/[escript]/trunk/paso/src/Smoother.c
ViewVC logotype

Contents of /trunk/paso/src/Smoother.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 3312 - (show annotations)
Tue Oct 26 07:54:58 2010 UTC (8 years, 4 months ago) by gross
File MIME type: text/plain
File size: 14600 byte(s)
last step for a clean up version of the AMG
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
17 /* Paso: Gauss-Seidel */
18
19 /**************************************************************/
20
21 /* Author: artak@uq.edu.au */
22
23 /**************************************************************/
24
25 #include "Paso.h"
26 #include "Preconditioner.h"
27 #include "PasoUtil.h"
28 #include "BlockOps.h"
29
30 #include <stdio.h>
31
32
33 /**************************************************************/
34
35 /* free all memory used by Smoother */
36
37 void Paso_Preconditioner_Smoother_free(Paso_Preconditioner_Smoother * in) {
38 if (in!=NULL) {
39 Paso_Preconditioner_LocalSmoother_free(in->localSmoother);
40 MEMFREE(in);
41 }
42 }
43 void Paso_Preconditioner_LocalSmoother_free(Paso_Preconditioner_LocalSmoother * in) {
44 if (in!=NULL) {
45 MEMFREE(in->diag);
46 MEMFREE(in->pivot);
47 MEMFREE(in->buffer);
48 MEMFREE(in);
49 }
50 }
51 /**************************************************************/
52
53 /* constructs the symmetric Gauss-Seidel preconditioner
54
55 */
56 Paso_Preconditioner_Smoother* Paso_Preconditioner_Smoother_alloc(Paso_SystemMatrix * A_p, const bool_t jacobi, const bool_t is_local, const bool_t verbose)
57 {
58
59 /* allocations: */
60 Paso_Preconditioner_Smoother* out=MEMALLOC(1,Paso_Preconditioner_Smoother);
61 if (! Esys_checkPtr(out)) {
62 out->localSmoother=Paso_Preconditioner_LocalSmoother_alloc(A_p->mainBlock,jacobi,verbose);
63 out->is_local=is_local;
64 }
65 if (Esys_MPIInfo_noError(A_p->mpi_info)) {
66 return out;
67 } else {
68 Paso_Preconditioner_Smoother_free(out);
69 return NULL;
70 }
71 }
72 Paso_Preconditioner_LocalSmoother* Paso_Preconditioner_LocalSmoother_alloc(Paso_SparseMatrix * A_p, const bool_t jacobi, bool_t verbose)
73 {
74
75 dim_t n=A_p->numRows;
76 dim_t n_block=A_p->row_block_size;
77 dim_t block_size=A_p->block_size;
78
79 double time0=Esys_timer();
80 /* allocations: */
81 Paso_Preconditioner_LocalSmoother* out=MEMALLOC(1,Paso_Preconditioner_LocalSmoother);
82 if (! Esys_checkPtr(out)) {
83
84 out->diag=MEMALLOC( ((size_t) n) * ((size_t) block_size),double);
85 out->pivot=MEMALLOC( ((size_t) n) * ((size_t) n_block), index_t);
86 out->buffer=MEMALLOC( ((size_t) n) * ((size_t) n_block), double);
87 out->Jacobi=jacobi;
88
89 if ( ! ( Esys_checkPtr(out->diag) || Esys_checkPtr(out->pivot) ) ) {
90 Paso_SparseMatrix_invMain(A_p, out->diag, out->pivot );
91 }
92
93 }
94 time0=Esys_timer()-time0;
95
96 if (Esys_noError()) {
97 return out;
98 } else {
99 Paso_Preconditioner_LocalSmoother_free(out);
100 return NULL;
101 }
102 }
103
104 /*
105
106 performs a few sweeps of the from
107
108 S (x_{k} - x_{k-1}) = b - A x_{k-1}
109
110 where x_{0}=0 and S provides some approximatioon of A.
