xref: /petsc/src/ksp/pc/impls/mg/mg.c (revision df48e8d96c19df54efdde2c76f874a16fcafd8fe)
1 #define PETSCKSP_DLL
2 
3 /*
4     Defines the multigrid preconditioner interface.
5 */
6 #include "src/ksp/pc/impls/mg/mgimpl.h"                    /*I "petscmg.h" I*/
7 
8 
9 #undef __FUNCT__
10 #define __FUNCT__ "PCMGMCycle_Private"
11 PetscErrorCode PCMGMCycle_Private(PC_MG **mglevels,PetscTruth *converged)
12 {
13   PC_MG          *mg = *mglevels,*mgc;
14   PetscErrorCode ierr;
15   PetscInt       cycles = mg->cycles;
16   PetscScalar    zero = 0.0;
17 
18   PetscFunctionBegin;
19   if (converged) *converged = PETSC_FALSE;
20 
21   if (mg->eventsolve) {ierr = PetscLogEventBegin(mg->eventsolve,0,0,0,0);CHKERRQ(ierr);}
22   ierr = KSPSolve(mg->smoothd,mg->b,mg->x);CHKERRQ(ierr);
23   if (mg->eventsolve) {ierr = PetscLogEventEnd(mg->eventsolve,0,0,0,0);CHKERRQ(ierr);}
24   if (mg->level) {  /* not the coarsest grid */
25     ierr = (*mg->residual)(mg->A,mg->b,mg->x,mg->r);CHKERRQ(ierr);
26 
27     /* if on finest level and have convergence criteria set */
28     if (mg->level == mg->levels-1 && mg->ttol) {
29       PetscReal rnorm;
30       ierr = VecNorm(mg->r,NORM_2,&rnorm);CHKERRQ(ierr);
31       if (rnorm <= mg->ttol) {
32         *converged = PETSC_TRUE;
33         if (rnorm < mg->abstol) {
34           ierr = PetscLogInfo((0,"PCMGMCycle_Private:Linear solver has converged. Residual norm %g is less than absolute tolerance %g\n",rnorm,mg->abstol));CHKERRQ(ierr);
35         } else {
36           ierr = PetscLogInfo((0,"PCMGMCycle_Private:Linear solver has converged. Residual norm %g is less than relative tolerance times initial residual norm %g\n",rnorm,mg->ttol));CHKERRQ(ierr);
37         }
38         PetscFunctionReturn(0);
39       }
40     }
41 
42     mgc = *(mglevels - 1);
43     ierr = MatRestrict(mg->restrct,mg->r,mgc->b);CHKERRQ(ierr);
44     ierr = VecSet(mgc->x,zero);CHKERRQ(ierr);
45     while (cycles--) {
46       ierr = PCMGMCycle_Private(mglevels-1,converged);CHKERRQ(ierr);
47     }
48     ierr = MatInterpolateAdd(mg->interpolate,mgc->x,mg->x,mg->x);CHKERRQ(ierr);
49     if (mg->eventsolve) {ierr = PetscLogEventBegin(mg->eventsolve,0,0,0,0);CHKERRQ(ierr);}
50     ierr = KSPSolve(mg->smoothu,mg->b,mg->x);CHKERRQ(ierr);
51     if (mg->eventsolve) {ierr = PetscLogEventEnd(mg->eventsolve,0,0,0,0);CHKERRQ(ierr);}
52   }
53   PetscFunctionReturn(0);
54 }
55 
56 /*
57        PCMGCreate_Private - Creates a PC_MG structure for use with the
58                multigrid code. Level 0 is the coarsest. (But the
59                finest level is stored first in the array).
60 
61 */
62 #undef __FUNCT__
63 #define __FUNCT__ "PCMGCreate_Private"
64 static PetscErrorCode PCMGCreate_Private(MPI_Comm comm,PetscInt levels,PC pc,MPI_Comm *comms,PC_MG ***result)
65 {
66   PC_MG          **mg;
67   PetscErrorCode ierr;
68   PetscInt       i;
69   PetscMPIInt    size;
70   const char     *prefix;
71   PC             ipc;
72 
73   PetscFunctionBegin;
74   ierr = PetscMalloc(levels*sizeof(PC_MG*),&mg);CHKERRQ(ierr);
75   ierr = PetscLogObjectMemory(pc,levels*(sizeof(PC_MG*)+sizeof(PC_MG)));CHKERRQ(ierr);
76 
77   ierr = PCGetOptionsPrefix(pc,&prefix);CHKERRQ(ierr);
78 
79   for (i=0; i<levels; i++) {
80     ierr = PetscNew(PC_MG,&mg[i]);CHKERRQ(ierr);
