Actual source code: mumps.c
petsc-3.9.4 2018-09-11
2: /*
3: Provides an interface to the MUMPS sparse solver
4: */
6: #include <../src/mat/impls/aij/mpi/mpiaij.h>
7: #include <../src/mat/impls/sbaij/mpi/mpisbaij.h>
8: #include <../src/mat/impls/sell/mpi/mpisell.h>
10: EXTERN_C_BEGIN
11: #if defined(PETSC_USE_COMPLEX)
12: #if defined(PETSC_USE_REAL_SINGLE)
13: #include <cmumps_c.h>
14: #else
15: #include <zmumps_c.h>
16: #endif
17: #else
18: #if defined(PETSC_USE_REAL_SINGLE)
19: #include <smumps_c.h>
20: #else
21: #include <dmumps_c.h>
22: #endif
23: #endif
24: EXTERN_C_END
25: #define JOB_INIT -1
26: #define JOB_FACTSYMBOLIC 1
27: #define JOB_FACTNUMERIC 2
28: #define JOB_SOLVE 3
29: #define JOB_END -2
31: /* calls to MUMPS */
32: #if defined(PETSC_USE_COMPLEX)
33: #if defined(PETSC_USE_REAL_SINGLE)
34: #define PetscMUMPS_c cmumps_c
35: #else
36: #define PetscMUMPS_c zmumps_c
37: #endif
38: #else
39: #if defined(PETSC_USE_REAL_SINGLE)
40: #define PetscMUMPS_c smumps_c
41: #else
42: #define PetscMUMPS_c dmumps_c
43: #endif
44: #endif
46: /* declare MumpsScalar */
47: #if defined(PETSC_USE_COMPLEX)
48: #if defined(PETSC_USE_REAL_SINGLE)
49: #define MumpsScalar mumps_complex
50: #else
51: #define MumpsScalar mumps_double_complex
52: #endif
53: #else
54: #define MumpsScalar PetscScalar
55: #endif
57: /* macros s.t. indices match MUMPS documentation */
58: #define ICNTL(I) icntl[(I)-1]
59: #define CNTL(I) cntl[(I)-1]
60: #define INFOG(I) infog[(I)-1]
61: #define INFO(I) info[(I)-1]
62: #define RINFOG(I) rinfog[(I)-1]
63: #define RINFO(I) rinfo[(I)-1]
65: typedef struct {
66: #if defined(PETSC_USE_COMPLEX)
67: #if defined(PETSC_USE_REAL_SINGLE)
68: CMUMPS_STRUC_C id;
69: #else
70: ZMUMPS_STRUC_C id;
71: #endif
72: #else
73: #if defined(PETSC_USE_REAL_SINGLE)
74: SMUMPS_STRUC_C id;
75: #else
76: DMUMPS_STRUC_C id;
77: #endif
78: #endif
80: MatStructure matstruc;
81: PetscMPIInt myid,size;
82: PetscInt *irn,*jcn,nz,sym;
83: PetscScalar *val;
84: MPI_Comm comm_mumps;
85: PetscInt ICNTL9_pre; /* check if ICNTL(9) is changed from previous MatSolve */
86: VecScatter scat_rhs, scat_sol; /* used by MatSolve() */
87: Vec b_seq,x_seq;
88: PetscInt ninfo,*info; /* display INFO */
89: PetscInt sizeredrhs;
90: PetscScalar *schur_sol;
91: PetscInt schur_sizesol;
93: PetscErrorCode (*ConvertToTriples)(Mat, int, MatReuse, int*, int**, int**, PetscScalar**);
94: } Mat_MUMPS;
96: extern PetscErrorCode MatDuplicate_MUMPS(Mat,MatDuplicateOption,Mat*);
98: static PetscErrorCode MatMumpsResetSchur_Private(Mat_MUMPS* mumps)
99: {
103: PetscFree2(mumps->id.listvar_schur,mumps->id.schur);
104: PetscFree(mumps->id.redrhs);
105: PetscFree(mumps->schur_sol);
106: mumps->id.size_schur = 0;
107: mumps->id.schur_lld = 0;
108: mumps->id.ICNTL(19) = 0;
109: return(0);
110: }
112: /* solve with rhs in mumps->id.redrhs and return in the same location */
113: static PetscErrorCode MatMumpsSolveSchur_Private(Mat F)
114: {
115: Mat_MUMPS *mumps=(Mat_MUMPS*)F->data;
116: Mat S,B,X;
117: MatFactorSchurStatus schurstatus;
118: PetscInt sizesol;
119: PetscErrorCode ierr;
122: MatFactorFactorizeSchurComplement(F);
123: MatFactorGetSchurComplement(F,&S,&schurstatus);
124: MatCreateSeqDense(PETSC_COMM_SELF,mumps->id.size_schur,mumps->id.nrhs,(PetscScalar*)mumps->id.redrhs,&B);
125: switch (schurstatus) {
126: case MAT_FACTOR_SCHUR_FACTORED:
127: MatCreateSeqDense(PETSC_COMM_SELF,mumps->id.size_schur,mumps->id.nrhs,(PetscScalar*)mumps->id.redrhs,&X);
128: if (!mumps->id.ICNTL(9)) { /* transpose solve */
129: MatMatSolveTranspose(S,B,X);
130: } else {
131: MatMatSolve(S,B,X);
132: }
133: break;
134: case MAT_FACTOR_SCHUR_INVERTED:
135: sizesol = mumps->id.nrhs*mumps->id.size_schur;
136: if (!mumps->schur_sol || sizesol > mumps->schur_sizesol) {
137: PetscFree(mumps->schur_sol);
138: PetscMalloc1(sizesol,&mumps->schur_sol);
139: mumps->schur_sizesol = sizesol;
140: }
141: MatCreateSeqDense(PETSC_COMM_SELF,mumps->id.size_schur,mumps->id.nrhs,mumps->schur_sol,&X);
142: if (!mumps->id.ICNTL(9)) { /* transpose solve */
143: MatTransposeMatMult(S,B,MAT_REUSE_MATRIX,PETSC_DEFAULT,&X);
144: } else {
145: MatMatMult(S,B,MAT_REUSE_MATRIX,PETSC_DEFAULT,&X);
146: }
147: MatCopy(X,B,SAME_NONZERO_PATTERN);
148: break;
149: default:
150: SETERRQ1(PetscObjectComm((PetscObject)F),PETSC_ERR_SUP,"Unhandled MatFactorSchurStatus %D",F->schur_status);
151: break;
152: }
153: MatFactorRestoreSchurComplement(F,&S,schurstatus);
154: MatDestroy(&B);
155: MatDestroy(&X);
156: return(0);
157: }
159: static PetscErrorCode MatMumpsHandleSchur_Private(Mat F, PetscBool expansion)
160: {
161: Mat_MUMPS *mumps=(Mat_MUMPS*)F->data;
165: if (!mumps->id.ICNTL(19)) { /* do nothing when Schur complement has not been computed */
166: return(0);
167: }
168: if (!expansion) { /* prepare for the condensation step */
169: PetscInt sizeredrhs = mumps->id.nrhs*mumps->id.size_schur;
170: /* allocate MUMPS internal array to store reduced right-hand sides */
171: if (!mumps->id.redrhs || sizeredrhs > mumps->sizeredrhs) {
172: PetscFree(mumps->id.redrhs);
173: mumps->id.lredrhs = mumps->id.size_schur;
174: PetscMalloc1(mumps->id.nrhs*mumps->id.lredrhs,&mumps->id.redrhs);
175: mumps->sizeredrhs = mumps->id.nrhs*mumps->id.lredrhs;
176: }
177: mumps->id.ICNTL(26) = 1; /* condensation phase */
178: } else { /* prepare for the expansion step */
179: /* solve Schur complement (this has to be done by the MUMPS user, so basically us) */
180: MatMumpsSolveSchur_Private(F);
181: mumps->id.ICNTL(26) = 2; /* expansion phase */
182: PetscMUMPS_c(&mumps->id);
183: if (mumps->id.INFOG(1) < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in solve phase: INFOG(1)=%d\n",mumps->id.INFOG(1));
184: /* restore defaults */
185: mumps->id.ICNTL(26) = -1;
186: /* free MUMPS internal array for redrhs if we have solved for multiple rhs in order to save memory space */
187: if (mumps->id.nrhs > 1) {
188: PetscFree(mumps->id.redrhs);
189: mumps->id.lredrhs = 0;
190: mumps->sizeredrhs = 0;
191: }
192: }
193: return(0);
194: }
196: /*
197: MatConvertToTriples_A_B - convert Petsc matrix to triples: row[nz], col[nz], val[nz]
199: input:
200: A - matrix in aij,baij or sbaij (bs=1) format
201: shift - 0: C style output triple; 1: Fortran style output triple.
202: reuse - MAT_INITIAL_MATRIX: spaces are allocated and values are set for the triple
203: MAT_REUSE_MATRIX: only the values in v array are updated
204: output:
205: nnz - dim of r, c, and v (number of local nonzero entries of A)
206: r, c, v - row and col index, matrix values (matrix triples)
208: The returned values r, c, and sometimes v are obtained in a single PetscMalloc(). Then in MatDestroy_MUMPS() it is
209: freed with PetscFree(mumps->irn); This is not ideal code, the fact that v is ONLY sometimes part of mumps->irn means
210: that the PetscMalloc() cannot easily be replaced with a PetscMalloc3().
212: */
214: PetscErrorCode MatConvertToTriples_seqaij_seqaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
215: {
216: const PetscInt *ai,*aj,*ajj,M=A->rmap->n;
217: PetscInt nz,rnz,i,j;
219: PetscInt *row,*col;
220: Mat_SeqAIJ *aa=(Mat_SeqAIJ*)A->data;
223: *v=aa->a;
224: if (reuse == MAT_INITIAL_MATRIX) {
225: nz = aa->nz;
226: ai = aa->i;
227: aj = aa->j;
228: *nnz = nz;
229: PetscMalloc1(2*nz, &row);
230: col = row + nz;
232: nz = 0;
233: for (i=0; i<M; i++) {
234: rnz = ai[i+1] - ai[i];
235: ajj = aj + ai[i];
236: for (j=0; j<rnz; j++) {
237: row[nz] = i+shift; col[nz++] = ajj[j] + shift;
238: }
239: }
240: *r = row; *c = col;
241: }
242: return(0);
243: }
245: PetscErrorCode MatConvertToTriples_seqsell_seqaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
246: {
247: Mat_SeqSELL *a=(Mat_SeqSELL*)A->data;
248: PetscInt *ptr;
251: *v = a->val;
252: if (reuse == MAT_INITIAL_MATRIX) {
253: PetscInt nz,i,j,row;
256: nz = a->sliidx[a->totalslices];
257: *nnz = nz;
258: PetscMalloc1(2*nz, &ptr);
259: *r = ptr;
260: *c = ptr + nz;
262: for (i=0; i<a->totalslices; i++) {
263: for (j=a->sliidx[i],row=0; j<a->sliidx[i+1]; j++,row=((row+1)&0x07)) {
264: *ptr++ = 8*i + row + shift;
265: }
266: }
267: for (i=0;i<nz;i++) *ptr++ = a->colidx[i] + shift;
268: }
269: return(0);
270: }
272: PetscErrorCode MatConvertToTriples_seqbaij_seqaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
273: {
274: Mat_SeqBAIJ *aa=(Mat_SeqBAIJ*)A->data;
275: const PetscInt *ai,*aj,*ajj,bs2 = aa->bs2;
276: PetscInt bs,M,nz,idx=0,rnz,i,j,k,m;
278: PetscInt *row,*col;
281: MatGetBlockSize(A,&bs);
282: M = A->rmap->N/bs;
283: *v = aa->a;
284: if (reuse == MAT_INITIAL_MATRIX) {
285: ai = aa->i; aj = aa->j;
286: nz = bs2*aa->nz;
287: *nnz = nz;
288: PetscMalloc1(2*nz, &row);
289: col = row + nz;
291: for (i=0; i<M; i++) {
292: ajj = aj + ai[i];
293: rnz = ai[i+1] - ai[i];
294: for (k=0; k<rnz; k++) {
295: for (j=0; j<bs; j++) {
296: for (m=0; m<bs; m++) {
297: row[idx] = i*bs + m + shift;
298: col[idx++] = bs*(ajj[k]) + j + shift;
299: }
300: }
301: }
302: }
303: *r = row; *c = col;
304: }
305: return(0);
306: }
308: PetscErrorCode MatConvertToTriples_seqsbaij_seqsbaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
309: {
310: const PetscInt *ai, *aj,*ajj,M=A->rmap->n;
311: PetscInt nz,rnz,i,j;
313: PetscInt *row,*col;
314: Mat_SeqSBAIJ *aa=(Mat_SeqSBAIJ*)A->data;
317: *v = aa->a;
318: if (reuse == MAT_INITIAL_MATRIX) {
319: nz = aa->nz;
320: ai = aa->i;
321: aj = aa->j;
322: *v = aa->a;
323: *nnz = nz;
324: PetscMalloc1(2*nz, &row);
325: col = row + nz;
327: nz = 0;
328: for (i=0; i<M; i++) {
329: rnz = ai[i+1] - ai[i];
330: ajj = aj + ai[i];
331: for (j=0; j<rnz; j++) {
332: row[nz] = i+shift; col[nz++] = ajj[j] + shift;
333: }
334: }
335: *r = row; *c = col;
336: }
337: return(0);
338: }
340: PetscErrorCode MatConvertToTriples_seqaij_seqsbaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
341: {
342: const PetscInt *ai,*aj,*ajj,*adiag,M=A->rmap->n;
343: PetscInt nz,rnz,i,j;
344: const PetscScalar *av,*v1;
345: PetscScalar *val;
346: PetscErrorCode ierr;
347: PetscInt *row,*col;
348: Mat_SeqAIJ *aa=(Mat_SeqAIJ*)A->data;
349: PetscBool missing;
352: ai = aa->i; aj = aa->j; av = aa->a;
353: adiag = aa->diag;
354: MatMissingDiagonal_SeqAIJ(A,&missing,&i);
355: if (reuse == MAT_INITIAL_MATRIX) {
356: /* count nz in the upper triangular part of A */
357: nz = 0;
358: if (missing) {
359: for (i=0; i<M; i++) {
360: if (PetscUnlikely(adiag[i] >= ai[i+1])) {
361: for (j=ai[i];j<ai[i+1];j++) {
362: if (aj[j] < i) continue;
363: nz++;
364: }
365: } else {
366: nz += ai[i+1] - adiag[i];
367: }
368: }
369: } else {
370: for (i=0; i<M; i++) nz += ai[i+1] - adiag[i];
371: }
372: *nnz = nz;
374: PetscMalloc((2*nz*sizeof(PetscInt)+nz*sizeof(PetscScalar)), &row);
375: col = row + nz;
376: val = (PetscScalar*)(col + nz);
378: nz = 0;
379: if (missing) {
380: for (i=0; i<M; i++) {
381: if (PetscUnlikely(adiag[i] >= ai[i+1])) {
382: for (j=ai[i];j<ai[i+1];j++) {
383: if (aj[j] < i) continue;
384: row[nz] = i+shift;
385: col[nz] = aj[j]+shift;
386: val[nz] = av[j];
387: nz++;
388: }
389: } else {
390: rnz = ai[i+1] - adiag[i];
391: ajj = aj + adiag[i];
392: v1 = av + adiag[i];
393: for (j=0; j<rnz; j++) {
394: row[nz] = i+shift; col[nz] = ajj[j] + shift; val[nz++] = v1[j];
395: }
396: }
397: }
398: } else {
399: for (i=0; i<M; i++) {
400: rnz = ai[i+1] - adiag[i];
401: ajj = aj + adiag[i];
402: v1 = av + adiag[i];
403: for (j=0; j<rnz; j++) {
