Actual source code: matrix.c
petsc-master 2020-10-23
1: /*
2: This is where the abstract matrix operations are defined
3: */
5: #include <petsc/private/matimpl.h>
6: #include <petsc/private/isimpl.h>
7: #include <petsc/private/vecimpl.h>
9: /* Logging support */
10: PetscClassId MAT_CLASSID;
11: PetscClassId MAT_COLORING_CLASSID;
12: PetscClassId MAT_FDCOLORING_CLASSID;
13: PetscClassId MAT_TRANSPOSECOLORING_CLASSID;
15: PetscLogEvent MAT_Mult, MAT_Mults, MAT_MultConstrained, MAT_MultAdd, MAT_MultTranspose;
16: PetscLogEvent MAT_MultTransposeConstrained, MAT_MultTransposeAdd, MAT_Solve, MAT_Solves, MAT_SolveAdd, MAT_SolveTranspose, MAT_MatSolve,MAT_MatTrSolve;
17: PetscLogEvent MAT_SolveTransposeAdd, MAT_SOR, MAT_ForwardSolve, MAT_BackwardSolve, MAT_LUFactor, MAT_LUFactorSymbolic;
18: PetscLogEvent MAT_LUFactorNumeric, MAT_CholeskyFactor, MAT_CholeskyFactorSymbolic, MAT_CholeskyFactorNumeric, MAT_ILUFactor;
19: PetscLogEvent MAT_ILUFactorSymbolic, MAT_ICCFactorSymbolic, MAT_Copy, MAT_Convert, MAT_Scale, MAT_AssemblyBegin;
20: PetscLogEvent MAT_AssemblyEnd, MAT_SetValues, MAT_GetValues, MAT_GetRow, MAT_GetRowIJ, MAT_CreateSubMats, MAT_GetOrdering, MAT_RedundantMat, MAT_GetSeqNonzeroStructure;
21: PetscLogEvent MAT_IncreaseOverlap, MAT_Partitioning, MAT_PartitioningND, MAT_Coarsen, MAT_ZeroEntries, MAT_Load, MAT_View, MAT_AXPY, MAT_FDColoringCreate;
22: PetscLogEvent MAT_FDColoringSetUp, MAT_FDColoringApply,MAT_Transpose,MAT_FDColoringFunction, MAT_CreateSubMat;
23: PetscLogEvent MAT_TransposeColoringCreate;
24: PetscLogEvent MAT_MatMult, MAT_MatMultSymbolic, MAT_MatMultNumeric;
25: PetscLogEvent MAT_PtAP, MAT_PtAPSymbolic, MAT_PtAPNumeric,MAT_RARt, MAT_RARtSymbolic, MAT_RARtNumeric;
26: PetscLogEvent MAT_MatTransposeMult, MAT_MatTransposeMultSymbolic, MAT_MatTransposeMultNumeric;
27: PetscLogEvent MAT_TransposeMatMult, MAT_TransposeMatMultSymbolic, MAT_TransposeMatMultNumeric;
28: PetscLogEvent MAT_MatMatMult, MAT_MatMatMultSymbolic, MAT_MatMatMultNumeric;
29: PetscLogEvent MAT_MultHermitianTranspose,MAT_MultHermitianTransposeAdd;
30: PetscLogEvent MAT_Getsymtranspose, MAT_Getsymtransreduced, MAT_GetBrowsOfAcols;
31: PetscLogEvent MAT_GetBrowsOfAocols, MAT_Getlocalmat, MAT_Getlocalmatcondensed, MAT_Seqstompi, MAT_Seqstompinum, MAT_Seqstompisym;
32: PetscLogEvent MAT_Applypapt, MAT_Applypapt_numeric, MAT_Applypapt_symbolic, MAT_GetSequentialNonzeroStructure;
33: PetscLogEvent MAT_GetMultiProcBlock;
34: PetscLogEvent MAT_CUSPARSECopyToGPU, MAT_CUSPARSEGenerateTranspose, MAT_SetValuesBatch;
35: PetscLogEvent MAT_ViennaCLCopyToGPU;
36: PetscLogEvent MAT_DenseCopyToGPU, MAT_DenseCopyFromGPU;
37: PetscLogEvent MAT_Merge,MAT_Residual,MAT_SetRandom;
38: PetscLogEvent MAT_FactorFactS,MAT_FactorInvS;
39: PetscLogEvent MATCOLORING_Apply,MATCOLORING_Comm,MATCOLORING_Local,MATCOLORING_ISCreate,MATCOLORING_SetUp,MATCOLORING_Weights;
41: const char *const MatFactorTypes[] = {"NONE","LU","CHOLESKY","ILU","ICC","ILUDT","MatFactorType","MAT_FACTOR_",NULL};
43: /*@
44: MatSetRandom - Sets all components of a matrix to random numbers. For sparse matrices that have been preallocated but not been assembled it randomly selects appropriate locations,
45: for sparse matrices that already have locations it fills the locations with random numbers
47: Logically Collective on Mat
49: Input Parameters:
50: + x - the matrix
51: - rctx - the random number context, formed by PetscRandomCreate(), or NULL and
52: it will create one internally.
54: Output Parameter:
55: . x - the matrix
57: Example of Usage:
58: .vb
59: PetscRandomCreate(PETSC_COMM_WORLD,&rctx);
60: MatSetRandom(x,rctx);
61: PetscRandomDestroy(rctx);
62: .ve
64: Level: intermediate
67: .seealso: MatZeroEntries(), MatSetValues(), PetscRandomCreate(), PetscRandomDestroy()
68: @*/
69: PetscErrorCode MatSetRandom(Mat x,PetscRandom rctx)
70: {
72: PetscRandom randObj = NULL;
79: if (!x->ops->setrandom) SETERRQ1(PetscObjectComm((PetscObject)x),PETSC_ERR_SUP,"Mat type %s",((PetscObject)x)->type_name);
81: if (!rctx) {
82: MPI_Comm comm;
83: PetscObjectGetComm((PetscObject)x,&comm);
84: PetscRandomCreate(comm,&randObj);
85: PetscRandomSetFromOptions(randObj);
86: rctx = randObj;
87: }
89: PetscLogEventBegin(MAT_SetRandom,x,rctx,0,0);
90: (*x->ops->setrandom)(x,rctx);
91: PetscLogEventEnd(MAT_SetRandom,x,rctx,0,0);
93: MatAssemblyBegin(x, MAT_FINAL_ASSEMBLY);
94: MatAssemblyEnd(x, MAT_FINAL_ASSEMBLY);
95: PetscRandomDestroy(&randObj);
96: return(0);
97: }
99: /*@
100: MatFactorGetErrorZeroPivot - returns the pivot value that was determined to be zero and the row it occurred in
102: Logically Collective on Mat
104: Input Parameters:
105: . mat - the factored matrix
107: Output Parameter:
108: + pivot - the pivot value computed
109: - row - the row that the zero pivot occurred. Note that this row must be interpreted carefully due to row reorderings and which processes
110: the share the matrix
112: Level: advanced
114: Notes:
115: This routine does not work for factorizations done with external packages.
117: This routine should only be called if MatGetFactorError() returns a value of MAT_FACTOR_NUMERIC_ZEROPIVOT
119: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
121: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatLUFactorSymbolic(), MatCholeskyFactorSymbolic(), MatFactorClearError(), MatFactorGetErrorZeroPivot()
122: @*/
123: PetscErrorCode MatFactorGetErrorZeroPivot(Mat mat,PetscReal *pivot,PetscInt *row)
124: {
127: *pivot = mat->factorerror_zeropivot_value;
128: *row = mat->factorerror_zeropivot_row;
129: return(0);
130: }
132: /*@
133: MatFactorGetError - gets the error code from a factorization
135: Logically Collective on Mat
137: Input Parameters:
138: . mat - the factored matrix
140: Output Parameter:
141: . err - the error code
143: Level: advanced
145: Notes:
146: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
148: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatLUFactorSymbolic(), MatCholeskyFactorSymbolic(), MatFactorClearError(), MatFactorGetErrorZeroPivot()
149: @*/
150: PetscErrorCode MatFactorGetError(Mat mat,MatFactorError *err)
151: {
154: *err = mat->factorerrortype;
155: return(0);
156: }
158: /*@
159: MatFactorClearError - clears the error code in a factorization
161: Logically Collective on Mat
163: Input Parameter:
164: . mat - the factored matrix
166: Level: developer
168: Notes:
169: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
171: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatLUFactorSymbolic(), MatCholeskyFactorSymbolic(), MatFactorGetError(), MatFactorGetErrorZeroPivot()
172: @*/
173: PetscErrorCode MatFactorClearError(Mat mat)
174: {
177: mat->factorerrortype = MAT_FACTOR_NOERROR;
178: mat->factorerror_zeropivot_value = 0.0;
179: mat->factorerror_zeropivot_row = 0;
180: return(0);
181: }
183: PETSC_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat mat,PetscBool cols,PetscReal tol,IS *nonzero)
184: {
185: PetscErrorCode ierr;
186: Vec r,l;
187: const PetscScalar *al;
188: PetscInt i,nz,gnz,N,n;
191: MatCreateVecs(mat,&r,&l);
192: if (!cols) { /* nonzero rows */
193: MatGetSize(mat,&N,NULL);
194: MatGetLocalSize(mat,&n,NULL);
195: VecSet(l,0.0);
196: VecSetRandom(r,NULL);
197: MatMult(mat,r,l);
198: VecGetArrayRead(l,&al);
199: } else { /* nonzero columns */
200: MatGetSize(mat,NULL,&N);
201: MatGetLocalSize(mat,NULL,&n);
202: VecSet(r,0.0);
203: VecSetRandom(l,NULL);
204: MatMultTranspose(mat,l,r);
205: VecGetArrayRead(r,&al);
206: }
207: if (tol <= 0.0) { for (i=0,nz=0;i<n;i++) if (al[i] != 0.0) nz++; }
208: else { for (i=0,nz=0;i<n;i++) if (PetscAbsScalar(al[i]) > tol) nz++; }
209: MPIU_Allreduce(&nz,&gnz,1,MPIU_INT,MPI_SUM,PetscObjectComm((PetscObject)mat));
210: if (gnz != N) {
211: PetscInt *nzr;
212: PetscMalloc1(nz,&nzr);
213: if (nz) {
214: if (tol < 0) { for (i=0,nz=0;i<n;i++) if (al[i] != 0.0) nzr[nz++] = i; }
215: else { for (i=0,nz=0;i<n;i++) if (PetscAbsScalar(al[i]) > tol) nzr[nz++] = i; }
216: }
217: ISCreateGeneral(PetscObjectComm((PetscObject)mat),nz,nzr,PETSC_OWN_POINTER,nonzero);
218: } else *nonzero = NULL;
219: if (!cols) { /* nonzero rows */
220: VecRestoreArrayRead(l,&al);
221: } else {
222: VecRestoreArrayRead(r,&al);
223: }
224: VecDestroy(&l);
225: VecDestroy(&r);
226: return(0);
227: }
229: /*@
230: MatFindNonzeroRows - Locate all rows that are not completely zero in the matrix
232: Input Parameter:
233: . A - the matrix
235: Output Parameter:
236: . keptrows - the rows that are not completely zero
238: Notes:
239: keptrows is set to NULL if all rows are nonzero.
241: Level: intermediate
243: @*/
244: PetscErrorCode MatFindNonzeroRows(Mat mat,IS *keptrows)
245: {
252: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
253: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
254: if (!mat->ops->findnonzerorows) {
255: MatFindNonzeroRowsOrCols_Basic(mat,PETSC_FALSE,0.0,keptrows);
256: } else {
257: (*mat->ops->findnonzerorows)(mat,keptrows);
258: }
259: return(0);
260: }
262: /*@
263: MatFindZeroRows - Locate all rows that are completely zero in the matrix
265: Input Parameter:
266: . A - the matrix
268: Output Parameter:
269: . zerorows - the rows that are completely zero
271: Notes:
272: zerorows is set to NULL if no rows are zero.
274: Level: intermediate
276: @*/
277: PetscErrorCode MatFindZeroRows(Mat mat,IS *zerorows)
278: {
280: IS keptrows;
281: PetscInt m, n;
286: MatFindNonzeroRows(mat, &keptrows);
287: /* MatFindNonzeroRows sets keptrows to NULL if there are no zero rows.
288: In keeping with this convention, we set zerorows to NULL if there are no zero
289: rows. */
290: if (keptrows == NULL) {
291: *zerorows = NULL;
292: } else {
293: MatGetOwnershipRange(mat,&m,&n);
294: ISComplement(keptrows,m,n,zerorows);
295: ISDestroy(&keptrows);
296: }
297: return(0);
298: }
300: /*@
301: MatGetDiagonalBlock - Returns the part of the matrix associated with the on-process coupling
303: Not Collective
305: Input Parameters:
306: . A - the matrix
308: Output Parameters:
309: . a - the diagonal part (which is a SEQUENTIAL matrix)
311: Notes:
312: see the manual page for MatCreateAIJ() for more information on the "diagonal part" of the matrix.
313: Use caution, as the reference count on the returned matrix is not incremented and it is used as
314: part of the containing MPI Mat's normal operation.
316: Level: advanced
318: @*/
319: PetscErrorCode MatGetDiagonalBlock(Mat A,Mat *a)
320: {
327: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
328: if (!A->ops->getdiagonalblock) {
329: PetscMPIInt size;
330: MPI_Comm_size(PetscObjectComm((PetscObject)A),&size);
331: if (size == 1) {
332: *a = A;
333: return(0);
334: } else SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Not coded for matrix type %s",((PetscObject)A)->type_name);
335: }
336: (*A->ops->getdiagonalblock)(A,a);
337: return(0);
338: }
340: /*@
341: MatGetTrace - Gets the trace of a matrix. The sum of the diagonal entries.
343: Collective on Mat
345: Input Parameters:
346: . mat - the matrix
348: Output Parameter:
349: . trace - the sum of the diagonal entries
351: Level: advanced
353: @*/
354: PetscErrorCode MatGetTrace(Mat mat,PetscScalar *trace)
355: {
357: Vec diag;
360: MatCreateVecs(mat,&diag,NULL);
361: MatGetDiagonal(mat,diag);
362: VecSum(diag,trace);
363: VecDestroy(&diag);
364: return(0);
365: }
367: /*@
368: MatRealPart - Zeros out the imaginary part of the matrix
370: Logically Collective on Mat
372: Input Parameters:
373: . mat - the matrix
375: Level: advanced
378: .seealso: MatImaginaryPart()
379: @*/
380: PetscErrorCode MatRealPart(Mat mat)
381: {
387: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
388: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
389: if (!mat->ops->realpart) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
390: MatCheckPreallocated(mat,1);
391: (*mat->ops->realpart)(mat);
392: return(0);
393: }
395: /*@C
396: MatGetGhosts - Get the global index of all ghost nodes defined by the sparse matrix
398: Collective on Mat
400: Input Parameter:
401: . mat - the matrix
403: Output Parameters:
404: + nghosts - number of ghosts (note for BAIJ matrices there is one ghost for each block)
405: - ghosts - the global indices of the ghost points
407: Notes:
408: the nghosts and ghosts are suitable to pass into VecCreateGhost()
410: Level: advanced
412: @*/
413: PetscErrorCode MatGetGhosts(Mat mat,PetscInt *nghosts,const PetscInt *ghosts[])
414: {
420: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
421: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
422: if (!mat->ops->getghosts) {
423: if (nghosts) *nghosts = 0;
424: if (ghosts) *ghosts = NULL;
425: } else {
426: (*mat->ops->getghosts)(mat,nghosts,ghosts);
427: }
428: return(0);
429: }
432: /*@
433: MatImaginaryPart - Moves the imaginary part of the matrix to the real part and zeros the imaginary part
435: Logically Collective on Mat
437: Input Parameters:
438: . mat - the matrix
440: Level: advanced
443: .seealso: MatRealPart()
444: @*/
445: PetscErrorCode MatImaginaryPart(Mat mat)
446: {
452: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
453: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
454: if (!mat->ops->imaginarypart) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
455: MatCheckPreallocated(mat,1);
456: (*mat->ops->imaginarypart)(mat);
457: return(0);
458: }
460: /*@
461: MatMissingDiagonal - Determine if sparse matrix is missing a diagonal entry (or block entry for BAIJ matrices)
463: Not Collective
465: Input Parameter:
466: . mat - the matrix
468: Output Parameters:
469: + missing - is any diagonal missing
470: - dd - first diagonal entry that is missing (optional) on this process
472: Level: advanced
475: .seealso: MatRealPart()
476: @*/
477: PetscErrorCode MatMissingDiagonal(Mat mat,PetscBool *missing,PetscInt *dd)
478: {
485: if (!mat->assembled) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix %s",((PetscObject)mat)->type_name);
486: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
487: if (!mat->ops->missingdiagonal) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
488: (*mat->ops->missingdiagonal)(mat,missing,dd);
489: return(0);
490: }
492: /*@C
493: MatGetRow - Gets a row of a matrix. You MUST call MatRestoreRow()
494: for each row that you get to ensure that your application does
495: not bleed memory.
497: Not Collective
499: Input Parameters:
500: + mat - the matrix
501: - row - the row to get
503: Output Parameters:
504: + ncols - if not NULL, the number of nonzeros in the row
505: . cols - if not NULL, the column numbers
506: - vals - if not NULL, the values
508: Notes:
509: This routine is provided for people who need to have direct access
510: to the structure of a matrix. We hope that we provide enough
511: high-level matrix routines that few users will need it.
513: MatGetRow() always returns 0-based column indices, regardless of
514: whether the internal representation is 0-based (default) or 1-based.
516: For better efficiency, set cols and/or vals to NULL if you do
517: not wish to extract these quantities.
519: The user can only examine the values extracted with MatGetRow();
520: the values cannot be altered. To change the matrix entries, one
521: must use MatSetValues().
523: You can only have one call to MatGetRow() outstanding for a particular
524: matrix at a time, per processor. MatGetRow() can only obtain rows
525: associated with the given processor, it cannot get rows from the
526: other processors; for that we suggest using MatCreateSubMatrices(), then
527: MatGetRow() on the submatrix. The row index passed to MatGetRow()
528: is in the global number of rows.
530: Fortran Notes:
531: The calling sequence from Fortran is
532: .vb
533: MatGetRow(matrix,row,ncols,cols,values,ierr)
534: Mat matrix (input)
535: integer row (input)
536: integer ncols (output)
537: integer cols(maxcols) (output)
538: double precision (or double complex) values(maxcols) output
539: .ve
540: where maxcols >= maximum nonzeros in any row of the matrix.
543: Caution:
544: Do not try to change the contents of the output arrays (cols and vals).
545: In some cases, this may corrupt the matrix.
547: Level: advanced
549: .seealso: MatRestoreRow(), MatSetValues(), MatGetValues(), MatCreateSubMatrices(), MatGetDiagonal()
550: @*/
551: PetscErrorCode MatGetRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[])
552: {
554: PetscInt incols;
559: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
560: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
561: if (!mat->ops->getrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
562: MatCheckPreallocated(mat,1);
563: PetscLogEventBegin(MAT_GetRow,mat,0,0,0);
564: (*mat->ops->getrow)(mat,row,&incols,(PetscInt**)cols,(PetscScalar**)vals);
565: if (ncols) *ncols = incols;
566: PetscLogEventEnd(MAT_GetRow,mat,0,0,0);
567: return(0);
568: }
570: /*@
571: MatConjugate - replaces the matrix values with their complex conjugates
573: Logically Collective on Mat
575: Input Parameters:
576: . mat - the matrix
578: Level: advanced
580: .seealso: VecConjugate()
581: @*/
582: PetscErrorCode MatConjugate(Mat mat)
583: {
584: #if defined(PETSC_USE_COMPLEX)
589: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
590: if (!mat->ops->conjugate) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not provided for matrix type %s, send email to petsc-maint@mcs.anl.gov",((PetscObject)mat)->type_name);
591: (*mat->ops->conjugate)(mat);
592: #else
594: #endif
595: return(0);
596: }
598: /*@C
599: MatRestoreRow - Frees any temporary space allocated by MatGetRow().
601: Not Collective
603: Input Parameters:
604: + mat - the matrix
605: . row - the row to get
606: . ncols, cols - the number of nonzeros and their columns
607: - vals - if nonzero the column values
609: Notes:
610: This routine should be called after you have finished examining the entries.
612: This routine zeros out ncols, cols, and vals. This is to prevent accidental
613: us of the array after it has been restored. If you pass NULL, it will
614: not zero the pointers. Use of cols or vals after MatRestoreRow is invalid.
616: Fortran Notes:
617: The calling sequence from Fortran is
618: .vb
619: MatRestoreRow(matrix,row,ncols,cols,values,ierr)
620: Mat matrix (input)
621: integer row (input)
622: integer ncols (output)
623: integer cols(maxcols) (output)
624: double precision (or double complex) values(maxcols) output
625: .ve
626: Where maxcols >= maximum nonzeros in any row of the matrix.
628: In Fortran MatRestoreRow() MUST be called after MatGetRow()
629: before another call to MatGetRow() can be made.
631: Level: advanced
633: .seealso: MatGetRow()
634: @*/
635: PetscErrorCode MatRestoreRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[])
636: {
642: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
643: if (!mat->ops->restorerow) return(0);
644: (*mat->ops->restorerow)(mat,row,ncols,(PetscInt **)cols,(PetscScalar **)vals);
645: if (ncols) *ncols = 0;
646: if (cols) *cols = NULL;
647: if (vals) *vals = NULL;
648: return(0);
649: }
651: /*@
652: MatGetRowUpperTriangular - Sets a flag to enable calls to MatGetRow() for matrix in MATSBAIJ format.
653: You should call MatRestoreRowUpperTriangular() after calling MatGetRow/MatRestoreRow() to disable the flag.
655: Not Collective
657: Input Parameters:
658: . mat - the matrix
660: Notes:
661: The flag is to ensure that users are aware of MatGetRow() only provides the upper triangular part of the row for the matrices in MATSBAIJ format.
663: Level: advanced
665: .seealso: MatRestoreRowUpperTriangular()
666: @*/
667: PetscErrorCode MatGetRowUpperTriangular(Mat mat)
668: {
674: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
675: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
676: MatCheckPreallocated(mat,1);
677: if (!mat->ops->getrowuppertriangular) return(0);
678: (*mat->ops->getrowuppertriangular)(mat);
679: return(0);
680: }
682: /*@
683: MatRestoreRowUpperTriangular - Disable calls to MatGetRow() for matrix in MATSBAIJ format.
685: Not Collective
687: Input Parameters:
688: . mat - the matrix
690: Notes:
691: This routine should be called after you have finished MatGetRow/MatRestoreRow().
694: Level: advanced
696: .seealso: MatGetRowUpperTriangular()
697: @*/
698: PetscErrorCode MatRestoreRowUpperTriangular(Mat mat)
699: {
705: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
706: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
707: MatCheckPreallocated(mat,1);
708: if (!mat->ops->restorerowuppertriangular) return(0);
709: (*mat->ops->restorerowuppertriangular)(mat);
710: return(0);
711: }
713: /*@C
714: MatSetOptionsPrefix - Sets the prefix used for searching for all
715: Mat options in the database.
717: Logically Collective on Mat
719: Input Parameter:
720: + A - the Mat context
721: - prefix - the prefix to prepend to all option names
723: Notes:
724: A hyphen (-) must NOT be given at the beginning of the prefix name.
725: The first character of all runtime options is AUTOMATICALLY the hyphen.
727: Level: advanced
729: .seealso: MatSetFromOptions()
730: @*/
731: PetscErrorCode MatSetOptionsPrefix(Mat A,const char prefix[])
732: {
737: PetscObjectSetOptionsPrefix((PetscObject)A,prefix);
738: return(0);
739: }
741: /*@C
742: MatAppendOptionsPrefix - Appends to the prefix used for searching for all
743: Mat options in the database.
745: Logically Collective on Mat
747: Input Parameters:
748: + A - the Mat context
749: - prefix - the prefix to prepend to all option names
751: Notes:
752: A hyphen (-) must NOT be given at the beginning of the prefix name.
753: The first character of all runtime options is AUTOMATICALLY the hyphen.
755: Level: advanced
757: .seealso: MatGetOptionsPrefix()
758: @*/
759: PetscErrorCode MatAppendOptionsPrefix(Mat A,const char prefix[])
760: {
765: PetscObjectAppendOptionsPrefix((PetscObject)A,prefix);
766: return(0);
767: }
769: /*@C
770: MatGetOptionsPrefix - Gets the prefix used for searching for all
771: Mat options in the database.
773: Not Collective
775: Input Parameter:
776: . A - the Mat context
778: Output Parameter:
779: . prefix - pointer to the prefix string used
781: Notes:
782: On the fortran side, the user should pass in a string 'prefix' of
783: sufficient length to hold the prefix.
785: Level: advanced
787: .seealso: MatAppendOptionsPrefix()
788: @*/
789: PetscErrorCode MatGetOptionsPrefix(Mat A,const char *prefix[])
790: {
795: PetscObjectGetOptionsPrefix((PetscObject)A,prefix);
796: return(0);
797: }
799: /*@
800: MatResetPreallocation - Reset mat to use the original nonzero pattern provided by users.
802: Collective on Mat
804: Input Parameters:
805: . A - the Mat context
807: Notes:
808: The allocated memory will be shrunk after calling MatAssembly with MAT_FINAL_ASSEMBLY. Users can reset the preallocation to access the original memory.
809: Currently support MPIAIJ and SEQAIJ.
811: Level: beginner
813: .seealso: MatSeqAIJSetPreallocation(), MatMPIAIJSetPreallocation(), MatXAIJSetPreallocation()
814: @*/
815: PetscErrorCode MatResetPreallocation(Mat A)
816: {
822: PetscUseMethod(A,"MatResetPreallocation_C",(Mat),(A));
823: return(0);
824: }
827: /*@
828: MatSetUp - Sets up the internal matrix data structures for later use.
830: Collective on Mat
832: Input Parameters:
833: . A - the Mat context
835: Notes:
836: If the user has not set preallocation for this matrix then a default preallocation that is likely to be inefficient is used.
838: If a suitable preallocation routine is used, this function does not need to be called.
840: See the Performance chapter of the PETSc users manual for how to preallocate matrices
842: Level: beginner
844: .seealso: MatCreate(), MatDestroy()
845: @*/
846: PetscErrorCode MatSetUp(Mat A)
847: {
848: PetscMPIInt size;
853: if (!((PetscObject)A)->type_name) {
854: MPI_Comm_size(PetscObjectComm((PetscObject)A), &size);
855: if (size == 1) {
856: MatSetType(A, MATSEQAIJ);
857: } else {
858: MatSetType(A, MATMPIAIJ);
859: }
860: }
861: if (!A->preallocated && A->ops->setup) {
862: PetscInfo(A,"Warning not preallocating matrix storage\n");
863: (*A->ops->setup)(A);
864: }
865: PetscLayoutSetUp(A->rmap);
866: PetscLayoutSetUp(A->cmap);
867: A->preallocated = PETSC_TRUE;
868: return(0);
869: }
871: #if defined(PETSC_HAVE_SAWS)
872: #include <petscviewersaws.h>
873: #endif
875: /*@C
876: MatViewFromOptions - View from Options
878: Collective on Mat
880: Input Parameters:
881: + A - the Mat context
882: . obj - Optional object
883: - name - command line option
885: Level: intermediate
886: .seealso: Mat, MatView, PetscObjectViewFromOptions(), MatCreate()
887: @*/
888: PetscErrorCode MatViewFromOptions(Mat A,PetscObject obj,const char name[])
889: {
894: PetscObjectViewFromOptions((PetscObject)A,obj,name);
895: return(0);
896: }
898: /*@C
899: MatView - Visualizes a matrix object.
901: Collective on Mat
903: Input Parameters:
904: + mat - the matrix
905: - viewer - visualization context
907: Notes:
908: The available visualization contexts include
909: + PETSC_VIEWER_STDOUT_SELF - for sequential matrices
910: . PETSC_VIEWER_STDOUT_WORLD - for parallel matrices created on PETSC_COMM_WORLD
911: . PETSC_VIEWER_STDOUT_(comm) - for matrices created on MPI communicator comm
912: - PETSC_VIEWER_DRAW_WORLD - graphical display of nonzero structure
914: The user can open alternative visualization contexts with
915: + PetscViewerASCIIOpen() - Outputs matrix to a specified file
916: . PetscViewerBinaryOpen() - Outputs matrix in binary to a
917: specified file; corresponding input uses MatLoad()
918: . PetscViewerDrawOpen() - Outputs nonzero matrix structure to
919: an X window display
920: - PetscViewerSocketOpen() - Outputs matrix to Socket viewer.
921: Currently only the sequential dense and AIJ
922: matrix types support the Socket viewer.
924: The user can call PetscViewerPushFormat() to specify the output
925: format of ASCII printed objects (when using PETSC_VIEWER_STDOUT_SELF,
926: PETSC_VIEWER_STDOUT_WORLD and PetscViewerASCIIOpen). Available formats include
927: + PETSC_VIEWER_DEFAULT - default, prints matrix contents
928: . PETSC_VIEWER_ASCII_MATLAB - prints matrix contents in Matlab format
929: . PETSC_VIEWER_ASCII_DENSE - prints entire matrix including zeros
930: . PETSC_VIEWER_ASCII_COMMON - prints matrix contents, using a sparse
931: format common among all matrix types
932: . PETSC_VIEWER_ASCII_IMPL - prints matrix contents, using an implementation-specific
933: format (which is in many cases the same as the default)
934: . PETSC_VIEWER_ASCII_INFO - prints basic information about the matrix
935: size and structure (not the matrix entries)
936: - PETSC_VIEWER_ASCII_INFO_DETAIL - prints more detailed information about
937: the matrix structure
939: Options Database Keys:
940: + -mat_view ::ascii_info - Prints info on matrix at conclusion of MatAssemblyEnd()
941: . -mat_view ::ascii_info_detail - Prints more detailed info
942: . -mat_view - Prints matrix in ASCII format
943: . -mat_view ::ascii_matlab - Prints matrix in Matlab format
944: . -mat_view draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX().
945: . -display <name> - Sets display name (default is host)
946: . -draw_pause <sec> - Sets number of seconds to pause after display
947: . -mat_view socket - Sends matrix to socket, can be accessed from Matlab (see Users-Manual: ch_matlab for details)
948: . -viewer_socket_machine <machine> -
949: . -viewer_socket_port <port> -
950: . -mat_view binary - save matrix to file in binary format
951: - -viewer_binary_filename <name> -
952: Level: beginner
954: Notes:
955: The ASCII viewers are only recommended for small matrices on at most a moderate number of processes,
956: the program will seemingly hang and take hours for larger matrices, for larger matrices one should use the binary format.
958: See the manual page for MatLoad() for the exact format of the binary file when the binary
959: viewer is used.
961: See share/petsc/matlab/PetscBinaryRead.m for a Matlab code that can read in the binary file when the binary
962: viewer is used and lib/petsc/bin/PetscBinaryIO.py for loading them into Python.
964: One can use '-mat_view draw -draw_pause -1' to pause the graphical display of matrix nonzero structure,
965: and then use the following mouse functions.
