Actual source code: mkl_cpardiso.c


  2: #include <petscsys.h>
  3: #include <../src/mat/impls/aij/mpi/mpiaij.h>
  4: #include <../src/mat/impls/sbaij/mpi/mpisbaij.h>

  6: #if defined(PETSC_HAVE_MKL_INTEL_ILP64)
  7:   #define MKL_ILP64
  8: #endif
  9: #include <mkl.h>
 10: #include <mkl_cluster_sparse_solver.h>

 12: /*
 13:  *  Possible mkl_cpardiso phases that controls the execution of the solver.
 14:  *  For more information check mkl_cpardiso manual.
 15:  */
 16: #define JOB_ANALYSIS                                                    11
 17: #define JOB_ANALYSIS_NUMERICAL_FACTORIZATION                            12
 18: #define JOB_ANALYSIS_NUMERICAL_FACTORIZATION_SOLVE_ITERATIVE_REFINEMENT 13
 19: #define JOB_NUMERICAL_FACTORIZATION                                     22
 20: #define JOB_NUMERICAL_FACTORIZATION_SOLVE_ITERATIVE_REFINEMENT          23
 21: #define JOB_SOLVE_ITERATIVE_REFINEMENT                                  33
 22: #define JOB_SOLVE_FORWARD_SUBSTITUTION                                  331
 23: #define JOB_SOLVE_DIAGONAL_SUBSTITUTION                                 332
 24: #define JOB_SOLVE_BACKWARD_SUBSTITUTION                                 333
 25: #define JOB_RELEASE_OF_LU_MEMORY                                        0
 26: #define JOB_RELEASE_OF_ALL_MEMORY                                       -1

 28: #define IPARM_SIZE 64
 29: #define INT_TYPE   MKL_INT

 31: static const char *Err_MSG_CPardiso(int errNo)
 32: {
 33:   switch (errNo) {
 34:   case -1:
 35:     return "input inconsistent";
 36:     break;
 37:   case -2:
 38:     return "not enough memory";
 39:     break;
 40:   case -3:
 41:     return "reordering problem";
 42:     break;
 43:   case -4:
 44:     return "zero pivot, numerical factorization or iterative refinement problem";
 45:     break;
 46:   case -5:
 47:     return "unclassified (internal) error";
 48:     break;
 49:   case -6:
 50:     return "preordering failed (matrix types 11, 13 only)";
 51:     break;
 52:   case -7:
 53:     return "diagonal matrix problem";
 54:     break;
 55:   case -8:
 56:     return "32-bit integer overflow problem";
 57:     break;
 58:   case -9:
 59:     return "not enough memory for OOC";
 60:     break;
 61:   case -10:
 62:     return "problems with opening OOC temporary files";
 63:     break;
 64:   case -11:
 65:     return "read/write problems with the OOC data file";
 66:     break;
 67:   default:
 68:     return "unknown error";
 69:   }
 70: }

 72: /*
 73:  *  Internal data structure.
 74:  *  For more information check mkl_cpardiso manual.
 75:  */

 77: typedef struct {
 78:   /* Configuration vector */
 79:   INT_TYPE iparm[IPARM_SIZE];

 81:   /*
 82:    * Internal mkl_cpardiso memory location.
 83:    * After the first call to mkl_cpardiso do not modify pt, as that could cause a serious memory leak.
 84:    */
 85:   void *pt[IPARM_SIZE];

 87:   MPI_Fint comm_mkl_cpardiso;

 89:   /* Basic mkl_cpardiso info*/
 90:   INT_TYPE phase, maxfct, mnum, mtype, n, nrhs, msglvl, err;

 92:   /* Matrix structure */
 93:   PetscScalar *a;

 95:   INT_TYPE *ia, *ja;

 97:   /* Number of non-zero elements */
 98:   INT_TYPE nz;

100:   /* Row permutaton vector*/
101:   INT_TYPE *perm;

103:   /* Define is matrix preserve sparce structure. */
104:   MatStructure matstruc;

106:   PetscErrorCode (*ConvertToTriples)(Mat, MatReuse, PetscInt *, PetscInt **, PetscInt **, PetscScalar **);

108:   /* True if mkl_cpardiso function have been used. */
109:   PetscBool CleanUp;
110: } Mat_MKL_CPARDISO;

112: /*
113:  * Copy the elements of matrix A.
114:  * Input:
115:  *   - Mat A: MATSEQAIJ matrix
116:  *   - int shift: matrix index.
117:  *     - 0 for c representation
118:  *     - 1 for fortran representation
119:  *   - MatReuse reuse:
120:  *     - MAT_INITIAL_MATRIX: Create a new aij representation
121:  *     - MAT_REUSE_MATRIX: Reuse all aij representation and just change values
122:  * Output:
123:  *   - int *nnz: Number of nonzero-elements.
124:  *   - int **r pointer to i index
125:  *   - int **c pointer to j elements
126:  *   - MATRIXTYPE **v: Non-zero elements
127:  */
128: PetscErrorCode MatCopy_seqaij_seqaij_MKL_CPARDISO(Mat A, MatReuse reuse, PetscInt *nnz, PetscInt **r, PetscInt **c, PetscScalar **v)
129: {
130:   Mat_SeqAIJ *aa = (Mat_SeqAIJ *)A->data;

132:   PetscFunctionBegin;
133:   *v = aa->a;
134:   if (reuse == MAT_INITIAL_MATRIX) {
135:     *r   = (INT_TYPE *)aa->i;
136:     *c   = (INT_TYPE *)aa->j;
137:     *nnz = aa->nz;
138:   }
139:   PetscFunctionReturn(PETSC_SUCCESS);
140: }

