← Back to branch summary

~vcs-imports/escript-finley/trunk

« back to all changes in this revision

Viewing changes to dudley/src/Assemble_PDE_Single2_1D.c

  • Committer: jfenwick
  • Date: 2010-10-11 01:48:14 UTC
  • Revision ID: svn-v4:77569008-7704-0410-b7a0-a92fef0b09fd:trunk:3259
Merging dudley and scons updates from branches

Show diffs side-by-side

added added

removed removed

Lines of Context:
dudley/src/Assemble_PDE_Single2_1D.c
 
1
 
 
2
/*******************************************************
 
3
*
 
4
* Copyright (c) 2003-2010 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
/**************************************************************/
 
15
 
 
16
/*    assembles the system of numEq PDEs into the stiffness matrix S right hand side F  */
 
17
/*    the shape functions for test and solution must be identical */
 
18
 
 
19
/*      -(A_{i,j} u_,j)_i-(B_{i} u)_i+C_{j} u_,j-D u_m  and -(X_,i)_i + Y */
 
20
 
 
21
/*    in a 1D domain. The shape functions for test and solution must be identical  */
 
22
/*    and row_NS == row_NN                                                         */
 
23
 
 
24
/*    Shape of the coefficients: */
 
25
 
 
26
/*      A = 1 x 1 */
 
27
/*      B = 1   */
 
28
/*      C = 1   */
 
29
/*      D = scalar  */
 
30
/*      X = 1  */
 
31
/*      Y = scalar   */
 
32
 
 
33
/**************************************************************/
 
34
 
 
35
#include "Assemble.h"
 
36
#include "Util.h"
 
37
#ifdef _OPENMP
 
38
#include <omp.h>
 
39
#endif
 
40
 
 
41
/**************************************************************/
 
42
 
 
43
void Dudley_Assemble_PDE_Single2_1D(Dudley_Assemble_Parameters p, Dudley_ElementFile * elements,
 
44
                                    Paso_SystemMatrix * Mat, escriptDataC * F,
 
45
                                    escriptDataC * A, escriptDataC * B, escriptDataC * C, escriptDataC * D,
 
46
                                    escriptDataC * X, escriptDataC * Y)
 
47
{
 
48
 
 
49
#define DIM 1
 
50
    index_t color;
 
51
    dim_t e;
 
52
    __const double *A_p, *B_p, *C_p, *D_p, *X_p, *Y_p, *A_q, *B_q, *C_q, *D_q, *X_q, *Y_q;
 
53
    double *EM_S, *EM_F, *DSDX;
 
54
    index_t *row_index;
 
55
    register dim_t q, s, r;
 
56
    register double rtmp;
 
57
    bool_t add_EM_F, add_EM_S;
 
58
 
 
59
    bool_t extendedA = isExpanded(A);
 
60
    bool_t extendedB = isExpanded(B);
 
61
    bool_t extendedC = isExpanded(C);
 
62
    bool_t extendedD = isExpanded(D);
 
63
    bool_t extendedX = isExpanded(X);
 
64
    bool_t extendedY = isExpanded(Y);
 
65
    double *F_p = (requireWrite(F), getSampleDataRW(F, 0));     /* use comma, to get around the mixed code and declarations thing */
 
66
    double *S = p.row_jac->BasisFunctions->S;
 
67
    dim_t len_EM_S = p.row_numShapesTotal * p.row_numShapesTotal;
 
68
    dim_t len_EM_F = p.row_numShapesTotal;
 
69
 
 
70
#pragma omp parallel private(color, EM_S, EM_F, Vol, DSDX, A_p, B_p, C_p, D_p, X_p, Y_p, A_q, B_q, C_q, D_q, X_q, Y_q, row_index, q, s,r,rtmp,add_EM_F, add_EM_S)
 
71
    {
 
72
        EM_S = THREAD_MEMALLOC(len_EM_S, double);
 
73
        EM_F = THREAD_MEMALLOC(len_EM_F, double);
 
74
        row_index = THREAD_MEMALLOC(p.row_numShapesTotal, index_t);
 
75
 
 
76
        if (!Dudley_checkPtr(EM_S) && !Dudley_checkPtr(EM_F) && !Dudley_checkPtr(row_index))
 