404: row[nz] = i+shift; col[nz] = ajj[j] + shift; val[nz++] = v1[j];
405: }
406: }
407: }
408: *r = row; *c = col; *v = val;
409: } else {
410: nz = 0; val = *v;
411: if (missing) {
412: for (i=0; i <M; i++) {
413: if (PetscUnlikely(adiag[i] >= ai[i+1])) {
414: for (j=ai[i];j<ai[i+1];j++) {
415: if (aj[j] < i) continue;
416: val[nz++] = av[j];
417: }
418: } else {
419: rnz = ai[i+1] - adiag[i];
420: v1 = av + adiag[i];
421: for (j=0; j<rnz; j++) {
422: val[nz++] = v1[j];
423: }
424: }
425: }
426: } else {
427: for (i=0; i <M; i++) {
428: rnz = ai[i+1] - adiag[i];
429: v1 = av + adiag[i];
430: for (j=0; j<rnz; j++) {
431: val[nz++] = v1[j];
432: }
433: }
434: }
435: }
436: return(0);
437: }
439: PetscErrorCode MatConvertToTriples_mpisbaij_mpisbaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
440: {
441: const PetscInt *ai, *aj, *bi, *bj,*garray,m=A->rmap->n,*ajj,*bjj;
442: PetscErrorCode ierr;
443: PetscInt rstart,nz,i,j,jj,irow,countA,countB;
444: PetscInt *row,*col;
445: const PetscScalar *av, *bv,*v1,*v2;
446: PetscScalar *val;
447: Mat_MPISBAIJ *mat = (Mat_MPISBAIJ*)A->data;
448: Mat_SeqSBAIJ *aa = (Mat_SeqSBAIJ*)(mat->A)->data;
449: Mat_SeqBAIJ *bb = (Mat_SeqBAIJ*)(mat->B)->data;
452: ai=aa->i; aj=aa->j; bi=bb->i; bj=bb->j; rstart= A->rmap->rstart;
453: av=aa->a; bv=bb->a;
455: garray = mat->garray;
457: if (reuse == MAT_INITIAL_MATRIX) {
458: nz = aa->nz + bb->nz;
459: *nnz = nz;
460: PetscMalloc((2*nz*sizeof(PetscInt)+nz*sizeof(PetscScalar)), &row);
461: col = row + nz;
462: val = (PetscScalar*)(col + nz);
464: *r = row; *c = col; *v = val;
465: } else {
466: row = *r; col = *c; val = *v;
467: }
469: jj = 0; irow = rstart;
470: for (i=0; i<m; i++) {
471: ajj = aj + ai[i]; /* ptr to the beginning of this row */
472: countA = ai[i+1] - ai[i];
473: countB = bi[i+1] - bi[i];
474: bjj = bj + bi[i];
475: v1 = av + ai[i];
476: v2 = bv + bi[i];
478: /* A-part */
479: for (j=0; j<countA; j++) {
480: if (reuse == MAT_INITIAL_MATRIX) {
481: row[jj] = irow + shift; col[jj] = rstart + ajj[j] + shift;
482: }
483: val[jj++] = v1[j];
484: }
486: /* B-part */
487: for (j=0; j < countB; j++) {
488: if (reuse == MAT_INITIAL_MATRIX) {
489: row[jj] = irow + shift; col[jj] = garray[bjj[j]] + shift;
490: }
491: val[jj++] = v2[j];
492: }
493: irow++;
494: }
495: return(0);
496: }
498: PetscErrorCode MatConvertToTriples_mpiaij_mpiaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
499: {
500: const PetscInt *ai, *aj, *bi, *bj,*garray,m=A->rmap->n,*ajj,*bjj;
501: PetscErrorCode ierr;
502: PetscInt rstart,nz,i,j,jj,irow,countA,countB;
503: PetscInt *row,*col;
504: const PetscScalar *av, *bv,*v1,*v2;
505: PetscScalar *val;
506: Mat_MPIAIJ *mat = (Mat_MPIAIJ*)A->data;
507: Mat_SeqAIJ *aa = (Mat_SeqAIJ*)(mat->A)->data;
508: Mat_SeqAIJ *bb = (Mat_SeqAIJ*)(mat->B)->data;
511: ai=aa->i; aj=aa->j; bi=bb->i; bj=bb->j; rstart= A->rmap->rstart;
512: av=aa->a; bv=bb->a;
514: garray = mat->garray;
516: if (reuse == MAT_INITIAL_MATRIX) {
517: nz = aa->nz + bb->nz;
518: *nnz = nz;
519: PetscMalloc((2*nz*sizeof(PetscInt)+nz*sizeof(PetscScalar)), &row);
520: col = row + nz;
521: val = (PetscScalar*)(col + nz);
523: *r = row; *c = col; *v = val;
524: } else {
525: row = *r; col = *c; val = *v;
526: }
528: jj = 0; irow = rstart;
529: for (i=0; i<m; i++) {
530: ajj = aj + ai[i]; /* ptr to the beginning of this row */
531: countA = ai[i+1] - ai[i];
532: countB = bi[i+1] - bi[i];
533: bjj = bj + bi[i];
534: v1 = av + ai[i];
535: v2 = bv + bi[i];
537: /* A-part */
538: for (j=0; j<countA; j++) {
539: if (reuse == MAT_INITIAL_MATRIX) {
540: row[jj] = irow + shift; col[jj] = rstart + ajj[j] + shift;
541: }
542: val[jj++] = v1[j];
543: }
545: /* B-part */
546: for (j=0; j < countB; j++) {
547: if (reuse == MAT_INITIAL_MATRIX) {
548: row[jj] = irow + shift; col[jj] = garray[bjj[j]] + shift;
549: }
550: val[jj++] = v2[j];
551: }
552: irow++;
553: }
554: return(0);
555: }
557: PetscErrorCode MatConvertToTriples_mpibaij_mpiaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
558: {
559: Mat_MPIBAIJ *mat = (Mat_MPIBAIJ*)A->data;
560: Mat_SeqBAIJ *aa = (Mat_SeqBAIJ*)(mat->A)->data;
561: Mat_SeqBAIJ *bb = (Mat_SeqBAIJ*)(mat->B)->data;
562: const PetscInt *ai = aa->i, *bi = bb->i, *aj = aa->j, *bj = bb->j,*ajj, *bjj;
563: const PetscInt *garray = mat->garray,mbs=mat->mbs,rstart=A->rmap->rstart;
564: const PetscInt bs2=mat->bs2;
565: PetscErrorCode ierr;
566: PetscInt bs,nz,i,j,k,n,jj,irow,countA,countB,idx;
567: PetscInt *row,*col;
568: const PetscScalar *av=aa->a, *bv=bb->a,*v1,*v2;
569: PetscScalar *val;
572: MatGetBlockSize(A,&bs);
573: if (reuse == MAT_INITIAL_MATRIX) {
574: nz = bs2*(aa->nz + bb->nz);
575: *nnz = nz;
576: PetscMalloc((2*nz*sizeof(PetscInt)+nz*sizeof(PetscScalar)), &row);
577: col = row + nz;
578: val = (PetscScalar*)(col + nz);
580: *r = row; *c = col; *v = val;
581: } else {
582: row = *r; col = *c; val = *v;
583: }
585: jj = 0; irow = rstart;
586: for (i=0; i<mbs; i++) {
587: countA = ai[i+1] - ai[i];
588: countB = bi[i+1] - bi[i];
589: ajj = aj + ai[i];
590: bjj = bj + bi[i];
591: v1 = av + bs2*ai[i];
592: v2 = bv + bs2*bi[i];
594: idx = 0;
595: /* A-part */
596: for (k=0; k<countA; k++) {
597: for (j=0; j<bs; j++) {
598: for (n=0; n<bs; n++) {
599: if (reuse == MAT_INITIAL_MATRIX) {
600: row[jj] = irow + n + shift;
601: col[jj] = rstart + bs*ajj[k] + j + shift;
602: }
603: val[jj++] = v1[idx++];
604: }
605: }
606: }
608: idx = 0;
609: /* B-part */
610: for (k=0; k<countB; k++) {
611: for (j=0; j<bs; j++) {
612: for (n=0; n<bs; n++) {
613: if (reuse == MAT_INITIAL_MATRIX) {
614: row[jj] = irow + n + shift;
615: col[jj] = bs*garray[bjj[k]] + j + shift;
616: }
617: val[jj++] = v2[idx++];
618: }
619: }
620: }
621: irow += bs;
622: }
623: return(0);
624: }
626: PetscErrorCode MatConvertToTriples_mpiaij_mpisbaij(Mat A,int shift,MatReuse reuse,int *nnz,int **r, int **c, PetscScalar **v)
627: {
628: const PetscInt *ai, *aj,*adiag, *bi, *bj,*garray,m=A->rmap->n,*ajj,*bjj;
629: PetscErrorCode ierr;
630: PetscInt rstart,nz,nza,nzb,i,j,jj,irow,countA,countB;
631: PetscInt *row,*col;
632: const PetscScalar *av, *bv,*v1,*v2;
633: PetscScalar *val;
634: Mat_MPIAIJ *mat = (Mat_MPIAIJ*)A->data;
635: Mat_SeqAIJ *aa =(Mat_SeqAIJ*)(mat->A)->data;
636: Mat_SeqAIJ *bb =(Mat_SeqAIJ*)(mat->B)->data;
639: ai=aa->i; aj=aa->j; adiag=aa->diag;
640: bi=bb->i; bj=bb->j; garray = mat->garray;
641: av=aa->a; bv=bb->a;
643: rstart = A->rmap->rstart;
645: if (reuse == MAT_INITIAL_MATRIX) {
646: nza = 0; /* num of upper triangular entries in mat->A, including diagonals */
647: nzb = 0; /* num of upper triangular entries in mat->B */
648: for (i=0; i<m; i++) {
649: nza += (ai[i+1] - adiag[i]);
650: countB = bi[i+1] - bi[i];
651: bjj = bj + bi[i];
652: for (j=0; j<countB; j++) {
653: if (garray[bjj[j]] > rstart) nzb++;
654: }
655: }
657: nz = nza + nzb; /* total nz of upper triangular part of mat */
658: *nnz = nz;
659: PetscMalloc((2*nz*sizeof(PetscInt)+nz*sizeof(PetscScalar)), &row);
660: col = row + nz;
661: val = (PetscScalar*)(col + nz);
663: *r = row; *c = col; *v = val;
664: } else {
665: row = *r; col = *c; val = *v;
666: }
668: jj = 0; irow = rstart;
669: for (i=0; i<m; i++) {
670: ajj = aj + adiag[i]; /* ptr to the beginning of the diagonal of this row */
671: v1 = av + adiag[i];
672: countA = ai[i+1] - adiag[i];
673: countB = bi[i+1] - bi[i];
674: bjj = bj + bi[i];
675: v2 = bv + bi[i];
677: /* A-part */
678: for (j=0; j<countA; j++) {
679: if (reuse == MAT_INITIAL_MATRIX) {
680: row[jj] = irow + shift; col[jj] = rstart + ajj[j] + shift;
681: }
682: val[jj++] = v1[j];
683: }
685: /* B-part */
686: for (j=0; j < countB; j++) {
687: if (garray[bjj[j]] > rstart) {
688: if (reuse == MAT_INITIAL_MATRIX) {
689: row[jj] = irow + shift; col[jj] = garray[bjj[j]] + shift;
690: }
691: val[jj++] = v2[j];
692: }
693: }
694: irow++;
695: }
696: return(0);
697: }
699: PetscErrorCode MatDestroy_MUMPS(Mat A)
700: {
701: Mat_MUMPS *mumps=(Mat_MUMPS*)A->data;
705: PetscFree2(mumps->id.sol_loc,mumps->id.isol_loc);
706: VecScatterDestroy(&mumps->scat_rhs);
707: VecScatterDestroy(&mumps->scat_sol);
708: VecDestroy(&mumps->b_seq);
709: VecDestroy(&mumps->x_seq);
710: PetscFree(mumps->id.perm_in);
711: PetscFree(mumps->irn);
712: PetscFree(mumps->info);
713: MatMumpsResetSchur_Private(mumps);
714: mumps->id.job = JOB_END;
715: PetscMUMPS_c(&mumps->id);
716: MPI_Comm_free(&mumps->comm_mumps);
717: PetscFree(A->data);
719: /* clear composed functions */
720: PetscObjectComposeFunction((PetscObject)A,"MatFactorGetSolverType_C",NULL);
721: PetscObjectComposeFunction((PetscObject)A,"MatFactorSetSchurIS_C",NULL);
722: PetscObjectComposeFunction((PetscObject)A,"MatFactorCreateSchurComplement_C",NULL);
723: PetscObjectComposeFunction((PetscObject)A,"MatMumpsSetIcntl_C",NULL);
724: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetIcntl_C",NULL);
725: PetscObjectComposeFunction((PetscObject)A,"MatMumpsSetCntl_C",NULL);
726: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetCntl_C",NULL);
727: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetInfo_C",NULL);
728: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetInfog_C",NULL);
729: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetRinfo_C",NULL);
730: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetRinfog_C",NULL);
731: PetscObjectComposeFunction((PetscObject)A,"MatMumpsGetInverse_C",NULL);
732: return(0);
733: }
735: PetscErrorCode MatSolve_MUMPS(Mat A,Vec b,Vec x)
736: {
737: Mat_MUMPS *mumps=(Mat_MUMPS*)A->data;
738: PetscScalar *array;
739: Vec b_seq;
740: IS is_iden,is_petsc;
741: PetscErrorCode ierr;
742: PetscInt i;
743: PetscBool second_solve = PETSC_FALSE;
744: static PetscBool cite1 = PETSC_FALSE,cite2 = PETSC_FALSE;
747: PetscCitationsRegister("@article{MUMPS01,\n author = {P.~R. Amestoy and I.~S. Duff and J.-Y. L'Excellent and J. Koster},\n title = {A fully asynchronous multifrontal solver using distributed dynamic scheduling},\n journal = {SIAM Journal on Matrix Analysis and Applications},\n volume = {23},\n number = {1},\n pages = {15--41},\n year = {2001}\n}\n",&cite1);
748: PetscCitationsRegister("@article{MUMPS02,\n author = {P.~R. Amestoy and A. Guermouche and J.-Y. L'Excellent and S. Pralet},\n title = {Hybrid scheduling for the parallel solution of linear systems},\n journal = {Parallel Computing},\n volume = {32},\n number = {2},\n pages = {136--156},\n year = {2006}\n}\n",&cite2);
750: if (A->factorerrortype) {
751: PetscInfo2(A,"MatSolve is called with singular matrix factor, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
752: VecSetInf(x);
753: return(0);
754: }
756: mumps->id.nrhs = 1;
757: b_seq = mumps->b_seq;
758: if (mumps->size > 1) {
759: /* MUMPS only supports centralized rhs. Scatter b into a seqential rhs vector */
760: VecScatterBegin(mumps->scat_rhs,b,b_seq,INSERT_VALUES,SCATTER_FORWARD);
761: VecScatterEnd(mumps->scat_rhs,b,b_seq,INSERT_VALUES,SCATTER_FORWARD);
762: if (!mumps->myid) {VecGetArray(b_seq,&array);}
763: } else { /* size == 1 */
764: VecCopy(b,x);
765: VecGetArray(x,&array);
766: }
767: if (!mumps->myid) { /* define rhs on the host */
768: mumps->id.nrhs = 1;
769: mumps->id.rhs = (MumpsScalar*)array;
770: }
772: /*
773: handle condensation step of Schur complement (if any)
774: We set by default ICNTL(26) == -1 when Schur indices have been provided by the user.