142: PetscErrorCode MatConvertToTriples_mpiaij_mpiaij_MKL_CPARDISO(Mat A, MatReuse reuse, PetscInt *nnz, PetscInt **r, PetscInt **c, PetscScalar **v)
143: {
144:   const PetscInt    *ai, *aj, *bi, *bj, *garray, m = A->rmap->n, *ajj, *bjj;
145:   PetscInt           rstart, nz, i, j, countA, countB;
146:   PetscInt          *row, *col;
147:   const PetscScalar *av, *bv;
148:   PetscScalar       *val;
149:   Mat_MPIAIJ        *mat = (Mat_MPIAIJ *)A->data;
150:   Mat_SeqAIJ        *aa  = (Mat_SeqAIJ *)(mat->A)->data;
151:   Mat_SeqAIJ        *bb  = (Mat_SeqAIJ *)(mat->B)->data;
152:   PetscInt           colA_start, jB, jcol;

154:   PetscFunctionBegin;
155:   ai     = aa->i;
156:   aj     = aa->j;
157:   bi     = bb->i;
158:   bj     = bb->j;
159:   rstart = A->rmap->rstart;
160:   av     = aa->a;
161:   bv     = bb->a;

163:   garray = mat->garray;

165:   if (reuse == MAT_INITIAL_MATRIX) {
166:     nz   = aa->nz + bb->nz;
167:     *nnz = nz;
168:     PetscCall(PetscMalloc3(m + 1, &row, nz, &col, nz, &val));
169:     *r = row;
170:     *c = col;
171:     *v = val;
172:   } else {
173:     row = *r;
174:     col = *c;
175:     val = *v;
176:   }

178:   nz = 0;
179:   for (i = 0; i < m; i++) {
180:     row[i] = nz;
181:     countA = ai[i + 1] - ai[i];
182:     countB = bi[i + 1] - bi[i];
183:     ajj    = aj + ai[i]; /* ptr to the beginning of this row */
184:     bjj    = bj + bi[i];

186:     /* B part, smaller col index */
187:     colA_start = rstart + ajj[0]; /* the smallest global col index of A */
188:     jB         = 0;
189:     for (j = 0; j < countB; j++) {
190:       jcol = garray[bjj[j]];
191:       if (jcol > colA_start) break;
192:       col[nz]   = jcol;
193:       val[nz++] = *bv++;
194:     }
195:     jB = j;

197:     /* A part */
198:     for (j = 0; j < countA; j++) {
199:       col[nz]   = rstart + ajj[j];
200:       val[nz++] = *av++;
201:     }

203:     /* B part, larger col index */
204:     for (j = jB; j < countB; j++) {
205:       col[nz]   = garray[bjj[j]];
206:       val[nz++] = *bv++;
207:     }
208:   }
209:   row[m] = nz;

211:   PetscFunctionReturn(PETSC_SUCCESS);
212: }

214: PetscErrorCode MatConvertToTriples_mpibaij_mpibaij_MKL_CPARDISO(Mat A, MatReuse reuse, PetscInt *nnz, PetscInt **r, PetscInt **c, PetscScalar **v)
215: {
216:   const PetscInt    *ai, *aj, *bi, *bj, *garray, bs = A->rmap->bs, bs2 = bs * bs, m = A->rmap->n / bs, *ajj, *bjj;
217:   PetscInt           rstart, nz, i, j, countA, countB;
218:   PetscInt          *row, *col;
219:   const PetscScalar *av, *bv;
220:   PetscScalar       *val;
221:   Mat_MPIBAIJ       *mat = (Mat_MPIBAIJ *)A->data;
222:   Mat_SeqBAIJ       *aa  = (Mat_SeqBAIJ *)(mat->A)->data;
223:   Mat_SeqBAIJ       *bb  = (Mat_SeqBAIJ *)(mat->B)->data;
224:   PetscInt           colA_start, jB, jcol;

226:   PetscFunctionBegin;
227:   ai     = aa->i;
228:   aj     = aa->j;
229:   bi     = bb->i;
230:   bj     = bb->j;
231:   rstart = A->rmap->rstart / bs;
232:   av     = aa->a;
233:   bv     = bb->a;

235:   garray = mat->garray;

237:   if (reuse == MAT_INITIAL_MATRIX) {
238:     nz   = aa->nz + bb->nz;
239:     *nnz = nz;
240:     PetscCall(PetscMalloc3(m + 1, &row, nz, &col, nz * bs2, &val));
241:     *r = row;
242:     *c = col;
243:     *v = val;
244:   } else {
245:     row = *r;
246:     col = *c;
247:     val = *v;
248:   }

250:   nz = 0;
251:   for (i = 0; i < m; i++) {
252:     row[i] = nz + 1;
253:     countA = ai[i + 1] - ai[i];
254:     countB = bi[i + 1] - bi[i];
255:     ajj    = aj + ai[i]; /* ptr to the beginning of this row */
256:     bjj    = bj + bi[i];

258:     /* B part, smaller col index */
259:     colA_start = rstart + (countA > 0 ? ajj[0] : 0); /* the smallest global col index of A */
260:     jB         = 0;
261:     for (j = 0; j < countB; j++) {
262:       jcol = garray[bjj[j]];
263:       if (jcol > colA_start) break;
264:       col[nz++] = jcol + 1;
265:     }
266:     jB = j;
267:     PetscCall(PetscArraycpy(val, bv, jB * bs2));
268:     val += jB * bs2;
269:     bv += jB * bs2;

271:     /* A part */
272:     for (j = 0; j < countA; j++) col[nz++] = rstart + ajj[j] + 1;
273:     PetscCall(PetscArraycpy(val, av, countA * bs2));
274:     val += countA * bs2;
275:     av += countA * bs2;

277:     /* B part, larger col index */
278:     for (j = jB; j < countB; j++) col[nz++] = garray[bjj[j]] + 1;
279:     PetscCall(PetscArraycpy(val, bv, (countB - jB) * bs2));
280:     val += (countB - jB) * bs2;
281:     bv += (countB - jB) * bs2;
282:   }
283:   row[m] = nz + 1;