966: + left mouse: zoom in
967: . middle mouse: zoom out
968: - right mouse: continue with the simulation
970: .seealso: PetscViewerPushFormat(), PetscViewerASCIIOpen(), PetscViewerDrawOpen(),
971: PetscViewerSocketOpen(), PetscViewerBinaryOpen(), MatLoad()
972: @*/
973: PetscErrorCode MatView(Mat mat,PetscViewer viewer)
974: {
975: PetscErrorCode ierr;
976: PetscInt rows,cols,rbs,cbs;
977: PetscBool isascii,isstring,issaws;
978: PetscViewerFormat format;
979: PetscMPIInt size;
984: if (!viewer) {PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)mat),&viewer);}
987: MatCheckPreallocated(mat,1);
989: PetscViewerGetFormat(viewer,&format);
990: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
991: if (size == 1 && format == PETSC_VIEWER_LOAD_BALANCE) return(0);
993: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSTRING,&isstring);
994: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&isascii);
995: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSAWS,&issaws);
996: if ((!isascii || (format != PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL)) && mat->factortype) {
997: SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"No viewers for factored matrix except ASCII info or info_detail");
998: }
1000: PetscLogEventBegin(MAT_View,mat,viewer,0,0);
1001: if (isascii) {
1002: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ORDER,"Must call MatAssemblyBegin/End() before viewing matrix");
1003: PetscObjectPrintClassNamePrefixType((PetscObject)mat,viewer);
1004: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
1005: MatNullSpace nullsp,transnullsp;
1007: PetscViewerASCIIPushTab(viewer);
1008: MatGetSize(mat,&rows,&cols);
1009: MatGetBlockSizes(mat,&rbs,&cbs);
1010: if (rbs != 1 || cbs != 1) {
1011: if (rbs != cbs) {PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D, rbs=%D, cbs=%D\n",rows,cols,rbs,cbs);}
1012: else {PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D, bs=%D\n",rows,cols,rbs);}
1013: } else {
1014: PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D\n",rows,cols);
1015: }
1016: if (mat->factortype) {
1017: MatSolverType solver;
1018: MatFactorGetSolverType(mat,&solver);
1019: PetscViewerASCIIPrintf(viewer,"package used to perform factorization: %s\n",solver);
1020: }
1021: if (mat->ops->getinfo) {
1022: MatInfo info;
1023: MatGetInfo(mat,MAT_GLOBAL_SUM,&info);
1024: PetscViewerASCIIPrintf(viewer,"total: nonzeros=%.f, allocated nonzeros=%.f\n",info.nz_used,info.nz_allocated);
1025: if (!mat->factortype) {
1026: PetscViewerASCIIPrintf(viewer,"total number of mallocs used during MatSetValues calls=%D\n",(PetscInt)info.mallocs);
1027: }
1028: }
1029: MatGetNullSpace(mat,&nullsp);
1030: MatGetTransposeNullSpace(mat,&transnullsp);
1031: if (nullsp) {PetscViewerASCIIPrintf(viewer," has attached null space\n");}
1032: if (transnullsp && transnullsp != nullsp) {PetscViewerASCIIPrintf(viewer," has attached transposed null space\n");}
1033: MatGetNearNullSpace(mat,&nullsp);
1034: if (nullsp) {PetscViewerASCIIPrintf(viewer," has attached near null space\n");}
1035: PetscViewerASCIIPushTab(viewer);
1036: MatProductView(mat,viewer);
1037: PetscViewerASCIIPopTab(viewer);
1038: }
1039: } else if (issaws) {
1040: #if defined(PETSC_HAVE_SAWS)
1041: PetscMPIInt rank;
1043: PetscObjectName((PetscObject)mat);
1044: MPI_Comm_rank(PETSC_COMM_WORLD,&rank);
1045: if (!((PetscObject)mat)->amsmem && !rank) {
1046: PetscObjectViewSAWs((PetscObject)mat,viewer);
1047: }
1048: #endif
1049: } else if (isstring) {
1050: const char *type;
1051: MatGetType(mat,&type);
1052: PetscViewerStringSPrintf(viewer," MatType: %-7.7s",type);
1053: if (mat->ops->view) {(*mat->ops->view)(mat,viewer);}
1054: }
1055: if ((format == PETSC_VIEWER_NATIVE || format == PETSC_VIEWER_LOAD_BALANCE) && mat->ops->viewnative) {
1056: PetscViewerASCIIPushTab(viewer);
1057: (*mat->ops->viewnative)(mat,viewer);
1058: PetscViewerASCIIPopTab(viewer);
1059: } else if (mat->ops->view) {
1060: PetscViewerASCIIPushTab(viewer);
1061: (*mat->ops->view)(mat,viewer);
1062: PetscViewerASCIIPopTab(viewer);
1063: }
1064: if (isascii) {
1065: PetscViewerGetFormat(viewer,&format);
1066: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
1067: PetscViewerASCIIPopTab(viewer);
1068: }
1069: }
1070: PetscLogEventEnd(MAT_View,mat,viewer,0,0);
1071: return(0);
1072: }
1074: #if defined(PETSC_USE_DEBUG)
1075: #include <../src/sys/totalview/tv_data_display.h>
1076: PETSC_UNUSED static int TV_display_type(const struct _p_Mat *mat)
1077: {
1078: TV_add_row("Local rows", "int", &mat->rmap->n);
1079: TV_add_row("Local columns", "int", &mat->cmap->n);
1080: TV_add_row("Global rows", "int", &mat->rmap->N);
1081: TV_add_row("Global columns", "int", &mat->cmap->N);
1082: TV_add_row("Typename", TV_ascii_string_type, ((PetscObject)mat)->type_name);
1083: return TV_format_OK;
1084: }
1085: #endif
1087: /*@C
1088: MatLoad - Loads a matrix that has been stored in binary/HDF5 format
1089: with MatView(). The matrix format is determined from the options database.
1090: Generates a parallel MPI matrix if the communicator has more than one
1091: processor. The default matrix type is AIJ.
1093: Collective on PetscViewer
1095: Input Parameters:
1096: + mat - the newly loaded matrix, this needs to have been created with MatCreate()
1097: or some related function before a call to MatLoad()
1098: - viewer - binary/HDF5 file viewer
1100: Options Database Keys:
1101: Used with block matrix formats (MATSEQBAIJ, ...) to specify
1102: block size
1103: . -matload_block_size <bs>
1105: Level: beginner
1107: Notes:
1108: If the Mat type has not yet been given then MATAIJ is used, call MatSetFromOptions() on the
1109: Mat before calling this routine if you wish to set it from the options database.
1111: MatLoad() automatically loads into the options database any options
1112: given in the file filename.info where filename is the name of the file
1113: that was passed to the PetscViewerBinaryOpen(). The options in the info
1114: file will be ignored if you use the -viewer_binary_skip_info option.
1116: If the type or size of mat is not set before a call to MatLoad, PETSc
1117: sets the default matrix type AIJ and sets the local and global sizes.
1118: If type and/or size is already set, then the same are used.
1120: In parallel, each processor can load a subset of rows (or the
1121: entire matrix). This routine is especially useful when a large
1122: matrix is stored on disk and only part of it is desired on each
1123: processor. For example, a parallel solver may access only some of
1124: the rows from each processor. The algorithm used here reads
1125: relatively small blocks of data rather than reading the entire
1126: matrix and then subsetting it.
1128: Viewer's PetscViewerType must be either PETSCVIEWERBINARY or PETSCVIEWERHDF5.
1129: Such viewer can be created using PetscViewerBinaryOpen()/PetscViewerHDF5Open(),
1130: or the sequence like
1131: $ PetscViewer v;
1132: $ PetscViewerCreate(PETSC_COMM_WORLD,&v);
1133: $ PetscViewerSetType(v,PETSCVIEWERBINARY);
1134: $ PetscViewerSetFromOptions(v);
1135: $ PetscViewerFileSetMode(v,FILE_MODE_READ);
1136: $ PetscViewerFileSetName(v,"datafile");
1137: The optional PetscViewerSetFromOptions() call allows to override PetscViewerSetType() using option
1138: $ -viewer_type {binary,hdf5}
1140: See the example src/ksp/ksp/tutorials/ex27.c with the first approach,
1141: and src/mat/tutorials/ex10.c with the second approach.
1143: Notes about the PETSc binary format:
1144: In case of PETSCVIEWERBINARY, a native PETSc binary format is used. Each of the blocks
1145: is read onto rank 0 and then shipped to its destination rank, one after another.
1146: Multiple objects, both matrices and vectors, can be stored within the same file.
1147: Their PetscObject name is ignored; they are loaded in the order of their storage.
1149: Most users should not need to know the details of the binary storage
1150: format, since MatLoad() and MatView() completely hide these details.
1151: But for anyone who's interested, the standard binary matrix storage
1152: format is
1154: $ PetscInt MAT_FILE_CLASSID
1155: $ PetscInt number of rows
1156: $ PetscInt number of columns
1157: $ PetscInt total number of nonzeros
1158: $ PetscInt *number nonzeros in each row
1159: $ PetscInt *column indices of all nonzeros (starting index is zero)
1160: $ PetscScalar *values of all nonzeros
1162: PETSc automatically does the byte swapping for
1163: machines that store the bytes reversed, e.g. DEC alpha, freebsd,
1164: linux, Windows and the paragon; thus if you write your own binary
1165: read/write routines you have to swap the bytes; see PetscBinaryRead()
1166: and PetscBinaryWrite() to see how this may be done.
1168: Notes about the HDF5 (MATLAB MAT-File Version 7.3) format:
1169: In case of PETSCVIEWERHDF5, a parallel HDF5 reader is used.
1170: Each processor's chunk is loaded independently by its owning rank.
1171: Multiple objects, both matrices and vectors, can be stored within the same file.
1172: They are looked up by their PetscObject name.
1174: As the MATLAB MAT-File Version 7.3 format is also a HDF5 flavor, we decided to use
1175: by default the same structure and naming of the AIJ arrays and column count
1176: within the HDF5 file. This means that a MAT file saved with -v7.3 flag, e.g.
1177: $ save example.mat A b -v7.3
1178: can be directly read by this routine (see Reference 1 for details).
1179: Note that depending on your MATLAB version, this format might be a default,
1180: otherwise you can set it as default in Preferences.
1182: Unless -nocompression flag is used to save the file in MATLAB,
1183: PETSc must be configured with ZLIB package.
1185: See also examples src/mat/tutorials/ex10.c and src/ksp/ksp/tutorials/ex27.c
1187: Current HDF5 (MAT-File) limitations:
1188: This reader currently supports only real MATSEQAIJ, MATMPIAIJ, MATSEQDENSE and MATMPIDENSE matrices.
1190: Corresponding MatView() is not yet implemented.
1192: The loaded matrix is actually a transpose of the original one in MATLAB,
1193: unless you push PETSC_VIEWER_HDF5_MAT format (see examples above).
1194: With this format, matrix is automatically transposed by PETSc,
1195: unless the matrix is marked as SPD or symmetric
1196: (see MatSetOption(), MAT_SPD, MAT_SYMMETRIC).
1198: References:
1199: 1. MATLAB(R) Documentation, manual page of save(), https://www.mathworks.com/help/matlab/ref/save.html#btox10b-1-version
1201: .seealso: PetscViewerBinaryOpen(), PetscViewerSetType(), MatView(), VecLoad()
1203: @*/
1204: PetscErrorCode MatLoad(Mat mat,PetscViewer viewer)
1205: {
1207: PetscBool flg;
1213: if (!((PetscObject)mat)->type_name) {
1214: MatSetType(mat,MATAIJ);
1215: }
1217: flg = PETSC_FALSE;
1218: PetscOptionsGetBool(((PetscObject)mat)->options,((PetscObject)mat)->prefix,"-matload_symmetric",&flg,NULL);
1219: if (flg) {
1220: MatSetOption(mat,MAT_SYMMETRIC,PETSC_TRUE);
1221: MatSetOption(mat,MAT_SYMMETRY_ETERNAL,PETSC_TRUE);
1222: }
1223: flg = PETSC_FALSE;
1224: PetscOptionsGetBool(((PetscObject)mat)->options,((PetscObject)mat)->prefix,"-matload_spd",&flg,NULL);
1225: if (flg) {
1226: MatSetOption(mat,MAT_SPD,PETSC_TRUE);
1227: }
1229: if (!mat->ops->load) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatLoad is not supported for type %s",((PetscObject)mat)->type_name);
1230: PetscLogEventBegin(MAT_Load,mat,viewer,0,0);
1231: (*mat->ops->load)(mat,viewer);
1232: PetscLogEventEnd(MAT_Load,mat,viewer,0,0);
1233: return(0);
1234: }
1236: static PetscErrorCode MatDestroy_Redundant(Mat_Redundant **redundant)
1237: {
1239: Mat_Redundant *redund = *redundant;
1240: PetscInt i;
1243: if (redund){
1244: if (redund->matseq) { /* via MatCreateSubMatrices() */
1245: ISDestroy(&redund->isrow);
1246: ISDestroy(&redund->iscol);
1247: MatDestroySubMatrices(1,&redund->matseq);
1248: } else {
1249: PetscFree2(redund->send_rank,redund->recv_rank);
1250: PetscFree(redund->sbuf_j);
1251: PetscFree(redund->sbuf_a);
1252: for (i=0; i<redund->nrecvs; i++) {
1253: PetscFree(redund->rbuf_j[i]);
1254: PetscFree(redund->rbuf_a[i]);
1255: }
1256: PetscFree4(redund->sbuf_nz,redund->rbuf_nz,redund->rbuf_j,redund->rbuf_a);
1257: }
1259: if (redund->subcomm) {
1260: PetscCommDestroy(&redund->subcomm);
1261: }
1262: PetscFree(redund);
1263: }
1264: return(0);
1265: }
1267: /*@
1268: MatDestroy - Frees space taken by a matrix.
1270: Collective on Mat
1272: Input Parameter:
1273: . A - the matrix
1275: Level: beginner
1277: @*/
1278: PetscErrorCode MatDestroy(Mat *A)
1279: {
1283: if (!*A) return(0);
1285: if (--((PetscObject)(*A))->refct > 0) {*A = NULL; return(0);}
1287: /* if memory was published with SAWs then destroy it */
1288: PetscObjectSAWsViewOff((PetscObject)*A);
1289: if ((*A)->ops->destroy) {
1290: (*(*A)->ops->destroy)(*A);
1291: }
1293: PetscFree((*A)->defaultvectype);
1294: PetscFree((*A)->bsizes);
1295: PetscFree((*A)->solvertype);
1296: MatDestroy_Redundant(&(*A)->redundant);
1297: MatProductClear(*A);
1298: MatNullSpaceDestroy(&(*A)->nullsp);
1299: MatNullSpaceDestroy(&(*A)->transnullsp);
1300: MatNullSpaceDestroy(&(*A)->nearnullsp);
1301: MatDestroy(&(*A)->schur);
1302: PetscLayoutDestroy(&(*A)->rmap);
1303: PetscLayoutDestroy(&(*A)->cmap);
1304: PetscHeaderDestroy(A);
1305: return(0);
1306: }
1308: /*@C
1309: MatSetValues - Inserts or adds a block of values into a matrix.
1310: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
1311: MUST be called after all calls to MatSetValues() have been completed.
1313: Not Collective
1315: Input Parameters:
1316: + mat - the matrix
1317: . v - a logically two-dimensional array of values
1318: . m, idxm - the number of rows and their global indices
1319: . n, idxn - the number of columns and their global indices
1320: - addv - either ADD_VALUES or INSERT_VALUES, where
1321: ADD_VALUES adds values to any existing entries, and
1322: INSERT_VALUES replaces existing entries with new values
1324: Notes:
1325: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
1326: MatSetUp() before using this routine
1328: By default the values, v, are row-oriented. See MatSetOption() for other options.
1330: Calls to MatSetValues() with the INSERT_VALUES and ADD_VALUES
1331: options cannot be mixed without intervening calls to the assembly
1332: routines.
1334: MatSetValues() uses 0-based row and column numbers in Fortran
1335: as well as in C.
1337: Negative indices may be passed in idxm and idxn, these rows and columns are
1338: simply ignored. This allows easily inserting element stiffness matrices
1339: with homogeneous Dirchlet boundary conditions that you don't want represented
1340: in the matrix.
1342: Efficiency Alert:
1343: The routine MatSetValuesBlocked() may offer much better efficiency
1344: for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ).
1346: Level: beginner
1348: Developer Notes:
1349: This is labeled with C so does not automatically generate Fortran stubs and interfaces
1350: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
1352: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1353: InsertMode, INSERT_VALUES, ADD_VALUES
1354: @*/
1355: PetscErrorCode MatSetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv)
1356: {
1362: if (!m || !n) return(0); /* no values to insert */
1365: MatCheckPreallocated(mat,1);
1367: if (mat->insertmode == NOT_SET_VALUES) {
1368: mat->insertmode = addv;
1369: } else if (PetscUnlikely(mat->insertmode != addv)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
1370: if (PetscDefined(USE_DEBUG)) {
1371: PetscInt i,j;
1373: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1374: if (!mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1376: for (i=0; i<m; i++) {
1377: for (j=0; j<n; j++) {
1378: if (mat->erroriffailure && PetscIsInfOrNanScalar(v[i*n+j]))
1379: #if defined(PETSC_USE_COMPLEX)
1380: SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_FP,"Inserting %g+ig at matrix entry (%D,%D)",(double)PetscRealPart(v[i*n+j]),(double)PetscImaginaryPart(v[i*n+j]),idxm[i],idxn[j]);
1381: #else
1382: SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_FP,"Inserting %g at matrix entry (%D,%D)",(double)v[i*n+j],idxm[i],idxn[j]);
1383: #endif
1384: }
1385: }
1386: }
1388: if (mat->assembled) {
1389: mat->was_assembled = PETSC_TRUE;
1390: mat->assembled = PETSC_FALSE;
1391: }
1392: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1393: (*mat->ops->setvalues)(mat,m,idxm,n,idxn,v,addv);
1394: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1395: return(0);
1396: }
1399: /*@
1400: MatSetValuesRowLocal - Inserts a row (block row for BAIJ matrices) of nonzero
1401: values into a matrix
1403: Not Collective
1405: Input Parameters:
1406: + mat - the matrix
1407: . row - the (block) row to set
1408: - v - a logically two-dimensional array of values
1410: Notes:
1411: By the values, v, are column-oriented (for the block version) and sorted
1413: All the nonzeros in the row must be provided
1415: The matrix must have previously had its column indices set
1417: The row must belong to this process
1419: Level: intermediate
1421: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1422: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues(), MatSetValuesRow(), MatSetLocalToGlobalMapping()
1423: @*/
1424: PetscErrorCode MatSetValuesRowLocal(Mat mat,PetscInt row,const PetscScalar v[])
1425: {
1427: PetscInt globalrow;
1433: ISLocalToGlobalMappingApply(mat->rmap->mapping,1,&row,&globalrow);
1434: MatSetValuesRow(mat,globalrow,v);
1435: return(0);
1436: }
1438: /*@
1439: MatSetValuesRow - Inserts a row (block row for BAIJ matrices) of nonzero
1440: values into a matrix
1442: Not Collective
1444: Input Parameters:
1445: + mat - the matrix
1446: . row - the (block) row to set
1447: - v - a logically two-dimensional (column major) array of values for block matrices with blocksize larger than one, otherwise a one dimensional array of values
1449: Notes:
1450: The values, v, are column-oriented for the block version.
1452: All the nonzeros in the row must be provided
1454: THE MATRIX MUST HAVE PREVIOUSLY HAD ITS COLUMN INDICES SET. IT IS RARE THAT THIS ROUTINE IS USED, usually MatSetValues() is used.
1456: The row must belong to this process
1458: Level: advanced
1460: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1461: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues()
1462: @*/
1463: PetscErrorCode MatSetValuesRow(Mat mat,PetscInt row,const PetscScalar v[])
1464: {
1470: MatCheckPreallocated(mat,1);
1472: if (PetscUnlikely(mat->insertmode == ADD_VALUES)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add and insert values");
1473: if (PetscUnlikely(mat->factortype)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1474: mat->insertmode = INSERT_VALUES;
1476: if (mat->assembled) {
1477: mat->was_assembled = PETSC_TRUE;
1478: mat->assembled = PETSC_FALSE;
1479: }
1480: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1481: if (!mat->ops->setvaluesrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1482: (*mat->ops->setvaluesrow)(mat,row,v);
1483: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1484: return(0);
1485: }
1487: /*@
1488: MatSetValuesStencil - Inserts or adds a block of values into a matrix.
1489: Using structured grid indexing
1491: Not Collective
1493: Input Parameters:
1494: + mat - the matrix
1495: . m - number of rows being entered
1496: . idxm - grid coordinates (and component number when dof > 1) for matrix rows being entered
1497: . n - number of columns being entered
1498: . idxn - grid coordinates (and component number when dof > 1) for matrix columns being entered
1499: . v - a logically two-dimensional array of values
1500: - addv - either ADD_VALUES or INSERT_VALUES, where
1501: ADD_VALUES adds values to any existing entries, and
1502: INSERT_VALUES replaces existing entries with new values
1504: Notes:
1505: By default the values, v, are row-oriented. See MatSetOption() for other options.
1507: Calls to MatSetValuesStencil() with the INSERT_VALUES and ADD_VALUES
1508: options cannot be mixed without intervening calls to the assembly
1509: routines.
1511: The grid coordinates are across the entire grid, not just the local portion
1513: MatSetValuesStencil() uses 0-based row and column numbers in Fortran
1514: as well as in C.
1516: For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine
1518: In order to use this routine you must either obtain the matrix with DMCreateMatrix()
1519: or call MatSetLocalToGlobalMapping() and MatSetStencil() first.
1521: The columns and rows in the stencil passed in MUST be contained within the
1522: ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example,
1523: if you create a DMDA with an overlap of one grid level and on a particular process its first
1524: local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the
1525: first i index you can use in your column and row indices in MatSetStencil() is 5.
1527: In Fortran idxm and idxn should be declared as
1528: $ MatStencil idxm(4,m),idxn(4,n)
1529: and the values inserted using
1530: $ idxm(MatStencil_i,1) = i
1531: $ idxm(MatStencil_j,1) = j
1532: $ idxm(MatStencil_k,1) = k
1533: $ idxm(MatStencil_c,1) = c
1534: etc
1536: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
1537: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
1538: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
1539: DM_BOUNDARY_PERIODIC boundary type.
1541: For indices that don't mean anything for your case (like the k index when working in 2d) or the c index when you have
1542: a single value per point) you can skip filling those indices.
1544: Inspired by the structured grid interface to the HYPRE package
1545: (https://computation.llnl.gov/projects/hypre-scalable-linear-solvers-multigrid-methods)
1547: Efficiency Alert:
1548: The routine MatSetValuesBlockedStencil() may offer much better efficiency
1549: for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ).
1551: Level: beginner
1553: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1554: MatSetValues(), MatSetValuesBlockedStencil(), MatSetStencil(), DMCreateMatrix(), DMDAVecGetArray(), MatStencil
1555: @*/
1556: PetscErrorCode MatSetValuesStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv)
1557: {
1559: PetscInt buf[8192],*bufm=NULL,*bufn=NULL,*jdxm,*jdxn;
1560: PetscInt j,i,dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp;
1561: PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc);
1564: if (!m || !n) return(0); /* no values to insert */
1570: if ((m+n) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1571: jdxm = buf; jdxn = buf+m;
1572: } else {
1573: PetscMalloc2(m,&bufm,n,&bufn);
1574: jdxm = bufm; jdxn = bufn;
1575: }
1576: for (i=0; i<m; i++) {
1577: for (j=0; j<3-sdim; j++) dxm++;
1578: tmp = *dxm++ - starts[0];
1579: for (j=0; j<dim-1; j++) {
1580: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1581: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
1582: }
1583: if (mat->stencil.noc) dxm++;
1584: jdxm[i] = tmp;
1585: }
1586: for (i=0; i<n; i++) {
1587: for (j=0; j<3-sdim; j++) dxn++;
1588: tmp = *dxn++ - starts[0];
1589: for (j=0; j<dim-1; j++) {
1590: if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1591: else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1];
1592: }
1593: if (mat->stencil.noc) dxn++;
1594: jdxn[i] = tmp;
1595: }
1596: MatSetValuesLocal(mat,m,jdxm,n,jdxn,v,addv);
1597: PetscFree2(bufm,bufn);
1598: return(0);
1599: }
1601: /*@
1602: MatSetValuesBlockedStencil - Inserts or adds a block of values into a matrix.
1603: Using structured grid indexing
1605: Not Collective
1607: Input Parameters:
1608: + mat - the matrix
1609: . m - number of rows being entered
1610: . idxm - grid coordinates for matrix rows being entered
1611: . n - number of columns being entered
1612: . idxn - grid coordinates for matrix columns being entered
1613: . v - a logically two-dimensional array of values
1614: - addv - either ADD_VALUES or INSERT_VALUES, where
1615: ADD_VALUES adds values to any existing entries, and
1616: INSERT_VALUES replaces existing entries with new values
1618: Notes:
1619: By default the values, v, are row-oriented and unsorted.
1620: See MatSetOption() for other options.
1622: Calls to MatSetValuesBlockedStencil() with the INSERT_VALUES and ADD_VALUES
1623: options cannot be mixed without intervening calls to the assembly
1624: routines.
1626: The grid coordinates are across the entire grid, not just the local portion
1628: MatSetValuesBlockedStencil() uses 0-based row and column numbers in Fortran
1629: as well as in C.
1631: For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine
1633: In order to use this routine you must either obtain the matrix with DMCreateMatrix()
1634: or call MatSetBlockSize(), MatSetLocalToGlobalMapping() and MatSetStencil() first.
1636: The columns and rows in the stencil passed in MUST be contained within the
1637: ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example,
1638: if you create a DMDA with an overlap of one grid level and on a particular process its first
1639: local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the
1640: first i index you can use in your column and row indices in MatSetStencil() is 5.
1642: In Fortran idxm and idxn should be declared as
1643: $ MatStencil idxm(4,m),idxn(4,n)
1644: and the values inserted using
1645: $ idxm(MatStencil_i,1) = i
1646: $ idxm(MatStencil_j,1) = j
1647: $ idxm(MatStencil_k,1) = k
1648: etc
1650: Negative indices may be passed in idxm and idxn, these rows and columns are
1651: simply ignored. This allows easily inserting element stiffness matrices
1652: with homogeneous Dirchlet boundary conditions that you don't want represented
1653: in the matrix.
1655: Inspired by the structured grid interface to the HYPRE package
1656: (https://computation.llnl.gov/projects/hypre-scalable-linear-solvers-multigrid-methods)
1658: Level: beginner
1660: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1661: MatSetValues(), MatSetValuesStencil(), MatSetStencil(), DMCreateMatrix(), DMDAVecGetArray(), MatStencil,
1662: MatSetBlockSize(), MatSetLocalToGlobalMapping()
1663: @*/
1664: PetscErrorCode MatSetValuesBlockedStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv)
1665: {
1667: PetscInt buf[8192],*bufm=NULL,*bufn=NULL,*jdxm,*jdxn;
1668: PetscInt j,i,dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp;
1669: PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc);
1672: if (!m || !n) return(0); /* no values to insert */
1679: if ((m+n) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1680: jdxm = buf; jdxn = buf+m;
1681: } else {
1682: PetscMalloc2(m,&bufm,n,&bufn);
1683: jdxm = bufm; jdxn = bufn;
1684: }
1685: for (i=0; i<m; i++) {
1686: for (j=0; j<3-sdim; j++) dxm++;
1687: tmp = *dxm++ - starts[0];
1688: for (j=0; j<sdim-1; j++) {
1689: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1690: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
1691: }
1692: dxm++;
1693: jdxm[i] = tmp;
1694: }
1695: for (i=0; i<n; i++) {
1696: for (j=0; j<3-sdim; j++) dxn++;
1697: tmp = *dxn++ - starts[0];
1698: for (j=0; j<sdim-1; j++) {
1699: if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1700: else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1];
1701: }
1702: dxn++;
1703: jdxn[i] = tmp;
1704: }
1705: MatSetValuesBlockedLocal(mat,m,jdxm,n,jdxn,v,addv);
1706: PetscFree2(bufm,bufn);
1707: return(0);
1708: }
1710: /*@
1711: MatSetStencil - Sets the grid information for setting values into a matrix via
1712: MatSetValuesStencil()
1714: Not Collective
1716: Input Parameters:
1717: + mat - the matrix
1718: . dim - dimension of the grid 1, 2, or 3
1719: . dims - number of grid points in x, y, and z direction, including ghost points on your processor
1720: . starts - starting point of ghost nodes on your processor in x, y, and z direction
1721: - dof - number of degrees of freedom per node
1724: Inspired by the structured grid interface to the HYPRE package
1725: (www.llnl.gov/CASC/hyper)
1727: For matrices generated with DMCreateMatrix() this routine is automatically called and so not needed by the
1728: user.
1730: Level: beginner
1732: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1733: MatSetValues(), MatSetValuesBlockedStencil(), MatSetValuesStencil()
1734: @*/
1735: PetscErrorCode MatSetStencil(Mat mat,PetscInt dim,const PetscInt dims[],const PetscInt starts[],PetscInt dof)
1736: {
1737: PetscInt i;
1744: mat->stencil.dim = dim + (dof > 1);
1745: for (i=0; i<dim; i++) {
1746: mat->stencil.dims[i] = dims[dim-i-1]; /* copy the values in backwards */
1747: mat->stencil.starts[i] = starts[dim-i-1];
1748: }
1749: mat->stencil.dims[dim] = dof;
1750: mat->stencil.starts[dim] = 0;
1751: mat->stencil.noc = (PetscBool)(dof == 1);
1752: return(0);
1753: }
1755: /*@C
1756: MatSetValuesBlocked - Inserts or adds a block of values into a matrix.
1758: Not Collective
1760: Input Parameters:
1761: + mat - the matrix
1762: . v - a logically two-dimensional array of values
1763: . m, idxm - the number of block rows and their global block indices
1764: . n, idxn - the number of block columns and their global block indices
1765: - addv - either ADD_VALUES or INSERT_VALUES, where
1766: ADD_VALUES adds values to any existing entries, and
1767: INSERT_VALUES replaces existing entries with new values
1769: Notes:
1770: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call
1771: MatXXXXSetPreallocation() or MatSetUp() before using this routine.
1773: The m and n count the NUMBER of blocks in the row direction and column direction,
1774: NOT the total number of rows/columns; for example, if the block size is 2 and
1775: you are passing in values for rows 2,3,4,5 then m would be 2 (not 4).
1776: The values in idxm would be 1 2; that is the first index for each block divided by
1777: the block size.
1779: Note that you must call MatSetBlockSize() when constructing this matrix (before
1780: preallocating it).
1782: By default the values, v, are row-oriented, so the layout of
1783: v is the same as for MatSetValues(). See MatSetOption() for other options.
1785: Calls to MatSetValuesBlocked() with the INSERT_VALUES and ADD_VALUES
1786: options cannot be mixed without intervening calls to the assembly
1787: routines.
1789: MatSetValuesBlocked() uses 0-based row and column numbers in Fortran
1790: as well as in C.
1792: Negative indices may be passed in idxm and idxn, these rows and columns are
1793: simply ignored. This allows easily inserting element stiffness matrices
1794: with homogeneous Dirchlet boundary conditions that you don't want represented
1795: in the matrix.
1797: Each time an entry is set within a sparse matrix via MatSetValues(),
1798: internal searching must be done to determine where to place the
1799: data in the matrix storage space. By instead inserting blocks of
1800: entries via MatSetValuesBlocked(), the overhead of matrix assembly is
1801: reduced.
1803: Example:
1804: $ Suppose m=n=2 and block size(bs) = 2 The array is
1805: $
1806: $ 1 2 | 3 4
1807: $ 5 6 | 7 8
1808: $ - - - | - - -
1809: $ 9 10 | 11 12
1810: $ 13 14 | 15 16
1811: $
1812: $ v[] should be passed in like
1813: $ v[] = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16]
1814: $
1815: $ If you are not using row oriented storage of v (that is you called MatSetOption(mat,MAT_ROW_ORIENTED,PETSC_FALSE)) then
1816: $ v[] = [1,5,9,13,2,6,10,14,3,7,11,15,4,8,12,16]
1818: Level: intermediate
1820: .seealso: MatSetBlockSize(), MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal()
1821: @*/
1822: PetscErrorCode MatSetValuesBlocked(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv)
1823: {
1829: if (!m || !n) return(0); /* no values to insert */
1833: MatCheckPreallocated(mat,1);
1834: if (mat->insertmode == NOT_SET_VALUES) {
1835: mat->insertmode = addv;
1836: } else if (PetscUnlikely(mat->insertmode != addv)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
1837: if (PetscDefined(USE_DEBUG)) {
1838: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1839: if (!mat->ops->setvaluesblocked && !mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1840: }
1842: if (mat->assembled) {
1843: mat->was_assembled = PETSC_TRUE;
1844: mat->assembled = PETSC_FALSE;
1845: }
1846: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1847: if (mat->ops->setvaluesblocked) {
1848: (*mat->ops->setvaluesblocked)(mat,m,idxm,n,idxn,v,addv);
1849: } else {
1850: PetscInt buf[8192],*bufr=NULL,*bufc=NULL,*iidxm,*iidxn;
1851: PetscInt i,j,bs,cbs;
1852: MatGetBlockSizes(mat,&bs,&cbs);
1853: if (m*bs+n*cbs <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1854: iidxm = buf; iidxn = buf + m*bs;
1855: } else {
1856: PetscMalloc2(m*bs,&bufr,n*cbs,&bufc);
1857: iidxm = bufr; iidxn = bufc;
1858: }
1859: for (i=0; i<m; i++) {
1860: for (j=0; j<bs; j++) {
1861: iidxm[i*bs+j] = bs*idxm[i] + j;
1862: }
1863: }
1864: for (i=0; i<n; i++) {
1865: for (j=0; j<cbs; j++) {
1866: iidxn[i*cbs+j] = cbs*idxn[i] + j;
1867: }
1868: }
1869: MatSetValues(mat,m*bs,iidxm,n*cbs,iidxn,v,addv);
1870: PetscFree2(bufr,bufc);
1871: }
1872: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1873: return(0);
1874: }
1876: /*@C
1877: MatGetValues - Gets a block of values from a matrix.