77
        {
 
78
 
 
79
            for (color = elements->minColor; color <= elements->maxColor; color++)
 
80
            {
 
81
                /*  open loop over all elements: */
 
82
#pragma omp for private(e) schedule(static)
 
83
                for (e = 0; e < elements->numElements; e++)
 
84
                {
 
85
                    if (elements->Color[e] == color)
 
86
                    {
 
87
 
 
88
                        A_p = getSampleDataRO(A, e);
 
89
                        C_p = getSampleDataRO(C, e);
 
90
                        B_p = getSampleDataRO(B, e);
 
91
                        D_p = getSampleDataRO(D, e);
 
92
                        X_p = getSampleDataRO(X, e);
 
93
                        Y_p = getSampleDataRO(Y, e);
 
94
 
 
95
                        double vol = p.row_jac->absD[e] * p.row_jac->quadweight;
 
96
                        DSDX = &(p.row_jac->DSDX[INDEX5(0, 0, 0, 0, e, p.row_numShapesTotal, DIM, p.numQuadTotal, 1)]);
 
97
                        for (q = 0; q < len_EM_S; ++q)
 
98
                            EM_S[q] = 0;
 
99
                        for (q = 0; q < len_EM_F; ++q)
 
100
                            EM_F[q] = 0;
 
101
                        add_EM_F = FALSE;
 
102
                        add_EM_S = FALSE;
 
103
                      /**************************************************************/
 
104
                        /*   process A: */
 
105
                      /**************************************************************/
 
106
                        if (NULL != A_p)
 
107
                        {
 
108
                            add_EM_S = TRUE;
 
109
                            if (extendedA)
 
110
                            {
 
111
                                A_q = &(A_p[INDEX4(0, 0, 0, 0, DIM, DIM, p.numQuadTotal)]);
 
112
                                for (s = 0; s < p.row_numShapes; s++)
 
113
                                {
 
114
                                    for (r = 0; r < p.row_numShapes; r++)
 
115
                                    {
 
116
                                        rtmp = 0;
 
117
                                        for (q = 0; q < p.numQuadTotal; q++)
 
118
                                        {
 
119
                                            rtmp +=
 
120
                                                vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)] *
 
121
                                                A_q[INDEX3(0, 0, q, DIM, DIM)] *
 
122
                                                DSDX[INDEX3(r, 0, q, p.row_numShapesTotal, DIM)];
 
123
                                        }
 
124
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] += rtmp;
 
125
                                    }
 
126
                                }
 
127
                            }
 
128
                            else
 
129
                            {
 
130
                                for (s = 0; s < p.row_numShapes; s++)
 
131
                                {
 
132
                                    for (r = 0; r < p.row_numShapes; r++)
 
133
                                    {
 
134
                                        rtmp = 0;
 
135
                                        for (q = 0; q < p.numQuadTotal; q++)
 
136
                                            rtmp +=
 
137
                                                vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)] *
 
138
                                                DSDX[INDEX3(r, 0, q, p.row_numShapesTotal, DIM)];
 
139
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] +=
 
140
                                            rtmp * A_p[INDEX2(0, 0, DIM)];
 
141
                                    }
 
142
                                }
 
143
                            }
 
144
                        }
 
145
                      /**************************************************************/
 
146
                        /*   process B: */
 
147
                      /**************************************************************/
 
148
                        if (NULL != B_p)
 
149
                        {
 
150
                            add_EM_S = TRUE;
 
151
                            if (extendedB)
 
152
                            {
 
153
                                B_q = &(B_p[INDEX3(0, 0, 0, DIM, p.numQuadTotal)]);
 
154
                                for (s = 0; s < p.row_numShapes; s++)
 
155
                                {
 
156
                                    for (r = 0; r < p.row_numShapes; r++)
 
157
                                    {
 
158
                                        rtmp = 0;
 
159
                                        for (q = 0; q < p.numQuadTotal; q++)
 
160
                                        {
 
161
                                            rtmp +=
 
162
                                                vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)] *
 
163
                                                B_q[INDEX2(0, q, DIM)] * S[INDEX2(r, q, p.row_numShapes)];
 
164
                                        }
 
165
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] += rtmp;
 
166
                                    }
 
167
                                }
 
168
                            }
 
169
                            else
 
170
                            {
 
171
                                for (s = 0; s < p.row_numShapes; s++)
 