285:   PetscFunctionReturn(PETSC_SUCCESS);
286: }

288: PetscErrorCode MatConvertToTriples_mpisbaij_mpisbaij_MKL_CPARDISO(Mat A, MatReuse reuse, PetscInt *nnz, PetscInt **r, PetscInt **c, PetscScalar **v)
289: {
290:   const PetscInt    *ai, *aj, *bi, *bj, *garray, bs = A->rmap->bs, bs2 = bs * bs, m = A->rmap->n / bs, *ajj, *bjj;
291:   PetscInt           rstart, nz, i, j, countA, countB;
292:   PetscInt          *row, *col;
293:   const PetscScalar *av, *bv;
294:   PetscScalar       *val;
295:   Mat_MPISBAIJ      *mat = (Mat_MPISBAIJ *)A->data;
296:   Mat_SeqSBAIJ      *aa  = (Mat_SeqSBAIJ *)(mat->A)->data;
297:   Mat_SeqBAIJ       *bb  = (Mat_SeqBAIJ *)(mat->B)->data;

299:   PetscFunctionBegin;
300:   ai     = aa->i;
301:   aj     = aa->j;
302:   bi     = bb->i;
303:   bj     = bb->j;
304:   rstart = A->rmap->rstart / bs;
305:   av     = aa->a;
306:   bv     = bb->a;

308:   garray = mat->garray;

310:   if (reuse == MAT_INITIAL_MATRIX) {
311:     nz   = aa->nz + bb->nz;
312:     *nnz = nz;
313:     PetscCall(PetscMalloc3(m + 1, &row, nz, &col, nz * bs2, &val));
314:     *r = row;
315:     *c = col;
316:     *v = val;
317:   } else {
318:     row = *r;
319:     col = *c;
320:     val = *v;
321:   }

323:   nz = 0;
324:   for (i = 0; i < m; i++) {
325:     row[i] = nz + 1;
326:     countA = ai[i + 1] - ai[i];
327:     countB = bi[i + 1] - bi[i];
328:     ajj    = aj + ai[i]; /* ptr to the beginning of this row */
329:     bjj    = bj + bi[i];

331:     /* A part */
332:     for (j = 0; j < countA; j++) col[nz++] = rstart + ajj[j] + 1;
333:     PetscCall(PetscArraycpy(val, av, countA * bs2));
334:     val += countA * bs2;
335:     av += countA * bs2;

337:     /* B part, larger col index */
338:     for (j = 0; j < countB; j++) col[nz++] = garray[bjj[j]] + 1;
339:     PetscCall(PetscArraycpy(val, bv, countB * bs2));
340:     val += countB * bs2;
341:     bv += countB * bs2;
342:   }
343:   row[m] = nz + 1;

345:   PetscFunctionReturn(PETSC_SUCCESS);
346: }

348: /*
349:  * Free memory for Mat_MKL_CPARDISO structure and pointers to objects.
350:  */
351: PetscErrorCode MatDestroy_MKL_CPARDISO(Mat A)
352: {
353:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)A->data;
354:   MPI_Comm          comm;

356:   PetscFunctionBegin;
357:   /* Terminate instance, deallocate memories */
358:   if (mat_mkl_cpardiso->CleanUp) {
359:     mat_mkl_cpardiso->phase = JOB_RELEASE_OF_ALL_MEMORY;

361:     cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, NULL, NULL, NULL, mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs,
362:                           mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, NULL, NULL, &mat_mkl_cpardiso->comm_mkl_cpardiso, (PetscInt *)&mat_mkl_cpardiso->err);
363:   }

365:   if (mat_mkl_cpardiso->ConvertToTriples != MatCopy_seqaij_seqaij_MKL_CPARDISO) PetscCall(PetscFree3(mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja, mat_mkl_cpardiso->a));
366:   comm = MPI_Comm_f2c(mat_mkl_cpardiso->comm_mkl_cpardiso);
367:   PetscCallMPI(MPI_Comm_free(&comm));
368:   PetscCall(PetscFree(A->data));

370:   /* clear composed functions */
371:   PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatFactorGetSolverType_C", NULL));
372:   PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMkl_CPardisoSetCntl_C", NULL));
373:   PetscFunctionReturn(PETSC_SUCCESS);
374: }

376: /*
377:  * Computes Ax = b
378:  */
379: PetscErrorCode MatSolve_MKL_CPARDISO(Mat A, Vec b, Vec x)
380: {
381:   Mat_MKL_CPARDISO  *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)(A)->data;
382:   PetscScalar       *xarray;
383:   const PetscScalar *barray;

385:   PetscFunctionBegin;
386:   mat_mkl_cpardiso->nrhs = 1;
387:   PetscCall(VecGetArray(x, &xarray));
388:   PetscCall(VecGetArrayRead(b, &barray));

390:   /* solve phase */
391:   mat_mkl_cpardiso->phase = JOB_SOLVE_ITERATIVE_REFINEMENT;
392:   cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, mat_mkl_cpardiso->a, mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja,
393:                         mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs, mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, (void *)barray, (void *)xarray, &mat_mkl_cpardiso->comm_mkl_cpardiso, (PetscInt *)&mat_mkl_cpardiso->err);

395:   PetscCheck(mat_mkl_cpardiso->err >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MKL_CPARDISO: err=%d, msg = \"%s\". Please check manual", mat_mkl_cpardiso->err, Err_MSG_CPardiso(mat_mkl_cpardiso->err));

397:   PetscCall(VecRestoreArray(x, &xarray));
398:   PetscCall(VecRestoreArrayRead(b, &barray));
399:   mat_mkl_cpardiso->CleanUp = PETSC_TRUE;
400:   PetscFunctionReturn(PETSC_SUCCESS);
401: }

403: PetscErrorCode MatSolveTranspose_MKL_CPARDISO(Mat A, Vec b, Vec x)
404: {
405:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)A->data;

407:   PetscFunctionBegin;
408: #if defined(PETSC_USE_COMPLEX)
409:   mat_mkl_cpardiso->iparm[12 - 1] = 1;
410: #else
411:   mat_mkl_cpardiso->iparm[12 - 1] = 2;
412: #endif
413:   PetscCall(MatSolve_MKL_CPARDISO(A, b, x));
414:   mat_mkl_cpardiso->iparm[12 - 1] = 0;
415:   PetscFunctionReturn(PETSC_SUCCESS);
416: }

418: PetscErrorCode MatMatSolve_MKL_CPARDISO(Mat A, Mat B, Mat X)
419: {
420:   Mat_MKL_CPARDISO  *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)(A)->data;
421:   PetscScalar       *xarray;
422:   const PetscScalar *barray;