1879: Not Collective; currently only returns a local block
1881: Input Parameters:
1882: + mat - the matrix
1883: . v - a logically two-dimensional array for storing the values
1884: . m, idxm - the number of rows and their global indices
1885: - n, idxn - the number of columns and their global indices
1887: Notes:
1888: The user must allocate space (m*n PetscScalars) for the values, v.
1889: The values, v, are then returned in a row-oriented format,
1890: analogous to that used by default in MatSetValues().
1892: MatGetValues() uses 0-based row and column numbers in
1893: Fortran as well as in C.
1895: MatGetValues() requires that the matrix has been assembled
1896: with MatAssemblyBegin()/MatAssemblyEnd(). Thus, calls to
1897: MatSetValues() and MatGetValues() CANNOT be made in succession
1898: without intermediate matrix assembly.
1900: Negative row or column indices will be ignored and those locations in v[] will be
1901: left unchanged.
1903: Level: advanced
1905: .seealso: MatGetRow(), MatCreateSubMatrices(), MatSetValues()
1906: @*/
1907: PetscErrorCode MatGetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[])
1908: {
1914: if (!m || !n) return(0);
1918: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
1919: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1920: if (!mat->ops->getvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1921: MatCheckPreallocated(mat,1);
1923: PetscLogEventBegin(MAT_GetValues,mat,0,0,0);
1924: (*mat->ops->getvalues)(mat,m,idxm,n,idxn,v);
1925: PetscLogEventEnd(MAT_GetValues,mat,0,0,0);
1926: return(0);
1927: }
1929: /*@C
1930: MatGetValuesLocal - retrieves values into certain locations of a matrix,
1931: using a local numbering of the nodes.
1933: Not Collective
1935: Input Parameters:
1936: + mat - the matrix
1937: . nrow, irow - number of rows and their local indices
1938: - ncol, icol - number of columns and their local indices
1940: Output Parameter:
1941: . y - a logically two-dimensional array of values
1943: Notes:
1944: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatSetLocalToGlobalMapping() before using this routine
1946: Level: advanced
1948: Developer Notes:
1949: This is labelled with C so does not automatically generate Fortran stubs and interfaces
1950: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
1952: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetLocalToGlobalMapping(),
1953: MatSetValuesLocal()
1954: @*/
1955: PetscErrorCode MatGetValuesLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],PetscScalar y[])
1956: {
1962: MatCheckPreallocated(mat,1);
1963: if (!nrow || !ncol) return(0); /* no values to retrieve */
1966: if (PetscDefined(USE_DEBUG)) {
1967: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1968: if (!mat->ops->getvalueslocal && !mat->ops->getvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1969: }
1970: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
1971: PetscLogEventBegin(MAT_GetValues,mat,0,0,0);
1972: if (mat->ops->getvalueslocal) {
1973: (*mat->ops->getvalueslocal)(mat,nrow,irow,ncol,icol,y);
1974: } else {
1975: PetscInt buf[8192],*bufr=NULL,*bufc=NULL,*irowm,*icolm;
1976: if ((nrow+ncol) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1977: irowm = buf; icolm = buf+nrow;
1978: } else {
1979: PetscMalloc2(nrow,&bufr,ncol,&bufc);
1980: irowm = bufr; icolm = bufc;
1981: }
1982: if (!mat->rmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MatGetValuesLocal() cannot proceed without local-to-global row mapping (See MatSetLocalToGlobalMapping()).");
1983: if (!mat->cmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MatGetValuesLocal() cannot proceed without local-to-global column mapping (See MatSetLocalToGlobalMapping()).");
1984: ISLocalToGlobalMappingApply(mat->rmap->mapping,nrow,irow,irowm);
1985: ISLocalToGlobalMappingApply(mat->cmap->mapping,ncol,icol,icolm);
1986: MatGetValues(mat,nrow,irowm,ncol,icolm,y);
1987: PetscFree2(bufr,bufc);
1988: }
1989: PetscLogEventEnd(MAT_GetValues,mat,0,0,0);
1990: return(0);
1991: }
1993: /*@
1994: MatSetValuesBatch - Adds (ADD_VALUES) many blocks of values into a matrix at once. The blocks must all be square and
1995: the same size. Currently, this can only be called once and creates the given matrix.
1997: Not Collective
1999: Input Parameters:
2000: + mat - the matrix
2001: . nb - the number of blocks
2002: . bs - the number of rows (and columns) in each block
2003: . rows - a concatenation of the rows for each block
2004: - v - a concatenation of logically two-dimensional arrays of values
2006: Notes:
2007: In the future, we will extend this routine to handle rectangular blocks, and to allow multiple calls for a given matrix.
2009: Level: advanced
2011: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
2012: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues()
2013: @*/
2014: PetscErrorCode MatSetValuesBatch(Mat mat, PetscInt nb, PetscInt bs, PetscInt rows[], const PetscScalar v[])
2015: {
2023: if (PetscUnlikelyDebug(mat->factortype)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2025: PetscLogEventBegin(MAT_SetValuesBatch,mat,0,0,0);
2026: if (mat->ops->setvaluesbatch) {
2027: (*mat->ops->setvaluesbatch)(mat,nb,bs,rows,v);
2028: } else {
2029: PetscInt b;
2030: for (b = 0; b < nb; ++b) {
2031: MatSetValues(mat, bs, &rows[b*bs], bs, &rows[b*bs], &v[b*bs*bs], ADD_VALUES);
2032: }
2033: }
2034: PetscLogEventEnd(MAT_SetValuesBatch,mat,0,0,0);
2035: return(0);
2036: }
2038: /*@
2039: MatSetLocalToGlobalMapping - Sets a local-to-global numbering for use by
2040: the routine MatSetValuesLocal() to allow users to insert matrix entries
2041: using a local (per-processor) numbering.
2043: Not Collective
2045: Input Parameters:
2046: + x - the matrix
2047: . rmapping - row mapping created with ISLocalToGlobalMappingCreate() or ISLocalToGlobalMappingCreateIS()
2048: - cmapping - column mapping
2050: Level: intermediate
2053: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesLocal()
2054: @*/
2055: PetscErrorCode MatSetLocalToGlobalMapping(Mat x,ISLocalToGlobalMapping rmapping,ISLocalToGlobalMapping cmapping)
2056: {
2065: if (x->ops->setlocaltoglobalmapping) {
2066: (*x->ops->setlocaltoglobalmapping)(x,rmapping,cmapping);
2067: } else {
2068: PetscLayoutSetISLocalToGlobalMapping(x->rmap,rmapping);
2069: PetscLayoutSetISLocalToGlobalMapping(x->cmap,cmapping);
2070: }
2071: return(0);
2072: }
2075: /*@
2076: MatGetLocalToGlobalMapping - Gets the local-to-global numbering set by MatSetLocalToGlobalMapping()
2078: Not Collective
2080: Input Parameters:
2081: . A - the matrix
2083: Output Parameters:
2084: + rmapping - row mapping
2085: - cmapping - column mapping
2087: Level: advanced
2090: .seealso: MatSetValuesLocal()
2091: @*/
2092: PetscErrorCode MatGetLocalToGlobalMapping(Mat A,ISLocalToGlobalMapping *rmapping,ISLocalToGlobalMapping *cmapping)
2093: {
2099: if (rmapping) *rmapping = A->rmap->mapping;
2100: if (cmapping) *cmapping = A->cmap->mapping;
2101: return(0);
2102: }
2104: /*@
2105: MatSetLayouts - Sets the PetscLayout objects for rows and columns of a matrix
2107: Logically Collective on A
2109: Input Parameters:
2110: + A - the matrix
2111: . rmap - row layout
2112: - cmap - column layout
2114: Level: advanced
2116: .seealso: MatCreateVecs(), MatGetLocalToGlobalMapping(), MatGetLayouts()
2117: @*/
2118: PetscErrorCode MatSetLayouts(Mat A,PetscLayout rmap,PetscLayout cmap)
2119: {
2125: PetscLayoutReference(rmap,&A->rmap);
2126: PetscLayoutReference(cmap,&A->cmap);
2127: return(0);
2128: }
2130: /*@
2131: MatGetLayouts - Gets the PetscLayout objects for rows and columns
2133: Not Collective
2135: Input Parameters:
2136: . A - the matrix
2138: Output Parameters:
2139: + rmap - row layout
2140: - cmap - column layout
2142: Level: advanced
2144: .seealso: MatCreateVecs(), MatGetLocalToGlobalMapping(), MatSetLayouts()
2145: @*/
2146: PetscErrorCode MatGetLayouts(Mat A,PetscLayout *rmap,PetscLayout *cmap)
2147: {
2153: if (rmap) *rmap = A->rmap;
2154: if (cmap) *cmap = A->cmap;
2155: return(0);
2156: }
2158: /*@C
2159: MatSetValuesLocal - Inserts or adds values into certain locations of a matrix,
2160: using a local numbering of the nodes.
2162: Not Collective
2164: Input Parameters:
2165: + mat - the matrix
2166: . nrow, irow - number of rows and their local indices
2167: . ncol, icol - number of columns and their local indices
2168: . y - a logically two-dimensional array of values
2169: - addv - either INSERT_VALUES or ADD_VALUES, where
2170: ADD_VALUES adds values to any existing entries, and
2171: INSERT_VALUES replaces existing entries with new values
2173: Notes:
2174: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
2175: MatSetUp() before using this routine
2177: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatSetLocalToGlobalMapping() before using this routine
2179: Calls to MatSetValuesLocal() with the INSERT_VALUES and ADD_VALUES
2180: options cannot be mixed without intervening calls to the assembly
2181: routines.
2183: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
2184: MUST be called after all calls to MatSetValuesLocal() have been completed.
2186: Level: intermediate
2188: Developer Notes:
2189: This is labeled with C so does not automatically generate Fortran stubs and interfaces
2190: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
2192: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetLocalToGlobalMapping(),
2193: MatSetValueLocal(), MatGetValuesLocal()
2194: @*/
2195: PetscErrorCode MatSetValuesLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv)
2196: {
2202: MatCheckPreallocated(mat,1);
2203: if (!nrow || !ncol) return(0); /* no values to insert */
2206: if (mat->insertmode == NOT_SET_VALUES) {
2207: mat->insertmode = addv;
2208: }
2209: else if (PetscUnlikely(mat->insertmode != addv)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
2210: if (PetscDefined(USE_DEBUG)) {
2211: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2212: if (!mat->ops->setvalueslocal && !mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2213: }
2215: if (mat->assembled) {
2216: mat->was_assembled = PETSC_TRUE;
2217: mat->assembled = PETSC_FALSE;
2218: }
2219: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
2220: if (mat->ops->setvalueslocal) {
2221: (*mat->ops->setvalueslocal)(mat,nrow,irow,ncol,icol,y,addv);
2222: } else {
2223: PetscInt buf[8192],*bufr=NULL,*bufc=NULL,*irowm,*icolm;
2224: if ((nrow+ncol) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
2225: irowm = buf; icolm = buf+nrow;
2226: } else {
2227: PetscMalloc2(nrow,&bufr,ncol,&bufc);
2228: irowm = bufr; icolm = bufc;
2229: }
2230: if (!mat->rmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MatSetValuesLocal() cannot proceed without local-to-global row mapping (See MatSetLocalToGlobalMapping()).");
2231: if (!mat->cmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MatSetValuesLocal() cannot proceed without local-to-global column mapping (See MatSetLocalToGlobalMapping()).");
2232: ISLocalToGlobalMappingApply(mat->rmap->mapping,nrow,irow,irowm);
2233: ISLocalToGlobalMappingApply(mat->cmap->mapping,ncol,icol,icolm);
2234: MatSetValues(mat,nrow,irowm,ncol,icolm,y,addv);
2235: PetscFree2(bufr,bufc);
2236: }
2237: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
2238: return(0);
2239: }
2241: /*@C
2242: MatSetValuesBlockedLocal - Inserts or adds values into certain locations of a matrix,
2243: using a local ordering of the nodes a block at a time.
2245: Not Collective
2247: Input Parameters:
2248: + x - the matrix
2249: . nrow, irow - number of rows and their local indices
2250: . ncol, icol - number of columns and their local indices
2251: . y - a logically two-dimensional array of values
2252: - addv - either INSERT_VALUES or ADD_VALUES, where
2253: ADD_VALUES adds values to any existing entries, and
2254: INSERT_VALUES replaces existing entries with new values
2256: Notes:
2257: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
2258: MatSetUp() before using this routine
2260: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatSetBlockSize() and MatSetLocalToGlobalMapping()
2261: before using this routineBefore calling MatSetValuesLocal(), the user must first set the
2263: Calls to MatSetValuesBlockedLocal() with the INSERT_VALUES and ADD_VALUES
2264: options cannot be mixed without intervening calls to the assembly
2265: routines.
2267: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
2268: MUST be called after all calls to MatSetValuesBlockedLocal() have been completed.
2270: Level: intermediate
2272: Developer Notes:
2273: This is labeled with C so does not automatically generate Fortran stubs and interfaces
2274: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
2276: .seealso: MatSetBlockSize(), MatSetLocalToGlobalMapping(), MatAssemblyBegin(), MatAssemblyEnd(),
2277: MatSetValuesLocal(), MatSetValuesBlocked()
2278: @*/
2279: PetscErrorCode MatSetValuesBlockedLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv)
2280: {
2286: MatCheckPreallocated(mat,1);
2287: if (!nrow || !ncol) return(0); /* no values to insert */
2291: if (mat->insertmode == NOT_SET_VALUES) {
2292: mat->insertmode = addv;
2293: } else if (PetscUnlikely(mat->insertmode != addv)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
2294: if (PetscDefined(USE_DEBUG)) {
2295: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2296: if (!mat->ops->setvaluesblockedlocal && !mat->ops->setvaluesblocked && !mat->ops->setvalueslocal && !mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2297: }
2299: if (mat->assembled) {
2300: mat->was_assembled = PETSC_TRUE;
2301: mat->assembled = PETSC_FALSE;
2302: }
2303: if (PetscUnlikelyDebug(mat->rmap->mapping)) { /* Condition on the mapping existing, because MatSetValuesBlockedLocal_IS does not require it to be set. */
2304: PetscInt irbs, rbs;
2305: MatGetBlockSizes(mat, &rbs, NULL);
2306: ISLocalToGlobalMappingGetBlockSize(mat->rmap->mapping,&irbs);
2307: if (rbs != irbs) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Different row block sizes! mat %D, row l2g map %D",rbs,irbs);
2308: }
2309: if (PetscUnlikelyDebug(mat->cmap->mapping)) {
2310: PetscInt icbs, cbs;
2311: MatGetBlockSizes(mat,NULL,&cbs);
2312: ISLocalToGlobalMappingGetBlockSize(mat->cmap->mapping,&icbs);
2313: if (cbs != icbs) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Different col block sizes! mat %D, col l2g map %D",cbs,icbs);
2314: }
2315: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
2316: if (mat->ops->setvaluesblockedlocal) {
2317: (*mat->ops->setvaluesblockedlocal)(mat,nrow,irow,ncol,icol,y,addv);
2318: } else {
2319: PetscInt buf[8192],*bufr=NULL,*bufc=NULL,*irowm,*icolm;
2320: if ((nrow+ncol) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
2321: irowm = buf; icolm = buf + nrow;
2322: } else {
2323: PetscMalloc2(nrow,&bufr,ncol,&bufc);
2324: irowm = bufr; icolm = bufc;
2325: }
2326: ISLocalToGlobalMappingApplyBlock(mat->rmap->mapping,nrow,irow,irowm);
2327: ISLocalToGlobalMappingApplyBlock(mat->cmap->mapping,ncol,icol,icolm);
2328: MatSetValuesBlocked(mat,nrow,irowm,ncol,icolm,y,addv);
2329: PetscFree2(bufr,bufc);
2330: }
2331: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
2332: return(0);
2333: }
2335: /*@
2336: MatMultDiagonalBlock - Computes the matrix-vector product, y = Dx. Where D is defined by the inode or block structure of the diagonal
2338: Collective on Mat
2340: Input Parameters:
2341: + mat - the matrix
2342: - x - the vector to be multiplied
2344: Output Parameters:
2345: . y - the result
2347: Notes:
2348: The vectors x and y cannot be the same. I.e., one cannot
2349: call MatMult(A,y,y).
2351: Level: developer
2353: .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2354: @*/
2355: PetscErrorCode MatMultDiagonalBlock(Mat mat,Vec x,Vec y)
2356: {
2365: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2366: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2367: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2368: MatCheckPreallocated(mat,1);
2370: if (!mat->ops->multdiagonalblock) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s does not have a multiply defined",((PetscObject)mat)->type_name);
2371: (*mat->ops->multdiagonalblock)(mat,x,y);
2372: PetscObjectStateIncrease((PetscObject)y);
2373: return(0);
2374: }
2376: /* --------------------------------------------------------*/
2377: /*@
2378: MatMult - Computes the matrix-vector product, y = Ax.
2380: Neighbor-wise Collective on Mat
2382: Input Parameters:
2383: + mat - the matrix
2384: - x - the vector to be multiplied
2386: Output Parameters:
2387: . y - the result
2389: Notes:
2390: The vectors x and y cannot be the same. I.e., one cannot
2391: call MatMult(A,y,y).
2393: Level: beginner
2395: .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2396: @*/
2397: PetscErrorCode MatMult(Mat mat,Vec x,Vec y)
2398: {
2406: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2407: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2408: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2409: #if !defined(PETSC_HAVE_CONSTRAINTS)
2410: if (mat->cmap->N != x->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
2411: if (mat->rmap->N != y->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap->N,y->map->N);
2412: if (mat->rmap->n != y->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: local dim %D %D",mat->rmap->n,y->map->n);
2413: #endif
2414: VecSetErrorIfLocked(y,3);
2415: if (mat->erroriffailure) {VecValidValues(x,2,PETSC_TRUE);}
2416: MatCheckPreallocated(mat,1);
2418: VecLockReadPush(x);
2419: if (!mat->ops->mult) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s does not have a multiply defined",((PetscObject)mat)->type_name);
2420: PetscLogEventBegin(MAT_Mult,mat,x,y,0);
2421: (*mat->ops->mult)(mat,x,y);
2422: PetscLogEventEnd(MAT_Mult,mat,x,y,0);
2423: if (mat->erroriffailure) {VecValidValues(y,3,PETSC_FALSE);}
2424: VecLockReadPop(x);
2425: return(0);
2426: }
2428: /*@
2429: MatMultTranspose - Computes matrix transpose times a vector y = A^T * x.
2431: Neighbor-wise Collective on Mat
2433: Input Parameters:
2434: + mat - the matrix
2435: - x - the vector to be multiplied
2437: Output Parameters:
2438: . y - the result
2440: Notes:
2441: The vectors x and y cannot be the same. I.e., one cannot
2442: call MatMultTranspose(A,y,y).
2444: For complex numbers this does NOT compute the Hermitian (complex conjugate) transpose multiple,
2445: use MatMultHermitianTranspose()
2447: Level: beginner
2449: .seealso: MatMult(), MatMultAdd(), MatMultTransposeAdd(), MatMultHermitianTranspose(), MatTranspose()
2450: @*/
2451: PetscErrorCode MatMultTranspose(Mat mat,Vec x,Vec y)
2452: {
2453: PetscErrorCode (*op)(Mat,Vec,Vec)=NULL,ierr;
2461: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2462: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2463: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2464: #if !defined(PETSC_HAVE_CONSTRAINTS)
2465: if (mat->rmap->N != x->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N);
2466: if (mat->cmap->N != y->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap->N,y->map->N);
2467: #endif
2468: if (mat->erroriffailure) {VecValidValues(x,2,PETSC_TRUE);}
2469: MatCheckPreallocated(mat,1);
2471: if (!mat->ops->multtranspose) {
2472: if (mat->symmetric && mat->ops->mult) op = mat->ops->mult;
2473: if (!op) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s does not have a multiply transpose defined or is symmetric and does not have a multiply defined",((PetscObject)mat)->type_name);
2474: } else op = mat->ops->multtranspose;
2475: PetscLogEventBegin(MAT_MultTranspose,mat,x,y,0);
2476: VecLockReadPush(x);
2477: (*op)(mat,x,y);
2478: VecLockReadPop(x);
2479: PetscLogEventEnd(MAT_MultTranspose,mat,x,y,0);
2480: PetscObjectStateIncrease((PetscObject)y);
2481: if (mat->erroriffailure) {VecValidValues(y,3,PETSC_FALSE);}
2482: return(0);
2483: }
2485: /*@
2486: MatMultHermitianTranspose - Computes matrix Hermitian transpose times a vector.
2488: Neighbor-wise Collective on Mat
2490: Input Parameters:
2491: + mat - the matrix
2492: - x - the vector to be multilplied
2494: Output Parameters:
2495: . y - the result
2497: Notes:
2498: The vectors x and y cannot be the same. I.e., one cannot
2499: call MatMultHermitianTranspose(A,y,y).
2501: Also called the conjugate transpose, complex conjugate transpose, or adjoint.
2503: For real numbers MatMultTranspose() and MatMultHermitianTranspose() are identical.
2505: Level: beginner
2507: .seealso: MatMult(), MatMultAdd(), MatMultHermitianTransposeAdd(), MatMultTranspose()
2508: @*/
2509: PetscErrorCode MatMultHermitianTranspose(Mat mat,Vec x,Vec y)
2510: {
2519: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2520: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2521: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2522: #if !defined(PETSC_HAVE_CONSTRAINTS)
2523: if (mat->rmap->N != x->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N);
2524: if (mat->cmap->N != y->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap->N,y->map->N);
2525: #endif
2526: MatCheckPreallocated(mat,1);
2528: PetscLogEventBegin(MAT_MultHermitianTranspose,mat,x,y,0);
2529: #if defined(PETSC_USE_COMPLEX)
2530: if (mat->ops->multhermitiantranspose || (mat->hermitian && mat->ops->mult)) {
2531: VecLockReadPush(x);
2532: if (mat->ops->multhermitiantranspose) {
2533: (*mat->ops->multhermitiantranspose)(mat,x,y);
2534: } else {
2535: (*mat->ops->mult)(mat,x,y);
2536: }
2537: VecLockReadPop(x);
2538: } else {
2539: Vec w;
2540: VecDuplicate(x,&w);
2541: VecCopy(x,w);
2542: VecConjugate(w);
2543: MatMultTranspose(mat,w,y);
2544: VecDestroy(&w);
2545: VecConjugate(y);
2546: }
2547: PetscObjectStateIncrease((PetscObject)y);
2548: #else
2549: MatMultTranspose(mat,x,y);
2550: #endif
2551: PetscLogEventEnd(MAT_MultHermitianTranspose,mat,x,y,0);
2552: return(0);
2553: }
2555: /*@
2556: MatMultAdd - Computes v3 = v2 + A * v1.
2558: Neighbor-wise Collective on Mat
2560: Input Parameters:
2561: + mat - the matrix
2562: - v1, v2 - the vectors
2564: Output Parameters:
2565: . v3 - the result
2567: Notes:
2568: The vectors v1 and v3 cannot be the same. I.e., one cannot
2569: call MatMultAdd(A,v1,v2,v1).
2571: Level: beginner
2573: .seealso: MatMultTranspose(), MatMult(), MatMultTransposeAdd()
2574: @*/
2575: PetscErrorCode MatMultAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2576: {
2586: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2587: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2588: if (mat->cmap->N != v1->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->cmap->N,v1->map->N);
2589: /* if (mat->rmap->N != v2->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->rmap->N,v2->map->N);
2590: if (mat->rmap->N != v3->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->rmap->N,v3->map->N); */
2591: if (mat->rmap->n != v3->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: local dim %D %D",mat->rmap->n,v3->map->n);
2592: if (mat->rmap->n != v2->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: local dim %D %D",mat->rmap->n,v2->map->n);
2593: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2594: MatCheckPreallocated(mat,1);
2596: if (!mat->ops->multadd) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"No MatMultAdd() for matrix type %s",((PetscObject)mat)->type_name);
2597: PetscLogEventBegin(MAT_MultAdd,mat,v1,v2,v3);
2598: VecLockReadPush(v1);
2599: (*mat->ops->multadd)(mat,v1,v2,v3);
2600: VecLockReadPop(v1);
2601: PetscLogEventEnd(MAT_MultAdd,mat,v1,v2,v3);
2602: PetscObjectStateIncrease((PetscObject)v3);
2603: return(0);
2604: }
2606: /*@
2607: MatMultTransposeAdd - Computes v3 = v2 + A' * v1.
2609: Neighbor-wise Collective on Mat
2611: Input Parameters:
2612: + mat - the matrix
2613: - v1, v2 - the vectors
2615: Output Parameters:
2616: . v3 - the result
2618: Notes:
2619: The vectors v1 and v3 cannot be the same. I.e., one cannot
2620: call MatMultTransposeAdd(A,v1,v2,v1).
2622: Level: beginner
2624: .seealso: MatMultTranspose(), MatMultAdd(), MatMult()
2625: @*/
2626: PetscErrorCode MatMultTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2627: {
2637: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2638: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2639: if (!mat->ops->multtransposeadd) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2640: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2641: if (mat->rmap->N != v1->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->rmap->N,v1->map->N);
2642: if (mat->cmap->N != v2->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->cmap->N,v2->map->N);
2643: if (mat->cmap->N != v3->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->cmap->N,v3->map->N);
2644: MatCheckPreallocated(mat,1);
2646: PetscLogEventBegin(MAT_MultTransposeAdd,mat,v1,v2,v3);
2647: VecLockReadPush(v1);
2648: (*mat->ops->multtransposeadd)(mat,v1,v2,v3);
2649: VecLockReadPop(v1);
2650: PetscLogEventEnd(MAT_MultTransposeAdd,mat,v1,v2,v3);
2651: PetscObjectStateIncrease((PetscObject)v3);
2652: return(0);
2653: }
2655: /*@
2656: MatMultHermitianTransposeAdd - Computes v3 = v2 + A^H * v1.
2658: Neighbor-wise Collective on Mat
2660: Input Parameters:
2661: + mat - the matrix
2662: - v1, v2 - the vectors
2664: Output Parameters:
2665: . v3 - the result
2667: Notes:
2668: The vectors v1 and v3 cannot be the same. I.e., one cannot
2669: call MatMultHermitianTransposeAdd(A,v1,v2,v1).
2671: Level: beginner
2673: .seealso: MatMultHermitianTranspose(), MatMultTranspose(), MatMultAdd(), MatMult()
2674: @*/
2675: PetscErrorCode MatMultHermitianTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2676: {
2686: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2687: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2688: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2689: if (mat->rmap->N != v1->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->rmap->N,v1->map->N);
2690: if (mat->cmap->N != v2->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->cmap->N,v2->map->N);
2691: if (mat->cmap->N != v3->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->cmap->N,v3->map->N);
2692: MatCheckPreallocated(mat,1);
2694: PetscLogEventBegin(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);
2695: VecLockReadPush(v1);
2696: if (mat->ops->multhermitiantransposeadd) {
2697: (*mat->ops->multhermitiantransposeadd)(mat,v1,v2,v3);
2698: } else {
2699: Vec w,z;
2700: VecDuplicate(v1,&w);
2701: VecCopy(v1,w);
2702: VecConjugate(w);
2703: VecDuplicate(v3,&z);
2704: MatMultTranspose(mat,w,z);
2705: VecDestroy(&w);
2706: VecConjugate(z);
2707: if (v2 != v3) {
2708: VecWAXPY(v3,1.0,v2,z);
2709: } else {
2710: VecAXPY(v3,1.0,z);
2711: }
2712: VecDestroy(&z);
2713: }
2714: VecLockReadPop(v1);
2715: PetscLogEventEnd(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);
2716: PetscObjectStateIncrease((PetscObject)v3);
2717: return(0);
2718: }
2720: /*@
2721: MatMultConstrained - The inner multiplication routine for a
2722: constrained matrix P^T A P.
2724: Neighbor-wise Collective on Mat
2726: Input Parameters:
2727: + mat - the matrix
2728: - x - the vector to be multilplied
2730: Output Parameters:
2731: . y - the result
2733: Notes:
2734: The vectors x and y cannot be the same. I.e., one cannot
2735: call MatMult(A,y,y).
2737: Level: beginner
2739: .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2740: @*/
2741: PetscErrorCode MatMultConstrained(Mat mat,Vec x,Vec y)
2742: {
2749: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2750: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2751: if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2752: if (mat->cmap->N != x->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
2753: if (mat->rmap->N != y->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap->N,y->map->N);
2754: if (mat->rmap->n != y->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: local dim %D %D",mat->rmap->n,y->map->n);
2756: PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);
2757: VecLockReadPush(x);
2758: (*mat->ops->multconstrained)(mat,x,y);
2759: VecLockReadPop(x);
2760: PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);
2761: PetscObjectStateIncrease((PetscObject)y);
2762: return(0);
2763: }
2765: /*@
2766: MatMultTransposeConstrained - The inner multiplication routine for a
2767: constrained matrix P^T A^T P.
2769: Neighbor-wise Collective on Mat
2771: Input Parameters:
2772: + mat - the matrix
2773: - x - the vector to be multilplied
2775: Output Parameters:
2776: . y - the result
2778: Notes:
2779: The vectors x and y cannot be the same. I.e., one cannot
2780: call MatMult(A,y,y).
2782: Level: beginner
2784: .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2785: @*/
2786: PetscErrorCode MatMultTransposeConstrained(Mat mat,Vec x,Vec y)
2787: {
2794: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2795: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2796: if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2797: if (mat->rmap->N != x->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
2798: if (mat->cmap->N != y->map->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap->N,y->map->N);
2800: PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);
2801: (*mat->ops->multtransposeconstrained)(mat,x,y);
2802: PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);
2803: PetscObjectStateIncrease((PetscObject)y);
2804: return(0);
2805: }
2807: /*@C
2808: MatGetFactorType - gets the type of factorization it is
2810: Not Collective
2812: Input Parameters:
2813: . mat - the matrix
2815: Output Parameters:
2816: . t - the type, one of MAT_FACTOR_NONE, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ILU, MAT_FACTOR_ICC,MAT_FACTOR_ILUDT
2818: Level: intermediate
2820: .seealso: MatFactorType, MatGetFactor(), MatSetFactorType()
2821: @*/
2822: PetscErrorCode MatGetFactorType(Mat mat,MatFactorType *t)
2823: {
2828: *t = mat->factortype;
2829: return(0);
2830: }
2832: /*@C
2833: MatSetFactorType - sets the type of factorization it is
2835: Logically Collective on Mat
2837: Input Parameters:
2838: + mat - the matrix
2839: - t - the type, one of MAT_FACTOR_NONE, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ILU, MAT_FACTOR_ICC,MAT_FACTOR_ILUDT
2841: Level: intermediate
2843: .seealso: MatFactorType, MatGetFactor(), MatGetFactorType()
2844: @*/
2845: PetscErrorCode MatSetFactorType(Mat mat, MatFactorType t)
2846: {
2850: mat->factortype = t;
2851: return(0);
2852: }
2854: /* ------------------------------------------------------------*/
2855: /*@C
2856: MatGetInfo - Returns information about matrix storage (number of
2857: nonzeros, memory, etc.).