81     mg[i]->level           = i;
82     mg[i]->levels          = levels;
83     mg[i]->cycles          = 1;
84     mg[i]->galerkin        = PETSC_FALSE;
85     mg[i]->galerkinused    = PETSC_FALSE;
86     mg[i]->default_smoothu = 1;
87     mg[i]->default_smoothd = 1;
88 
89     if (comms) comm = comms[i];
90     ierr = KSPCreate(comm,&mg[i]->smoothd);CHKERRQ(ierr);
91     ierr = KSPSetTolerances(mg[i]->smoothd,PETSC_DEFAULT,PETSC_DEFAULT,PETSC_DEFAULT, mg[i]->default_smoothd);CHKERRQ(ierr);
92     ierr = KSPSetOptionsPrefix(mg[i]->smoothd,prefix);CHKERRQ(ierr);
93 
94     /* do special stuff for coarse grid */
95     if (!i && levels > 1) {
96       ierr = KSPAppendOptionsPrefix(mg[0]->smoothd,"mg_coarse_");CHKERRQ(ierr);
97 
98       /* coarse solve is (redundant) LU by default */
99       ierr = KSPSetType(mg[0]->smoothd,KSPPREONLY);CHKERRQ(ierr);
100       ierr = KSPGetPC(mg[0]->smoothd,&ipc);CHKERRQ(ierr);
101       ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr);
102       if (size > 1) {
103         ierr = PCSetType(ipc,PCREDUNDANT);CHKERRQ(ierr);
104         ierr = PCRedundantGetPC(ipc,&ipc);CHKERRQ(ierr);
105       }
106       ierr = PCSetType(ipc,PCLU);CHKERRQ(ierr);
107 
108     } else {
109       char tprefix[128];
110       sprintf(tprefix,"mg_levels_%d_",(int)i);
111       ierr = KSPAppendOptionsPrefix(mg[i]->smoothd,tprefix);CHKERRQ(ierr);
112     }
113     ierr = PetscLogObjectParent(pc,mg[i]->smoothd);CHKERRQ(ierr);
114     mg[i]->smoothu         = mg[i]->smoothd;
115     mg[i]->rtol = 0.0;
116     mg[i]->abstol = 0.0;
117     mg[i]->dtol = 0.0;
118     mg[i]->ttol = 0.0;
119     mg[i]->eventsetup = 0;
120     mg[i]->eventsolve = 0;
121   }
122   *result = mg;
123   PetscFunctionReturn(0);
124 }
125 
126 #undef __FUNCT__
127 #define __FUNCT__ "PCDestroy_MG"
128 static PetscErrorCode PCDestroy_MG(PC pc)
129 {
130   PC_MG          **mg = (PC_MG**)pc->data;
131   PetscErrorCode ierr;
132   PetscInt       i,n = mg[0]->levels;
133 
134   PetscFunctionBegin;
135   if (mg[0]->galerkinused) {
136     Mat B;
137     for (i=0; i<n-1; i++) {
138       ierr = KSPGetOperators(mg[i]->smoothd,0,&B,0);CHKERRQ(ierr);
139       ierr = MatDestroy(B);CHKERRQ(ierr);
140     }
141   }
142 
143   for (i=0; i<n-1; i++) {
144     if (mg[i+1]->r) {ierr = VecDestroy(mg[i+1]->r);CHKERRQ(ierr);}
145     if (mg[i]->b) {ierr = VecDestroy(mg[i]->b);CHKERRQ(ierr);}
146     if (mg[i]->x) {ierr = VecDestroy(mg[i]->x);CHKERRQ(ierr);}
147     if (mg[i+1]->restrct) {ierr = MatDestroy(mg[i+1]->restrct);CHKERRQ(ierr);}
148     if (mg[i+1]->interpolate) {ierr = MatDestroy(mg[i+1]->interpolate);CHKERRQ(ierr);}
149   }
150 
151   for (i=0; i<n; i++) {
152     if (mg[i]->smoothd != mg[i]->smoothu) {
153       ierr = KSPDestroy(mg[i]->smoothd);CHKERRQ(ierr);
154     }
155     ierr = KSPDestroy(mg[i]->smoothu);CHKERRQ(ierr);
156     ierr = PetscFree(mg[i]);CHKERRQ(ierr);
157   }
158   ierr = PetscFree(mg);CHKERRQ(ierr);
159   PetscFunctionReturn(0);
160 }
161 
162 
163 
164 EXTERN PetscErrorCode PCMGACycle_Private(PC_MG**);
165 EXTERN PetscErrorCode PCMGFCycle_Private(PC_MG**);
166 EXTERN PetscErrorCode PCMGKCycle_Private(PC_MG**);
167 
168 /*
169    PCApply_MG - Runs either an additive, multiplicative, Kaskadic
170              or full cycle of multigrid.