775: According to MUMPS (5.0.0) manual, any value should be harmful during the factorization phase
776: Unless the user provides a valid value for ICNTL(26), MatSolve and MatMatSolve routines solve the full system.
777: This requires an extra call to PetscMUMPS_c and the computation of the factors for S
778: */
779: if (mumps->id.size_schur > 0 && (mumps->id.ICNTL(26) < 0 || mumps->id.ICNTL(26) > 2)) {
780: if (mumps->size > 1) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Parallel Schur complements not yet supported from PETSc\n");
781: second_solve = PETSC_TRUE;
782: MatMumpsHandleSchur_Private(A,PETSC_FALSE);
783: }
784: /* solve phase */
785: /*-------------*/
786: mumps->id.job = JOB_SOLVE;
787: PetscMUMPS_c(&mumps->id);
788: if (mumps->id.INFOG(1) < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in solve phase: INFOG(1)=%d\n",mumps->id.INFOG(1));
790: /* handle expansion step of Schur complement (if any) */
791: if (second_solve) {
792: MatMumpsHandleSchur_Private(A,PETSC_TRUE);
793: }
795: if (mumps->size > 1) { /* convert mumps distributed solution to petsc mpi x */
796: if (mumps->scat_sol && mumps->ICNTL9_pre != mumps->id.ICNTL(9)) {
797: /* when id.ICNTL(9) changes, the contents of lsol_loc may change (not its size, lsol_loc), recreates scat_sol */
798: VecScatterDestroy(&mumps->scat_sol);
799: }
800: if (!mumps->scat_sol) { /* create scatter scat_sol */
801: ISCreateStride(PETSC_COMM_SELF,mumps->id.lsol_loc,0,1,&is_iden); /* from */
802: for (i=0; i<mumps->id.lsol_loc; i++) {
803: mumps->id.isol_loc[i] -= 1; /* change Fortran style to C style */
804: }
805: ISCreateGeneral(PETSC_COMM_SELF,mumps->id.lsol_loc,mumps->id.isol_loc,PETSC_COPY_VALUES,&is_petsc); /* to */
806: VecScatterCreate(mumps->x_seq,is_iden,x,is_petsc,&mumps->scat_sol);
807: ISDestroy(&is_iden);
808: ISDestroy(&is_petsc);
810: mumps->ICNTL9_pre = mumps->id.ICNTL(9); /* save current value of id.ICNTL(9) */
811: }
813: VecScatterBegin(mumps->scat_sol,mumps->x_seq,x,INSERT_VALUES,SCATTER_FORWARD);
814: VecScatterEnd(mumps->scat_sol,mumps->x_seq,x,INSERT_VALUES,SCATTER_FORWARD);
815: }
816: PetscLogFlops(2.0*mumps->id.RINFO(3));
817: return(0);
818: }
820: PetscErrorCode MatSolveTranspose_MUMPS(Mat A,Vec b,Vec x)
821: {
822: Mat_MUMPS *mumps=(Mat_MUMPS*)A->data;
826: mumps->id.ICNTL(9) = 0;
827: MatSolve_MUMPS(A,b,x);
828: mumps->id.ICNTL(9) = 1;
829: return(0);
830: }
832: PetscErrorCode MatMatSolve_MUMPS(Mat A,Mat B,Mat X)
833: {
835: Mat Bt = NULL;
836: PetscBool flg, flgT;
837: Mat_MUMPS *mumps=(Mat_MUMPS*)A->data;
838: PetscInt i,nrhs,M;
839: PetscScalar *array,*bray;
842: PetscObjectTypeCompareAny((PetscObject)B,&flg,MATSEQDENSE,MATMPIDENSE,NULL);
843: PetscObjectTypeCompare((PetscObject)B,MATTRANSPOSEMAT,&flgT);
844: if (flgT) {
845: if (mumps->size > 1) SETERRQ(PetscObjectComm((PetscObject)B),PETSC_ERR_ARG_WRONG,"Matrix B must be MATDENSE matrix");
846: MatTransposeGetMat(B,&Bt);
847: } else {
848: if (!flg) SETERRQ(PetscObjectComm((PetscObject)B),PETSC_ERR_ARG_WRONG,"Matrix B must be MATDENSE matrix");
849: }
851: MatGetSize(B,&M,&nrhs);
852: mumps->id.nrhs = nrhs;
853: mumps->id.lrhs = M;
855: PetscObjectTypeCompareAny((PetscObject)X,&flg,MATSEQDENSE,MATMPIDENSE,NULL);
856: if (!flg) SETERRQ(PetscObjectComm((PetscObject)X),PETSC_ERR_ARG_WRONG,"Matrix X must be MATDENSE matrix");
858: if (B->rmap->n != X->rmap->n) SETERRQ(PetscObjectComm((PetscObject)B),PETSC_ERR_ARG_WRONG,"Matrix B and X must have same row distribution");
860: if (mumps->size == 1) {
861: PetscScalar *aa;
862: PetscInt spnr,*ia,*ja;
863: PetscBool second_solve = PETSC_FALSE;
865: /* copy B to X */
866: MatDenseGetArray(X,&array);
867: mumps->id.rhs = (MumpsScalar*)array;
868: if (!Bt) {
869: MatDenseGetArray(B,&bray);
870: PetscMemcpy(array,bray,M*nrhs*sizeof(PetscScalar));
871: MatDenseRestoreArray(B,&bray);
872: } else {
873: PetscBool done;
875: MatSeqAIJGetArray(Bt,&aa);
876: MatGetRowIJ(Bt,1,PETSC_FALSE,PETSC_FALSE,&spnr,(const PetscInt**)&ia,(const PetscInt**)&ja,&done);
877: if (!done) SETERRQ(PetscObjectComm((PetscObject)Bt),PETSC_ERR_ARG_WRONG,"Cannot get IJ structure");
878: mumps->id.irhs_ptr = ia;
879: mumps->id.irhs_sparse = ja;
880: mumps->id.nz_rhs = ia[spnr] - 1;
881: mumps->id.rhs_sparse = (MumpsScalar*)aa;
882: mumps->id.ICNTL(20) = 1;
883: }
884: /* handle condensation step of Schur complement (if any) */
885: if (mumps->id.size_schur > 0 && (mumps->id.ICNTL(26) < 0 || mumps->id.ICNTL(26) > 2)) {
886: second_solve = PETSC_TRUE;
887: MatMumpsHandleSchur_Private(A,PETSC_FALSE);
888: }
889: /* solve phase */
890: /*-------------*/
891: mumps->id.job = JOB_SOLVE;
892: PetscMUMPS_c(&mumps->id);
893: if (mumps->id.INFOG(1) < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in solve phase: INFOG(1)=%d\n",mumps->id.INFOG(1));
895: /* handle expansion step of Schur complement (if any) */
896: if (second_solve) {
897: MatMumpsHandleSchur_Private(A,PETSC_TRUE);
898: }
899: if (Bt) {
900: PetscBool done;
902: MatSeqAIJRestoreArray(Bt,&aa);
903: MatRestoreRowIJ(Bt,1,PETSC_FALSE,PETSC_FALSE,&spnr,(const PetscInt**)&ia,(const PetscInt**)&ja,&done);
904: if (!done) SETERRQ(PetscObjectComm((PetscObject)Bt),PETSC_ERR_ARG_WRONG,"Cannot restore IJ structure");
905: mumps->id.ICNTL(20) = 0;
906: }
907: MatDenseRestoreArray(X,&array);
908: } else { /*--------- parallel case --------*/
909: PetscInt lsol_loc,nlsol_loc,*isol_loc,*idx,*iidx,*idxx,*isol_loc_save;
910: MumpsScalar *sol_loc,*sol_loc_save;
911: IS is_to,is_from;
912: PetscInt k,proc,j,m;
913: const PetscInt *rstart;
914: Vec v_mpi,b_seq,x_seq;
915: VecScatter scat_rhs,scat_sol;
917: if (mumps->size > 1 && mumps->id.ICNTL(19)) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Parallel Schur complements not yet supported from PETSc\n");
919: /* create x_seq to hold local solution */
920: isol_loc_save = mumps->id.isol_loc; /* save it for MatSovle() */
921: sol_loc_save = mumps->id.sol_loc;
923: lsol_loc = mumps->id.INFO(23);
924: nlsol_loc = nrhs*lsol_loc; /* length of sol_loc */
925: PetscMalloc2(nlsol_loc,&sol_loc,nlsol_loc,&isol_loc);
926: mumps->id.sol_loc = (MumpsScalar*)sol_loc;
927: mumps->id.isol_loc = isol_loc;
929: VecCreateSeqWithArray(PETSC_COMM_SELF,1,nlsol_loc,(PetscScalar*)sol_loc,&x_seq);
931: /* copy rhs matrix B into vector v_mpi */
932: MatGetLocalSize(B,&m,NULL);
933: MatDenseGetArray(B,&bray);
934: VecCreateMPIWithArray(PetscObjectComm((PetscObject)B),1,nrhs*m,nrhs*M,(const PetscScalar*)bray,&v_mpi);
935: MatDenseRestoreArray(B,&bray);
937: /* scatter v_mpi to b_seq because MUMPS only supports centralized rhs */
938: /* idx: maps from k-th index of v_mpi to (i,j)-th global entry of B;
939: iidx: inverse of idx, will be used by scattering xx_seq -> X */
940: PetscMalloc2(nrhs*M,&idx,nrhs*M,&iidx);
941: MatGetOwnershipRanges(B,&rstart);
942: k = 0;
943: for (proc=0; proc<mumps->size; proc++){
944: for (j=0; j<nrhs; j++){
945: for (i=rstart[proc]; i<rstart[proc+1]; i++){
946: iidx[j*M + i] = k;
947: idx[k++] = j*M + i;
948: }
949: }
950: }
952: if (!mumps->myid) {
953: VecCreateSeq(PETSC_COMM_SELF,nrhs*M,&b_seq);
954: ISCreateGeneral(PETSC_COMM_SELF,nrhs*M,idx,PETSC_COPY_VALUES,&is_to);
955: ISCreateStride(PETSC_COMM_SELF,nrhs*M,0,1,&is_from);
956: } else {
957: VecCreateSeq(PETSC_COMM_SELF,0,&b_seq);
958: ISCreateStride(PETSC_COMM_SELF,0,0,1,&is_to);
959: ISCreateStride(PETSC_COMM_SELF,0,0,1,&is_from);
960: }
961: VecScatterCreate(v_mpi,is_from,b_seq,is_to,&scat_rhs);
962: VecScatterBegin(scat_rhs,v_mpi,b_seq,INSERT_VALUES,SCATTER_FORWARD);
963: ISDestroy(&is_to);
964: ISDestroy(&is_from);
965: VecScatterEnd(scat_rhs,v_mpi,b_seq,INSERT_VALUES,SCATTER_FORWARD);
967: if (!mumps->myid) { /* define rhs on the host */
968: VecGetArray(b_seq,&bray);
969: mumps->id.rhs = (MumpsScalar*)bray;
970: VecRestoreArray(b_seq,&bray);
971: }
973: /* solve phase */
974: /*-------------*/
975: mumps->id.job = JOB_SOLVE;
976: PetscMUMPS_c(&mumps->id);
977: if (mumps->id.INFOG(1) < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in solve phase: INFOG(1)=%d\n",mumps->id.INFOG(1));
979: /* scatter mumps distributed solution to petsc vector v_mpi, which shares local arrays with solution matrix X */
980: MatDenseGetArray(X,&array);
981: VecPlaceArray(v_mpi,array);
983: /* create scatter scat_sol */
984: PetscMalloc1(nlsol_loc,&idxx);
985: ISCreateStride(PETSC_COMM_SELF,nlsol_loc,0,1,&is_from);
986: for (i=0; i<lsol_loc; i++) {
987: isol_loc[i] -= 1; /* change Fortran style to C style */
988: idxx[i] = iidx[isol_loc[i]];
989: for (j=1; j<nrhs; j++){
990: idxx[j*lsol_loc+i] = iidx[isol_loc[i]+j*M];
991: }
992: }
993: ISCreateGeneral(PETSC_COMM_SELF,nlsol_loc,idxx,PETSC_COPY_VALUES,&is_to);
994: VecScatterCreate(x_seq,is_from,v_mpi,is_to,&scat_sol);
995: VecScatterBegin(scat_sol,x_seq,v_mpi,INSERT_VALUES,SCATTER_FORWARD);
996: ISDestroy(&is_from);
997: ISDestroy(&is_to);
998: VecScatterEnd(scat_sol,x_seq,v_mpi,INSERT_VALUES,SCATTER_FORWARD);
999: MatDenseRestoreArray(X,&array);
1001: /* free spaces */
1002: mumps->id.sol_loc = sol_loc_save;
1003: mumps->id.isol_loc = isol_loc_save;
1005: PetscFree2(sol_loc,isol_loc);
1006: PetscFree2(idx,iidx);
1007: PetscFree(idxx);
1008: VecDestroy(&x_seq);
1009: VecDestroy(&v_mpi);
1010: VecDestroy(&b_seq);
1011: VecScatterDestroy(&scat_rhs);
1012: VecScatterDestroy(&scat_sol);
1013: }
1014: PetscLogFlops(2.0*nrhs*mumps->id.RINFO(3));
1015: return(0);
1016: }
1018: #if !defined(PETSC_USE_COMPLEX)
1019: /*
1020: input:
1021: F: numeric factor
1022: output:
1023: nneg: total number of negative pivots
1024: nzero: total number of zero pivots
1025: npos: (global dimension of F) - nneg - nzero
1026: */
1027: PetscErrorCode MatGetInertia_SBAIJMUMPS(Mat F,int *nneg,int *nzero,int *npos)
1028: {
1029: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1031: PetscMPIInt size;
1034: MPI_Comm_size(PetscObjectComm((PetscObject)F),&size);
1035: /* MUMPS 4.3.1 calls ScaLAPACK when ICNTL(13)=0 (default), which does not offer the possibility to compute the inertia of a dense matrix. Set ICNTL(13)=1 to skip ScaLAPACK */
1036: if (size > 1 && mumps->id.ICNTL(13) != 1) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"ICNTL(13)=%d. -mat_mumps_icntl_13 must be set as 1 for correct global matrix inertia\n",mumps->id.INFOG(13));
1038: if (nneg) *nneg = mumps->id.INFOG(12);
1039: if (nzero || npos) {
1040: if (mumps->id.ICNTL(24) != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"-mat_mumps_icntl_24 must be set as 1 for null pivot row detection");
1041: if (nzero) *nzero = mumps->id.INFOG(28);
1042: if (npos) *npos = F->rmap->N - (mumps->id.INFOG(12) + mumps->id.INFOG(28));
1043: }
1044: return(0);
1045: }
1046: #endif
1048: PetscErrorCode MatFactorNumeric_MUMPS(Mat F,Mat A,const MatFactorInfo *info)
1049: {
1050: Mat_MUMPS *mumps =(Mat_MUMPS*)(F)->data;
1052: PetscBool isMPIAIJ;
1055: if (mumps->id.INFOG(1) < 0) {
1056: if (mumps->id.INFOG(1) == -6) {
1057: PetscInfo2(A,"MatFactorNumeric is called with singular matrix structure, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1058: }
1059: PetscInfo2(A,"MatFactorNumeric is called after analysis phase fails, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1060: return(0);
1061: }
1063: (*mumps->ConvertToTriples)(A, 1, MAT_REUSE_MATRIX, &mumps->nz, &mumps->irn, &mumps->jcn, &mumps->val);
1065: /* numerical factorization phase */
1066: /*-------------------------------*/
1067: mumps->id.job = JOB_FACTNUMERIC;
1068: if (!mumps->id.ICNTL(18)) { /* A is centralized */
1069: if (!mumps->myid) {
1070: mumps->id.a = (MumpsScalar*)mumps->val;
1071: }
1072: } else {
1073: mumps->id.a_loc = (MumpsScalar*)mumps->val;
1074: }
1075: PetscMUMPS_c(&mumps->id);
1076: if (mumps->id.INFOG(1) < 0) {
1077: if (A->erroriffailure) {
1078: SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1079: } else {
1080: if (mumps->id.INFOG(1) == -10) { /* numerically singular matrix */
1081: PetscInfo2(F,"matrix is numerically singular, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1082: F->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
1083: } else if (mumps->id.INFOG(1) == -13) {
1084: PetscInfo2(F,"MUMPS in numerical factorization phase: INFOG(1)=%d, cannot allocate required memory %d megabytes\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1085: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1086: } else if (mumps->id.INFOG(1) == -8 || mumps->id.INFOG(1) == -9 || (-16 < mumps->id.INFOG(1) && mumps->id.INFOG(1) < -10) ) {