2859: Collective on Mat if MAT_GLOBAL_MAX or MAT_GLOBAL_SUM is used as the flag
2861: Input Parameters:
2862: . mat - the matrix
2864: Output Parameters:
2865: + flag - flag indicating the type of parameters to be returned
2866: (MAT_LOCAL - local matrix, MAT_GLOBAL_MAX - maximum over all processors,
2867: MAT_GLOBAL_SUM - sum over all processors)
2868: - info - matrix information context
2870: Notes:
2871: The MatInfo context contains a variety of matrix data, including
2872: number of nonzeros allocated and used, number of mallocs during
2873: matrix assembly, etc. Additional information for factored matrices
2874: is provided (such as the fill ratio, number of mallocs during
2875: factorization, etc.). Much of this info is printed to PETSC_STDOUT
2876: when using the runtime options
2877: $ -info -mat_view ::ascii_info
2879: Example for C/C++ Users:
2880: See the file ${PETSC_DIR}/include/petscmat.h for a complete list of
2881: data within the MatInfo context. For example,
2882: .vb
2883: MatInfo info;
2884: Mat A;
2885: double mal, nz_a, nz_u;
2887: MatGetInfo(A,MAT_LOCAL,&info);
2888: mal = info.mallocs;
2889: nz_a = info.nz_allocated;
2890: .ve
2892: Example for Fortran Users:
2893: Fortran users should declare info as a double precision
2894: array of dimension MAT_INFO_SIZE, and then extract the parameters
2895: of interest. See the file ${PETSC_DIR}/include/petsc/finclude/petscmat.h
2896: a complete list of parameter names.
2897: .vb
2898: double precision info(MAT_INFO_SIZE)
2899: double precision mal, nz_a
2900: Mat A
2901: integer ierr
2903: call MatGetInfo(A,MAT_LOCAL,info,ierr)
2904: mal = info(MAT_INFO_MALLOCS)
2905: nz_a = info(MAT_INFO_NZ_ALLOCATED)
2906: .ve
2908: Level: intermediate
2910: Developer Note: fortran interface is not autogenerated as the f90
2911: interface defintion cannot be generated correctly [due to MatInfo]
2913: .seealso: MatStashGetInfo()
2915: @*/
2916: PetscErrorCode MatGetInfo(Mat mat,MatInfoType flag,MatInfo *info)
2917: {
2924: if (!mat->ops->getinfo) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2925: MatCheckPreallocated(mat,1);
2926: (*mat->ops->getinfo)(mat,flag,info);
2927: return(0);
2928: }
2930: /*
2931: This is used by external packages where it is not easy to get the info from the actual
2932: matrix factorization.
2933: */
2934: PetscErrorCode MatGetInfo_External(Mat A,MatInfoType flag,MatInfo *info)
2935: {
2939: PetscMemzero(info,sizeof(MatInfo));
2940: return(0);
2941: }
2943: /* ----------------------------------------------------------*/
2945: /*@C
2946: MatLUFactor - Performs in-place LU factorization of matrix.
2948: Collective on Mat
2950: Input Parameters:
2951: + mat - the matrix
2952: . row - row permutation
2953: . col - column permutation
2954: - info - options for factorization, includes
2955: $ fill - expected fill as ratio of original fill.
2956: $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting)
2957: $ Run with the option -info to determine an optimal value to use
2959: Notes:
2960: Most users should employ the simplified KSP interface for linear solvers
2961: instead of working directly with matrix algebra routines such as this.
2962: See, e.g., KSPCreate().
2964: This changes the state of the matrix to a factored matrix; it cannot be used
2965: for example with MatSetValues() unless one first calls MatSetUnfactored().
2967: Level: developer
2969: .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(),
2970: MatGetOrdering(), MatSetUnfactored(), MatFactorInfo, MatGetFactor()
2972: Developer Note: fortran interface is not autogenerated as the f90
2973: interface defintion cannot be generated correctly [due to MatFactorInfo]
2975: @*/
2976: PetscErrorCode MatLUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info)
2977: {
2979: MatFactorInfo tinfo;
2987: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2988: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2989: if (!mat->ops->lufactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2990: MatCheckPreallocated(mat,1);
2991: if (!info) {
2992: MatFactorInfoInitialize(&tinfo);
2993: info = &tinfo;
2994: }
2996: PetscLogEventBegin(MAT_LUFactor,mat,row,col,0);
2997: (*mat->ops->lufactor)(mat,row,col,info);
2998: PetscLogEventEnd(MAT_LUFactor,mat,row,col,0);
2999: PetscObjectStateIncrease((PetscObject)mat);
3000: return(0);
3001: }
3003: /*@C
3004: MatILUFactor - Performs in-place ILU factorization of matrix.
3006: Collective on Mat
3008: Input Parameters:
3009: + mat - the matrix
3010: . row - row permutation
3011: . col - column permutation
3012: - info - structure containing
3013: $ levels - number of levels of fill.
3014: $ expected fill - as ratio of original fill.
3015: $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices
3016: missing diagonal entries)
3018: Notes:
3019: Probably really in-place only when level of fill is zero, otherwise allocates
3020: new space to store factored matrix and deletes previous memory.
3022: Most users should employ the simplified KSP interface for linear solvers
3023: instead of working directly with matrix algebra routines such as this.
3024: See, e.g., KSPCreate().
3026: Level: developer
3028: .seealso: MatILUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo
3030: Developer Note: fortran interface is not autogenerated as the f90
3031: interface defintion cannot be generated correctly [due to MatFactorInfo]
3033: @*/
3034: PetscErrorCode MatILUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info)
3035: {
3044: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"matrix must be square");
3045: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3046: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3047: if (!mat->ops->ilufactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3048: MatCheckPreallocated(mat,1);
3050: PetscLogEventBegin(MAT_ILUFactor,mat,row,col,0);
3051: (*mat->ops->ilufactor)(mat,row,col,info);
3052: PetscLogEventEnd(MAT_ILUFactor,mat,row,col,0);
3053: PetscObjectStateIncrease((PetscObject)mat);
3054: return(0);
3055: }
3057: /*@C
3058: MatLUFactorSymbolic - Performs symbolic LU factorization of matrix.
3059: Call this routine before calling MatLUFactorNumeric().
3061: Collective on Mat
3063: Input Parameters:
3064: + fact - the factor matrix obtained with MatGetFactor()
3065: . mat - the matrix
3066: . row, col - row and column permutations
3067: - info - options for factorization, includes
3068: $ fill - expected fill as ratio of original fill.
3069: $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting)
3070: $ Run with the option -info to determine an optimal value to use
3073: Notes:
3074: See Users-Manual: ch_mat for additional information about choosing the fill factor for better efficiency.
3076: Most users should employ the simplified KSP interface for linear solvers
3077: instead of working directly with matrix algebra routines such as this.
3078: See, e.g., KSPCreate().
3080: Level: developer
3082: .seealso: MatLUFactor(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo, MatFactorInfoInitialize()
3084: Developer Note: fortran interface is not autogenerated as the f90
3085: interface defintion cannot be generated correctly [due to MatFactorInfo]
3087: @*/
3088: PetscErrorCode MatLUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info)
3089: {
3091: MatFactorInfo tinfo;
3100: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3101: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3102: if (!(fact)->ops->lufactorsymbolic) {
3103: MatSolverType stype;
3104: MatFactorGetSolverType(fact,&stype);
3105: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic LU using solver package %s",((PetscObject)mat)->type_name,stype);
3106: }
3107: MatCheckPreallocated(mat,2);
3108: if (!info) {
3109: MatFactorInfoInitialize(&tinfo);
3110: info = &tinfo;
3111: }
3113: PetscLogEventBegin(MAT_LUFactorSymbolic,mat,row,col,0);
3114: (fact->ops->lufactorsymbolic)(fact,mat,row,col,info);
3115: PetscLogEventEnd(MAT_LUFactorSymbolic,mat,row,col,0);
3116: PetscObjectStateIncrease((PetscObject)fact);
3117: return(0);
3118: }
3120: /*@C
3121: MatLUFactorNumeric - Performs numeric LU factorization of a matrix.
3122: Call this routine after first calling MatLUFactorSymbolic().
3124: Collective on Mat
3126: Input Parameters:
3127: + fact - the factor matrix obtained with MatGetFactor()
3128: . mat - the matrix
3129: - info - options for factorization
3131: Notes:
3132: See MatLUFactor() for in-place factorization. See
3133: MatCholeskyFactorNumeric() for the symmetric, positive definite case.
3135: Most users should employ the simplified KSP interface for linear solvers
3136: instead of working directly with matrix algebra routines such as this.
3137: See, e.g., KSPCreate().
3139: Level: developer
3141: .seealso: MatLUFactorSymbolic(), MatLUFactor(), MatCholeskyFactor()
3143: Developer Note: fortran interface is not autogenerated as the f90
3144: interface defintion cannot be generated correctly [due to MatFactorInfo]
3146: @*/
3147: PetscErrorCode MatLUFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info)
3148: {
3149: MatFactorInfo tinfo;
3157: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3158: if (mat->rmap->N != (fact)->rmap->N || mat->cmap->N != (fact)->cmap->N) SETERRQ4(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Mat fact: global dimensions are different %D should = %D %D should = %D",mat->rmap->N,(fact)->rmap->N,mat->cmap->N,(fact)->cmap->N);
3160: if (!(fact)->ops->lufactornumeric) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s numeric LU",((PetscObject)mat)->type_name);
3161: MatCheckPreallocated(mat,2);
3162: if (!info) {
3163: MatFactorInfoInitialize(&tinfo);
3164: info = &tinfo;
3165: }
3167: PetscLogEventBegin(MAT_LUFactorNumeric,mat,fact,0,0);
3168: (fact->ops->lufactornumeric)(fact,mat,info);
3169: PetscLogEventEnd(MAT_LUFactorNumeric,mat,fact,0,0);
3170: MatViewFromOptions(fact,NULL,"-mat_factor_view");
3171: PetscObjectStateIncrease((PetscObject)fact);
3172: return(0);
3173: }
3175: /*@C
3176: MatCholeskyFactor - Performs in-place Cholesky factorization of a
3177: symmetric matrix.
3179: Collective on Mat
3181: Input Parameters:
3182: + mat - the matrix
3183: . perm - row and column permutations
3184: - f - expected fill as ratio of original fill
3186: Notes:
3187: See MatLUFactor() for the nonsymmetric case. See also
3188: MatCholeskyFactorSymbolic(), and MatCholeskyFactorNumeric().
3190: Most users should employ the simplified KSP interface for linear solvers
3191: instead of working directly with matrix algebra routines such as this.
3192: See, e.g., KSPCreate().
3194: Level: developer
3196: .seealso: MatLUFactor(), MatCholeskyFactorSymbolic(), MatCholeskyFactorNumeric()
3197: MatGetOrdering()
3199: Developer Note: fortran interface is not autogenerated as the f90
3200: interface defintion cannot be generated correctly [due to MatFactorInfo]
3202: @*/
3203: PetscErrorCode MatCholeskyFactor(Mat mat,IS perm,const MatFactorInfo *info)
3204: {
3206: MatFactorInfo tinfo;
3213: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"Matrix must be square");
3214: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3215: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3216: if (!mat->ops->choleskyfactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"In-place factorization for Mat type %s is not supported, try out-of-place factorization. See MatCholeskyFactorSymbolic/Numeric",((PetscObject)mat)->type_name);
3217: MatCheckPreallocated(mat,1);
3218: if (!info) {
3219: MatFactorInfoInitialize(&tinfo);
3220: info = &tinfo;
3221: }
3223: PetscLogEventBegin(MAT_CholeskyFactor,mat,perm,0,0);
3224: (*mat->ops->choleskyfactor)(mat,perm,info);
3225: PetscLogEventEnd(MAT_CholeskyFactor,mat,perm,0,0);
3226: PetscObjectStateIncrease((PetscObject)mat);
3227: return(0);
3228: }
3230: /*@C
3231: MatCholeskyFactorSymbolic - Performs symbolic Cholesky factorization
3232: of a symmetric matrix.
3234: Collective on Mat
3236: Input Parameters:
3237: + fact - the factor matrix obtained with MatGetFactor()
3238: . mat - the matrix
3239: . perm - row and column permutations
3240: - info - options for factorization, includes
3241: $ fill - expected fill as ratio of original fill.
3242: $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting)
3243: $ Run with the option -info to determine an optimal value to use
3245: Notes:
3246: See MatLUFactorSymbolic() for the nonsymmetric case. See also
3247: MatCholeskyFactor() and MatCholeskyFactorNumeric().
3249: Most users should employ the simplified KSP interface for linear solvers
3250: instead of working directly with matrix algebra routines such as this.
3251: See, e.g., KSPCreate().
3253: Level: developer
3255: .seealso: MatLUFactorSymbolic(), MatCholeskyFactor(), MatCholeskyFactorNumeric()
3256: MatGetOrdering()
3258: Developer Note: fortran interface is not autogenerated as the f90
3259: interface defintion cannot be generated correctly [due to MatFactorInfo]
3261: @*/
3262: PetscErrorCode MatCholeskyFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info)
3263: {
3265: MatFactorInfo tinfo;
3273: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"Matrix must be square");
3274: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3275: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3276: if (!(fact)->ops->choleskyfactorsymbolic) {
3277: MatSolverType stype;
3278: MatFactorGetSolverType(fact,&stype);
3279: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s symbolic factor Cholesky using solver package %s",((PetscObject)mat)->type_name,stype);
3280: }
3281: MatCheckPreallocated(mat,2);
3282: if (!info) {
3283: MatFactorInfoInitialize(&tinfo);
3284: info = &tinfo;
3285: }
3287: PetscLogEventBegin(MAT_CholeskyFactorSymbolic,mat,perm,0,0);
3288: (fact->ops->choleskyfactorsymbolic)(fact,mat,perm,info);
3289: PetscLogEventEnd(MAT_CholeskyFactorSymbolic,mat,perm,0,0);
3290: PetscObjectStateIncrease((PetscObject)fact);
3291: return(0);
3292: }
3294: /*@C
3295: MatCholeskyFactorNumeric - Performs numeric Cholesky factorization
3296: of a symmetric matrix. Call this routine after first calling
3297: MatCholeskyFactorSymbolic().
3299: Collective on Mat
3301: Input Parameters:
3302: + fact - the factor matrix obtained with MatGetFactor()
3303: . mat - the initial matrix
3304: . info - options for factorization
3305: - fact - the symbolic factor of mat
3308: Notes:
3309: Most users should employ the simplified KSP interface for linear solvers
3310: instead of working directly with matrix algebra routines such as this.
3311: See, e.g., KSPCreate().
3313: Level: developer
3315: .seealso: MatCholeskyFactorSymbolic(), MatCholeskyFactor(), MatLUFactorNumeric()
3317: Developer Note: fortran interface is not autogenerated as the f90
3318: interface defintion cannot be generated correctly [due to MatFactorInfo]
3320: @*/
3321: PetscErrorCode MatCholeskyFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info)
3322: {
3323: MatFactorInfo tinfo;
3331: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3332: if (!(fact)->ops->choleskyfactornumeric) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s numeric factor Cholesky",((PetscObject)mat)->type_name);
3333: if (mat->rmap->N != (fact)->rmap->N || mat->cmap->N != (fact)->cmap->N) SETERRQ4(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Mat fact: global dim %D should = %D %D should = %D",mat->rmap->N,(fact)->rmap->N,mat->cmap->N,(fact)->cmap->N);
3334: MatCheckPreallocated(mat,2);
3335: if (!info) {
3336: MatFactorInfoInitialize(&tinfo);
3337: info = &tinfo;
3338: }
3340: PetscLogEventBegin(MAT_CholeskyFactorNumeric,mat,fact,0,0);
3341: (fact->ops->choleskyfactornumeric)(fact,mat,info);
3342: PetscLogEventEnd(MAT_CholeskyFactorNumeric,mat,fact,0,0);
3343: MatViewFromOptions(fact,NULL,"-mat_factor_view");
3344: PetscObjectStateIncrease((PetscObject)fact);
3345: return(0);
3346: }
3348: /* ----------------------------------------------------------------*/
3349: /*@
3350: MatSolve - Solves A x = b, given a factored matrix.
3352: Neighbor-wise Collective on Mat
3354: Input Parameters:
3355: + mat - the factored matrix
3356: - b - the right-hand-side vector
3358: Output Parameter:
3359: . x - the result vector
3361: Notes:
3362: The vectors b and x cannot be the same. I.e., one cannot
3363: call MatSolve(A,x,x).
3365: Notes:
3366: Most users should employ the simplified KSP interface for linear solvers
3367: instead of working directly with matrix algebra routines such as this.
3368: See, e.g., KSPCreate().
3370: Level: developer
3372: .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd()
3373: @*/
3374: PetscErrorCode MatSolve(Mat mat,Vec b,Vec x)
3375: {
3385: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3386: if (mat->cmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
3387: if (mat->rmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N);
3388: if (mat->rmap->n != b->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap->n,b->map->n);
3389: if (!mat->rmap->N && !mat->cmap->N) return(0);
3390: MatCheckPreallocated(mat,1);
3392: PetscLogEventBegin(MAT_Solve,mat,b,x,0);
3393: if (mat->factorerrortype) {
3394: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3395: VecSetInf(x);
3396: } else {
3397: if (!mat->ops->solve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3398: (*mat->ops->solve)(mat,b,x);
3399: }
3400: PetscLogEventEnd(MAT_Solve,mat,b,x,0);
3401: PetscObjectStateIncrease((PetscObject)x);
3402: return(0);
3403: }
3405: static PetscErrorCode MatMatSolve_Basic(Mat A,Mat B,Mat X,PetscBool trans)
3406: {
3408: Vec b,x;
3409: PetscInt m,N,i;
3410: PetscScalar *bb,*xx;
3413: MatDenseGetArrayRead(B,(const PetscScalar**)&bb);
3414: MatDenseGetArray(X,&xx);
3415: MatGetLocalSize(B,&m,NULL); /* number local rows */
3416: MatGetSize(B,NULL,&N); /* total columns in dense matrix */
3417: MatCreateVecs(A,&x,&b);
3418: for (i=0; i<N; i++) {
3419: VecPlaceArray(b,bb + i*m);
3420: VecPlaceArray(x,xx + i*m);
3421: if (trans) {
3422: MatSolveTranspose(A,b,x);
3423: } else {
3424: MatSolve(A,b,x);
3425: }
3426: VecResetArray(x);
3427: VecResetArray(b);
3428: }
3429: VecDestroy(&b);
3430: VecDestroy(&x);
3431: MatDenseRestoreArrayRead(B,(const PetscScalar**)&bb);
3432: MatDenseRestoreArray(X,&xx);
3433: return(0);
3434: }
3436: /*@
3437: MatMatSolve - Solves A X = B, given a factored matrix.
3439: Neighbor-wise Collective on Mat
3441: Input Parameters:
3442: + A - the factored matrix
3443: - B - the right-hand-side matrix MATDENSE (or sparse -- when using MUMPS)
3445: Output Parameter:
3446: . X - the result matrix (dense matrix)
3448: Notes:
3449: If B is a MATDENSE matrix then one can call MatMatSolve(A,B,B) except with MKL_CPARDISO;
3450: otherwise, B and X cannot be the same.
3452: Notes:
3453: Most users should usually employ the simplified KSP interface for linear solvers
3454: instead of working directly with matrix algebra routines such as this.
3455: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3456: at a time.
3458: Level: developer
3460: .seealso: MatMatSolveTranspose(), MatLUFactor(), MatCholeskyFactor()
3461: @*/
3462: PetscErrorCode MatMatSolve(Mat A,Mat B,Mat X)
3463: {
3473: if (A->cmap->N != X->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat X: global dim %D %D",A->cmap->N,X->rmap->N);
3474: if (A->rmap->N != B->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim %D %D",A->rmap->N,B->rmap->N);
3475: if (X->cmap->N != B->cmap->N) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Solution matrix must have same number of columns as rhs matrix");
3476: if (!A->rmap->N && !A->cmap->N) return(0);
3477: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3478: MatCheckPreallocated(A,1);
3480: PetscLogEventBegin(MAT_MatSolve,A,B,X,0);
3481: if (!A->ops->matsolve) {
3482: PetscInfo1(A,"Mat type %s using basic MatMatSolve\n",((PetscObject)A)->type_name);
3483: MatMatSolve_Basic(A,B,X,PETSC_FALSE);
3484: } else {
3485: (*A->ops->matsolve)(A,B,X);
3486: }
3487: PetscLogEventEnd(MAT_MatSolve,A,B,X,0);
3488: PetscObjectStateIncrease((PetscObject)X);
3489: return(0);
3490: }
3492: /*@
3493: MatMatSolveTranspose - Solves A^T X = B, given a factored matrix.
3495: Neighbor-wise Collective on Mat
3497: Input Parameters:
3498: + A - the factored matrix
3499: - B - the right-hand-side matrix (dense matrix)
3501: Output Parameter:
3502: . X - the result matrix (dense matrix)
3504: Notes:
3505: The matrices B and X cannot be the same. I.e., one cannot
3506: call MatMatSolveTranspose(A,X,X).
3508: Notes:
3509: Most users should usually employ the simplified KSP interface for linear solvers
3510: instead of working directly with matrix algebra routines such as this.
3511: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3512: at a time.
3514: When using SuperLU_Dist or MUMPS as a parallel solver, PETSc will use their functionality to solve multiple right hand sides simultaneously.
3516: Level: developer
3518: .seealso: MatMatSolve(), MatLUFactor(), MatCholeskyFactor()
3519: @*/
3520: PetscErrorCode MatMatSolveTranspose(Mat A,Mat B,Mat X)
3521: {
3531: if (X == B) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_IDN,"X and B must be different matrices");
3532: if (A->cmap->N != X->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat X: global dim %D %D",A->cmap->N,X->rmap->N);
3533: if (A->rmap->N != B->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim %D %D",A->rmap->N,B->rmap->N);
3534: if (A->rmap->n != B->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat A,Mat B: local dim %D %D",A->rmap->n,B->rmap->n);
3535: if (X->cmap->N < B->cmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Solution matrix must have same number of columns as rhs matrix");
3536: if (!A->rmap->N && !A->cmap->N) return(0);
3537: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3538: MatCheckPreallocated(A,1);
3540: PetscLogEventBegin(MAT_MatSolve,A,B,X,0);
3541: if (!A->ops->matsolvetranspose) {
3542: PetscInfo1(A,"Mat type %s using basic MatMatSolveTranspose\n",((PetscObject)A)->type_name);
3543: MatMatSolve_Basic(A,B,X,PETSC_TRUE);
3544: } else {
3545: (*A->ops->matsolvetranspose)(A,B,X);
3546: }
3547: PetscLogEventEnd(MAT_MatSolve,A,B,X,0);
3548: PetscObjectStateIncrease((PetscObject)X);
3549: return(0);
3550: }
3552: /*@
3553: MatMatTransposeSolve - Solves A X = B^T, given a factored matrix.
3555: Neighbor-wise Collective on Mat
3557: Input Parameters:
3558: + A - the factored matrix
3559: - Bt - the transpose of right-hand-side matrix
3561: Output Parameter:
3562: . X - the result matrix (dense matrix)
3564: Notes:
3565: Most users should usually employ the simplified KSP interface for linear solvers
3566: instead of working directly with matrix algebra routines such as this.
3567: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3568: at a time.
3570: For MUMPS, it only supports centralized sparse compressed column format on the host processor for right hand side matrix. User must create B^T in sparse compressed row format on the host processor and call MatMatTransposeSolve() to implement MUMPS' MatMatSolve().
3572: Level: developer
3574: .seealso: MatMatSolve(), MatMatSolveTranspose(), MatLUFactor(), MatCholeskyFactor()
3575: @*/
3576: PetscErrorCode MatMatTransposeSolve(Mat A,Mat Bt,Mat X)
3577: {
3588: if (X == Bt) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_IDN,"X and B must be different matrices");
3589: if (A->cmap->N != X->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat X: global dim %D %D",A->cmap->N,X->rmap->N);
3590: if (A->rmap->N != Bt->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat Bt: global dim %D %D",A->rmap->N,Bt->cmap->N);
3591: if (X->cmap->N < Bt->rmap->N) SETERRQ(PetscObjectComm((PetscObject)X),PETSC_ERR_ARG_SIZ,"Solution matrix must have same number of columns as row number of the rhs matrix");
3592: if (!A->rmap->N && !A->cmap->N) return(0);
3593: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3594: MatCheckPreallocated(A,1);
3596: if (!A->ops->mattransposesolve) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)A)->type_name);
3597: PetscLogEventBegin(MAT_MatTrSolve,A,Bt,X,0);
3598: (*A->ops->mattransposesolve)(A,Bt,X);
3599: PetscLogEventEnd(MAT_MatTrSolve,A,Bt,X,0);
3600: PetscObjectStateIncrease((PetscObject)X);
3601: return(0);
3602: }
3604: /*@
3605: MatForwardSolve - Solves L x = b, given a factored matrix, A = LU, or
3606: U^T*D^(1/2) x = b, given a factored symmetric matrix, A = U^T*D*U,
3608: Neighbor-wise Collective on Mat
3610: Input Parameters:
3611: + mat - the factored matrix
3612: - b - the right-hand-side vector
3614: Output Parameter:
3615: . x - the result vector
3617: Notes:
3618: MatSolve() should be used for most applications, as it performs
3619: a forward solve followed by a backward solve.
3621: The vectors b and x cannot be the same, i.e., one cannot
3622: call MatForwardSolve(A,x,x).
3624: For matrix in seqsbaij format with block size larger than 1,
3625: the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet.
3626: MatForwardSolve() solves U^T*D y = b, and
3627: MatBackwardSolve() solves U x = y.
3628: Thus they do not provide a symmetric preconditioner.
3630: Most users should employ the simplified KSP interface for linear solvers
3631: instead of working directly with matrix algebra routines such as this.
3632: See, e.g., KSPCreate().
3634: Level: developer
3636: .seealso: MatSolve(), MatBackwardSolve()
3637: @*/
3638: PetscErrorCode MatForwardSolve(Mat mat,Vec b,Vec x)
3639: {
3649: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3650: if (mat->cmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
3651: if (mat->rmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N);
3652: if (mat->rmap->n != b->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap->n,b->map->n);
3653: if (!mat->rmap->N && !mat->cmap->N) return(0);
3654: MatCheckPreallocated(mat,1);
3656: if (!mat->ops->forwardsolve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3657: PetscLogEventBegin(MAT_ForwardSolve,mat,b,x,0);
3658: (*mat->ops->forwardsolve)(mat,b,x);
3659: PetscLogEventEnd(MAT_ForwardSolve,mat,b,x,0);
3660: PetscObjectStateIncrease((PetscObject)x);
3661: return(0);
3662: }
3664: /*@
3665: MatBackwardSolve - Solves U x = b, given a factored matrix, A = LU.
3666: D^(1/2) U x = b, given a factored symmetric matrix, A = U^T*D*U,
3668: Neighbor-wise Collective on Mat
3670: Input Parameters:
3671: + mat - the factored matrix
3672: - b - the right-hand-side vector
3674: Output Parameter:
3675: . x - the result vector
3677: Notes:
3678: MatSolve() should be used for most applications, as it performs
3679: a forward solve followed by a backward solve.
3681: The vectors b and x cannot be the same. I.e., one cannot
3682: call MatBackwardSolve(A,x,x).
3684: For matrix in seqsbaij format with block size larger than 1,
3685: the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet.
3686: MatForwardSolve() solves U^T*D y = b, and
3687: MatBackwardSolve() solves U x = y.
3688: Thus they do not provide a symmetric preconditioner.
3690: Most users should employ the simplified KSP interface for linear solvers
3691: instead of working directly with matrix algebra routines such as this.
3692: See, e.g., KSPCreate().
3694: Level: developer
3696: .seealso: MatSolve(), MatForwardSolve()
3697: @*/
3698: PetscErrorCode MatBackwardSolve(Mat mat,Vec b,Vec x)
3699: {
3709: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3710: if (mat->cmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
3711: if (mat->rmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N);
3712: if (mat->rmap->n != b->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap->n,b->map->n);
3713: if (!mat->rmap->N && !mat->cmap->N) return(0);
3714: MatCheckPreallocated(mat,1);
3716: if (!mat->ops->backwardsolve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3717: PetscLogEventBegin(MAT_BackwardSolve,mat,b,x,0);
3718: (*mat->ops->backwardsolve)(mat,b,x);
3719: PetscLogEventEnd(MAT_BackwardSolve,mat,b,x,0);
3720: PetscObjectStateIncrease((PetscObject)x);
3721: return(0);
3722: }
3724: /*@
3725: MatSolveAdd - Computes x = y + inv(A)*b, given a factored matrix.
3727: Neighbor-wise Collective on Mat
3729: Input Parameters:
3730: + mat - the factored matrix
3731: . b - the right-hand-side vector
3732: - y - the vector to be added to
3734: Output Parameter:
3735: . x - the result vector
3737: Notes:
3738: The vectors b and x cannot be the same. I.e., one cannot
3739: call MatSolveAdd(A,x,y,x).
3741: Most users should employ the simplified KSP interface for linear solvers
3742: instead of working directly with matrix algebra routines such as this.
3743: See, e.g., KSPCreate().
3745: Level: developer
3747: .seealso: MatSolve(), MatSolveTranspose(), MatSolveTransposeAdd()
3748: @*/
3749: PetscErrorCode MatSolveAdd(Mat mat,Vec b,Vec y,Vec x)
3750: {
3751: PetscScalar one = 1.0;
3752: Vec tmp;
3764: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3765: if (mat->cmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
3766: if (mat->rmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N);
3767: if (mat->rmap->N != y->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap->N,y->map->N);
3768: if (mat->rmap->n != b->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap->n,b->map->n);
3769: if (x->map->n != y->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Vec x,Vec y: local dim %D %D",x->map->n,y->map->n);
3770: if (!mat->rmap->N && !mat->cmap->N) return(0);
3771: MatCheckPreallocated(mat,1);
3773: PetscLogEventBegin(MAT_SolveAdd,mat,b,x,y);
3774: if (mat->factorerrortype) {
3775: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3776: VecSetInf(x);
3777: } else if (mat->ops->solveadd) {
3778: (*mat->ops->solveadd)(mat,b,y,x);
3779: } else {
3780: /* do the solve then the add manually */
3781: if (x != y) {
3782: MatSolve(mat,b,x);
3783: VecAXPY(x,one,y);
3784: } else {
3785: VecDuplicate(x,&tmp);
3786: PetscLogObjectParent((PetscObject)mat,(PetscObject)tmp);
3787: VecCopy(x,tmp);
3788: MatSolve(mat,b,x);
3789: VecAXPY(x,one,tmp);
3790: VecDestroy(&tmp);
3791: }
3792: }
3793: PetscLogEventEnd(MAT_SolveAdd,mat,b,x,y);
3794: PetscObjectStateIncrease((PetscObject)x);
3795: return(0);
3796: }
3798: /*@
3799: MatSolveTranspose - Solves A' x = b, given a factored matrix.
3801: Neighbor-wise Collective on Mat
3803: Input Parameters:
3804: + mat - the factored matrix
3805: - b - the right-hand-side vector
3807: Output Parameter:
3808: . x - the result vector
3810: Notes:
3811: The vectors b and x cannot be the same. I.e., one cannot
3812: call MatSolveTranspose(A,x,x).
3814: Most users should employ the simplified KSP interface for linear solvers
3815: instead of working directly with matrix algebra routines such as this.
3816: See, e.g., KSPCreate().
3818: Level: developer
3820: .seealso: MatSolve(), MatSolveAdd(), MatSolveTransposeAdd()
3821: @*/
3822: PetscErrorCode MatSolveTranspose(Mat mat,Vec b,Vec x)
3823: {
3833: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3834: if (mat->rmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N);
3835: if (mat->cmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap->N,b->map->N);
3836: if (!mat->rmap->N && !mat->cmap->N) return(0);
3837: MatCheckPreallocated(mat,1);
3838: PetscLogEventBegin(MAT_SolveTranspose,mat,b,x,0);
3839: if (mat->factorerrortype) {
3840: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3841: VecSetInf(x);
3842: } else {
3843: if (!mat->ops->solvetranspose) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s",((PetscObject)mat)->type_name);
3844: (*mat->ops->solvetranspose)(mat,b,x);
3845: }
3846: PetscLogEventEnd(MAT_SolveTranspose,mat,b,x,0);
3847: PetscObjectStateIncrease((PetscObject)x);
3848: return(0);
3849: }
3851: /*@
3852: MatSolveTransposeAdd - Computes x = y + inv(Transpose(A)) b, given a
3853: factored matrix.
3855: Neighbor-wise Collective on Mat
3857: Input Parameters:
3858: + mat - the factored matrix
3859: . b - the right-hand-side vector
3860: - y - the vector to be added to
3862: Output Parameter:
3863: . x - the result vector
3865: Notes:
3866: The vectors b and x cannot be the same. I.e., one cannot
3867: call MatSolveTransposeAdd(A,x,y,x).
3869: Most users should employ the simplified KSP interface for linear solvers
3870: instead of working directly with matrix algebra routines such as this.
3871: See, e.g., KSPCreate().