111
112 Under MPI S is build on using A_p->mainBlock only.
113 if Smoother is local the defect b - A x_{k-1} is calculated using A_p->mainBlock only.
114
115 */
116
117 void Paso_Preconditioner_Smoother_solve(Paso_SystemMatrix* A_p, Paso_Preconditioner_Smoother * smoother, double * x, const double * b,
118 const dim_t sweeps, const bool_t x_is_initial)
119 {
120 const dim_t n= (A_p->mainBlock->numRows) * (A_p->mainBlock->row_block_size);
121
122 double *b_new = smoother->localSmoother->buffer;
123 dim_t nsweeps=sweeps;
124 if (smoother->is_local) {
125 Paso_Preconditioner_LocalSmoother_solve(A_p->mainBlock,smoother->localSmoother,x,b,sweeps,x_is_initial);
126 } else {
127 if (! x_is_initial) {
128 Paso_Copy(n, x, b);
129 Paso_Preconditioner_LocalSmoother_Sweep(A_p->mainBlock,smoother->localSmoother,x);
130 nsweeps--;
131 }
132 while (nsweeps > 0 ) {
133 Paso_Copy(n, b_new, b);
134 Paso_SystemMatrix_MatrixVector((-1.), A_p, x, 1., b_new); /* b_new = b - A*x */
135 Paso_Preconditioner_LocalSmoother_Sweep(A_p->mainBlock,smoother->localSmoother,b_new);
136 Paso_AXPY(n, x, 1., b_new);
137 nsweeps--;
138 }
139
140 }
141 }
142 void Paso_Preconditioner_LocalSmoother_solve(Paso_SparseMatrix* A_p, Paso_Preconditioner_LocalSmoother * smoother, double * x, const double * b,
143 const dim_t sweeps, const bool_t x_is_initial)
144 {
145 const dim_t n= (A_p->numRows) * (A_p->row_block_size);
146 double *b_new = smoother->buffer;
147 dim_t nsweeps=sweeps;
148
149 if (! x_is_initial) {
150 Paso_Copy(n, x, b);
151 Paso_Preconditioner_LocalSmoother_Sweep(A_p,smoother,x);
152 nsweeps--;
153 }
154
155 while (nsweeps > 0 ) {
156 Paso_Copy(n, b_new, b);
157
158 Paso_SparseMatrix_MatrixVector_CSR_OFFSET0((-1.), A_p, x, 1., b_new); /* b_new = b - A*x */
159 Paso_Preconditioner_LocalSmoother_Sweep(A_p,smoother,b_new);
160 Paso_AXPY(n, x, 1., b_new);
161 nsweeps--;
162 }
163 }
164
165 void Paso_Preconditioner_LocalSmoother_Sweep(Paso_SparseMatrix* A, Paso_Preconditioner_LocalSmoother * smoother, double * x)
166 {
167 const dim_t nt=omp_get_max_threads();
168 if (smoother->Jacobi) {
169 Paso_BlockOps_allMV(A->row_block_size,A->numRows,smoother->diag,smoother->pivot,x);
170 } else {
171 if (nt < 2) {
172 Paso_Preconditioner_LocalSmoother_Sweep_sequential(A,smoother,x);
173 } else {
174 Paso_Preconditioner_LocalSmoother_Sweep_colored(A,smoother,x);
175 }
176 }
177 }
178
179 /* inplace Gauss-Seidel sweep in seqential mode: */
180
181 void Paso_Preconditioner_LocalSmoother_Sweep_sequential(Paso_SparseMatrix* A_p, Paso_Preconditioner_LocalSmoother * smoother, double * x)
182 {
183 const dim_t n=A_p->numRows;
184 const dim_t n_block=A_p->row_block_size;
185 const double *diag = smoother->diag;
186 /* const index_t* pivot = smoother->pivot;
