmp1_1 + tmp68_1 + tmp69_1 + tmp70_1 + tmp71_1;
1737 EM_S[INDEX2(0,3,4)]+=tmp40_1 + tmp41_1 + tmp7_1 + tmp8_1;
1738 EM_S[INDEX2(1,1,4)]+=tmp72_1 + tmp73_1 + tmp74_1 + tmp75_1 + tmp76_1 + tmp77_1 + tmp78_1 + tmp79_1;
1739 } else { /* constant data */
1740 const register double B_0 = B_p[0];
1741 const register double B_1 = B_p[1];
1742 const register double tmp6_1 = B_1*w50;
1743 const register double tmp1_1 = B_1*w45;
1744 const register double tmp5_1 = B_1*w49;
1745 const register double tmp4_1 = B_1*w48;
1746 const register double tmp0_1 = B_0*w44;
1747 const register double tmp2_1 = B_0*w46;
1748 const register double tmp7_1 = B_0*w51;
1749 const register double tmp3_1 = B_0*w47;
1750 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1;
1751 EM_S[INDEX2(1,2,4)]+=tmp1_1 + tmp2_1;
1752 EM_S[INDEX2(3,2,4)]+=tmp3_1 + tmp4_1;
1753 EM_S[INDEX2(0,0,4)]+=tmp0_1 + tmp5_1;
1754 EM_S[INDEX2(3,3,4)]+=tmp3_1 + tmp6_1;
1755 EM_S[INDEX2(3,0,4)]+=tmp2_1 + tmp4_1;
1756 EM_S[INDEX2(3,1,4)]+=tmp2_1 + tmp6_1;
1757 EM_S[INDEX2(2,1,4)]+=tmp4_1 + tmp7_1;
1758 EM_S[INDEX2(0,2,4)]+=tmp5_1 + tmp7_1;
1759 EM_S[INDEX2(2,0,4)]+=tmp6_1 + tmp7_1;
1760 EM_S[INDEX2(1,3,4)]+=tmp2_1 + tmp5_1;
1761 EM_S[INDEX2(2,3,4)]+=tmp0_1 + tmp4_1;
1762 EM_S[INDEX2(2,2,4)]+=tmp0_1 + tmp6_1;
1763 EM_S[INDEX2(1,0,4)]+=tmp1_1 + tmp3_1;
1764 EM_S[INDEX2(0,3,4)]+=tmp1_1 + tmp7_1;
1765 EM_S[INDEX2(1,1,4)]+=tmp3_1 + tmp5_1;
1766 }
1767 }
1768 ///////////////
1769 // process C //
1770 ///////////////
1771 if (!C.isEmpty()) {
1772 add_EM_S=true;
1773 const double* C_p=const_cast<escript::Data*>(&C)->getSampleDataRO(e);
1774 if (C.actsExpanded()) {
1775 const register double C_0_0 = C_p[INDEX2(0,0,2)];
1776 const register double C_1_0 = C_p[INDEX2(1,0,2)];
1777 const register double C_0_1 = C_p[INDEX2(0,1,2)];
1778 const register double C_1_1 = C_p[INDEX2(1,1,2)];
1779 const register double C_0_2 = C_p[INDEX2(0,2,2)];
1780 const register double C_1_2 = C_p[INDEX2(1,2,2)];
1781 const register double C_0_3 = C_p[INDEX2(0,3,2)];
1782 const register double C_1_3 = C_p[INDEX2(1,3,2)];
1783 const register double tmp2_0 = C_0_1 + C_0_3;
1784 const register double tmp1_0 = C_1_2 + C_1_3;
1785 const register double tmp3_0 = C_0_0 + C_0_2;
1786 const register double tmp0_0 = C_1_0 + C_1_1;
1787 const register double tmp64_1 = C_0_2*w30;
1788 const register double tmp14_1 = C_0_2*w28;
1789 const register double tmp19_1 = C_0_3*w31;
1790 const register double tmp22_1 = C_1_0*w40;
1791 const register double tmp37_1 = tmp3_0*w35;
1792 const register double tmp29_1 = C_0_1*w35;
1793 const register double tmp73_1 = C_0_2*w37;
1794 const register double tmp74_1 = C_1_2*w41;
1795 const register double tmp52_1 = C_1_3*w39;
1796 const register double tmp25_1 = C_1_1*w39;