1087: PetscInfo2(F,"MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d, problem with workarray \n",mumps->id.INFOG(1),mumps->id.INFO(2));
1088: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1089: } else {
1090: PetscInfo2(F,"MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1091: F->factorerrortype = MAT_FACTOR_OTHER;
1092: }
1093: }
1094: }
1095: if (!mumps->myid && mumps->id.ICNTL(16) > 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB," mumps->id.ICNTL(16):=%d\n",mumps->id.INFOG(16));
1097: F->assembled = PETSC_TRUE;
1098: mumps->matstruc = SAME_NONZERO_PATTERN;
1099: if (F->schur) { /* reset Schur status to unfactored */
1100: if (mumps->id.ICNTL(19) == 1) { /* stored by rows */
1101: mumps->id.ICNTL(19) = 2;
1102: MatTranspose(F->schur,MAT_INPLACE_MATRIX,&F->schur);
1103: }
1104: MatFactorRestoreSchurComplement(F,NULL,MAT_FACTOR_SCHUR_UNFACTORED);
1105: }
1107: /* just to be sure that ICNTL(19) value returned by a call from MatMumpsGetIcntl is always consistent */
1108: if (!mumps->sym && mumps->id.ICNTL(19) && mumps->id.ICNTL(19) != 1) mumps->id.ICNTL(19) = 3;
1110: if (mumps->size > 1) {
1111: PetscInt lsol_loc;
1112: PetscScalar *sol_loc;
1114: PetscObjectTypeCompare((PetscObject)A,MATMPIAIJ,&isMPIAIJ);
1116: /* distributed solution; Create x_seq=sol_loc for repeated use */
1117: if (mumps->x_seq) {
1118: VecScatterDestroy(&mumps->scat_sol);
1119: PetscFree2(mumps->id.sol_loc,mumps->id.isol_loc);
1120: VecDestroy(&mumps->x_seq);
1121: }
1122: lsol_loc = mumps->id.INFO(23); /* length of sol_loc */
1123: PetscMalloc2(lsol_loc,&sol_loc,lsol_loc,&mumps->id.isol_loc);
1124: mumps->id.lsol_loc = lsol_loc;
1125: mumps->id.sol_loc = (MumpsScalar*)sol_loc;
1126: VecCreateSeqWithArray(PETSC_COMM_SELF,1,lsol_loc,sol_loc,&mumps->x_seq);
1127: }
1128: PetscLogFlops(mumps->id.RINFO(2));
1129: return(0);
1130: }
1132: /* Sets MUMPS options from the options database */
1133: PetscErrorCode PetscSetMUMPSFromOptions(Mat F, Mat A)
1134: {
1135: Mat_MUMPS *mumps = (Mat_MUMPS*)F->data;
1137: PetscInt icntl,info[40],i,ninfo=40;
1138: PetscBool flg;
1141: PetscOptionsBegin(PetscObjectComm((PetscObject)A),((PetscObject)A)->prefix,"MUMPS Options","Mat");
1142: PetscOptionsInt("-mat_mumps_icntl_1","ICNTL(1): output stream for error messages","None",mumps->id.ICNTL(1),&icntl,&flg);
1143: if (flg) mumps->id.ICNTL(1) = icntl;
1144: PetscOptionsInt("-mat_mumps_icntl_2","ICNTL(2): output stream for diagnostic printing, statistics, and warning","None",mumps->id.ICNTL(2),&icntl,&flg);
1145: if (flg) mumps->id.ICNTL(2) = icntl;
1146: PetscOptionsInt("-mat_mumps_icntl_3","ICNTL(3): output stream for global information, collected on the host","None",mumps->id.ICNTL(3),&icntl,&flg);
1147: if (flg) mumps->id.ICNTL(3) = icntl;
1149: PetscOptionsInt("-mat_mumps_icntl_4","ICNTL(4): level of printing (0 to 4)","None",mumps->id.ICNTL(4),&icntl,&flg);
1150: if (flg) mumps->id.ICNTL(4) = icntl;
1151: if (mumps->id.ICNTL(4) || PetscLogPrintInfo) mumps->id.ICNTL(3) = 6; /* resume MUMPS default id.ICNTL(3) = 6 */
1153: PetscOptionsInt("-mat_mumps_icntl_6","ICNTL(6): permutes to a zero-free diagonal and/or scale the matrix (0 to 7)","None",mumps->id.ICNTL(6),&icntl,&flg);
1154: if (flg) mumps->id.ICNTL(6) = icntl;
1156: PetscOptionsInt("-mat_mumps_icntl_7","ICNTL(7): computes a symmetric permutation in sequential analysis (0 to 7). 3=Scotch, 4=PORD, 5=Metis","None",mumps->id.ICNTL(7),&icntl,&flg);
1157: if (flg) {
1158: if (icntl== 1 && mumps->size > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"pivot order be set by the user in PERM_IN -- not supported by the PETSc/MUMPS interface\n");
1159: else mumps->id.ICNTL(7) = icntl;
1160: }
1162: PetscOptionsInt("-mat_mumps_icntl_8","ICNTL(8): scaling strategy (-2 to 8 or 77)","None",mumps->id.ICNTL(8),&mumps->id.ICNTL(8),NULL);
1163: /* PetscOptionsInt("-mat_mumps_icntl_9","ICNTL(9): computes the solution using A or A^T","None",mumps->id.ICNTL(9),&mumps->id.ICNTL(9),NULL); handled by MatSolveTranspose_MUMPS() */
1164: PetscOptionsInt("-mat_mumps_icntl_10","ICNTL(10): max num of refinements","None",mumps->id.ICNTL(10),&mumps->id.ICNTL(10),NULL);
1165: PetscOptionsInt("-mat_mumps_icntl_11","ICNTL(11): statistics related to an error analysis (via -ksp_view)","None",mumps->id.ICNTL(11),&mumps->id.ICNTL(11),NULL);
1166: PetscOptionsInt("-mat_mumps_icntl_12","ICNTL(12): an ordering strategy for symmetric matrices (0 to 3)","None",mumps->id.ICNTL(12),&mumps->id.ICNTL(12),NULL);
1167: PetscOptionsInt("-mat_mumps_icntl_13","ICNTL(13): parallelism of the root node (enable ScaLAPACK) and its splitting","None",mumps->id.ICNTL(13),&mumps->id.ICNTL(13),NULL);
1168: PetscOptionsInt("-mat_mumps_icntl_14","ICNTL(14): percentage increase in the estimated working space","None",mumps->id.ICNTL(14),&mumps->id.ICNTL(14),NULL);
1169: PetscOptionsInt("-mat_mumps_icntl_19","ICNTL(19): computes the Schur complement","None",mumps->id.ICNTL(19),&mumps->id.ICNTL(19),NULL);
1170: if (mumps->id.ICNTL(19) <= 0 || mumps->id.ICNTL(19) > 3) { /* reset any schur data (if any) */
1171: MatDestroy(&F->schur);
1172: MatMumpsResetSchur_Private(mumps);
1173: }
1174: /* PetscOptionsInt("-mat_mumps_icntl_20","ICNTL(20): the format (dense or sparse) of the right-hand sides","None",mumps->id.ICNTL(20),&mumps->id.ICNTL(20),NULL); -- sparse rhs is not supported in PETSc API */
1175: /* PetscOptionsInt("-mat_mumps_icntl_21","ICNTL(21): the distribution (centralized or distributed) of the solution vectors","None",mumps->id.ICNTL(21),&mumps->id.ICNTL(21),NULL); we only use distributed solution vector */
1177: PetscOptionsInt("-mat_mumps_icntl_22","ICNTL(22): in-core/out-of-core factorization and solve (0 or 1)","None",mumps->id.ICNTL(22),&mumps->id.ICNTL(22),NULL);
1178: PetscOptionsInt("-mat_mumps_icntl_23","ICNTL(23): max size of the working memory (MB) that can allocate per processor","None",mumps->id.ICNTL(23),&mumps->id.ICNTL(23),NULL);
1179: PetscOptionsInt("-mat_mumps_icntl_24","ICNTL(24): detection of null pivot rows (0 or 1)","None",mumps->id.ICNTL(24),&mumps->id.ICNTL(24),NULL);
1180: if (mumps->id.ICNTL(24)) {
1181: mumps->id.ICNTL(13) = 1; /* turn-off ScaLAPACK to help with the correct detection of null pivots */
1182: }
1184: PetscOptionsInt("-mat_mumps_icntl_25","ICNTL(25): computes a solution of a deficient matrix and a null space basis","None",mumps->id.ICNTL(25),&mumps->id.ICNTL(25),NULL);
1185: PetscOptionsInt("-mat_mumps_icntl_26","ICNTL(26): drives the solution phase if a Schur complement matrix","None",mumps->id.ICNTL(26),&mumps->id.ICNTL(26),NULL);
1186: PetscOptionsInt("-mat_mumps_icntl_27","ICNTL(27): the blocking size for multiple right-hand sides","None",mumps->id.ICNTL(27),&mumps->id.ICNTL(27),NULL);
1187: PetscOptionsInt("-mat_mumps_icntl_28","ICNTL(28): use 1 for sequential analysis and ictnl(7) ordering, or 2 for parallel analysis and ictnl(29) ordering","None",mumps->id.ICNTL(28),&mumps->id.ICNTL(28),NULL);
1188: PetscOptionsInt("-mat_mumps_icntl_29","ICNTL(29): parallel ordering 1 = ptscotch, 2 = parmetis","None",mumps->id.ICNTL(29),&mumps->id.ICNTL(29),NULL);
1189: /* PetscOptionsInt("-mat_mumps_icntl_30","ICNTL(30): compute user-specified set of entries in inv(A)","None",mumps->id.ICNTL(30),&mumps->id.ICNTL(30),NULL); */ /* call MatMumpsGetInverse() directly */
1190: PetscOptionsInt("-mat_mumps_icntl_31","ICNTL(31): indicates which factors may be discarded during factorization","None",mumps->id.ICNTL(31),&mumps->id.ICNTL(31),NULL);
1191: /* PetscOptionsInt("-mat_mumps_icntl_32","ICNTL(32): performs the forward elemination of the right-hand sides during factorization","None",mumps->id.ICNTL(32),&mumps->id.ICNTL(32),NULL); -- not supported by PETSc API */
1192: PetscOptionsInt("-mat_mumps_icntl_33","ICNTL(33): compute determinant","None",mumps->id.ICNTL(33),&mumps->id.ICNTL(33),NULL);
1193: PetscOptionsInt("-mat_mumps_icntl_35","ICNTL(35): activates Block Lock Rank (BLR) based factorization","None",mumps->id.ICNTL(35),&mumps->id.ICNTL(35),NULL);
1195: PetscOptionsReal("-mat_mumps_cntl_1","CNTL(1): relative pivoting threshold","None",mumps->id.CNTL(1),&mumps->id.CNTL(1),NULL);
1196: PetscOptionsReal("-mat_mumps_cntl_2","CNTL(2): stopping criterion of refinement","None",mumps->id.CNTL(2),&mumps->id.CNTL(2),NULL);
1197: PetscOptionsReal("-mat_mumps_cntl_3","CNTL(3): absolute pivoting threshold","None",mumps->id.CNTL(3),&mumps->id.CNTL(3),NULL);
1198: PetscOptionsReal("-mat_mumps_cntl_4","CNTL(4): value for static pivoting","None",mumps->id.CNTL(4),&mumps->id.CNTL(4),NULL);
1199: PetscOptionsReal("-mat_mumps_cntl_5","CNTL(5): fixation for null pivots","None",mumps->id.CNTL(5),&mumps->id.CNTL(5),NULL);
1200: PetscOptionsReal("-mat_mumps_cntl_7","CNTL(7): dropping parameter used during BLR","None",mumps->id.CNTL(7),&mumps->id.CNTL(7),NULL);
1202: PetscOptionsString("-mat_mumps_ooc_tmpdir", "out of core directory", "None", mumps->id.ooc_tmpdir, mumps->id.ooc_tmpdir, 256, NULL);
1204: PetscOptionsIntArray("-mat_mumps_view_info","request INFO local to each processor","",info,&ninfo,NULL);
1205: if (ninfo) {
1206: if (ninfo > 40) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_USER,"number of INFO %d must <= 40\n",ninfo);
1207: PetscMalloc1(ninfo,&mumps->info);
1208: mumps->ninfo = ninfo;
1209: for (i=0; i<ninfo; i++) {
1210: if (info[i] < 0 || info[i]>40) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_USER,"index of INFO %d must between 1 and 40\n",ninfo);
1211: else mumps->info[i] = info[i];
1212: }
1213: }
1215: PetscOptionsEnd();
1216: return(0);
1217: }
1219: PetscErrorCode PetscInitializeMUMPS(Mat A,Mat_MUMPS *mumps)
1220: {
1224: MPI_Comm_rank(PetscObjectComm((PetscObject)A), &mumps->myid);
1225: MPI_Comm_size(PetscObjectComm((PetscObject)A),&mumps->size);
1226: MPI_Comm_dup(PetscObjectComm((PetscObject)A),&(mumps->comm_mumps));
1228: mumps->id.comm_fortran = MPI_Comm_c2f(mumps->comm_mumps);
1230: mumps->id.job = JOB_INIT;
1231: mumps->id.par = 1; /* host participates factorizaton and solve */
1232: mumps->id.sym = mumps->sym;
1233: PetscMUMPS_c(&mumps->id);
1235: mumps->scat_rhs = NULL;
1236: mumps->scat_sol = NULL;
1238: /* set PETSc-MUMPS default options - override MUMPS default */
1239: mumps->id.ICNTL(3) = 0;
1240: mumps->id.ICNTL(4) = 0;
1241: if (mumps->size == 1) {
1242: mumps->id.ICNTL(18) = 0; /* centralized assembled matrix input */
1243: } else {
1244: mumps->id.ICNTL(18) = 3; /* distributed assembled matrix input */
1245: mumps->id.ICNTL(20) = 0; /* rhs is in dense format */
1246: mumps->id.ICNTL(21) = 1; /* distributed solution */
1247: }
1249: /* schur */
1250: mumps->id.size_schur = 0;
1251: mumps->id.listvar_schur = NULL;
1252: mumps->id.schur = NULL;
1253: mumps->sizeredrhs = 0;
1254: mumps->schur_sol = NULL;
1255: mumps->schur_sizesol = 0;
1256: return(0);
1257: }
1259: PetscErrorCode MatFactorSymbolic_MUMPS_ReportIfError(Mat F,Mat A,const MatFactorInfo *info,Mat_MUMPS *mumps)
1260: {
1264: if (mumps->id.INFOG(1) < 0) {
1265: if (A->erroriffailure) {
1266: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in analysis phase: INFOG(1)=%d\n",mumps->id.INFOG(1));
1267: } else {
1268: if (mumps->id.INFOG(1) == -6) {
1269: PetscInfo2(F,"matrix is singular in structure, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1270: F->factorerrortype = MAT_FACTOR_STRUCT_ZEROPIVOT;
1271: } else if (mumps->id.INFOG(1) == -5 || mumps->id.INFOG(1) == -7) {
1272: PetscInfo2(F,"problem of workspace, INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1273: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1274: } else {
1275: PetscInfo2(F,"Error reported by MUMPS in analysis phase: INFOG(1)=%d, INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
1276: F->factorerrortype = MAT_FACTOR_OTHER;
1277: }
1278: }
1279: }
1280: return(0);
1281: }
1283: /* Note Petsc r(=c) permutation is used when mumps->id.ICNTL(7)==1 with centralized assembled matrix input; otherwise r and c are ignored */
1284: PetscErrorCode MatLUFactorSymbolic_AIJMUMPS(Mat F,Mat A,IS r,IS c,const MatFactorInfo *info)
1285: {
1286: Mat_MUMPS *mumps = (Mat_MUMPS*)F->data;
1288: Vec b;
1289: IS is_iden;
1290: const PetscInt M = A->rmap->N;
1293: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
