Actual source code: mumps.c
2: /*
3: Provides an interface to the MUMPS sparse solver
4: */
5: #include <petscpkg_version.h>
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_NULL 0
27: #define JOB_FACTSYMBOLIC 1
28: #define JOB_FACTNUMERIC 2
29: #define JOB_SOLVE 3
30: #define JOB_END -2
32: /* calls to MUMPS */
33: #if defined(PETSC_USE_COMPLEX)
34: #if defined(PETSC_USE_REAL_SINGLE)
35: #define MUMPS_c cmumps_c
36: #else
37: #define MUMPS_c zmumps_c
38: #endif
39: #else
40: #if defined(PETSC_USE_REAL_SINGLE)
41: #define MUMPS_c smumps_c
42: #else
43: #define MUMPS_c dmumps_c
44: #endif
45: #endif
47: /* MUMPS uses MUMPS_INT for nonzero indices such as irn/jcn, irn_loc/jcn_loc and uses int64_t for
48: number of nonzeros such as nnz, nnz_loc. We typedef MUMPS_INT to PetscMUMPSInt to follow the
49: naming convention in PetscMPIInt, PetscBLASInt etc.
50: */
51: typedef MUMPS_INT PetscMUMPSInt;
53: #if PETSC_PKG_MUMPS_VERSION_GE(5, 3, 0)
54: #if defined(MUMPS_INTSIZE64) /* MUMPS_INTSIZE64 is in MUMPS headers if it is built in full 64-bit mode, therefore the macro is more reliable */
55: #error "Petsc has not been tested with full 64-bit MUMPS and we choose to error out"
56: #endif
57: #else
58: #if defined(INTSIZE64) /* INTSIZE64 is a command line macro one used to build MUMPS in full 64-bit mode */
59: #error "Petsc has not been tested with full 64-bit MUMPS and we choose to error out"
60: #endif
61: #endif
63: #define MPIU_MUMPSINT MPI_INT
64: #define PETSC_MUMPS_INT_MAX 2147483647
65: #define PETSC_MUMPS_INT_MIN -2147483648
67: /* Cast PetscInt to PetscMUMPSInt. Usually there is no overflow since <a> is row/col indices or some small integers*/
68: static inline PetscErrorCode PetscMUMPSIntCast(PetscInt a, PetscMUMPSInt *b)
69: {
70: PetscFunctionBegin;
71: #if PetscDefined(USE_64BIT_INDICES)
72: PetscAssert(a <= PETSC_MUMPS_INT_MAX && a >= PETSC_MUMPS_INT_MIN, PETSC_COMM_SELF, PETSC_ERR_ARG_OUTOFRANGE, "PetscInt too long for PetscMUMPSInt");
73: #endif
74: *b = (PetscMUMPSInt)(a);
75: PetscFunctionReturn(PETSC_SUCCESS);
76: }
78: /* Put these utility routines here since they are only used in this file */
79: static inline PetscErrorCode PetscOptionsMUMPSInt_Private(PetscOptionItems *PetscOptionsObject, const char opt[], const char text[], const char man[], PetscMUMPSInt currentvalue, PetscMUMPSInt *value, PetscBool *set, PetscMUMPSInt lb, PetscMUMPSInt ub)
80: {
81: PetscInt myval;
82: PetscBool myset;
83: PetscFunctionBegin;
84: /* PetscInt's size should be always >= PetscMUMPSInt's. It is safe to call PetscOptionsInt_Private to read a PetscMUMPSInt */
85: PetscCall(PetscOptionsInt_Private(PetscOptionsObject, opt, text, man, (PetscInt)currentvalue, &myval, &myset, lb, ub));
86: if (myset) PetscCall(PetscMUMPSIntCast(myval, value));
87: if (set) *set = myset;
88: PetscFunctionReturn(PETSC_SUCCESS);
89: }
90: #define PetscOptionsMUMPSInt(a, b, c, d, e, f) PetscOptionsMUMPSInt_Private(PetscOptionsObject, a, b, c, d, e, f, PETSC_MUMPS_INT_MIN, PETSC_MUMPS_INT_MAX)
92: /* if using PETSc OpenMP support, we only call MUMPS on master ranks. Before/after the call, we change/restore CPUs the master ranks can run on */
93: #if defined(PETSC_HAVE_OPENMP_SUPPORT)
94: #define PetscMUMPS_c(mumps) \
95: do { \
96: if (mumps->use_petsc_omp_support) { \
97: if (mumps->is_omp_master) { \
98: PetscCall(PetscOmpCtrlOmpRegionOnMasterBegin(mumps->omp_ctrl)); \
99: PetscCall(PetscFPTrapPush(PETSC_FP_TRAP_OFF)); \
100: PetscStackCallExternalVoid(PetscStringize(MUMPS_c), MUMPS_c(&mumps->id)); \
101: PetscCall(PetscFPTrapPop()); \
102: PetscCall(PetscOmpCtrlOmpRegionOnMasterEnd(mumps->omp_ctrl)); \
103: } \
104: PetscCall(PetscOmpCtrlBarrier(mumps->omp_ctrl)); \
105: /* Global info is same on all processes so we Bcast it within omp_comm. Local info is specific \
106: to processes, so we only Bcast info[1], an error code and leave others (since they do not have \
107: an easy translation between omp_comm and petsc_comm). See MUMPS-5.1.2 manual p82. \
108: omp_comm is a small shared memory communicator, hence doing multiple Bcast as shown below is OK. \
109: */ \
110: PetscCallMPI(MPI_Bcast(mumps->id.infog, 40, MPIU_MUMPSINT, 0, mumps->omp_comm)); \
111: PetscCallMPI(MPI_Bcast(mumps->id.rinfog, 20, MPIU_REAL, 0, mumps->omp_comm)); \
112: PetscCallMPI(MPI_Bcast(mumps->id.info, 1, MPIU_MUMPSINT, 0, mumps->omp_comm)); \
113: } else { \
114: PetscCall(PetscFPTrapPush(PETSC_FP_TRAP_OFF)); \
115: PetscStackCallExternalVoid(PetscStringize(MUMPS_c), MUMPS_c(&mumps->id)); \
116: PetscCall(PetscFPTrapPop()); \
117: } \
118: } while (0)
119: #else
120: #define PetscMUMPS_c(mumps) \
121: do { \
122: PetscCall(PetscFPTrapPush(PETSC_FP_TRAP_OFF)); \
123: PetscStackCallExternalVoid(PetscStringize(MUMPS_c), MUMPS_c(&mumps->id)); \
124: PetscCall(PetscFPTrapPop()); \
125: } while (0)
126: #endif
128: /* declare MumpsScalar */
129: #if defined(PETSC_USE_COMPLEX)
130: #if defined(PETSC_USE_REAL_SINGLE)
131: #define MumpsScalar mumps_complex
132: #else
133: #define MumpsScalar mumps_double_complex
134: #endif
135: #else
136: #define MumpsScalar PetscScalar
137: #endif
139: /* macros s.t. indices match MUMPS documentation */
140: #define ICNTL(I) icntl[(I)-1]
141: #define CNTL(I) cntl[(I)-1]
142: #define INFOG(I) infog[(I)-1]
143: #define INFO(I) info[(I)-1]
144: #define RINFOG(I) rinfog[(I)-1]
145: #define RINFO(I) rinfo[(I)-1]
147: typedef struct Mat_MUMPS Mat_MUMPS;
148: struct Mat_MUMPS {
149: #if defined(PETSC_USE_COMPLEX)
150: #if defined(PETSC_USE_REAL_SINGLE)
151: CMUMPS_STRUC_C id;
152: #else
153: ZMUMPS_STRUC_C id;
154: #endif
155: #else
156: #if defined(PETSC_USE_REAL_SINGLE)
157: SMUMPS_STRUC_C id;
158: #else
159: DMUMPS_STRUC_C id;
160: #endif
161: #endif
163: MatStructure matstruc;
164: PetscMPIInt myid, petsc_size;
165: PetscMUMPSInt *irn, *jcn; /* the (i,j,v) triplets passed to mumps. */
166: PetscScalar *val, *val_alloc; /* For some matrices, we can directly access their data array without a buffer. For others, we need a buffer. So comes val_alloc. */
167: PetscInt64 nnz; /* number of nonzeros. The type is called selective 64-bit in mumps */
168: PetscMUMPSInt sym;
169: MPI_Comm mumps_comm;
170: PetscMUMPSInt *ICNTL_pre;
171: PetscReal *CNTL_pre;
172: PetscMUMPSInt ICNTL9_pre; /* check if ICNTL(9) is changed from previous MatSolve */
173: VecScatter scat_rhs, scat_sol; /* used by MatSolve() */
174: PetscMUMPSInt ICNTL20; /* use centralized (0) or distributed (10) dense RHS */
175: PetscMUMPSInt lrhs_loc, nloc_rhs, *irhs_loc;
176: #if defined(PETSC_HAVE_OPENMP_SUPPORT)
177: PetscInt *rhs_nrow, max_nrhs;
178: PetscMPIInt *rhs_recvcounts, *rhs_disps;
179: PetscScalar *rhs_loc, *rhs_recvbuf;
180: #endif
181: Vec b_seq, x_seq;
182: PetscInt ninfo, *info; /* which INFO to display */
183: PetscInt sizeredrhs;
184: PetscScalar *schur_sol;
185: PetscInt schur_sizesol;
186: PetscMUMPSInt *ia_alloc, *ja_alloc; /* work arrays used for the CSR struct for sparse rhs */
187: PetscInt64 cur_ilen, cur_jlen; /* current len of ia_alloc[], ja_alloc[] */
188: PetscErrorCode (*ConvertToTriples)(Mat, PetscInt, MatReuse, Mat_MUMPS *);
190: /* stuff used by petsc/mumps OpenMP support*/
191: PetscBool use_petsc_omp_support;
192: PetscOmpCtrl omp_ctrl; /* an OpenMP controller that blocked processes will release their CPU (MPI_Barrier does not have this guarantee) */
193: MPI_Comm petsc_comm, omp_comm; /* petsc_comm is petsc matrix's comm */
194: PetscInt64 *recvcount; /* a collection of nnz on omp_master */
195: PetscMPIInt tag, omp_comm_size;
196: PetscBool is_omp_master; /* is this rank the master of omp_comm */
197: MPI_Request *reqs;
198: };
200: /* Cast a 1-based CSR represented by (nrow, ia, ja) of type PetscInt to a CSR of type PetscMUMPSInt.
201: Here, nrow is number of rows, ia[] is row pointer and ja[] is column indices.
202: */
203: static PetscErrorCode PetscMUMPSIntCSRCast(Mat_MUMPS *mumps, PetscInt nrow, PetscInt *ia, PetscInt *ja, PetscMUMPSInt **ia_mumps, PetscMUMPSInt **ja_mumps, PetscMUMPSInt *nnz_mumps)
204: {
205: PetscInt nnz = ia[nrow] - 1; /* mumps uses 1-based indices. Uses PetscInt instead of PetscInt64 since mumps only uses PetscMUMPSInt for rhs */
207: PetscFunctionBegin;
208: #if defined(PETSC_USE_64BIT_INDICES)
209: {
210: PetscInt i;
211: if (nrow + 1 > mumps->cur_ilen) { /* realloc ia_alloc/ja_alloc to fit ia/ja */
212: PetscCall(PetscFree(mumps->ia_alloc));
213: PetscCall(PetscMalloc1(nrow + 1, &mumps->ia_alloc));
214: mumps->cur_ilen = nrow + 1;
215: }
216: if (nnz > mumps->cur_jlen) {
217: PetscCall(PetscFree(mumps->ja_alloc));
218: PetscCall(PetscMalloc1(nnz, &mumps->ja_alloc));
219: mumps->cur_jlen = nnz;
220: }
221: for (i = 0; i < nrow + 1; i++) PetscCall(PetscMUMPSIntCast(ia[i], &(mumps->ia_alloc[i])));
222: for (i = 0; i < nnz; i++) PetscCall(PetscMUMPSIntCast(ja[i], &(mumps->ja_alloc[i])));
223: *ia_mumps = mumps->ia_alloc;
224: *ja_mumps = mumps->ja_alloc;
225: }
226: #else
227: *ia_mumps = ia;
228: *ja_mumps = ja;
229: #endif
230: PetscCall(PetscMUMPSIntCast(nnz, nnz_mumps));
231: PetscFunctionReturn(PETSC_SUCCESS);
232: }
234: static PetscErrorCode MatMumpsResetSchur_Private(Mat_MUMPS *mumps)
235: {
236: PetscFunctionBegin;
237: PetscCall(PetscFree(mumps->id.listvar_schur));
238: PetscCall(PetscFree(mumps->id.redrhs));
239: PetscCall(PetscFree(mumps->schur_sol));
240: mumps->id.size_schur = 0;
241: mumps->id.schur_lld = 0;
242: mumps->id.ICNTL(19) = 0;
243: PetscFunctionReturn(PETSC_SUCCESS);
244: }
246: /* solve with rhs in mumps->id.redrhs and return in the same location */
247: static PetscErrorCode MatMumpsSolveSchur_Private(Mat F)
248: {
249: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
250: Mat S, B, X;
251: MatFactorSchurStatus schurstatus;
252: PetscInt sizesol;
254: PetscFunctionBegin;
255: PetscCall(MatFactorFactorizeSchurComplement(F));
256: PetscCall(MatFactorGetSchurComplement(F, &S, &schurstatus));
257: PetscCall(MatCreateSeqDense(PETSC_COMM_SELF, mumps->id.size_schur, mumps->id.nrhs, (PetscScalar *)mumps->id.redrhs, &B));
258: PetscCall(MatSetType(B, ((PetscObject)S)->type_name));
259: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
260: PetscCall(MatBindToCPU(B, S->boundtocpu));
261: #endif
262: switch (schurstatus) {
263: case MAT_FACTOR_SCHUR_FACTORED:
264: PetscCall(MatCreateSeqDense(PETSC_COMM_SELF, mumps->id.size_schur, mumps->id.nrhs, (PetscScalar *)mumps->id.redrhs, &X));
265: PetscCall(MatSetType(X, ((PetscObject)S)->type_name));
266: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
267: PetscCall(MatBindToCPU(X, S->boundtocpu));
268: #endif
269: if (!mumps->id.ICNTL(9)) { /* transpose solve */
270: PetscCall(MatMatSolveTranspose(S, B, X));
271: } else {
272: PetscCall(MatMatSolve(S, B, X));
273: }
274: break;
275: case MAT_FACTOR_SCHUR_INVERTED:
276: sizesol = mumps->id.nrhs * mumps->id.size_schur;
277: if (!mumps->schur_sol || sizesol > mumps->schur_sizesol) {
278: PetscCall(PetscFree(mumps->schur_sol));
279: PetscCall(PetscMalloc1(sizesol, &mumps->schur_sol));
280: mumps->schur_sizesol = sizesol;
281: }
282: PetscCall(MatCreateSeqDense(PETSC_COMM_SELF, mumps->id.size_schur, mumps->id.nrhs, mumps->schur_sol, &X));
283: PetscCall(MatSetType(X, ((PetscObject)S)->type_name));
284: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
285: PetscCall(MatBindToCPU(X, S->boundtocpu));
286: #endif
287: PetscCall(MatProductCreateWithMat(S, B, NULL, X));
288: if (!mumps->id.ICNTL(9)) { /* transpose solve */
289: PetscCall(MatProductSetType(X, MATPRODUCT_AtB));
290: } else {
291: PetscCall(MatProductSetType(X, MATPRODUCT_AB));
292: }
293: PetscCall(MatProductSetFromOptions(X));
294: PetscCall(MatProductSymbolic(X));
295: PetscCall(MatProductNumeric(X));
297: PetscCall(MatCopy(X, B, SAME_NONZERO_PATTERN));
298: break;
299: default:
300: SETERRQ(PetscObjectComm((PetscObject)F), PETSC_ERR_SUP, "Unhandled MatFactorSchurStatus %d", F->schur_status);
301: }
302: PetscCall(MatFactorRestoreSchurComplement(F, &S, schurstatus));
303: PetscCall(MatDestroy(&B));
304: PetscCall(MatDestroy(&X));
305: PetscFunctionReturn(PETSC_SUCCESS);
306: }
308: static PetscErrorCode MatMumpsHandleSchur_Private(Mat F, PetscBool expansion)
309: {
310: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
312: PetscFunctionBegin;
313: if (!mumps->id.ICNTL(19)) { /* do nothing when Schur complement has not been computed */
314: PetscFunctionReturn(PETSC_SUCCESS);
315: }
316: if (!expansion) { /* prepare for the condensation step */
317: PetscInt sizeredrhs = mumps->id.nrhs * mumps->id.size_schur;
318: /* allocate MUMPS internal array to store reduced right-hand sides */
319: if (!mumps->id.redrhs || sizeredrhs > mumps->sizeredrhs) {
320: PetscCall(PetscFree(mumps->id.redrhs));
321: mumps->id.lredrhs = mumps->id.size_schur;
322: PetscCall(PetscMalloc1(mumps->id.nrhs * mumps->id.lredrhs, &mumps->id.redrhs));
323: mumps->sizeredrhs = mumps->id.nrhs * mumps->id.lredrhs;
324: }
325: mumps->id.ICNTL(26) = 1; /* condensation phase */
326: } else { /* prepare for the expansion step */
327: /* solve Schur complement (this has to be done by the MUMPS user, so basically us) */
328: PetscCall(MatMumpsSolveSchur_Private(F));
329: mumps->id.ICNTL(26) = 2; /* expansion phase */
330: PetscMUMPS_c(mumps);
331: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in solve phase: INFOG(1)=%d", mumps->id.INFOG(1));
332: /* restore defaults */
333: mumps->id.ICNTL(26) = -1;
334: /* free MUMPS internal array for redrhs if we have solved for multiple rhs in order to save memory space */
335: if (mumps->id.nrhs > 1) {
336: PetscCall(PetscFree(mumps->id.redrhs));
337: mumps->id.lredrhs = 0;
338: mumps->sizeredrhs = 0;
339: }
340: }
341: PetscFunctionReturn(PETSC_SUCCESS);
342: }
344: /*
345: MatConvertToTriples_A_B - convert Petsc matrix to triples: row[nz], col[nz], val[nz]
347: input:
348: A - matrix in aij,baij or sbaij format
349: shift - 0: C style output triple; 1: Fortran style output triple.
350: reuse - MAT_INITIAL_MATRIX: spaces are allocated and values are set for the triple
351: MAT_REUSE_MATRIX: only the values in v array are updated
352: output:
353: nnz - dim of r, c, and v (number of local nonzero entries of A)
354: r, c, v - row and col index, matrix values (matrix triples)
356: The returned values r, c, and sometimes v are obtained in a single PetscMalloc(). Then in MatDestroy_MUMPS() it is
357: freed with PetscFree(mumps->irn); This is not ideal code, the fact that v is ONLY sometimes part of mumps->irn means
358: that the PetscMalloc() cannot easily be replaced with a PetscMalloc3().