1797 const register double tmp62_1 = C_0_1*w31;
1798 const register double tmp79_1 = C_1_1*w40;
1799 const register double tmp43_1 = C_1_1*w36;
1800 const register double tmp27_1 = C_0_3*w38;
1801 const register double tmp28_1 = C_1_0*w42;
1802 const register double tmp63_1 = C_1_1*w42;
1803 const register double tmp59_1 = C_0_3*w34;
1804 const register double tmp72_1 = C_1_3*w29;
1805 const register double tmp40_1 = tmp2_0*w35;
1806 const register double tmp13_1 = C_0_3*w30;
1807 const register double tmp51_1 = C_1_2*w40;
1808 const register double tmp54_1 = C_1_2*w42;
1809 const register double tmp12_1 = C_0_0*w31;
1810 const register double tmp2_1 = tmp1_0*w32;
1811 const register double tmp68_1 = C_0_2*w31;
1812 const register double tmp75_1 = C_1_0*w32;
1813 const register double tmp49_1 = C_1_1*w41;
1814 const register double tmp4_1 = C_0_2*w35;
1815 const register double tmp66_1 = C_0_3*w28;
1816 const register double tmp56_1 = C_0_1*w37;
1817 const register double tmp5_1 = C_0_1*w34;
1818 const register double tmp38_1 = tmp2_0*w34;
1819 const register double tmp76_1 = C_0_1*w38;
1820 const register double tmp21_1 = C_0_1*w28;
1821 const register double tmp69_1 = C_0_1*w30;
1822 const register double tmp53_1 = C_1_0*w36;
1823 const register double tmp42_1 = C_1_2*w39;
1824 const register double tmp32_1 = tmp1_0*w29;
1825 const register double tmp45_1 = C_1_0*w43;
1826 const register double tmp33_1 = tmp0_0*w32;
1827 const register double tmp35_1 = C_1_0*w41;
1828 const register double tmp26_1 = C_1_2*w36;
1829 const register double tmp67_1 = C_0_0*w33;
1830 const register double tmp31_1 = C_0_2*w34;
1831 const register double tmp20_1 = C_0_0*w30;
1832 const register double tmp70_1 = C_0_0*w28;
1833 const register double tmp8_1 = tmp0_0*w39;
1834 const register double tmp30_1 = C_1_3*w43;
1835 const register double tmp0_1 = tmp0_0*w29;
1836 const register double tmp17_1 = C_1_3*w41;
1837 const register double tmp58_1 = C_0_0*w35;
1838 const register double tmp9_1 = tmp3_0*w33;
1839 const register double tmp61_1 = C_1_3*w36;
1840 const register double tmp41_1 = tmp3_0*w34;
1841 const register double tmp50_1 = C_1_3*w32;
1842 const register double tmp18_1 = C_1_1*w32;
1843 const register double tmp6_1 = tmp1_0*w36;
1844 const register double tmp3_1 = C_0_0*w38;
1845 const register double tmp34_1 = C_1_1*w29;
1846 const register double tmp77_1 = C_0_3*w35;
1847 const register double tmp65_1 = C_1_2*w43;
1848 const register double tmp71_1 = C_0_3*w33;
1849 const register double tmp55_1 = C_1_1*w43;
1850 const register double tmp46_1 = tmp3_0*w28;
1851 const register double tmp24_1 = C_0_0*w37;
1852 const register double tmp10_1 = tmp1_0*w39;