1295: /* Set MUMPS options from the options database */
1296: PetscSetMUMPSFromOptions(F,A);
1298: (*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, &mumps->nz, &mumps->irn, &mumps->jcn, &mumps->val);
1300: /* analysis phase */
1301: /*----------------*/
1302: mumps->id.job = JOB_FACTSYMBOLIC;
1303: mumps->id.n = M;
1304: switch (mumps->id.ICNTL(18)) {
1305: case 0: /* centralized assembled matrix input */
1306: if (!mumps->myid) {
1307: mumps->id.nz =mumps->nz; mumps->id.irn=mumps->irn; mumps->id.jcn=mumps->jcn;
1308: if (mumps->id.ICNTL(6)>1) {
1309: mumps->id.a = (MumpsScalar*)mumps->val;
1310: }
1311: if (mumps->id.ICNTL(7) == 1) { /* use user-provide matrix ordering - assuming r = c ordering */
1312: /*
1313: PetscBool flag;
1314: ISEqual(r,c,&flag);
1315: if (!flag) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_USER,"row_perm != col_perm");
1316: ISView(r,PETSC_VIEWER_STDOUT_SELF);
1317: */
1318: if (!mumps->myid) {
1319: const PetscInt *idx;
1320: PetscInt i,*perm_in;
1322: PetscMalloc1(M,&perm_in);
1323: ISGetIndices(r,&idx);
1325: mumps->id.perm_in = perm_in;
1326: for (i=0; i<M; i++) perm_in[i] = idx[i]+1; /* perm_in[]: start from 1, not 0! */
1327: ISRestoreIndices(r,&idx);
1328: }
1329: }
1330: }
1331: break;
1332: case 3: /* distributed assembled matrix input (size>1) */
1333: mumps->id.nz_loc = mumps->nz;
1334: mumps->id.irn_loc=mumps->irn; mumps->id.jcn_loc=mumps->jcn;
1335: if (mumps->id.ICNTL(6)>1) {
1336: mumps->id.a_loc = (MumpsScalar*)mumps->val;
1337: }
1338: /* MUMPS only supports centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
1339: if (!mumps->myid) {
1340: VecCreateSeq(PETSC_COMM_SELF,A->rmap->N,&mumps->b_seq);
1341: ISCreateStride(PETSC_COMM_SELF,A->rmap->N,0,1,&is_iden);
1342: } else {
1343: VecCreateSeq(PETSC_COMM_SELF,0,&mumps->b_seq);
1344: ISCreateStride(PETSC_COMM_SELF,0,0,1,&is_iden);
1345: }
1346: MatCreateVecs(A,NULL,&b);
1347: VecScatterCreate(b,is_iden,mumps->b_seq,is_iden,&mumps->scat_rhs);
1348: ISDestroy(&is_iden);
1349: VecDestroy(&b);
1350: break;
1351: }
1352: PetscMUMPS_c(&mumps->id);
1353: MatFactorSymbolic_MUMPS_ReportIfError(F,A,info,mumps);
1355: F->ops->lufactornumeric = MatFactorNumeric_MUMPS;
1356: F->ops->solve = MatSolve_MUMPS;
1357: F->ops->solvetranspose = MatSolveTranspose_MUMPS;
1358: F->ops->matsolve = MatMatSolve_MUMPS;
1359: return(0);
1360: }
1362: /* Note the Petsc r and c permutations are ignored */
1363: PetscErrorCode MatLUFactorSymbolic_BAIJMUMPS(Mat F,Mat A,IS r,IS c,const MatFactorInfo *info)
1364: {
1365: Mat_MUMPS *mumps = (Mat_MUMPS*)F->data;
1367: Vec b;
1368: IS is_iden;
1369: const PetscInt M = A->rmap->N;
1372: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
1374: /* Set MUMPS options from the options database */
1375: PetscSetMUMPSFromOptions(F,A);
1377: (*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, &mumps->nz, &mumps->irn, &mumps->jcn, &mumps->val);
1379: /* analysis phase */
1380: /*----------------*/
1381: mumps->id.job = JOB_FACTSYMBOLIC;
1382: mumps->id.n = M;
1383: switch (mumps->id.ICNTL(18)) {
1384: case 0: /* centralized assembled matrix input */
1385: if (!mumps->myid) {
1386: mumps->id.nz =mumps->nz; mumps->id.irn=mumps->irn; mumps->id.jcn=mumps->jcn;
1387: if (mumps->id.ICNTL(6)>1) {
1388: mumps->id.a = (MumpsScalar*)mumps->val;
1389: }
1390: }
1391: break;
1392: case 3: /* distributed assembled matrix input (size>1) */
1393: mumps->id.nz_loc = mumps->nz;
1394: mumps->id.irn_loc=mumps->irn; mumps->id.jcn_loc=mumps->jcn;
1395: if (mumps->id.ICNTL(6)>1) {
1396: mumps->id.a_loc = (MumpsScalar*)mumps->val;
1397: }
1398: /* MUMPS only supports centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
1399: if (!mumps->myid) {
1400: VecCreateSeq(PETSC_COMM_SELF,A->cmap->N,&mumps->b_seq);
1401: ISCreateStride(PETSC_COMM_SELF,A->cmap->N,0,1,&is_iden);
1402: } else {
1403: VecCreateSeq(PETSC_COMM_SELF,0,&mumps->b_seq);
1404: ISCreateStride(PETSC_COMM_SELF,0,0,1,&is_iden);
1405: }
1406: MatCreateVecs(A,NULL,&b);
1407: VecScatterCreate(b,is_iden,mumps->b_seq,is_iden,&mumps->scat_rhs);
1408: ISDestroy(&is_iden);
1409: VecDestroy(&b);
1410: break;
1411: }
1412: PetscMUMPS_c(&mumps->id);
1413: MatFactorSymbolic_MUMPS_ReportIfError(F,A,info,mumps);
1415: F->ops->lufactornumeric = MatFactorNumeric_MUMPS;
1416: F->ops->solve = MatSolve_MUMPS;
1417: F->ops->solvetranspose = MatSolveTranspose_MUMPS;
1418: return(0);
1419: }
1421: /* Note the Petsc r permutation and factor info are ignored */
1422: PetscErrorCode MatCholeskyFactorSymbolic_MUMPS(Mat F,Mat A,IS r,const MatFactorInfo *info)
1423: {
1424: Mat_MUMPS *mumps = (Mat_MUMPS*)F->data;
1426: Vec b;
1427: IS is_iden;
1428: const PetscInt M = A->rmap->N;
1431: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
1433: /* Set MUMPS options from the options database */
1434: PetscSetMUMPSFromOptions(F,A);
1436: (*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, &mumps->nz, &mumps->irn, &mumps->jcn, &mumps->val);
1438: /* analysis phase */
1439: /*----------------*/
1440: mumps->id.job = JOB_FACTSYMBOLIC;
1441: mumps->id.n = M;
1442: switch (mumps->id.ICNTL(18)) {
1443: case 0: /* centralized assembled matrix input */
1444: if (!mumps->myid) {
1445: mumps->id.nz =mumps->nz; mumps->id.irn=mumps->irn; mumps->id.jcn=mumps->jcn;
1446: if (mumps->id.ICNTL(6)>1) {
1447: mumps->id.a = (MumpsScalar*)mumps->val;
1448: }
1449: }
1450: break;
1451: case 3: /* distributed assembled matrix input (size>1) */
1452: mumps->id.nz_loc = mumps->nz;
1453: mumps->id.irn_loc=mumps->irn; mumps->id.jcn_loc=mumps->jcn;
1454: if (mumps->id.ICNTL(6)>1) {
1455: mumps->id.a_loc = (MumpsScalar*)mumps->val;
1456: }
1457: /* MUMPS only supports centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
1458: if (!mumps->myid) {
1459: VecCreateSeq(PETSC_COMM_SELF,A->cmap->N,&mumps->b_seq);
1460: ISCreateStride(PETSC_COMM_SELF,A->cmap->N,0,1,&is_iden);
1461: } else {
1462: VecCreateSeq(PETSC_COMM_SELF,0,&mumps->b_seq);
1463: ISCreateStride(PETSC_COMM_SELF,0,0,1,&is_iden);
1464: }
1465: MatCreateVecs(A,NULL,&b);
1466: VecScatterCreate(b,is_iden,mumps->b_seq,is_iden,&mumps->scat_rhs);
1467: ISDestroy(&is_iden);
1468: VecDestroy(&b);
1469: break;
1470: }
1471: PetscMUMPS_c(&mumps->id);
1472: MatFactorSymbolic_MUMPS_ReportIfError(F,A,info,mumps);
1474: F->ops->choleskyfactornumeric = MatFactorNumeric_MUMPS;
1475: F->ops->solve = MatSolve_MUMPS;
1476: F->ops->solvetranspose = MatSolve_MUMPS;
1477: F->ops->matsolve = MatMatSolve_MUMPS;
1478: #if defined(PETSC_USE_COMPLEX)
1479: F->ops->getinertia = NULL;
1480: #else
1481: F->ops->getinertia = MatGetInertia_SBAIJMUMPS;
1482: #endif
1483: return(0);
1484: }
1486: PetscErrorCode MatView_MUMPS(Mat A,PetscViewer viewer)
1487: {
1488: PetscErrorCode ierr;
1489: PetscBool iascii;
1490: PetscViewerFormat format;
1491: Mat_MUMPS *mumps=(Mat_MUMPS*)A->data;
1494: /* check if matrix is mumps type */
1495: if (A->ops->solve != MatSolve_MUMPS) return(0);
1497: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);
1498: if (iascii) {
1499: PetscViewerGetFormat(viewer,&format);
1500: if (format == PETSC_VIEWER_ASCII_INFO) {
1501: PetscViewerASCIIPrintf(viewer,"MUMPS run parameters:\n");
1502: PetscViewerASCIIPrintf(viewer," SYM (matrix type): %d \n",mumps->id.sym);
1503: PetscViewerASCIIPrintf(viewer," PAR (host participation): %d \n",mumps->id.par);
1504: PetscViewerASCIIPrintf(viewer," ICNTL(1) (output for error): %d \n",mumps->id.ICNTL(1));
1505: PetscViewerASCIIPrintf(viewer," ICNTL(2) (output of diagnostic msg): %d \n",mumps->id.ICNTL(2));
1506: PetscViewerASCIIPrintf(viewer," ICNTL(3) (output for global info): %d \n",mumps->id.ICNTL(3));
1507: PetscViewerASCIIPrintf(viewer," ICNTL(4) (level of printing): %d \n",mumps->id.ICNTL(4));
1508: PetscViewerASCIIPrintf(viewer," ICNTL(5) (input mat struct): %d \n",mumps->id.ICNTL(5));
1509: PetscViewerASCIIPrintf(viewer," ICNTL(6) (matrix prescaling): %d \n",mumps->id.ICNTL(6));
1510: PetscViewerASCIIPrintf(viewer," ICNTL(7) (sequential matrix ordering):%d \n",mumps->id.ICNTL(7));
1511: PetscViewerASCIIPrintf(viewer," ICNTL(8) (scaling strategy): %d \n",mumps->id.ICNTL(8));
1512: PetscViewerASCIIPrintf(viewer," ICNTL(10) (max num of refinements): %d \n",mumps->id.ICNTL(10));
1513: PetscViewerASCIIPrintf(viewer," ICNTL(11) (error analysis): %d \n",mumps->id.ICNTL(11));
1514: if (mumps->id.ICNTL(11)>0) {
1515: PetscViewerASCIIPrintf(viewer," RINFOG(4) (inf norm of input mat): %g\n",mumps->id.RINFOG(4));
1516: PetscViewerASCIIPrintf(viewer," RINFOG(5) (inf norm of solution): %g\n",mumps->id.RINFOG(5));
1517: PetscViewerASCIIPrintf(viewer," RINFOG(6) (inf norm of residual): %g\n",mumps->id.RINFOG(6));
1518: PetscViewerASCIIPrintf(viewer," RINFOG(7),RINFOG(8) (backward error est): %g, %g\n",mumps->id.RINFOG(7),mumps->id.RINFOG(8));
1519: PetscViewerASCIIPrintf(viewer," RINFOG(9) (error estimate): %g \n",mumps->id.RINFOG(9));
1520: PetscViewerASCIIPrintf(viewer," RINFOG(10),RINFOG(11)(condition numbers): %g, %g\n",mumps->id.RINFOG(10),mumps->id.RINFOG(11));
1521: }
1522: PetscViewerASCIIPrintf(viewer," ICNTL(12) (efficiency control): %d \n",mumps->id.ICNTL(12));
1523: PetscViewerASCIIPrintf(viewer," ICNTL(13) (efficiency control): %d \n",mumps->id.ICNTL(13));
1524: PetscViewerASCIIPrintf(viewer," ICNTL(14) (percentage of estimated workspace increase): %d \n",mumps->id.ICNTL(14));
1525: /* ICNTL(15-17) not used */
1526: PetscViewerASCIIPrintf(viewer," ICNTL(18) (input mat struct): %d \n",mumps->id.ICNTL(18));
1527: PetscViewerASCIIPrintf(viewer," ICNTL(19) (Schur complement info): %d \n",mumps->id.ICNTL(19));
1528: PetscViewerASCIIPrintf(viewer," ICNTL(20) (rhs sparse pattern): %d \n",mumps->id.ICNTL(20));
1529: PetscViewerASCIIPrintf(viewer," ICNTL(21) (solution struct): %d \n",mumps->id.ICNTL(21));
1530: PetscViewerASCIIPrintf(viewer," ICNTL(22) (in-core/out-of-core facility): %d \n",mumps->id.ICNTL(22));
1531: PetscViewerASCIIPrintf(viewer," ICNTL(23) (max size of memory can be allocated locally):%d \n",mumps->id.ICNTL(23));
1533: PetscViewerASCIIPrintf(viewer," ICNTL(24) (detection of null pivot rows): %d \n",mumps->id.ICNTL(24));
1534: PetscViewerASCIIPrintf(viewer," ICNTL(25) (computation of a null space basis): %d \n",mumps->id.ICNTL(25));
1535: PetscViewerASCIIPrintf(viewer," ICNTL(26) (Schur options for rhs or solution): %d \n",mumps->id.ICNTL(26));
1536: PetscViewerASCIIPrintf(viewer," ICNTL(27) (experimental parameter): %d \n",mumps->id.ICNTL(27));
1537: PetscViewerASCIIPrintf(viewer," ICNTL(28) (use parallel or sequential ordering): %d \n",mumps->id.ICNTL(28));
1538: PetscViewerASCIIPrintf(viewer," ICNTL(29) (parallel ordering): %d \n",mumps->id.ICNTL(29));
1540: PetscViewerASCIIPrintf(viewer," ICNTL(30) (user-specified set of entries in inv(A)): %d \n",mumps->id.ICNTL(30));
1541: PetscViewerASCIIPrintf(viewer," ICNTL(31) (factors is discarded in the solve phase): %d \n",mumps->id.ICNTL(31));
1542: PetscViewerASCIIPrintf(viewer," ICNTL(33) (compute determinant): %d \n",mumps->id.ICNTL(33));
1543: PetscViewerASCIIPrintf(viewer," ICNTL(35) (activate BLR based factorization): %d \n",mumps->id.ICNTL(35));
1545: PetscViewerASCIIPrintf(viewer," CNTL(1) (relative pivoting threshold): %g \n",mumps->id.CNTL(1));
1546: PetscViewerASCIIPrintf(viewer," CNTL(2) (stopping criterion of refinement): %g \n",mumps->id.CNTL(2));
1547: PetscViewerASCIIPrintf(viewer," CNTL(3) (absolute pivoting threshold): %g \n",mumps->id.CNTL(3));
1548: PetscViewerASCIIPrintf(viewer," CNTL(4) (value of static pivoting): %g \n",mumps->id.CNTL(4));
1549: PetscViewerASCIIPrintf(viewer," CNTL(5) (fixation for null pivots): %g \n",mumps->id.CNTL(5));
1550: PetscViewerASCIIPrintf(viewer," CNTL(7) (dropping parameter for BLR): %g \n",mumps->id.CNTL(7));
1552: /* infomation local to each processor */
1553: PetscViewerASCIIPrintf(viewer, " RINFO(1) (local estimated flops for the elimination after analysis): \n");
1554: PetscViewerASCIIPushSynchronized(viewer);
1555: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %g \n",mumps->myid,mumps->id.RINFO(1));
1556: PetscViewerFlush(viewer);
1557: PetscViewerASCIIPrintf(viewer, " RINFO(2) (local estimated flops for the assembly after factorization): \n");
1558: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %g \n",mumps->myid,mumps->id.RINFO(2));
1559: PetscViewerFlush(viewer);