172
                                {
 
173
                                    for (r = 0; r < p.row_numShapes; r++)
 
174
                                    {
 
175
                                        rtmp = 0;
 
176
                                        for (q = 0; q < p.numQuadTotal; q++)
 
177
                                            rtmp +=
 
178
                                                vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)] *
 
179
                                                S[INDEX2(r, q, p.row_numShapes)];
 
180
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] +=
 
181
                                            rtmp * B_p[0];
 
182
                                    }
 
183
                                }
 
184
                            }
 
185
                        }
 
186
                      /**************************************************************/
 
187
                        /*   process C: */
 
188
                      /**************************************************************/
 
189
                        if (NULL != C_p)
 
190
                        {
 
191
                            add_EM_S = TRUE;
 
192
                            if (extendedC)
 
193
                            {
 
194
                                C_q = &(C_p[INDEX3(0, 0, 0, DIM, p.numQuadTotal)]);
 
195
                                for (s = 0; s < p.row_numShapes; s++)
 
196
                                {
 
197
                                    for (r = 0; r < p.row_numShapes; r++)
 
198
                                    {
 
199
                                        rtmp = 0;
 
200
                                        for (q = 0; q < p.numQuadTotal; q++)
 
201
                                        {
 
202
                                            rtmp +=
 
203
                                                vol * S[INDEX2(s, q, p.row_numShapes)] * C_q[INDEX2(0, q, DIM)] *
 
204
                                                DSDX[INDEX3(r, 0, q, p.row_numShapesTotal, DIM)];
 
205
                                        }
 
206
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] += rtmp;
 
207
                                    }
 
208
                                }
 
209
                            }
 
210
                            else
 
211
                            {
 
212
                                for (s = 0; s < p.row_numShapes; s++)
 
213
                                {
 
214
                                    for (r = 0; r < p.row_numShapes; r++)
 
215
                                    {
 
216
                                        rtmp = 0;
 
217
                                        for (q = 0; q < p.numQuadTotal; q++)
 
218
                                            rtmp +=
 
219
                                                vol * S[INDEX2(s, q, p.row_numShapes)] *
 
220
                                                DSDX[INDEX3(r, 0, q, p.row_numShapesTotal, DIM)];
 
221
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] +=
 
222
                                            rtmp * C_p[0];
 
223
                                    }
 
224
                                }
 
225
                            }
 
226
                        }
 
227
                      /************************************************************* */
 
228
                        /* process D */
 
229
                      /**************************************************************/
 
230
                        if (NULL != D_p)
 
231
                        {
 
232
                            add_EM_S = TRUE;
 
233
                            if (extendedD)
 
234
                            {
 
235
                                D_q = &(D_p[INDEX2(0, 0, p.numQuadTotal)]);
 
236
                                for (s = 0; s < p.row_numShapes; s++)
 
237
                                {
 
238
                                    for (r = 0; r < p.row_numShapes; r++)
 
239
                                    {
 
240
                                        rtmp = 0;
 
241
                                        for (q = 0; q < p.numQuadTotal; q++)
 
242
                                        {
 
243
                                            rtmp +=
 
244
                                                vol * S[INDEX2(s, q, p.row_numShapes)] * D_q[q] *
 
245
                                                S[INDEX2(r, q, p.row_numShapes)];
 
246
                                        }
 
247
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] += rtmp;
 
248
                                    }
 
249
                                }
 
250
                            }
 
251
                            else
 
252
                            {
 
253
                                for (s = 0; s < p.row_numShapes; s++)
 
254
                                {
 
255
                                    for (r = 0; r < p.row_numShapes; r++)
 
256
                                    {
 
257
                                        rtmp = 0;
 
258
                                        for (q = 0; q < p.numQuadTotal; q++)
 
259
                                            rtmp +=
 
260
                                                vol * S[INDEX2(s, q, p.row_numShapes)] *
 
261
                                                S[INDEX2(r, q, p.row_numShapes)];
 
262
                                        EM_S[INDEX4(0, 0, s, r, p.numEqu, p.numComp, p.row_numShapesTotal)] +=
 
263
                                            rtmp * D_p[0];
 
264
                                    }
 
265
                                }
 
266
                            }
 
267
                        }
 
268
                      /**************************************************************/
 