1853 const register double tmp48_1 = C_1_0*w29;
1854 const register double tmp15_1 = C_0_1*w33;
1855 const register double tmp36_1 = C_1_2*w32;
1856 const register double tmp60_1 = C_1_0*w39;
1857 const register double tmp47_1 = tmp2_0*w33;
1858 const register double tmp16_1 = C_1_2*w29;
1859 const register double tmp1_1 = C_0_3*w37;
1860 const register double tmp7_1 = tmp2_0*w28;
1861 const register double tmp39_1 = C_1_3*w40;
1862 const register double tmp44_1 = C_1_3*w42;
1863 const register double tmp57_1 = C_0_2*w38;
1864 const register double tmp78_1 = C_0_0*w34;
1865 const register double tmp11_1 = tmp0_0*w36;
1866 const register double tmp23_1 = C_0_2*w33;
1867 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1 + tmp4_1 + tmp5_1;
1868 EM_S[INDEX2(1,2,4)]+=tmp6_1 + tmp7_1 + tmp8_1 + tmp9_1;
1869 EM_S[INDEX2(3,2,4)]+=tmp10_1 + tmp11_1 + tmp12_1 + tmp13_1 + tmp14_1 + tmp15_1;
1870 EM_S[INDEX2(0,0,4)]+=tmp16_1 + tmp17_1 + tmp18_1 + tmp19_1 + tmp20_1 + tmp21_1 + tmp22_1 + tmp23_1;
1871 EM_S[INDEX2(3,3,4)]+=tmp24_1 + tmp25_1 + tmp26_1 + tmp27_1 + tmp28_1 + tmp29_1 + tmp30_1 + tmp31_1;
1872 EM_S[INDEX2(3,0,4)]+=tmp32_1 + tmp33_1 + tmp7_1 + tmp9_1;
1873 EM_S[INDEX2(3,1,4)]+=tmp34_1 + tmp35_1 + tmp36_1 + tmp37_1 + tmp38_1 + tmp39_1;
1874 EM_S[INDEX2(2,1,4)]+=tmp32_1 + tmp33_1 + tmp40_1 + tmp41_1;
1875 EM_S[INDEX2(0,2,4)]+=tmp42_1 + tmp43_1 + tmp44_1 + tmp45_1 + tmp46_1 + tmp47_1;
1876 EM_S[INDEX2(2,0,4)]+=tmp46_1 + tmp47_1 + tmp48_1 + tmp49_1 + tmp50_1 + tmp51_1;
1877 EM_S[INDEX2(1,3,4)]+=tmp37_1 + tmp38_1 + tmp52_1 + tmp53_1 + tmp54_1 + tmp55_1;
1878 EM_S[INDEX2(2,3,4)]+=tmp10_1 + tmp11_1 + tmp56_1 + tmp57_1 + tmp58_1 + tmp59_1;
1879 EM_S[INDEX2(2,2,4)]+=tmp60_1 + tmp61_1 + tmp62_1 + tmp63_1 + tmp64_1 + tmp65_1 + tmp66_1 + tmp67_1;
1880 EM_S[INDEX2(1,0,4)]+=tmp0_1 + tmp2_1 + tmp68_1 + tmp69_1 + tmp70_1 + tmp71_1;
1881 EM_S[INDEX2(0,3,4)]+=tmp40_1 + tmp41_1 + tmp6_1 + tmp8_1;
1882 EM_S[INDEX2(1,1,4)]+=tmp72_1 + tmp73_1 + tmp74_1 + tmp75_1 + tmp76_1 + tmp77_1 + tmp78_1 + tmp79_1;
1883 } else { /* constant data */
1884 const register double C_0 = C_p[0];
1885 const register double C_1 = C_p[1];
1886 const register double tmp1_1 = C_1*w45;
1887 const register double tmp3_1 = C_0*w51;
1888 const register double tmp4_1 = C_0*w44;
1889 const register double tmp7_1 = C_0*w46;
1890 const register double tmp5_1 = C_1*w49;
1891 const register double tmp2_1 = C_1*w48;
1892 const register double tmp0_1 = C_0*w47;
1893 const register double tmp6_1 = C_1*w50;
1894 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1;
1895 EM_S[INDEX2(1,2,4)]+=tmp2_1 + tmp3_1;
1896 EM_S[INDEX2(3,2,4)]+=tmp2_1 + tmp4_1;
1897 EM_S[INDEX2(0,0,4)]+=tmp4_1 + tmp5_1;
1898 EM_S[INDEX2(3,3,4)]+=tmp0_1 + tmp6_1;