424:   PetscFunctionBegin;
425:   PetscCall(MatGetSize(B, NULL, (PetscInt *)&mat_mkl_cpardiso->nrhs));

427:   if (mat_mkl_cpardiso->nrhs > 0) {
428:     PetscCall(MatDenseGetArrayRead(B, &barray));
429:     PetscCall(MatDenseGetArray(X, &xarray));

431:     PetscCheck(barray != xarray, PETSC_COMM_SELF, PETSC_ERR_SUP, "B and X cannot share the same memory location");

433:     /* solve phase */
434:     mat_mkl_cpardiso->phase = JOB_SOLVE_ITERATIVE_REFINEMENT;
435:     cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, mat_mkl_cpardiso->a, mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja,
436:                           mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs, mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, (void *)barray, (void *)xarray, &mat_mkl_cpardiso->comm_mkl_cpardiso, (PetscInt *)&mat_mkl_cpardiso->err);
437:     PetscCheck(mat_mkl_cpardiso->err >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MKL_CPARDISO: err=%d, msg = \"%s\". Please check manual", mat_mkl_cpardiso->err, Err_MSG_CPardiso(mat_mkl_cpardiso->err));
438:     PetscCall(MatDenseRestoreArrayRead(B, &barray));
439:     PetscCall(MatDenseRestoreArray(X, &xarray));
440:   }
441:   mat_mkl_cpardiso->CleanUp = PETSC_TRUE;
442:   PetscFunctionReturn(PETSC_SUCCESS);
443: }

445: /*
446:  * LU Decomposition
447:  */
448: PetscErrorCode MatFactorNumeric_MKL_CPARDISO(Mat F, Mat A, const MatFactorInfo *info)
449: {
450:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)(F)->data;

452:   PetscFunctionBegin;
453:   mat_mkl_cpardiso->matstruc = SAME_NONZERO_PATTERN;
454:   PetscCall((*mat_mkl_cpardiso->ConvertToTriples)(A, MAT_REUSE_MATRIX, &mat_mkl_cpardiso->nz, &mat_mkl_cpardiso->ia, &mat_mkl_cpardiso->ja, &mat_mkl_cpardiso->a));

456:   mat_mkl_cpardiso->phase = JOB_NUMERICAL_FACTORIZATION;
457:   cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, mat_mkl_cpardiso->a, mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja,
458:                         mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs, mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, NULL, NULL, &mat_mkl_cpardiso->comm_mkl_cpardiso, &mat_mkl_cpardiso->err);
459:   PetscCheck(mat_mkl_cpardiso->err >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MKL_CPARDISO: err=%d, msg = \"%s\". Please check manual", mat_mkl_cpardiso->err, Err_MSG_CPardiso(mat_mkl_cpardiso->err));

461:   mat_mkl_cpardiso->matstruc = SAME_NONZERO_PATTERN;
462:   mat_mkl_cpardiso->CleanUp  = PETSC_TRUE;
463:   PetscFunctionReturn(PETSC_SUCCESS);
464: }

466: /* Sets mkl_cpardiso options from the options database */
467: PetscErrorCode MatSetFromOptions_MKL_CPARDISO(Mat F, Mat A)
468: {
469:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)F->data;
470:   PetscInt          icntl, threads;
471:   PetscBool         flg;

473:   PetscFunctionBegin;
474:   PetscOptionsBegin(PetscObjectComm((PetscObject)F), ((PetscObject)F)->prefix, "MKL_CPARDISO Options", "Mat");
475:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_65", "Suggested number of threads to use within MKL_CPARDISO", "None", threads, &threads, &flg));
476:   if (flg) mkl_set_num_threads((int)threads);

478:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_66", "Maximum number of factors with identical sparsity structure that must be kept in memory at the same time", "None", mat_mkl_cpardiso->maxfct, &icntl, &flg));
479:   if (flg) mat_mkl_cpardiso->maxfct = icntl;

481:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_67", "Indicates the actual matrix for the solution phase", "None", mat_mkl_cpardiso->mnum, &icntl, &flg));
482:   if (flg) mat_mkl_cpardiso->mnum = icntl;

484:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_68", "Message level information", "None", mat_mkl_cpardiso->msglvl, &icntl, &flg));
485:   if (flg) mat_mkl_cpardiso->msglvl = icntl;

487:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_69", "Defines the matrix type", "None", mat_mkl_cpardiso->mtype, &icntl, &flg));
488:   if (flg) mat_mkl_cpardiso->mtype = icntl;
489:   PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_1", "Use default values", "None", mat_mkl_cpardiso->iparm[0], &icntl, &flg));

491:   if (flg && icntl != 0) {
492:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_2", "Fill-in reducing ordering for the input matrix", "None", mat_mkl_cpardiso->iparm[1], &icntl, &flg));
493:     if (flg) mat_mkl_cpardiso->iparm[1] = icntl;

495:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_4", "Preconditioned CGS/CG", "None", mat_mkl_cpardiso->iparm[3], &icntl, &flg));
496:     if (flg) mat_mkl_cpardiso->iparm[3] = icntl;

498:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_5", "User permutation", "None", mat_mkl_cpardiso->iparm[4], &icntl, &flg));
499:     if (flg) mat_mkl_cpardiso->iparm[4] = icntl;

501:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_6", "Write solution on x", "None", mat_mkl_cpardiso->iparm[5], &icntl, &flg));
502:     if (flg) mat_mkl_cpardiso->iparm[5] = icntl;

504:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_8", "Iterative refinement step", "None", mat_mkl_cpardiso->iparm[7], &icntl, &flg));
505:     if (flg) mat_mkl_cpardiso->iparm[7] = icntl;

507:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_10", "Pivoting perturbation", "None", mat_mkl_cpardiso->iparm[9], &icntl, &flg));
508:     if (flg) mat_mkl_cpardiso->iparm[9] = icntl;

510:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_11", "Scaling vectors", "None", mat_mkl_cpardiso->iparm[10], &icntl, &flg));
511:     if (flg) mat_mkl_cpardiso->iparm[10] = icntl;

513:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_12", "Solve with transposed or conjugate transposed matrix A", "None", mat_mkl_cpardiso->iparm[11], &icntl, &flg));
514:     if (flg) mat_mkl_cpardiso->iparm[11] = icntl;

516:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_13", "Improved accuracy using (non-) symmetric weighted matching", "None", mat_mkl_cpardiso->iparm[12], &icntl, &flg));
517:     if (flg) mat_mkl_cpardiso->iparm[12] = icntl;

519:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_18", "Numbers of non-zero elements", "None", mat_mkl_cpardiso->iparm[17], &icntl, &flg));
520:     if (flg) mat_mkl_cpardiso->iparm[17] = icntl;

522:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_19", "Report number of floating point operations", "None", mat_mkl_cpardiso->iparm[18], &icntl, &flg));
523:     if (flg) mat_mkl_cpardiso->iparm[18] = icntl;

525:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_21", "Pivoting for symmetric indefinite matrices", "None", mat_mkl_cpardiso->iparm[20], &icntl, &flg));
526:     if (flg) mat_mkl_cpardiso->iparm[20] = icntl;

528:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_24", "Parallel factorization control", "None", mat_mkl_cpardiso->iparm[23], &icntl, &flg));
529:     if (flg) mat_mkl_cpardiso->iparm[23] = icntl;

531:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_25", "Parallel forward/backward solve control", "None", mat_mkl_cpardiso->iparm[24], &icntl, &flg));
532:     if (flg) mat_mkl_cpardiso->iparm[24] = icntl;

534:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_27", "Matrix checker", "None", mat_mkl_cpardiso->iparm[26], &icntl, &flg));
535:     if (flg) mat_mkl_cpardiso->iparm[26] = icntl;

537:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_31", "Partial solve and computing selected components of the solution vectors", "None", mat_mkl_cpardiso->iparm[30], &icntl, &flg));
538:     if (flg) mat_mkl_cpardiso->iparm[30] = icntl;

540:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_34", "Optimal number of threads for conditional numerical reproducibility (CNR) mode", "None", mat_mkl_cpardiso->iparm[33], &icntl, &flg));
541:     if (flg) mat_mkl_cpardiso->iparm[33] = icntl;

543:     PetscCall(PetscOptionsInt("-mat_mkl_cpardiso_60", "Intel MKL_CPARDISO mode", "None", mat_mkl_cpardiso->iparm[59], &icntl, &flg));
544:     if (flg) mat_mkl_cpardiso->iparm[59] = icntl;
545:   }

547:   PetscOptionsEnd();
548:   PetscFunctionReturn(PETSC_SUCCESS);
549: }

551: PetscErrorCode PetscInitialize_MKL_CPARDISO(Mat A, Mat_MKL_CPARDISO *mat_mkl_cpardiso)
552: {
553:   PetscInt    bs;
554:   PetscBool   match;
555:   PetscMPIInt size;
556:   MPI_Comm    comm;

558:   PetscFunctionBegin;

560:   PetscCallMPI(MPI_Comm_dup(PetscObjectComm((PetscObject)A), &comm));
561:   PetscCallMPI(MPI_Comm_size(comm, &size));
562:   mat_mkl_cpardiso->comm_mkl_cpardiso = MPI_Comm_c2f(comm);

564:   mat_mkl_cpardiso->CleanUp = PETSC_FALSE;
565:   mat_mkl_cpardiso->maxfct  = 1;
566:   mat_mkl_cpardiso->mnum    = 1;
567:   mat_mkl_cpardiso->n       = A->rmap->N;
568:   if (mat_mkl_cpardiso->iparm[36]) mat_mkl_cpardiso->n /= mat_mkl_cpardiso->iparm[36];
569:   mat_mkl_cpardiso->msglvl = 0;
570:   mat_mkl_cpardiso->nrhs   = 1;
571:   mat_mkl_cpardiso->err    = 0;
572:   mat_mkl_cpardiso->phase  = -1;
573: #if defined(PETSC_USE_COMPLEX)
574:   mat_mkl_cpardiso->mtype = 13;
575: #else
576:   mat_mkl_cpardiso->mtype         = 11;
577: #endif

579: #if defined(PETSC_USE_REAL_SINGLE)
580:   mat_mkl_cpardiso->iparm[27] = 1;
581: #else
582:   mat_mkl_cpardiso->iparm[27]     = 0;
583: #endif

585:   mat_mkl_cpardiso->iparm[0]  = 1;  /* Solver default parameters overridden with provided by iparm */
586:   mat_mkl_cpardiso->iparm[1]  = 2;  /* Use METIS for fill-in reordering */
587:   mat_mkl_cpardiso->iparm[5]  = 0;  /* Write solution into x */
588:   mat_mkl_cpardiso->iparm[7]  = 2;  /* Max number of iterative refinement steps */
589:   mat_mkl_cpardiso->iparm[9]  = 13; /* Perturb the pivot elements with 1E-13 */
590:   mat_mkl_cpardiso->iparm[10] = 1;  /* Use nonsymmetric permutation and scaling MPS */
591:   mat_mkl_cpardiso->iparm[12] = 1;  /* Switch on Maximum Weighted Matching algorithm (default for non-symmetric) */
592:   mat_mkl_cpardiso->iparm[17] = -1; /* Output: Number of nonzeros in the factor LU */
593:   mat_mkl_cpardiso->iparm[18] = -1; /* Output: Mflops for LU factorization */
594:   mat_mkl_cpardiso->iparm[26] = 1;  /* Check input data for correctness */