360: */
362: PetscErrorCode MatConvertToTriples_seqaij_seqaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
363: {
364: const PetscScalar *av;
365: const PetscInt *ai, *aj, *ajj, M = A->rmap->n;
366: PetscInt64 nz, rnz, i, j, k;
367: PetscMUMPSInt *row, *col;
368: Mat_SeqAIJ *aa = (Mat_SeqAIJ *)A->data;
370: PetscFunctionBegin;
371: PetscCall(MatSeqAIJGetArrayRead(A, &av));
372: mumps->val = (PetscScalar *)av;
373: if (reuse == MAT_INITIAL_MATRIX) {
374: nz = aa->nz;
375: ai = aa->i;
376: aj = aa->j;
377: PetscCall(PetscMalloc2(nz, &row, nz, &col));
378: for (i = k = 0; i < M; i++) {
379: rnz = ai[i + 1] - ai[i];
380: ajj = aj + ai[i];
381: for (j = 0; j < rnz; j++) {
382: PetscCall(PetscMUMPSIntCast(i + shift, &row[k]));
383: PetscCall(PetscMUMPSIntCast(ajj[j] + shift, &col[k]));
384: k++;
385: }
386: }
387: mumps->irn = row;
388: mumps->jcn = col;
389: mumps->nnz = nz;
390: }
391: PetscCall(MatSeqAIJRestoreArrayRead(A, &av));
392: PetscFunctionReturn(PETSC_SUCCESS);
393: }
395: PetscErrorCode MatConvertToTriples_seqsell_seqaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
396: {
397: PetscInt64 nz, i, j, k, r;
398: Mat_SeqSELL *a = (Mat_SeqSELL *)A->data;
399: PetscMUMPSInt *row, *col;
401: PetscFunctionBegin;
402: mumps->val = a->val;
403: if (reuse == MAT_INITIAL_MATRIX) {
404: nz = a->sliidx[a->totalslices];
405: PetscCall(PetscMalloc2(nz, &row, nz, &col));
406: for (i = k = 0; i < a->totalslices; i++) {
407: for (j = a->sliidx[i], r = 0; j < a->sliidx[i + 1]; j++, r = ((r + 1) & 0x07)) PetscCall(PetscMUMPSIntCast(8 * i + r + shift, &row[k++]));
408: }
409: for (i = 0; i < nz; i++) PetscCall(PetscMUMPSIntCast(a->colidx[i] + shift, &col[i]));
410: mumps->irn = row;
411: mumps->jcn = col;
412: mumps->nnz = nz;
413: }
414: PetscFunctionReturn(PETSC_SUCCESS);
415: }
417: PetscErrorCode MatConvertToTriples_seqbaij_seqaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
418: {
419: Mat_SeqBAIJ *aa = (Mat_SeqBAIJ *)A->data;
420: const PetscInt *ai, *aj, *ajj, bs2 = aa->bs2;
421: PetscInt64 M, nz, idx = 0, rnz, i, j, k, m;
422: PetscInt bs;
423: PetscMUMPSInt *row, *col;
425: PetscFunctionBegin;
426: PetscCall(MatGetBlockSize(A, &bs));
427: M = A->rmap->N / bs;
428: mumps->val = aa->a;
429: if (reuse == MAT_INITIAL_MATRIX) {
430: ai = aa->i;
431: aj = aa->j;
432: nz = bs2 * aa->nz;
433: PetscCall(PetscMalloc2(nz, &row, nz, &col));
434: for (i = 0; i < M; i++) {
435: ajj = aj + ai[i];
436: rnz = ai[i + 1] - ai[i];
437: for (k = 0; k < rnz; k++) {
438: for (j = 0; j < bs; j++) {
439: for (m = 0; m < bs; m++) {
440: PetscCall(PetscMUMPSIntCast(i * bs + m + shift, &row[idx]));
441: PetscCall(PetscMUMPSIntCast(bs * ajj[k] + j + shift, &col[idx]));
442: idx++;
443: }
444: }
445: }
446: }
447: mumps->irn = row;
448: mumps->jcn = col;
449: mumps->nnz = nz;
450: }
451: PetscFunctionReturn(PETSC_SUCCESS);
452: }
454: PetscErrorCode MatConvertToTriples_seqsbaij_seqsbaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
455: {
456: const PetscInt *ai, *aj, *ajj;
457: PetscInt bs;
458: PetscInt64 nz, rnz, i, j, k, m;
459: PetscMUMPSInt *row, *col;
460: PetscScalar *val;
461: Mat_SeqSBAIJ *aa = (Mat_SeqSBAIJ *)A->data;
462: const PetscInt bs2 = aa->bs2, mbs = aa->mbs;
463: #if defined(PETSC_USE_COMPLEX)
464: PetscBool isset, hermitian;
465: #endif
467: PetscFunctionBegin;
468: #if defined(PETSC_USE_COMPLEX)
469: PetscCall(MatIsHermitianKnown(A, &isset, &hermitian));
470: PetscCheck(!isset || !hermitian, PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "MUMPS does not support Hermitian symmetric matrices for Choleksy");
471: #endif
472: ai = aa->i;
473: aj = aa->j;
474: PetscCall(MatGetBlockSize(A, &bs));
475: if (reuse == MAT_INITIAL_MATRIX) {
476: const PetscInt64 alloc_size = aa->nz * bs2;
478: PetscCall(PetscMalloc2(alloc_size, &row, alloc_size, &col));
479: if (bs > 1) {
480: PetscCall(PetscMalloc1(alloc_size, &mumps->val_alloc));
481: mumps->val = mumps->val_alloc;
482: } else {
483: mumps->val = aa->a;
484: }
485: mumps->irn = row;
486: mumps->jcn = col;
487: } else {
488: if (bs == 1) mumps->val = aa->a;
489: row = mumps->irn;
490: col = mumps->jcn;
491: }
492: val = mumps->val;
494: nz = 0;
495: if (bs > 1) {
496: for (i = 0; i < mbs; i++) {
497: rnz = ai[i + 1] - ai[i];
498: ajj = aj + ai[i];
499: for (j = 0; j < rnz; j++) {
500: for (k = 0; k < bs; k++) {
501: for (m = 0; m < bs; m++) {
502: if (ajj[j] > i || k >= m) {
503: if (reuse == MAT_INITIAL_MATRIX) {
504: PetscCall(PetscMUMPSIntCast(i * bs + m + shift, &row[nz]));
505: PetscCall(PetscMUMPSIntCast(ajj[j] * bs + k + shift, &col[nz]));
506: }
507: val[nz++] = aa->a[(ai[i] + j) * bs2 + m + k * bs];
508: }
509: }
510: }
511: }
512: }
513: } else if (reuse == MAT_INITIAL_MATRIX) {
514: for (i = 0; i < mbs; i++) {
515: rnz = ai[i + 1] - ai[i];
516: ajj = aj + ai[i];
517: for (j = 0; j < rnz; j++) {
518: PetscCall(PetscMUMPSIntCast(i + shift, &row[nz]));
519: PetscCall(PetscMUMPSIntCast(ajj[j] + shift, &col[nz]));
520: nz++;
521: }
522: }
523: PetscCheck(nz == aa->nz, PETSC_COMM_SELF, PETSC_ERR_PLIB, "Different numbers of nonzeros %" PetscInt64_FMT " != %" PetscInt_FMT, nz, aa->nz);
524: }
525: if (reuse == MAT_INITIAL_MATRIX) mumps->nnz = nz;
526: PetscFunctionReturn(PETSC_SUCCESS);
527: }
529: PetscErrorCode MatConvertToTriples_seqaij_seqsbaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
530: {
531: const PetscInt *ai, *aj, *ajj, *adiag, M = A->rmap->n;
532: PetscInt64 nz, rnz, i, j;
533: const PetscScalar *av, *v1;
534: PetscScalar *val;
535: PetscMUMPSInt *row, *col;
536: Mat_SeqAIJ *aa = (Mat_SeqAIJ *)A->data;
537: PetscBool missing;
538: #if defined(PETSC_USE_COMPLEX)
539: PetscBool hermitian, isset;
540: #endif
542: PetscFunctionBegin;
543: #if defined(PETSC_USE_COMPLEX)
544: PetscCall(MatIsHermitianKnown(A, &isset, &hermitian));
545: PetscCheck(!isset || !hermitian, PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "MUMPS does not support Hermitian symmetric matrices for Choleksy");
546: #endif
547: PetscCall(MatSeqAIJGetArrayRead(A, &av));
548: ai = aa->i;
549: aj = aa->j;
550: adiag = aa->diag;
551: PetscCall(MatMissingDiagonal_SeqAIJ(A, &missing, NULL));
552: if (reuse == MAT_INITIAL_MATRIX) {
553: /* count nz in the upper triangular part of A */
554: nz = 0;
555: if (missing) {
556: for (i = 0; i < M; i++) {
557: if (PetscUnlikely(adiag[i] >= ai[i + 1])) {
558: for (j = ai[i]; j < ai[i + 1]; j++) {
559: if (aj[j] < i) continue;
560: nz++;
561: }
562: } else {
563: nz += ai[i + 1] - adiag[i];
564: }
565: }
566: } else {
567: for (i = 0; i < M; i++) nz += ai[i + 1] - adiag[i];
568: }
569: PetscCall(PetscMalloc2(nz, &row, nz, &col));
570: PetscCall(PetscMalloc1(nz, &val));
571: mumps->nnz = nz;
572: mumps->irn = row;
573: mumps->jcn = col;
574: mumps->val = mumps->val_alloc = val;
576: nz = 0;
577: if (missing) {
578: for (i = 0; i < M; i++) {
579: if (PetscUnlikely(adiag[i] >= ai[i + 1])) {
580: for (j = ai[i]; j < ai[i + 1]; j++) {
581: if (aj[j] < i) continue;
582: PetscCall(PetscMUMPSIntCast(i + shift, &row[nz]));
583: PetscCall(PetscMUMPSIntCast(aj[j] + shift, &col[nz]));
584: val[nz] = av[j];
585: nz++;
586: }
587: } else {
588: rnz = ai[i + 1] - adiag[i];
589: ajj = aj + adiag[i];
590: v1 = av + adiag[i];
591: for (j = 0; j < rnz; j++) {
592: PetscCall(PetscMUMPSIntCast(i + shift, &row[nz]));
593: PetscCall(PetscMUMPSIntCast(ajj[j] + shift, &col[nz]));
594: val[nz++] = v1[j];
595: }
596: }
597: }
598: } else {
599: for (i = 0; i < M; i++) {
600: rnz = ai[i + 1] - adiag[i];
601: ajj = aj + adiag[i];
602: v1 = av + adiag[i];
603: for (j = 0; j < rnz; j++) {
604: PetscCall(PetscMUMPSIntCast(i + shift, &row[nz]));
605: PetscCall(PetscMUMPSIntCast(ajj[j] + shift, &col[nz]));
606: val[nz++] = v1[j];
607: }
608: }
609: }
610: } else {
611: nz = 0;
612: val = mumps->val;
613: if (missing) {
614: for (i = 0; i < M; i++) {
615: if (PetscUnlikely(adiag[i] >= ai[i + 1])) {
616: for (j = ai[i]; j < ai[i + 1]; j++) {
617: if (aj[j] < i) continue;
618: val[nz++] = av[j];
619: }
620: } else {
621: rnz = ai[i + 1] - adiag[i];
622: v1 = av + adiag[i];
623: for (j = 0; j < rnz; j++) val[nz++] = v1[j];
624: }
625: }
626: } else {
627: for (i = 0; i < M; i++) {
628: rnz = ai[i + 1] - adiag[i];
629: v1 = av + adiag[i];
630: for (j = 0; j < rnz; j++) val[nz++] = v1[j];
631: }
632: }
633: }
634: PetscCall(MatSeqAIJRestoreArrayRead(A, &av));
635: PetscFunctionReturn(PETSC_SUCCESS);
636: }
638: PetscErrorCode MatConvertToTriples_mpisbaij_mpisbaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
639: {
640: const PetscInt *ai, *aj, *bi, *bj, *garray, *ajj, *bjj;
641: PetscInt bs;
642: PetscInt64 rstart, nz, i, j, k, m, jj, irow, countA, countB;
643: PetscMUMPSInt *row, *col;
644: const PetscScalar *av, *bv, *v1, *v2;
645: PetscScalar *val;
646: Mat_MPISBAIJ *mat = (Mat_MPISBAIJ *)A->data;
647: Mat_SeqSBAIJ *aa = (Mat_SeqSBAIJ *)(mat->A)->data;
648: Mat_SeqBAIJ *bb = (Mat_SeqBAIJ *)(mat->B)->data;
649: const PetscInt bs2 = aa->bs2, mbs = aa->mbs;
650: #if defined(PETSC_USE_COMPLEX)
651: PetscBool hermitian, isset;
652: #endif
654: PetscFunctionBegin;
655: #if defined(PETSC_USE_COMPLEX)
656: PetscCall(MatIsHermitianKnown(A, &isset, &hermitian));
657: PetscCheck(!isset || !hermitian, PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "MUMPS does not support Hermitian symmetric matrices for Choleksy");
658: #endif
659: PetscCall(MatGetBlockSize(A, &bs));
660: rstart = A->rmap->rstart;
661: ai = aa->i;
662: aj = aa->j;
663: bi = bb->i;
664: bj = bb->j;
665: av = aa->a;
666: bv = bb->a;
668: garray = mat->garray;
670: if (reuse == MAT_INITIAL_MATRIX) {
671: nz = (aa->nz + bb->nz) * bs2; /* just a conservative estimate */
672: PetscCall(PetscMalloc2(nz, &row, nz, &col));
673: PetscCall(PetscMalloc1(nz, &val));
674: /* can not decide the exact mumps->nnz now because of the SBAIJ */
675: mumps->irn = row;
676: mumps->jcn = col;
677: mumps->val = mumps->val_alloc = val;
678: } else {
679: val = mumps->val;
680: }
682: jj = 0;
683: irow = rstart;
684: for (i = 0; i < mbs; i++) {
685: ajj = aj + ai[i]; /* ptr to the beginning of this row */
686: countA = ai[i + 1] - ai[i];
687: countB = bi[i + 1] - bi[i];
688: bjj = bj + bi[i];
689: v1 = av + ai[i] * bs2;
690: v2 = bv + bi[i] * bs2;
692: if (bs > 1) {
693: /* A-part */
694: for (j = 0; j < countA; j++) {
695: for (k = 0; k < bs; k++) {
696: for (m = 0; m < bs; m++) {
697: if (rstart + ajj[j] * bs > irow || k >= m) {
698: if (reuse == MAT_INITIAL_MATRIX) {
699: PetscCall(PetscMUMPSIntCast(irow + m + shift, &row[jj]));
700: PetscCall(PetscMUMPSIntCast(rstart + ajj[j] * bs + k + shift, &col[jj]));
701: }
702: val[jj++] = v1[j * bs2 + m + k * bs];
703: }
704: }
705: }
706: }
708: /* B-part */
709: for (j = 0; j < countB; j++) {
710: for (k = 0; k < bs; k++) {
711: for (m = 0; m < bs; m++) {
712: if (reuse == MAT_INITIAL_MATRIX) {
713: PetscCall(PetscMUMPSIntCast(irow + m + shift, &row[jj]));
714: PetscCall(PetscMUMPSIntCast(garray[bjj[j]] * bs + k + shift, &col[jj]));
715: }
716: val[jj++] = v2[j * bs2 + m + k * bs];
717: }
718: }
719: }
720: } else {
721: /* A-part */
722: for (j = 0; j < countA; j++) {
723: if (reuse == MAT_INITIAL_MATRIX) {
724: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
725: PetscCall(PetscMUMPSIntCast(rstart + ajj[j] + shift, &col[jj]));
726: }
727: val[jj++] = v1[j];
728: }
730: /* B-part */
731: for (j = 0; j < countB; j++) {
732: if (reuse == MAT_INITIAL_MATRIX) {
733: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
734: PetscCall(PetscMUMPSIntCast(garray[bjj[j]] + shift, &col[jj]));
735: }
736: val[jj++] = v2[j];
737: }
738: }
739: irow += bs;
740: }
741: mumps->nnz = jj;
742: PetscFunctionReturn(PETSC_SUCCESS);
743: }
745: PetscErrorCode MatConvertToTriples_mpiaij_mpiaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
746: {
747: const PetscInt *ai, *aj, *bi, *bj, *garray, m = A->rmap->n, *ajj, *bjj;
748: PetscInt64 rstart, nz, i, j, jj, irow, countA, countB;
749: PetscMUMPSInt *row, *col;
750: const PetscScalar *av, *bv, *v1, *v2;
751: PetscScalar *val;
752: Mat Ad, Ao;
753: Mat_SeqAIJ *aa;
754: Mat_SeqAIJ *bb;
756: PetscFunctionBegin;
757: PetscCall(MatMPIAIJGetSeqAIJ(A, &Ad, &Ao, &garray));
758: PetscCall(MatSeqAIJGetArrayRead(Ad, &av));
759: PetscCall(MatSeqAIJGetArrayRead(Ao, &bv));
761: aa = (Mat_SeqAIJ *)(Ad)->data;
762: bb = (Mat_SeqAIJ *)(Ao)->data;
763: ai = aa->i;
764: aj = aa->j;
765: bi = bb->i;
766: bj = bb->j;
768: rstart = A->rmap->rstart;
770: if (reuse == MAT_INITIAL_MATRIX) {
771: nz = (PetscInt64)aa->nz + bb->nz; /* make sure the sum won't overflow PetscInt */
772: PetscCall(PetscMalloc2(nz, &row, nz, &col));
773: PetscCall(PetscMalloc1(nz, &val));
774: mumps->nnz = nz;
775: mumps->irn = row;
776: mumps->jcn = col;
777: mumps->val = mumps->val_alloc = val;
778: } else {
779: val = mumps->val;
780: }
782: jj = 0;
783: irow = rstart;
784: for (i = 0; i < m; i++) {
785: ajj = aj + ai[i]; /* ptr to the beginning of this row */
786: countA = ai[i + 1] - ai[i];
787: countB = bi[i + 1] - bi[i];
788: bjj = bj + bi[i];
789: v1 = av + ai[i];
790: v2 = bv + bi[i];
792: /* A-part */
793: for (j = 0; j < countA; j++) {
794: if (reuse == MAT_INITIAL_MATRIX) {
795: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
796: PetscCall(PetscMUMPSIntCast(rstart + ajj[j] + shift, &col[jj]));
797: }
798: val[jj++] = v1[j];
799: }
801: /* B-part */
802: for (j = 0; j < countB; j++) {
803: if (reuse == MAT_INITIAL_MATRIX) {
804: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
805: PetscCall(PetscMUMPSIntCast(garray[bjj[j]] + shift, &col[jj]));
806: }
807: val[jj++] = v2[j];
808: }
809: irow++;
810: }
811: PetscCall(MatSeqAIJRestoreArrayRead(Ad, &av));
812: PetscCall(MatSeqAIJRestoreArrayRead(Ao, &bv));
813: PetscFunctionReturn(PETSC_SUCCESS);
814: }
816: PetscErrorCode MatConvertToTriples_mpibaij_mpiaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
817: {
818: Mat_MPIBAIJ *mat = (Mat_MPIBAIJ *)A->data;
819: Mat_SeqBAIJ *aa = (Mat_SeqBAIJ *)(mat->A)->data;
820: Mat_SeqBAIJ *bb = (Mat_SeqBAIJ *)(mat->B)->data;
821: const PetscInt *ai = aa->i, *bi = bb->i, *aj = aa->j, *bj = bb->j, *ajj, *bjj;
822: const PetscInt *garray = mat->garray, mbs = mat->mbs, rstart = A->rmap->rstart;
823: const PetscInt bs2 = mat->bs2;
824: PetscInt bs;
825: PetscInt64 nz, i, j, k, n, jj, irow, countA, countB, idx;
826: PetscMUMPSInt *row, *col;
827: const PetscScalar *av = aa->a, *bv = bb->a, *v1, *v2;
828: PetscScalar *val;
830: PetscFunctionBegin;
831: PetscCall(MatGetBlockSize(A, &bs));
832: if (reuse == MAT_INITIAL_MATRIX) {
833: nz = bs2 * (aa->nz + bb->nz);
834: PetscCall(PetscMalloc2(nz, &row, nz, &col));
835: PetscCall(PetscMalloc1(nz, &val));
836: mumps->nnz = nz;
837: mumps->irn = row;
838: mumps->jcn = col;
839: mumps->val = mumps->val_alloc = val;
840: } else {
841: val = mumps->val;
842: }
844: jj = 0;
845: irow = rstart;
846: for (i = 0; i < mbs; i++) {
847: countA = ai[i + 1] - ai[i];
848: countB = bi[i + 1] - bi[i];
849: ajj = aj + ai[i];
850: bjj = bj + bi[i];
851: v1 = av + bs2 * ai[i];
852: v2 = bv + bs2 * bi[i];
854: idx = 0;
855: /* A-part */
856: for (k = 0; k < countA; k++) {
857: for (j = 0; j < bs; j++) {
858: for (n = 0; n < bs; n++) {
859: if (reuse == MAT_INITIAL_MATRIX) {
860: PetscCall(PetscMUMPSIntCast(irow + n + shift, &row[jj]));
861: PetscCall(PetscMUMPSIntCast(rstart + bs * ajj[k] + j + shift, &col[jj]));
862: }
863: val[jj++] = v1[idx++];
864: }
865: }
866: }
868: idx = 0;
869: /* B-part */
870: for (k = 0; k < countB; k++) {
871: for (j = 0; j < bs; j++) {
872: for (n = 0; n < bs; n++) {
873: if (reuse == MAT_INITIAL_MATRIX) {
874: PetscCall(PetscMUMPSIntCast(irow + n + shift, &row[jj]));
875: PetscCall(PetscMUMPSIntCast(bs * garray[bjj[k]] + j + shift, &col[jj]));
876: }
877: val[jj++] = v2[idx++];
878: }
879: }
880: }
881: irow += bs;
882: }
883: PetscFunctionReturn(PETSC_SUCCESS);
884: }
886: PetscErrorCode MatConvertToTriples_mpiaij_mpisbaij(Mat A, PetscInt shift, MatReuse reuse, Mat_MUMPS *mumps)
887: {
888: const PetscInt *ai, *aj, *adiag, *bi, *bj, *garray, m = A->rmap->n, *ajj, *bjj;
889: PetscInt64 rstart, nz, nza, nzb, i, j, jj, irow, countA, countB;
890: PetscMUMPSInt *row, *col;
891: const PetscScalar *av, *bv, *v1, *v2;
892: PetscScalar *val;
893: Mat Ad, Ao;
894: Mat_SeqAIJ *aa;
895: Mat_SeqAIJ *bb;
896: #if defined(PETSC_USE_COMPLEX)
897: PetscBool hermitian, isset;
898: #endif
900: PetscFunctionBegin;
901: #if defined(PETSC_USE_COMPLEX)
902: PetscCall(MatIsHermitianKnown(A, &isset, &hermitian));
903: PetscCheck(!isset || !hermitian, PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "MUMPS does not support Hermitian symmetric matrices for Choleksy");
904: #endif
905: PetscCall(MatMPIAIJGetSeqAIJ(A, &Ad, &Ao, &garray));
906: PetscCall(MatSeqAIJGetArrayRead(Ad, &av));
907: PetscCall(MatSeqAIJGetArrayRead(Ao, &bv));
909: aa = (Mat_SeqAIJ *)(Ad)->data;
910: bb = (Mat_SeqAIJ *)(Ao)->data;
911: ai = aa->i;
912: aj = aa->j;
913: adiag = aa->diag;
914: bi = bb->i;
915: bj = bb->j;
917: rstart = A->rmap->rstart;
919: if (reuse == MAT_INITIAL_MATRIX) {
920: nza = 0; /* num of upper triangular entries in mat->A, including diagonals */
921: nzb = 0; /* num of upper triangular entries in mat->B */
922: for (i = 0; i < m; i++) {
923: nza += (ai[i + 1] - adiag[i]);
924: countB = bi[i + 1] - bi[i];
925: bjj = bj + bi[i];
926: for (j = 0; j < countB; j++) {
927: if (garray[bjj[j]] > rstart) nzb++;
928: }
929: }
931: nz = nza + nzb; /* total nz of upper triangular part of mat */
932: PetscCall(PetscMalloc2(nz, &row, nz, &col));
933: PetscCall(PetscMalloc1(nz, &val));
934: mumps->nnz = nz;
935: mumps->irn = row;
936: mumps->jcn = col;
937: mumps->val = mumps->val_alloc = val;
938: } else {
939: val = mumps->val;
940: }
942: jj = 0;
943: irow = rstart;
944: for (i = 0; i < m; i++) {
945: ajj = aj + adiag[i]; /* ptr to the beginning of the diagonal of this row */
946: v1 = av + adiag[i];
947: countA = ai[i + 1] - adiag[i];
948: countB = bi[i + 1] - bi[i];
949: bjj = bj + bi[i];
950: v2 = bv + bi[i];
952: /* A-part */
953: for (j = 0; j < countA; j++) {
954: if (reuse == MAT_INITIAL_MATRIX) {
955: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
956: PetscCall(PetscMUMPSIntCast(rstart + ajj[j] + shift, &col[jj]));
957: }
958: val[jj++] = v1[j];
959: }
961: /* B-part */
962: for (j = 0; j < countB; j++) {
963: if (garray[bjj[j]] > rstart) {
964: if (reuse == MAT_INITIAL_MATRIX) {
965: PetscCall(PetscMUMPSIntCast(irow + shift, &row[jj]));
966: PetscCall(PetscMUMPSIntCast(garray[bjj[j]] + shift, &col[jj]));