187 const dim_t block_size=A_p->block_size; use for block size >3*/
188
189 register dim_t i,k;
190 register index_t iptr_ik, mm;
191
192 const index_t* ptr_main = Paso_SparseMatrix_borrowMainDiagonalPointer(A_p);
193 /* forward substitution */
194
195 if (n_block==1) {
196 Paso_BlockOps_MV_1(&x[0], &diag[0], &x[0]);
197 for (i = 1; i < n; ++i) {
198 mm=ptr_main[i];
199 /* x_i=x_i-a_ik*x_k (with k<i) */
200 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<mm; ++iptr_ik) {
201 k=A_p->pattern->index[iptr_ik];
202 Paso_BlockOps_SMV_1(&x[i], &A_p->val[iptr_ik], &x[k]);
203 }
204 Paso_BlockOps_MV_1(&x[i], &diag[i], &x[i]);
205 }
206 } else if (n_block==2) {
207 Paso_BlockOps_MV_2(&x[0], &diag[0], &x[0]);
208 for (i = 1; i < n; ++i) {
209 mm=ptr_main[i];
210 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<mm; ++iptr_ik) {
211 k=A_p->pattern->index[iptr_ik];
212 Paso_BlockOps_SMV_2(&x[2*i], &A_p->val[4*iptr_ik], &x[2*k]);
213 }
214 Paso_BlockOps_MV_2(&x[2*i], &diag[4*i], &x[2*i]);
215 }
216 } else if (n_block==3) {
217 Paso_BlockOps_MV_3(&x[0], &diag[0], &x[0]);
218 for (i = 1; i < n; ++i) {
219 mm=ptr_main[i];
220 /* x_i=x_i-a_ik*x_k */
221 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<mm; ++iptr_ik) {
222 k=A_p->pattern->index[iptr_ik];
223 Paso_BlockOps_SMV_3(&x[3*i], &A_p->val[9*iptr_ik], &x[3*k]);
224 }
225 Paso_BlockOps_MV_3(&x[3*i], &diag[9*i], &x[3*i]);
226 }
227 } /* add block size >3 */
228
229 /* backward substitution */
230
231 if (n_block==1) {
232 for (i = n-2; i > -1; --i) {
233 mm=ptr_main[i];
234 Paso_BlockOps_MV_1(&x[i], &A_p->val[mm], &x[i]);
235 for (iptr_ik=mm+1; iptr_ik < A_p->pattern->ptr[i+1]; ++iptr_ik) {
236 k=A_p->pattern->index[iptr_ik];
237 Paso_BlockOps_SMV_1(&x[i], &A_p->val[iptr_ik], &x[k]);
238 }
239 Paso_BlockOps_MV_1(&x[i], &diag[i], &x[i]);
240 }
241
242 } else if (n_block==2) {
243 for (i = n-2; i > -1; --i) {
244 mm=ptr_main[i];
245 Paso_BlockOps_MV_2(&x[2*i], &A_p->val[4*mm], &x[2*i]);
246 for (iptr_ik=mm+1; iptr_ik < A_p->pattern->ptr[i+1]; ++iptr_ik) {
247 k=A_p->pattern->index[iptr_ik];
248 Paso_BlockOps_SMV_2(&x[2*i], &A_p->val[4*iptr_ik], &x[2*k]);
249 }
250 Paso_BlockOps_MV_2(&x[2*i], &diag[i*4], &x[2*i]);
251
252 }
253 } else if (n_block==3) {
254 for (i = n-2; i > -1; --i) {
255 mm=ptr_main[i];
256 Paso_BlockOps_MV_3(&x[3*i], &A_p->val[9*mm], &x[3*i]);
257
258 for (iptr_ik=mm+1; iptr_ik < A_p->pattern->ptr[i+1]; ++iptr_ik) {
259 k=A_p->pattern->index[iptr_ik];
260 Paso_BlockOps_SMV_3(&x[3*i], &A_p->val[9*iptr_ik], &x[3*k]);
261 }
262 Paso_BlockOps_MV_3(&x[3*i], &diag[i*9], &x[3*i]);
263 }
264
265 } /* add block size >3 */
266
267 return;
268 }
269
270 void Paso_Preconditioner_LocalSmoother_Sweep_colored(Paso_SparseMatrix* A_p, Paso_Preconditioner_LocalSmoother * smoother, double * x)