1560: PetscViewerASCIIPrintf(viewer, " RINFO(3) (local estimated flops for the elimination after factorization): \n");
1561: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %g \n",mumps->myid,mumps->id.RINFO(3));
1562: PetscViewerFlush(viewer);
1564: PetscViewerASCIIPrintf(viewer, " INFO(15) (estimated size of (in MB) MUMPS internal data for running numerical factorization): \n");
1565: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %d \n",mumps->myid,mumps->id.INFO(15));
1566: PetscViewerFlush(viewer);
1568: PetscViewerASCIIPrintf(viewer, " INFO(16) (size of (in MB) MUMPS internal data used during numerical factorization): \n");
1569: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %d \n",mumps->myid,mumps->id.INFO(16));
1570: PetscViewerFlush(viewer);
1572: PetscViewerASCIIPrintf(viewer, " INFO(23) (num of pivots eliminated on this processor after factorization): \n");
1573: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %d \n",mumps->myid,mumps->id.INFO(23));
1574: PetscViewerFlush(viewer);
1576: if (mumps->ninfo && mumps->ninfo <= 40){
1577: PetscInt i;
1578: for (i=0; i<mumps->ninfo; i++){
1579: PetscViewerASCIIPrintf(viewer, " INFO(%d): \n",mumps->info[i]);
1580: PetscViewerASCIISynchronizedPrintf(viewer," [%d] %d \n",mumps->myid,mumps->id.INFO(mumps->info[i]));
1581: PetscViewerFlush(viewer);
1582: }
1583: }
1584: PetscViewerASCIIPopSynchronized(viewer);
1586: if (!mumps->myid) { /* information from the host */
1587: PetscViewerASCIIPrintf(viewer," RINFOG(1) (global estimated flops for the elimination after analysis): %g \n",mumps->id.RINFOG(1));
1588: PetscViewerASCIIPrintf(viewer," RINFOG(2) (global estimated flops for the assembly after factorization): %g \n",mumps->id.RINFOG(2));
1589: PetscViewerASCIIPrintf(viewer," RINFOG(3) (global estimated flops for the elimination after factorization): %g \n",mumps->id.RINFOG(3));
1590: PetscViewerASCIIPrintf(viewer," (RINFOG(12) RINFOG(13))*2^INFOG(34) (determinant): (%g,%g)*(2^%d)\n",mumps->id.RINFOG(12),mumps->id.RINFOG(13),mumps->id.INFOG(34));
1592: PetscViewerASCIIPrintf(viewer," INFOG(3) (estimated real workspace for factors on all processors after analysis): %d \n",mumps->id.INFOG(3));
1593: PetscViewerASCIIPrintf(viewer," INFOG(4) (estimated integer workspace for factors on all processors after analysis): %d \n",mumps->id.INFOG(4));
1594: PetscViewerASCIIPrintf(viewer," INFOG(5) (estimated maximum front size in the complete tree): %d \n",mumps->id.INFOG(5));
1595: PetscViewerASCIIPrintf(viewer," INFOG(6) (number of nodes in the complete tree): %d \n",mumps->id.INFOG(6));
1596: PetscViewerASCIIPrintf(viewer," INFOG(7) (ordering option effectively use after analysis): %d \n",mumps->id.INFOG(7));
1597: PetscViewerASCIIPrintf(viewer," INFOG(8) (structural symmetry in percent of the permuted matrix after analysis): %d \n",mumps->id.INFOG(8));
1598: PetscViewerASCIIPrintf(viewer," INFOG(9) (total real/complex workspace to store the matrix factors after factorization): %d \n",mumps->id.INFOG(9));
1599: PetscViewerASCIIPrintf(viewer," INFOG(10) (total integer space store the matrix factors after factorization): %d \n",mumps->id.INFOG(10));
1600: PetscViewerASCIIPrintf(viewer," INFOG(11) (order of largest frontal matrix after factorization): %d \n",mumps->id.INFOG(11));
1601: PetscViewerASCIIPrintf(viewer," INFOG(12) (number of off-diagonal pivots): %d \n",mumps->id.INFOG(12));
1602: PetscViewerASCIIPrintf(viewer," INFOG(13) (number of delayed pivots after factorization): %d \n",mumps->id.INFOG(13));
1603: PetscViewerASCIIPrintf(viewer," INFOG(14) (number of memory compress after factorization): %d \n",mumps->id.INFOG(14));
1604: PetscViewerASCIIPrintf(viewer," INFOG(15) (number of steps of iterative refinement after solution): %d \n",mumps->id.INFOG(15));
1605: PetscViewerASCIIPrintf(viewer," INFOG(16) (estimated size (in MB) of all MUMPS internal data for factorization after analysis: value on the most memory consuming processor): %d \n",mumps->id.INFOG(16));
1606: PetscViewerASCIIPrintf(viewer," INFOG(17) (estimated size of all MUMPS internal data for factorization after analysis: sum over all processors): %d \n",mumps->id.INFOG(17));
1607: PetscViewerASCIIPrintf(viewer," INFOG(18) (size of all MUMPS internal data allocated during factorization: value on the most memory consuming processor): %d \n",mumps->id.INFOG(18));
1608: PetscViewerASCIIPrintf(viewer," INFOG(19) (size of all MUMPS internal data allocated during factorization: sum over all processors): %d \n",mumps->id.INFOG(19));
1609: PetscViewerASCIIPrintf(viewer," INFOG(20) (estimated number of entries in the factors): %d \n",mumps->id.INFOG(20));
1610: PetscViewerASCIIPrintf(viewer," INFOG(21) (size in MB of memory effectively used during factorization - value on the most memory consuming processor): %d \n",mumps->id.INFOG(21));
1611: PetscViewerASCIIPrintf(viewer," INFOG(22) (size in MB of memory effectively used during factorization - sum over all processors): %d \n",mumps->id.INFOG(22));
1612: PetscViewerASCIIPrintf(viewer," INFOG(23) (after analysis: value of ICNTL(6) effectively used): %d \n",mumps->id.INFOG(23));
1613: PetscViewerASCIIPrintf(viewer," INFOG(24) (after analysis: value of ICNTL(12) effectively used): %d \n",mumps->id.INFOG(24));
1614: PetscViewerASCIIPrintf(viewer," INFOG(25) (after factorization: number of pivots modified by static pivoting): %d \n",mumps->id.INFOG(25));
1615: PetscViewerASCIIPrintf(viewer," INFOG(28) (after factorization: number of null pivots encountered): %d\n",mumps->id.INFOG(28));
1616: PetscViewerASCIIPrintf(viewer," INFOG(29) (after factorization: effective number of entries in the factors (sum over all processors)): %d\n",mumps->id.INFOG(29));
1617: PetscViewerASCIIPrintf(viewer," INFOG(30, 31) (after solution: size in Mbytes of memory used during solution phase): %d, %d\n",mumps->id.INFOG(30),mumps->id.INFOG(31));
1618: PetscViewerASCIIPrintf(viewer," INFOG(32) (after analysis: type of analysis done): %d\n",mumps->id.INFOG(32));
1619: PetscViewerASCIIPrintf(viewer," INFOG(33) (value used for ICNTL(8)): %d\n",mumps->id.INFOG(33));
1620: PetscViewerASCIIPrintf(viewer," INFOG(34) (exponent of the determinant if determinant is requested): %d\n",mumps->id.INFOG(34));
1621: }
1622: }
1623: }
1624: return(0);
1625: }
1627: PetscErrorCode MatGetInfo_MUMPS(Mat A,MatInfoType flag,MatInfo *info)
1628: {
1629: Mat_MUMPS *mumps =(Mat_MUMPS*)A->data;
1632: info->block_size = 1.0;
1633: info->nz_allocated = mumps->id.INFOG(20);
1634: info->nz_used = mumps->id.INFOG(20);
1635: info->nz_unneeded = 0.0;
1636: info->assemblies = 0.0;
1637: info->mallocs = 0.0;
1638: info->memory = 0.0;
1639: info->fill_ratio_given = 0;
1640: info->fill_ratio_needed = 0;
1641: info->factor_mallocs = 0;
1642: return(0);
1643: }
1645: /* -------------------------------------------------------------------------------------------*/
1646: PetscErrorCode MatFactorSetSchurIS_MUMPS(Mat F, IS is)
1647: {
1648: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1649: const PetscInt *idxs;
1650: PetscInt size,i;
1654: ISGetLocalSize(is,&size);
1655: if (mumps->size > 1) {
1656: PetscBool ls,gs;
1658: ls = mumps->myid ? (size ? PETSC_FALSE : PETSC_TRUE) : PETSC_TRUE;
1659: MPI_Allreduce(&ls,&gs,1,MPIU_BOOL,MPI_LAND,mumps->comm_mumps);
1660: if (!gs) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"MUMPS distributed parallel Schur complements not yet supported from PETSc\n");
1661: }
1662: if (mumps->id.size_schur != size) {
1663: PetscFree2(mumps->id.listvar_schur,mumps->id.schur);
1664: mumps->id.size_schur = size;
1665: mumps->id.schur_lld = size;
1666: PetscMalloc2(size,&mumps->id.listvar_schur,size*size,&mumps->id.schur);
1667: }
1669: /* Schur complement matrix */
1670: MatCreateSeqDense(PETSC_COMM_SELF,mumps->id.size_schur,mumps->id.size_schur,(PetscScalar*)mumps->id.schur,&F->schur);
1671: if (mumps->sym == 1) {
1672: MatSetOption(F->schur,MAT_SPD,PETSC_TRUE);
1673: }
1675: /* MUMPS expects Fortran style indices */
1676: ISGetIndices(is,&idxs);
1677: PetscMemcpy(mumps->id.listvar_schur,idxs,size*sizeof(PetscInt));
1678: for (i=0;i<size;i++) mumps->id.listvar_schur[i]++;
1679: ISRestoreIndices(is,&idxs);
1680: if (mumps->size > 1) {
1681: mumps->id.ICNTL(19) = 1; /* MUMPS returns Schur centralized on the host */
1682: } else {
1683: if (F->factortype == MAT_FACTOR_LU) {
1684: mumps->id.ICNTL(19) = 3; /* MUMPS returns full matrix */
1685: } else {
1686: mumps->id.ICNTL(19) = 2; /* MUMPS returns lower triangular part */
1687: }
1688: }
1689: /* set a special value of ICNTL (not handled my MUMPS) to be used in the solve phase by PETSc */
1690: mumps->id.ICNTL(26) = -1;
1691: return(0);
1692: }
1694: /* -------------------------------------------------------------------------------------------*/
1695: PetscErrorCode MatFactorCreateSchurComplement_MUMPS(Mat F,Mat* S)
1696: {
1697: Mat St;
1698: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1699: PetscScalar *array;
1700: #if defined(PETSC_USE_COMPLEX)
1701: PetscScalar im = PetscSqrtScalar((PetscScalar)-1.0);
1702: #endif
1706: if (!mumps->id.ICNTL(19)) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ORDER,"Schur complement mode not selected! You should call MatFactorSetSchurIS to enable it");
1707: MatCreate(PETSC_COMM_SELF,&St);
1708: MatSetSizes(St,PETSC_DECIDE,PETSC_DECIDE,mumps->id.size_schur,mumps->id.size_schur);
1709: MatSetType(St,MATDENSE);
1710: MatSetUp(St);
1711: MatDenseGetArray(St,&array);
1712: if (!mumps->sym) { /* MUMPS always return a full matrix */
1713: if (mumps->id.ICNTL(19) == 1) { /* stored by rows */
1714: PetscInt i,j,N=mumps->id.size_schur;
1715: for (i=0;i<N;i++) {
1716: for (j=0;j<N;j++) {
1717: #if !defined(PETSC_USE_COMPLEX)
1718: PetscScalar val = mumps->id.schur[i*N+j];
1719: #else
1720: PetscScalar val = mumps->id.schur[i*N+j].r + im*mumps->id.schur[i*N+j].i;
1721: #endif
1722: array[j*N+i] = val;
1723: }
1724: }
1725: } else { /* stored by columns */
1726: PetscMemcpy(array,mumps->id.schur,mumps->id.size_schur*mumps->id.size_schur*sizeof(PetscScalar));
1727: }
1728: } else { /* either full or lower-triangular (not packed) */
1729: if (mumps->id.ICNTL(19) == 2) { /* lower triangular stored by columns */
1730: PetscInt i,j,N=mumps->id.size_schur;
1731: for (i=0;i<N;i++) {
1732: for (j=i;j<N;j++) {
1733: #if !defined(PETSC_USE_COMPLEX)
1734: PetscScalar val = mumps->id.schur[i*N+j];
1735: #else
1736: PetscScalar val = mumps->id.schur[i*N+j].r + im*mumps->id.schur[i*N+j].i;
1737: #endif
1738: array[i*N+j] = val;
1739: array[j*N+i] = val;
1740: }
1741: }
1742: } else if (mumps->id.ICNTL(19) == 3) { /* full matrix */
1743: PetscMemcpy(array,mumps->id.schur,mumps->id.size_schur*mumps->id.size_schur*sizeof(PetscScalar));
1744: } else { /* ICNTL(19) == 1 lower triangular stored by rows */
1745: PetscInt i,j,N=mumps->id.size_schur;
1746: for (i=0;i<N;i++) {
1747: for (j=0;j<i+1;j++) {
1748: #if !defined(PETSC_USE_COMPLEX)
1749: PetscScalar val = mumps->id.schur[i*N+j];
1750: #else
1751: PetscScalar val = mumps->id.schur[i*N+j].r + im*mumps->id.schur[i*N+j].i;
1752: #endif
1753: array[i*N+j] = val;
1754: array[j*N+i] = val;
1755: }
1756: }
1757: }
1758: }
1759: MatDenseRestoreArray(St,&array);
1760: *S = St;
1761: return(0);
1762: }
1764: /* -------------------------------------------------------------------------------------------*/
1765: PetscErrorCode MatMumpsSetIcntl_MUMPS(Mat F,PetscInt icntl,PetscInt ival)
1766: {
1767: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1770: mumps->id.ICNTL(icntl) = ival;
1771: return(0);
1772: }
1774: PetscErrorCode MatMumpsGetIcntl_MUMPS(Mat F,PetscInt icntl,PetscInt *ival)
1775: {
1776: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1779: *ival = mumps->id.ICNTL(icntl);
1780: return(0);
1781: }
1783: /*@
1784: MatMumpsSetIcntl - Set MUMPS parameter ICNTL()
1786: Logically Collective on Mat
1788: Input Parameters:
1789: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
1790: . icntl - index of MUMPS parameter array ICNTL()
1791: - ival - value of MUMPS ICNTL(icntl)
1793: Options Database:
1794: . -mat_mumps_icntl_<icntl> <ival>
1796: Level: beginner
1798: References:
1799: . MUMPS Users' Guide
1801: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
1802: @*/
1803: PetscErrorCode MatMumpsSetIcntl(Mat F,PetscInt icntl,PetscInt ival)