967: }
968: val[jj++] = v2[j];
969: }
970: }
971: irow++;
972: }
973: PetscCall(MatSeqAIJRestoreArrayRead(Ad, &av));
974: PetscCall(MatSeqAIJRestoreArrayRead(Ao, &bv));
975: PetscFunctionReturn(PETSC_SUCCESS);
976: }
978: PetscErrorCode MatDestroy_MUMPS(Mat A)
979: {
980: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
982: PetscFunctionBegin;
983: PetscCall(PetscFree2(mumps->id.sol_loc, mumps->id.isol_loc));
984: PetscCall(VecScatterDestroy(&mumps->scat_rhs));
985: PetscCall(VecScatterDestroy(&mumps->scat_sol));
986: PetscCall(VecDestroy(&mumps->b_seq));
987: PetscCall(VecDestroy(&mumps->x_seq));
988: PetscCall(PetscFree(mumps->id.perm_in));
989: PetscCall(PetscFree2(mumps->irn, mumps->jcn));
990: PetscCall(PetscFree(mumps->val_alloc));
991: PetscCall(PetscFree(mumps->info));
992: PetscCall(PetscFree(mumps->ICNTL_pre));
993: PetscCall(PetscFree(mumps->CNTL_pre));
994: PetscCall(MatMumpsResetSchur_Private(mumps));
995: if (mumps->id.job != JOB_NULL) { /* cannot call PetscMUMPS_c() if JOB_INIT has never been called for this instance */
996: mumps->id.job = JOB_END;
997: PetscMUMPS_c(mumps);
998: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in MatDestroy_MUMPS: INFOG(1)=%d", mumps->id.INFOG(1));
999: if (mumps->mumps_comm != MPI_COMM_NULL) {
1000: if (PetscDefined(HAVE_OPENMP_SUPPORT) && mumps->use_petsc_omp_support) PetscCallMPI(MPI_Comm_free(&mumps->mumps_comm));
1001: else PetscCall(PetscCommRestoreComm(PetscObjectComm((PetscObject)A), &mumps->mumps_comm));
1002: }
1003: }
1004: #if defined(PETSC_HAVE_OPENMP_SUPPORT)
1005: if (mumps->use_petsc_omp_support) {
1006: PetscCall(PetscOmpCtrlDestroy(&mumps->omp_ctrl));
1007: PetscCall(PetscFree2(mumps->rhs_loc, mumps->rhs_recvbuf));
1008: PetscCall(PetscFree3(mumps->rhs_nrow, mumps->rhs_recvcounts, mumps->rhs_disps));
1009: }
1010: #endif
1011: PetscCall(PetscFree(mumps->ia_alloc));
1012: PetscCall(PetscFree(mumps->ja_alloc));
1013: PetscCall(PetscFree(mumps->recvcount));
1014: PetscCall(PetscFree(mumps->reqs));
1015: PetscCall(PetscFree(mumps->irhs_loc));
1016: PetscCall(PetscFree(A->data));
1018: /* clear composed functions */
1019: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatFactorGetSolverType_C", NULL));
1020: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatFactorSetSchurIS_C", NULL));
1021: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatFactorCreateSchurComplement_C", NULL));
1022: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsSetIcntl_C", NULL));
1023: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetIcntl_C", NULL));
1024: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsSetCntl_C", NULL));
1025: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetCntl_C", NULL));
1026: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetInfo_C", NULL));
1027: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetInfog_C", NULL));
1028: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetRinfo_C", NULL));
1029: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetRinfog_C", NULL));
1030: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetNullPivots_C", NULL));
1031: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetInverse_C", NULL));
1032: PetscCall(PetscObjectComposeFunction((PetscObject)A, "MatMumpsGetInverseTranspose_C", NULL));
1033: PetscFunctionReturn(PETSC_SUCCESS);
1034: }
1036: /* Set up the distributed RHS info for MUMPS. <nrhs> is the number of RHS. <array> points to start of RHS on the local processor. */
1037: static PetscErrorCode MatMumpsSetUpDistRHSInfo(Mat A, PetscInt nrhs, const PetscScalar *array)
1038: {
1039: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
1040: const PetscMPIInt ompsize = mumps->omp_comm_size;
1041: PetscInt i, m, M, rstart;
1043: PetscFunctionBegin;
1044: PetscCall(MatGetSize(A, &M, NULL));
1045: PetscCall(MatGetLocalSize(A, &m, NULL));
1046: PetscCheck(M <= PETSC_MUMPS_INT_MAX, PETSC_COMM_SELF, PETSC_ERR_ARG_OUTOFRANGE, "PetscInt too long for PetscMUMPSInt");
1047: if (ompsize == 1) {
1048: if (!mumps->irhs_loc) {
1049: mumps->nloc_rhs = m;
1050: PetscCall(PetscMalloc1(m, &mumps->irhs_loc));
1051: PetscCall(MatGetOwnershipRange(A, &rstart, NULL));
1052: for (i = 0; i < m; i++) mumps->irhs_loc[i] = rstart + i + 1; /* use 1-based indices */
1053: }
1054: mumps->id.rhs_loc = (MumpsScalar *)array;
1055: } else {
1056: #if defined(PETSC_HAVE_OPENMP_SUPPORT)
1057: const PetscInt *ranges;
1058: PetscMPIInt j, k, sendcount, *petsc_ranks, *omp_ranks;
1059: MPI_Group petsc_group, omp_group;
1060: PetscScalar *recvbuf = NULL;
1062: if (mumps->is_omp_master) {
1063: /* Lazily initialize the omp stuff for distributed rhs */
1064: if (!mumps->irhs_loc) {
1065: PetscCall(PetscMalloc2(ompsize, &omp_ranks, ompsize, &petsc_ranks));
1066: PetscCall(PetscMalloc3(ompsize, &mumps->rhs_nrow, ompsize, &mumps->rhs_recvcounts, ompsize, &mumps->rhs_disps));
1067: PetscCallMPI(MPI_Comm_group(mumps->petsc_comm, &petsc_group));
1068: PetscCallMPI(MPI_Comm_group(mumps->omp_comm, &omp_group));
1069: for (j = 0; j < ompsize; j++) omp_ranks[j] = j;
1070: PetscCallMPI(MPI_Group_translate_ranks(omp_group, ompsize, omp_ranks, petsc_group, petsc_ranks));
1072: /* Populate mumps->irhs_loc[], rhs_nrow[] */
1073: mumps->nloc_rhs = 0;
1074: PetscCall(MatGetOwnershipRanges(A, &ranges));
1075: for (j = 0; j < ompsize; j++) {
1076: mumps->rhs_nrow[j] = ranges[petsc_ranks[j] + 1] - ranges[petsc_ranks[j]];
1077: mumps->nloc_rhs += mumps->rhs_nrow[j];
1078: }
1079: PetscCall(PetscMalloc1(mumps->nloc_rhs, &mumps->irhs_loc));
1080: for (j = k = 0; j < ompsize; j++) {
1081: for (i = ranges[petsc_ranks[j]]; i < ranges[petsc_ranks[j] + 1]; i++, k++) mumps->irhs_loc[k] = i + 1; /* uses 1-based indices */
1082: }
1084: PetscCall(PetscFree2(omp_ranks, petsc_ranks));
1085: PetscCallMPI(MPI_Group_free(&petsc_group));
1086: PetscCallMPI(MPI_Group_free(&omp_group));
1087: }
1089: /* Realloc buffers when current nrhs is bigger than what we have met */
1090: if (nrhs > mumps->max_nrhs) {
1091: PetscCall(PetscFree2(mumps->rhs_loc, mumps->rhs_recvbuf));
1092: PetscCall(PetscMalloc2(mumps->nloc_rhs * nrhs, &mumps->rhs_loc, mumps->nloc_rhs * nrhs, &mumps->rhs_recvbuf));
1093: mumps->max_nrhs = nrhs;
1094: }
1096: /* Setup recvcounts[], disps[], recvbuf on omp rank 0 for the upcoming MPI_Gatherv */
1097: for (j = 0; j < ompsize; j++) PetscCall(PetscMPIIntCast(mumps->rhs_nrow[j] * nrhs, &mumps->rhs_recvcounts[j]));
1098: mumps->rhs_disps[0] = 0;
1099: for (j = 1; j < ompsize; j++) {
1100: mumps->rhs_disps[j] = mumps->rhs_disps[j - 1] + mumps->rhs_recvcounts[j - 1];
1101: PetscCheck(mumps->rhs_disps[j] >= 0, PETSC_COMM_SELF, PETSC_ERR_ARG_OUTOFRANGE, "PetscMPIInt overflow!");
1102: }
1103: recvbuf = (nrhs == 1) ? mumps->rhs_loc : mumps->rhs_recvbuf; /* Directly use rhs_loc[] as recvbuf. Single rhs is common in Ax=b */
1104: }
1106: PetscCall(PetscMPIIntCast(m * nrhs, &sendcount));
1107: PetscCallMPI(MPI_Gatherv(array, sendcount, MPIU_SCALAR, recvbuf, mumps->rhs_recvcounts, mumps->rhs_disps, MPIU_SCALAR, 0, mumps->omp_comm));
1109: if (mumps->is_omp_master) {
1110: if (nrhs > 1) { /* Copy & re-arrange data from rhs_recvbuf[] to mumps->rhs_loc[] only when there are multiple rhs */
1111: PetscScalar *dst, *dstbase = mumps->rhs_loc;
1112: for (j = 0; j < ompsize; j++) {
1113: const PetscScalar *src = mumps->rhs_recvbuf + mumps->rhs_disps[j];
1114: dst = dstbase;
1115: for (i = 0; i < nrhs; i++) {
1116: PetscCall(PetscArraycpy(dst, src, mumps->rhs_nrow[j]));
1117: src += mumps->rhs_nrow[j];
1118: dst += mumps->nloc_rhs;
1119: }
1120: dstbase += mumps->rhs_nrow[j];
1121: }
1122: }
1123: mumps->id.rhs_loc = (MumpsScalar *)mumps->rhs_loc;
1124: }
1125: #endif /* PETSC_HAVE_OPENMP_SUPPORT */
1126: }
1127: mumps->id.nrhs = nrhs;
1128: mumps->id.nloc_rhs = mumps->nloc_rhs;
1129: mumps->id.lrhs_loc = mumps->nloc_rhs;
1130: mumps->id.irhs_loc = mumps->irhs_loc;
1131: PetscFunctionReturn(PETSC_SUCCESS);
1132: }
1134: PetscErrorCode MatSolve_MUMPS(Mat A, Vec b, Vec x)
1135: {
1136: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
1137: const PetscScalar *rarray = NULL;
1138: PetscScalar *array;
1139: IS is_iden, is_petsc;
1140: PetscInt i;
1141: PetscBool second_solve = PETSC_FALSE;
1142: static PetscBool cite1 = PETSC_FALSE, cite2 = PETSC_FALSE;
1144: PetscFunctionBegin;
1145: PetscCall(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 "
1146: "Journal on Matrix Analysis and Applications},\n volume = {23},\n number = {1},\n pages = {15--41},\n year = {2001}\n}\n",
1147: &cite1));
1148: PetscCall(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 "
1149: "Computing},\n volume = {32},\n number = {2},\n pages = {136--156},\n year = {2006}\n}\n",
1150: &cite2));
1152: if (A->factorerrortype) {
1153: PetscCall(PetscInfo(A, "MatSolve is called with singular matrix factor, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1154: PetscCall(VecSetInf(x));
1155: PetscFunctionReturn(PETSC_SUCCESS);
1156: }
1158: mumps->id.nrhs = 1;
1159: if (mumps->petsc_size > 1) {
1160: if (mumps->ICNTL20 == 10) {
1161: mumps->id.ICNTL(20) = 10; /* dense distributed RHS */
1162: PetscCall(VecGetArrayRead(b, &rarray));
1163: PetscCall(MatMumpsSetUpDistRHSInfo(A, 1, rarray));
1164: } else {
1165: mumps->id.ICNTL(20) = 0; /* dense centralized RHS; Scatter b into a sequential rhs vector*/
1166: PetscCall(VecScatterBegin(mumps->scat_rhs, b, mumps->b_seq, INSERT_VALUES, SCATTER_FORWARD));
1167: PetscCall(VecScatterEnd(mumps->scat_rhs, b, mumps->b_seq, INSERT_VALUES, SCATTER_FORWARD));
1168: if (!mumps->myid) {
1169: PetscCall(VecGetArray(mumps->b_seq, &array));
1170: mumps->id.rhs = (MumpsScalar *)array;
1171: }
1172: }
1173: } else { /* petsc_size == 1 */
1174: mumps->id.ICNTL(20) = 0; /* dense centralized RHS */
1175: PetscCall(VecCopy(b, x));
1176: PetscCall(VecGetArray(x, &array));
1177: mumps->id.rhs = (MumpsScalar *)array;
1178: }
1180: /*
1181: handle condensation step of Schur complement (if any)
1182: We set by default ICNTL(26) == -1 when Schur indices have been provided by the user.
1183: According to MUMPS (5.0.0) manual, any value should be harmful during the factorization phase
1184: Unless the user provides a valid value for ICNTL(26), MatSolve and MatMatSolve routines solve the full system.
1185: This requires an extra call to PetscMUMPS_c and the computation of the factors for S
1186: */
1187: if (mumps->id.size_schur > 0 && (mumps->id.ICNTL(26) < 0 || mumps->id.ICNTL(26) > 2)) {
1188: PetscCheck(mumps->petsc_size <= 1, PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "Parallel Schur complements not yet supported from PETSc");
1189: second_solve = PETSC_TRUE;
1190: PetscCall(MatMumpsHandleSchur_Private(A, PETSC_FALSE));
1191: }
1192: /* solve phase */
1193: mumps->id.job = JOB_SOLVE;
1194: PetscMUMPS_c(mumps);
1195: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in solve phase: INFOG(1)=%d", mumps->id.INFOG(1));
1197: /* handle expansion step of Schur complement (if any) */
1198: if (second_solve) PetscCall(MatMumpsHandleSchur_Private(A, PETSC_TRUE));
1200: if (mumps->petsc_size > 1) { /* convert mumps distributed solution to petsc mpi x */
1201: if (mumps->scat_sol && mumps->ICNTL9_pre != mumps->id.ICNTL(9)) {
1202: /* when id.ICNTL(9) changes, the contents of lsol_loc may change (not its size, lsol_loc), recreates scat_sol */
1203: PetscCall(VecScatterDestroy(&mumps->scat_sol));
1204: }
1205: if (!mumps->scat_sol) { /* create scatter scat_sol */
1206: PetscInt *isol2_loc = NULL;
1207: PetscCall(ISCreateStride(PETSC_COMM_SELF, mumps->id.lsol_loc, 0, 1, &is_iden)); /* from */
1208: PetscCall(PetscMalloc1(mumps->id.lsol_loc, &isol2_loc));
1209: for (i = 0; i < mumps->id.lsol_loc; i++) isol2_loc[i] = mumps->id.isol_loc[i] - 1; /* change Fortran style to C style */
1210: PetscCall(ISCreateGeneral(PETSC_COMM_SELF, mumps->id.lsol_loc, isol2_loc, PETSC_OWN_POINTER, &is_petsc)); /* to */
1211: PetscCall(VecScatterCreate(mumps->x_seq, is_iden, x, is_petsc, &mumps->scat_sol));
1212: PetscCall(ISDestroy(&is_iden));
1213: PetscCall(ISDestroy(&is_petsc));
1214: mumps->ICNTL9_pre = mumps->id.ICNTL(9); /* save current value of id.ICNTL(9) */
1215: }
1217: PetscCall(VecScatterBegin(mumps->scat_sol, mumps->x_seq, x, INSERT_VALUES, SCATTER_FORWARD));
1218: PetscCall(VecScatterEnd(mumps->scat_sol, mumps->x_seq, x, INSERT_VALUES, SCATTER_FORWARD));
1219: }
1221: if (mumps->petsc_size > 1) {
1222: if (mumps->ICNTL20 == 10) {
1223: PetscCall(VecRestoreArrayRead(b, &rarray));
1224: } else if (!mumps->myid) {
1225: PetscCall(VecRestoreArray(mumps->b_seq, &array));
1226: }
1227: } else PetscCall(VecRestoreArray(x, &array));
1229: PetscCall(PetscLogFlops(2.0 * mumps->id.RINFO(3)));
1230: PetscFunctionReturn(PETSC_SUCCESS);
1231: }
1233: PetscErrorCode MatSolveTranspose_MUMPS(Mat A, Vec b, Vec x)
1234: {
1235: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
1237: PetscFunctionBegin;
1238: mumps->id.ICNTL(9) = 0;
1239: PetscCall(MatSolve_MUMPS(A, b, x));
1240: mumps->id.ICNTL(9) = 1;
1241: PetscFunctionReturn(PETSC_SUCCESS);
1242: }
1244: PetscErrorCode MatMatSolve_MUMPS(Mat A, Mat B, Mat X)
1245: {
1246: Mat Bt = NULL;
1247: PetscBool denseX, denseB, flg, flgT;
1248: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
1249: PetscInt i, nrhs, M;
1250: PetscScalar *array;
1251: const PetscScalar *rbray;
1252: PetscInt lsol_loc, nlsol_loc, *idxx, iidx = 0;
1253: PetscMUMPSInt *isol_loc, *isol_loc_save;
1254: PetscScalar *bray, *sol_loc, *sol_loc_save;
1255: IS is_to, is_from;
1256: PetscInt k, proc, j, m, myrstart;
1257: const PetscInt *rstart;
1258: Vec v_mpi, msol_loc;
1259: VecScatter scat_sol;
1260: Vec b_seq;
1261: VecScatter scat_rhs;
1262: PetscScalar *aa;
1263: PetscInt spnr, *ia, *ja;
1264: Mat_MPIAIJ *b = NULL;
1266: PetscFunctionBegin;
1267: PetscCall(PetscObjectTypeCompareAny((PetscObject)X, &denseX, MATSEQDENSE, MATMPIDENSE, NULL));
1268: PetscCheck(denseX, PetscObjectComm((PetscObject)X), PETSC_ERR_ARG_WRONG, "Matrix X must be MATDENSE matrix");
1270: PetscCall(PetscObjectTypeCompareAny((PetscObject)B, &denseB, MATSEQDENSE, MATMPIDENSE, NULL));
1271: if (denseB) {
1272: PetscCheck(B->rmap->n == X->rmap->n, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Matrix B and X must have same row distribution");
1273: mumps->id.ICNTL(20) = 0; /* dense RHS */
1274: } else { /* sparse B */
1275: PetscCheck(X != B, PetscObjectComm((PetscObject)A), PETSC_ERR_ARG_IDN, "X and B must be different matrices");
1276: PetscCall(PetscObjectTypeCompare((PetscObject)B, MATTRANSPOSEVIRTUAL, &flgT));
1277: if (flgT) { /* input B is transpose of actual RHS matrix,
1278: because mumps requires sparse compressed COLUMN storage! See MatMatTransposeSolve_MUMPS() */
1279: PetscCall(MatTransposeGetMat(B, &Bt));
1280: } else SETERRQ(PetscObjectComm((PetscObject)B), PETSC_ERR_ARG_WRONG, "Matrix B must be MATTRANSPOSEVIRTUAL matrix");
1281: mumps->id.ICNTL(20) = 1; /* sparse RHS */
1282: }
1284: PetscCall(MatGetSize(B, &M, &nrhs));
1285: mumps->id.nrhs = nrhs;
1286: mumps->id.lrhs = M;
1287: mumps->id.rhs = NULL;
1289: if (mumps->petsc_size == 1) {
1290: PetscScalar *aa;
1291: PetscInt spnr, *ia, *ja;
1292: PetscBool second_solve = PETSC_FALSE;
1294: PetscCall(MatDenseGetArray(X, &array));
1295: mumps->id.rhs = (MumpsScalar *)array;
1297: if (denseB) {
1298: /* copy B to X */
1299: PetscCall(MatDenseGetArrayRead(B, &rbray));
1300: PetscCall(PetscArraycpy(array, rbray, M * nrhs));
1301: PetscCall(MatDenseRestoreArrayRead(B, &rbray));
1302: } else { /* sparse B */
1303: PetscCall(MatSeqAIJGetArray(Bt, &aa));
1304: PetscCall(MatGetRowIJ(Bt, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
1305: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot get IJ structure");
1306: PetscCall(PetscMUMPSIntCSRCast(mumps, spnr, ia, ja, &mumps->id.irhs_ptr, &mumps->id.irhs_sparse, &mumps->id.nz_rhs));
1307: mumps->id.rhs_sparse = (MumpsScalar *)aa;
1308: }
1309: /* handle condensation step of Schur complement (if any) */
1310: if (mumps->id.size_schur > 0 && (mumps->id.ICNTL(26) < 0 || mumps->id.ICNTL(26) > 2)) {
1311: second_solve = PETSC_TRUE;
1312: PetscCall(MatMumpsHandleSchur_Private(A, PETSC_FALSE));
1313: }
1314: /* solve phase */
1315: mumps->id.job = JOB_SOLVE;
1316: PetscMUMPS_c(mumps);
1317: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in solve phase: INFOG(1)=%d", mumps->id.INFOG(1));
1319: /* handle expansion step of Schur complement (if any) */
1320: if (second_solve) PetscCall(MatMumpsHandleSchur_Private(A, PETSC_TRUE));
1321: if (!denseB) { /* sparse B */
1322: PetscCall(MatSeqAIJRestoreArray(Bt, &aa));
1323: PetscCall(MatRestoreRowIJ(Bt, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
1324: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot restore IJ structure");
1325: }
1326: PetscCall(MatDenseRestoreArray(X, &array));
1327: PetscFunctionReturn(PETSC_SUCCESS);
1328: }
1330: /* parallel case: MUMPS requires rhs B to be centralized on the host! */
1331: PetscCheck(mumps->petsc_size <= 1 || !mumps->id.ICNTL(19), PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "Parallel Schur complements not yet supported from PETSc");
1333: /* create msol_loc to hold mumps local solution */
1334: isol_loc_save = mumps->id.isol_loc; /* save it for MatSolve() */
1335: sol_loc_save = (PetscScalar *)mumps->id.sol_loc;
1337: lsol_loc = mumps->id.lsol_loc;
1338: nlsol_loc = nrhs * lsol_loc; /* length of sol_loc */
1339: PetscCall(PetscMalloc2(nlsol_loc, &sol_loc, lsol_loc, &isol_loc));
1340: mumps->id.sol_loc = (MumpsScalar *)sol_loc;
1341: mumps->id.isol_loc = isol_loc;
1343: PetscCall(VecCreateSeqWithArray(PETSC_COMM_SELF, 1, nlsol_loc, (PetscScalar *)sol_loc, &msol_loc));
1345: if (denseB) {
1346: if (mumps->ICNTL20 == 10) {
1347: mumps->id.ICNTL(20) = 10; /* dense distributed RHS */
1348: PetscCall(MatDenseGetArrayRead(B, &rbray));
1349: PetscCall(MatMumpsSetUpDistRHSInfo(A, nrhs, rbray));
1350: PetscCall(MatDenseRestoreArrayRead(B, &rbray));
1351: PetscCall(MatGetLocalSize(B, &m, NULL));
1352: PetscCall(VecCreateMPIWithArray(PetscObjectComm((PetscObject)B), 1, nrhs * m, nrhs * M, NULL, &v_mpi));
1353: } else {
1354: mumps->id.ICNTL(20) = 0; /* dense centralized RHS */
1355: /* TODO: Because of non-contiguous indices, the created vecscatter scat_rhs is not done in MPI_Gather, resulting in
1356: very inefficient communication. An optimization is to use VecScatterCreateToZero to gather B to rank 0. Then on rank
1357: 0, re-arrange B into desired order, which is a local operation.