171 
172   Note:
173   A simple wrapper which calls PCMGMCycle(),PCMGACycle(), or PCMGFCycle().
174 */
175 #undef __FUNCT__
176 #define __FUNCT__ "PCApply_MG"
177 static PetscErrorCode PCApply_MG(PC pc,Vec b,Vec x)
178 {
179   PC_MG          **mg = (PC_MG**)pc->data;
180   PetscScalar    zero = 0.0;
181   PetscErrorCode ierr;
182   PetscInt       levels = mg[0]->levels;
183 
184   PetscFunctionBegin;
185   mg[levels-1]->b = b;
186   mg[levels-1]->x = x;
187   if (!mg[levels-1]->r && mg[0]->am == PC_MG_ADDITIVE) {
188     Vec tvec;
189     ierr = VecDuplicate(mg[levels-1]->b,&tvec);CHKERRQ(ierr);
190     ierr = PCMGSetR(pc,levels-1,tvec);CHKERRQ(ierr);
191     ierr = VecDestroy(tvec);CHKERRQ(ierr);
192   }
193   if (mg[0]->am == PC_MG_MULTIPLICATIVE) {
194     ierr = VecSet(x,zero);CHKERRQ(ierr);
195     ierr = PCMGMCycle_Private(mg+levels-1,PETSC_NULL);CHKERRQ(ierr);
196   }
197   else if (mg[0]->am == PC_MG_ADDITIVE) {
198     ierr = PCMGACycle_Private(mg);CHKERRQ(ierr);
199   }
200   else if (mg[0]->am == PC_MG_KASKADE) {
201     ierr = PCMGKCycle_Private(mg);CHKERRQ(ierr);
202   }
203   else {
204     ierr = PCMGFCycle_Private(mg);CHKERRQ(ierr);
205   }
206   PetscFunctionReturn(0);
207 }
208 
209 #undef __FUNCT__
210 #define __FUNCT__ "PCApplyRichardson_MG"
211 static PetscErrorCode PCApplyRichardson_MG(PC pc,Vec b,Vec x,Vec w,PetscReal rtol,PetscReal abstol, PetscReal dtol,PetscInt its)
212 {
213   PC_MG          **mg = (PC_MG**)pc->data;
214   PetscErrorCode ierr;
215   PetscInt       levels = mg[0]->levels;
216   PetscTruth     converged = PETSC_FALSE;
217 
218   PetscFunctionBegin;
219   mg[levels-1]->b    = b;
220   mg[levels-1]->x    = x;
221 
222   mg[levels-1]->rtol = rtol;
223   mg[levels-1]->abstol = abstol;
224   mg[levels-1]->dtol = dtol;
225   if (rtol) {
226     /* compute initial residual norm for relative convergence test */
227     PetscReal rnorm;
228     ierr               = (*mg[levels-1]->residual)(mg[levels-1]->A,b,x,w);CHKERRQ(ierr);
229     ierr               = VecNorm(w,NORM_2,&rnorm);CHKERRQ(ierr);
230     mg[levels-1]->ttol = PetscMax(rtol*rnorm,abstol);
231   } else if (abstol) {
232     mg[levels-1]->ttol = abstol;
233   } else {
234     mg[levels-1]->ttol = 0.0;
235   }
236 
237   while (its-- && !converged) {
238     ierr = PCMGMCycle_Private(mg+levels-1,&converged);CHKERRQ(ierr);
239   }
240   PetscFunctionReturn(0);
241 }
242 
243 #undef __FUNCT__
244 #define __FUNCT__ "PCSetFromOptions_MG"
245 PetscErrorCode PCSetFromOptions_MG(PC pc)
246 {
247   PetscErrorCode ierr;
248   PetscInt       m,levels = 1;
249   PetscTruth     flg;
250   PC_MG          **mg = (PC_MG**)pc->data;
251   PCMGType       mgtype = mg[0]->am;;
252 
253   PetscFunctionBegin;
254 
255   ierr = PetscOptionsHead("Multigrid options");CHKERRQ(ierr);
256     if (!pc->data) {
257       ierr = PetscOptionsInt("-pc_mg_levels","Number of Levels","PCMGSetLevels",levels,&levels,&flg);CHKERRQ(ierr);
258       ierr = PCMGSetLevels(pc,levels,PETSC_NULL);CHKERRQ(ierr);
259     }
260     ierr = PetscOptionsInt("-pc_mg_cycles","1 for V cycle, 2 for W-cycle","PCMGSetCycles",1,&m,&flg);CHKERRQ(ierr);
261     if (flg) {
262       ierr = PCMGSetCycles(pc,m);CHKERRQ(ierr);
263     }
264     ierr = PetscOptionsName("-pc_mg_galerkin","Use Galerkin process to compute coarser operators","PCMGSetGalerkin",&flg);CHKERRQ(ierr);
265     if (flg) {
266       ierr = PCMGSetGalerkin(pc);CHKERRQ(ierr);
267     }
268     ierr = PetscOptionsInt("-pc_mg_smoothup","Number of post-smoothing steps","PCMGSetNumberSmoothUp",1,&m,&flg);CHKERRQ(ierr);
269     if (flg) {
270       ierr = PCMGSetNumberSmoothUp(pc,m);CHKERRQ(ierr);
271     }
272     ierr = PetscOptionsInt("-pc_mg_smoothdown","Number of pre-smoothing steps","PCMGSetNumberSmoothDown",1,&m,&flg);CHKERRQ(ierr);
273     if (flg) {
274       ierr = PCMGSetNumberSmoothDown(pc,m);CHKERRQ(ierr);
275     }
276     ierr = PetscOptionsEnum("-pc_mg_type","Multigrid type","PCMGSetType",PCMGTypes,(PetscEnum)mgtype,(PetscEnum*)&mgtype,&flg);CHKERRQ(ierr);
277     if (flg) {ierr = PCMGSetType(pc,mgtype);CHKERRQ(ierr);}
278     ierr = PetscOptionsName("-pc_mg_log","Log times for each multigrid level","None",&flg);CHKERRQ(ierr);
279     if (flg) {
280       PetscInt i;
281       char     eventname[128];
282       if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