596:   mat_mkl_cpardiso->iparm[39] = 0;
597:   if (size > 1) {
598:     mat_mkl_cpardiso->iparm[39] = 2;
599:     mat_mkl_cpardiso->iparm[40] = A->rmap->rstart;
600:     mat_mkl_cpardiso->iparm[41] = A->rmap->rend - 1;
601:   }
602:   PetscCall(PetscObjectTypeCompareAny((PetscObject)A, &match, MATMPIBAIJ, MATMPISBAIJ, ""));
603:   if (match) {
604:     PetscCall(MatGetBlockSize(A, &bs));
605:     mat_mkl_cpardiso->iparm[36] = bs;
606:     mat_mkl_cpardiso->iparm[40] /= bs;
607:     mat_mkl_cpardiso->iparm[41] /= bs;
608:     mat_mkl_cpardiso->iparm[40]++;
609:     mat_mkl_cpardiso->iparm[41]++;
610:     mat_mkl_cpardiso->iparm[34] = 0; /* Fortran style */
611:   } else {
612:     mat_mkl_cpardiso->iparm[34] = 1; /* C style */
613:   }

615:   mat_mkl_cpardiso->perm = 0;
616:   PetscFunctionReturn(PETSC_SUCCESS);
617: }

619: /*
620:  * Symbolic decomposition. Mkl_Pardiso analysis phase.
621:  */
622: PetscErrorCode MatLUFactorSymbolic_AIJMKL_CPARDISO(Mat F, Mat A, IS r, IS c, const MatFactorInfo *info)
623: {
624:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)F->data;

626:   PetscFunctionBegin;
627:   mat_mkl_cpardiso->matstruc = DIFFERENT_NONZERO_PATTERN;

629:   /* Set MKL_CPARDISO options from the options database */
630:   PetscCall(MatSetFromOptions_MKL_CPARDISO(F, A));
631:   PetscCall((*mat_mkl_cpardiso->ConvertToTriples)(A, MAT_INITIAL_MATRIX, &mat_mkl_cpardiso->nz, &mat_mkl_cpardiso->ia, &mat_mkl_cpardiso->ja, &mat_mkl_cpardiso->a));

633:   mat_mkl_cpardiso->n = A->rmap->N;
634:   if (mat_mkl_cpardiso->iparm[36]) mat_mkl_cpardiso->n /= mat_mkl_cpardiso->iparm[36];

636:   /* analysis phase */
637:   mat_mkl_cpardiso->phase = JOB_ANALYSIS;

639:   cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, mat_mkl_cpardiso->a, mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja,
640:                         mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs, mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, NULL, NULL, &mat_mkl_cpardiso->comm_mkl_cpardiso, (PetscInt *)&mat_mkl_cpardiso->err);

642:   PetscCheck(mat_mkl_cpardiso->err >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MKL_CPARDISO: err=%d, msg = \"%s\".Check manual", mat_mkl_cpardiso->err, Err_MSG_CPardiso(mat_mkl_cpardiso->err));

644:   mat_mkl_cpardiso->CleanUp = PETSC_TRUE;
645:   F->ops->lufactornumeric   = MatFactorNumeric_MKL_CPARDISO;
646:   F->ops->solve             = MatSolve_MKL_CPARDISO;
647:   F->ops->solvetranspose    = MatSolveTranspose_MKL_CPARDISO;
648:   F->ops->matsolve          = MatMatSolve_MKL_CPARDISO;
649:   PetscFunctionReturn(PETSC_SUCCESS);
650: }

652: PetscErrorCode MatCholeskyFactorSymbolic_AIJMKL_CPARDISO(Mat F, Mat A, IS perm, const MatFactorInfo *info)
653: {
654:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)F->data;

656:   PetscFunctionBegin;
657:   mat_mkl_cpardiso->matstruc = DIFFERENT_NONZERO_PATTERN;

659:   /* Set MKL_CPARDISO options from the options database */
660:   PetscCall(MatSetFromOptions_MKL_CPARDISO(F, A));
661:   PetscCall((*mat_mkl_cpardiso->ConvertToTriples)(A, MAT_INITIAL_MATRIX, &mat_mkl_cpardiso->nz, &mat_mkl_cpardiso->ia, &mat_mkl_cpardiso->ja, &mat_mkl_cpardiso->a));

663:   mat_mkl_cpardiso->n = A->rmap->N;
664:   if (mat_mkl_cpardiso->iparm[36]) mat_mkl_cpardiso->n /= mat_mkl_cpardiso->iparm[36];
665:   PetscCheck(!PetscDefined(USE_COMPLEX), PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "No support for PARDISO CHOLESKY with complex scalars! Use MAT_FACTOR_LU instead");
666:   if (A->spd == PETSC_BOOL3_TRUE) mat_mkl_cpardiso->mtype = 2;
667:   else mat_mkl_cpardiso->mtype = -2;

669:   /* analysis phase */
670:   mat_mkl_cpardiso->phase = JOB_ANALYSIS;

672:   cluster_sparse_solver(mat_mkl_cpardiso->pt, &mat_mkl_cpardiso->maxfct, &mat_mkl_cpardiso->mnum, &mat_mkl_cpardiso->mtype, &mat_mkl_cpardiso->phase, &mat_mkl_cpardiso->n, mat_mkl_cpardiso->a, mat_mkl_cpardiso->ia, mat_mkl_cpardiso->ja,
673:                         mat_mkl_cpardiso->perm, &mat_mkl_cpardiso->nrhs, mat_mkl_cpardiso->iparm, &mat_mkl_cpardiso->msglvl, NULL, NULL, &mat_mkl_cpardiso->comm_mkl_cpardiso, (PetscInt *)&mat_mkl_cpardiso->err);

675:   PetscCheck(mat_mkl_cpardiso->err >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MKL_CPARDISO: err=%d, msg = \"%s\".Check manual", mat_mkl_cpardiso->err, Err_MSG_CPardiso(mat_mkl_cpardiso->err));

677:   mat_mkl_cpardiso->CleanUp     = PETSC_TRUE;
678:   F->ops->choleskyfactornumeric = MatFactorNumeric_MKL_CPARDISO;
679:   F->ops->solve                 = MatSolve_MKL_CPARDISO;
680:   F->ops->solvetranspose        = MatSolveTranspose_MKL_CPARDISO;
681:   F->ops->matsolve              = MatMatSolve_MKL_CPARDISO;
682:   PetscFunctionReturn(PETSC_SUCCESS);
683: }

685: PetscErrorCode MatView_MKL_CPARDISO(Mat A, PetscViewer viewer)
686: {
687:   PetscBool         iascii;
688:   PetscViewerFormat format;
689:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)A->data;
690:   PetscInt          i;