1358: */
1360: /* scatter v_mpi to b_seq because MUMPS before 5.3.0 only supports centralized rhs */
1361: /* wrap dense rhs matrix B into a vector v_mpi */
1362: PetscCall(MatGetLocalSize(B, &m, NULL));
1363: PetscCall(MatDenseGetArray(B, &bray));
1364: PetscCall(VecCreateMPIWithArray(PetscObjectComm((PetscObject)B), 1, nrhs * m, nrhs * M, (const PetscScalar *)bray, &v_mpi));
1365: PetscCall(MatDenseRestoreArray(B, &bray));
1367: /* scatter v_mpi to b_seq in proc[0]. MUMPS requires rhs to be centralized on the host! */
1368: if (!mumps->myid) {
1369: PetscInt *idx;
1370: /* idx: maps from k-th index of v_mpi to (i,j)-th global entry of B */
1371: PetscCall(PetscMalloc1(nrhs * M, &idx));
1372: PetscCall(MatGetOwnershipRanges(B, &rstart));
1373: k = 0;
1374: for (proc = 0; proc < mumps->petsc_size; proc++) {
1375: for (j = 0; j < nrhs; j++) {
1376: for (i = rstart[proc]; i < rstart[proc + 1]; i++) idx[k++] = j * M + i;
1377: }
1378: }
1380: PetscCall(VecCreateSeq(PETSC_COMM_SELF, nrhs * M, &b_seq));
1381: PetscCall(ISCreateGeneral(PETSC_COMM_SELF, nrhs * M, idx, PETSC_OWN_POINTER, &is_to));
1382: PetscCall(ISCreateStride(PETSC_COMM_SELF, nrhs * M, 0, 1, &is_from));
1383: } else {
1384: PetscCall(VecCreateSeq(PETSC_COMM_SELF, 0, &b_seq));
1385: PetscCall(ISCreateStride(PETSC_COMM_SELF, 0, 0, 1, &is_to));
1386: PetscCall(ISCreateStride(PETSC_COMM_SELF, 0, 0, 1, &is_from));
1387: }
1388: PetscCall(VecScatterCreate(v_mpi, is_from, b_seq, is_to, &scat_rhs));
1389: PetscCall(VecScatterBegin(scat_rhs, v_mpi, b_seq, INSERT_VALUES, SCATTER_FORWARD));
1390: PetscCall(ISDestroy(&is_to));
1391: PetscCall(ISDestroy(&is_from));
1392: PetscCall(VecScatterEnd(scat_rhs, v_mpi, b_seq, INSERT_VALUES, SCATTER_FORWARD));
1394: if (!mumps->myid) { /* define rhs on the host */
1395: PetscCall(VecGetArray(b_seq, &bray));
1396: mumps->id.rhs = (MumpsScalar *)bray;
1397: PetscCall(VecRestoreArray(b_seq, &bray));
1398: }
1399: }
1400: } else { /* sparse B */
1401: b = (Mat_MPIAIJ *)Bt->data;
1403: /* wrap dense X into a vector v_mpi */
1404: PetscCall(MatGetLocalSize(X, &m, NULL));
1405: PetscCall(MatDenseGetArray(X, &bray));
1406: PetscCall(VecCreateMPIWithArray(PetscObjectComm((PetscObject)X), 1, nrhs * m, nrhs * M, (const PetscScalar *)bray, &v_mpi));
1407: PetscCall(MatDenseRestoreArray(X, &bray));
1409: if (!mumps->myid) {
1410: PetscCall(MatSeqAIJGetArray(b->A, &aa));
1411: PetscCall(MatGetRowIJ(b->A, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
1412: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot get IJ structure");
1413: PetscCall(PetscMUMPSIntCSRCast(mumps, spnr, ia, ja, &mumps->id.irhs_ptr, &mumps->id.irhs_sparse, &mumps->id.nz_rhs));
1414: mumps->id.rhs_sparse = (MumpsScalar *)aa;
1415: } else {
1416: mumps->id.irhs_ptr = NULL;
1417: mumps->id.irhs_sparse = NULL;
1418: mumps->id.nz_rhs = 0;
1419: mumps->id.rhs_sparse = NULL;
1420: }
1421: }
1423: /* solve phase */
1424: mumps->id.job = JOB_SOLVE;
1425: PetscMUMPS_c(mumps);
1426: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in solve phase: INFOG(1)=%d", mumps->id.INFOG(1));
1428: /* scatter mumps distributed solution to petsc vector v_mpi, which shares local arrays with solution matrix X */
1429: PetscCall(MatDenseGetArray(X, &array));
1430: PetscCall(VecPlaceArray(v_mpi, array));
1432: /* create scatter scat_sol */
1433: PetscCall(MatGetOwnershipRanges(X, &rstart));
1434: /* iidx: index for scatter mumps solution to petsc X */
1436: PetscCall(ISCreateStride(PETSC_COMM_SELF, nlsol_loc, 0, 1, &is_from));
1437: PetscCall(PetscMalloc1(nlsol_loc, &idxx));
1438: for (i = 0; i < lsol_loc; i++) {
1439: isol_loc[i] -= 1; /* change Fortran style to C style. isol_loc[i+j*lsol_loc] contains x[isol_loc[i]] in j-th vector */
1441: for (proc = 0; proc < mumps->petsc_size; proc++) {
1442: if (isol_loc[i] >= rstart[proc] && isol_loc[i] < rstart[proc + 1]) {
1443: myrstart = rstart[proc];
1444: k = isol_loc[i] - myrstart; /* local index on 1st column of petsc vector X */
1445: iidx = k + myrstart * nrhs; /* maps mumps isol_loc[i] to petsc index in X */
1446: m = rstart[proc + 1] - rstart[proc]; /* rows of X for this proc */
1447: break;
1448: }
1449: }
1451: for (j = 0; j < nrhs; j++) idxx[i + j * lsol_loc] = iidx + j * m;
1452: }
1453: PetscCall(ISCreateGeneral(PETSC_COMM_SELF, nlsol_loc, idxx, PETSC_COPY_VALUES, &is_to));
1454: PetscCall(VecScatterCreate(msol_loc, is_from, v_mpi, is_to, &scat_sol));
1455: PetscCall(VecScatterBegin(scat_sol, msol_loc, v_mpi, INSERT_VALUES, SCATTER_FORWARD));
1456: PetscCall(ISDestroy(&is_from));
1457: PetscCall(ISDestroy(&is_to));
1458: PetscCall(VecScatterEnd(scat_sol, msol_loc, v_mpi, INSERT_VALUES, SCATTER_FORWARD));
1459: PetscCall(MatDenseRestoreArray(X, &array));
1461: /* free spaces */
1462: mumps->id.sol_loc = (MumpsScalar *)sol_loc_save;
1463: mumps->id.isol_loc = isol_loc_save;
1465: PetscCall(PetscFree2(sol_loc, isol_loc));
1466: PetscCall(PetscFree(idxx));
1467: PetscCall(VecDestroy(&msol_loc));
1468: PetscCall(VecDestroy(&v_mpi));
1469: if (!denseB) {
1470: if (!mumps->myid) {
1471: b = (Mat_MPIAIJ *)Bt->data;
1472: PetscCall(MatSeqAIJRestoreArray(b->A, &aa));
1473: PetscCall(MatRestoreRowIJ(b->A, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
1474: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot restore IJ structure");
1475: }
1476: } else {
1477: if (mumps->ICNTL20 == 0) {
1478: PetscCall(VecDestroy(&b_seq));
1479: PetscCall(VecScatterDestroy(&scat_rhs));
1480: }
1481: }
1482: PetscCall(VecScatterDestroy(&scat_sol));
1483: PetscCall(PetscLogFlops(2.0 * nrhs * mumps->id.RINFO(3)));
1484: PetscFunctionReturn(PETSC_SUCCESS);
1485: }
1487: PetscErrorCode MatMatSolveTranspose_MUMPS(Mat A, Mat B, Mat X)
1488: {
1489: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
1490: PetscMUMPSInt oldvalue = mumps->id.ICNTL(9);
1492: PetscFunctionBegin;
1493: mumps->id.ICNTL(9) = 0;
1494: PetscCall(MatMatSolve_MUMPS(A, B, X));
1495: mumps->id.ICNTL(9) = oldvalue;
1496: PetscFunctionReturn(PETSC_SUCCESS);
1497: }
1499: PetscErrorCode MatMatTransposeSolve_MUMPS(Mat A, Mat Bt, Mat X)
1500: {
1501: PetscBool flg;
1502: Mat B;
1504: PetscFunctionBegin;
1505: PetscCall(PetscObjectTypeCompareAny((PetscObject)Bt, &flg, MATSEQAIJ, MATMPIAIJ, NULL));
1506: PetscCheck(flg, PetscObjectComm((PetscObject)Bt), PETSC_ERR_ARG_WRONG, "Matrix Bt must be MATAIJ matrix");
1508: /* Create B=Bt^T that uses Bt's data structure */
1509: PetscCall(MatCreateTranspose(Bt, &B));
1511: PetscCall(MatMatSolve_MUMPS(A, B, X));
1512: PetscCall(MatDestroy(&B));
1513: PetscFunctionReturn(PETSC_SUCCESS);
1514: }
1516: #if !defined(PETSC_USE_COMPLEX)
1517: /*
1518: input:
1519: F: numeric factor
1520: output:
1521: nneg: total number of negative pivots
1522: nzero: total number of zero pivots
1523: npos: (global dimension of F) - nneg - nzero
1524: */
1525: PetscErrorCode MatGetInertia_SBAIJMUMPS(Mat F, PetscInt *nneg, PetscInt *nzero, PetscInt *npos)
1526: {
1527: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
1528: PetscMPIInt size;
1530: PetscFunctionBegin;
1531: PetscCallMPI(MPI_Comm_size(PetscObjectComm((PetscObject)F), &size));
1532: /* 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 */
1533: PetscCheck(size <= 1 || mumps->id.ICNTL(13) == 1, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "ICNTL(13)=%d. -mat_mumps_icntl_13 must be set as 1 for correct global matrix inertia", mumps->id.INFOG(13));
1535: if (nneg) *nneg = mumps->id.INFOG(12);
1536: if (nzero || npos) {
1537: PetscCheck(mumps->id.ICNTL(24) == 1, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "-mat_mumps_icntl_24 must be set as 1 for null pivot row detection");
1538: if (nzero) *nzero = mumps->id.INFOG(28);
1539: if (npos) *npos = F->rmap->N - (mumps->id.INFOG(12) + mumps->id.INFOG(28));
1540: }
1541: PetscFunctionReturn(PETSC_SUCCESS);
1542: }
1543: #endif
1545: PetscErrorCode MatMumpsGatherNonzerosOnMaster(MatReuse reuse, Mat_MUMPS *mumps)
1546: {
1547: PetscInt i, nreqs;
1548: PetscMUMPSInt *irn, *jcn;
1549: PetscMPIInt count;
1550: PetscInt64 totnnz, remain;
1551: const PetscInt osize = mumps->omp_comm_size;
1552: PetscScalar *val;
1554: PetscFunctionBegin;
1555: if (osize > 1) {
1556: if (reuse == MAT_INITIAL_MATRIX) {
1557: /* master first gathers counts of nonzeros to receive */
1558: if (mumps->is_omp_master) PetscCall(PetscMalloc1(osize, &mumps->recvcount));
1559: PetscCallMPI(MPI_Gather(&mumps->nnz, 1, MPIU_INT64, mumps->recvcount, 1, MPIU_INT64, 0 /*master*/, mumps->omp_comm));
1561: /* Then each computes number of send/recvs */
1562: if (mumps->is_omp_master) {
1563: /* Start from 1 since self communication is not done in MPI */
1564: nreqs = 0;
1565: for (i = 1; i < osize; i++) nreqs += (mumps->recvcount[i] + PETSC_MPI_INT_MAX - 1) / PETSC_MPI_INT_MAX;
1566: } else {
1567: nreqs = (mumps->nnz + PETSC_MPI_INT_MAX - 1) / PETSC_MPI_INT_MAX;
1568: }
1569: PetscCall(PetscMalloc1(nreqs * 3, &mumps->reqs)); /* Triple the requests since we send irn, jcn and val separately */
1571: /* The following code is doing a very simple thing: omp_master rank gathers irn/jcn/val from others.
1572: MPI_Gatherv would be enough if it supports big counts > 2^31-1. Since it does not, and mumps->nnz
1573: might be a prime number > 2^31-1, we have to slice the message. Note omp_comm_size
1574: is very small, the current approach should have no extra overhead compared to MPI_Gatherv.
1575: */
1576: nreqs = 0; /* counter for actual send/recvs */
1577: if (mumps->is_omp_master) {
1578: for (i = 0, totnnz = 0; i < osize; i++) totnnz += mumps->recvcount[i]; /* totnnz = sum of nnz over omp_comm */
1579: PetscCall(PetscMalloc2(totnnz, &irn, totnnz, &jcn));
1580: PetscCall(PetscMalloc1(totnnz, &val));
1582: /* Self communication */
1583: PetscCall(PetscArraycpy(irn, mumps->irn, mumps->nnz));
1584: PetscCall(PetscArraycpy(jcn, mumps->jcn, mumps->nnz));
1585: PetscCall(PetscArraycpy(val, mumps->val, mumps->nnz));
1587: /* Replace mumps->irn/jcn etc on master with the newly allocated bigger arrays */
1588: PetscCall(PetscFree2(mumps->irn, mumps->jcn));
1589: PetscCall(PetscFree(mumps->val_alloc));
1590: mumps->nnz = totnnz;
1591: mumps->irn = irn;
1592: mumps->jcn = jcn;
1593: mumps->val = mumps->val_alloc = val;
1595: irn += mumps->recvcount[0]; /* recvcount[0] is old mumps->nnz on omp rank 0 */
1596: jcn += mumps->recvcount[0];
1597: val += mumps->recvcount[0];
1599: /* Remote communication */
1600: for (i = 1; i < osize; i++) {
1601: count = PetscMin(mumps->recvcount[i], PETSC_MPI_INT_MAX);
1602: remain = mumps->recvcount[i] - count;
1603: while (count > 0) {
1604: PetscCallMPI(MPI_Irecv(irn, count, MPIU_MUMPSINT, i, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1605: PetscCallMPI(MPI_Irecv(jcn, count, MPIU_MUMPSINT, i, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1606: PetscCallMPI(MPI_Irecv(val, count, MPIU_SCALAR, i, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1607: irn += count;
1608: jcn += count;
1609: val += count;
1610: count = PetscMin(remain, PETSC_MPI_INT_MAX);
1611: remain -= count;
1612: }
1613: }
1614: } else {
1615: irn = mumps->irn;
1616: jcn = mumps->jcn;
1617: val = mumps->val;
1618: count = PetscMin(mumps->nnz, PETSC_MPI_INT_MAX);
1619: remain = mumps->nnz - count;
1620: while (count > 0) {
1621: PetscCallMPI(MPI_Isend(irn, count, MPIU_MUMPSINT, 0, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1622: PetscCallMPI(MPI_Isend(jcn, count, MPIU_MUMPSINT, 0, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1623: PetscCallMPI(MPI_Isend(val, count, MPIU_SCALAR, 0, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1624: irn += count;
1625: jcn += count;
1626: val += count;
1627: count = PetscMin(remain, PETSC_MPI_INT_MAX);
1628: remain -= count;
1629: }
1630: }
1631: } else {
1632: nreqs = 0;
1633: if (mumps->is_omp_master) {
1634: val = mumps->val + mumps->recvcount[0];
1635: for (i = 1; i < osize; i++) { /* Remote communication only since self data is already in place */
1636: count = PetscMin(mumps->recvcount[i], PETSC_MPI_INT_MAX);
1637: remain = mumps->recvcount[i] - count;
1638: while (count > 0) {
1639: PetscCallMPI(MPI_Irecv(val, count, MPIU_SCALAR, i, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1640: val += count;
1641: count = PetscMin(remain, PETSC_MPI_INT_MAX);
1642: remain -= count;
1643: }
1644: }
1645: } else {
1646: val = mumps->val;
1647: count = PetscMin(mumps->nnz, PETSC_MPI_INT_MAX);
1648: remain = mumps->nnz - count;
1649: while (count > 0) {
1650: PetscCallMPI(MPI_Isend(val, count, MPIU_SCALAR, 0, mumps->tag, mumps->omp_comm, &mumps->reqs[nreqs++]));
1651: val += count;
1652: count = PetscMin(remain, PETSC_MPI_INT_MAX);
1653: remain -= count;
1654: }
1655: }
1656: }
1657: PetscCallMPI(MPI_Waitall(nreqs, mumps->reqs, MPI_STATUSES_IGNORE));
1658: mumps->tag++; /* It is totally fine for above send/recvs to share one mpi tag */
1659: }
1660: PetscFunctionReturn(PETSC_SUCCESS);
1661: }
1663: PetscErrorCode MatFactorNumeric_MUMPS(Mat F, Mat A, const MatFactorInfo *info)
1664: {
1665: Mat_MUMPS *mumps = (Mat_MUMPS *)(F)->data;
1666: PetscBool isMPIAIJ;
1668: PetscFunctionBegin;
1669: if (mumps->id.INFOG(1) < 0 && !(mumps->id.INFOG(1) == -16 && mumps->id.INFOG(1) == 0)) {
1670: if (mumps->id.INFOG(1) == -6) PetscCall(PetscInfo(A, "MatFactorNumeric is called with singular matrix structure, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1671: PetscCall(PetscInfo(A, "MatFactorNumeric is called after analysis phase fails, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1672: PetscFunctionReturn(PETSC_SUCCESS);
1673: }
1675: PetscCall((*mumps->ConvertToTriples)(A, 1, MAT_REUSE_MATRIX, mumps));
1676: PetscCall(MatMumpsGatherNonzerosOnMaster(MAT_REUSE_MATRIX, mumps));
1678: /* numerical factorization phase */
1679: mumps->id.job = JOB_FACTNUMERIC;
1680: if (!mumps->id.ICNTL(18)) { /* A is centralized */
1681: if (!mumps->myid) mumps->id.a = (MumpsScalar *)mumps->val;
1682: } else {
1683: mumps->id.a_loc = (MumpsScalar *)mumps->val;
1684: }
1685: PetscMUMPS_c(mumps);
1686: if (mumps->id.INFOG(1) < 0) {
1687: PetscCheck(!A->erroriffailure, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d", mumps->id.INFOG(1), mumps->id.INFO(2));
1688: if (mumps->id.INFOG(1) == -10) { /* numerically singular matrix */
1689: PetscCall(PetscInfo(F, "matrix is numerically singular, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1690: F->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
1691: } else if (mumps->id.INFOG(1) == -13) {
1692: PetscCall(PetscInfo(F, "MUMPS in numerical factorization phase: INFOG(1)=%d, cannot allocate required memory %d megabytes\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1693: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1694: } else if (mumps->id.INFOG(1) == -8 || mumps->id.INFOG(1) == -9 || (-16 < mumps->id.INFOG(1) && mumps->id.INFOG(1) < -10)) {
1695: PetscCall(PetscInfo(F, "MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d, problem with workarray\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1696: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1697: } else {
1698: PetscCall(PetscInfo(F, "MUMPS in numerical factorization phase: INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1699: F->factorerrortype = MAT_FACTOR_OTHER;
1700: }
1701: }
1702: PetscCheck(mumps->myid || mumps->id.ICNTL(16) <= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, " mumps->id.ICNTL(16):=%d", mumps->id.INFOG(16));
1704: F->assembled = PETSC_TRUE;
1706: if (F->schur) { /* reset Schur status to unfactored */
1707: #if defined(PETSC_HAVE_CUDA)
1708: F->schur->offloadmask = PETSC_OFFLOAD_CPU;
1709: #endif
1710: if (mumps->id.ICNTL(19) == 1) { /* stored by rows */
1711: mumps->id.ICNTL(19) = 2;
1712: PetscCall(MatTranspose(F->schur, MAT_INPLACE_MATRIX, &F->schur));
1713: }
1714: PetscCall(MatFactorRestoreSchurComplement(F, NULL, MAT_FACTOR_SCHUR_UNFACTORED));
1715: }
1717: /* just to be sure that ICNTL(19) value returned by a call from MatMumpsGetIcntl is always consistent */
1718: if (!mumps->sym && mumps->id.ICNTL(19) && mumps->id.ICNTL(19) != 1) mumps->id.ICNTL(19) = 3;
1720: if (!mumps->is_omp_master) mumps->id.INFO(23) = 0;
1721: if (mumps->petsc_size > 1) {
1722: PetscInt lsol_loc;
1723: PetscScalar *sol_loc;
1725: PetscCall(PetscObjectTypeCompare((PetscObject)A, MATMPIAIJ, &isMPIAIJ));
1727: /* distributed solution; Create x_seq=sol_loc for repeated use */
1728: if (mumps->x_seq) {
1729: PetscCall(VecScatterDestroy(&mumps->scat_sol));
1730: PetscCall(PetscFree2(mumps->id.sol_loc, mumps->id.isol_loc));
1731: PetscCall(VecDestroy(&mumps->x_seq));
1732: }
1733: lsol_loc = mumps->id.INFO(23); /* length of sol_loc */
1734: PetscCall(PetscMalloc2(lsol_loc, &sol_loc, lsol_loc, &mumps->id.isol_loc));
1735: mumps->id.lsol_loc = lsol_loc;
1736: mumps->id.sol_loc = (MumpsScalar *)sol_loc;
1737: PetscCall(VecCreateSeqWithArray(PETSC_COMM_SELF, 1, lsol_loc, sol_loc, &mumps->x_seq));
1738: }
1739: PetscCall(PetscLogFlops(mumps->id.RINFO(2)));
1740: PetscFunctionReturn(PETSC_SUCCESS);
1741: }
1743: /* Sets MUMPS options from the options database */
1744: PetscErrorCode MatSetFromOptions_MUMPS(Mat F, Mat A)
1745: {
1746: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
1747: PetscMUMPSInt icntl = 0, size, *listvar_schur;
1748: PetscInt info[80], i, ninfo = 80, rbs, cbs;
1749: PetscBool flg = PETSC_FALSE, schur = (PetscBool)(mumps->id.ICNTL(26) == -1);
1750: MumpsScalar *arr;
1752: PetscFunctionBegin;
1753: PetscOptionsBegin(PetscObjectComm((PetscObject)F), ((PetscObject)F)->prefix, "MUMPS Options", "Mat");
1754: if (mumps->id.job == JOB_NULL) { /* MatSetFromOptions_MUMPS() has never been called before */
1755: PetscInt nthreads = 0;
1756: PetscInt nCNTL_pre = mumps->CNTL_pre ? mumps->CNTL_pre[0] : 0;
1757: PetscInt nICNTL_pre = mumps->ICNTL_pre ? mumps->ICNTL_pre[0] : 0;
1759: mumps->petsc_comm = PetscObjectComm((PetscObject)A);
1760: PetscCallMPI(MPI_Comm_size(mumps->petsc_comm, &mumps->petsc_size));
1761: PetscCallMPI(MPI_Comm_rank(mumps->petsc_comm, &mumps->myid)); /* "if (!myid)" still works even if mumps_comm is different */
1763: PetscCall(PetscOptionsName("-mat_mumps_use_omp_threads", "Convert MPI processes into OpenMP threads", "None", &mumps->use_petsc_omp_support));
1764: if (mumps->use_petsc_omp_support) nthreads = -1; /* -1 will let PetscOmpCtrlCreate() guess a proper value when user did not supply one */
1765: /* do not use PetscOptionsInt() so that the option -mat_mumps_use_omp_threads is not displayed twice in the help */
1766: PetscCall(PetscOptionsGetInt(NULL, ((PetscObject)F)->prefix, "-mat_mumps_use_omp_threads", &nthreads, NULL));
1767: if (mumps->use_petsc_omp_support) {
1768: PetscCheck(PetscDefined(HAVE_OPENMP_SUPPORT), PETSC_COMM_SELF, PETSC_ERR_SUP_SYS, "The system does not have PETSc OpenMP support but you added the -%smat_mumps_use_omp_threads option. Configure PETSc with --with-openmp --download-hwloc (or --with-hwloc) to enable it, see more in MATSOLVERMUMPS manual",
1769: ((PetscObject)F)->prefix ? ((PetscObject)F)->prefix : "");
1770: PetscCheck(!schur, PETSC_COMM_SELF, PETSC_ERR_SUP, "Cannot use -%smat_mumps_use_omp_threads with the Schur complement feature", ((PetscObject)F)->prefix ? ((PetscObject)F)->prefix : "");
1771: #if defined(PETSC_HAVE_OPENMP_SUPPORT)
1772: PetscCall(PetscOmpCtrlCreate(mumps->petsc_comm, nthreads, &mumps->omp_ctrl));
1773: PetscCall(PetscOmpCtrlGetOmpComms(mumps->omp_ctrl, &mumps->omp_comm, &mumps->mumps_comm, &mumps->is_omp_master));
1774: #endif
1775: } else {
1776: mumps->omp_comm = PETSC_COMM_SELF;
1777: mumps->mumps_comm = mumps->petsc_comm;
1778: mumps->is_omp_master = PETSC_TRUE;
1779: }
1780: PetscCallMPI(MPI_Comm_size(mumps->omp_comm, &mumps->omp_comm_size));
1781: mumps->reqs = NULL;
1782: mumps->tag = 0;
1784: if (mumps->mumps_comm != MPI_COMM_NULL) {
1785: if (PetscDefined(HAVE_OPENMP_SUPPORT) && mumps->use_petsc_omp_support) {
1786: /* It looks like MUMPS does not dup the input comm. Dup a new comm for MUMPS to avoid any tag mismatches. */
1787: MPI_Comm comm;
1788: PetscCallMPI(MPI_Comm_dup(mumps->mumps_comm, &comm));
1789: mumps->mumps_comm = comm;
1790: } else PetscCall(PetscCommGetComm(mumps->petsc_comm, &mumps->mumps_comm));
1791: }
1793: mumps->id.comm_fortran = MPI_Comm_c2f(mumps->mumps_comm);
1794: mumps->id.job = JOB_INIT;
1795: mumps->id.par = 1; /* host participates factorizaton and solve */
1796: mumps->id.sym = mumps->sym;
1798: size = mumps->id.size_schur;
1799: arr = mumps->id.schur;
1800: listvar_schur = mumps->id.listvar_schur;
1801: PetscMUMPS_c(mumps);
1802: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS: INFOG(1)=%d", mumps->id.INFOG(1));
1803: /* restore cached ICNTL and CNTL values */
1804: for (icntl = 0; icntl < nICNTL_pre; ++icntl) mumps->id.ICNTL(mumps->ICNTL_pre[1 + 2 * icntl]) = mumps->ICNTL_pre[2 + 2 * icntl];
1805: for (icntl = 0; icntl < nCNTL_pre; ++icntl) mumps->id.CNTL((PetscInt)mumps->CNTL_pre[1 + 2 * icntl]) = mumps->CNTL_pre[2 + 2 * icntl];
1806: PetscCall(PetscFree(mumps->ICNTL_pre));
1807: PetscCall(PetscFree(mumps->CNTL_pre));
1809: if (schur) {
1810: mumps->id.size_schur = size;
1811: mumps->id.schur_lld = size;
1812: mumps->id.schur = arr;
1813: mumps->id.listvar_schur = listvar_schur;
1814: if (mumps->petsc_size > 1) {
1815: PetscBool gs; /* gs is false if any rank other than root has non-empty IS */
1817: mumps->id.ICNTL(19) = 1; /* MUMPS returns Schur centralized on the host */
1818: gs = mumps->myid ? (mumps->id.size_schur ? PETSC_FALSE : PETSC_TRUE) : PETSC_TRUE; /* always true on root; false on others if their size != 0 */
1819: PetscCall(MPIU_Allreduce(MPI_IN_PLACE, &gs, 1, MPIU_BOOL, MPI_LAND, mumps->petsc_comm));
1820: PetscCheck(gs, PETSC_COMM_SELF, PETSC_ERR_SUP, "MUMPS distributed parallel Schur complements not yet supported from PETSc");
1821: } else {
1822: if (F->factortype == MAT_FACTOR_LU) {
1823: mumps->id.ICNTL(19) = 3; /* MUMPS returns full matrix */
1824: } else {
1825: mumps->id.ICNTL(19) = 2; /* MUMPS returns lower triangular part */
1826: }
1827: }
1828: mumps->id.ICNTL(26) = -1;
1829: }
1831: /* copy MUMPS default control values from master to slaves. Although slaves do not call MUMPS, they may access these values in code.