283       levels = mg[0]->levels;
284       for (i=0; i<levels; i++) {
285         sprintf(eventname,"MSetup Level %d",(int)i);
286         ierr = PetscLogEventRegister(&mg[i]->eventsetup,eventname,pc->cookie);CHKERRQ(ierr);
287         sprintf(eventname,"MGSolve Level %d to 0",(int)i);
288         ierr = PetscLogEventRegister(&mg[i]->eventsolve,eventname,pc->cookie);CHKERRQ(ierr);
289       }
290     }
291   ierr = PetscOptionsTail();CHKERRQ(ierr);
292   PetscFunctionReturn(0);
293 }
294 
295 const char *PCMGTypes[] = {"MULTIPLICATIVE","ADDITIVE","FULL","KASKADE","PCMGType","PC_MG",0};
296 
297 #undef __FUNCT__
298 #define __FUNCT__ "PCView_MG"
299 static PetscErrorCode PCView_MG(PC pc,PetscViewer viewer)
300 {
301   PC_MG          **mg = (PC_MG**)pc->data;
302   PetscErrorCode ierr;
303   PetscInt       levels = mg[0]->levels,i;
304   PetscTruth     iascii;
305 
306   PetscFunctionBegin;
307   ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr);
308   if (iascii) {
309     ierr = PetscViewerASCIIPrintf(viewer,"  MG: type is %s, levels=%D cycles=%D, pre-smooths=%D, post-smooths=%D\n",
310                       PCMGTypes[mg[0]->am],levels,mg[0]->cycles,mg[0]->default_smoothd,mg[0]->default_smoothu);CHKERRQ(ierr);
311     if (mg[0]->galerkin) {
312       ierr = PetscViewerASCIIPrintf(viewer,"    Using Galerkin computed coarse grid matrices\n");CHKERRQ(ierr);
313     }
314     for (i=0; i<levels; i++) {
315       if (!i) {
316         ierr = PetscViewerASCIIPrintf(viewer,"Coarse gride solver -- level %D -------------------------------\n",i);CHKERRQ(ierr);
317       } else {
318         ierr = PetscViewerASCIIPrintf(viewer,"Down solver (pre-smoother) on level %D -------------------------------\n",i);CHKERRQ(ierr);
319       }
320       ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr);
321       ierr = KSPView(mg[i]->smoothd,viewer);CHKERRQ(ierr);
322       ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr);
323       if (i && mg[i]->smoothd == mg[i]->smoothu) {
324         ierr = PetscViewerASCIIPrintf(viewer,"Up solver (post-smoother) same as down solver (pre-smoother)\n");CHKERRQ(ierr);
325       } else if (i){
326         ierr = PetscViewerASCIIPrintf(viewer,"Up solver (post-smoother) on level %D -------------------------------\n",i);CHKERRQ(ierr);
327         ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr);
328         ierr = KSPView(mg[i]->smoothu,viewer);CHKERRQ(ierr);
329         ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr);
330       }
331     }
332   } else {
333     SETERRQ1(PETSC_ERR_SUP,"Viewer type %s not supported for PCMG",((PetscObject)viewer)->type_name);
334   }
335   PetscFunctionReturn(0);
336 }
337 
338 /*
339     Calls setup for the KSP on each level
340 */
341 #undef __FUNCT__
342 #define __FUNCT__ "PCSetUp_MG"
343 static PetscErrorCode PCSetUp_MG(PC pc)
344 {
345   PC_MG          **mg = (PC_MG**)pc->data;
346   PetscErrorCode ierr;
347   PetscInt       i,n = mg[0]->levels;
348   PC             cpc;
349   PetscTruth     preonly,lu,redundant,cholesky,monitor = PETSC_FALSE,dump;
350   PetscViewer    ascii;
351   MPI_Comm       comm;
352   Mat            dA,dB;
353   MatStructure   uflag;
354   Vec            tvec;
355 
356   PetscFunctionBegin;
357   if (!pc->setupcalled) {
358     ierr = PetscOptionsHasName(0,"-pc_mg_monitor",&monitor);CHKERRQ(ierr);
359 
360     for (i=0; i<n; i++) {
361       if (monitor) {
362         ierr = PetscObjectGetComm((PetscObject)mg[i]->smoothd,&comm);CHKERRQ(ierr);
363         ierr = PetscViewerASCIIOpen(comm,"stdout",&ascii);CHKERRQ(ierr);
364         ierr = PetscViewerASCIISetTab(ascii,n-i);CHKERRQ(ierr);
365         ierr = KSPSetMonitor(mg[i]->smoothd,KSPDefaultMonitor,ascii,(PetscErrorCode(*)(void*))PetscViewerDestroy);CHKERRQ(ierr);
366       }
367       ierr = KSPSetFromOptions(mg[i]->smoothd);CHKERRQ(ierr);
368     }
369     for (i=1; i<n; i++) {
370       if (mg[i]->smoothu && mg[i]->smoothu != mg[i]->smoothd) {
371         if (monitor) {
372           ierr = PetscObjectGetComm((PetscObject)mg[i]->smoothu,&comm);CHKERRQ(ierr);
373           ierr = PetscViewerASCIIOpen(comm,"stdout",&ascii);CHKERRQ(ierr);
374           ierr = PetscViewerASCIISetTab(ascii,n-i);CHKERRQ(ierr);
375           ierr = KSPSetMonitor(mg[i]->smoothu,KSPDefaultMonitor,ascii,(PetscErrorCode(*)(void*))PetscViewerDestroy);CHKERRQ(ierr);