692:   PetscFunctionBegin;
693:   /* check if matrix is mkl_cpardiso type */
694:   if (A->ops->solve != MatSolve_MKL_CPARDISO) PetscFunctionReturn(PETSC_SUCCESS);

696:   PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &iascii));
697:   if (iascii) {
698:     PetscCall(PetscViewerGetFormat(viewer, &format));
699:     if (format == PETSC_VIEWER_ASCII_INFO) {
700:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO run parameters:\n"));
701:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO phase:             %d \n", mat_mkl_cpardiso->phase));
702:       for (i = 1; i <= 64; i++) PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO iparm[%d]:     %d \n", i, mat_mkl_cpardiso->iparm[i - 1]));
703:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO maxfct:     %d \n", mat_mkl_cpardiso->maxfct));
704:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO mnum:     %d \n", mat_mkl_cpardiso->mnum));
705:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO mtype:     %d \n", mat_mkl_cpardiso->mtype));
706:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO n:     %d \n", mat_mkl_cpardiso->n));
707:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO nrhs:     %d \n", mat_mkl_cpardiso->nrhs));
708:       PetscCall(PetscViewerASCIIPrintf(viewer, "MKL_CPARDISO msglvl:     %d \n", mat_mkl_cpardiso->msglvl));
709:     }
710:   }
711:   PetscFunctionReturn(PETSC_SUCCESS);
712: }

714: PetscErrorCode MatGetInfo_MKL_CPARDISO(Mat A, MatInfoType flag, MatInfo *info)
715: {
716:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)A->data;

718:   PetscFunctionBegin;
719:   info->block_size        = 1.0;
720:   info->nz_allocated      = mat_mkl_cpardiso->nz + 0.0;
721:   info->nz_unneeded       = 0.0;
722:   info->assemblies        = 0.0;
723:   info->mallocs           = 0.0;
724:   info->memory            = 0.0;
725:   info->fill_ratio_given  = 0;
726:   info->fill_ratio_needed = 0;
727:   info->factor_mallocs    = 0;
728:   PetscFunctionReturn(PETSC_SUCCESS);
729: }

731: PetscErrorCode MatMkl_CPardisoSetCntl_MKL_CPARDISO(Mat F, PetscInt icntl, PetscInt ival)
732: {
733:   Mat_MKL_CPARDISO *mat_mkl_cpardiso = (Mat_MKL_CPARDISO *)F->data;

735:   PetscFunctionBegin;
736:   if (icntl <= 64) {
737:     mat_mkl_cpardiso->iparm[icntl - 1] = ival;
738:   } else {
739:     if (icntl == 65) mkl_set_num_threads((int)ival);
740:     else if (icntl == 66) mat_mkl_cpardiso->maxfct = ival;
741:     else if (icntl == 67) mat_mkl_cpardiso->mnum = ival;
742:     else if (icntl == 68) mat_mkl_cpardiso->msglvl = ival;
743:     else if (icntl == 69) mat_mkl_cpardiso->mtype = ival;
744:   }
745:   PetscFunctionReturn(PETSC_SUCCESS);
746: }

748: /*@
749:   MatMkl_CPardisoSetCntl - Set Mkl_Pardiso parameters

751:    Logically Collective

753:    Input Parameters:
754: +  F - the factored matrix obtained by calling `MatGetFactor()`
755: .  icntl - index of Mkl_Pardiso parameter
756: -  ival - value of Mkl_Pardiso parameter

758:   Options Database Key:
759: .   -mat_mkl_cpardiso_<icntl> <ival> - set the option numbered icntl to ival

761:    Level: Intermediate

763:    Note:
764:     This routine cannot be used if you are solving the linear system with `TS`, `SNES`, or `KSP`, only if you directly call `MatGetFactor()` so use the options
765:           database approach when working with `TS`, `SNES`, or `KSP`. See `MATSOLVERMKL_CPARDISO` for the options

767:    References:
768: .  * - Mkl_Pardiso Users' Guide

770: .seealso: [](ch_matrices), `Mat`, `MatGetFactor()`, `MATMPIAIJ`, `MATSOLVERMKL_CPARDISO`
771: @*/
772: PetscErrorCode MatMkl_CPardisoSetCntl(Mat F, PetscInt icntl, PetscInt ival)
773: {
774:   PetscFunctionBegin;
775:   PetscTryMethod(F, "MatMkl_CPardisoSetCntl_C", (Mat, PetscInt, PetscInt), (F, icntl, ival));
776:   PetscFunctionReturn(PETSC_SUCCESS);
777: }

779: /*MC
780:   MATSOLVERMKL_CPARDISO -  A matrix type providing direct solvers (LU) for parallel matrices via the external package MKL_CPARDISO.

782:   Works with `MATMPIAIJ` matrices

784:   Use `-pc_type lu` `-pc_factor_mat_solver_type mkl_cpardiso` to use this direct solver

786:   Options Database Keys:
787: + -mat_mkl_cpardiso_65 - Suggested number of threads to use within MKL_CPARDISO
788: . -mat_mkl_cpardiso_66 - Maximum number of factors with identical sparsity structure that must be kept in memory at the same time
789: . -mat_mkl_cpardiso_67 - Indicates the actual matrix for the solution phase
790: . -mat_mkl_cpardiso_68 - Message level information, use 1 to get detailed information on the solver options
791: . -mat_mkl_cpardiso_69 - Defines the matrix type. IMPORTANT: When you set this flag, iparm parameters are going to be set to the default ones for the matrix type
792: . -mat_mkl_cpardiso_1  - Use default values
793: . -mat_mkl_cpardiso_2  - Fill-in reducing ordering for the input matrix
794: . -mat_mkl_cpardiso_4  - Preconditioned CGS/CG
795: . -mat_mkl_cpardiso_5  - User permutation
796: . -mat_mkl_cpardiso_6  - Write solution on x
797: . -mat_mkl_cpardiso_8  - Iterative refinement step
798: . -mat_mkl_cpardiso_10 - Pivoting perturbation
799: . -mat_mkl_cpardiso_11 - Scaling vectors
800: . -mat_mkl_cpardiso_12 - Solve with transposed or conjugate transposed matrix A
801: . -mat_mkl_cpardiso_13 - Improved accuracy using (non-) symmetric weighted matching
802: . -mat_mkl_cpardiso_18 - Numbers of non-zero elements
803: . -mat_mkl_cpardiso_19 - Report number of floating point operations
804: . -mat_mkl_cpardiso_21 - Pivoting for symmetric indefinite matrices
805: . -mat_mkl_cpardiso_24 - Parallel factorization control
806: . -mat_mkl_cpardiso_25 - Parallel forward/backward solve control
807: . -mat_mkl_cpardiso_27 - Matrix checker
808: . -mat_mkl_cpardiso_31 - Partial solve and computing selected components of the solution vectors
809: . -mat_mkl_cpardiso_34 - Optimal number of threads for conditional numerical reproducibility (CNR) mode
810: - -mat_mkl_cpardiso_60 - Intel MKL_CPARDISO mode