3873: Level: developer
3875: .seealso: MatSolve(), MatSolveAdd(), MatSolveTranspose()
3876: @*/
3877: PetscErrorCode MatSolveTransposeAdd(Mat mat,Vec b,Vec y,Vec x)
3878: {
3879: PetscScalar one = 1.0;
3881: Vec tmp;
3892: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3893: if (mat->rmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N);
3894: if (mat->cmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap->N,b->map->N);
3895: if (mat->cmap->N != y->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap->N,y->map->N);
3896: if (x->map->n != y->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Vec x,Vec y: local dim %D %D",x->map->n,y->map->n);
3897: if (!mat->rmap->N && !mat->cmap->N) return(0);
3898: MatCheckPreallocated(mat,1);
3900: PetscLogEventBegin(MAT_SolveTransposeAdd,mat,b,x,y);
3901: if (mat->factorerrortype) {
3902: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3903: VecSetInf(x);
3904: } else if (mat->ops->solvetransposeadd){
3905: (*mat->ops->solvetransposeadd)(mat,b,y,x);
3906: } else {
3907: /* do the solve then the add manually */
3908: if (x != y) {
3909: MatSolveTranspose(mat,b,x);
3910: VecAXPY(x,one,y);
3911: } else {
3912: VecDuplicate(x,&tmp);
3913: PetscLogObjectParent((PetscObject)mat,(PetscObject)tmp);
3914: VecCopy(x,tmp);
3915: MatSolveTranspose(mat,b,x);
3916: VecAXPY(x,one,tmp);
3917: VecDestroy(&tmp);
3918: }
3919: }
3920: PetscLogEventEnd(MAT_SolveTransposeAdd,mat,b,x,y);
3921: PetscObjectStateIncrease((PetscObject)x);
3922: return(0);
3923: }
3924: /* ----------------------------------------------------------------*/
3926: /*@
3927: MatSOR - Computes relaxation (SOR, Gauss-Seidel) sweeps.
3929: Neighbor-wise Collective on Mat
3931: Input Parameters:
3932: + mat - the matrix
3933: . b - the right hand side
3934: . omega - the relaxation factor
3935: . flag - flag indicating the type of SOR (see below)
3936: . shift - diagonal shift
3937: . its - the number of iterations
3938: - lits - the number of local iterations
3940: Output Parameters:
3941: . x - the solution (can contain an initial guess, use option SOR_ZERO_INITIAL_GUESS to indicate no guess)
3943: SOR Flags:
3944: + SOR_FORWARD_SWEEP - forward SOR
3945: . SOR_BACKWARD_SWEEP - backward SOR
3946: . SOR_SYMMETRIC_SWEEP - SSOR (symmetric SOR)
3947: . SOR_LOCAL_FORWARD_SWEEP - local forward SOR
3948: . SOR_LOCAL_BACKWARD_SWEEP - local forward SOR
3949: . SOR_LOCAL_SYMMETRIC_SWEEP - local SSOR
3950: . SOR_APPLY_UPPER, SOR_APPLY_LOWER - applies
3951: upper/lower triangular part of matrix to
3952: vector (with omega)
3953: - SOR_ZERO_INITIAL_GUESS - zero initial guess
3955: Notes:
3956: SOR_LOCAL_FORWARD_SWEEP, SOR_LOCAL_BACKWARD_SWEEP, and
3957: SOR_LOCAL_SYMMETRIC_SWEEP perform separate independent smoothings
3958: on each processor.
3960: Application programmers will not generally use MatSOR() directly,
3961: but instead will employ the KSP/PC interface.
3963: Notes:
3964: for BAIJ, SBAIJ, and AIJ matrices with Inodes this does a block SOR smoothing, otherwise it does a pointwise smoothing
3966: Notes for Advanced Users:
3967: The flags are implemented as bitwise inclusive or operations.
3968: For example, use (SOR_ZERO_INITIAL_GUESS | SOR_SYMMETRIC_SWEEP)
3969: to specify a zero initial guess for SSOR.
3971: Most users should employ the simplified KSP interface for linear solvers
3972: instead of working directly with matrix algebra routines such as this.
3973: See, e.g., KSPCreate().
3975: Vectors x and b CANNOT be the same
3977: Developer Note: We should add block SOR support for AIJ matrices with block size set to great than one and no inodes
3979: Level: developer
3981: @*/
3982: PetscErrorCode MatSOR(Mat mat,Vec b,PetscReal omega,MatSORType flag,PetscReal shift,PetscInt its,PetscInt lits,Vec x)
3983: {
3993: if (!mat->ops->sor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3994: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3995: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3996: if (mat->cmap->N != x->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N);
3997: if (mat->rmap->N != b->map->N) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N);
3998: if (mat->rmap->n != b->map->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap->n,b->map->n);
3999: if (its <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D positive",its);
4000: if (lits <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires local its %D positive",lits);
4001: if (b == x) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_IDN,"b and x vector cannot be the same");
4003: MatCheckPreallocated(mat,1);
4004: PetscLogEventBegin(MAT_SOR,mat,b,x,0);
4005: ierr =(*mat->ops->sor)(mat,b,omega,flag,shift,its,lits,x);
4006: PetscLogEventEnd(MAT_SOR,mat,b,x,0);
4007: PetscObjectStateIncrease((PetscObject)x);
4008: return(0);
4009: }
4011: /*
4012: Default matrix copy routine.
4013: */
4014: PetscErrorCode MatCopy_Basic(Mat A,Mat B,MatStructure str)
4015: {
4016: PetscErrorCode ierr;
4017: PetscInt i,rstart = 0,rend = 0,nz;
4018: const PetscInt *cwork;
4019: const PetscScalar *vwork;
4022: if (B->assembled) {
4023: MatZeroEntries(B);
4024: }
4025: if (str == SAME_NONZERO_PATTERN) {
4026: MatGetOwnershipRange(A,&rstart,&rend);
4027: for (i=rstart; i<rend; i++) {
4028: MatGetRow(A,i,&nz,&cwork,&vwork);
4029: MatSetValues(B,1,&i,nz,cwork,vwork,INSERT_VALUES);
4030: MatRestoreRow(A,i,&nz,&cwork,&vwork);
4031: }
4032: } else {
4033: MatAYPX(B,0.0,A,str);
4034: }
4035: MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
4036: MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);
4037: return(0);
4038: }
4040: /*@
4041: MatCopy - Copies a matrix to another matrix.
4043: Collective on Mat
4045: Input Parameters:
4046: + A - the matrix
4047: - str - SAME_NONZERO_PATTERN or DIFFERENT_NONZERO_PATTERN
4049: Output Parameter:
4050: . B - where the copy is put
4052: Notes:
4053: If you use SAME_NONZERO_PATTERN then the two matrices had better have the
4054: same nonzero pattern or the routine will crash.
4056: MatCopy() copies the matrix entries of a matrix to another existing
4057: matrix (after first zeroing the second matrix). A related routine is
4058: MatConvert(), which first creates a new matrix and then copies the data.
4060: Level: intermediate
4062: .seealso: MatConvert(), MatDuplicate()
4064: @*/
4065: PetscErrorCode MatCopy(Mat A,Mat B,MatStructure str)
4066: {
4068: PetscInt i;
4076: MatCheckPreallocated(B,2);
4077: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4078: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4079: if (A->rmap->N != B->rmap->N || A->cmap->N != B->cmap->N) SETERRQ4(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim (%D,%D) (%D,%D)",A->rmap->N,B->rmap->N,A->cmap->N,B->cmap->N);
4080: MatCheckPreallocated(A,1);
4081: if (A == B) return(0);
4083: PetscLogEventBegin(MAT_Copy,A,B,0,0);
4084: if (A->ops->copy) {
4085: (*A->ops->copy)(A,B,str);
4086: } else { /* generic conversion */
4087: MatCopy_Basic(A,B,str);
4088: }
4090: B->stencil.dim = A->stencil.dim;
4091: B->stencil.noc = A->stencil.noc;
4092: for (i=0; i<=A->stencil.dim; i++) {
4093: B->stencil.dims[i] = A->stencil.dims[i];
4094: B->stencil.starts[i] = A->stencil.starts[i];
4095: }
4097: PetscLogEventEnd(MAT_Copy,A,B,0,0);
4098: PetscObjectStateIncrease((PetscObject)B);
4099: return(0);
4100: }
4102: /*@C
4103: MatConvert - Converts a matrix to another matrix, either of the same
4104: or different type.
4106: Collective on Mat
4108: Input Parameters:
4109: + mat - the matrix
4110: . newtype - new matrix type. Use MATSAME to create a new matrix of the
4111: same type as the original matrix.
4112: - reuse - denotes if the destination matrix is to be created or reused.
4113: Use MAT_INPLACE_MATRIX for inplace conversion (that is when you want the input mat to be changed to contain the matrix in the new format), otherwise use
4114: MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX (can only be used after the first call was made with MAT_INITIAL_MATRIX, causes the matrix space in M to be reused).
4116: Output Parameter:
4117: . M - pointer to place new matrix
4119: Notes:
4120: MatConvert() first creates a new matrix and then copies the data from
4121: the first matrix. A related routine is MatCopy(), which copies the matrix
4122: entries of one matrix to another already existing matrix context.
4124: Cannot be used to convert a sequential matrix to parallel or parallel to sequential,
4125: the MPI communicator of the generated matrix is always the same as the communicator
4126: of the input matrix.
4128: Level: intermediate
4130: .seealso: MatCopy(), MatDuplicate()
4131: @*/
4132: PetscErrorCode MatConvert(Mat mat, MatType newtype,MatReuse reuse,Mat *M)
4133: {
4135: PetscBool sametype,issame,flg,issymmetric,ishermitian;
4136: char convname[256],mtype[256];
4137: Mat B;
4143: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4144: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4145: MatCheckPreallocated(mat,1);
4147: PetscOptionsGetString(((PetscObject)mat)->options,((PetscObject)mat)->prefix,"-matconvert_type",mtype,sizeof(mtype),&flg);
4148: if (flg) newtype = mtype;
4150: PetscObjectTypeCompare((PetscObject)mat,newtype,&sametype);
4151: PetscStrcmp(newtype,"same",&issame);
4152: if ((reuse == MAT_INPLACE_MATRIX) && (mat != *M)) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"MAT_INPLACE_MATRIX requires same input and output matrix");
4153: if ((reuse == MAT_REUSE_MATRIX) && (mat == *M)) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"MAT_REUSE_MATRIX means reuse matrix in final argument, perhaps you mean MAT_INPLACE_MATRIX");
4155: if ((reuse == MAT_INPLACE_MATRIX) && (issame || sametype)) {
4156: PetscInfo3(mat,"Early return for inplace %s %d %d\n",((PetscObject)mat)->type_name,sametype,issame);
4157: return(0);
4158: }
4160: /* Cache Mat options because some converter use MatHeaderReplace */
4161: issymmetric = mat->symmetric;
4162: ishermitian = mat->hermitian;
4164: if ((sametype || issame) && (reuse==MAT_INITIAL_MATRIX) && mat->ops->duplicate) {
4165: PetscInfo3(mat,"Calling duplicate for initial matrix %s %d %d\n",((PetscObject)mat)->type_name,sametype,issame);
4166: (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);
4167: } else {
4168: PetscErrorCode (*conv)(Mat, MatType,MatReuse,Mat*)=NULL;
4169: const char *prefix[3] = {"seq","mpi",""};
4170: PetscInt i;
4171: /*
4172: Order of precedence:
4173: 0) See if newtype is a superclass of the current matrix.
4174: 1) See if a specialized converter is known to the current matrix.
4175: 2) See if a specialized converter is known to the desired matrix class.
4176: 3) See if a good general converter is registered for the desired class
4177: (as of 6/27/03 only MATMPIADJ falls into this category).
4178: 4) See if a good general converter is known for the current matrix.
4179: 5) Use a really basic converter.
4180: */
4182: /* 0) See if newtype is a superclass of the current matrix.
4183: i.e mat is mpiaij and newtype is aij */
4184: for (i=0; i<2; i++) {
4185: PetscStrncpy(convname,prefix[i],sizeof(convname));
4186: PetscStrlcat(convname,newtype,sizeof(convname));
4187: PetscStrcmp(convname,((PetscObject)mat)->type_name,&flg);
4188: PetscInfo3(mat,"Check superclass %s %s -> %d\n",convname,((PetscObject)mat)->type_name,flg);
4189: if (flg) {
4190: if (reuse == MAT_INPLACE_MATRIX) {
4191: PetscInfo(mat,"Early return\n");
4192: return(0);
4193: } else if (reuse == MAT_INITIAL_MATRIX && mat->ops->duplicate) {
4194: PetscInfo(mat,"Calling MatDuplicate\n");
4195: (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);
4196: return(0);
4197: } else if (reuse == MAT_REUSE_MATRIX && mat->ops->copy) {
4198: PetscInfo(mat,"Calling MatCopy\n");
4199: MatCopy(mat,*M,SAME_NONZERO_PATTERN);
4200: return(0);
4201: }
4202: }
4203: }
4204: /* 1) See if a specialized converter is known to the current matrix and the desired class */
4205: for (i=0; i<3; i++) {
4206: PetscStrncpy(convname,"MatConvert_",sizeof(convname));
4207: PetscStrlcat(convname,((PetscObject)mat)->type_name,sizeof(convname));
4208: PetscStrlcat(convname,"_",sizeof(convname));
4209: PetscStrlcat(convname,prefix[i],sizeof(convname));
4210: PetscStrlcat(convname,issame ? ((PetscObject)mat)->type_name : newtype,sizeof(convname));
4211: PetscStrlcat(convname,"_C",sizeof(convname));
4212: PetscObjectQueryFunction((PetscObject)mat,convname,&conv);
4213: PetscInfo3(mat,"Check specialized (1) %s (%s) -> %d\n",convname,((PetscObject)mat)->type_name,!!conv);
4214: if (conv) goto foundconv;
4215: }
4217: /* 2) See if a specialized converter is known to the desired matrix class. */
4218: MatCreate(PetscObjectComm((PetscObject)mat),&B);
4219: MatSetSizes(B,mat->rmap->n,mat->cmap->n,mat->rmap->N,mat->cmap->N);
4220: MatSetType(B,newtype);
4221: for (i=0; i<3; i++) {
4222: PetscStrncpy(convname,"MatConvert_",sizeof(convname));
4223: PetscStrlcat(convname,((PetscObject)mat)->type_name,sizeof(convname));
4224: PetscStrlcat(convname,"_",sizeof(convname));
4225: PetscStrlcat(convname,prefix[i],sizeof(convname));
4226: PetscStrlcat(convname,newtype,sizeof(convname));
4227: PetscStrlcat(convname,"_C",sizeof(convname));
4228: PetscObjectQueryFunction((PetscObject)B,convname,&conv);
4229: PetscInfo3(mat,"Check specialized (2) %s (%s) -> %d\n",convname,((PetscObject)B)->type_name,!!conv);
4230: if (conv) {
4231: MatDestroy(&B);
4232: goto foundconv;
4233: }
4234: }
4236: /* 3) See if a good general converter is registered for the desired class */
4237: conv = B->ops->convertfrom;
4238: PetscInfo2(mat,"Check convertfrom (%s) -> %d\n",((PetscObject)B)->type_name,!!conv);
4239: MatDestroy(&B);
4240: if (conv) goto foundconv;
4242: /* 4) See if a good general converter is known for the current matrix */
4243: if (mat->ops->convert) conv = mat->ops->convert;
4245: PetscInfo2(mat,"Check general convert (%s) -> %d\n",((PetscObject)mat)->type_name,!!conv);
4246: if (conv) goto foundconv;
4248: /* 5) Use a really basic converter. */
4249: PetscInfo(mat,"Using MatConvert_Basic\n");
4250: conv = MatConvert_Basic;
4252: foundconv:
4253: PetscLogEventBegin(MAT_Convert,mat,0,0,0);
4254: (*conv)(mat,newtype,reuse,M);
4255: if (mat->rmap->mapping && mat->cmap->mapping && !(*M)->rmap->mapping && !(*M)->cmap->mapping) {
4256: /* the block sizes must be same if the mappings are copied over */
4257: (*M)->rmap->bs = mat->rmap->bs;
4258: (*M)->cmap->bs = mat->cmap->bs;
4259: PetscObjectReference((PetscObject)mat->rmap->mapping);
4260: PetscObjectReference((PetscObject)mat->cmap->mapping);
4261: (*M)->rmap->mapping = mat->rmap->mapping;
4262: (*M)->cmap->mapping = mat->cmap->mapping;
4263: }
4264: (*M)->stencil.dim = mat->stencil.dim;
4265: (*M)->stencil.noc = mat->stencil.noc;
4266: for (i=0; i<=mat->stencil.dim; i++) {
4267: (*M)->stencil.dims[i] = mat->stencil.dims[i];
4268: (*M)->stencil.starts[i] = mat->stencil.starts[i];
4269: }
4270: PetscLogEventEnd(MAT_Convert,mat,0,0,0);
4271: }
4272: PetscObjectStateIncrease((PetscObject)*M);
4274: /* Copy Mat options */
4275: if (issymmetric) {
4276: MatSetOption(*M,MAT_SYMMETRIC,PETSC_TRUE);
4277: }
4278: if (ishermitian) {
4279: MatSetOption(*M,MAT_HERMITIAN,PETSC_TRUE);
4280: }
4281: return(0);
4282: }
4284: /*@C
4285: MatFactorGetSolverType - Returns name of the package providing the factorization routines
4287: Not Collective
4289: Input Parameter:
4290: . mat - the matrix, must be a factored matrix
4292: Output Parameter:
4293: . type - the string name of the package (do not free this string)
4295: Notes:
4296: In Fortran you pass in a empty string and the package name will be copied into it.
4297: (Make sure the string is long enough)
4299: Level: intermediate
4301: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatGetFactor()
4302: @*/
4303: PetscErrorCode MatFactorGetSolverType(Mat mat, MatSolverType *type)
4304: {
4305: PetscErrorCode ierr, (*conv)(Mat,MatSolverType*);
4310: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
4311: PetscObjectQueryFunction((PetscObject)mat,"MatFactorGetSolverType_C",&conv);
4312: if (!conv) {
4313: *type = MATSOLVERPETSC;
4314: } else {
4315: (*conv)(mat,type);
4316: }
4317: return(0);
4318: }
4320: typedef struct _MatSolverTypeForSpecifcType* MatSolverTypeForSpecifcType;
4321: struct _MatSolverTypeForSpecifcType {
4322: MatType mtype;
4323: PetscErrorCode (*createfactor[4])(Mat,MatFactorType,Mat*);
4324: MatSolverTypeForSpecifcType next;
4325: };
4327: typedef struct _MatSolverTypeHolder* MatSolverTypeHolder;
4328: struct _MatSolverTypeHolder {
4329: char *name;
4330: MatSolverTypeForSpecifcType handlers;
4331: MatSolverTypeHolder next;
4332: };
4334: static MatSolverTypeHolder MatSolverTypeHolders = NULL;
4336: /*@C
4337: MatSolveTypeRegister - Registers a MatSolverType that works for a particular matrix type
4339: Input Parameters:
4340: + package - name of the package, for example petsc or superlu
4341: . mtype - the matrix type that works with this package
4342: . ftype - the type of factorization supported by the package
4343: - createfactor - routine that will create the factored matrix ready to be used
4345: Level: intermediate
4347: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatGetFactor()
4348: @*/
4349: PetscErrorCode MatSolverTypeRegister(MatSolverType package,MatType mtype,MatFactorType ftype,PetscErrorCode (*createfactor)(Mat,MatFactorType,Mat*))
4350: {
4351: PetscErrorCode ierr;
4352: MatSolverTypeHolder next = MatSolverTypeHolders,prev = NULL;
4353: PetscBool flg;
4354: MatSolverTypeForSpecifcType inext,iprev = NULL;
4357: MatInitializePackage();
4358: if (!next) {
4359: PetscNew(&MatSolverTypeHolders);
4360: PetscStrallocpy(package,&MatSolverTypeHolders->name);
4361: PetscNew(&MatSolverTypeHolders->handlers);
4362: PetscStrallocpy(mtype,(char **)&MatSolverTypeHolders->handlers->mtype);
4363: MatSolverTypeHolders->handlers->createfactor[(int)ftype-1] = createfactor;
4364: return(0);
4365: }
4366: while (next) {
4367: PetscStrcasecmp(package,next->name,&flg);
4368: if (flg) {
4369: if (!next->handlers) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"MatSolverTypeHolder is missing handlers");
4370: inext = next->handlers;
4371: while (inext) {
4372: PetscStrcasecmp(mtype,inext->mtype,&flg);
4373: if (flg) {
4374: inext->createfactor[(int)ftype-1] = createfactor;
4375: return(0);
4376: }
4377: iprev = inext;
4378: inext = inext->next;
4379: }
4380: PetscNew(&iprev->next);
4381: PetscStrallocpy(mtype,(char **)&iprev->next->mtype);
4382: iprev->next->createfactor[(int)ftype-1] = createfactor;
4383: return(0);
4384: }
4385: prev = next;
4386: next = next->next;
4387: }
4388: PetscNew(&prev->next);
4389: PetscStrallocpy(package,&prev->next->name);
4390: PetscNew(&prev->next->handlers);
4391: PetscStrallocpy(mtype,(char **)&prev->next->handlers->mtype);
4392: prev->next->handlers->createfactor[(int)ftype-1] = createfactor;
4393: return(0);
4394: }
4396: /*@C
4397: MatSolveTypeGet - Gets the function that creates the factor matrix if it exist
4399: Input Parameters:
4400: + type - name of the package, for example petsc or superlu
4401: . ftype - the type of factorization supported by the type
4402: - mtype - the matrix type that works with this type
4404: Output Parameters:
4405: + foundtype - PETSC_TRUE if the type was registered
4406: . foundmtype - PETSC_TRUE if the type supports the requested mtype
4407: - createfactor - routine that will create the factored matrix ready to be used or NULL if not found
4409: Level: intermediate
4411: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatSolvePackageRegister), MatGetFactor()
4412: @*/
4413: PetscErrorCode MatSolverTypeGet(MatSolverType type,MatType mtype,MatFactorType ftype,PetscBool *foundtype,PetscBool *foundmtype,PetscErrorCode (**createfactor)(Mat,MatFactorType,Mat*))
4414: {
4415: PetscErrorCode ierr;
4416: MatSolverTypeHolder next = MatSolverTypeHolders;
4417: PetscBool flg;
4418: MatSolverTypeForSpecifcType inext;
4421: if (foundtype) *foundtype = PETSC_FALSE;
4422: if (foundmtype) *foundmtype = PETSC_FALSE;
4423: if (createfactor) *createfactor = NULL;
4425: if (type) {
4426: while (next) {
4427: PetscStrcasecmp(type,next->name,&flg);
4428: if (flg) {
4429: if (foundtype) *foundtype = PETSC_TRUE;
4430: inext = next->handlers;
4431: while (inext) {
4432: PetscStrbeginswith(mtype,inext->mtype,&flg);
4433: if (flg) {
4434: if (foundmtype) *foundmtype = PETSC_TRUE;
4435: if (createfactor) *createfactor = inext->createfactor[(int)ftype-1];
4436: return(0);
4437: }
4438: inext = inext->next;
4439: }
4440: }
4441: next = next->next;
4442: }
4443: } else {
4444: while (next) {
4445: inext = next->handlers;
4446: while (inext) {
4447: PetscStrbeginswith(mtype,inext->mtype,&flg);
4448: if (flg && inext->createfactor[(int)ftype-1]) {
4449: if (foundtype) *foundtype = PETSC_TRUE;
4450: if (foundmtype) *foundmtype = PETSC_TRUE;
4451: if (createfactor) *createfactor = inext->createfactor[(int)ftype-1];
4452: return(0);
4453: }
4454: inext = inext->next;
4455: }
4456: next = next->next;
4457: }
4458: }
4459: return(0);
4460: }
4462: PetscErrorCode MatSolverTypeDestroy(void)
4463: {
4464: PetscErrorCode ierr;
4465: MatSolverTypeHolder next = MatSolverTypeHolders,prev;
4466: MatSolverTypeForSpecifcType inext,iprev;
4469: while (next) {
4470: PetscFree(next->name);
4471: inext = next->handlers;
4472: while (inext) {
4473: PetscFree(inext->mtype);
4474: iprev = inext;
4475: inext = inext->next;
4476: PetscFree(iprev);
4477: }
4478: prev = next;
4479: next = next->next;
4480: PetscFree(prev);
4481: }
4482: MatSolverTypeHolders = NULL;
4483: return(0);
4484: }
4486: /*@C
4487: MatFactorGetUseOrdering - Indicates if the factorization uses the ordering provided in MatLUFactorSymbolic(), MatCholeskyFactorSymbolic()
4489: Logically Collective on Mat
4491: Input Parameters:
4492: . mat - the matrix
4494: Output Parameters:
4495: . flg - PETSC_TRUE if uses the ordering
4497: Notes:
4498: Most internal PETSc factorizations use the ordering past to the factorization routine but external
4499: packages do no, thus we want to skip the ordering when it is not needed.
4501: Level: developer
4503: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatGetFactor(), MatLUFactorSymbolic(), MatCholeskyFactorSymbolic()
4504: @*/
4505: PetscErrorCode MatFactorGetUseOrdering(Mat mat, PetscBool *flg)
4506: {
4508: *flg = mat->useordering;
4509: return(0);
4510: }
4512: /*@C
4513: MatGetFactor - Returns a matrix suitable to calls to MatXXFactorSymbolic()
4515: Collective on Mat
4517: Input Parameters:
4518: + mat - the matrix
4519: . type - name of solver type, for example, superlu, petsc (to use PETSc's default)
4520: - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU,
4522: Output Parameters:
4523: . f - the factor matrix used with MatXXFactorSymbolic() calls
4525: Notes:
4526: Some PETSc matrix formats have alternative solvers available that are contained in alternative packages
4527: such as pastix, superlu, mumps etc.
4529: PETSc must have been ./configure to use the external solver, using the option --download-package
4531: Developer Notes:
4532: This should actually be called MatCreateFactor() since it creates a new factor object
4534: Level: intermediate
4536: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatFactorGetUseOrdering(), MatSolverTypeRegister()
4537: @*/
4538: PetscErrorCode MatGetFactor(Mat mat, MatSolverType type,MatFactorType ftype,Mat *f)
4539: {
4540: PetscErrorCode ierr,(*conv)(Mat,MatFactorType,Mat*);
4541: PetscBool foundtype,foundmtype;
4547: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4548: MatCheckPreallocated(mat,1);
4550: MatSolverTypeGet(type,((PetscObject)mat)->type_name,ftype,&foundtype,&foundmtype,&conv);
4551: if (!foundtype) {
4552: if (type) {
4553: SETERRQ4(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"Could not locate solver type %s for factorization type %s and matrix type %s. Perhaps you must ./configure with --download-%s",type,MatFactorTypes[ftype],((PetscObject)mat)->type_name,type);
4554: } else {
4555: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"Could not locate a solver type for factorization type %s and matrix type %s.",MatFactorTypes[ftype],((PetscObject)mat)->type_name);
4556: }
4557: }
4558: if (!foundmtype) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"MatSolverType %s does not support matrix type %s",type,((PetscObject)mat)->type_name);
4559: if (!conv) SETERRQ3(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"MatSolverType %s does not support factorization type %s for matrix type %s",type,MatFactorTypes[ftype],((PetscObject)mat)->type_name);
4561: (*conv)(mat,ftype,f);
4562: return(0);
4563: }
4565: /*@C
4566: MatGetFactorAvailable - Returns a a flag if matrix supports particular type and factor type
4568: Not Collective
4570: Input Parameters:
4571: + mat - the matrix
4572: . type - name of solver type, for example, superlu, petsc (to use PETSc's default)
4573: - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU,
4575: Output Parameter:
4576: . flg - PETSC_TRUE if the factorization is available
4578: Notes:
4579: Some PETSc matrix formats have alternative solvers available that are contained in alternative packages
4580: such as pastix, superlu, mumps etc.
4582: PETSc must have been ./configure to use the external solver, using the option --download-package
4584: Developer Notes:
4585: This should actually be called MatCreateFactorAvailable() since MatGetFactor() creates a new factor object
4587: Level: intermediate
4589: .seealso: MatCopy(), MatDuplicate(), MatGetFactor(), MatSolverTypeRegister()
4590: @*/
4591: PetscErrorCode MatGetFactorAvailable(Mat mat, MatSolverType type,MatFactorType ftype,PetscBool *flg)
4592: {
4593: PetscErrorCode ierr, (*gconv)(Mat,MatFactorType,Mat*);
4599: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4600: MatCheckPreallocated(mat,1);
4602: *flg = PETSC_FALSE;
4603: MatSolverTypeGet(type,((PetscObject)mat)->type_name,ftype,NULL,NULL,&gconv);
4604: if (gconv) {
4605: *flg = PETSC_TRUE;
4606: }
4607: return(0);
4608: }
4610: #include <petscdmtypes.h>
4612: /*@
4613: MatDuplicate - Duplicates a matrix including the non-zero structure.
4615: Collective on Mat
4617: Input Parameters:
4618: + mat - the matrix
4619: - op - One of MAT_DO_NOT_COPY_VALUES, MAT_COPY_VALUES, or MAT_SHARE_NONZERO_PATTERN.
4620: See the manual page for MatDuplicateOption for an explanation of these options.
4622: Output Parameter:
4623: . M - pointer to place new matrix
4625: Level: intermediate
4627: Notes:
4628: You cannot change the nonzero pattern for the parent or child matrix if you use MAT_SHARE_NONZERO_PATTERN.
4629: When original mat is a product of matrix operation, e.g., an output of MatMatMult() or MatCreateSubMatrix(), only the simple matrix data structure of mat is duplicated and the internal data structures created for the reuse of previous matrix operations are not duplicated. User should not use MatDuplicate() to create new matrix M if M is intended to be reused as the product of matrix operation.
4631: .seealso: MatCopy(), MatConvert(), MatDuplicateOption
4632: @*/
4633: PetscErrorCode MatDuplicate(Mat mat,MatDuplicateOption op,Mat *M)
4634: {
4636: Mat B;
4637: PetscInt i;
4638: DM dm;
4639: void (*viewf)(void);
4645: if (op == MAT_COPY_VALUES && !mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MAT_COPY_VALUES not allowed for unassembled matrix");
4646: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4647: MatCheckPreallocated(mat,1);
4649: *M = NULL;
4650: if (!mat->ops->duplicate) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not written for matrix type %s\n",((PetscObject)mat)->type_name);
4651: PetscLogEventBegin(MAT_Convert,mat,0,0,0);
4652: (*mat->ops->duplicate)(mat,op,M);
4653: B = *M;
4655: MatGetOperation(mat,MATOP_VIEW,&viewf);
4656: if (viewf) {
4657: MatSetOperation(B,MATOP_VIEW,viewf);
4658: }
4660: B->stencil.dim = mat->stencil.dim;
4661: B->stencil.noc = mat->stencil.noc;
4662: for (i=0; i<=mat->stencil.dim; i++) {
4663: B->stencil.dims[i] = mat->stencil.dims[i];
4664: B->stencil.starts[i] = mat->stencil.starts[i];
4665: }
4667: B->nooffproczerorows = mat->nooffproczerorows;
4668: B->nooffprocentries = mat->nooffprocentries;
4670: PetscObjectQuery((PetscObject) mat, "__PETSc_dm", (PetscObject*) &dm);
4671: if (dm) {
4672: PetscObjectCompose((PetscObject) B, "__PETSc_dm", (PetscObject) dm);
4673: }
4674: PetscLogEventEnd(MAT_Convert,mat,0,0,0);
4675: PetscObjectStateIncrease((PetscObject)B);
4676: return(0);
4677: }
4679: /*@
4680: MatGetDiagonal - Gets the diagonal of a matrix.
4682: Logically Collective on Mat
4684: Input Parameters:
4685: + mat - the matrix
4686: - v - the vector for storing the diagonal
4688: Output Parameter:
4689: . v - the diagonal of the matrix
4691: Level: intermediate
4693: Note:
4694: Currently only correct in parallel for square matrices.