376         }
377         ierr = KSPSetFromOptions(mg[i]->smoothu);CHKERRQ(ierr);
378       }
379     }
380     for (i=1; i<n; i++) {
381       if (mg[i]->restrct && !mg[i]->interpolate) {
382         ierr = PCMGSetInterpolate(pc,i,mg[i]->restrct);CHKERRQ(ierr);
383       }
384       if (!mg[i]->restrct && mg[i]->interpolate) {
385         ierr = PCMGSetRestriction(pc,i,mg[i]->interpolate);CHKERRQ(ierr);
386       }
387 #if defined(PETSC_USE_DEBUG)
388       if (!mg[i]->restrct || !mg[i]->interpolate) {
389         SETERRQ1(PETSC_ERR_ARG_WRONGSTATE,"Need to set restriction or interpolation on level %d",(int)i);
390       }
391 #endif
392     }
393     for (i=0; i<n-1; i++) {
394       if (!mg[i]->r && i) {
395         ierr = VecDuplicate(mg[i]->b,&tvec);CHKERRQ(ierr);
396         ierr = PCMGSetR(pc,i,tvec);CHKERRQ(ierr);
397         ierr = VecDestroy(tvec);CHKERRQ(ierr);
398       }
399       if (!mg[i]->x) {
400         ierr = VecDuplicate(mg[i]->b,&tvec);CHKERRQ(ierr);
401         ierr = PCMGSetX(pc,i,tvec);CHKERRQ(ierr);
402         ierr = VecDestroy(tvec);CHKERRQ(ierr);
403       }
404     }
405   }
406 
407   /* If user did not provide fine grid operators, use those from PC */
408   /* BUG BUG BUG This will work ONLY the first time called: hence if the user changes
409      the PC matrices between solves PCMG will continue to use first set provided */
410   ierr = KSPGetOperators(mg[n-1]->smoothd,&dA,&dB,&uflag);CHKERRQ(ierr);
411   if (!dA  && !dB) {
412     ierr = PetscLogInfo((pc,"PCSetUp_MG: Using outer operators to define finest grid operator \n  because PCMGGetSmoother(pc,nlevels-1,&ksp);KSPSetOperators(ksp,...); was not called.\n"));
413     ierr = KSPSetOperators(mg[n-1]->smoothd,pc->mat,pc->pmat,uflag);CHKERRQ(ierr);
414   }
415 
416   if (mg[0]->galerkin) {
417     Mat B;
418     mg[0]->galerkinused = PETSC_TRUE;
419     /* currently only handle case where mat and pmat are the same on coarser levels */
420     ierr = KSPGetOperators(mg[n-1]->smoothd,&dA,&dB,&uflag);CHKERRQ(ierr);
421     if (!pc->setupcalled) {
422       for (i=n-2; i>-1; i--) {
423         ierr = MatPtAP(dB,mg[i+1]->interpolate,MAT_INITIAL_MATRIX,1.0,&B);CHKERRQ(ierr);
424         ierr = KSPSetOperators(mg[i]->smoothd,B,B,uflag);CHKERRQ(ierr);
425         dB   = B;
426       }
427     } else {
428       for (i=n-2; i>-1; i--) {
429         ierr = KSPGetOperators(mg[i]->smoothd,0,&B,0);CHKERRQ(ierr);
430         ierr = MatPtAP(dB,mg[i]->interpolate,MAT_REUSE_MATRIX,1.0,&B);CHKERRQ(ierr);
431         ierr = KSPSetOperators(mg[i]->smoothd,B,B,uflag);CHKERRQ(ierr);
432         dB   = B;
433       }
434     }
435   }
436 
437   for (i=1; i<n; i++) {
438     if (mg[i]->smoothu == mg[i]->smoothd) {
439       /* if doing only down then initial guess is zero */
440       ierr = KSPSetInitialGuessNonzero(mg[i]->smoothd,PETSC_TRUE);CHKERRQ(ierr);
441     }
442     if (mg[i]->eventsetup) {ierr = PetscLogEventBegin(mg[i]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
443     ierr = KSPSetUp(mg[i]->smoothd);CHKERRQ(ierr);
444     if (mg[i]->eventsetup) {ierr = PetscLogEventEnd(mg[i]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
445   }
446   for (i=1; i<n; i++) {
447     if (mg[i]->smoothu && mg[i]->smoothu != mg[i]->smoothd) {
448       PC           uppc,downpc;
449       Mat          downmat,downpmat,upmat,uppmat;
450       MatStructure matflag;
451 
452       /* check if operators have been set for up, if not use down operators to set them */
453       ierr = KSPGetPC(mg[i]->smoothu,&uppc);CHKERRQ(ierr);
454       ierr = PCGetOperators(uppc,&upmat,&uppmat,PETSC_NULL);CHKERRQ(ierr);
455       if (!upmat) {
456         ierr = KSPGetPC(mg[i]->smoothd,&downpc);CHKERRQ(ierr);
457         ierr = PCGetOperators(downpc,&downmat,&downpmat,&matflag);CHKERRQ(ierr);
458         ierr = KSPSetOperators(mg[i]->smoothu,downmat,downpmat,matflag);CHKERRQ(ierr);
459       }
460 
461       ierr = KSPSetInitialGuessNonzero(mg[i]->smoothu,PETSC_TRUE);CHKERRQ(ierr);
462       if (mg[i]->eventsetup) {ierr = PetscLogEventBegin(mg[i]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
463       ierr = KSPSetUp(mg[i]->smoothu);CHKERRQ(ierr);
464       if (mg[i]->eventsetup) {ierr = PetscLogEventEnd(mg[i]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