1832: For example, ICNTL(9) is initialized to 1 by MUMPS and slaves check ICNTL(9) in MatSolve_MUMPS.
1833: */
1834: PetscCallMPI(MPI_Bcast(mumps->id.icntl, 40, MPI_INT, 0, mumps->omp_comm));
1835: PetscCallMPI(MPI_Bcast(mumps->id.cntl, 15, MPIU_REAL, 0, mumps->omp_comm));
1837: mumps->scat_rhs = NULL;
1838: mumps->scat_sol = NULL;
1840: /* set PETSc-MUMPS default options - override MUMPS default */
1841: mumps->id.ICNTL(3) = 0;
1842: mumps->id.ICNTL(4) = 0;
1843: if (mumps->petsc_size == 1) {
1844: mumps->id.ICNTL(18) = 0; /* centralized assembled matrix input */
1845: mumps->id.ICNTL(7) = 7; /* automatic choice of ordering done by the package */
1846: } else {
1847: mumps->id.ICNTL(18) = 3; /* distributed assembled matrix input */
1848: mumps->id.ICNTL(21) = 1; /* distributed solution */
1849: }
1850: }
1851: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_1", "ICNTL(1): output stream for error messages", "None", mumps->id.ICNTL(1), &icntl, &flg));
1852: if (flg) mumps->id.ICNTL(1) = icntl;
1853: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_2", "ICNTL(2): output stream for diagnostic printing, statistics, and warning", "None", mumps->id.ICNTL(2), &icntl, &flg));
1854: if (flg) mumps->id.ICNTL(2) = icntl;
1855: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_3", "ICNTL(3): output stream for global information, collected on the host", "None", mumps->id.ICNTL(3), &icntl, &flg));
1856: if (flg) mumps->id.ICNTL(3) = icntl;
1858: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_4", "ICNTL(4): level of printing (0 to 4)", "None", mumps->id.ICNTL(4), &icntl, &flg));
1859: if (flg) mumps->id.ICNTL(4) = icntl;
1860: if (mumps->id.ICNTL(4) || PetscLogPrintInfo) mumps->id.ICNTL(3) = 6; /* resume MUMPS default id.ICNTL(3) = 6 */
1862: PetscCall(PetscOptionsMUMPSInt("-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));
1863: if (flg) mumps->id.ICNTL(6) = icntl;
1865: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_7", "ICNTL(7): computes a symmetric permutation in sequential analysis. 0=AMD, 2=AMF, 3=Scotch, 4=PORD, 5=Metis, 6=QAMD, and 7=auto(default)", "None", mumps->id.ICNTL(7), &icntl, &flg));
1866: if (flg) {
1867: PetscCheck(icntl != 1 && icntl >= 0 && icntl <= 7, PETSC_COMM_SELF, PETSC_ERR_ARG_OUTOFRANGE, "Valid values are 0=AMD, 2=AMF, 3=Scotch, 4=PORD, 5=Metis, 6=QAMD, and 7=auto");
1868: mumps->id.ICNTL(7) = icntl;
1869: }
1871: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_8", "ICNTL(8): scaling strategy (-2 to 8 or 77)", "None", mumps->id.ICNTL(8), &mumps->id.ICNTL(8), NULL));
1872: /* PetscCall(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() */
1873: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_10", "ICNTL(10): max num of refinements", "None", mumps->id.ICNTL(10), &mumps->id.ICNTL(10), NULL));
1874: PetscCall(PetscOptionsMUMPSInt("-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));
1875: PetscCall(PetscOptionsMUMPSInt("-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));
1876: PetscCall(PetscOptionsMUMPSInt("-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));
1877: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_14", "ICNTL(14): percentage increase in the estimated working space", "None", mumps->id.ICNTL(14), &mumps->id.ICNTL(14), NULL));
1878: PetscCall(MatGetBlockSizes(A, &rbs, &cbs));
1879: if (rbs == cbs && rbs > 1) mumps->id.ICNTL(15) = -rbs;
1880: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_15", "ICNTL(15): compression of the input matrix resulting from a block format", "None", mumps->id.ICNTL(15), &mumps->id.ICNTL(15), &flg));
1881: if (flg) {
1882: PetscCheck(mumps->id.ICNTL(15) <= 0, PETSC_COMM_SELF, PETSC_ERR_SUP, "Positive -mat_mumps_icntl_15 not handled");
1883: PetscCheck((-mumps->id.ICNTL(15) % cbs == 0) && (-mumps->id.ICNTL(15) % rbs == 0), PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "The opposite of -mat_mumps_icntl_15 must be a multiple of the column and row blocksizes");
1884: }
1885: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_19", "ICNTL(19): computes the Schur complement", "None", mumps->id.ICNTL(19), &mumps->id.ICNTL(19), NULL));
1886: if (mumps->id.ICNTL(19) <= 0 || mumps->id.ICNTL(19) > 3) { /* reset any schur data (if any) */
1887: PetscCall(MatDestroy(&F->schur));
1888: PetscCall(MatMumpsResetSchur_Private(mumps));
1889: }
1891: /* Two MPICH Fortran MPI_IN_PLACE binding bugs prevented the use of 'mpich + mumps'. One happened with "mpi4py + mpich + mumps",
1892: and was reported by Firedrake. See https://bitbucket.org/mpi4py/mpi4py/issues/162/mpi4py-initialization-breaks-fortran
1893: and a petsc-maint mailing list thread with subject 'MUMPS segfaults in parallel because of ...'
1894: This bug was fixed by https://github.com/pmodels/mpich/pull/4149. But the fix brought a new bug,
1895: see https://github.com/pmodels/mpich/issues/5589. This bug was fixed by https://github.com/pmodels/mpich/pull/5590.
1896: In short, we could not use distributed RHS with MPICH until v4.0b1.
1897: */
1898: #if PETSC_PKG_MUMPS_VERSION_LT(5, 3, 0) || (defined(PETSC_HAVE_MPICH_NUMVERSION) && (PETSC_HAVE_MPICH_NUMVERSION < 40000101))
1899: mumps->ICNTL20 = 0; /* Centralized dense RHS*/
1900: #else
1901: mumps->ICNTL20 = 10; /* Distributed dense RHS*/
1902: #endif
1903: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_20", "ICNTL(20): give mumps centralized (0) or distributed (10) dense right-hand sides", "None", mumps->ICNTL20, &mumps->ICNTL20, &flg));
1904: PetscCheck(!flg || mumps->ICNTL20 == 10 || mumps->ICNTL20 == 0, PETSC_COMM_SELF, PETSC_ERR_SUP, "ICNTL(20)=%d is not supported by the PETSc/MUMPS interface. Allowed values are 0, 10", (int)mumps->ICNTL20);
1905: #if PETSC_PKG_MUMPS_VERSION_LT(5, 3, 0)
1906: PetscCheck(!flg || mumps->ICNTL20 != 10, PETSC_COMM_SELF, PETSC_ERR_SUP, "ICNTL(20)=10 is not supported before MUMPS-5.3.0");
1907: #endif
1908: /* PetscCall(PetscOptionsMUMPSInt("-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 */
1910: PetscCall(PetscOptionsMUMPSInt("-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));
1911: PetscCall(PetscOptionsMUMPSInt("-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));
1912: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_24", "ICNTL(24): detection of null pivot rows (0 or 1)", "None", mumps->id.ICNTL(24), &mumps->id.ICNTL(24), NULL));
1913: if (mumps->id.ICNTL(24)) { mumps->id.ICNTL(13) = 1; /* turn-off ScaLAPACK to help with the correct detection of null pivots */ }
1915: PetscCall(PetscOptionsMUMPSInt("-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));
1916: PetscCall(PetscOptionsMUMPSInt("-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));
1917: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_27", "ICNTL(27): controls the blocking size for multiple right-hand sides", "None", mumps->id.ICNTL(27), &mumps->id.ICNTL(27), NULL));
1918: PetscCall(PetscOptionsMUMPSInt("-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));
1919: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_29", "ICNTL(29): parallel ordering 1 = ptscotch, 2 = parmetis", "None", mumps->id.ICNTL(29), &mumps->id.ICNTL(29), NULL));
1920: /* PetscCall(PetscOptionsMUMPSInt("-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 */
1921: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_31", "ICNTL(31): indicates which factors may be discarded during factorization", "None", mumps->id.ICNTL(31), &mumps->id.ICNTL(31), NULL));
1922: /* PetscCall(PetscOptionsMUMPSInt("-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 */
1923: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_33", "ICNTL(33): compute determinant", "None", mumps->id.ICNTL(33), &mumps->id.ICNTL(33), NULL));
1924: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_35", "ICNTL(35): activates Block Low Rank (BLR) based factorization", "None", mumps->id.ICNTL(35), &mumps->id.ICNTL(35), NULL));
1925: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_36", "ICNTL(36): choice of BLR factorization variant", "None", mumps->id.ICNTL(36), &mumps->id.ICNTL(36), NULL));
1926: PetscCall(PetscOptionsMUMPSInt("-mat_mumps_icntl_38", "ICNTL(38): estimated compression rate of LU factors with BLR", "None", mumps->id.ICNTL(38), &mumps->id.ICNTL(38), NULL));
1928: PetscCall(PetscOptionsReal("-mat_mumps_cntl_1", "CNTL(1): relative pivoting threshold", "None", mumps->id.CNTL(1), &mumps->id.CNTL(1), NULL));
1929: PetscCall(PetscOptionsReal("-mat_mumps_cntl_2", "CNTL(2): stopping criterion of refinement", "None", mumps->id.CNTL(2), &mumps->id.CNTL(2), NULL));
1930: PetscCall(PetscOptionsReal("-mat_mumps_cntl_3", "CNTL(3): absolute pivoting threshold", "None", mumps->id.CNTL(3), &mumps->id.CNTL(3), NULL));
1931: PetscCall(PetscOptionsReal("-mat_mumps_cntl_4", "CNTL(4): value for static pivoting", "None", mumps->id.CNTL(4), &mumps->id.CNTL(4), NULL));
1932: PetscCall(PetscOptionsReal("-mat_mumps_cntl_5", "CNTL(5): fixation for null pivots", "None", mumps->id.CNTL(5), &mumps->id.CNTL(5), NULL));
1933: PetscCall(PetscOptionsReal("-mat_mumps_cntl_7", "CNTL(7): dropping parameter used during BLR", "None", mumps->id.CNTL(7), &mumps->id.CNTL(7), NULL));
1935: PetscCall(PetscOptionsString("-mat_mumps_ooc_tmpdir", "out of core directory", "None", mumps->id.ooc_tmpdir, mumps->id.ooc_tmpdir, sizeof(mumps->id.ooc_tmpdir), NULL));
1937: PetscCall(PetscOptionsIntArray("-mat_mumps_view_info", "request INFO local to each processor", "", info, &ninfo, NULL));
1938: if (ninfo) {
1939: PetscCheck(ninfo <= 80, PETSC_COMM_SELF, PETSC_ERR_USER, "number of INFO %" PetscInt_FMT " must <= 80", ninfo);
1940: PetscCall(PetscMalloc1(ninfo, &mumps->info));
1941: mumps->ninfo = ninfo;
1942: for (i = 0; i < ninfo; i++) {
1943: PetscCheck(info[i] >= 0 && info[i] <= 80, PETSC_COMM_SELF, PETSC_ERR_USER, "index of INFO %" PetscInt_FMT " must between 1 and 80", ninfo);
1944: mumps->info[i] = info[i];
1945: }
1946: }
1947: PetscOptionsEnd();
1948: PetscFunctionReturn(PETSC_SUCCESS);
1949: }
1951: PetscErrorCode MatFactorSymbolic_MUMPS_ReportIfError(Mat F, Mat A, const MatFactorInfo *info, Mat_MUMPS *mumps)
1952: {
1953: PetscFunctionBegin;
1954: if (mumps->id.INFOG(1) < 0) {
1955: PetscCheck(!A->erroriffailure, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in analysis phase: INFOG(1)=%d", mumps->id.INFOG(1));
1956: if (mumps->id.INFOG(1) == -6) {
1957: PetscCall(PetscInfo(F, "matrix is singular in structure, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1958: F->factorerrortype = MAT_FACTOR_STRUCT_ZEROPIVOT;
1959: } else if (mumps->id.INFOG(1) == -5 || mumps->id.INFOG(1) == -7) {
1960: PetscCall(PetscInfo(F, "problem of workspace, INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1961: F->factorerrortype = MAT_FACTOR_OUTMEMORY;
1962: } else if (mumps->id.INFOG(1) == -16 && mumps->id.INFOG(1) == 0) {
1963: PetscCall(PetscInfo(F, "Empty matrix\n"));
1964: } else {
1965: PetscCall(PetscInfo(F, "Error reported by MUMPS in analysis phase: INFOG(1)=%d, INFO(2)=%d\n", mumps->id.INFOG(1), mumps->id.INFO(2)));
1966: F->factorerrortype = MAT_FACTOR_OTHER;
1967: }
1968: }
1969: PetscFunctionReturn(PETSC_SUCCESS);
1970: }
1972: PetscErrorCode MatLUFactorSymbolic_AIJMUMPS(Mat F, Mat A, IS r, IS c, const MatFactorInfo *info)
1973: {
1974: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
1975: Vec b;
1976: const PetscInt M = A->rmap->N;
1978: PetscFunctionBegin;
1979: if (mumps->matstruc == SAME_NONZERO_PATTERN) {
1980: /* F is assembled by a previous call of MatLUFactorSymbolic_AIJMUMPS() */
1981: PetscFunctionReturn(PETSC_SUCCESS);
1982: }
1984: /* Set MUMPS options from the options database */
1985: PetscCall(MatSetFromOptions_MUMPS(F, A));
1987: PetscCall((*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, mumps));
1988: PetscCall(MatMumpsGatherNonzerosOnMaster(MAT_INITIAL_MATRIX, mumps));
1990: /* analysis phase */
1991: mumps->id.job = JOB_FACTSYMBOLIC;
1992: mumps->id.n = M;
1993: switch (mumps->id.ICNTL(18)) {
1994: case 0: /* centralized assembled matrix input */
1995: if (!mumps->myid) {
1996: mumps->id.nnz = mumps->nnz;
1997: mumps->id.irn = mumps->irn;
1998: mumps->id.jcn = mumps->jcn;
1999: if (mumps->id.ICNTL(6) > 1) mumps->id.a = (MumpsScalar *)mumps->val;
2000: if (r) {
2001: mumps->id.ICNTL(7) = 1;
2002: if (!mumps->myid) {
2003: const PetscInt *idx;
2004: PetscInt i;
2006: PetscCall(PetscMalloc1(M, &mumps->id.perm_in));
2007: PetscCall(ISGetIndices(r, &idx));
2008: for (i = 0; i < M; i++) PetscCall(PetscMUMPSIntCast(idx[i] + 1, &(mumps->id.perm_in[i]))); /* perm_in[]: start from 1, not 0! */
2009: PetscCall(ISRestoreIndices(r, &idx));
2010: }
2011: }
2012: }
2013: break;
2014: case 3: /* distributed assembled matrix input (size>1) */
2015: mumps->id.nnz_loc = mumps->nnz;
2016: mumps->id.irn_loc = mumps->irn;
2017: mumps->id.jcn_loc = mumps->jcn;
2018: if (mumps->id.ICNTL(6) > 1) mumps->id.a_loc = (MumpsScalar *)mumps->val;
2019: if (mumps->ICNTL20 == 0) { /* Centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
2020: PetscCall(MatCreateVecs(A, NULL, &b));
2021: PetscCall(VecScatterCreateToZero(b, &mumps->scat_rhs, &mumps->b_seq));
2022: PetscCall(VecDestroy(&b));
2023: }
2024: break;
2025: }
2026: PetscMUMPS_c(mumps);
2027: PetscCall(MatFactorSymbolic_MUMPS_ReportIfError(F, A, info, mumps));
2029: F->ops->lufactornumeric = MatFactorNumeric_MUMPS;
2030: F->ops->solve = MatSolve_MUMPS;
2031: F->ops->solvetranspose = MatSolveTranspose_MUMPS;
2032: F->ops->matsolve = MatMatSolve_MUMPS;
2033: F->ops->mattransposesolve = MatMatTransposeSolve_MUMPS;
2034: F->ops->matsolvetranspose = MatMatSolveTranspose_MUMPS;
2036: mumps->matstruc = SAME_NONZERO_PATTERN;
2037: PetscFunctionReturn(PETSC_SUCCESS);
2038: }
2040: /* Note the Petsc r and c permutations are ignored */
2041: PetscErrorCode MatLUFactorSymbolic_BAIJMUMPS(Mat F, Mat A, IS r, IS c, const MatFactorInfo *info)
2042: {
2043: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2044: Vec b;
2045: const PetscInt M = A->rmap->N;
2047: PetscFunctionBegin;
2048: if (mumps->matstruc == SAME_NONZERO_PATTERN) {
2049: /* F is assembled by a previous call of MatLUFactorSymbolic_AIJMUMPS() */
2050: PetscFunctionReturn(PETSC_SUCCESS);
2051: }
2053: /* Set MUMPS options from the options database */
2054: PetscCall(MatSetFromOptions_MUMPS(F, A));
2056: PetscCall((*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, mumps));
2057: PetscCall(MatMumpsGatherNonzerosOnMaster(MAT_INITIAL_MATRIX, mumps));
2059: /* analysis phase */
2060: mumps->id.job = JOB_FACTSYMBOLIC;
2061: mumps->id.n = M;
2062: switch (mumps->id.ICNTL(18)) {
2063: case 0: /* centralized assembled matrix input */
2064: if (!mumps->myid) {
2065: mumps->id.nnz = mumps->nnz;
2066: mumps->id.irn = mumps->irn;
2067: mumps->id.jcn = mumps->jcn;
2068: if (mumps->id.ICNTL(6) > 1) mumps->id.a = (MumpsScalar *)mumps->val;
2069: }
2070: break;
2071: case 3: /* distributed assembled matrix input (size>1) */
2072: mumps->id.nnz_loc = mumps->nnz;
2073: mumps->id.irn_loc = mumps->irn;
2074: mumps->id.jcn_loc = mumps->jcn;
2075: if (mumps->id.ICNTL(6) > 1) mumps->id.a_loc = (MumpsScalar *)mumps->val;
2076: if (mumps->ICNTL20 == 0) { /* Centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
2077: PetscCall(MatCreateVecs(A, NULL, &b));
2078: PetscCall(VecScatterCreateToZero(b, &mumps->scat_rhs, &mumps->b_seq));
2079: PetscCall(VecDestroy(&b));
2080: }
2081: break;
2082: }
2083: PetscMUMPS_c(mumps);
2084: PetscCall(MatFactorSymbolic_MUMPS_ReportIfError(F, A, info, mumps));
2086: F->ops->lufactornumeric = MatFactorNumeric_MUMPS;
2087: F->ops->solve = MatSolve_MUMPS;
2088: F->ops->solvetranspose = MatSolveTranspose_MUMPS;
2089: F->ops->matsolvetranspose = MatMatSolveTranspose_MUMPS;
2091: mumps->matstruc = SAME_NONZERO_PATTERN;
2092: PetscFunctionReturn(PETSC_SUCCESS);
2093: }
2095: /* Note the Petsc r permutation and factor info are ignored */
2096: PetscErrorCode MatCholeskyFactorSymbolic_MUMPS(Mat F, Mat A, IS r, const MatFactorInfo *info)
2097: {
2098: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2099: Vec b;
2100: const PetscInt M = A->rmap->N;
2102: PetscFunctionBegin;
2103: if (mumps->matstruc == SAME_NONZERO_PATTERN) {
2104: /* F is assembled by a previous call of MatLUFactorSymbolic_AIJMUMPS() */
2105: PetscFunctionReturn(PETSC_SUCCESS);
2106: }
2108: /* Set MUMPS options from the options database */
2109: PetscCall(MatSetFromOptions_MUMPS(F, A));
2111: PetscCall((*mumps->ConvertToTriples)(A, 1, MAT_INITIAL_MATRIX, mumps));
2112: PetscCall(MatMumpsGatherNonzerosOnMaster(MAT_INITIAL_MATRIX, mumps));
2114: /* analysis phase */
2115: mumps->id.job = JOB_FACTSYMBOLIC;
2116: mumps->id.n = M;
2117: switch (mumps->id.ICNTL(18)) {
2118: case 0: /* centralized assembled matrix input */
2119: if (!mumps->myid) {
2120: mumps->id.nnz = mumps->nnz;
2121: mumps->id.irn = mumps->irn;
2122: mumps->id.jcn = mumps->jcn;
2123: if (mumps->id.ICNTL(6) > 1) mumps->id.a = (MumpsScalar *)mumps->val;
2124: }
2125: break;
2126: case 3: /* distributed assembled matrix input (size>1) */
2127: mumps->id.nnz_loc = mumps->nnz;
2128: mumps->id.irn_loc = mumps->irn;
2129: mumps->id.jcn_loc = mumps->jcn;
2130: if (mumps->id.ICNTL(6) > 1) mumps->id.a_loc = (MumpsScalar *)mumps->val;
2131: if (mumps->ICNTL20 == 0) { /* Centralized rhs. Create scatter scat_rhs for repeated use in MatSolve() */
2132: PetscCall(MatCreateVecs(A, NULL, &b));
2133: PetscCall(VecScatterCreateToZero(b, &mumps->scat_rhs, &mumps->b_seq));
2134: PetscCall(VecDestroy(&b));
2135: }
2136: break;
2137: }
2138: PetscMUMPS_c(mumps);
2139: PetscCall(MatFactorSymbolic_MUMPS_ReportIfError(F, A, info, mumps));
2141: F->ops->choleskyfactornumeric = MatFactorNumeric_MUMPS;
2142: F->ops->solve = MatSolve_MUMPS;
2143: F->ops->solvetranspose = MatSolve_MUMPS;
2144: F->ops->matsolve = MatMatSolve_MUMPS;
2145: F->ops->mattransposesolve = MatMatTransposeSolve_MUMPS;
2146: F->ops->matsolvetranspose = MatMatSolveTranspose_MUMPS;
2147: #if defined(PETSC_USE_COMPLEX)
2148: F->ops->getinertia = NULL;
2149: #else
2150: F->ops->getinertia = MatGetInertia_SBAIJMUMPS;
2151: #endif
2153: mumps->matstruc = SAME_NONZERO_PATTERN;
2154: PetscFunctionReturn(PETSC_SUCCESS);
2155: }
2157: PetscErrorCode MatView_MUMPS(Mat A, PetscViewer viewer)
2158: {
2159: PetscBool iascii;
2160: PetscViewerFormat format;
2161: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
2163: PetscFunctionBegin;
2164: /* check if matrix is mumps type */
2165: if (A->ops->solve != MatSolve_MUMPS) PetscFunctionReturn(PETSC_SUCCESS);
2167: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &iascii));
2168: if (iascii) {
2169: PetscCall(PetscViewerGetFormat(viewer, &format));
2170: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
2171: PetscCall(PetscViewerASCIIPrintf(viewer, "MUMPS run parameters:\n"));
2172: if (format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
2173: PetscCall(PetscViewerASCIIPrintf(viewer, " SYM (matrix type): %d\n", mumps->id.sym));
2174: PetscCall(PetscViewerASCIIPrintf(viewer, " PAR (host participation): %d\n", mumps->id.par));
2175: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(1) (output for error): %d\n", mumps->id.ICNTL(1)));
2176: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(2) (output of diagnostic msg): %d\n", mumps->id.ICNTL(2)));
2177: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(3) (output for global info): %d\n", mumps->id.ICNTL(3)));
2178: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(4) (level of printing): %d\n", mumps->id.ICNTL(4)));