271 {
272 const dim_t n=A_p->numRows;
273 const dim_t n_block=A_p->row_block_size;
274 const double *diag = smoother->diag;
275 index_t* pivot = smoother->pivot;
276 const dim_t block_size=A_p->block_size;
277
278 register dim_t i,k;
279 register index_t color,iptr_ik, mm;
280
281 const index_t* coloring = Paso_Pattern_borrowColoringPointer(A_p->pattern);
282 const dim_t num_colors = Paso_Pattern_getNumColors(A_p->pattern);
283 const index_t* ptr_main = Paso_SparseMatrix_borrowMainDiagonalPointer(A_p);
284
285 /* forward substitution */
286
287
288 /* color = 0 */
289 if (n_block==1) {
290 #pragma omp parallel for schedule(static) private(i)
291 for (i = 0; i < n; ++i) {
292 if (coloring[i]== 0 ) Paso_BlockOps_MV_1(&x[i], &diag[i], &x[i]);
293 }
294 } else if (n_block==2) {
295 #pragma omp parallel for schedule(static) private(i)
296 for (i = 0; i < n; ++i) {
297 if (coloring[i]== 0 ) Paso_BlockOps_MV_2(&x[2*i], &diag[i*4], &x[2*i]);
298 }
299 } else if (n_block==3) {
300 #pragma omp parallel for schedule(static) private(i)
301 for (i = 0; i < n; ++i) {
302 if (coloring[i]==0) Paso_BlockOps_MV_3(&x[3*i], &diag[i*9], &x[3*i]);
303 }
304 } else {
305 #pragma omp parallel for schedule(static) private(i)
306 for (i = 0; i < n; ++i) {
307 if (coloring[i]==0) Paso_BlockOps_Solve_N(n_block, &x[n_block*i], &diag[block_size*i], &pivot[n_block*i], &x[n_block*i]);
308 }
309 }
310
311 for (color=1;color<num_colors;++color) {
312
313 if (n_block==1) {
314 #pragma omp parallel for schedule(static) private(i,iptr_ik,k)
315 for (i = 0; i < n; ++i) {
316 if (coloring[i]==color) {
317 /* x_i=x_i-a_ik*x_k */
318 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
319 k=A_p->pattern->index[iptr_ik];
320 if (coloring[k]<color) Paso_BlockOps_SMV_1(&x[i], &A_p->val[iptr_ik], &x[k]);
321 }
322 Paso_BlockOps_MV_1(&x[i], &diag[i], &x[i]);
323 }
324 }
325 } else if (n_block==2) {
326 #pragma omp parallel for schedule(static) private(i,iptr_ik,k)
327 for (i = 0; i < n; ++i) {
328 if (coloring[i]==color) {
329 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
330 k=A_p->pattern->index[iptr_ik];
331 if (coloring[k]<color) Paso_BlockOps_SMV_2(&x[2*i], &A_p->val[4*iptr_ik], &x[2*k]);
332 }
333 Paso_BlockOps_MV_2(&x[2*i], &diag[4*i], &x[2*i]);
334 }
335 }
336 } else if (n_block==3) {
337 #pragma omp parallel for schedule(static) private(i,iptr_ik,k)
338 for (i = 0; i < n; ++i) {
339 if (coloring[i]==color) {
340 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
341 k=A_p->pattern->index[iptr_ik];
342 if (coloring[k]<color) Paso_BlockOps_SMV_3(&x[3*i], &A_p->val[9*iptr_ik], &x[3*k]);
343 }
344 Paso_BlockOps_MV_3(&x[3*i], &diag[9*i], &x[3*i]);
345 }
346 }
347 } else {