269
                        /*   process X: */
 
270
                      /**************************************************************/
 
271
                        if (NULL != X_p)
 
272
                        {
 
273
                            add_EM_F = TRUE;
 
274
                            if (extendedX)
 
275
                            {
 
276
                                X_q = &(X_p[INDEX3(0, 0, 0, DIM, p.numQuadTotal)]);
 
277
                                for (s = 0; s < p.row_numShapes; s++)
 
278
                                {
 
279
                                    rtmp = 0;
 
280
                                    for (q = 0; q < p.numQuadTotal; q++)
 
281
                                        rtmp +=
 
282
                                            vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)] *
 
283
                                            X_q[INDEX2(0, q, DIM)];
 
284
                                    EM_F[INDEX2(0, s, p.numEqu)] += rtmp;
 
285
                                }
 
286
                            }
 
287
                            else
 
288
                            {
 
289
                                for (s = 0; s < p.row_numShapes; s++)
 
290
                                {
 
291
                                    rtmp = 0;
 
292
                                    for (q = 0; q < p.numQuadTotal; q++)
 
293
                                        rtmp += vol * DSDX[INDEX3(s, 0, q, p.row_numShapesTotal, DIM)];
 
294
                                    EM_F[INDEX2(0, s, p.numEqu)] += rtmp * X_p[0];
 
295
                                }
 
296
                            }
 
297
                        }
 
298
                     /**************************************************************/
 
299
                        /*   process Y: */
 
300
                     /**************************************************************/
 
301
                        if (NULL != Y_p)
 
302
                        {
 
303
                            add_EM_F = TRUE;
 
304
                            if (extendedY)
 
305
                            {
 
306
                                Y_q = &(Y_p[INDEX2(0, 0, p.numQuadTotal)]);
 
307
                                for (s = 0; s < p.row_numShapes; s++)
 
308
                                {
 
309
                                    rtmp = 0;
 
310
                                    for (q = 0; q < p.numQuadTotal; q++)
 
311
                                        rtmp += vol * S[INDEX2(s, q, p.row_numShapes)] * Y_q[q];
 
312
                                    EM_F[INDEX2(0, s, p.numEqu)] += rtmp;
 
313
                                }
 
314
                            }
 
315
                            else
 
316
                            {
 
317
                                for (s = 0; s < p.row_numShapes; s++)
 
318
                                {
 
319
                                    rtmp = 0;
 
320
                                    for (q = 0; q < p.numQuadTotal; q++)
 
321
                                        rtmp += vol * S[INDEX2(s, q, p.row_numShapes)];
 
322
                                    EM_F[INDEX2(0, s, p.numEqu)] += rtmp * Y_p[0];
 
323
                                }
 
324
                            }
 
325
                        }
 
326
                       /***********************************************************************************************/
 
327
                        /* add the element matrices onto the matrix and right hand side                                */
 
328
                       /***********************************************************************************************/
 
329
                        for (q = 0; q < p.row_numShapesTotal; q++)
 
330
                            row_index[q] = p.row_DOF[elements->Nodes[INDEX2(q, e, p.NN)]];
 
331
 
 
332
                        if (add_EM_F)
 
333
                            Dudley_Util_AddScatter(p.row_numShapesTotal, row_index, p.numEqu, EM_F, F_p,
 
334
                                                   p.row_DOF_UpperBound);
 
335
                        if (add_EM_S)
 
336
                            Dudley_Assemble_addToSystemMatrix(Mat, p.row_numShapesTotal, row_index, p.numEqu,
 
337
                                                              p.row_numShapesTotal, row_index, p.numComp, EM_S);
 
338
 
 
339
                    }           /* end color check */
 
340
                }               /* end element loop */
 
341
            }                   /* end color loop */
 
342
 
 
343
            THREAD_MEMFREE(EM_S);       /* these FREEs appear to be inside the if because if any of the allocs */
 
344
            THREAD_MEMFREE(EM_F);       /* failed it means an out of memory (which is not recoverable anyway) */
 
345
            THREAD_MEMFREE(row_index);
 
346
 
 
347
        }                       /* end of pointer check */
 
348
    }                           /* end parallel region */
 
349
}
 
350
 
 
351
/*
 
352
 * $Log$
 
353
 */