1804: {
1806:
1809: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
1812: PetscTryMethod(F,"MatMumpsSetIcntl_C",(Mat,PetscInt,PetscInt),(F,icntl,ival));
1813: return(0);
1814: }
1816: /*@
1817: MatMumpsGetIcntl - Get MUMPS parameter ICNTL()
1819: Logically Collective on Mat
1821: Input Parameters:
1822: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
1823: - icntl - index of MUMPS parameter array ICNTL()
1825: Output Parameter:
1826: . ival - value of MUMPS ICNTL(icntl)
1828: Level: beginner
1830: References:
1831: . MUMPS Users' Guide
1833: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
1834: @*/
1835: PetscErrorCode MatMumpsGetIcntl(Mat F,PetscInt icntl,PetscInt *ival)
1836: {
1841: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
1844: PetscUseMethod(F,"MatMumpsGetIcntl_C",(Mat,PetscInt,PetscInt*),(F,icntl,ival));
1845: return(0);
1846: }
1848: /* -------------------------------------------------------------------------------------------*/
1849: PetscErrorCode MatMumpsSetCntl_MUMPS(Mat F,PetscInt icntl,PetscReal val)
1850: {
1851: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1854: mumps->id.CNTL(icntl) = val;
1855: return(0);
1856: }
1858: PetscErrorCode MatMumpsGetCntl_MUMPS(Mat F,PetscInt icntl,PetscReal *val)
1859: {
1860: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1863: *val = mumps->id.CNTL(icntl);
1864: return(0);
1865: }
1867: /*@
1868: MatMumpsSetCntl - Set MUMPS parameter CNTL()
1870: Logically Collective on Mat
1872: Input Parameters:
1873: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
1874: . icntl - index of MUMPS parameter array CNTL()
1875: - val - value of MUMPS CNTL(icntl)
1877: Options Database:
1878: . -mat_mumps_cntl_<icntl> <val>
1880: Level: beginner
1882: References:
1883: . MUMPS Users' Guide
1885: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
1886: @*/
1887: PetscErrorCode MatMumpsSetCntl(Mat F,PetscInt icntl,PetscReal val)
1888: {
1893: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
1896: PetscTryMethod(F,"MatMumpsSetCntl_C",(Mat,PetscInt,PetscReal),(F,icntl,val));
1897: return(0);
1898: }
1900: /*@
1901: MatMumpsGetCntl - Get MUMPS parameter CNTL()
1903: Logically Collective on Mat
1905: Input Parameters:
1906: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
1907: - icntl - index of MUMPS parameter array CNTL()
1909: Output Parameter:
1910: . val - value of MUMPS CNTL(icntl)
1912: Level: beginner
1914: References:
1915: . MUMPS Users' Guide
1917: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
1918: @*/
1919: PetscErrorCode MatMumpsGetCntl(Mat F,PetscInt icntl,PetscReal *val)
1920: {
1925: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
1928: PetscUseMethod(F,"MatMumpsGetCntl_C",(Mat,PetscInt,PetscReal*),(F,icntl,val));
1929: return(0);
1930: }
1932: PetscErrorCode MatMumpsGetInfo_MUMPS(Mat F,PetscInt icntl,PetscInt *info)
1933: {
1934: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1937: *info = mumps->id.INFO(icntl);
1938: return(0);
1939: }
1941: PetscErrorCode MatMumpsGetInfog_MUMPS(Mat F,PetscInt icntl,PetscInt *infog)
1942: {
1943: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1946: *infog = mumps->id.INFOG(icntl);
1947: return(0);
1948: }
1950: PetscErrorCode MatMumpsGetRinfo_MUMPS(Mat F,PetscInt icntl,PetscReal *rinfo)
1951: {
1952: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1955: *rinfo = mumps->id.RINFO(icntl);
1956: return(0);
1957: }
1959: PetscErrorCode MatMumpsGetRinfog_MUMPS(Mat F,PetscInt icntl,PetscReal *rinfog)
1960: {
1961: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1964: *rinfog = mumps->id.RINFOG(icntl);
1965: return(0);
1966: }
1968: PetscErrorCode MatMumpsGetInverse_MUMPS(Mat F,Mat spRHS)
1969: {
1971: Mat Bt = NULL;
1972: PetscBool flgT;
1973: Mat_MUMPS *mumps =(Mat_MUMPS*)F->data;
1974: PetscBool done;
1975: PetscScalar *aa;
1976: PetscInt spnr,*ia,*ja;
1979: if (!mumps->myid) {
1981: PetscObjectTypeCompare((PetscObject)spRHS,MATTRANSPOSEMAT,&flgT);
1982: if (flgT) {
1983: MatTransposeGetMat(spRHS,&Bt);
1984: } else {
1985: SETERRQ(PetscObjectComm((PetscObject)spRHS),PETSC_ERR_ARG_WRONG,"Matrix spRHS must be type MATTRANSPOSEMAT matrix");
1986: }
1987: }
1989: MatMumpsSetIcntl(F,30,1);
1991: if (!mumps->myid) {
1992: MatSeqAIJGetArray(Bt,&aa);
1993: MatGetRowIJ(Bt,1,PETSC_FALSE,PETSC_FALSE,&spnr,(const PetscInt**)&ia,(const PetscInt**)&ja,&done);
1994: if (!done) SETERRQ(PetscObjectComm((PetscObject)Bt),PETSC_ERR_ARG_WRONG,"Cannot get IJ structure");
1996: mumps->id.irhs_ptr = ia;
1997: mumps->id.irhs_sparse = ja;
1998: mumps->id.nz_rhs = ia[spnr] - 1;
1999: mumps->id.rhs_sparse = (MumpsScalar*)aa;
2000: } else {
2001: mumps->id.irhs_ptr = NULL;
2002: mumps->id.irhs_sparse = NULL;
2003: mumps->id.nz_rhs = 0;
2004: mumps->id.rhs_sparse = NULL;
2005: }
2006: mumps->id.ICNTL(20) = 1; /* rhs is sparse */
2007: mumps->id.ICNTL(21) = 0; /* solution is in assembled centralized format */
2009: /* solve phase */
2010: /*-------------*/
2011: mumps->id.job = JOB_SOLVE;
2012: PetscMUMPS_c(&mumps->id);
2013: if (mumps->id.INFOG(1) < 0)
2014: SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_LIB,"Error reported by MUMPS in solve phase: INFOG(1)=%d INFO(2)=%d\n",mumps->id.INFOG(1),mumps->id.INFO(2));
2016: if (!mumps->myid) {
2017: MatSeqAIJRestoreArray(Bt,&aa);
2018: MatRestoreRowIJ(Bt,1,PETSC_FALSE,PETSC_FALSE,&spnr,(const PetscInt**)&ia,(const PetscInt**)&ja,&done);
2019: }
2020: return(0);
2021: }
2023: /*@
2024: MatMumpsGetInverse - Get user-specified set of entries in inverse of A
2026: Logically Collective on Mat
2028: Input Parameters:
2029: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
2030: - spRHS - sequential sparse matrix in MATTRANSPOSEMAT format holding specified indices in processor[0]
2032: Output Parameter:
2033: . spRHS - requested entries of inverse of A
2035: Level: beginner
2037: References:
2038: . MUMPS Users' Guide
2040: .seealso: MatGetFactor(), MatCreateTranspose()
2041: @*/
2042: PetscErrorCode MatMumpsGetInverse(Mat F,Mat spRHS)
2043: {
2048: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
2049: PetscUseMethod(F,"MatMumpsGetInverse_C",(Mat,Mat),(F,spRHS));
2050: return(0);
2051: }
2053: /*@
2054: MatMumpsGetInfo - Get MUMPS parameter INFO()
2056: Logically Collective on Mat
2058: Input Parameters:
2059: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
2060: - icntl - index of MUMPS parameter array INFO()
2062: Output Parameter:
2063: . ival - value of MUMPS INFO(icntl)
2065: Level: beginner
2067: References:
2068: . MUMPS Users' Guide
2070: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
2071: @*/
2072: PetscErrorCode MatMumpsGetInfo(Mat F,PetscInt icntl,PetscInt *ival)
2073: {
2078: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
2080: PetscUseMethod(F,"MatMumpsGetInfo_C",(Mat,PetscInt,PetscInt*),(F,icntl,ival));
2081: return(0);
2082: }
2084: /*@
2085: MatMumpsGetInfog - Get MUMPS parameter INFOG()
2087: Logically Collective on Mat
2089: Input Parameters:
2090: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
2091: - icntl - index of MUMPS parameter array INFOG()
2093: Output Parameter:
2094: . ival - value of MUMPS INFOG(icntl)
2096: Level: beginner
2098: References:
2099: . MUMPS Users' Guide
2101: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
2102: @*/
2103: PetscErrorCode MatMumpsGetInfog(Mat F,PetscInt icntl,PetscInt *ival)
2104: {
2109: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
2111: PetscUseMethod(F,"MatMumpsGetInfog_C",(Mat,PetscInt,PetscInt*),(F,icntl,ival));
2112: return(0);
2113: }
2115: /*@
2116: MatMumpsGetRinfo - Get MUMPS parameter RINFO()
2118: Logically Collective on Mat
2120: Input Parameters:
2121: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
2122: - icntl - index of MUMPS parameter array RINFO()
2124: Output Parameter:
2125: . val - value of MUMPS RINFO(icntl)
2127: Level: beginner
2129: References:
2130: . MUMPS Users' Guide
2132: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
2133: @*/
2134: PetscErrorCode MatMumpsGetRinfo(Mat F,PetscInt icntl,PetscReal *val)
2135: {
2140: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
2142: PetscUseMethod(F,"MatMumpsGetRinfo_C",(Mat,PetscInt,PetscReal*),(F,icntl,val));
2143: return(0);
2144: }
2146: /*@
2147: MatMumpsGetRinfog - Get MUMPS parameter RINFOG()
2149: Logically Collective on Mat
2151: Input Parameters:
2152: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
2153: - icntl - index of MUMPS parameter array RINFOG()
2155: Output Parameter:
2156: . val - value of MUMPS RINFOG(icntl)
2158: Level: beginner
2160: References:
2161: . MUMPS Users' Guide
2163: .seealso: MatGetFactor(), MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog()
2164: @*/
2165: PetscErrorCode MatMumpsGetRinfog(Mat F,PetscInt icntl,PetscReal *val)
2166: {
2171: if (!F->factortype) SETERRQ(PetscObjectComm((PetscObject)F),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
2173: PetscUseMethod(F,"MatMumpsGetRinfog_C",(Mat,PetscInt,PetscReal*),(F,icntl,val));
2174: return(0);
2175: }
2177: /*MC
2178: MATSOLVERMUMPS - A matrix type providing direct solvers (LU and Cholesky) for
2179: distributed and sequential matrices via the external package MUMPS.
2181: Works with MATAIJ and MATSBAIJ matrices
2183: Use ./configure --download-mumps --download-scalapack --download-parmetis --download-metis --download-ptscotch to have PETSc installed with MUMPS
2185: Use -pc_type cholesky or lu -pc_factor_mat_solver_type mumps to use this direct solver
2187: Options Database Keys:
2188: + -mat_mumps_icntl_1 - ICNTL(1): output stream for error messages
2189: . -mat_mumps_icntl_2 - ICNTL(2): output stream for diagnostic printing, statistics, and warning
2190: . -mat_mumps_icntl_3 - ICNTL(3): output stream for global information, collected on the host
2191: . -mat_mumps_icntl_4 - ICNTL(4): level of printing (0 to 4)
2192: . -mat_mumps_icntl_6 - ICNTL(6): permutes to a zero-free diagonal and/or scale the matrix (0 to 7)
2193: . -mat_mumps_icntl_7 - ICNTL(7): computes a symmetric permutation in sequential analysis (0 to 7). 3=Scotch, 4=PORD, 5=Metis
2194: . -mat_mumps_icntl_8 - ICNTL(8): scaling strategy (-2 to 8 or 77)
2195: . -mat_mumps_icntl_10 - ICNTL(10): max num of refinements
2196: . -mat_mumps_icntl_11 - ICNTL(11): statistics related to an error analysis (via -ksp_view)
2197: . -mat_mumps_icntl_12 - ICNTL(12): an ordering strategy for symmetric matrices (0 to 3)
2198: . -mat_mumps_icntl_13 - ICNTL(13): parallelism of the root node (enable ScaLAPACK) and its splitting
2199: . -mat_mumps_icntl_14 - ICNTL(14): percentage increase in the estimated working space
2200: . -mat_mumps_icntl_19 - ICNTL(19): computes the Schur complement
2201: . -mat_mumps_icntl_22 - ICNTL(22): in-core/out-of-core factorization and solve (0 or 1)
2202: . -mat_mumps_icntl_23 - ICNTL(23): max size of the working memory (MB) that can allocate per processor
2203: . -mat_mumps_icntl_24 - ICNTL(24): detection of null pivot rows (0 or 1)
2204: . -mat_mumps_icntl_25 - ICNTL(25): compute a solution of a deficient matrix and a null space basis
2205: . -mat_mumps_icntl_26 - ICNTL(26): drives the solution phase if a Schur complement matrix
2206: . -mat_mumps_icntl_28 - ICNTL(28): use 1 for sequential analysis and ictnl(7) ordering, or 2 for parallel analysis and ictnl(29) ordering
2207: . -mat_mumps_icntl_29 - ICNTL(29): parallel ordering 1 = ptscotch, 2 = parmetis
2208: . -mat_mumps_icntl_30 - ICNTL(30): compute user-specified set of entries in inv(A)
2209: . -mat_mumps_icntl_31 - ICNTL(31): indicates which factors may be discarded during factorization
2210: . -mat_mumps_icntl_33 - ICNTL(33): compute determinant
2211: . -mat_mumps_cntl_1 - CNTL(1): relative pivoting threshold
2212: . -mat_mumps_cntl_2 - CNTL(2): stopping criterion of refinement
2213: . -mat_mumps_cntl_3 - CNTL(3): absolute pivoting threshold
2214: . -mat_mumps_cntl_4 - CNTL(4): value for static pivoting
2215: - -mat_mumps_cntl_5 - CNTL(5): fixation for null pivots
2217: Level: beginner
2219: Notes: When a MUMPS factorization fails inside a KSP solve, for example with a KSP_DIVERGED_PCSETUP_FAILED, one can find the MUMPS information about the failure by calling
2220: $ KSPGetPC(ksp,&pc);
2221: $ PCFactorGetMatrix(pc,&mat);
2222: $ MatMumpsGetInfo(mat,....);
2223: $ MatMumpsGetInfog(mat,....); etc.