4696: .seealso: MatGetRow(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs()
4697: @*/
4698: PetscErrorCode MatGetDiagonal(Mat mat,Vec v)
4699: {
4706: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4707: if (!mat->ops->getdiagonal) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4708: MatCheckPreallocated(mat,1);
4710: (*mat->ops->getdiagonal)(mat,v);
4711: PetscObjectStateIncrease((PetscObject)v);
4712: return(0);
4713: }
4715: /*@C
4716: MatGetRowMin - Gets the minimum value (of the real part) of each
4717: row of the matrix
4719: Logically Collective on Mat
4721: Input Parameters:
4722: . mat - the matrix
4724: Output Parameter:
4725: + v - the vector for storing the maximums
4726: - idx - the indices of the column found for each row (optional)
4728: Level: intermediate
4730: Notes:
4731: The result of this call are the same as if one converted the matrix to dense format
4732: and found the minimum value in each row (i.e. the implicit zeros are counted as zeros).
4734: This code is only implemented for a couple of matrix formats.
4736: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs(),
4737: MatGetRowMax()
4738: @*/
4739: PetscErrorCode MatGetRowMin(Mat mat,Vec v,PetscInt idx[])
4740: {
4747: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4749: if (!mat->cmap->N) {
4750: VecSet(v,PETSC_MAX_REAL);
4751: if (idx) {
4752: PetscInt i,m = mat->rmap->n;
4753: for (i=0; i<m; i++) idx[i] = -1;
4754: }
4755: } else {
4756: if (!mat->ops->getrowmin) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4757: MatCheckPreallocated(mat,1);
4758: }
4759: (*mat->ops->getrowmin)(mat,v,idx);
4760: PetscObjectStateIncrease((PetscObject)v);
4761: return(0);
4762: }
4764: /*@C
4765: MatGetRowMinAbs - Gets the minimum value (in absolute value) of each
4766: row of the matrix
4768: Logically Collective on Mat
4770: Input Parameters:
4771: . mat - the matrix
4773: Output Parameter:
4774: + v - the vector for storing the minimums
4775: - idx - the indices of the column found for each row (or NULL if not needed)
4777: Level: intermediate
4779: Notes:
4780: if a row is completely empty or has only 0.0 values then the idx[] value for that
4781: row is 0 (the first column).
4783: This code is only implemented for a couple of matrix formats.
4785: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMaxAbs(), MatGetRowMin()
4786: @*/
4787: PetscErrorCode MatGetRowMinAbs(Mat mat,Vec v,PetscInt idx[])
4788: {
4795: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4796: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4798: if (!mat->cmap->N) {
4799: VecSet(v,0.0);
4800: if (idx) {
4801: PetscInt i,m = mat->rmap->n;
4802: for (i=0; i<m; i++) idx[i] = -1;
4803: }
4804: } else {
4805: if (!mat->ops->getrowminabs) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4806: MatCheckPreallocated(mat,1);
4807: if (idx) {PetscArrayzero(idx,mat->rmap->n);}
4808: (*mat->ops->getrowminabs)(mat,v,idx);
4809: }
4810: PetscObjectStateIncrease((PetscObject)v);
4811: return(0);
4812: }
4814: /*@C
4815: MatGetRowMax - Gets the maximum value (of the real part) of each
4816: row of the matrix
4818: Logically Collective on Mat
4820: Input Parameters:
4821: . mat - the matrix
4823: Output Parameter:
4824: + v - the vector for storing the maximums
4825: - idx - the indices of the column found for each row (optional)
4827: Level: intermediate
4829: Notes:
4830: The result of this call are the same as if one converted the matrix to dense format
4831: and found the minimum value in each row (i.e. the implicit zeros are counted as zeros).
4833: This code is only implemented for a couple of matrix formats.
4835: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs(), MatGetRowMin()
4836: @*/
4837: PetscErrorCode MatGetRowMax(Mat mat,Vec v,PetscInt idx[])
4838: {
4845: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4847: if (!mat->cmap->N) {
4848: VecSet(v,PETSC_MIN_REAL);
4849: if (idx) {
4850: PetscInt i,m = mat->rmap->n;
4851: for (i=0; i<m; i++) idx[i] = -1;
4852: }
4853: } else {
4854: if (!mat->ops->getrowmax) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4855: MatCheckPreallocated(mat,1);
4856: (*mat->ops->getrowmax)(mat,v,idx);
4857: }
4858: PetscObjectStateIncrease((PetscObject)v);
4859: return(0);
4860: }
4862: /*@C
4863: MatGetRowMaxAbs - Gets the maximum value (in absolute value) of each
4864: row of the matrix
4866: Logically Collective on Mat
4868: Input Parameters:
4869: . mat - the matrix
4871: Output Parameter:
4872: + v - the vector for storing the maximums
4873: - idx - the indices of the column found for each row (or NULL if not needed)
4875: Level: intermediate
4877: Notes:
4878: if a row is completely empty or has only 0.0 values then the idx[] value for that
4879: row is 0 (the first column).
4881: This code is only implemented for a couple of matrix formats.
4883: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMin()
4884: @*/
4885: PetscErrorCode MatGetRowMaxAbs(Mat mat,Vec v,PetscInt idx[])
4886: {
4893: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4895: if (!mat->cmap->N) {
4896: VecSet(v,0.0);
4897: if (idx) {
4898: PetscInt i,m = mat->rmap->n;
4899: for (i=0; i<m; i++) idx[i] = -1;
4900: }
4901: } else {
4902: if (!mat->ops->getrowmaxabs) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4903: MatCheckPreallocated(mat,1);
4904: if (idx) {PetscArrayzero(idx,mat->rmap->n);}
4905: (*mat->ops->getrowmaxabs)(mat,v,idx);
4906: }
4907: PetscObjectStateIncrease((PetscObject)v);
4908: return(0);
4909: }
4911: /*@
4912: MatGetRowSum - Gets the sum of each row of the matrix
4914: Logically or Neighborhood Collective on Mat
4916: Input Parameters:
4917: . mat - the matrix
4919: Output Parameter:
4920: . v - the vector for storing the sum of rows
4922: Level: intermediate
4924: Notes:
4925: This code is slow since it is not currently specialized for different formats
4927: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMin()
4928: @*/
4929: PetscErrorCode MatGetRowSum(Mat mat, Vec v)
4930: {
4931: Vec ones;
4938: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4939: MatCheckPreallocated(mat,1);
4940: MatCreateVecs(mat,&ones,NULL);
4941: VecSet(ones,1.);
4942: MatMult(mat,ones,v);
4943: VecDestroy(&ones);
4944: return(0);
4945: }
4947: /*@
4948: MatTranspose - Computes an in-place or out-of-place transpose of a matrix.
4950: Collective on Mat
4952: Input Parameter:
4953: + mat - the matrix to transpose
4954: - reuse - either MAT_INITIAL_MATRIX, MAT_REUSE_MATRIX, or MAT_INPLACE_MATRIX
4956: Output Parameters:
4957: . B - the transpose
4959: Notes:
4960: If you use MAT_INPLACE_MATRIX then you must pass in &mat for B
4962: MAT_REUSE_MATRIX causes the B matrix from a previous call to this function with MAT_INITIAL_MATRIX to be used
4964: Consider using MatCreateTranspose() instead if you only need a matrix that behaves like the transpose, but don't need the storage to be changed.
4966: Level: intermediate
4968: .seealso: MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse
4969: @*/
4970: PetscErrorCode MatTranspose(Mat mat,MatReuse reuse,Mat *B)
4971: {
4977: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4978: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4979: if (!mat->ops->transpose) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4980: if (reuse == MAT_INPLACE_MATRIX && mat != *B) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"MAT_INPLACE_MATRIX requires last matrix to match first");
4981: if (reuse == MAT_REUSE_MATRIX && mat == *B) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Perhaps you mean MAT_INPLACE_MATRIX");
4982: MatCheckPreallocated(mat,1);
4984: PetscLogEventBegin(MAT_Transpose,mat,0,0,0);
4985: (*mat->ops->transpose)(mat,reuse,B);
4986: PetscLogEventEnd(MAT_Transpose,mat,0,0,0);
4987: if (B) {PetscObjectStateIncrease((PetscObject)*B);}
4988: return(0);
4989: }
4991: /*@
4992: MatIsTranspose - Test whether a matrix is another one's transpose,
4993: or its own, in which case it tests symmetry.
4995: Collective on Mat
4997: Input Parameter:
4998: + A - the matrix to test
4999: - B - the matrix to test against, this can equal the first parameter
5001: Output Parameters:
5002: . flg - the result
5004: Notes:
5005: Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm
5006: has a running time of the order of the number of nonzeros; the parallel
5007: test involves parallel copies of the block-offdiagonal parts of the matrix.
5009: Level: intermediate
5011: .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian()
5012: @*/
5013: PetscErrorCode MatIsTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg)
5014: {
5015: PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool*),(*g)(Mat,Mat,PetscReal,PetscBool*);
5021: PetscObjectQueryFunction((PetscObject)A,"MatIsTranspose_C",&f);
5022: PetscObjectQueryFunction((PetscObject)B,"MatIsTranspose_C",&g);
5023: *flg = PETSC_FALSE;
5024: if (f && g) {
5025: if (f == g) {
5026: (*f)(A,B,tol,flg);
5027: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for symmetry test");
5028: } else {
5029: MatType mattype;
5030: if (!f) {
5031: MatGetType(A,&mattype);
5032: } else {
5033: MatGetType(B,&mattype);
5034: }
5035: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type %s does not support checking for transpose",mattype);
5036: }
5037: return(0);
5038: }
5040: /*@
5041: MatHermitianTranspose - Computes an in-place or out-of-place transpose of a matrix in complex conjugate.
5043: Collective on Mat
5045: Input Parameter:
5046: + mat - the matrix to transpose and complex conjugate
5047: - reuse - MAT_INITIAL_MATRIX to create a new matrix, MAT_INPLACE_MATRIX to reuse the first argument to store the transpose
5049: Output Parameters:
5050: . B - the Hermitian
5052: Level: intermediate
5054: .seealso: MatTranspose(), MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse
5055: @*/
5056: PetscErrorCode MatHermitianTranspose(Mat mat,MatReuse reuse,Mat *B)
5057: {
5061: MatTranspose(mat,reuse,B);
5062: #if defined(PETSC_USE_COMPLEX)
5063: MatConjugate(*B);
5064: #endif
5065: return(0);
5066: }
5068: /*@
5069: MatIsHermitianTranspose - Test whether a matrix is another one's Hermitian transpose,
5071: Collective on Mat
5073: Input Parameter:
5074: + A - the matrix to test
5075: - B - the matrix to test against, this can equal the first parameter
5077: Output Parameters:
5078: . flg - the result
5080: Notes:
5081: Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm
5082: has a running time of the order of the number of nonzeros; the parallel
5083: test involves parallel copies of the block-offdiagonal parts of the matrix.
5085: Level: intermediate
5087: .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian(), MatIsTranspose()
5088: @*/
5089: PetscErrorCode MatIsHermitianTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg)
5090: {
5091: PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool*),(*g)(Mat,Mat,PetscReal,PetscBool*);
5097: PetscObjectQueryFunction((PetscObject)A,"MatIsHermitianTranspose_C",&f);
5098: PetscObjectQueryFunction((PetscObject)B,"MatIsHermitianTranspose_C",&g);
5099: if (f && g) {
5100: if (f==g) {
5101: (*f)(A,B,tol,flg);
5102: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for Hermitian test");
5103: }
5104: return(0);
5105: }
5107: /*@
5108: MatPermute - Creates a new matrix with rows and columns permuted from the
5109: original.
5111: Collective on Mat
5113: Input Parameters:
5114: + mat - the matrix to permute
5115: . row - row permutation, each processor supplies only the permutation for its rows
5116: - col - column permutation, each processor supplies only the permutation for its columns
5118: Output Parameters:
5119: . B - the permuted matrix
5121: Level: advanced
5123: Note:
5124: The index sets map from row/col of permuted matrix to row/col of original matrix.
5125: The index sets should be on the same communicator as Mat and have the same local sizes.
5127: .seealso: MatGetOrdering(), ISAllGather()
5129: @*/
5130: PetscErrorCode MatPermute(Mat mat,IS row,IS col,Mat *B)
5131: {
5140: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5141: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5142: if (!mat->ops->permute) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatPermute not available for Mat type %s",((PetscObject)mat)->type_name);
5143: MatCheckPreallocated(mat,1);
5145: (*mat->ops->permute)(mat,row,col,B);
5146: PetscObjectStateIncrease((PetscObject)*B);
5147: return(0);
5148: }
5150: /*@
5151: MatEqual - Compares two matrices.
5153: Collective on Mat
5155: Input Parameters:
5156: + A - the first matrix
5157: - B - the second matrix
5159: Output Parameter:
5160: . flg - PETSC_TRUE if the matrices are equal; PETSC_FALSE otherwise.
5162: Level: intermediate
5164: @*/
5165: PetscErrorCode MatEqual(Mat A,Mat B,PetscBool *flg)
5166: {
5176: MatCheckPreallocated(B,2);
5177: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5178: if (!B->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5179: if (A->rmap->N != B->rmap->N || A->cmap->N != B->cmap->N) SETERRQ4(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim %D %D %D %D",A->rmap->N,B->rmap->N,A->cmap->N,B->cmap->N);
5180: if (!A->ops->equal) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)A)->type_name);
5181: if (!B->ops->equal) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)B)->type_name);
5182: if (A->ops->equal != B->ops->equal) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_INCOMP,"A is type: %s\nB is type: %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name);
5183: MatCheckPreallocated(A,1);
5185: (*A->ops->equal)(A,B,flg);
5186: return(0);
5187: }
5189: /*@
5190: MatDiagonalScale - Scales a matrix on the left and right by diagonal
5191: matrices that are stored as vectors. Either of the two scaling
5192: matrices can be NULL.
5194: Collective on Mat
5196: Input Parameters:
5197: + mat - the matrix to be scaled
5198: . l - the left scaling vector (or NULL)
5199: - r - the right scaling vector (or NULL)
5201: Notes:
5202: MatDiagonalScale() computes A = LAR, where
5203: L = a diagonal matrix (stored as a vector), R = a diagonal matrix (stored as a vector)
5204: The L scales the rows of the matrix, the R scales the columns of the matrix.
5206: Level: intermediate
5209: .seealso: MatScale(), MatShift(), MatDiagonalSet()
5210: @*/
5211: PetscErrorCode MatDiagonalScale(Mat mat,Vec l,Vec r)
5212: {
5220: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5221: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5222: MatCheckPreallocated(mat,1);
5223: if (!l && !r) return(0);
5225: if (!mat->ops->diagonalscale) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5226: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
5227: (*mat->ops->diagonalscale)(mat,l,r);
5228: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
5229: PetscObjectStateIncrease((PetscObject)mat);
5230: return(0);
5231: }
5233: /*@
5234: MatScale - Scales all elements of a matrix by a given number.
5236: Logically Collective on Mat
5238: Input Parameters:
5239: + mat - the matrix to be scaled
5240: - a - the scaling value
5242: Output Parameter:
5243: . mat - the scaled matrix
5245: Level: intermediate
5247: .seealso: MatDiagonalScale()
5248: @*/
5249: PetscErrorCode MatScale(Mat mat,PetscScalar a)
5250: {
5256: if (a != (PetscScalar)1.0 && !mat->ops->scale) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5257: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5258: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5260: MatCheckPreallocated(mat,1);
5262: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
5263: if (a != (PetscScalar)1.0) {
5264: (*mat->ops->scale)(mat,a);
5265: PetscObjectStateIncrease((PetscObject)mat);
5266: }
5267: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
5268: return(0);
5269: }
5271: /*@
5272: MatNorm - Calculates various norms of a matrix.
5274: Collective on Mat
5276: Input Parameters:
5277: + mat - the matrix
5278: - type - the type of norm, NORM_1, NORM_FROBENIUS, NORM_INFINITY
5280: Output Parameters:
5281: . nrm - the resulting norm
5283: Level: intermediate
5285: @*/
5286: PetscErrorCode MatNorm(Mat mat,NormType type,PetscReal *nrm)
5287: {
5295: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5296: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5297: if (!mat->ops->norm) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5298: MatCheckPreallocated(mat,1);
5300: (*mat->ops->norm)(mat,type,nrm);
5301: return(0);
5302: }
5304: /*
5305: This variable is used to prevent counting of MatAssemblyBegin() that
5306: are called from within a MatAssemblyEnd().
5307: */
5308: static PetscInt MatAssemblyEnd_InUse = 0;
5309: /*@
5310: MatAssemblyBegin - Begins assembling the matrix. This routine should
5311: be called after completing all calls to MatSetValues().
5313: Collective on Mat
5315: Input Parameters:
5316: + mat - the matrix
5317: - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY
5319: Notes:
5320: MatSetValues() generally caches the values. The matrix is ready to
5321: use only after MatAssemblyBegin() and MatAssemblyEnd() have been called.
5322: Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES
5323: in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before
5324: using the matrix.
5326: ALL processes that share a matrix MUST call MatAssemblyBegin() and MatAssemblyEnd() the SAME NUMBER of times, and each time with the
5327: same flag of MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY for all processes. Thus you CANNOT locally change from ADD_VALUES to INSERT_VALUES, that is
5328: a global collective operation requring all processes that share the matrix.
5330: Space for preallocated nonzeros that is not filled by a call to MatSetValues() or a related routine are compressed
5331: out by assembly. If you intend to use that extra space on a subsequent assembly, be sure to insert explicit zeros
5332: before MAT_FINAL_ASSEMBLY so the space is not compressed out.
5334: Level: beginner
5336: .seealso: MatAssemblyEnd(), MatSetValues(), MatAssembled()
5337: @*/
5338: PetscErrorCode MatAssemblyBegin(Mat mat,MatAssemblyType type)
5339: {
5345: MatCheckPreallocated(mat,1);
5346: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix.\nDid you forget to call MatSetUnfactored()?");
5347: if (mat->assembled) {
5348: mat->was_assembled = PETSC_TRUE;
5349: mat->assembled = PETSC_FALSE;
5350: }
5352: if (!MatAssemblyEnd_InUse) {
5353: PetscLogEventBegin(MAT_AssemblyBegin,mat,0,0,0);
5354: if (mat->ops->assemblybegin) {(*mat->ops->assemblybegin)(mat,type);}
5355: PetscLogEventEnd(MAT_AssemblyBegin,mat,0,0,0);
5356: } else if (mat->ops->assemblybegin) {
5357: (*mat->ops->assemblybegin)(mat,type);
5358: }
5359: return(0);
5360: }
5362: /*@
5363: MatAssembled - Indicates if a matrix has been assembled and is ready for
5364: use; for example, in matrix-vector product.
5366: Not Collective
5368: Input Parameter:
5369: . mat - the matrix
5371: Output Parameter:
5372: . assembled - PETSC_TRUE or PETSC_FALSE
5374: Level: advanced
5376: .seealso: MatAssemblyEnd(), MatSetValues(), MatAssemblyBegin()
5377: @*/
5378: PetscErrorCode MatAssembled(Mat mat,PetscBool *assembled)
5379: {
5383: *assembled = mat->assembled;
5384: return(0);
5385: }
5387: /*@
5388: MatAssemblyEnd - Completes assembling the matrix. This routine should
5389: be called after MatAssemblyBegin().
5391: Collective on Mat
5393: Input Parameters:
5394: + mat - the matrix
5395: - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY
5397: Options Database Keys:
5398: + -mat_view ::ascii_info - Prints info on matrix at conclusion of MatEndAssembly()
5399: . -mat_view ::ascii_info_detail - Prints more detailed info
5400: . -mat_view - Prints matrix in ASCII format
5401: . -mat_view ::ascii_matlab - Prints matrix in Matlab format
5402: . -mat_view draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX().
5403: . -display <name> - Sets display name (default is host)
5404: . -draw_pause <sec> - Sets number of seconds to pause after display
5405: . -mat_view socket - Sends matrix to socket, can be accessed from Matlab (See Users-Manual: ch_matlab)
5406: . -viewer_socket_machine <machine> - Machine to use for socket
5407: . -viewer_socket_port <port> - Port number to use for socket
5408: - -mat_view binary:filename[:append] - Save matrix to file in binary format
5410: Notes:
5411: MatSetValues() generally caches the values. The matrix is ready to
5412: use only after MatAssemblyBegin() and MatAssemblyEnd() have been called.
5413: Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES
5414: in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before
5415: using the matrix.
5417: Space for preallocated nonzeros that is not filled by a call to MatSetValues() or a related routine are compressed
5418: out by assembly. If you intend to use that extra space on a subsequent assembly, be sure to insert explicit zeros
5419: before MAT_FINAL_ASSEMBLY so the space is not compressed out.
5421: Level: beginner
5423: .seealso: MatAssemblyBegin(), MatSetValues(), PetscDrawOpenX(), PetscDrawCreate(), MatView(), MatAssembled(), PetscViewerSocketOpen()
5424: @*/
5425: PetscErrorCode MatAssemblyEnd(Mat mat,MatAssemblyType type)
5426: {
5427: PetscErrorCode ierr;
5428: static PetscInt inassm = 0;
5429: PetscBool flg = PETSC_FALSE;
5435: inassm++;
5436: MatAssemblyEnd_InUse++;
5437: if (MatAssemblyEnd_InUse == 1) { /* Do the logging only the first time through */
5438: PetscLogEventBegin(MAT_AssemblyEnd,mat,0,0,0);
5439: if (mat->ops->assemblyend) {
5440: (*mat->ops->assemblyend)(mat,type);
5441: }
5442: PetscLogEventEnd(MAT_AssemblyEnd,mat,0,0,0);
5443: } else if (mat->ops->assemblyend) {
5444: (*mat->ops->assemblyend)(mat,type);
5445: }
5447: /* Flush assembly is not a true assembly */
5448: if (type != MAT_FLUSH_ASSEMBLY) {
5449: mat->num_ass++;
5450: mat->assembled = PETSC_TRUE;
5451: mat->ass_nonzerostate = mat->nonzerostate;
5452: }
5454: mat->insertmode = NOT_SET_VALUES;
5455: MatAssemblyEnd_InUse--;
5456: PetscObjectStateIncrease((PetscObject)mat);
5457: if (!mat->symmetric_eternal) {
5458: mat->symmetric_set = PETSC_FALSE;
5459: mat->hermitian_set = PETSC_FALSE;
5460: mat->structurally_symmetric_set = PETSC_FALSE;
5461: }
5462: if (inassm == 1 && type != MAT_FLUSH_ASSEMBLY) {
5463: MatViewFromOptions(mat,NULL,"-mat_view");
5465: if (mat->checksymmetryonassembly) {
5466: MatIsSymmetric(mat,mat->checksymmetrytol,&flg);
5467: if (flg) {
5468: PetscPrintf(PetscObjectComm((PetscObject)mat),"Matrix is symmetric (tolerance %g)\n",(double)mat->checksymmetrytol);
5469: } else {
5470: PetscPrintf(PetscObjectComm((PetscObject)mat),"Matrix is not symmetric (tolerance %g)\n",(double)mat->checksymmetrytol);
5471: }
5472: }
5473: if (mat->nullsp && mat->checknullspaceonassembly) {
5474: MatNullSpaceTest(mat->nullsp,mat,NULL);
5475: }
5476: }
5477: inassm--;
5478: return(0);
5479: }
5481: /*@
5482: MatSetOption - Sets a parameter option for a matrix. Some options
5483: may be specific to certain storage formats. Some options
5484: determine how values will be inserted (or added). Sorted,
5485: row-oriented input will generally assemble the fastest. The default
5486: is row-oriented.
5488: Logically Collective on Mat for certain operations, such as MAT_SPD, not collective for MAT_ROW_ORIENTED, see MatOption
5490: Input Parameters:
5491: + mat - the matrix
5492: . option - the option, one of those listed below (and possibly others),
5493: - flg - turn the option on (PETSC_TRUE) or off (PETSC_FALSE)
5495: Options Describing Matrix Structure:
5496: + MAT_SPD - symmetric positive definite
5497: . MAT_SYMMETRIC - symmetric in terms of both structure and value
5498: . MAT_HERMITIAN - transpose is the complex conjugation
5499: . MAT_STRUCTURALLY_SYMMETRIC - symmetric nonzero structure
5500: - MAT_SYMMETRY_ETERNAL - if you would like the symmetry/Hermitian flag
5501: you set to be kept with all future use of the matrix
5502: including after MatAssemblyBegin/End() which could
5503: potentially change the symmetry structure, i.e. you
5504: KNOW the matrix will ALWAYS have the property you set.
5505: Note that setting this flag alone implies nothing about whether the matrix is symmetric/Hermitian;
5506: the relevant flags must be set independently.
5509: Options For Use with MatSetValues():
5510: Insert a logically dense subblock, which can be
5511: . MAT_ROW_ORIENTED - row-oriented (default)
5513: Note these options reflect the data you pass in with MatSetValues(); it has
5514: nothing to do with how the data is stored internally in the matrix
5515: data structure.
5517: When (re)assembling a matrix, we can restrict the input for
5518: efficiency/debugging purposes. These options include:
5519: + MAT_NEW_NONZERO_LOCATIONS - additional insertions will be allowed if they generate a new nonzero (slow)
5520: . MAT_NEW_DIAGONALS - new diagonals will be allowed (for block diagonal format only)
5521: . MAT_IGNORE_OFF_PROC_ENTRIES - drops off-processor entries
5522: . MAT_NEW_NONZERO_LOCATION_ERR - generates an error for new matrix entry
5523: . MAT_USE_HASH_TABLE - uses a hash table to speed up matrix assembly
5524: . MAT_NO_OFF_PROC_ENTRIES - you know each process will only set values for its own rows, will generate an error if
5525: any process sets values for another process. This avoids all reductions in the MatAssembly routines and thus improves
5526: performance for very large process counts.
5527: - MAT_SUBSET_OFF_PROC_ENTRIES - you know that the first assembly after setting this flag will set a superset
5528: of the off-process entries required for all subsequent assemblies. This avoids a rendezvous step in the MatAssembly
5529: functions, instead sending only neighbor messages.
5531: Notes:
5532: Except for MAT_UNUSED_NONZERO_LOCATION_ERR and MAT_ROW_ORIENTED all processes that share the matrix must pass the same value in flg!
5534: Some options are relevant only for particular matrix types and
5535: are thus ignored by others. Other options are not supported by
5536: certain matrix types and will generate an error message if set.
5538: If using a Fortran 77 module to compute a matrix, one may need to
5539: use the column-oriented option (or convert to the row-oriented
5540: format).
5542: MAT_NEW_NONZERO_LOCATIONS set to PETSC_FALSE indicates that any add or insertion
5543: that would generate a new entry in the nonzero structure is instead
5544: ignored. Thus, if memory has not alredy been allocated for this particular
5545: data, then the insertion is ignored. For dense matrices, in which
5546: the entire array is allocated, no entries are ever ignored.
5547: Set after the first MatAssemblyEnd(). If this option is set then the MatAssemblyBegin/End() processes has one less global reduction
5549: MAT_NEW_NONZERO_LOCATION_ERR set to PETSC_TRUE indicates that any add or insertion
5550: that would generate a new entry in the nonzero structure instead produces
5551: an error. (Currently supported for AIJ and BAIJ formats only.) If this option is set then the MatAssemblyBegin/End() processes has one less global reduction
5553: MAT_NEW_NONZERO_ALLOCATION_ERR set to PETSC_TRUE indicates that any add or insertion
5554: that would generate a new entry that has not been preallocated will
5555: instead produce an error. (Currently supported for AIJ and BAIJ formats
5556: only.) This is a useful flag when debugging matrix memory preallocation.
5557: If this option is set then the MatAssemblyBegin/End() processes has one less global reduction
5559: MAT_IGNORE_OFF_PROC_ENTRIES set to PETSC_TRUE indicates entries destined for
5560: other processors should be dropped, rather than stashed.
5561: This is useful if you know that the "owning" processor is also
5562: always generating the correct matrix entries, so that PETSc need
5563: not transfer duplicate entries generated on another processor.
5565: MAT_USE_HASH_TABLE indicates that a hash table be used to improve the
5566: searches during matrix assembly. When this flag is set, the hash table
5567: is created during the first Matrix Assembly. This hash table is
5568: used the next time through, during MatSetVaules()/MatSetVaulesBlocked()
5569: to improve the searching of indices. MAT_NEW_NONZERO_LOCATIONS flag
5570: should be used with MAT_USE_HASH_TABLE flag. This option is currently
5571: supported by MATMPIBAIJ format only.
5573: MAT_KEEP_NONZERO_PATTERN indicates when MatZeroRows() is called the zeroed entries
5574: are kept in the nonzero structure
5576: MAT_IGNORE_ZERO_ENTRIES - for AIJ/IS matrices this will stop zero values from creating
5577: a zero location in the matrix
5579: MAT_USE_INODES - indicates using inode version of the code - works with AIJ matrix types
5581: MAT_NO_OFF_PROC_ZERO_ROWS - you know each process will only zero its own rows. This avoids all reductions in the
5582: zero row routines and thus improves performance for very large process counts.
5584: MAT_IGNORE_LOWER_TRIANGULAR - For SBAIJ matrices will ignore any insertions you make in the lower triangular
5585: part of the matrix (since they should match the upper triangular part).
5587: MAT_SORTED_FULL - each process provides exactly its local rows; all column indices for a given row are passed in a
5588: single call to MatSetValues(), preallocation is perfect, row oriented, INSERT_VALUES is used. Common
5589: with finite difference schemes with non-periodic boundary conditions.
5590: Notes:
5591: Can only be called after MatSetSizes() and MatSetType() have been set.
5593: Level: intermediate
5595: .seealso: MatOption, Mat
5597: @*/
5598: PetscErrorCode MatSetOption(Mat mat,MatOption op,PetscBool flg)
5599: {
5605: if (op > 0) {
5608: }
5610: if (((int) op) <= MAT_OPTION_MIN || ((int) op) >= MAT_OPTION_MAX) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Options %d is out of range",(int)op);
5611: if (!((PetscObject)mat)->type_name) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_TYPENOTSET,"Cannot set options until type and size have been set, see MatSetType() and MatSetSizes()");
5613: switch (op) {
5614: case MAT_NO_OFF_PROC_ENTRIES:
5615: mat->nooffprocentries = flg;
5616: return(0);
5617: break;
5618: case MAT_SUBSET_OFF_PROC_ENTRIES:
5619: mat->assembly_subset = flg;
5620: if (!mat->assembly_subset) { /* See the same logic in VecAssembly wrt VEC_SUBSET_OFF_PROC_ENTRIES */
5621: #if !defined(PETSC_HAVE_MPIUNI)
5622: MatStashScatterDestroy_BTS(&mat->stash);
5623: #endif
5624: mat->stash.first_assembly_done = PETSC_FALSE;
5625: }
5626: return(0);
5627: case MAT_NO_OFF_PROC_ZERO_ROWS:
5628: mat->nooffproczerorows = flg;
5629: return(0);
5630: break;
5631: case MAT_SPD:
5632: mat->spd_set = PETSC_TRUE;
5633: mat->spd = flg;
5634: if (flg) {
5635: mat->symmetric = PETSC_TRUE;
5636: mat->structurally_symmetric = PETSC_TRUE;
5637: mat->symmetric_set = PETSC_TRUE;
5638: mat->structurally_symmetric_set = PETSC_TRUE;
5639: }
5640: break;
5641: case MAT_SYMMETRIC:
5642: mat->symmetric = flg;
5643: if (flg) mat->structurally_symmetric = PETSC_TRUE;
5644: mat->symmetric_set = PETSC_TRUE;
5645: mat->structurally_symmetric_set = flg;
5646: #if !defined(PETSC_USE_COMPLEX)
5647: mat->hermitian = flg;
5648: mat->hermitian_set = PETSC_TRUE;
5649: #endif
5650: break;
5651: case MAT_HERMITIAN:
5652: mat->hermitian = flg;
5653: if (flg) mat->structurally_symmetric = PETSC_TRUE;
5654: mat->hermitian_set = PETSC_TRUE;
5655: mat->structurally_symmetric_set = flg;
5656: #if !defined(PETSC_USE_COMPLEX)
5657: mat->symmetric = flg;
5658: mat->symmetric_set = PETSC_TRUE;
5659: #endif
5660: break;
5661: case MAT_STRUCTURALLY_SYMMETRIC:
5662: mat->structurally_symmetric = flg;
5663: mat->structurally_symmetric_set = PETSC_TRUE;
5664: break;
5665: case MAT_SYMMETRY_ETERNAL:
5666: mat->symmetric_eternal = flg;
5667: break;
5668: case MAT_STRUCTURE_ONLY:
5669: mat->structure_only = flg;
5670: break;
5671: case MAT_SORTED_FULL:
5672: mat->sortedfull = flg;
5673: break;
5674: default:
5675: break;
5676: }
5677: if (mat->ops->setoption) {
5678: (*mat->ops->setoption)(mat,op,flg);
5679: }
5680: return(0);
5681: }
5683: /*@
5684: MatGetOption - Gets a parameter option that has been set for a matrix.