1899 EM_S[INDEX2(3,0,4)]+=tmp1_1 + tmp3_1;
1900 EM_S[INDEX2(3,1,4)]+=tmp5_1 + tmp7_1;
1901 EM_S[INDEX2(2,1,4)]+=tmp1_1 + tmp7_1;
1902 EM_S[INDEX2(0,2,4)]+=tmp3_1 + tmp6_1;
1903 EM_S[INDEX2(2,0,4)]+=tmp3_1 + tmp5_1;
1904 EM_S[INDEX2(1,3,4)]+=tmp6_1 + tmp7_1;
1905 EM_S[INDEX2(2,3,4)]+=tmp0_1 + tmp2_1;
1906 EM_S[INDEX2(2,2,4)]+=tmp4_1 + tmp6_1;
1907 EM_S[INDEX2(1,0,4)]+=tmp1_1 + tmp4_1;
1908 EM_S[INDEX2(0,3,4)]+=tmp2_1 + tmp7_1;
1909 EM_S[INDEX2(1,1,4)]+=tmp0_1 + tmp5_1;
1910 }
1911 }
1912 ///////////////
1913 // process D //
1914 ///////////////
1915 if (!D.isEmpty()) {
1916 add_EM_S=true;
1917 const double* D_p=const_cast<escript::Data*>(&D)->getSampleDataRO(e);
1918 if (D.actsExpanded()) {
1919 const register double D_0 = D_p[0];
1920 const register double D_1 = D_p[1];
1921 const register double D_2 = D_p[2];
1922 const register double D_3 = D_p[3];
1923 const register double tmp4_0 = D_1 + D_3;
1924 const register double tmp2_0 = D_0 + D_1 + D_2 + D_3;
1925 const register double tmp5_0 = D_0 + D_2;
1926 const register double tmp0_0 = D_0 + D_1;
1927 const register double tmp6_0 = D_0 + D_3;
1928 const register double tmp1_0 = D_2 + D_3;
1929 const register double tmp3_0 = D_1 + D_2;
1930 const register double tmp16_1 = D_1*w56;
1931 const register double tmp14_1 = tmp6_0*w54;
1932 const register double tmp8_1 = D_3*w55;
1933 const register double tmp2_1 = tmp2_0*w54;
1934 const register double tmp12_1 = tmp5_0*w52;
1935 const register double tmp4_1 = tmp0_0*w53;
1936 const register double tmp3_1 = tmp1_0*w52;
1937 const register double tmp13_1 = tmp4_0*w53;
1938 const register double tmp10_1 = tmp4_0*w52;
1939 const register double tmp1_1 = tmp1_0*w53;
1940 const register double tmp7_1 = D_3*w56;
1941 const register double tmp0_1 = tmp0_0*w52;
1942 const register double tmp11_1 = tmp5_0*w53;
1943 const register double tmp9_1 = D_0*w56;
1944 const register double tmp5_1 = tmp3_0*w54;
1945 const register double tmp18_1 = D_2*w56;
1946 const register double tmp17_1 = D_1*w55;
1947 const register double tmp6_1 = D_0*w55;
1948 const register double tmp15_1 = D_2*w55;
1949 EM_S[INDEX2(0,1,4)]+=tmp0_1 + tmp1_1;
1950 EM_S[INDEX2(1,2,4)]+=tmp2_1;
1951 EM_S[INDEX2(3,2,4)]+=tmp3_1 + tmp4_1;
1952 EM_S[INDEX2(0,0,4)]+=tmp5_1 + tmp6_1 + tmp7_1;
1953 EM_S[INDEX2(3,3,4)]+=tmp5_1 + tmp8_1 + tmp9_1;
1954 EM_S[INDEX2(3,0,4)]+=tmp2_1;
1955 EM_S[INDEX2(3,1,4)]+=tmp10_1 + tmp11_1;
1956 EM_S[INDEX2(2,1,4)]+=tmp2_1;
1957 EM_S[INDEX2(0,2,4)]+=tmp12_1 + tmp13_1;
1958 EM_S[INDEX2(2,0,4)]+=tmp12_1 + tmp13_1;
1959 EM_S[INDEX2(1,3,4)]+=tmp10_1 + tmp11_1;
1960 EM_S[INDEX2(2,3,4)]+=tmp3_1 + tmp4_1;
1961 EM_S[INDEX2(2,2,4)]+=tmp14_1 + tmp15_1 + tmp16_1;
1962 EM_S[INDEX2(1,0,4)]+=tmp0_1 + tmp1_1;
1963 EM_S[INDEX2(0,3,4)]+=tmp2_1;