465     }
466   }
467 
468   /*
469       If coarse solver is not direct method then DO NOT USE preonly
470   */
471   ierr = PetscTypeCompare((PetscObject)mg[0]->smoothd,KSPPREONLY,&preonly);CHKERRQ(ierr);
472   if (preonly) {
473     ierr = KSPGetPC(mg[0]->smoothd,&cpc);CHKERRQ(ierr);
474     ierr = PetscTypeCompare((PetscObject)cpc,PCLU,&lu);CHKERRQ(ierr);
475     ierr = PetscTypeCompare((PetscObject)cpc,PCREDUNDANT,&redundant);CHKERRQ(ierr);
476     ierr = PetscTypeCompare((PetscObject)cpc,PCCHOLESKY,&cholesky);CHKERRQ(ierr);
477     if (!lu && !redundant && !cholesky) {
478       ierr = KSPSetType(mg[0]->smoothd,KSPGMRES);CHKERRQ(ierr);
479     }
480   }
481 
482   if (!pc->setupcalled) {
483     if (monitor) {
484       ierr = PetscObjectGetComm((PetscObject)mg[0]->smoothd,&comm);CHKERRQ(ierr);
485       ierr = PetscViewerASCIIOpen(comm,"stdout",&ascii);CHKERRQ(ierr);
486       ierr = PetscViewerASCIISetTab(ascii,n);CHKERRQ(ierr);
487       ierr = KSPSetMonitor(mg[0]->smoothd,KSPDefaultMonitor,ascii,(PetscErrorCode(*)(void*))PetscViewerDestroy);CHKERRQ(ierr);
488     }
489     ierr = KSPSetFromOptions(mg[0]->smoothd);CHKERRQ(ierr);
490   }
491 
492   if (mg[0]->eventsetup) {ierr = PetscLogEventBegin(mg[0]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
493   ierr = KSPSetUp(mg[0]->smoothd);CHKERRQ(ierr);
494   if (mg[0]->eventsetup) {ierr = PetscLogEventEnd(mg[0]->eventsetup,0,0,0,0);CHKERRQ(ierr);}
495 
496 #if defined(PETSC_USE_SOCKET_VIEWER)
497   /*
498      Dump the interpolation/restriction matrices to matlab plus the
499    Jacobian/stiffness on each level. This allows Matlab users to
500    easily check if the Galerkin condition A_c = R A_f R^T is satisfied */
501   ierr = PetscOptionsHasName(pc->prefix,"-pc_mg_dump_matlab",&dump);CHKERRQ(ierr);
502   if (dump) {
503     for (i=1; i<n; i++) {
504       ierr = MatView(mg[i]->restrct,PETSC_VIEWER_SOCKET_(pc->comm));CHKERRQ(ierr);
505     }
506     for (i=0; i<n; i++) {
507       ierr = KSPGetPC(mg[i]->smoothd,&pc);CHKERRQ(ierr);
508       ierr = MatView(pc->mat,PETSC_VIEWER_SOCKET_(pc->comm));CHKERRQ(ierr);
509     }
510   }
511 #endif
512 
513   ierr = PetscOptionsHasName(pc->prefix,"-pc_mg_dump_binary",&dump);CHKERRQ(ierr);
514   if (dump) {
515     for (i=1; i<n; i++) {
516       ierr = MatView(mg[i]->restrct,PETSC_VIEWER_BINARY_(pc->comm));CHKERRQ(ierr);
517     }
518     for (i=0; i<n; i++) {
519       ierr = KSPGetPC(mg[i]->smoothd,&pc);CHKERRQ(ierr);
520       ierr = MatView(pc->mat,PETSC_VIEWER_BINARY_(pc->comm));CHKERRQ(ierr);
521     }
522   }
523   PetscFunctionReturn(0);
524 }
525 
526 /* -------------------------------------------------------------------------------------*/
527 
528 #undef __FUNCT__
529 #define __FUNCT__ "PCMGSetLevels"
530 /*@C
531    PCMGSetLevels - Sets the number of levels to use with MG.
532    Must be called before any other MG routine.
533 
534    Collective on PC
535 
536    Input Parameters:
537 +  pc - the preconditioner context
538 .  levels - the number of levels
539 -  comms - optional communicators for each level; this is to allow solving the coarser problems
540            on smaller sets of processors. Use PETSC_NULL_OBJECT for default in Fortran
541 
542    Level: intermediate
543 
544    Notes:
545      If the number of levels is one then the multigrid uses the -mg_levels prefix
546   for setting the level options rather than the -mg_coarse prefix.
547 
548 .keywords: MG, set, levels, multigrid
549 
550 .seealso: PCMGSetType(), PCMGGetLevels()
551 @*/
552 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetLevels(PC pc,PetscInt levels,MPI_Comm *comms)
553 {
554   PetscErrorCode ierr;
555   PC_MG          **mg;
556 
557   PetscFunctionBegin;
558   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
559 
560   if (pc->data) {
561     SETERRQ(PETSC_ERR_ORDER,"Number levels already set for MG\n\
562     make sure that you call PCMGSetLevels() before KSPSetFromOptions()");
563   }
564   ierr                     = PCMGCreate_Private(pc->comm,levels,pc,comms,&mg);CHKERRQ(ierr);
565   mg[0]->am                = PC_MG_MULTIPLICATIVE;
566   pc->data                 = (void*)mg;
567   pc->ops->applyrichardson = PCApplyRichardson_MG;
568   PetscFunctionReturn(0);
569 }
570 
571 #undef __FUNCT__
572 #define __FUNCT__ "PCMGGetLevels"