2224: Or you can run with -ksp_error_if_not_converged and the program will be stopped and the information printed in the error message.
2226: .seealso: PCFactorSetMatSolverType(), MatSolverType, MatMumpsSetICntl(), MatMumpsGetIcntl(), MatMumpsSetCntl(), MatMumpsGetCntl(), MatMumpsGetInfo(), MatMumpsGetInfog(), MatMumpsGetRinfo(), MatMumpsGetRinfog(), KSPGetPC(), PCGetFactor(), PCFactorGetMatrix()
2228: M*/
2230: static PetscErrorCode MatFactorGetSolverType_mumps(Mat A,MatSolverType *type)
2231: {
2233: *type = MATSOLVERMUMPS;
2234: return(0);
2235: }
2237: /* MatGetFactor for Seq and MPI AIJ matrices */
2238: static PetscErrorCode MatGetFactor_aij_mumps(Mat A,MatFactorType ftype,Mat *F)
2239: {
2240: Mat B;
2242: Mat_MUMPS *mumps;
2243: PetscBool isSeqAIJ;
2246: /* Create the factorization matrix */
2247: PetscObjectTypeCompare((PetscObject)A,MATSEQAIJ,&isSeqAIJ);
2248: MatCreate(PetscObjectComm((PetscObject)A),&B);
2249: MatSetSizes(B,A->rmap->n,A->cmap->n,A->rmap->N,A->cmap->N);
2250: PetscStrallocpy("mumps",&((PetscObject)B)->type_name);
2251: MatSetUp(B);
2253: PetscNewLog(B,&mumps);
2255: B->ops->view = MatView_MUMPS;
2256: B->ops->getinfo = MatGetInfo_MUMPS;
2258: PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverType_C",MatFactorGetSolverType_mumps);
2259: PetscObjectComposeFunction((PetscObject)B,"MatFactorSetSchurIS_C",MatFactorSetSchurIS_MUMPS);
2260: PetscObjectComposeFunction((PetscObject)B,"MatFactorCreateSchurComplement_C",MatFactorCreateSchurComplement_MUMPS);
2261: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetIcntl_C",MatMumpsSetIcntl_MUMPS);
2262: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetIcntl_C",MatMumpsGetIcntl_MUMPS);
2263: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetCntl_C",MatMumpsSetCntl_MUMPS);
2264: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetCntl_C",MatMumpsGetCntl_MUMPS);
2265: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfo_C",MatMumpsGetInfo_MUMPS);
2266: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfog_C",MatMumpsGetInfog_MUMPS);
2267: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfo_C",MatMumpsGetRinfo_MUMPS);
2268: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfog_C",MatMumpsGetRinfog_MUMPS);
2269: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInverse_C",MatMumpsGetInverse_MUMPS);
2271: if (ftype == MAT_FACTOR_LU) {
2272: B->ops->lufactorsymbolic = MatLUFactorSymbolic_AIJMUMPS;
2273: B->factortype = MAT_FACTOR_LU;
2274: if (isSeqAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqaij_seqaij;
2275: else mumps->ConvertToTriples = MatConvertToTriples_mpiaij_mpiaij;
2276: mumps->sym = 0;
2277: } else {
2278: B->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_MUMPS;
2279: B->factortype = MAT_FACTOR_CHOLESKY;
2280: if (isSeqAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqaij_seqsbaij;
2281: else mumps->ConvertToTriples = MatConvertToTriples_mpiaij_mpisbaij;
2282: #if defined(PETSC_USE_COMPLEX)
2283: mumps->sym = 2;
2284: #else
2285: if (A->spd_set && A->spd) mumps->sym = 1;
2286: else mumps->sym = 2;
2287: #endif
2288: }
2290: /* set solvertype */
2291: PetscFree(B->solvertype);
2292: PetscStrallocpy(MATSOLVERMUMPS,&B->solvertype);
2294: B->ops->destroy = MatDestroy_MUMPS;
2295: B->data = (void*)mumps;
2297: PetscInitializeMUMPS(A,mumps);
2299: *F = B;
2300: return(0);
2301: }
2303: /* MatGetFactor for Seq and MPI SBAIJ matrices */
2304: static PetscErrorCode MatGetFactor_sbaij_mumps(Mat A,MatFactorType ftype,Mat *F)
2305: {
2306: Mat B;
2308: Mat_MUMPS *mumps;
2309: PetscBool isSeqSBAIJ;
2312: if (ftype != MAT_FACTOR_CHOLESKY) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Cannot use PETSc SBAIJ matrices with MUMPS LU, use AIJ matrix");
2313: if (A->rmap->bs > 1) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Cannot use PETSc SBAIJ matrices with block size > 1 with MUMPS Cholesky, use AIJ matrix instead");
2314: PetscObjectTypeCompare((PetscObject)A,MATSEQSBAIJ,&isSeqSBAIJ);
2315: /* Create the factorization matrix */
2316: MatCreate(PetscObjectComm((PetscObject)A),&B);
2317: MatSetSizes(B,A->rmap->n,A->cmap->n,A->rmap->N,A->cmap->N);
2318: PetscStrallocpy("mumps",&((PetscObject)B)->type_name);
2319: MatSetUp(B);
2321: PetscNewLog(B,&mumps);
2322: if (isSeqSBAIJ) {
2323: mumps->ConvertToTriples = MatConvertToTriples_seqsbaij_seqsbaij;
2324: } else {
2325: mumps->ConvertToTriples = MatConvertToTriples_mpisbaij_mpisbaij;
2326: }
2328: B->ops->getinfo = MatGetInfo_External;
2329: B->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_MUMPS;
2330: B->ops->view = MatView_MUMPS;
2332: PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverType_C",MatFactorGetSolverType_mumps);
2333: PetscObjectComposeFunction((PetscObject)B,"MatFactorSetSchurIS_C",MatFactorSetSchurIS_MUMPS);
2334: PetscObjectComposeFunction((PetscObject)B,"MatFactorCreateSchurComplement_C",MatFactorCreateSchurComplement_MUMPS);
2335: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetIcntl_C",MatMumpsSetIcntl_MUMPS);
2336: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetIcntl_C",MatMumpsGetIcntl_MUMPS);
2337: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetCntl_C",MatMumpsSetCntl_MUMPS);
2338: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetCntl_C",MatMumpsGetCntl_MUMPS);
2339: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfo_C",MatMumpsGetInfo_MUMPS);
2340: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfog_C",MatMumpsGetInfog_MUMPS);
2341: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfo_C",MatMumpsGetRinfo_MUMPS);
2342: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfog_C",MatMumpsGetRinfog_MUMPS);
2343: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInverse_C",MatMumpsGetInverse_MUMPS);
2345: B->factortype = MAT_FACTOR_CHOLESKY;
2346: #if defined(PETSC_USE_COMPLEX)
2347: mumps->sym = 2;
2348: #else
2349: if (A->spd_set && A->spd) mumps->sym = 1;
2350: else mumps->sym = 2;
2351: #endif
2353: /* set solvertype */
2354: PetscFree(B->solvertype);
2355: PetscStrallocpy(MATSOLVERMUMPS,&B->solvertype);
2357: B->ops->destroy = MatDestroy_MUMPS;
2358: B->data = (void*)mumps;
2360: PetscInitializeMUMPS(A,mumps);
2362: *F = B;
2363: return(0);
2364: }
2366: static PetscErrorCode MatGetFactor_baij_mumps(Mat A,MatFactorType ftype,Mat *F)
2367: {
2368: Mat B;
2370: Mat_MUMPS *mumps;
2371: PetscBool isSeqBAIJ;
2374: /* Create the factorization matrix */
2375: PetscObjectTypeCompare((PetscObject)A,MATSEQBAIJ,&isSeqBAIJ);
2376: MatCreate(PetscObjectComm((PetscObject)A),&B);
2377: MatSetSizes(B,A->rmap->n,A->cmap->n,A->rmap->N,A->cmap->N);
2378: PetscStrallocpy("mumps",&((PetscObject)B)->type_name);
2379: MatSetUp(B);
2381: PetscNewLog(B,&mumps);
2382: if (ftype == MAT_FACTOR_LU) {
2383: B->ops->lufactorsymbolic = MatLUFactorSymbolic_BAIJMUMPS;
2384: B->factortype = MAT_FACTOR_LU;
2385: if (isSeqBAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqbaij_seqaij;
2386: else mumps->ConvertToTriples = MatConvertToTriples_mpibaij_mpiaij;
2387: mumps->sym = 0;
2388: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Cannot use PETSc BAIJ matrices with MUMPS Cholesky, use SBAIJ or AIJ matrix instead\n");
2390: B->ops->getinfo = MatGetInfo_External;
2391: B->ops->view = MatView_MUMPS;
2393: PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverType_C",MatFactorGetSolverType_mumps);
2394: PetscObjectComposeFunction((PetscObject)B,"MatFactorSetSchurIS_C",MatFactorSetSchurIS_MUMPS);
2395: PetscObjectComposeFunction((PetscObject)B,"MatFactorCreateSchurComplement_C",MatFactorCreateSchurComplement_MUMPS);
2396: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetIcntl_C",MatMumpsSetIcntl_MUMPS);
2397: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetIcntl_C",MatMumpsGetIcntl_MUMPS);
2398: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetCntl_C",MatMumpsSetCntl_MUMPS);
2399: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetCntl_C",MatMumpsGetCntl_MUMPS);
2400: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfo_C",MatMumpsGetInfo_MUMPS);
2401: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfog_C",MatMumpsGetInfog_MUMPS);
2402: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfo_C",MatMumpsGetRinfo_MUMPS);
2403: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfog_C",MatMumpsGetRinfog_MUMPS);
2404: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInverse_C",MatMumpsGetInverse_MUMPS);
2406: /* set solvertype */
2407: PetscFree(B->solvertype);
2408: PetscStrallocpy(MATSOLVERMUMPS,&B->solvertype);
2410: B->ops->destroy = MatDestroy_MUMPS;
2411: B->data = (void*)mumps;
2413: PetscInitializeMUMPS(A,mumps);
2415: *F = B;
2416: return(0);
2417: }
2419: /* MatGetFactor for Seq and MPI SELL matrices */
2420: static PetscErrorCode MatGetFactor_sell_mumps(Mat A,MatFactorType ftype,Mat *F)
2421: {
2422: Mat B;
2424: Mat_MUMPS *mumps;
2425: PetscBool isSeqSELL;
2428: /* Create the factorization matrix */
2429: PetscObjectTypeCompare((PetscObject)A,MATSEQSELL,&isSeqSELL);
2430: MatCreate(PetscObjectComm((PetscObject)A),&B);
2431: MatSetSizes(B,A->rmap->n,A->cmap->n,A->rmap->N,A->cmap->N);
2432: PetscStrallocpy("mumps",&((PetscObject)B)->type_name);
2433: MatSetUp(B);
2435: PetscNewLog(B,&mumps);
2437: B->ops->view = MatView_MUMPS;
2438: B->ops->getinfo = MatGetInfo_MUMPS;
2440: PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverType_C",MatFactorGetSolverType_mumps);
2441: PetscObjectComposeFunction((PetscObject)B,"MatFactorSetSchurIS_C",MatFactorSetSchurIS_MUMPS);
2442: PetscObjectComposeFunction((PetscObject)B,"MatFactorCreateSchurComplement_C",MatFactorCreateSchurComplement_MUMPS);
2443: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetIcntl_C",MatMumpsSetIcntl_MUMPS);
2444: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetIcntl_C",MatMumpsGetIcntl_MUMPS);
2445: PetscObjectComposeFunction((PetscObject)B,"MatMumpsSetCntl_C",MatMumpsSetCntl_MUMPS);
2446: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetCntl_C",MatMumpsGetCntl_MUMPS);
2447: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfo_C",MatMumpsGetInfo_MUMPS);
2448: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetInfog_C",MatMumpsGetInfog_MUMPS);
2449: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfo_C",MatMumpsGetRinfo_MUMPS);
2450: PetscObjectComposeFunction((PetscObject)B,"MatMumpsGetRinfog_C",MatMumpsGetRinfog_MUMPS);
2452: if (ftype == MAT_FACTOR_LU) {
2453: B->ops->lufactorsymbolic = MatLUFactorSymbolic_AIJMUMPS;
2454: B->factortype = MAT_FACTOR_LU;
2455: if (isSeqSELL) mumps->ConvertToTriples = MatConvertToTriples_seqsell_seqaij;
2456: else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"To be implemented");
2457: mumps->sym = 0;
2458: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"To be implemented");
2460: /* set solvertype */
2461: PetscFree(B->solvertype);
2462: PetscStrallocpy(MATSOLVERMUMPS,&B->solvertype);
2464: B->ops->destroy = MatDestroy_MUMPS;
2465: B->data = (void*)mumps;
2467: PetscInitializeMUMPS(A,mumps);
2469: *F = B;
2470: return(0);
2471: }
2473: PETSC_EXTERN PetscErrorCode MatSolverTypeRegister_MUMPS(void)
2474: {
2478: MatSolverTypeRegister(MATSOLVERMUMPS,MATMPIAIJ,MAT_FACTOR_LU,MatGetFactor_aij_mumps);
2479: MatSolverTypeRegister(MATSOLVERMUMPS,MATMPIAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_aij_mumps);
2480: MatSolverTypeRegister(MATSOLVERMUMPS,MATMPIBAIJ,MAT_FACTOR_LU,MatGetFactor_baij_mumps);
2481: MatSolverTypeRegister(MATSOLVERMUMPS,MATMPIBAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_baij_mumps);
2482: MatSolverTypeRegister(MATSOLVERMUMPS,MATMPISBAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_sbaij_mumps);
2483: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQAIJ,MAT_FACTOR_LU,MatGetFactor_aij_mumps);
2484: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_aij_mumps);
2485: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQBAIJ,MAT_FACTOR_LU,MatGetFactor_baij_mumps);
2486: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQBAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_baij_mumps);
2487: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQSBAIJ,MAT_FACTOR_CHOLESKY,MatGetFactor_sbaij_mumps);
2488: MatSolverTypeRegister(MATSOLVERMUMPS,MATSEQSELL,MAT_FACTOR_LU,MatGetFactor_sell_mumps);
2489: return(0);
2490: }