5686: Logically Collective on Mat for certain operations, such as MAT_SPD, not collective for MAT_ROW_ORIENTED, see MatOption
5688: Input Parameters:
5689: + mat - the matrix
5690: - option - the option, this only responds to certain options, check the code for which ones
5692: Output Parameter:
5693: . flg - turn the option on (PETSC_TRUE) or off (PETSC_FALSE)
5695: Notes:
5696: Can only be called after MatSetSizes() and MatSetType() have been set.
5698: Level: intermediate
5700: .seealso: MatOption, MatSetOption()
5702: @*/
5703: PetscErrorCode MatGetOption(Mat mat,MatOption op,PetscBool *flg)
5704: {
5709: if (((int) op) <= MAT_OPTION_MIN || ((int) op) >= MAT_OPTION_MAX) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Options %d is out of range",(int)op);
5710: if (!((PetscObject)mat)->type_name) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_TYPENOTSET,"Cannot get options until type and size have been set, see MatSetType() and MatSetSizes()");
5712: switch (op) {
5713: case MAT_NO_OFF_PROC_ENTRIES:
5714: *flg = mat->nooffprocentries;
5715: break;
5716: case MAT_NO_OFF_PROC_ZERO_ROWS:
5717: *flg = mat->nooffproczerorows;
5718: break;
5719: case MAT_SYMMETRIC:
5720: *flg = mat->symmetric;
5721: break;
5722: case MAT_HERMITIAN:
5723: *flg = mat->hermitian;
5724: break;
5725: case MAT_STRUCTURALLY_SYMMETRIC:
5726: *flg = mat->structurally_symmetric;
5727: break;
5728: case MAT_SYMMETRY_ETERNAL:
5729: *flg = mat->symmetric_eternal;
5730: break;
5731: case MAT_SPD:
5732: *flg = mat->spd;
5733: break;
5734: default:
5735: break;
5736: }
5737: return(0);
5738: }
5740: /*@
5741: MatZeroEntries - Zeros all entries of a matrix. For sparse matrices
5742: this routine retains the old nonzero structure.
5744: Logically Collective on Mat
5746: Input Parameters:
5747: . mat - the matrix
5749: Level: intermediate
5751: Notes:
5752: If the matrix was not preallocated then a default, likely poor preallocation will be set in the matrix, so this should be called after the preallocation phase.
5753: See the Performance chapter of the users manual for information on preallocating matrices.
5755: .seealso: MatZeroRows()
5756: @*/
5757: PetscErrorCode MatZeroEntries(Mat mat)
5758: {
5764: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5765: if (mat->insertmode != NOT_SET_VALUES) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for matrices where you have set values but not yet assembled");
5766: if (!mat->ops->zeroentries) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5767: MatCheckPreallocated(mat,1);
5769: PetscLogEventBegin(MAT_ZeroEntries,mat,0,0,0);
5770: (*mat->ops->zeroentries)(mat);
5771: PetscLogEventEnd(MAT_ZeroEntries,mat,0,0,0);
5772: PetscObjectStateIncrease((PetscObject)mat);
5773: return(0);
5774: }
5776: /*@
5777: MatZeroRowsColumns - Zeros all entries (except possibly the main diagonal)
5778: of a set of rows and columns of a matrix.
5780: Collective on Mat
5782: Input Parameters:
5783: + mat - the matrix
5784: . numRows - the number of rows to remove
5785: . rows - the global row indices
5786: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5787: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5788: - b - optional vector of right hand side, that will be adjusted by provided solution
5790: Notes:
5791: This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix.
5793: The user can set a value in the diagonal entry (or for the AIJ and
5794: row formats can optionally remove the main diagonal entry from the
5795: nonzero structure as well, by passing 0.0 as the final argument).
5797: For the parallel case, all processes that share the matrix (i.e.,
5798: those in the communicator used for matrix creation) MUST call this
5799: routine, regardless of whether any rows being zeroed are owned by
5800: them.
5802: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5803: list only rows local to itself).
5805: The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine.
5807: Level: intermediate
5809: .seealso: MatZeroRowsIS(), MatZeroRows(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5810: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5811: @*/
5812: PetscErrorCode MatZeroRowsColumns(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
5813: {
5820: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5821: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5822: if (!mat->ops->zerorowscolumns) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5823: MatCheckPreallocated(mat,1);
5825: (*mat->ops->zerorowscolumns)(mat,numRows,rows,diag,x,b);
5826: MatViewFromOptions(mat,NULL,"-mat_view");
5827: PetscObjectStateIncrease((PetscObject)mat);
5828: return(0);
5829: }
5831: /*@
5832: MatZeroRowsColumnsIS - Zeros all entries (except possibly the main diagonal)
5833: of a set of rows and columns of a matrix.
5835: Collective on Mat
5837: Input Parameters:
5838: + mat - the matrix
5839: . is - the rows to zero
5840: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5841: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5842: - b - optional vector of right hand side, that will be adjusted by provided solution
5844: Notes:
5845: This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix.
5847: The user can set a value in the diagonal entry (or for the AIJ and
5848: row formats can optionally remove the main diagonal entry from the
5849: nonzero structure as well, by passing 0.0 as the final argument).
5851: For the parallel case, all processes that share the matrix (i.e.,
5852: those in the communicator used for matrix creation) MUST call this
5853: routine, regardless of whether any rows being zeroed are owned by
5854: them.
5856: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5857: list only rows local to itself).
5859: The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine.
5861: Level: intermediate
5863: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5864: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRows(), MatZeroRowsColumnsStencil()
5865: @*/
5866: PetscErrorCode MatZeroRowsColumnsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
5867: {
5869: PetscInt numRows;
5870: const PetscInt *rows;
5877: ISGetLocalSize(is,&numRows);
5878: ISGetIndices(is,&rows);
5879: MatZeroRowsColumns(mat,numRows,rows,diag,x,b);
5880: ISRestoreIndices(is,&rows);
5881: return(0);
5882: }
5884: /*@
5885: MatZeroRows - Zeros all entries (except possibly the main diagonal)
5886: of a set of rows of a matrix.
5888: Collective on Mat
5890: Input Parameters:
5891: + mat - the matrix
5892: . numRows - the number of rows to remove
5893: . rows - the global row indices
5894: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5895: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5896: - b - optional vector of right hand side, that will be adjusted by provided solution
5898: Notes:
5899: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5900: but does not release memory. For the dense and block diagonal
5901: formats this does not alter the nonzero structure.
5903: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5904: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5905: merely zeroed.
5907: The user can set a value in the diagonal entry (or for the AIJ and
5908: row formats can optionally remove the main diagonal entry from the
5909: nonzero structure as well, by passing 0.0 as the final argument).
5911: For the parallel case, all processes that share the matrix (i.e.,
5912: those in the communicator used for matrix creation) MUST call this
5913: routine, regardless of whether any rows being zeroed are owned by
5914: them.
5916: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5917: list only rows local to itself).
5919: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
5920: owns that are to be zeroed. This saves a global synchronization in the implementation.
5922: Level: intermediate
5924: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5925: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5926: @*/
5927: PetscErrorCode MatZeroRows(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
5928: {
5935: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5936: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5937: if (!mat->ops->zerorows) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5938: MatCheckPreallocated(mat,1);
5940: (*mat->ops->zerorows)(mat,numRows,rows,diag,x,b);
5941: MatViewFromOptions(mat,NULL,"-mat_view");
5942: PetscObjectStateIncrease((PetscObject)mat);
5943: return(0);
5944: }
5946: /*@
5947: MatZeroRowsIS - Zeros all entries (except possibly the main diagonal)
5948: of a set of rows of a matrix.
5950: Collective on Mat
5952: Input Parameters:
5953: + mat - the matrix
5954: . is - index set of rows to remove
5955: . diag - value put in all diagonals of eliminated rows
5956: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5957: - b - optional vector of right hand side, that will be adjusted by provided solution
5959: Notes:
5960: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5961: but does not release memory. For the dense and block diagonal
5962: formats this does not alter the nonzero structure.
5964: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5965: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5966: merely zeroed.
5968: The user can set a value in the diagonal entry (or for the AIJ and
5969: row formats can optionally remove the main diagonal entry from the
5970: nonzero structure as well, by passing 0.0 as the final argument).
5972: For the parallel case, all processes that share the matrix (i.e.,
5973: those in the communicator used for matrix creation) MUST call this
5974: routine, regardless of whether any rows being zeroed are owned by
5975: them.
5977: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5978: list only rows local to itself).
5980: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
5981: owns that are to be zeroed. This saves a global synchronization in the implementation.
5983: Level: intermediate
5985: .seealso: MatZeroRows(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5986: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5987: @*/
5988: PetscErrorCode MatZeroRowsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
5989: {
5990: PetscInt numRows;
5991: const PetscInt *rows;
5998: ISGetLocalSize(is,&numRows);
5999: ISGetIndices(is,&rows);
6000: MatZeroRows(mat,numRows,rows,diag,x,b);
6001: ISRestoreIndices(is,&rows);
6002: return(0);
6003: }
6005: /*@
6006: MatZeroRowsStencil - Zeros all entries (except possibly the main diagonal)
6007: of a set of rows of a matrix. These rows must be local to the process.
6009: Collective on Mat
6011: Input Parameters:
6012: + mat - the matrix
6013: . numRows - the number of rows to remove
6014: . rows - the grid coordinates (and component number when dof > 1) for matrix rows
6015: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
6016: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6017: - b - optional vector of right hand side, that will be adjusted by provided solution
6019: Notes:
6020: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
6021: but does not release memory. For the dense and block diagonal
6022: formats this does not alter the nonzero structure.
6024: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6025: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6026: merely zeroed.
6028: The user can set a value in the diagonal entry (or for the AIJ and
6029: row formats can optionally remove the main diagonal entry from the
6030: nonzero structure as well, by passing 0.0 as the final argument).
6032: For the parallel case, all processes that share the matrix (i.e.,
6033: those in the communicator used for matrix creation) MUST call this
6034: routine, regardless of whether any rows being zeroed are owned by
6035: them.
6037: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
6038: list only rows local to itself).
6040: The grid coordinates are across the entire grid, not just the local portion
6042: In Fortran idxm and idxn should be declared as
6043: $ MatStencil idxm(4,m)
6044: and the values inserted using
6045: $ idxm(MatStencil_i,1) = i
6046: $ idxm(MatStencil_j,1) = j
6047: $ idxm(MatStencil_k,1) = k
6048: $ idxm(MatStencil_c,1) = c
6049: etc
6051: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
6052: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
6053: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
6054: DM_BOUNDARY_PERIODIC boundary type.
6056: For indices that don't mean anything for your case (like the k index when working in 2d) or the c index when you have
6057: a single value per point) you can skip filling those indices.
6059: Level: intermediate
6061: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsl(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6062: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6063: @*/
6064: PetscErrorCode MatZeroRowsStencil(Mat mat,PetscInt numRows,const MatStencil rows[],PetscScalar diag,Vec x,Vec b)
6065: {
6066: PetscInt dim = mat->stencil.dim;
6067: PetscInt sdim = dim - (1 - (PetscInt) mat->stencil.noc);
6068: PetscInt *dims = mat->stencil.dims+1;
6069: PetscInt *starts = mat->stencil.starts;
6070: PetscInt *dxm = (PetscInt*) rows;
6071: PetscInt *jdxm, i, j, tmp, numNewRows = 0;
6079: PetscMalloc1(numRows, &jdxm);
6080: for (i = 0; i < numRows; ++i) {
6081: /* Skip unused dimensions (they are ordered k, j, i, c) */
6082: for (j = 0; j < 3-sdim; ++j) dxm++;
6083: /* Local index in X dir */
6084: tmp = *dxm++ - starts[0];
6085: /* Loop over remaining dimensions */
6086: for (j = 0; j < dim-1; ++j) {
6087: /* If nonlocal, set index to be negative */
6088: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT;
6089: /* Update local index */
6090: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
6091: }
6092: /* Skip component slot if necessary */
6093: if (mat->stencil.noc) dxm++;
6094: /* Local row number */
6095: if (tmp >= 0) {
6096: jdxm[numNewRows++] = tmp;
6097: }
6098: }
6099: MatZeroRowsLocal(mat,numNewRows,jdxm,diag,x,b);
6100: PetscFree(jdxm);
6101: return(0);
6102: }
6104: /*@
6105: MatZeroRowsColumnsStencil - Zeros all row and column entries (except possibly the main diagonal)
6106: of a set of rows and columns of a matrix.
6108: Collective on Mat
6110: Input Parameters:
6111: + mat - the matrix
6112: . numRows - the number of rows/columns to remove
6113: . rows - the grid coordinates (and component number when dof > 1) for matrix rows
6114: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
6115: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6116: - b - optional vector of right hand side, that will be adjusted by provided solution
6118: Notes:
6119: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
6120: but does not release memory. For the dense and block diagonal
6121: formats this does not alter the nonzero structure.
6123: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6124: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6125: merely zeroed.
6127: The user can set a value in the diagonal entry (or for the AIJ and
6128: row formats can optionally remove the main diagonal entry from the
6129: nonzero structure as well, by passing 0.0 as the final argument).
6131: For the parallel case, all processes that share the matrix (i.e.,
6132: those in the communicator used for matrix creation) MUST call this
6133: routine, regardless of whether any rows being zeroed are owned by
6134: them.
6136: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
6137: list only rows local to itself, but the row/column numbers are given in local numbering).
6139: The grid coordinates are across the entire grid, not just the local portion
6141: In Fortran idxm and idxn should be declared as
6142: $ MatStencil idxm(4,m)
6143: and the values inserted using
6144: $ idxm(MatStencil_i,1) = i
6145: $ idxm(MatStencil_j,1) = j
6146: $ idxm(MatStencil_k,1) = k
6147: $ idxm(MatStencil_c,1) = c
6148: etc
6150: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
6151: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
6152: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
6153: DM_BOUNDARY_PERIODIC boundary type.
6155: For indices that don't mean anything for your case (like the k index when working in 2d) or the c index when you have
6156: a single value per point) you can skip filling those indices.
6158: Level: intermediate
6160: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6161: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRows()
6162: @*/
6163: PetscErrorCode MatZeroRowsColumnsStencil(Mat mat,PetscInt numRows,const MatStencil rows[],PetscScalar diag,Vec x,Vec b)
6164: {
6165: PetscInt dim = mat->stencil.dim;
6166: PetscInt sdim = dim - (1 - (PetscInt) mat->stencil.noc);
6167: PetscInt *dims = mat->stencil.dims+1;
6168: PetscInt *starts = mat->stencil.starts;
6169: PetscInt *dxm = (PetscInt*) rows;
6170: PetscInt *jdxm, i, j, tmp, numNewRows = 0;
6178: PetscMalloc1(numRows, &jdxm);
6179: for (i = 0; i < numRows; ++i) {
6180: /* Skip unused dimensions (they are ordered k, j, i, c) */
6181: for (j = 0; j < 3-sdim; ++j) dxm++;
6182: /* Local index in X dir */
6183: tmp = *dxm++ - starts[0];
6184: /* Loop over remaining dimensions */
6185: for (j = 0; j < dim-1; ++j) {
6186: /* If nonlocal, set index to be negative */
6187: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT;
6188: /* Update local index */
6189: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
6190: }
6191: /* Skip component slot if necessary */
6192: if (mat->stencil.noc) dxm++;
6193: /* Local row number */
6194: if (tmp >= 0) {
6195: jdxm[numNewRows++] = tmp;
6196: }
6197: }
6198: MatZeroRowsColumnsLocal(mat,numNewRows,jdxm,diag,x,b);
6199: PetscFree(jdxm);
6200: return(0);
6201: }
6203: /*@C
6204: MatZeroRowsLocal - Zeros all entries (except possibly the main diagonal)
6205: of a set of rows of a matrix; using local numbering of rows.
6207: Collective on Mat
6209: Input Parameters:
6210: + mat - the matrix
6211: . numRows - the number of rows to remove
6212: . rows - the global row indices
6213: . diag - value put in all diagonals of eliminated rows
6214: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6215: - b - optional vector of right hand side, that will be adjusted by provided solution
6217: Notes:
6218: Before calling MatZeroRowsLocal(), the user must first set the
6219: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6221: For the AIJ matrix formats this removes the old nonzero structure,
6222: but does not release memory. For the dense and block diagonal
6223: formats this does not alter the nonzero structure.
6225: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6226: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6227: merely zeroed.
6229: The user can set a value in the diagonal entry (or for the AIJ and
6230: row formats can optionally remove the main diagonal entry from the
6231: nonzero structure as well, by passing 0.0 as the final argument).
6233: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
6234: owns that are to be zeroed. This saves a global synchronization in the implementation.
6236: Level: intermediate
6238: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRows(), MatSetOption(),
6239: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6240: @*/
6241: PetscErrorCode MatZeroRowsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
6242: {
6249: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6250: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6251: MatCheckPreallocated(mat,1);
6253: if (mat->ops->zerorowslocal) {
6254: (*mat->ops->zerorowslocal)(mat,numRows,rows,diag,x,b);
6255: } else {
6256: IS is, newis;
6257: const PetscInt *newRows;
6259: if (!mat->rmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first");
6260: ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);
6261: ISLocalToGlobalMappingApplyIS(mat->rmap->mapping,is,&newis);
6262: ISGetIndices(newis,&newRows);
6263: (*mat->ops->zerorows)(mat,numRows,newRows,diag,x,b);
6264: ISRestoreIndices(newis,&newRows);
6265: ISDestroy(&newis);
6266: ISDestroy(&is);
6267: }
6268: PetscObjectStateIncrease((PetscObject)mat);
6269: return(0);
6270: }
6272: /*@
6273: MatZeroRowsLocalIS - Zeros all entries (except possibly the main diagonal)
6274: of a set of rows of a matrix; using local numbering of rows.
6276: Collective on Mat
6278: Input Parameters:
6279: + mat - the matrix
6280: . is - index set of rows to remove
6281: . diag - value put in all diagonals of eliminated rows
6282: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6283: - b - optional vector of right hand side, that will be adjusted by provided solution
6285: Notes:
6286: Before calling MatZeroRowsLocalIS(), the user must first set the
6287: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6289: For the AIJ matrix formats this removes the old nonzero structure,
6290: but does not release memory. For the dense and block diagonal
6291: formats this does not alter the nonzero structure.
6293: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6294: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6295: merely zeroed.
6297: The user can set a value in the diagonal entry (or for the AIJ and
6298: row formats can optionally remove the main diagonal entry from the
6299: nonzero structure as well, by passing 0.0 as the final argument).
6301: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
6302: owns that are to be zeroed. This saves a global synchronization in the implementation.
6304: Level: intermediate
6306: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRows(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6307: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6308: @*/
6309: PetscErrorCode MatZeroRowsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
6310: {
6312: PetscInt numRows;
6313: const PetscInt *rows;
6319: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6320: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6321: MatCheckPreallocated(mat,1);
6323: ISGetLocalSize(is,&numRows);
6324: ISGetIndices(is,&rows);
6325: MatZeroRowsLocal(mat,numRows,rows,diag,x,b);
6326: ISRestoreIndices(is,&rows);
6327: return(0);
6328: }
6330: /*@
6331: MatZeroRowsColumnsLocal - Zeros all entries (except possibly the main diagonal)
6332: of a set of rows and columns of a matrix; using local numbering of rows.
6334: Collective on Mat
6336: Input Parameters:
6337: + mat - the matrix
6338: . numRows - the number of rows to remove
6339: . rows - the global row indices
6340: . diag - value put in all diagonals of eliminated rows
6341: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6342: - b - optional vector of right hand side, that will be adjusted by provided solution
6344: Notes:
6345: Before calling MatZeroRowsColumnsLocal(), the user must first set the
6346: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6348: The user can set a value in the diagonal entry (or for the AIJ and
6349: row formats can optionally remove the main diagonal entry from the
6350: nonzero structure as well, by passing 0.0 as the final argument).
6352: Level: intermediate
6354: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6355: MatZeroRows(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6356: @*/
6357: PetscErrorCode MatZeroRowsColumnsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
6358: {
6360: IS is, newis;
6361: const PetscInt *newRows;
6367: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6368: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6369: MatCheckPreallocated(mat,1);
6371: if (!mat->cmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first");
6372: ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);
6373: ISLocalToGlobalMappingApplyIS(mat->cmap->mapping,is,&newis);
6374: ISGetIndices(newis,&newRows);
6375: (*mat->ops->zerorowscolumns)(mat,numRows,newRows,diag,x,b);
6376: ISRestoreIndices(newis,&newRows);
6377: ISDestroy(&newis);
6378: ISDestroy(&is);
6379: PetscObjectStateIncrease((PetscObject)mat);
6380: return(0);
6381: }
6383: /*@
6384: MatZeroRowsColumnsLocalIS - Zeros all entries (except possibly the main diagonal)
6385: of a set of rows and columns of a matrix; using local numbering of rows.
6387: Collective on Mat
6389: Input Parameters:
6390: + mat - the matrix
6391: . is - index set of rows to remove
6392: . diag - value put in all diagonals of eliminated rows
6393: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6394: - b - optional vector of right hand side, that will be adjusted by provided solution
6396: Notes:
6397: Before calling MatZeroRowsColumnsLocalIS(), the user must first set the
6398: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6400: The user can set a value in the diagonal entry (or for the AIJ and
6401: row formats can optionally remove the main diagonal entry from the
6402: nonzero structure as well, by passing 0.0 as the final argument).
6404: Level: intermediate
6406: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6407: MatZeroRowsColumnsLocal(), MatZeroRows(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6408: @*/
6409: PetscErrorCode MatZeroRowsColumnsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
6410: {
6412: PetscInt numRows;
6413: const PetscInt *rows;
6419: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6420: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6421: MatCheckPreallocated(mat,1);
6423: ISGetLocalSize(is,&numRows);
6424: ISGetIndices(is,&rows);
6425: MatZeroRowsColumnsLocal(mat,numRows,rows,diag,x,b);
6426: ISRestoreIndices(is,&rows);
6427: return(0);
6428: }
6430: /*@C
6431: MatGetSize - Returns the numbers of rows and columns in a matrix.
6433: Not Collective
6435: Input Parameter:
6436: . mat - the matrix
6438: Output Parameters:
6439: + m - the number of global rows
6440: - n - the number of global columns
6442: Note: both output parameters can be NULL on input.
6444: Level: beginner
6446: .seealso: MatGetLocalSize()
6447: @*/
6448: PetscErrorCode MatGetSize(Mat mat,PetscInt *m,PetscInt *n)
6449: {
6452: if (m) *m = mat->rmap->N;
6453: if (n) *n = mat->cmap->N;
6454: return(0);
6455: }
6457: /*@C
6458: MatGetLocalSize - Returns the number of local rows and local columns
6459: of a matrix, that is the local size of the left and right vectors as returned by MatCreateVecs().
6461: Not Collective
6463: Input Parameters:
6464: . mat - the matrix
6466: Output Parameters:
6467: + m - the number of local rows
6468: - n - the number of local columns
6470: Note: both output parameters can be NULL on input.
6472: Level: beginner
6474: .seealso: MatGetSize()
6475: @*/
6476: PetscErrorCode MatGetLocalSize(Mat mat,PetscInt *m,PetscInt *n)
6477: {
6482: if (m) *m = mat->rmap->n;
6483: if (n) *n = mat->cmap->n;
6484: return(0);
6485: }
6487: /*@C
6488: MatGetOwnershipRangeColumn - Returns the range of matrix columns associated with rows of a vector one multiplies by that owned by
6489: this processor. (The columns of the "diagonal block")
6491: Not Collective, unless matrix has not been allocated, then collective on Mat
6493: Input Parameters:
6494: . mat - the matrix
6496: Output Parameters:
6497: + m - the global index of the first local column
6498: - n - one more than the global index of the last local column
6500: Notes:
6501: both output parameters can be NULL on input.
6503: Level: developer
6505: .seealso: MatGetOwnershipRange(), MatGetOwnershipRanges(), MatGetOwnershipRangesColumn()
6507: @*/
6508: PetscErrorCode MatGetOwnershipRangeColumn(Mat mat,PetscInt *m,PetscInt *n)
6509: {
6515: MatCheckPreallocated(mat,1);
6516: if (m) *m = mat->cmap->rstart;
6517: if (n) *n = mat->cmap->rend;
6518: return(0);
6519: }
6521: /*@C
6522: MatGetOwnershipRange - Returns the range of matrix rows owned by
6523: this processor, assuming that the matrix is laid out with the first
6524: n1 rows on the first processor, the next n2 rows on the second, etc.
6525: For certain parallel layouts this range may not be well defined.
6527: Not Collective
6529: Input Parameters:
6530: . mat - the matrix
6532: Output Parameters:
6533: + m - the global index of the first local row
6534: - n - one more than the global index of the last local row
6536: Note: Both output parameters can be NULL on input.
6537: $ This function requires that the matrix be preallocated. If you have not preallocated, consider using
6538: $ PetscSplitOwnership(MPI_Comm comm, PetscInt *n, PetscInt *N)
6539: $ and then MPI_Scan() to calculate prefix sums of the local sizes.
6541: Level: beginner
6543: .seealso: MatGetOwnershipRanges(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn(), PetscSplitOwnership(), PetscSplitOwnershipBlock()
6545: @*/
6546: PetscErrorCode MatGetOwnershipRange(Mat mat,PetscInt *m,PetscInt *n)
6547: {
6553: MatCheckPreallocated(mat,1);
6554: if (m) *m = mat->rmap->rstart;
6555: if (n) *n = mat->rmap->rend;
6556: return(0);
6557: }
6559: /*@C
6560: MatGetOwnershipRanges - Returns the range of matrix rows owned by
6561: each process
6563: Not Collective, unless matrix has not been allocated, then collective on Mat
6565: Input Parameters:
6566: . mat - the matrix
6568: Output Parameters:
6569: . ranges - start of each processors portion plus one more than the total length at the end
6571: Level: beginner
6573: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn()
6575: @*/
6576: PetscErrorCode MatGetOwnershipRanges(Mat mat,const PetscInt **ranges)
6577: {
6583: MatCheckPreallocated(mat,1);
6584: PetscLayoutGetRanges(mat->rmap,ranges);
6585: return(0);
6586: }
6588: /*@C
6589: MatGetOwnershipRangesColumn - Returns the range of matrix columns associated with rows of a vector one multiplies by that owned by
6590: this processor. (The columns of the "diagonal blocks" for each process)
6592: Not Collective, unless matrix has not been allocated, then collective on Mat
6594: Input Parameters:
6595: . mat - the matrix
6597: Output Parameters:
6598: . ranges - start of each processors portion plus one more then the total length at the end
6600: Level: beginner
6602: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRanges()
6604: @*/
6605: PetscErrorCode MatGetOwnershipRangesColumn(Mat mat,const PetscInt **ranges)
6606: {
6612: MatCheckPreallocated(mat,1);
6613: PetscLayoutGetRanges(mat->cmap,ranges);
6614: return(0);
6615: }
6617: /*@C
6618: MatGetOwnershipIS - Get row and column ownership as index sets
6620: Not Collective
6622: Input Arguments:
6623: . A - matrix of type Elemental or ScaLAPACK
6625: Output Arguments:
6626: + rows - rows in which this process owns elements
6627: - cols - columns in which this process owns elements
6629: Level: intermediate
6631: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatSetValues(), MATELEMENTAL
6632: @*/
6633: PetscErrorCode MatGetOwnershipIS(Mat A,IS *rows,IS *cols)
6634: {
6635: PetscErrorCode ierr,(*f)(Mat,IS*,IS*);
6638: MatCheckPreallocated(A,1);
6639: PetscObjectQueryFunction((PetscObject)A,"MatGetOwnershipIS_C",&f);
6640: if (f) {
6641: (*f)(A,rows,cols);
6642: } else { /* Create a standard row-based partition, each process is responsible for ALL columns in their row block */
6643: if (rows) {ISCreateStride(PETSC_COMM_SELF,A->rmap->n,A->rmap->rstart,1,rows);}
6644: if (cols) {ISCreateStride(PETSC_COMM_SELF,A->cmap->N,0,1,cols);}
6645: }
6646: return(0);
6647: }
6649: /*@C
6650: MatILUFactorSymbolic - Performs symbolic ILU factorization of a matrix.
6651: Uses levels of fill only, not drop tolerance. Use MatLUFactorNumeric()
6652: to complete the factorization.
6654: Collective on Mat
6656: Input Parameters:
6657: + mat - the matrix
6658: . row - row permutation
6659: . column - column permutation
6660: - info - structure containing
6661: $ levels - number of levels of fill.
6662: $ expected fill - as ratio of original fill.
6663: $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices
6664: missing diagonal entries)
6666: Output Parameters:
6667: . fact - new matrix that has been symbolically factored
6669: Notes:
6670: See Users-Manual: ch_mat for additional information about choosing the fill factor for better efficiency.
6672: Most users should employ the simplified KSP interface for linear solvers
6673: instead of working directly with matrix algebra routines such as this.
6674: See, e.g., KSPCreate().
6676: Level: developer
6678: .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor()
6679: MatGetOrdering(), MatFactorInfo
6681: Note: this uses the definition of level of fill as in Y. Saad, 2003
6683: Developer Note: fortran interface is not autogenerated as the f90
6684: interface defintion cannot be generated correctly [due to MatFactorInfo]
6686: References:
6687: Y. Saad, Iterative methods for sparse linear systems Philadelphia: Society for Industrial and Applied Mathematics, 2003
6688: @*/
6689: PetscErrorCode MatILUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info)
6690: {
6700: if (info->levels < 0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Levels of fill negative %D",(PetscInt)info->levels);
6701: if (info->fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %g",(double)info->fill);
6702: if (!fact->ops->ilufactorsymbolic) {
6703: MatSolverType stype;
6704: MatFactorGetSolverType(fact,&stype);
6705: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic ILU using solver type %s",((PetscObject)mat)->type_name,stype);
6706: }
6707: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6708: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6709: MatCheckPreallocated(mat,2);
6711: PetscLogEventBegin(MAT_ILUFactorSymbolic,mat,row,col,0);
6712: (fact->ops->ilufactorsymbolic)(fact,mat,row,col,info);
6713: PetscLogEventEnd(MAT_ILUFactorSymbolic,mat,row,col,0);
6714: return(0);
6715: }
6717: /*@C
6718: MatICCFactorSymbolic - Performs symbolic incomplete
6719: Cholesky factorization for a symmetric matrix. Use
6720: MatCholeskyFactorNumeric() to complete the factorization.
6722: Collective on Mat
6724: Input Parameters:
6725: + mat - the matrix
6726: . perm - row and column permutation
6727: - info - structure containing
6728: $ levels - number of levels of fill.
6729: $ expected fill - as ratio of original fill.
6731: Output Parameter:
6732: . fact - the factored matrix
6734: Notes:
6735: Most users should employ the KSP interface for linear solvers
6736: instead of working directly with matrix algebra routines such as this.
6737: See, e.g., KSPCreate().
6739: Level: developer
6741: .seealso: MatCholeskyFactorNumeric(), MatCholeskyFactor(), MatFactorInfo
6743: Note: this uses the definition of level of fill as in Y. Saad, 2003
6745: Developer Note: fortran interface is not autogenerated as the f90
6746: interface defintion cannot be generated correctly [due to MatFactorInfo]
6748: References:
6749: Y. Saad, Iterative methods for sparse linear systems Philadelphia: Society for Industrial and Applied Mathematics, 2003
6750: @*/
6751: PetscErrorCode MatICCFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info)
6752: {
6761: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6762: if (info->levels < 0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Levels negative %D",(PetscInt) info->levels);
6763: if (info->fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %g",(double)info->fill);
6764: if (!(fact)->ops->iccfactorsymbolic) {
6765: MatSolverType stype;
6766: MatFactorGetSolverType(fact,&stype);
6767: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic ICC using solver type %s",((PetscObject)mat)->type_name,stype);
6768: }
6769: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6770: MatCheckPreallocated(mat,2);
6772: PetscLogEventBegin(MAT_ICCFactorSymbolic,mat,perm,0,0);
6773: (fact->ops->iccfactorsymbolic)(fact,mat,perm,info);
6774: PetscLogEventEnd(MAT_ICCFactorSymbolic,mat,perm,0,0);
6775: return(0);
6776: }
6778: /*@C
6779: MatCreateSubMatrices - Extracts several submatrices from a matrix. If submat
6780: points to an array of valid matrices, they may be reused to store the new
6781: submatrices.