812:   Level: beginner

814:   Notes:
815:     Use `-mat_mkl_cpardiso_68 1` to display the number of threads the solver is using. MKL does not provide a way to directly access this
816:     information.

818:     For more information on the options check the MKL_CPARDISO manual

820: .seealso: [](ch_matrices), `Mat`, `PCFactorSetMatSolverType()`, `MatSolverType`, `MatMkl_CPardisoSetCntl()`, `MatGetFactor()`
821: M*/

823: static PetscErrorCode MatFactorGetSolverType_mkl_cpardiso(Mat A, MatSolverType *type)
824: {
825:   PetscFunctionBegin;
826:   *type = MATSOLVERMKL_CPARDISO;
827:   PetscFunctionReturn(PETSC_SUCCESS);
828: }

830: /* MatGetFactor for MPI AIJ matrices */
831: static PetscErrorCode MatGetFactor_mpiaij_mkl_cpardiso(Mat A, MatFactorType ftype, Mat *F)
832: {
833:   Mat               B;
834:   Mat_MKL_CPARDISO *mat_mkl_cpardiso;
835:   PetscBool         isSeqAIJ, isMPIBAIJ, isMPISBAIJ;

837:   PetscFunctionBegin;
838:   /* Create the factorization matrix */

840:   PetscCall(PetscObjectTypeCompare((PetscObject)A, MATSEQAIJ, &isSeqAIJ));
841:   PetscCall(PetscObjectTypeCompare((PetscObject)A, MATMPIBAIJ, &isMPIBAIJ));
842:   PetscCall(PetscObjectTypeCompare((PetscObject)A, MATMPISBAIJ, &isMPISBAIJ));
843:   PetscCall(MatCreate(PetscObjectComm((PetscObject)A), &B));
844:   PetscCall(MatSetSizes(B, A->rmap->n, A->cmap->n, A->rmap->N, A->cmap->N));
845:   PetscCall(PetscStrallocpy("mkl_cpardiso", &((PetscObject)B)->type_name));
846:   PetscCall(MatSetUp(B));

848:   PetscCall(PetscNew(&mat_mkl_cpardiso));

850:   if (isSeqAIJ) mat_mkl_cpardiso->ConvertToTriples = MatCopy_seqaij_seqaij_MKL_CPARDISO;
851:   else if (isMPIBAIJ) mat_mkl_cpardiso->ConvertToTriples = MatConvertToTriples_mpibaij_mpibaij_MKL_CPARDISO;
852:   else if (isMPISBAIJ) mat_mkl_cpardiso->ConvertToTriples = MatConvertToTriples_mpisbaij_mpisbaij_MKL_CPARDISO;
853:   else mat_mkl_cpardiso->ConvertToTriples = MatConvertToTriples_mpiaij_mpiaij_MKL_CPARDISO;

855:   if (ftype == MAT_FACTOR_LU) B->ops->lufactorsymbolic = MatLUFactorSymbolic_AIJMKL_CPARDISO;
856:   else B->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_AIJMKL_CPARDISO;
857:   B->ops->destroy = MatDestroy_MKL_CPARDISO;

859:   B->ops->view    = MatView_MKL_CPARDISO;
860:   B->ops->getinfo = MatGetInfo_MKL_CPARDISO;

862:   B->factortype = ftype;
863:   B->assembled  = PETSC_TRUE; /* required by -ksp_view */

865:   B->data = mat_mkl_cpardiso;

867:   /* set solvertype */
868:   PetscCall(PetscFree(B->solvertype));
869:   PetscCall(PetscStrallocpy(MATSOLVERMKL_CPARDISO, &B->solvertype));

871:   PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorGetSolverType_C", MatFactorGetSolverType_mkl_cpardiso));
872:   PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMkl_CPardisoSetCntl_C", MatMkl_CPardisoSetCntl_MKL_CPARDISO));
873:   PetscCall(PetscInitialize_MKL_CPARDISO(A, mat_mkl_cpardiso));

875:   *F = B;
876:   PetscFunctionReturn(PETSC_SUCCESS);
877: }

879: PETSC_EXTERN PetscErrorCode MatSolverTypeRegister_MKL_CPardiso(void)
880: {
881:   PetscFunctionBegin;
882:   PetscCall(MatSolverTypeRegister(MATSOLVERMKL_CPARDISO, MATMPIAIJ, MAT_FACTOR_LU, MatGetFactor_mpiaij_mkl_cpardiso));
883:   PetscCall(MatSolverTypeRegister(MATSOLVERMKL_CPARDISO, MATSEQAIJ, MAT_FACTOR_LU, MatGetFactor_mpiaij_mkl_cpardiso));
884:   PetscCall(MatSolverTypeRegister(MATSOLVERMKL_CPARDISO, MATMPIBAIJ, MAT_FACTOR_LU, MatGetFactor_mpiaij_mkl_cpardiso));
885:   PetscCall(MatSolverTypeRegister(MATSOLVERMKL_CPARDISO, MATMPISBAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_mpiaij_mkl_cpardiso));
886:   PetscFunctionReturn(PETSC_SUCCESS);
887: }