2179: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(5) (input mat struct): %d\n", mumps->id.ICNTL(5)));
2180: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(6) (matrix prescaling): %d\n", mumps->id.ICNTL(6)));
2181: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(7) (sequential matrix ordering):%d\n", mumps->id.ICNTL(7)));
2182: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(8) (scaling strategy): %d\n", mumps->id.ICNTL(8)));
2183: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(10) (max num of refinements): %d\n", mumps->id.ICNTL(10)));
2184: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(11) (error analysis): %d\n", mumps->id.ICNTL(11)));
2185: if (mumps->id.ICNTL(11) > 0) {
2186: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(4) (inf norm of input mat): %g\n", mumps->id.RINFOG(4)));
2187: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(5) (inf norm of solution): %g\n", mumps->id.RINFOG(5)));
2188: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(6) (inf norm of residual): %g\n", mumps->id.RINFOG(6)));
2189: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(7),RINFOG(8) (backward error est): %g, %g\n", mumps->id.RINFOG(7), mumps->id.RINFOG(8)));
2190: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(9) (error estimate): %g\n", mumps->id.RINFOG(9)));
2191: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(10),RINFOG(11)(condition numbers): %g, %g\n", mumps->id.RINFOG(10), mumps->id.RINFOG(11)));
2192: }
2193: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(12) (efficiency control): %d\n", mumps->id.ICNTL(12)));
2194: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(13) (sequential factorization of the root node): %d\n", mumps->id.ICNTL(13)));
2195: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(14) (percentage of estimated workspace increase): %d\n", mumps->id.ICNTL(14)));
2196: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(15) (compression of the input matrix): %d\n", mumps->id.ICNTL(15)));
2197: /* ICNTL(15-17) not used */
2198: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(18) (input mat struct): %d\n", mumps->id.ICNTL(18)));
2199: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(19) (Schur complement info): %d\n", mumps->id.ICNTL(19)));
2200: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(20) (RHS sparse pattern): %d\n", mumps->id.ICNTL(20)));
2201: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(21) (solution struct): %d\n", mumps->id.ICNTL(21)));
2202: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(22) (in-core/out-of-core facility): %d\n", mumps->id.ICNTL(22)));
2203: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(23) (max size of memory can be allocated locally):%d\n", mumps->id.ICNTL(23)));
2205: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(24) (detection of null pivot rows): %d\n", mumps->id.ICNTL(24)));
2206: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(25) (computation of a null space basis): %d\n", mumps->id.ICNTL(25)));
2207: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(26) (Schur options for RHS or solution): %d\n", mumps->id.ICNTL(26)));
2208: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(27) (blocking size for multiple RHS): %d\n", mumps->id.ICNTL(27)));
2209: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(28) (use parallel or sequential ordering): %d\n", mumps->id.ICNTL(28)));
2210: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(29) (parallel ordering): %d\n", mumps->id.ICNTL(29)));
2212: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(30) (user-specified set of entries in inv(A)): %d\n", mumps->id.ICNTL(30)));
2213: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(31) (factors is discarded in the solve phase): %d\n", mumps->id.ICNTL(31)));
2214: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(33) (compute determinant): %d\n", mumps->id.ICNTL(33)));
2215: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(35) (activate BLR based factorization): %d\n", mumps->id.ICNTL(35)));
2216: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(36) (choice of BLR factorization variant): %d\n", mumps->id.ICNTL(36)));
2217: PetscCall(PetscViewerASCIIPrintf(viewer, " ICNTL(38) (estimated compression rate of LU factors): %d\n", mumps->id.ICNTL(38)));
2219: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(1) (relative pivoting threshold): %g\n", mumps->id.CNTL(1)));
2220: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(2) (stopping criterion of refinement): %g\n", mumps->id.CNTL(2)));
2221: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(3) (absolute pivoting threshold): %g\n", mumps->id.CNTL(3)));
2222: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(4) (value of static pivoting): %g\n", mumps->id.CNTL(4)));
2223: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(5) (fixation for null pivots): %g\n", mumps->id.CNTL(5)));
2224: PetscCall(PetscViewerASCIIPrintf(viewer, " CNTL(7) (dropping parameter for BLR): %g\n", mumps->id.CNTL(7)));
2226: /* information local to each processor */
2227: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFO(1) (local estimated flops for the elimination after analysis):\n"));
2228: PetscCall(PetscViewerASCIIPushSynchronized(viewer));
2229: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %g\n", mumps->myid, mumps->id.RINFO(1)));
2230: PetscCall(PetscViewerFlush(viewer));
2231: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFO(2) (local estimated flops for the assembly after factorization):\n"));
2232: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %g\n", mumps->myid, mumps->id.RINFO(2)));
2233: PetscCall(PetscViewerFlush(viewer));
2234: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFO(3) (local estimated flops for the elimination after factorization):\n"));
2235: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %g\n", mumps->myid, mumps->id.RINFO(3)));
2236: PetscCall(PetscViewerFlush(viewer));
2238: PetscCall(PetscViewerASCIIPrintf(viewer, " INFO(15) (estimated size of (in MB) MUMPS internal data for running numerical factorization):\n"));
2239: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %d\n", mumps->myid, mumps->id.INFO(15)));
2240: PetscCall(PetscViewerFlush(viewer));
2242: PetscCall(PetscViewerASCIIPrintf(viewer, " INFO(16) (size of (in MB) MUMPS internal data used during numerical factorization):\n"));
2243: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %d\n", mumps->myid, mumps->id.INFO(16)));
2244: PetscCall(PetscViewerFlush(viewer));
2246: PetscCall(PetscViewerASCIIPrintf(viewer, " INFO(23) (num of pivots eliminated on this processor after factorization):\n"));
2247: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %d\n", mumps->myid, mumps->id.INFO(23)));
2248: PetscCall(PetscViewerFlush(viewer));
2250: if (mumps->ninfo && mumps->ninfo <= 80) {
2251: PetscInt i;
2252: for (i = 0; i < mumps->ninfo; i++) {
2253: PetscCall(PetscViewerASCIIPrintf(viewer, " INFO(%" PetscInt_FMT "):\n", mumps->info[i]));
2254: PetscCall(PetscViewerASCIISynchronizedPrintf(viewer, " [%d] %d\n", mumps->myid, mumps->id.INFO(mumps->info[i])));
2255: PetscCall(PetscViewerFlush(viewer));
2256: }
2257: }
2258: PetscCall(PetscViewerASCIIPopSynchronized(viewer));
2259: } else PetscCall(PetscViewerASCIIPrintf(viewer, " Use -%sksp_view ::ascii_info_detail to display information for all processes\n", ((PetscObject)A)->prefix ? ((PetscObject)A)->prefix : ""));
2261: if (mumps->myid == 0) { /* information from the host */
2262: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(1) (global estimated flops for the elimination after analysis): %g\n", mumps->id.RINFOG(1)));
2263: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(2) (global estimated flops for the assembly after factorization): %g\n", mumps->id.RINFOG(2)));
2264: PetscCall(PetscViewerASCIIPrintf(viewer, " RINFOG(3) (global estimated flops for the elimination after factorization): %g\n", mumps->id.RINFOG(3)));
2265: PetscCall(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)));
2267: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(3) (estimated real workspace for factors on all processors after analysis): %d\n", mumps->id.INFOG(3)));
2268: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(4) (estimated integer workspace for factors on all processors after analysis): %d\n", mumps->id.INFOG(4)));
2269: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(5) (estimated maximum front size in the complete tree): %d\n", mumps->id.INFOG(5)));
2270: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(6) (number of nodes in the complete tree): %d\n", mumps->id.INFOG(6)));
2271: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(7) (ordering option effectively used after analysis): %d\n", mumps->id.INFOG(7)));
2272: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(8) (structural symmetry in percent of the permuted matrix after analysis): %d\n", mumps->id.INFOG(8)));
2273: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(9) (total real/complex workspace to store the matrix factors after factorization): %d\n", mumps->id.INFOG(9)));
2274: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(10) (total integer space store the matrix factors after factorization): %d\n", mumps->id.INFOG(10)));
2275: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(11) (order of largest frontal matrix after factorization): %d\n", mumps->id.INFOG(11)));
2276: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(12) (number of off-diagonal pivots): %d\n", mumps->id.INFOG(12)));
2277: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(13) (number of delayed pivots after factorization): %d\n", mumps->id.INFOG(13)));
2278: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(14) (number of memory compress after factorization): %d\n", mumps->id.INFOG(14)));
2279: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(15) (number of steps of iterative refinement after solution): %d\n", mumps->id.INFOG(15)));
2280: PetscCall(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)));
2281: PetscCall(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)));
2282: PetscCall(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)));
2283: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(19) (size of all MUMPS internal data allocated during factorization: sum over all processors): %d\n", mumps->id.INFOG(19)));
2284: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(20) (estimated number of entries in the factors): %d\n", mumps->id.INFOG(20)));
2285: PetscCall(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)));
2286: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(22) (size in MB of memory effectively used during factorization - sum over all processors): %d\n", mumps->id.INFOG(22)));
2287: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(23) (after analysis: value of ICNTL(6) effectively used): %d\n", mumps->id.INFOG(23)));
2288: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(24) (after analysis: value of ICNTL(12) effectively used): %d\n", mumps->id.INFOG(24)));
2289: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(25) (after factorization: number of pivots modified by static pivoting): %d\n", mumps->id.INFOG(25)));
2290: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(28) (after factorization: number of null pivots encountered): %d\n", mumps->id.INFOG(28)));
2291: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(29) (after factorization: effective number of entries in the factors (sum over all processors)): %d\n", mumps->id.INFOG(29)));
2292: PetscCall(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)));
2293: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(32) (after analysis: type of analysis done): %d\n", mumps->id.INFOG(32)));
2294: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(33) (value used for ICNTL(8)): %d\n", mumps->id.INFOG(33)));
2295: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(34) (exponent of the determinant if determinant is requested): %d\n", mumps->id.INFOG(34)));
2296: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(35) (after factorization: number of entries taking into account BLR factor compression - sum over all processors): %d\n", mumps->id.INFOG(35)));
2297: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(36) (after analysis: estimated size of all MUMPS internal data for running BLR in-core - value on the most memory consuming processor): %d\n", mumps->id.INFOG(36)));
2298: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(37) (after analysis: estimated size of all MUMPS internal data for running BLR in-core - sum over all processors): %d\n", mumps->id.INFOG(37)));
2299: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(38) (after analysis: estimated size of all MUMPS internal data for running BLR out-of-core - value on the most memory consuming processor): %d\n", mumps->id.INFOG(38)));
2300: PetscCall(PetscViewerASCIIPrintf(viewer, " INFOG(39) (after analysis: estimated size of all MUMPS internal data for running BLR out-of-core - sum over all processors): %d\n", mumps->id.INFOG(39)));
2301: }
2302: }
2303: }
2304: PetscFunctionReturn(PETSC_SUCCESS);
2305: }
2307: PetscErrorCode MatGetInfo_MUMPS(Mat A, MatInfoType flag, MatInfo *info)
2308: {
2309: Mat_MUMPS *mumps = (Mat_MUMPS *)A->data;
2311: PetscFunctionBegin;
2312: info->block_size = 1.0;
2313: info->nz_allocated = mumps->id.INFOG(20);
2314: info->nz_used = mumps->id.INFOG(20);
2315: info->nz_unneeded = 0.0;
2316: info->assemblies = 0.0;
2317: info->mallocs = 0.0;
2318: info->memory = 0.0;
2319: info->fill_ratio_given = 0;
2320: info->fill_ratio_needed = 0;
2321: info->factor_mallocs = 0;
2322: PetscFunctionReturn(PETSC_SUCCESS);
2323: }
2325: PetscErrorCode MatFactorSetSchurIS_MUMPS(Mat F, IS is)
2326: {
2327: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2328: const PetscScalar *arr;
2329: const PetscInt *idxs;
2330: PetscInt size, i;
2332: PetscFunctionBegin;
2333: PetscCall(ISGetLocalSize(is, &size));
2334: /* Schur complement matrix */
2335: PetscCall(MatDestroy(&F->schur));
2336: PetscCall(MatCreateSeqDense(PETSC_COMM_SELF, size, size, NULL, &F->schur));
2337: PetscCall(MatDenseGetArrayRead(F->schur, &arr));
2338: mumps->id.schur = (MumpsScalar *)arr;
2339: mumps->id.size_schur = size;
2340: mumps->id.schur_lld = size;
2341: PetscCall(MatDenseRestoreArrayRead(F->schur, &arr));
2342: if (mumps->sym == 1) PetscCall(MatSetOption(F->schur, MAT_SPD, PETSC_TRUE));
2344: /* MUMPS expects Fortran style indices */
2345: PetscCall(PetscFree(mumps->id.listvar_schur));
2346: PetscCall(PetscMalloc1(size, &mumps->id.listvar_schur));
2347: PetscCall(ISGetIndices(is, &idxs));
2348: for (i = 0; i < size; i++) PetscCall(PetscMUMPSIntCast(idxs[i] + 1, &(mumps->id.listvar_schur[i])));
2349: PetscCall(ISRestoreIndices(is, &idxs));
2350: /* set a special value of ICNTL (not handled my MUMPS) to be used in the solve phase by PETSc */
2351: mumps->id.ICNTL(26) = -1;
2352: PetscFunctionReturn(PETSC_SUCCESS);
2353: }
2355: PetscErrorCode MatFactorCreateSchurComplement_MUMPS(Mat F, Mat *S)
2356: {
2357: Mat St;
2358: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2359: PetscScalar *array;
2360: #if defined(PETSC_USE_COMPLEX)
2361: PetscScalar im = PetscSqrtScalar((PetscScalar)-1.0);
2362: #endif
2364: PetscFunctionBegin;
2365: PetscCheck(mumps->id.ICNTL(19), PetscObjectComm((PetscObject)F), PETSC_ERR_ORDER, "Schur complement mode not selected! You should call MatFactorSetSchurIS to enable it");
2366: PetscCall(MatCreate(PETSC_COMM_SELF, &St));
2367: PetscCall(MatSetSizes(St, PETSC_DECIDE, PETSC_DECIDE, mumps->id.size_schur, mumps->id.size_schur));
2368: PetscCall(MatSetType(St, MATDENSE));
2369: PetscCall(MatSetUp(St));
2370: PetscCall(MatDenseGetArray(St, &array));
2371: if (!mumps->sym) { /* MUMPS always return a full matrix */
2372: if (mumps->id.ICNTL(19) == 1) { /* stored by rows */
2373: PetscInt i, j, N = mumps->id.size_schur;
2374: for (i = 0; i < N; i++) {
2375: for (j = 0; j < N; j++) {
2376: #if !defined(PETSC_USE_COMPLEX)
2377: PetscScalar val = mumps->id.schur[i * N + j];
2378: #else
2379: PetscScalar val = mumps->id.schur[i * N + j].r + im * mumps->id.schur[i * N + j].i;
2380: #endif
2381: array[j * N + i] = val;
2382: }
2383: }
2384: } else { /* stored by columns */
2385: PetscCall(PetscArraycpy(array, mumps->id.schur, mumps->id.size_schur * mumps->id.size_schur));
2386: }
2387: } else { /* either full or lower-triangular (not packed) */
2388: if (mumps->id.ICNTL(19) == 2) { /* lower triangular stored by columns */
2389: PetscInt i, j, N = mumps->id.size_schur;
2390: for (i = 0; i < N; i++) {
2391: for (j = i; j < N; j++) {
2392: #if !defined(PETSC_USE_COMPLEX)
2393: PetscScalar val = mumps->id.schur[i * N + j];
2394: #else
2395: PetscScalar val = mumps->id.schur[i * N + j].r + im * mumps->id.schur[i * N + j].i;
2396: #endif
2397: array[i * N + j] = val;
2398: array[j * N + i] = val;
2399: }
2400: }
2401: } else if (mumps->id.ICNTL(19) == 3) { /* full matrix */
2402: PetscCall(PetscArraycpy(array, mumps->id.schur, mumps->id.size_schur * mumps->id.size_schur));
2403: } else { /* ICNTL(19) == 1 lower triangular stored by rows */
2404: PetscInt i, j, N = mumps->id.size_schur;
2405: for (i = 0; i < N; i++) {
2406: for (j = 0; j < i + 1; j++) {
2407: #if !defined(PETSC_USE_COMPLEX)
2408: PetscScalar val = mumps->id.schur[i * N + j];
2409: #else
2410: PetscScalar val = mumps->id.schur[i * N + j].r + im * mumps->id.schur[i * N + j].i;
2411: #endif
2412: array[i * N + j] = val;
2413: array[j * N + i] = val;
2414: }
2415: }
2416: }
2417: }
2418: PetscCall(MatDenseRestoreArray(St, &array));
2419: *S = St;
2420: PetscFunctionReturn(PETSC_SUCCESS);
2421: }
2423: PetscErrorCode MatMumpsSetIcntl_MUMPS(Mat F, PetscInt icntl, PetscInt ival)
2424: {
2425: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2427: PetscFunctionBegin;
2428: if (mumps->id.job == JOB_NULL) { /* need to cache icntl and ival since PetscMUMPS_c() has never been called */
2429: PetscInt i, nICNTL_pre = mumps->ICNTL_pre ? mumps->ICNTL_pre[0] : 0; /* number of already cached ICNTL */
2430: for (i = 0; i < nICNTL_pre; ++i)
2431: if (mumps->ICNTL_pre[1 + 2 * i] == icntl) break; /* is this ICNTL already cached? */
2432: if (i == nICNTL_pre) { /* not already cached */
2433: if (i > 0) PetscCall(PetscRealloc(sizeof(PetscMUMPSInt) * (2 * nICNTL_pre + 3), &mumps->ICNTL_pre));
2434: else PetscCall(PetscCalloc(sizeof(PetscMUMPSInt) * 3, &mumps->ICNTL_pre));
2435: mumps->ICNTL_pre[0]++;
2436: }
2437: mumps->ICNTL_pre[1 + 2 * i] = icntl;
2438: PetscCall(PetscMUMPSIntCast(ival, mumps->ICNTL_pre + 2 + 2 * i));
2439: } else PetscCall(PetscMUMPSIntCast(ival, &mumps->id.ICNTL(icntl)));
2440: PetscFunctionReturn(PETSC_SUCCESS);
2441: }
2443: PetscErrorCode MatMumpsGetIcntl_MUMPS(Mat F, PetscInt icntl, PetscInt *ival)
2444: {
2445: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2447: PetscFunctionBegin;
2448: if (mumps->id.job == JOB_NULL) {
2449: PetscInt i, nICNTL_pre = mumps->ICNTL_pre ? mumps->ICNTL_pre[0] : 0;
2450: *ival = 0;
2451: for (i = 0; i < nICNTL_pre; ++i) {
2452: if (mumps->ICNTL_pre[1 + 2 * i] == icntl) *ival = mumps->ICNTL_pre[2 + 2 * i];
2453: }
2454: } else *ival = mumps->id.ICNTL(icntl);
2455: PetscFunctionReturn(PETSC_SUCCESS);
2456: }
2458: /*@
2459: MatMumpsSetIcntl - Set MUMPS parameter ICNTL()
2461: Logically Collective
2463: Input Parameters:
2464: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2465: . icntl - index of MUMPS parameter array ICNTL()
2466: - ival - value of MUMPS ICNTL(icntl)
2468: Options Database Key:
2469: . -mat_mumps_icntl_<icntl> <ival> - change the option numbered icntl to ival
2471: Level: beginner
2473: References:
2474: . * - MUMPS Users' Guide
2476: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2477: @*/
2478: PetscErrorCode MatMumpsSetIcntl(Mat F, PetscInt icntl, PetscInt ival)
2479: {
2480: PetscFunctionBegin;
2482: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2485: PetscCheck(icntl >= 1 && icntl <= 38, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONG, "Unsupported ICNTL value %" PetscInt_FMT, icntl);
2486: PetscTryMethod(F, "MatMumpsSetIcntl_C", (Mat, PetscInt, PetscInt), (F, icntl, ival));
2487: PetscFunctionReturn(PETSC_SUCCESS);
2488: }
2490: /*@
2491: MatMumpsGetIcntl - Get MUMPS parameter ICNTL()
2493: Logically Collective
2495: Input Parameters:
2496: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2497: - icntl - index of MUMPS parameter array ICNTL()
2499: Output Parameter:
2500: . ival - value of MUMPS ICNTL(icntl)
2502: Level: beginner
2504: References:
2505: . * - MUMPS Users' Guide
2507: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2508: @*/
2509: PetscErrorCode MatMumpsGetIcntl(Mat F, PetscInt icntl, PetscInt *ival)
2510: {
2511: PetscFunctionBegin;
2513: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2516: PetscCheck(icntl >= 1 && icntl <= 38, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONG, "Unsupported ICNTL value %" PetscInt_FMT, icntl);
2517: PetscUseMethod(F, "MatMumpsGetIcntl_C", (Mat, PetscInt, PetscInt *), (F, icntl, ival));
2518: PetscFunctionReturn(PETSC_SUCCESS);
2519: }
2521: PetscErrorCode MatMumpsSetCntl_MUMPS(Mat F, PetscInt icntl, PetscReal val)
2522: {
2523: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2525: PetscFunctionBegin;
2526: if (mumps->id.job == JOB_NULL) {
2527: PetscInt i, nCNTL_pre = mumps->CNTL_pre ? mumps->CNTL_pre[0] : 0;
2528: for (i = 0; i < nCNTL_pre; ++i)
2529: if (mumps->CNTL_pre[1 + 2 * i] == icntl) break;
2530: if (i == nCNTL_pre) {
2531: if (i > 0) PetscCall(PetscRealloc(sizeof(PetscReal) * (2 * nCNTL_pre + 3), &mumps->CNTL_pre));
2532: else PetscCall(PetscCalloc(sizeof(PetscReal) * 3, &mumps->CNTL_pre));
2533: mumps->CNTL_pre[0]++;
2534: }
2535: mumps->CNTL_pre[1 + 2 * i] = icntl;
2536: mumps->CNTL_pre[2 + 2 * i] = val;
2537: } else mumps->id.CNTL(icntl) = val;
2538: PetscFunctionReturn(PETSC_SUCCESS);
2539: }
2541: PetscErrorCode MatMumpsGetCntl_MUMPS(Mat F, PetscInt icntl, PetscReal *val)
2542: {
2543: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2545: PetscFunctionBegin;
2546: if (mumps->id.job == JOB_NULL) {
2547: PetscInt i, nCNTL_pre = mumps->CNTL_pre ? mumps->CNTL_pre[0] : 0;
2548: *val = 0.0;
2549: for (i = 0; i < nCNTL_pre; ++i) {
2550: if (mumps->CNTL_pre[1 + 2 * i] == icntl) *val = mumps->CNTL_pre[2 + 2 * i];
2551: }
2552: } else *val = mumps->id.CNTL(icntl);
2553: PetscFunctionReturn(PETSC_SUCCESS);
2554: }
2556: /*@
2557: MatMumpsSetCntl - Set MUMPS parameter CNTL()
2559: Logically Collective
2561: Input Parameters:
2562: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2563: . icntl - index of MUMPS parameter array CNTL()
2564: - val - value of MUMPS CNTL(icntl)
2566: Options Database Key:
2567: . -mat_mumps_cntl_<icntl> <val> - change the option numbered icntl to ival
2569: Level: beginner
2571: References:
2572: . * - MUMPS Users' Guide
2574: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2575: @*/
2576: PetscErrorCode MatMumpsSetCntl(Mat F, PetscInt icntl, PetscReal val)
2577: {
2578: PetscFunctionBegin;
2580: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2583: PetscCheck(icntl >= 1 && icntl <= 7, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONG, "Unsupported CNTL value %" PetscInt_FMT, icntl);
2584: PetscTryMethod(F, "MatMumpsSetCntl_C", (Mat, PetscInt, PetscReal), (F, icntl, val));
2585: PetscFunctionReturn(PETSC_SUCCESS);
2586: }
2588: /*@
2589: MatMumpsGetCntl - Get MUMPS parameter CNTL()
2591: Logically Collective
2593: Input Parameters:
2594: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2595: - icntl - index of MUMPS parameter array CNTL()
2597: Output Parameter:
2598: . val - value of MUMPS CNTL(icntl)
2600: Level: beginner
2602: References:
2603: . * - MUMPS Users' Guide
2605: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2606: @*/
2607: PetscErrorCode MatMumpsGetCntl(Mat F, PetscInt icntl, PetscReal *val)
2608: {
2609: PetscFunctionBegin;
2611: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2614: PetscCheck(icntl >= 1 && icntl <= 7, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONG, "Unsupported CNTL value %" PetscInt_FMT, icntl);
2615: PetscUseMethod(F, "MatMumpsGetCntl_C", (Mat, PetscInt, PetscReal *), (F, icntl, val));
2616: PetscFunctionReturn(PETSC_SUCCESS);
2617: }
2619: PetscErrorCode MatMumpsGetInfo_MUMPS(Mat F, PetscInt icntl, PetscInt *info)
2620: {
2621: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2623: PetscFunctionBegin;
2624: *info = mumps->id.INFO(icntl);
2625: PetscFunctionReturn(PETSC_SUCCESS);
2626: }
2628: PetscErrorCode MatMumpsGetInfog_MUMPS(Mat F, PetscInt icntl, PetscInt *infog)
2629: {
2630: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2632: PetscFunctionBegin;
2633: *infog = mumps->id.INFOG(icntl);
2634: PetscFunctionReturn(PETSC_SUCCESS);
2635: }
2637: PetscErrorCode MatMumpsGetRinfo_MUMPS(Mat F, PetscInt icntl, PetscReal *rinfo)
2638: {
2639: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2641: PetscFunctionBegin;
2642: *rinfo = mumps->id.RINFO(icntl);
2643: PetscFunctionReturn(PETSC_SUCCESS);
2644: }
2646: PetscErrorCode MatMumpsGetRinfog_MUMPS(Mat F, PetscInt icntl, PetscReal *rinfog)
2647: {
2648: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2650: PetscFunctionBegin;
2651: *rinfog = mumps->id.RINFOG(icntl);
2652: PetscFunctionReturn(PETSC_SUCCESS);
2653: }
2655: PetscErrorCode MatMumpsGetNullPivots_MUMPS(Mat F, PetscInt *size, PetscInt **array)
2656: {
2657: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2659: PetscFunctionBegin;
2660: PetscCheck(mumps->id.ICNTL(24) == 1, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "-mat_mumps_icntl_24 must be set as 1 for null pivot row detection");
2661: *size = 0;
2662: *array = NULL;
2663: if (!mumps->myid) {
2664: *size = mumps->id.INFOG(28);
2665: PetscCall(PetscMalloc1(*size, array));
2666: for (int i = 0; i < *size; i++) (*array)[i] = mumps->id.pivnul_list[i] - 1;
2667: }
2668: PetscFunctionReturn(PETSC_SUCCESS);
2669: }
2671: PetscErrorCode MatMumpsGetInverse_MUMPS(Mat F, Mat spRHS)
2672: {
2673: Mat Bt = NULL, Btseq = NULL;
2674: PetscBool flg;
2675: Mat_MUMPS *mumps = (Mat_MUMPS *)F->data;
2676: PetscScalar *aa;
2677: PetscInt spnr, *ia, *ja, M, nrhs;
2679: PetscFunctionBegin;
2681: PetscCall(PetscObjectTypeCompare((PetscObject)spRHS, MATTRANSPOSEVIRTUAL, &flg));
2682: if (flg) {
2683: PetscCall(MatTransposeGetMat(spRHS, &Bt));
2684: } else SETERRQ(PetscObjectComm((PetscObject)spRHS), PETSC_ERR_ARG_WRONG, "Matrix spRHS must be type MATTRANSPOSEVIRTUAL matrix");
2686: PetscCall(MatMumpsSetIcntl(F, 30, 1));
2688: if (mumps->petsc_size > 1) {
2689: Mat_MPIAIJ *b = (Mat_MPIAIJ *)Bt->data;
2690: Btseq = b->A;
2691: } else {
2692: Btseq = Bt;
2693: }
2695: PetscCall(MatGetSize(spRHS, &M, &nrhs));
2696: mumps->id.nrhs = nrhs;
2697: mumps->id.lrhs = M;
2698: mumps->id.rhs = NULL;
2700: if (!mumps->myid) {
2701: PetscCall(MatSeqAIJGetArray(Btseq, &aa));
2702: PetscCall(MatGetRowIJ(Btseq, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
2703: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot get IJ structure");
2704: PetscCall(PetscMUMPSIntCSRCast(mumps, spnr, ia, ja, &mumps->id.irhs_ptr, &mumps->id.irhs_sparse, &mumps->id.nz_rhs));
2705: mumps->id.rhs_sparse = (MumpsScalar *)aa;
2706: } else {
2707: mumps->id.irhs_ptr = NULL;
2708: mumps->id.irhs_sparse = NULL;
2709: mumps->id.nz_rhs = 0;
2710: mumps->id.rhs_sparse = NULL;
2711: }
2712: mumps->id.ICNTL(20) = 1; /* rhs is sparse */
2713: mumps->id.ICNTL(21) = 0; /* solution is in assembled centralized format */
2715: /* solve phase */
2716: mumps->id.job = JOB_SOLVE;
2717: PetscMUMPS_c(mumps);
2718: PetscCheck(mumps->id.INFOG(1) >= 0, PETSC_COMM_SELF, PETSC_ERR_LIB, "Error reported by MUMPS in solve phase: INFOG(1)=%d INFO(2)=%d", mumps->id.INFOG(1), mumps->id.INFO(2));
2720: if (!mumps->myid) {
2721: PetscCall(MatSeqAIJRestoreArray(Btseq, &aa));
2722: PetscCall(MatRestoreRowIJ(Btseq, 1, PETSC_FALSE, PETSC_FALSE, &spnr, (const PetscInt **)&ia, (const PetscInt **)&ja, &flg));
2723: PetscCheck(flg, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Cannot get IJ structure");
2724: }
2725: PetscFunctionReturn(PETSC_SUCCESS);
2726: }
2728: /*@
2729: MatMumpsGetInverse - Get user-specified set of entries in inverse of `A`
2731: Logically Collective
2733: Input Parameter:
2734: . F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2736: Output Parameter:
2737: . spRHS - sequential sparse matrix in `MATTRANSPOSEVIRTUAL` format with requested entries of inverse of `A`
2739: Level: beginner
2741: References:
2742: . * - MUMPS Users' Guide
2744: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatCreateTranspose()`
2745: @*/
2746: PetscErrorCode MatMumpsGetInverse(Mat F, Mat spRHS)
2747: {
2748: PetscFunctionBegin;
2750: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2751: PetscUseMethod(F, "MatMumpsGetInverse_C", (Mat, Mat), (F, spRHS));
2752: PetscFunctionReturn(PETSC_SUCCESS);
2753: }
2755: PetscErrorCode MatMumpsGetInverseTranspose_MUMPS(Mat F, Mat spRHST)
2756: {
2757: Mat spRHS;
2759: PetscFunctionBegin;
2760: PetscCall(MatCreateTranspose(spRHST, &spRHS));
2761: PetscCall(MatMumpsGetInverse_MUMPS(F, spRHS));
2762: PetscCall(MatDestroy(&spRHS));
2763: PetscFunctionReturn(PETSC_SUCCESS);
2764: }
2766: /*@
2767: MatMumpsGetInverseTranspose - Get user-specified set of entries in inverse of matrix `A`^T
2769: Logically Collective
2771: Input Parameter:
2772: . F - the factored matrix of A obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2774: Output Parameter:
2775: . spRHST - sequential sparse matrix in `MATAIJ` format containing the requested entries of inverse of `A`^T
2777: Level: beginner
2779: References:
2780: . * - MUMPS Users' Guide
2782: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatCreateTranspose()`, `MatMumpsGetInverse()`
2783: @*/
2784: PetscErrorCode MatMumpsGetInverseTranspose(Mat F, Mat spRHST)
2785: {
2786: PetscBool flg;
2788: PetscFunctionBegin;
2790: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2791: PetscCall(PetscObjectTypeCompareAny((PetscObject)spRHST, &flg, MATSEQAIJ, MATMPIAIJ, NULL));
2792: PetscCheck(flg, PetscObjectComm((PetscObject)spRHST), PETSC_ERR_ARG_WRONG, "Matrix spRHST must be MATAIJ matrix");
2794: PetscUseMethod(F, "MatMumpsGetInverseTranspose_C", (Mat, Mat), (F, spRHST));
2795: PetscFunctionReturn(PETSC_SUCCESS);
2796: }
2798: /*@
2799: MatMumpsGetInfo - Get MUMPS parameter INFO()
2801: Logically Collective
2803: Input Parameters:
2804: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2805: - icntl - index of MUMPS parameter array INFO()
2807: Output Parameter:
2808: . ival - value of MUMPS INFO(icntl)
2810: Level: beginner
2812: References:
2813: . * - MUMPS Users' Guide
2815: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2816: @*/
2817: PetscErrorCode MatMumpsGetInfo(Mat F, PetscInt icntl, PetscInt *ival)
2818: {
2819: PetscFunctionBegin;
2821: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2823: PetscUseMethod(F, "MatMumpsGetInfo_C", (Mat, PetscInt, PetscInt *), (F, icntl, ival));
2824: PetscFunctionReturn(PETSC_SUCCESS);
2825: }
2827: /*@
2828: MatMumpsGetInfog - Get MUMPS parameter INFOG()
2830: Logically Collective
2832: Input Parameters:
2833: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2834: - icntl - index of MUMPS parameter array INFOG()
2836: Output Parameter:
2837: . ival - value of MUMPS INFOG(icntl)
2839: Level: beginner
2841: References:
2842: . * - MUMPS Users' Guide
2844: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`
2845: @*/
2846: PetscErrorCode MatMumpsGetInfog(Mat F, PetscInt icntl, PetscInt *ival)
2847: {
2848: PetscFunctionBegin;
2850: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2852: PetscUseMethod(F, "MatMumpsGetInfog_C", (Mat, PetscInt, PetscInt *), (F, icntl, ival));
2853: PetscFunctionReturn(PETSC_SUCCESS);
2854: }
2856: /*@
2857: MatMumpsGetRinfo - Get MUMPS parameter RINFO()
2859: Logically Collective
2861: Input Parameters:
2862: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2863: - icntl - index of MUMPS parameter array RINFO()
2865: Output Parameter:
2866: . val - value of MUMPS RINFO(icntl)
2868: Level: beginner
2870: References:
2871: . * - MUMPS Users' Guide
2873: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfog()`
2874: @*/
2875: PetscErrorCode MatMumpsGetRinfo(Mat F, PetscInt icntl, PetscReal *val)
2876: {
2877: PetscFunctionBegin;
2879: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2881: PetscUseMethod(F, "MatMumpsGetRinfo_C", (Mat, PetscInt, PetscReal *), (F, icntl, val));
2882: PetscFunctionReturn(PETSC_SUCCESS);
2883: }
2885: /*@
2886: MatMumpsGetRinfog - Get MUMPS parameter RINFOG()
2888: Logically Collective
2890: Input Parameters:
2891: + F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2892: - icntl - index of MUMPS parameter array RINFOG()
2894: Output Parameter:
2895: . val - value of MUMPS RINFOG(icntl)
2897: Level: beginner
2899: References:
2900: . * - MUMPS Users' Guide
2902: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`
2903: @*/
2904: PetscErrorCode MatMumpsGetRinfog(Mat F, PetscInt icntl, PetscReal *val)
2905: {
2906: PetscFunctionBegin;
2908: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2910: PetscUseMethod(F, "MatMumpsGetRinfog_C", (Mat, PetscInt, PetscReal *), (F, icntl, val));
2911: PetscFunctionReturn(PETSC_SUCCESS);
2912: }
2914: /*@
2915: MatMumpsGetNullPivots - Get MUMPS parameter PIVNUL_LIST()
2917: Logically Collective
2919: Input Parameter:
2920: . F - the factored matrix obtained by calling `MatGetFactor()` from PETSc-MUMPS interface
2922: Output Parameters:
2923: + size - local size of the array. The size of the array is non-zero only on the host.
2924: - array - array of rows with null pivot, these rows follow 0-based indexing. The array gets allocated within the function and the user is responsible
2925: for freeing this array.
2927: Level: beginner
2929: References:
2930: . * - MUMPS Users' Guide
2932: .seealso: [](chapter_matrices), `Mat`, `MatGetFactor()`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`
2933: @*/
2934: PetscErrorCode MatMumpsGetNullPivots(Mat F, PetscInt *size, PetscInt **array)
2935: {
2936: PetscFunctionBegin;
2938: PetscCheck(F->factortype, PetscObjectComm((PetscObject)F), PETSC_ERR_ARG_WRONGSTATE, "Only for factored matrix");
2941: PetscUseMethod(F, "MatMumpsGetNullPivots_C", (Mat, PetscInt *, PetscInt **), (F, size, array));
2942: PetscFunctionReturn(PETSC_SUCCESS);
2943: }
2945: /*MC
2946: MATSOLVERMUMPS - A matrix type providing direct solvers (LU and Cholesky) for
2947: distributed and sequential matrices via the external package MUMPS.
2949: Works with `MATAIJ` and `MATSBAIJ` matrices
2951: Use ./configure --download-mumps --download-scalapack --download-parmetis --download-metis --download-ptscotch to have PETSc installed with MUMPS
2953: Use ./configure --with-openmp --download-hwloc (or --with-hwloc) to enable running MUMPS in MPI+OpenMP hybrid mode and non-MUMPS in flat-MPI mode.
2954: See details below.
2956: Use `-pc_type cholesky` or `lu` `-pc_factor_mat_solver_type mumps` to use this direct solver
2958: Options Database Keys:
2959: + -mat_mumps_icntl_1 - ICNTL(1): output stream for error messages
2960: . -mat_mumps_icntl_2 - ICNTL(2): output stream for diagnostic printing, statistics, and warning
2961: . -mat_mumps_icntl_3 - ICNTL(3): output stream for global information, collected on the host
2962: . -mat_mumps_icntl_4 - ICNTL(4): level of printing (0 to 4)
2963: . -mat_mumps_icntl_6 - ICNTL(6): permutes to a zero-free diagonal and/or scale the matrix (0 to 7)
2964: . -mat_mumps_icntl_7 - ICNTL(7): computes a symmetric permutation in sequential analysis, 0=AMD, 2=AMF, 3=Scotch, 4=PORD, 5=Metis, 6=QAMD, and 7=auto
2965: Use -pc_factor_mat_ordering_type <type> to have PETSc perform the ordering (sequential only)
2966: . -mat_mumps_icntl_8 - ICNTL(8): scaling strategy (-2 to 8 or 77)
2967: . -mat_mumps_icntl_10 - ICNTL(10): max num of refinements
2968: . -mat_mumps_icntl_11 - ICNTL(11): statistics related to an error analysis (via -ksp_view)
2969: . -mat_mumps_icntl_12 - ICNTL(12): an ordering strategy for symmetric matrices (0 to 3)
2970: . -mat_mumps_icntl_13 - ICNTL(13): parallelism of the root node (enable ScaLAPACK) and its splitting
2971: . -mat_mumps_icntl_14 - ICNTL(14): percentage increase in the estimated working space
2972: . -mat_mumps_icntl_15 - ICNTL(15): compression of the input matrix resulting from a block format
2973: . -mat_mumps_icntl_19 - ICNTL(19): computes the Schur complement
2974: . -mat_mumps_icntl_20 - ICNTL(20): give MUMPS centralized (0) or distributed (10) dense RHS
2975: . -mat_mumps_icntl_22 - ICNTL(22): in-core/out-of-core factorization and solve (0 or 1)
2976: . -mat_mumps_icntl_23 - ICNTL(23): max size of the working memory (MB) that can allocate per processor
2977: . -mat_mumps_icntl_24 - ICNTL(24): detection of null pivot rows (0 or 1)
2978: . -mat_mumps_icntl_25 - ICNTL(25): compute a solution of a deficient matrix and a null space basis
2979: . -mat_mumps_icntl_26 - ICNTL(26): drives the solution phase if a Schur complement matrix
2980: . -mat_mumps_icntl_28 - ICNTL(28): use 1 for sequential analysis and ictnl(7) ordering, or 2 for parallel analysis and ictnl(29) ordering
2981: . -mat_mumps_icntl_29 - ICNTL(29): parallel ordering 1 = ptscotch, 2 = parmetis
2982: . -mat_mumps_icntl_30 - ICNTL(30): compute user-specified set of entries in inv(A)
2983: . -mat_mumps_icntl_31 - ICNTL(31): indicates which factors may be discarded during factorization
2984: . -mat_mumps_icntl_33 - ICNTL(33): compute determinant
2985: . -mat_mumps_icntl_35 - ICNTL(35): level of activation of BLR (Block Low-Rank) feature
2986: . -mat_mumps_icntl_36 - ICNTL(36): controls the choice of BLR factorization variant
2987: . -mat_mumps_icntl_38 - ICNTL(38): sets the estimated compression rate of LU factors with BLR
2988: . -mat_mumps_cntl_1 - CNTL(1): relative pivoting threshold
2989: . -mat_mumps_cntl_2 - CNTL(2): stopping criterion of refinement
2990: . -mat_mumps_cntl_3 - CNTL(3): absolute pivoting threshold
2991: . -mat_mumps_cntl_4 - CNTL(4): value for static pivoting
2992: . -mat_mumps_cntl_5 - CNTL(5): fixation for null pivots
2993: . -mat_mumps_cntl_7 - CNTL(7): precision of the dropping parameter used during BLR factorization
2994: - -mat_mumps_use_omp_threads [m] - run MUMPS in MPI+OpenMP hybrid mode as if omp_set_num_threads(m) is called before calling MUMPS.
2995: Default might be the number of cores per CPU package (socket) as reported by hwloc and suggested by the MUMPS manual.
2997: Level: beginner
2999: Notes:
3000: MUMPS Cholesky does not handle (complex) Hermitian matrices (see User's Guide at https://mumps-solver.org/index.php?page=doc) so using it will
3001: error if the matrix is Hermitian.