1964 EM_S[INDEX2(1,1,4)]+=tmp14_1 + tmp17_1 + tmp18_1;
1965 } else { /* constant data */
1966 const register double D_0 = D_p[0];
1967 const register double tmp2_1 = D_0*w59;
1968 const register double tmp1_1 = D_0*w58;
1969 const register double tmp0_1 = D_0*w57;
1970 EM_S[INDEX2(0,1,4)]+=tmp0_1;
1971 EM_S[INDEX2(1,2,4)]+=tmp1_1;
1972 EM_S[INDEX2(3,2,4)]+=tmp0_1;
1973 EM_S[INDEX2(0,0,4)]+=tmp2_1;
1974 EM_S[INDEX2(3,3,4)]+=tmp2_1;
1975 EM_S[INDEX2(3,0,4)]+=tmp1_1;
1976 EM_S[INDEX2(3,1,4)]+=tmp0_1;
1977 EM_S[INDEX2(2,1,4)]+=tmp1_1;
1978 EM_S[INDEX2(0,2,4)]+=tmp0_1;
1979 EM_S[INDEX2(2,0,4)]+=tmp0_1;
1980 EM_S[INDEX2(1,3,4)]+=tmp0_1;
1981 EM_S[INDEX2(2,3,4)]+=tmp0_1;
1982 EM_S[INDEX2(2,2,4)]+=tmp2_1;
1983 EM_S[INDEX2(1,0,4)]+=tmp0_1;
1984 EM_S[INDEX2(0,3,4)]+=tmp1_1;
1985 EM_S[INDEX2(1,1,4)]+=tmp2_1;
1986 }
1987 }
1988 ///////////////
1989 // process X //
1990 ///////////////
1991 if (!X.isEmpty()) {
1992 add_EM_F=true;
1993 const double* X_p=const_cast<escript::Data*>(&X)->getSampleDataRO(e);
1994 if (X.actsExpanded()) {
1995 const register double X_0_0 = X_p[INDEX2(0,0,2)];
1996 const register double X_1_0 = X_p[INDEX2(1,0,2)];
1997 const register double X_0_1 = X_p[INDEX2(0,1,2)];
1998 const register double X_1_1 = X_p[INDEX2(1,1,2)];
1999 const register double X_0_2 = X_p[INDEX2(0,2,2)];
2000 const register double X_1_2 = X_p[INDEX2(1,2,2)];
2001 const register double X_0_3 = X_p[INDEX2(0,3,2)];
2002 const register double X_1_3 = X_p[INDEX2(1,3,2)];
2003 const register double tmp1_0 = X_1_1 + X_1_3;
2004 const register double tmp3_0 = X_0_0 + X_0_1;
2005 const register double tmp2_0 = X_1_0 + X_1_2;
2006 const register double tmp0_0 = X_0_2 + X_0_3;
2007 const register double tmp8_1 = tmp2_0*w66;
2008 const register double tmp5_1 = tmp3_0*w64;
2009 const register double tmp14_1 = tmp0_0*w64;
2010 const register double tmp3_1 = tmp3_0*w60;
2011 const register double tmp9_1 = tmp3_0*w63;
2012 const register double tmp13_1 = tmp3_0*w65;
2013 const register double tmp12_1 = tmp1_0*w66;
2014 const register double tmp10_1 = tmp0_0*w60;
2015 const register double tmp2_1 = tmp2_0*w61;
2016 const register double tmp6_1 = tmp2_0*w62;
2017 const register double tmp4_1 = tmp0_0*w65;
2018 const register double tmp11_1 = tmp1_0*w67;
2019 const register double tmp1_1 = tmp1_0*w62;
2020 const register double tmp7_1 = tmp1_0*w61;
2021 const register double tmp0_1 = tmp0_0*w63;
2022 const register double tmp15_1 = tmp2_0*w67;
2023 EM_F[0]+=tmp0_1 + tmp1_1 + tmp2_1 + tmp3_1;
2024 EM_F[1]+=tmp4_1 + tmp5_1 + tmp6_1 + tmp7_1;
2025 EM_F[2]+=tmp10_1 + tmp11_1 + tmp8_1 + tmp9_1;
2026 EM_F[3]+=tmp12_1 + tmp13_1 + tmp14_1 + tmp15_1;
2027 } else { /* constant data */
2028 const register double X_0 = X_p[0];