348 #pragma omp parallel for schedule(static) private(i,iptr_ik,k)
349 for (i = 0; i < n; ++i) {
350 if (coloring[i] == color) {
351 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
352 k=A_p->pattern->index[iptr_ik];
353 if (coloring[k]<color) Paso_BlockOps_SMV_N(n_block, &x[n_block*i], &A_p->val[block_size*iptr_ik], &x[n_block*k]);
354 }
355 Paso_BlockOps_Solve_N(n_block, &x[n_block*i], &diag[block_size*i], &pivot[n_block*i], &x[n_block*i]);
356 }
357 }
358 }
359 } /* end of coloring loop */
360
361 /* backward substitution */
362 for (color=(num_colors)-2 ;color>-1;--color) { /* Note: color=(num_colors)-1 is not required */
363 if (n_block==1) {
364 #pragma omp parallel for schedule(static) private(mm, i,iptr_ik,k)
365 for (i = 0; i < n; ++i) {
366 if (coloring[i]==color) {
367 mm=ptr_main[i];
368 Paso_BlockOps_MV_1(&x[i], &A_p->val[mm], &x[i]);
369 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
370 k=A_p->pattern->index[iptr_ik];
371 if (coloring[k]>color) Paso_BlockOps_SMV_1(&x[i], &A_p->val[iptr_ik], &x[k]);
372 }
373 Paso_BlockOps_MV_1(&x[i], &diag[i], &x[i]);
374 }
375 }
376 } else if (n_block==2) {
377 #pragma omp parallel for schedule(static) private(mm, i,iptr_ik,k)
378 for (i = 0; i < n; ++i) {
379 if (coloring[i]==color) {
380 mm=ptr_main[i];
381 Paso_BlockOps_MV_2(&x[2*i], &A_p->val[4*mm], &x[2*i]);
382 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
383 k=A_p->pattern->index[iptr_ik];
384 if (coloring[k]>color) Paso_BlockOps_SMV_2(&x[2*i], &A_p->val[4*iptr_ik], &x[2*k]);
385 }
386 Paso_BlockOps_MV_2(&x[2*i], &diag[4*i], &x[2*i]);
387 }
388 }
389 } else if (n_block==3) {
390 #pragma omp parallel for schedule(static) private(mm, i,iptr_ik,k)
391 for (i = 0; i < n; ++i) {
392 if (coloring[i]==color) {
393 mm=ptr_main[i];
394 Paso_BlockOps_MV_3(&x[3*i], &A_p->val[9*mm], &x[3*i]);
395 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
396 k=A_p->pattern->index[iptr_ik];
397 if (coloring[k]>color) Paso_BlockOps_SMV_3(&x[3*i], &A_p->val[9*iptr_ik], &x[3*k]);
398 }
399 Paso_BlockOps_MV_3(&x[3*i], &diag[9*i], &x[3*i]);
400 }
401 }
402 } else {
403 #pragma omp parallel for schedule(static) private(mm, i,iptr_ik,k)
404 for (i = 0; i < n; ++i) {
405 if (coloring[i]==color) {
406 mm=ptr_main[i];
407 Paso_BlockOps_MV_N( n_block, &x[n_block*i], &A_p->val[block_size*mm], &x[n_block*i] );
408 for (iptr_ik=A_p->pattern->ptr[i];iptr_ik<A_p->pattern->ptr[i+1]; ++iptr_ik) {
409 k=A_p->pattern->index[iptr_ik];
410 if (coloring[k]>color) Paso_BlockOps_SMV_N(n_block, &x[n_block*i], &A_p->val[block_size*iptr_ik], &x[n_block*k]);
411 }
412 Paso_BlockOps_Solve_N(n_block, &x[n_block*i], &diag[block_size*i], &pivot[n_block*i], &x[n_block*i]);
413 }
414 }
415 }
416 }
417 return;
418 }

  ViewVC Help
Powered by ViewVC 1.1.26