573 /*@
574    PCMGGetLevels - Gets the number of levels to use with MG.
575 
576    Not Collective
577 
578    Input Parameter:
579 .  pc - the preconditioner context
580 
581    Output parameter:
582 .  levels - the number of levels
583 
584    Level: advanced
585 
586 .keywords: MG, get, levels, multigrid
587 
588 .seealso: PCMGSetLevels()
589 @*/
590 PetscErrorCode PETSCKSP_DLLEXPORT PCMGGetLevels(PC pc,PetscInt *levels)
591 {
592   PC_MG  **mg;
593 
594   PetscFunctionBegin;
595   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
596   PetscValidIntPointer(levels,2);
597 
598   mg      = (PC_MG**)pc->data;
599   *levels = mg[0]->levels;
600   PetscFunctionReturn(0);
601 }
602 
603 #undef __FUNCT__
604 #define __FUNCT__ "PCMGSetType"
605 /*@
606    PCMGSetType - Determines the form of multigrid to use:
607    multiplicative, additive, full, or the Kaskade algorithm.
608 
609    Collective on PC
610 
611    Input Parameters:
612 +  pc - the preconditioner context
613 -  form - multigrid form, one of PC_MG_MULTIPLICATIVE, PC_MG_ADDITIVE,
614    PC_MG_FULL, PC_MG_KASKADE
615 
616    Options Database Key:
617 .  -pc_mg_type <form> - Sets <form>, one of multiplicative,
618    additive, full, kaskade
619 
620    Level: advanced
621 
622 .keywords: MG, set, method, multiplicative, additive, full, Kaskade, multigrid
623 
624 .seealso: PCMGSetLevels()
625 @*/
626 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetType(PC pc,PCMGType form)
627 {
628   PC_MG **mg;
629 
630   PetscFunctionBegin;
631   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
632   mg = (PC_MG**)pc->data;
633 
634   if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
635   mg[0]->am = form;
636   if (form == PC_MG_MULTIPLICATIVE) pc->ops->applyrichardson = PCApplyRichardson_MG;
637   else pc->ops->applyrichardson = 0;
638   PetscFunctionReturn(0);
639 }
640 
641 #undef __FUNCT__
642 #define __FUNCT__ "PCMGSetCycles"
643 /*@
644    PCMGSetCycles - Sets the type cycles to use.  Use PCMGSetCyclesOnLevel() for more
645    complicated cycling.
646 
647    Collective on PC
648 
649    Input Parameters:
650 +  pc - the multigrid context
651 -  n - the number of cycles
652 
653    Options Database Key:
654 $  -pc_mg_cycles n - 1 denotes a V-cycle; 2 denotes a W-cycle.
655 
656    Level: advanced
657 
658 .keywords: MG, set, cycles, V-cycle, W-cycle, multigrid
659 
660 .seealso: PCMGSetCyclesOnLevel()
661 @*/
662 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetCycles(PC pc,PetscInt n)
663 {
664   PC_MG    **mg;
665   PetscInt i,levels;
666 
667   PetscFunctionBegin;
668   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
669   mg     = (PC_MG**)pc->data;
670   if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
671   levels = mg[0]->levels;
672 
673   for (i=0; i<levels; i++) {
674     mg[i]->cycles  = n;
675   }
676   PetscFunctionReturn(0);
677 }
678 
679 #undef __FUNCT__
680 #define __FUNCT__ "PCMGSetGalerkin"
681 /*@
682    PCMGSetGalerkin - Causes the coarser grid matrices to be computed from the
683       finest grid via the Galerkin process: A_i-1 = r_i * A_i * r_i^t
684 
685    Collective on PC
686 
687    Input Parameters:
688 +  pc - the multigrid context
689 -  n - the number of cycles
690 
691    Options Database Key:
692 $  -pc_mg_galerkin
693 
694    Level: intermediate
695 
696 .keywords: MG, set, Galerkin
697 
698 @*/
699 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetGalerkin(PC pc)
700 {
701   PC_MG    **mg;
702   PetscInt i,levels;
703 
704   PetscFunctionBegin;
705   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
706   mg     = (PC_MG**)pc->data;
707   if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
708   levels = mg[0]->levels;
709 
710   for (i=0; i<levels; i++) {
711     mg[i]->galerkin = PETSC_TRUE;
712   }
713   PetscFunctionReturn(0);
714 }
715 
716 #undef __FUNCT__
717 #define __FUNCT__ "PCMGSetNumberSmoothDown"
718 /*@
719    PCMGSetNumberSmoothDown - Sets the number of pre-smoothing steps to
720    use on all levels. Use PCMGGetSmootherDown() to set different
721    pre-smoothing steps on different levels.