2029 const register double X_1 = X_p[1];
2030 const register double tmp3_1 = X_1*w71;
2031 const register double tmp2_1 = X_0*w70;
2032 const register double tmp1_1 = X_0*w68;
2033 const register double tmp0_1 = X_1*w69;
2034 EM_F[0]+=tmp0_1 + tmp1_1;
2035 EM_F[1]+=tmp0_1 + tmp2_1;
2036 EM_F[2]+=tmp1_1 + tmp3_1;
2037 EM_F[3]+=tmp2_1 + tmp3_1;
2038 }
2039 }
2040 ///////////////
2041 // process Y //
2042 ///////////////
2043 if (!Y.isEmpty()) {
2044 add_EM_F=true;
2045 const double* Y_p=const_cast<escript::Data*>(&Y)->getSampleDataRO(e);
2046 if (Y.actsExpanded()) {
2047 const register double Y_0 = Y_p[0];
2048 const register double Y_1 = Y_p[1];
2049 const register double Y_2 = Y_p[2];
2050 const register double Y_3 = Y_p[3];
2051 const register double tmp0_0 = Y_1 + Y_2;
2052 const register double tmp1_0 = Y_0 + Y_3;
2053 const register double tmp9_1 = Y_0*w74;
2054 const register double tmp4_1 = tmp1_0*w73;
2055 const register double tmp5_1 = Y_2*w74;
2056 const register double tmp7_1 = Y_1*w74;
2057 const register double tmp6_1 = Y_2*w72;
2058 const register double tmp2_1 = Y_3*w74;
2059 const register double tmp8_1 = Y_3*w72;
2060 const register double tmp3_1 = Y_1*w72;
2061 const register double tmp0_1 = Y_0*w72;
2062 const register double tmp1_1 = tmp0_0*w73;
2063 EM_F[0]+=tmp0_1 + tmp1_1 + tmp2_1;
2064 EM_F[1]+=tmp3_1 + tmp4_1 + tmp5_1;
2065 EM_F[2]+=tmp4_1 + tmp6_1 + tmp7_1;
2066 EM_F[3]+=tmp1_1 + tmp8_1 + tmp9_1;
2067 } else { /* constant data */
2068 const register double Y_0 = Y_p[0];
2069 const register double tmp0_1 = Y_0*w75;
2070 EM_F[0]+=tmp0_1;
2071 EM_F[1]+=tmp0_1;
2072 EM_F[2]+=tmp0_1;
2073 EM_F[3]+=tmp0_1;
2074 }
2075 }
2076 /* GENERATOR SNIP_PDE_SINGLE BOTTOM */
2077
2078 // add to matrix (if add_EM_S) and RHS (if add_EM_F)
2079 const index_t firstNode=m_N0*k1+k0;
2080 IndexVector rowIndex;
2081 rowIndex.push_back(m_dofMap[firstNode]);
2082 rowIndex.push_back(m_dofMap[firstNode+1]);
2083 rowIndex.push_back(m_dofMap[firstNode+m_N0]);
2084 rowIndex.push_back(m_dofMap[firstNode+m_N0+1]);
2085 if (add_EM_F) {
2086 //cout << "-- AddtoRHS -- " << endl;
2087 double *F_p=rhs.getSampleDataRW(0);
2088 for (index_t i=0; i<rowIndex.size(); i++) {
2089 if (rowIndex[i]<getNumDOF()) {
2090 F_p[rowIndex[i]]+=EM_F[i];
2091 //cout << "F[" << rowIndex[i] << "]=" << F_p[rowIndex[i]] << endl;
2092 }
2093 }
2094 //cout << "---"<<endl;
2095 }
2096 if (add_EM_S) {
2097 //cout << "-- AddtoSystem -- " << endl;
2098 addToSystemMatrix(mat, rowIndex, 1, rowIndex, 1, EM_S);
2099 }
2100
2101 } // end k0 loop
2102 } // end k1 loop
2103 } // end of coloring
2104 } // end of parallel region
2105 }
2106
2107 } // end of namespace ripley
2108
  ViewVC Help
Powered by ViewVC 1.1.26