3003: When used within a `KSP`/`PC` solve the options are prefixed with that of the `PC`. Otherwise one can set the options prefix by calling
3004: `MatSetOptionsPrefixFactor()` on the matrix from which the factor was obtained or `MatSetOptionsPrefix()` on the factor matrix.
3006: When a MUMPS factorization fails inside a KSP solve, for example with a `KSP_DIVERGED_PC_FAILED`, one can find the MUMPS information about
3007: the failure with
3008: .vb
3009: KSPGetPC(ksp,&pc);
3010: PCFactorGetMatrix(pc,&mat);
3011: MatMumpsGetInfo(mat,....);
3012: MatMumpsGetInfog(mat,....); etc.
3013: .ve
3014: Or run with `-ksp_error_if_not_converged` and the program will be stopped and the information printed in the error message.
3016: MUMPS provides 64-bit integer support in two build modes:
3017: full 64-bit: here MUMPS is built with C preprocessing flag -DINTSIZE64 and Fortran compiler option -i8, -fdefault-integer-8 or equivalent, and
3018: requires all dependent libraries MPI, ScaLAPACK, LAPACK and BLAS built the same way with 64-bit integers (for example ILP64 Intel MKL and MPI).
3020: selective 64-bit: with the default MUMPS build, 64-bit integers have been introduced where needed. In compressed sparse row (CSR) storage of matrices,
3021: MUMPS stores column indices in 32-bit, but row offsets in 64-bit, so you can have a huge number of non-zeros, but must have less than 2^31 rows and
3022: columns. This can lead to significant memory and performance gains with respect to a full 64-bit integer MUMPS version. This requires a regular (32-bit
3023: integer) build of all dependent libraries MPI, ScaLAPACK, LAPACK and BLAS.
3025: With --download-mumps=1, PETSc always build MUMPS in selective 64-bit mode, which can be used by both --with-64-bit-indices=0/1 variants of PETSc.
3027: Two modes to run MUMPS/PETSc with OpenMP
3028: .vb
3029: Set OMP_NUM_THREADS and run with fewer MPI ranks than cores. For example, if you want to have 16 OpenMP
3030: threads per rank, then you may use "export OMP_NUM_THREADS=16 && mpirun -n 4 ./test".
3031: .ve
3033: .vb
3034: -mat_mumps_use_omp_threads [m] and run your code with as many MPI ranks as the number of cores. For example,
3035: if a compute node has 32 cores and you run on two nodes, you may use "mpirun -n 64 ./test -mat_mumps_use_omp_threads 16"
3036: .ve
3038: To run MUMPS in MPI+OpenMP hybrid mode (i.e., enable multithreading in MUMPS), but still run the non-MUMPS part
3039: (i.e., PETSc part) of your code in the so-called flat-MPI (aka pure-MPI) mode, you need to configure PETSc with `--with-openmp` `--download-hwloc`
3040: (or `--with-hwloc`), and have an MPI that supports MPI-3.0's process shared memory (which is usually available). Since MUMPS calls BLAS
3041: libraries, to really get performance, you should have multithreaded BLAS libraries such as Intel MKL, AMD ACML, Cray libSci or OpenBLAS
3042: (PETSc will automatically try to utilized a threaded BLAS if --with-openmp is provided).
3044: If you run your code through a job submission system, there are caveats in MPI rank mapping. We use MPI_Comm_split_type() to obtain MPI
3045: processes on each compute node. Listing the processes in rank ascending order, we split processes on a node into consecutive groups of
3046: size m and create a communicator called omp_comm for each group. Rank 0 in an omp_comm is called the master rank, and others in the omp_comm
3047: are called slave ranks (or slaves). Only master ranks are seen to MUMPS and slaves are not. We will free CPUs assigned to slaves (might be set
3048: by CPU binding policies in job scripts) and make the CPUs available to the master so that OMP threads spawned by MUMPS can run on the CPUs.
3049: In a multi-socket compute node, MPI rank mapping is an issue. Still use the above example and suppose your compute node has two sockets,
3050: if you interleave MPI ranks on the two sockets, in other words, even ranks are placed on socket 0, and odd ranks are on socket 1, and bind
3051: MPI ranks to cores, then with -mat_mumps_use_omp_threads 16, a master rank (and threads it spawns) will use half cores in socket 0, and half
3052: cores in socket 1, that definitely hurts locality. On the other hand, if you map MPI ranks consecutively on the two sockets, then the
3053: problem will not happen. Therefore, when you use -mat_mumps_use_omp_threads, you need to keep an eye on your MPI rank mapping and CPU binding.
3054: For example, with the Slurm job scheduler, one can use srun --cpu-bind=verbose -m block:block to map consecutive MPI ranks to sockets and
3055: examine the mapping result.
3057: PETSc does not control thread binding in MUMPS. So to get best performance, one still has to set `OMP_PROC_BIND` and `OMP_PLACES` in job scripts,
3058: for example, export `OMP_PLACES`=threads and export `OMP_PROC_BIND`=spread. One does not need to export `OMP_NUM_THREADS`=m in job scripts as PETSc
3059: calls `omp_set_num_threads`(m) internally before calling MUMPS.
3061: References:
3062: + * - Heroux, Michael A., R. Brightwell, and Michael M. Wolf. "Bi-modal MPI and MPI+ threads computing on scalable multicore systems." IJHPCA (Submitted) (2011).
3063: - * - Gutierrez, Samuel K., et al. "Accommodating Thread-Level Heterogeneity in Coupled Parallel Applications." Parallel and Distributed Processing Symposium (IPDPS), 2017 IEEE International. IEEE, 2017.
3065: .seealso: [](chapter_matrices), `Mat`, `PCFactorSetMatSolverType()`, `MatSolverType`, `MatMumpsSetIcntl()`, `MatMumpsGetIcntl()`, `MatMumpsSetCntl()`, `MatMumpsGetCntl()`, `MatMumpsGetInfo()`, `MatMumpsGetInfog()`, `MatMumpsGetRinfo()`, `MatMumpsGetRinfog()`, `KSPGetPC()`, `PCFactorGetMatrix()`
3066: M*/
3068: static PetscErrorCode MatFactorGetSolverType_mumps(Mat A, MatSolverType *type)
3069: {
3070: PetscFunctionBegin;
3071: *type = MATSOLVERMUMPS;
3072: PetscFunctionReturn(PETSC_SUCCESS);
3073: }
3075: /* MatGetFactor for Seq and MPI AIJ matrices */
3076: static PetscErrorCode MatGetFactor_aij_mumps(Mat A, MatFactorType ftype, Mat *F)
3077: {
3078: Mat B;
3079: Mat_MUMPS *mumps;
3080: PetscBool isSeqAIJ;
3081: PetscMPIInt size;
3083: PetscFunctionBegin;
3084: #if defined(PETSC_USE_COMPLEX)
3085: PetscCheck(A->hermitian != PETSC_BOOL3_TRUE || A->symmetric == PETSC_BOOL3_TRUE || ftype != MAT_FACTOR_CHOLESKY, PETSC_COMM_SELF, PETSC_ERR_SUP, "Hermitian CHOLESKY Factor is not supported");
3086: #endif
3087: /* Create the factorization matrix */
3088: PetscCall(PetscObjectBaseTypeCompare((PetscObject)A, MATSEQAIJ, &isSeqAIJ));
3089: PetscCall(MatCreate(PetscObjectComm((PetscObject)A), &B));
3090: PetscCall(MatSetSizes(B, A->rmap->n, A->cmap->n, A->rmap->N, A->cmap->N));
3091: PetscCall(PetscStrallocpy("mumps", &((PetscObject)B)->type_name));
3092: PetscCall(MatSetUp(B));
3094: PetscCall(PetscNew(&mumps));
3096: B->ops->view = MatView_MUMPS;
3097: B->ops->getinfo = MatGetInfo_MUMPS;
3099: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorGetSolverType_C", MatFactorGetSolverType_mumps));
3100: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorSetSchurIS_C", MatFactorSetSchurIS_MUMPS));
3101: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorCreateSchurComplement_C", MatFactorCreateSchurComplement_MUMPS));
3102: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetIcntl_C", MatMumpsSetIcntl_MUMPS));
3103: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetIcntl_C", MatMumpsGetIcntl_MUMPS));
3104: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetCntl_C", MatMumpsSetCntl_MUMPS));
3105: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetCntl_C", MatMumpsGetCntl_MUMPS));
3106: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfo_C", MatMumpsGetInfo_MUMPS));
3107: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfog_C", MatMumpsGetInfog_MUMPS));
3108: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfo_C", MatMumpsGetRinfo_MUMPS));
3109: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfog_C", MatMumpsGetRinfog_MUMPS));
3110: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetNullPivots_C", MatMumpsGetNullPivots_MUMPS));
3111: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverse_C", MatMumpsGetInverse_MUMPS));
3112: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverseTranspose_C", MatMumpsGetInverseTranspose_MUMPS));
3114: if (ftype == MAT_FACTOR_LU) {
3115: B->ops->lufactorsymbolic = MatLUFactorSymbolic_AIJMUMPS;
3116: B->factortype = MAT_FACTOR_LU;
3117: if (isSeqAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqaij_seqaij;
3118: else mumps->ConvertToTriples = MatConvertToTriples_mpiaij_mpiaij;
3119: PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL, (char **)&B->preferredordering[MAT_FACTOR_LU]));
3120: mumps->sym = 0;
3121: } else {
3122: B->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_MUMPS;
3123: B->factortype = MAT_FACTOR_CHOLESKY;
3124: if (isSeqAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqaij_seqsbaij;
3125: else mumps->ConvertToTriples = MatConvertToTriples_mpiaij_mpisbaij;
3126: PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL, (char **)&B->preferredordering[MAT_FACTOR_CHOLESKY]));
3127: #if defined(PETSC_USE_COMPLEX)
3128: mumps->sym = 2;
3129: #else
3130: if (A->spd == PETSC_BOOL3_TRUE) mumps->sym = 1;
3131: else mumps->sym = 2;
3132: #endif
3133: }
3135: /* set solvertype */
3136: PetscCall(PetscFree(B->solvertype));
3137: PetscCall(PetscStrallocpy(MATSOLVERMUMPS, &B->solvertype));
3138: PetscCallMPI(MPI_Comm_size(PetscObjectComm((PetscObject)A), &size));
3139: if (size == 1) {
3140: /* MUMPS option -mat_mumps_icntl_7 1 is automatically set if PETSc ordering is passed into symbolic factorization */
3141: B->canuseordering = PETSC_TRUE;
3142: }
3143: B->ops->destroy = MatDestroy_MUMPS;
3144: B->data = (void *)mumps;
3146: *F = B;
3147: mumps->id.job = JOB_NULL;
3148: mumps->ICNTL_pre = NULL;
3149: mumps->CNTL_pre = NULL;
3150: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
3151: PetscFunctionReturn(PETSC_SUCCESS);
3152: }
3154: /* MatGetFactor for Seq and MPI SBAIJ matrices */
3155: static PetscErrorCode MatGetFactor_sbaij_mumps(Mat A, MatFactorType ftype, Mat *F)
3156: {
3157: Mat B;
3158: Mat_MUMPS *mumps;
3159: PetscBool isSeqSBAIJ;
3160: PetscMPIInt size;
3162: PetscFunctionBegin;
3163: #if defined(PETSC_USE_COMPLEX)
3164: PetscCheck(A->hermitian != PETSC_BOOL3_TRUE || A->symmetric == PETSC_BOOL3_TRUE, PETSC_COMM_SELF, PETSC_ERR_SUP, "Hermitian CHOLESKY Factor is not supported");
3165: #endif
3166: PetscCall(MatCreate(PetscObjectComm((PetscObject)A), &B));
3167: PetscCall(MatSetSizes(B, A->rmap->n, A->cmap->n, A->rmap->N, A->cmap->N));
3168: PetscCall(PetscStrallocpy("mumps", &((PetscObject)B)->type_name));
3169: PetscCall(MatSetUp(B));
3171: PetscCall(PetscNew(&mumps));
3172: PetscCall(PetscObjectTypeCompare((PetscObject)A, MATSEQSBAIJ, &isSeqSBAIJ));
3173: if (isSeqSBAIJ) {
3174: mumps->ConvertToTriples = MatConvertToTriples_seqsbaij_seqsbaij;
3175: } else {
3176: mumps->ConvertToTriples = MatConvertToTriples_mpisbaij_mpisbaij;
3177: }
3179: B->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_MUMPS;
3180: B->ops->view = MatView_MUMPS;
3181: B->ops->getinfo = MatGetInfo_MUMPS;
3183: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorGetSolverType_C", MatFactorGetSolverType_mumps));
3184: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorSetSchurIS_C", MatFactorSetSchurIS_MUMPS));
3185: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorCreateSchurComplement_C", MatFactorCreateSchurComplement_MUMPS));
3186: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetIcntl_C", MatMumpsSetIcntl_MUMPS));
3187: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetIcntl_C", MatMumpsGetIcntl_MUMPS));
3188: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetCntl_C", MatMumpsSetCntl_MUMPS));
3189: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetCntl_C", MatMumpsGetCntl_MUMPS));
3190: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfo_C", MatMumpsGetInfo_MUMPS));
3191: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfog_C", MatMumpsGetInfog_MUMPS));
3192: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfo_C", MatMumpsGetRinfo_MUMPS));
3193: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfog_C", MatMumpsGetRinfog_MUMPS));
3194: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetNullPivots_C", MatMumpsGetNullPivots_MUMPS));
3195: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverse_C", MatMumpsGetInverse_MUMPS));
3196: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverseTranspose_C", MatMumpsGetInverseTranspose_MUMPS));
3198: B->factortype = MAT_FACTOR_CHOLESKY;
3199: #if defined(PETSC_USE_COMPLEX)
3200: mumps->sym = 2;
3201: #else
3202: if (A->spd == PETSC_BOOL3_TRUE) mumps->sym = 1;
3203: else mumps->sym = 2;
3204: #endif
3206: /* set solvertype */
3207: PetscCall(PetscFree(B->solvertype));
3208: PetscCall(PetscStrallocpy(MATSOLVERMUMPS, &B->solvertype));
3209: PetscCallMPI(MPI_Comm_size(PetscObjectComm((PetscObject)A), &size));
3210: if (size == 1) {
3211: /* MUMPS option -mat_mumps_icntl_7 1 is automatically set if PETSc ordering is passed into symbolic factorization */
3212: B->canuseordering = PETSC_TRUE;
3213: }
3214: PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL, (char **)&B->preferredordering[MAT_FACTOR_CHOLESKY]));
3215: B->ops->destroy = MatDestroy_MUMPS;
3216: B->data = (void *)mumps;
3218: *F = B;
3219: mumps->id.job = JOB_NULL;
3220: mumps->ICNTL_pre = NULL;
3221: mumps->CNTL_pre = NULL;
3222: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
3223: PetscFunctionReturn(PETSC_SUCCESS);
3224: }
3226: static PetscErrorCode MatGetFactor_baij_mumps(Mat A, MatFactorType ftype, Mat *F)
3227: {
3228: Mat B;
3229: Mat_MUMPS *mumps;
3230: PetscBool isSeqBAIJ;
3231: PetscMPIInt size;
3233: PetscFunctionBegin;
3234: /* Create the factorization matrix */
3235: PetscCall(PetscObjectTypeCompare((PetscObject)A, MATSEQBAIJ, &isSeqBAIJ));
3236: PetscCall(MatCreate(PetscObjectComm((PetscObject)A), &B));
3237: PetscCall(MatSetSizes(B, A->rmap->n, A->cmap->n, A->rmap->N, A->cmap->N));
3238: PetscCall(PetscStrallocpy("mumps", &((PetscObject)B)->type_name));
3239: PetscCall(MatSetUp(B));
3241: PetscCall(PetscNew(&mumps));
3242: if (ftype == MAT_FACTOR_LU) {
3243: B->ops->lufactorsymbolic = MatLUFactorSymbolic_BAIJMUMPS;
3244: B->factortype = MAT_FACTOR_LU;
3245: if (isSeqBAIJ) mumps->ConvertToTriples = MatConvertToTriples_seqbaij_seqaij;
3246: else mumps->ConvertToTriples = MatConvertToTriples_mpibaij_mpiaij;
3247: mumps->sym = 0;
3248: PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL, (char **)&B->preferredordering[MAT_FACTOR_LU]));
3249: } else SETERRQ(PETSC_COMM_SELF, PETSC_ERR_SUP, "Cannot use PETSc BAIJ matrices with MUMPS Cholesky, use SBAIJ or AIJ matrix instead");
3251: B->ops->view = MatView_MUMPS;
3252: B->ops->getinfo = MatGetInfo_MUMPS;
3254: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorGetSolverType_C", MatFactorGetSolverType_mumps));
3255: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorSetSchurIS_C", MatFactorSetSchurIS_MUMPS));
3256: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorCreateSchurComplement_C", MatFactorCreateSchurComplement_MUMPS));
3257: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetIcntl_C", MatMumpsSetIcntl_MUMPS));
3258: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetIcntl_C", MatMumpsGetIcntl_MUMPS));
3259: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetCntl_C", MatMumpsSetCntl_MUMPS));
3260: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetCntl_C", MatMumpsGetCntl_MUMPS));
3261: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfo_C", MatMumpsGetInfo_MUMPS));
3262: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfog_C", MatMumpsGetInfog_MUMPS));
3263: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfo_C", MatMumpsGetRinfo_MUMPS));
3264: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfog_C", MatMumpsGetRinfog_MUMPS));
3265: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetNullPivots_C", MatMumpsGetNullPivots_MUMPS));
3266: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverse_C", MatMumpsGetInverse_MUMPS));
3267: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInverseTranspose_C", MatMumpsGetInverseTranspose_MUMPS));
3269: /* set solvertype */
3270: PetscCall(PetscFree(B->solvertype));
3271: PetscCall(PetscStrallocpy(MATSOLVERMUMPS, &B->solvertype));
3272: PetscCallMPI(MPI_Comm_size(PetscObjectComm((PetscObject)A), &size));
3273: if (size == 1) {
3274: /* MUMPS option -mat_mumps_icntl_7 1 is automatically set if PETSc ordering is passed into symbolic factorization */
3275: B->canuseordering = PETSC_TRUE;
3276: }
3277: B->ops->destroy = MatDestroy_MUMPS;
3278: B->data = (void *)mumps;
3280: *F = B;
3281: mumps->id.job = JOB_NULL;
3282: mumps->ICNTL_pre = NULL;
3283: mumps->CNTL_pre = NULL;
3284: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
3285: PetscFunctionReturn(PETSC_SUCCESS);
3286: }
3288: /* MatGetFactor for Seq and MPI SELL matrices */
3289: static PetscErrorCode MatGetFactor_sell_mumps(Mat A, MatFactorType ftype, Mat *F)
3290: {
3291: Mat B;
3292: Mat_MUMPS *mumps;
3293: PetscBool isSeqSELL;
3294: PetscMPIInt size;
3296: PetscFunctionBegin;
3297: /* Create the factorization matrix */
3298: PetscCall(PetscObjectTypeCompare((PetscObject)A, MATSEQSELL, &isSeqSELL));
3299: PetscCall(MatCreate(PetscObjectComm((PetscObject)A), &B));
3300: PetscCall(MatSetSizes(B, A->rmap->n, A->cmap->n, A->rmap->N, A->cmap->N));
3301: PetscCall(PetscStrallocpy("mumps", &((PetscObject)B)->type_name));
3302: PetscCall(MatSetUp(B));
3304: PetscCall(PetscNew(&mumps));
3306: B->ops->view = MatView_MUMPS;
3307: B->ops->getinfo = MatGetInfo_MUMPS;
3309: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorGetSolverType_C", MatFactorGetSolverType_mumps));
3310: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorSetSchurIS_C", MatFactorSetSchurIS_MUMPS));
3311: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatFactorCreateSchurComplement_C", MatFactorCreateSchurComplement_MUMPS));
3312: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetIcntl_C", MatMumpsSetIcntl_MUMPS));
3313: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetIcntl_C", MatMumpsGetIcntl_MUMPS));
3314: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsSetCntl_C", MatMumpsSetCntl_MUMPS));
3315: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetCntl_C", MatMumpsGetCntl_MUMPS));
3316: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfo_C", MatMumpsGetInfo_MUMPS));
3317: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetInfog_C", MatMumpsGetInfog_MUMPS));
3318: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfo_C", MatMumpsGetRinfo_MUMPS));
3319: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetRinfog_C", MatMumpsGetRinfog_MUMPS));
3320: PetscCall(PetscObjectComposeFunction((PetscObject)B, "MatMumpsGetNullPivots_C", MatMumpsGetNullPivots_MUMPS));
3322: if (ftype == MAT_FACTOR_LU) {
3323: B->ops->lufactorsymbolic = MatLUFactorSymbolic_AIJMUMPS;
3324: B->factortype = MAT_FACTOR_LU;
3325: if (isSeqSELL) mumps->ConvertToTriples = MatConvertToTriples_seqsell_seqaij;
3326: else SETERRQ(PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "To be implemented");
3327: mumps->sym = 0;
3328: PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL, (char **)&B->preferredordering[MAT_FACTOR_LU]));
3329: } else SETERRQ(PetscObjectComm((PetscObject)A), PETSC_ERR_SUP, "To be implemented");
3331: /* set solvertype */
3332: PetscCall(PetscFree(B->solvertype));
3333: PetscCall(PetscStrallocpy(MATSOLVERMUMPS, &B->solvertype));
3334: PetscCallMPI(MPI_Comm_size(PetscObjectComm((PetscObject)A), &size));
3335: if (size == 1) {
3336: /* MUMPS option -mat_mumps_icntl_7 1 is automatically set if PETSc ordering is passed into symbolic factorization */
3337: B->canuseordering = PETSC_TRUE;
3338: }
3339: B->ops->destroy = MatDestroy_MUMPS;
3340: B->data = (void *)mumps;
3342: *F = B;
3343: mumps->id.job = JOB_NULL;
3344: mumps->ICNTL_pre = NULL;
3345: mumps->CNTL_pre = NULL;
3346: mumps->matstruc = DIFFERENT_NONZERO_PATTERN;
3347: PetscFunctionReturn(PETSC_SUCCESS);
3348: }
3350: PETSC_EXTERN PetscErrorCode MatSolverTypeRegister_MUMPS(void)
3351: {
3352: PetscFunctionBegin;
3353: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATMPIAIJ, MAT_FACTOR_LU, MatGetFactor_aij_mumps));
3354: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATMPIAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_aij_mumps));
3355: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATMPIBAIJ, MAT_FACTOR_LU, MatGetFactor_baij_mumps));
3356: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATMPIBAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_baij_mumps));
3357: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATMPISBAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_sbaij_mumps));
3358: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQAIJ, MAT_FACTOR_LU, MatGetFactor_aij_mumps));
3359: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_aij_mumps));
3360: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQBAIJ, MAT_FACTOR_LU, MatGetFactor_baij_mumps));
3361: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQBAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_baij_mumps));
3362: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQSBAIJ, MAT_FACTOR_CHOLESKY, MatGetFactor_sbaij_mumps));
3363: PetscCall(MatSolverTypeRegister(MATSOLVERMUMPS, MATSEQSELL, MAT_FACTOR_LU, MatGetFactor_sell_mumps));
3364: PetscFunctionReturn(PETSC_SUCCESS);
3365: }