722 
723    Collective on PC
724 
725    Input Parameters:
726 +  mg - the multigrid context
727 -  n - the number of smoothing steps
728 
729    Options Database Key:
730 .  -pc_mg_smoothdown <n> - Sets number of pre-smoothing steps
731 
732    Level: advanced
733 
734 .keywords: MG, smooth, down, pre-smoothing, steps, multigrid
735 
736 .seealso: PCMGSetNumberSmoothUp()
737 @*/
738 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetNumberSmoothDown(PC pc,PetscInt n)
739 {
740   PC_MG          **mg;
741   PetscErrorCode ierr;
742   PetscInt       i,levels;
743 
744   PetscFunctionBegin;
745   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
746   mg     = (PC_MG**)pc->data;
747   if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
748   levels = mg[0]->levels;
749 
750   for (i=0; i<levels; i++) {
751     /* make sure smoother up and down are different */
752     ierr = PCMGGetSmootherUp(pc,i,PETSC_NULL);CHKERRQ(ierr);
753     ierr = KSPSetTolerances(mg[i]->smoothd,PETSC_DEFAULT,PETSC_DEFAULT,PETSC_DEFAULT,n);CHKERRQ(ierr);
754     mg[i]->default_smoothd = n;
755   }
756   PetscFunctionReturn(0);
757 }
758 
759 #undef __FUNCT__
760 #define __FUNCT__ "PCMGSetNumberSmoothUp"
761 /*@
762    PCMGSetNumberSmoothUp - Sets the number of post-smoothing steps to use
763    on all levels. Use PCMGGetSmootherUp() to set different numbers of
764    post-smoothing steps on different levels.
765 
766    Collective on PC
767 
768    Input Parameters:
769 +  mg - the multigrid context
770 -  n - the number of smoothing steps
771 
772    Options Database Key:
773 .  -pc_mg_smoothup <n> - Sets number of post-smoothing steps
774 
775    Level: advanced
776 
777    Note: this does not set a value on the coarsest grid, since we assume that
778     there is no seperate smooth up on the coarsest grid.
779 
780 .keywords: MG, smooth, up, post-smoothing, steps, multigrid
781 
782 .seealso: PCMGSetNumberSmoothDown()
783 @*/
784 PetscErrorCode PETSCKSP_DLLEXPORT PCMGSetNumberSmoothUp(PC pc,PetscInt n)
785 {
786   PC_MG          **mg;
787   PetscErrorCode ierr;
788   PetscInt       i,levels;
789 
790   PetscFunctionBegin;
791   PetscValidHeaderSpecific(pc,PC_COOKIE,1);
792   mg     = (PC_MG**)pc->data;
793   if (!mg) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Must set MG levels before calling");
794   levels = mg[0]->levels;
795 
796   for (i=1; i<levels; i++) {
797     /* make sure smoother up and down are different */
798     ierr = PCMGGetSmootherUp(pc,i,PETSC_NULL);CHKERRQ(ierr);
799     ierr = KSPSetTolerances(mg[i]->smoothu,PETSC_DEFAULT,PETSC_DEFAULT,PETSC_DEFAULT,n);CHKERRQ(ierr);
800     mg[i]->default_smoothu = n;
801   }
802   PetscFunctionReturn(0);
803 }
804 
805 /* ----------------------------------------------------------------------------------------*/
806 
807 /*MC
808    PCMG - Use geometric multigrid preconditioning. This preconditioner requires you provide additional
809     information about the coarser grid matrices and restriction/interpolation operators.
810 
811    Options Database Keys:
812 +  -pc_mg_levels <nlevels> - number of levels including finest
813 .  -pc_mg_cycles 1 or 2 - for V or W-cycle
814 .  -pc_mg_smoothup <n> - number of smoothing steps after interpolation
815 .  -pc_mg_smoothdown <n> - number of smoothing steps before applying restriction operator
816 .  -pc_mg_type <additive,multiplicative,full,cascade> - multiplicative is the default
817 .  -pc_mg_log - log information about time spent on each level of the solver
818 .  -pc_mg_monitor - print information on the multigrid convergence
819 .  -pc_mg_galerkin - use Galerkin process to compute coarser operators
820 -  -pc_mg_dump_matlab - dumps the matrices for each level and the restriction/interpolation matrices
821                         to the Socket viewer for reading from Matlab.
822 
823    Notes:
824 
825    Level: intermediate
826 
827    Concepts: multigrid
828 
829 .seealso:  PCCreate(), PCSetType(), PCType (for list of available types), PC, PCMGType,
830            PCMGSetLevels(), PCMGGetLevels(), PCMGSetType(), PCMGSetCycles(), PCMGSetNumberSmoothDown(),
831            PCMGSetNumberSmoothUp(), PCMGGetCoarseSolve(), PCMGSetResidual(), PCMGSetInterpolation(),
832            PCMGSetRestriction(), PCMGGetSmoother(), PCMGGetSmootherUp(), PCMGGetSmootherDown(),
833            PCMGSetCyclesOnLevel(), PCMGSetRhs(), PCMGSetX(), PCMGSetR()
834 M*/
835 
836 EXTERN_C_BEGIN
837 #undef __FUNCT__
838 #define __FUNCT__ "PCCreate_MG"
839 PetscErrorCode PETSCKSP_DLLEXPORT PCCreate_MG(PC pc)
840 {
841   PetscFunctionBegin;
842   pc->ops->apply          = PCApply_MG;
843   pc->ops->setup          = PCSetUp_MG;
844   pc->ops->destroy        = PCDestroy_MG;
845   pc->ops->setfromoptions = PCSetFromOptions_MG;
846   pc->ops->view           = PCView_MG;
847 
848   pc->data                = (void*)0;
849   PetscFunctionReturn(0);
850 }
851 EXTERN_C_END
852