6783: Collective on Mat
6785: Input Parameters:
6786: + mat - the matrix
6787: . n - the number of submatrixes to be extracted (on this processor, may be zero)
6788: . irow, icol - index sets of rows and columns to extract
6789: - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
6791: Output Parameter:
6792: . submat - the array of submatrices
6794: Notes:
6795: MatCreateSubMatrices() can extract ONLY sequential submatrices
6796: (from both sequential and parallel matrices). Use MatCreateSubMatrix()
6797: to extract a parallel submatrix.
6799: Some matrix types place restrictions on the row and column
6800: indices, such as that they be sorted or that they be equal to each other.
6802: The index sets may not have duplicate entries.
6804: When extracting submatrices from a parallel matrix, each processor can
6805: form a different submatrix by setting the rows and columns of its
6806: individual index sets according to the local submatrix desired.
6808: When finished using the submatrices, the user should destroy
6809: them with MatDestroySubMatrices().
6811: MAT_REUSE_MATRIX can only be used when the nonzero structure of the
6812: original matrix has not changed from that last call to MatCreateSubMatrices().
6814: This routine creates the matrices in submat; you should NOT create them before
6815: calling it. It also allocates the array of matrix pointers submat.
6817: For BAIJ matrices the index sets must respect the block structure, that is if they
6818: request one row/column in a block, they must request all rows/columns that are in
6819: that block. For example, if the block size is 2 you cannot request just row 0 and
6820: column 0.
6822: Fortran Note:
6823: The Fortran interface is slightly different from that given below; it
6824: requires one to pass in as submat a Mat (integer) array of size at least n+1.
6826: Level: advanced
6829: .seealso: MatDestroySubMatrices(), MatCreateSubMatrix(), MatGetRow(), MatGetDiagonal(), MatReuse
6830: @*/
6831: PetscErrorCode MatCreateSubMatrices(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[])
6832: {
6834: PetscInt i;
6835: PetscBool eq;
6840: if (n) {
6845: }
6847: if (n && scall == MAT_REUSE_MATRIX) {
6850: }
6851: if (!mat->ops->createsubmatrices) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
6852: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6853: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6854: MatCheckPreallocated(mat,1);
6856: PetscLogEventBegin(MAT_CreateSubMats,mat,0,0,0);
6857: (*mat->ops->createsubmatrices)(mat,n,irow,icol,scall,submat);
6858: PetscLogEventEnd(MAT_CreateSubMats,mat,0,0,0);
6859: for (i=0; i<n; i++) {
6860: (*submat)[i]->factortype = MAT_FACTOR_NONE; /* in case in place factorization was previously done on submatrix */
6861: ISEqualUnsorted(irow[i],icol[i],&eq);
6862: if (eq) {
6863: MatPropagateSymmetryOptions(mat,(*submat)[i]);
6864: }
6865: }
6866: return(0);
6867: }
6869: /*@C
6870: MatCreateSubMatricesMPI - Extracts MPI submatrices across a sub communicator of mat (by pairs of IS that may live on subcomms).
6872: Collective on Mat
6874: Input Parameters:
6875: + mat - the matrix
6876: . n - the number of submatrixes to be extracted
6877: . irow, icol - index sets of rows and columns to extract
6878: - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
6880: Output Parameter:
6881: . submat - the array of submatrices
6883: Level: advanced
6886: .seealso: MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRow(), MatGetDiagonal(), MatReuse
6887: @*/
6888: PetscErrorCode MatCreateSubMatricesMPI(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[])
6889: {
6891: PetscInt i;
6892: PetscBool eq;
6897: if (n) {
6902: }
6904: if (n && scall == MAT_REUSE_MATRIX) {
6907: }
6908: if (!mat->ops->createsubmatricesmpi) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
6909: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6910: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6911: MatCheckPreallocated(mat,1);
6913: PetscLogEventBegin(MAT_CreateSubMats,mat,0,0,0);
6914: (*mat->ops->createsubmatricesmpi)(mat,n,irow,icol,scall,submat);
6915: PetscLogEventEnd(MAT_CreateSubMats,mat,0,0,0);
6916: for (i=0; i<n; i++) {
6917: ISEqualUnsorted(irow[i],icol[i],&eq);
6918: if (eq) {
6919: MatPropagateSymmetryOptions(mat,(*submat)[i]);
6920: }
6921: }
6922: return(0);
6923: }
6925: /*@C
6926: MatDestroyMatrices - Destroys an array of matrices.
6928: Collective on Mat
6930: Input Parameters:
6931: + n - the number of local matrices
6932: - mat - the matrices (note that this is a pointer to the array of matrices)
6934: Level: advanced
6936: Notes:
6937: Frees not only the matrices, but also the array that contains the matrices
6938: In Fortran will not free the array.
6940: .seealso: MatCreateSubMatrices() MatDestroySubMatrices()
6941: @*/
6942: PetscErrorCode MatDestroyMatrices(PetscInt n,Mat *mat[])
6943: {
6945: PetscInt i;
6948: if (!*mat) return(0);
6949: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n);
6952: for (i=0; i<n; i++) {
6953: MatDestroy(&(*mat)[i]);
6954: }
6956: /* memory is allocated even if n = 0 */
6957: PetscFree(*mat);
6958: return(0);
6959: }
6961: /*@C
6962: MatDestroySubMatrices - Destroys a set of matrices obtained with MatCreateSubMatrices().
6964: Collective on Mat
6966: Input Parameters:
6967: + n - the number of local matrices
6968: - mat - the matrices (note that this is a pointer to the array of matrices, just to match the calling
6969: sequence of MatCreateSubMatrices())
6971: Level: advanced
6973: Notes:
6974: Frees not only the matrices, but also the array that contains the matrices
6975: In Fortran will not free the array.
6977: .seealso: MatCreateSubMatrices()
6978: @*/
6979: PetscErrorCode MatDestroySubMatrices(PetscInt n,Mat *mat[])
6980: {
6982: Mat mat0;
6985: if (!*mat) return(0);
6986: /* mat[] is an array of length n+1, see MatCreateSubMatrices_xxx() */
6987: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n);
6990: mat0 = (*mat)[0];
6991: if (mat0 && mat0->ops->destroysubmatrices) {
6992: (mat0->ops->destroysubmatrices)(n,mat);
6993: } else {
6994: MatDestroyMatrices(n,mat);
6995: }
6996: return(0);
6997: }
6999: /*@C
7000: MatGetSeqNonzeroStructure - Extracts the sequential nonzero structure from a matrix.
7002: Collective on Mat
7004: Input Parameters:
7005: . mat - the matrix
7007: Output Parameter:
7008: . matstruct - the sequential matrix with the nonzero structure of mat
7010: Level: intermediate
7012: .seealso: MatDestroySeqNonzeroStructure(), MatCreateSubMatrices(), MatDestroyMatrices()
7013: @*/
7014: PetscErrorCode MatGetSeqNonzeroStructure(Mat mat,Mat *matstruct)
7015: {
7023: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
7024: MatCheckPreallocated(mat,1);
7026: if (!mat->ops->getseqnonzerostructure) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not for matrix type %s\n",((PetscObject)mat)->type_name);
7027: PetscLogEventBegin(MAT_GetSeqNonzeroStructure,mat,0,0,0);
7028: (*mat->ops->getseqnonzerostructure)(mat,matstruct);
7029: PetscLogEventEnd(MAT_GetSeqNonzeroStructure,mat,0,0,0);
7030: return(0);
7031: }
7033: /*@C
7034: MatDestroySeqNonzeroStructure - Destroys matrix obtained with MatGetSeqNonzeroStructure().
7036: Collective on Mat
7038: Input Parameters:
7039: . mat - the matrix (note that this is a pointer to the array of matrices, just to match the calling
7040: sequence of MatGetSequentialNonzeroStructure())
7042: Level: advanced
7044: Notes:
7045: Frees not only the matrices, but also the array that contains the matrices
7047: .seealso: MatGetSeqNonzeroStructure()
7048: @*/
7049: PetscErrorCode MatDestroySeqNonzeroStructure(Mat *mat)
7050: {
7055: MatDestroy(mat);
7056: return(0);
7057: }
7059: /*@
7060: MatIncreaseOverlap - Given a set of submatrices indicated by index sets,
7061: replaces the index sets by larger ones that represent submatrices with
7062: additional overlap.
7064: Collective on Mat
7066: Input Parameters:
7067: + mat - the matrix
7068: . n - the number of index sets
7069: . is - the array of index sets (these index sets will changed during the call)
7070: - ov - the additional overlap requested
7072: Options Database:
7073: . -mat_increase_overlap_scalable - use a scalable algorithm to compute the overlap (supported by MPIAIJ matrix)
7075: Level: developer
7078: .seealso: MatCreateSubMatrices()
7079: @*/
7080: PetscErrorCode MatIncreaseOverlap(Mat mat,PetscInt n,IS is[],PetscInt ov)
7081: {
7087: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n);
7088: if (n) {
7091: }
7092: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
7093: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
7094: MatCheckPreallocated(mat,1);
7096: if (!ov) return(0);
7097: if (!mat->ops->increaseoverlap) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
7098: PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);
7099: (*mat->ops->increaseoverlap)(mat,n,is,ov);
7100: PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);
7101: return(0);
7102: }
7105: PetscErrorCode MatIncreaseOverlapSplit_Single(Mat,IS*,PetscInt);
7107: /*@
7108: MatIncreaseOverlapSplit - Given a set of submatrices indicated by index sets across
7109: a sub communicator, replaces the index sets by larger ones that represent submatrices with
7110: additional overlap.
7112: Collective on Mat
7114: Input Parameters:
7115: + mat - the matrix
7116: . n - the number of index sets
7117: . is - the array of index sets (these index sets will changed during the call)
7118: - ov - the additional overlap requested
7120: Options Database:
7121: . -mat_increase_overlap_scalable - use a scalable algorithm to compute the overlap (supported by MPIAIJ matrix)
7123: Level: developer
7126: .seealso: MatCreateSubMatrices()
7127: @*/
7128: PetscErrorCode MatIncreaseOverlapSplit(Mat mat,PetscInt n,IS is[],PetscInt ov)
7129: {
7130: PetscInt i;
7136: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n);
7137: if (n) {
7140: }
7141: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
7142: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
7143: MatCheckPreallocated(mat,1);
7144: if (!ov) return(0);
7145: PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);
7146: for (i=0; i<n; i++){
7147: MatIncreaseOverlapSplit_Single(mat,&is[i],ov);
7148: }
7149: PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);
7150: return(0);
7151: }
7156: /*@
7157: MatGetBlockSize - Returns the matrix block size.
7159: Not Collective
7161: Input Parameter:
7162: . mat - the matrix
7164: Output Parameter:
7165: . bs - block size
7167: Notes:
7168: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7170: If the block size has not been set yet this routine returns 1.
7172: Level: intermediate
7174: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSizes()
7175: @*/
7176: PetscErrorCode MatGetBlockSize(Mat mat,PetscInt *bs)
7177: {
7181: *bs = PetscAbs(mat->rmap->bs);
7182: return(0);
7183: }
7185: /*@
7186: MatGetBlockSizes - Returns the matrix block row and column sizes.
7188: Not Collective
7190: Input Parameter:
7191: . mat - the matrix
7193: Output Parameter:
7194: + rbs - row block size
7195: - cbs - column block size
7197: Notes:
7198: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7199: If you pass a different block size for the columns than the rows, the row block size determines the square block storage.
7201: If a block size has not been set yet this routine returns 1.
7203: Level: intermediate
7205: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSize(), MatSetBlockSizes()
7206: @*/
7207: PetscErrorCode MatGetBlockSizes(Mat mat,PetscInt *rbs, PetscInt *cbs)
7208: {
7213: if (rbs) *rbs = PetscAbs(mat->rmap->bs);
7214: if (cbs) *cbs = PetscAbs(mat->cmap->bs);
7215: return(0);
7216: }
7218: /*@
7219: MatSetBlockSize - Sets the matrix block size.
7221: Logically Collective on Mat
7223: Input Parameters:
7224: + mat - the matrix
7225: - bs - block size
7227: Notes:
7228: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7229: This must be called before MatSetUp() or MatXXXSetPreallocation() (or will default to 1) and the block size cannot be changed later.
7231: For MATMPIAIJ and MATSEQAIJ matrix formats, this function can be called at a later stage, provided that the specified block size
7232: is compatible with the matrix local sizes.
7234: Level: intermediate
7236: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes()
7237: @*/
7238: PetscErrorCode MatSetBlockSize(Mat mat,PetscInt bs)
7239: {
7245: MatSetBlockSizes(mat,bs,bs);
7246: return(0);
7247: }
7249: /*@
7250: MatSetVariableBlockSizes - Sets a diagonal blocks of the matrix that need not be of the same size
7252: Logically Collective on Mat
7254: Input Parameters:
7255: + mat - the matrix
7256: . nblocks - the number of blocks on this process
7257: - bsizes - the block sizes
7259: Notes:
7260: Currently used by PCVPBJACOBI for SeqAIJ matrices
7262: Level: intermediate
7264: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes(), MatGetVariableBlockSizes()
7265: @*/
7266: PetscErrorCode MatSetVariableBlockSizes(Mat mat,PetscInt nblocks,PetscInt *bsizes)
7267: {
7269: PetscInt i,ncnt = 0, nlocal;
7273: if (nblocks < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Number of local blocks must be great than or equal to zero");
7274: MatGetLocalSize(mat,&nlocal,NULL);
7275: for (i=0; i<nblocks; i++) ncnt += bsizes[i];
7276: if (ncnt != nlocal) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Sum of local block sizes %D does not equal local size of matrix %D",ncnt,nlocal);
7277: PetscFree(mat->bsizes);
7278: mat->nblocks = nblocks;
7279: PetscMalloc1(nblocks,&mat->bsizes);
7280: PetscArraycpy(mat->bsizes,bsizes,nblocks);
7281: return(0);
7282: }
7284: /*@C
7285: MatGetVariableBlockSizes - Gets a diagonal blocks of the matrix that need not be of the same size
7287: Logically Collective on Mat
7289: Input Parameters:
7290: . mat - the matrix
7292: Output Parameters:
7293: + nblocks - the number of blocks on this process
7294: - bsizes - the block sizes
7296: Notes: Currently not supported from Fortran
7298: Level: intermediate
7300: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes(), MatSetVariableBlockSizes()
7301: @*/
7302: PetscErrorCode MatGetVariableBlockSizes(Mat mat,PetscInt *nblocks,const PetscInt **bsizes)
7303: {
7306: *nblocks = mat->nblocks;
7307: *bsizes = mat->bsizes;
7308: return(0);
7309: }
7311: /*@
7312: MatSetBlockSizes - Sets the matrix block row and column sizes.
7314: Logically Collective on Mat
7316: Input Parameters:
7317: + mat - the matrix
7318: . rbs - row block size
7319: - cbs - column block size
7321: Notes:
7322: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7323: If you pass a different block size for the columns than the rows, the row block size determines the square block storage.
7324: This must be called before MatSetUp() or MatXXXSetPreallocation() (or will default to 1) and the block size cannot be changed later.
7326: For MATMPIAIJ and MATSEQAIJ matrix formats, this function can be called at a later stage, provided that the specified block sizes
7327: are compatible with the matrix local sizes.
7329: The row and column block size determine the blocksize of the "row" and "column" vectors returned by MatCreateVecs().
7331: Level: intermediate
7333: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSize(), MatGetBlockSizes()
7334: @*/
7335: PetscErrorCode MatSetBlockSizes(Mat mat,PetscInt rbs,PetscInt cbs)
7336: {
7343: if (mat->ops->setblocksizes) {
7344: (*mat->ops->setblocksizes)(mat,rbs,cbs);
7345: }
7346: if (mat->rmap->refcnt) {
7347: ISLocalToGlobalMapping l2g = NULL;
7348: PetscLayout nmap = NULL;
7350: PetscLayoutDuplicate(mat->rmap,&nmap);
7351: if (mat->rmap->mapping) {
7352: ISLocalToGlobalMappingDuplicate(mat->rmap->mapping,&l2g);
7353: }
7354: PetscLayoutDestroy(&mat->rmap);
7355: mat->rmap = nmap;
7356: mat->rmap->mapping = l2g;
7357: }
7358: if (mat->cmap->refcnt) {
7359: ISLocalToGlobalMapping l2g = NULL;
7360: PetscLayout nmap = NULL;
7362: PetscLayoutDuplicate(mat->cmap,&nmap);
7363: if (mat->cmap->mapping) {
7364: ISLocalToGlobalMappingDuplicate(mat->cmap->mapping,&l2g);
7365: }
7366: PetscLayoutDestroy(&mat->cmap);
7367: mat->cmap = nmap;
7368: mat->cmap->mapping = l2g;
7369: }
7370: PetscLayoutSetBlockSize(mat->rmap,rbs);
7371: PetscLayoutSetBlockSize(mat->cmap,cbs);
7372: return(0);
7373: }
7375: /*@
7376: MatSetBlockSizesFromMats - Sets the matrix block row and column sizes to match a pair of matrices
7378: Logically Collective on Mat
7380: Input Parameters:
7381: + mat - the matrix
7382: . fromRow - matrix from which to copy row block size
7383: - fromCol - matrix from which to copy column block size (can be same as fromRow)
7385: Level: developer
7387: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes()
7388: @*/
7389: PetscErrorCode MatSetBlockSizesFromMats(Mat mat,Mat fromRow,Mat fromCol)
7390: {
7397: if (fromRow->rmap->bs > 0) {PetscLayoutSetBlockSize(mat->rmap,fromRow->rmap->bs);}
7398: if (fromCol->cmap->bs > 0) {PetscLayoutSetBlockSize(mat->cmap,fromCol->cmap->bs);}
7399: return(0);
7400: }
7402: /*@
7403: MatResidual - Default routine to calculate the residual.
7405: Collective on Mat
7407: Input Parameters:
7408: + mat - the matrix
7409: . b - the right-hand-side
7410: - x - the approximate solution
7412: Output Parameter:
7413: . r - location to store the residual
7415: Level: developer
7417: .seealso: PCMGSetResidual()
7418: @*/
7419: PetscErrorCode MatResidual(Mat mat,Vec b,Vec x,Vec r)
7420: {
7429: MatCheckPreallocated(mat,1);
7430: PetscLogEventBegin(MAT_Residual,mat,0,0,0);
7431: if (!mat->ops->residual) {
7432: MatMult(mat,x,r);
7433: VecAYPX(r,-1.0,b);
7434: } else {
7435: (*mat->ops->residual)(mat,b,x,r);
7436: }
7437: PetscLogEventEnd(MAT_Residual,mat,0,0,0);
7438: return(0);
7439: }
7441: /*@C
7442: MatGetRowIJ - Returns the compressed row storage i and j indices for sequential matrices.
7444: Collective on Mat
7446: Input Parameters:
7447: + mat - the matrix
7448: . shift - 0 or 1 indicating we want the indices starting at 0 or 1
7449: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be symmetrized
7450: - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7451: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7452: always used.
7454: Output Parameters:
7455: + n - number of rows in the (possibly compressed) matrix
7456: . ia - the row pointers; that is ia[0] = 0, ia[row] = ia[row-1] + number of elements in that row of the matrix
7457: . ja - the column indices
7458: - done - indicates if the routine actually worked and returned appropriate ia[] and ja[] arrays; callers
7459: are responsible for handling the case when done == PETSC_FALSE and ia and ja are not set
7461: Level: developer
7463: Notes:
7464: You CANNOT change any of the ia[] or ja[] values.
7466: Use MatRestoreRowIJ() when you are finished accessing the ia[] and ja[] values.
7468: Fortran Notes:
7469: In Fortran use
7470: $
7471: $ PetscInt ia(1), ja(1)
7472: $ PetscOffset iia, jja
7473: $ call MatGetRowIJ(mat,shift,symmetric,inodecompressed,n,ia,iia,ja,jja,done,ierr)
7474: $ ! Access the ith and jth entries via ia(iia + i) and ja(jja + j)
7476: or
7477: $
7478: $ PetscInt, pointer :: ia(:),ja(:)
7479: $ call MatGetRowIJF90(mat,shift,symmetric,inodecompressed,n,ia,ja,done,ierr)
7480: $ ! Access the ith and jth entries via ia(i) and ja(j)
7482: .seealso: MatGetColumnIJ(), MatRestoreRowIJ(), MatSeqAIJGetArray()
7483: @*/
7484: PetscErrorCode MatGetRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7485: {
7495: MatCheckPreallocated(mat,1);
7496: if (!mat->ops->getrowij) *done = PETSC_FALSE;
7497: else {
7498: *done = PETSC_TRUE;
7499: PetscLogEventBegin(MAT_GetRowIJ,mat,0,0,0);
7500: (*mat->ops->getrowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7501: PetscLogEventEnd(MAT_GetRowIJ,mat,0,0,0);
7502: }
7503: return(0);
7504: }
7506: /*@C
7507: MatGetColumnIJ - Returns the compressed column storage i and j indices for sequential matrices.
7509: Collective on Mat
7511: Input Parameters:
7512: + mat - the matrix
7513: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7514: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7515: symmetrized
7516: . inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7517: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7518: always used.
7519: . n - number of columns in the (possibly compressed) matrix
7520: . ia - the column pointers; that is ia[0] = 0, ia[col] = i[col-1] + number of elements in that col of the matrix
7521: - ja - the row indices
7523: Output Parameters:
7524: . done - PETSC_TRUE or PETSC_FALSE, indicating whether the values have been returned
7526: Level: developer
7528: .seealso: MatGetRowIJ(), MatRestoreColumnIJ()
7529: @*/
7530: PetscErrorCode MatGetColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7531: {
7541: MatCheckPreallocated(mat,1);
7542: if (!mat->ops->getcolumnij) *done = PETSC_FALSE;
7543: else {
7544: *done = PETSC_TRUE;
7545: (*mat->ops->getcolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7546: }
7547: return(0);
7548: }
7550: /*@C
7551: MatRestoreRowIJ - Call after you are completed with the ia,ja indices obtained with
7552: MatGetRowIJ().
7554: Collective on Mat
7556: Input Parameters:
7557: + mat - the matrix
7558: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7559: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7560: symmetrized
7561: . inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7562: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7563: always used.
7564: . n - size of (possibly compressed) matrix
7565: . ia - the row pointers
7566: - ja - the column indices
7568: Output Parameters:
7569: . done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned
7571: Note:
7572: This routine zeros out n, ia, and ja. This is to prevent accidental
7573: us of the array after it has been restored. If you pass NULL, it will
7574: not zero the pointers. Use of ia or ja after MatRestoreRowIJ() is invalid.
7576: Level: developer
7578: .seealso: MatGetRowIJ(), MatRestoreColumnIJ()
7579: @*/
7580: PetscErrorCode MatRestoreRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7581: {
7590: MatCheckPreallocated(mat,1);
7592: if (!mat->ops->restorerowij) *done = PETSC_FALSE;
7593: else {
7594: *done = PETSC_TRUE;
7595: (*mat->ops->restorerowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7596: if (n) *n = 0;
7597: if (ia) *ia = NULL;
7598: if (ja) *ja = NULL;
7599: }
7600: return(0);
7601: }
7603: /*@C
7604: MatRestoreColumnIJ - Call after you are completed with the ia,ja indices obtained with
7605: MatGetColumnIJ().
7607: Collective on Mat
7609: Input Parameters:
7610: + mat - the matrix
7611: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7612: - symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7613: symmetrized
7614: - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7615: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7616: always used.
7618: Output Parameters:
7619: + n - size of (possibly compressed) matrix
7620: . ia - the column pointers
7621: . ja - the row indices
7622: - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned
7624: Level: developer
7626: .seealso: MatGetColumnIJ(), MatRestoreRowIJ()
7627: @*/
7628: PetscErrorCode MatRestoreColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7629: {
7638: MatCheckPreallocated(mat,1);
7640: if (!mat->ops->restorecolumnij) *done = PETSC_FALSE;
7641: else {
7642: *done = PETSC_TRUE;
7643: (*mat->ops->restorecolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7644: if (n) *n = 0;
7645: if (ia) *ia = NULL;
7646: if (ja) *ja = NULL;
7647: }
7648: return(0);
7649: }
7651: /*@C
7652: MatColoringPatch -Used inside matrix coloring routines that
7653: use MatGetRowIJ() and/or MatGetColumnIJ().
7655: Collective on Mat
7657: Input Parameters:
7658: + mat - the matrix
7659: . ncolors - max color value
7660: . n - number of entries in colorarray
7661: - colorarray - array indicating color for each column
7663: Output Parameters:
7664: . iscoloring - coloring generated using colorarray information
7666: Level: developer
7668: .seealso: MatGetRowIJ(), MatGetColumnIJ()
7670: @*/
7671: PetscErrorCode MatColoringPatch(Mat mat,PetscInt ncolors,PetscInt n,ISColoringValue colorarray[],ISColoring *iscoloring)
7672: {
7680: MatCheckPreallocated(mat,1);
7682: if (!mat->ops->coloringpatch) {
7683: ISColoringCreate(PetscObjectComm((PetscObject)mat),ncolors,n,colorarray,PETSC_OWN_POINTER,iscoloring);
7684: } else {
7685: (*mat->ops->coloringpatch)(mat,ncolors,n,colorarray,iscoloring);
7686: }
7687: return(0);
7688: }
7691: /*@
7692: MatSetUnfactored - Resets a factored matrix to be treated as unfactored.
7694: Logically Collective on Mat
7696: Input Parameter:
7697: . mat - the factored matrix to be reset
7699: Notes:
7700: This routine should be used only with factored matrices formed by in-place
7701: factorization via ILU(0) (or by in-place LU factorization for the MATSEQDENSE
7702: format). This option can save memory, for example, when solving nonlinear
7703: systems with a matrix-free Newton-Krylov method and a matrix-based, in-place
7704: ILU(0) preconditioner.
7706: Note that one can specify in-place ILU(0) factorization by calling
7707: .vb
7708: PCType(pc,PCILU);
7709: PCFactorSeUseInPlace(pc);
7710: .ve
7711: or by using the options -pc_type ilu -pc_factor_in_place
7713: In-place factorization ILU(0) can also be used as a local
7714: solver for the blocks within the block Jacobi or additive Schwarz
7715: methods (runtime option: -sub_pc_factor_in_place). See Users-Manual: ch_pc
7716: for details on setting local solver options.
7718: Most users should employ the simplified KSP interface for linear solvers
7719: instead of working directly with matrix algebra routines such as this.
7720: See, e.g., KSPCreate().
7722: Level: developer
7724: .seealso: PCFactorSetUseInPlace(), PCFactorGetUseInPlace()
7726: @*/
7727: PetscErrorCode MatSetUnfactored(Mat mat)
7728: {
7734: MatCheckPreallocated(mat,1);
7735: mat->factortype = MAT_FACTOR_NONE;
7736: if (!mat->ops->setunfactored) return(0);
7737: (*mat->ops->setunfactored)(mat);
7738: return(0);
7739: }
7741: /*MC
7742: MatDenseGetArrayF90 - Accesses a matrix array from Fortran90.
7744: Synopsis:
7745: MatDenseGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:,:)},integer ierr)
7747: Not collective
7749: Input Parameter:
7750: . x - matrix
7752: Output Parameters:
7753: + xx_v - the Fortran90 pointer to the array
7754: - ierr - error code
7756: Example of Usage:
7757: .vb
7758: PetscScalar, pointer xx_v(:,:)
7759: ....
7760: call MatDenseGetArrayF90(x,xx_v,ierr)
7761: a = xx_v(3)
7762: call MatDenseRestoreArrayF90(x,xx_v,ierr)
7763: .ve
7765: Level: advanced
7767: .seealso: MatDenseRestoreArrayF90(), MatDenseGetArray(), MatDenseRestoreArray(), MatSeqAIJGetArrayF90()
7769: M*/
7771: /*MC
7772: MatDenseRestoreArrayF90 - Restores a matrix array that has been
7773: accessed with MatDenseGetArrayF90().
7775: Synopsis:
7776: MatDenseRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:,:)},integer ierr)
7778: Not collective
7780: Input Parameters:
7781: + x - matrix
7782: - xx_v - the Fortran90 pointer to the array
7784: Output Parameter:
7785: . ierr - error code
7787: Example of Usage:
7788: .vb
7789: PetscScalar, pointer xx_v(:,:)
7790: ....
7791: call MatDenseGetArrayF90(x,xx_v,ierr)
7792: a = xx_v(3)
7793: call MatDenseRestoreArrayF90(x,xx_v,ierr)
7794: .ve
7796: Level: advanced
7798: .seealso: MatDenseGetArrayF90(), MatDenseGetArray(), MatDenseRestoreArray(), MatSeqAIJRestoreArrayF90()
7800: M*/
7803: /*MC
7804: MatSeqAIJGetArrayF90 - Accesses a matrix array from Fortran90.
7806: Synopsis:
7807: MatSeqAIJGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr)
7809: Not collective
7811: Input Parameter:
7812: . x - matrix
7814: Output Parameters:
7815: + xx_v - the Fortran90 pointer to the array
7816: - ierr - error code
7818: Example of Usage:
7819: .vb
7820: PetscScalar, pointer xx_v(:)
7821: ....
7822: call MatSeqAIJGetArrayF90(x,xx_v,ierr)
7823: a = xx_v(3)
7824: call MatSeqAIJRestoreArrayF90(x,xx_v,ierr)
7825: .ve
7827: Level: advanced
7829: .seealso: MatSeqAIJRestoreArrayF90(), MatSeqAIJGetArray(), MatSeqAIJRestoreArray(), MatDenseGetArrayF90()
7831: M*/
7833: /*MC
7834: MatSeqAIJRestoreArrayF90 - Restores a matrix array that has been
7835: accessed with MatSeqAIJGetArrayF90().
7837: Synopsis:
7838: MatSeqAIJRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr)
7840: Not collective
7842: Input Parameters:
7843: + x - matrix
7844: - xx_v - the Fortran90 pointer to the array
7846: Output Parameter:
7847: . ierr - error code
7849: Example of Usage:
7850: .vb
7851: PetscScalar, pointer xx_v(:)
7852: ....
7853: call MatSeqAIJGetArrayF90(x,xx_v,ierr)
7854: a = xx_v(3)
7855: call MatSeqAIJRestoreArrayF90(x,xx_v,ierr)
7856: .ve
7858: Level: advanced
7860: .seealso: MatSeqAIJGetArrayF90(), MatSeqAIJGetArray(), MatSeqAIJRestoreArray(), MatDenseRestoreArrayF90()
7862: M*/
7865: /*@
7866: MatCreateSubMatrix - Gets a single submatrix on the same number of processors
7867: as the original matrix.
7869: Collective on Mat
7871: Input Parameters:
7872: + mat - the original matrix
7873: . isrow - parallel IS containing the rows this processor should obtain
7874: . iscol - parallel IS containing all columns you wish to keep. Each process should list the columns that will be in IT's "diagonal part" in the new matrix.
7875: - cll - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
7877: Output Parameter:
7878: . newmat - the new submatrix, of the same type as the old
7880: Level: advanced
7882: Notes:
7883: The submatrix will be able to be multiplied with vectors using the same layout as iscol.
7885: Some matrix types place restrictions on the row and column indices, such
7886: as that they be sorted or that they be equal to each other.
7888: The index sets may not have duplicate entries.
7890: The first time this is called you should use a cll of MAT_INITIAL_MATRIX,
7891: the MatCreateSubMatrix() routine will create the newmat for you. Any additional calls
7892: to this routine with a mat of the same nonzero structure and with a call of MAT_REUSE_MATRIX
7893: will reuse the matrix generated the first time. You should call MatDestroy() on newmat when
7894: you are finished using it.
7896: The communicator of the newly obtained matrix is ALWAYS the same as the communicator of
7897: the input matrix.
7899: If iscol is NULL then all columns are obtained (not supported in Fortran).
7901: Example usage:
7902: Consider the following 8x8 matrix with 34 non-zero values, that is
7903: assembled across 3 processors. Let's assume that proc0 owns 3 rows,
7904: proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown
7905: as follows:
7907: .vb
7908: 1 2 0 | 0 3 0 | 0 4
7909: Proc0 0 5 6 | 7 0 0 | 8 0
7910: 9 0 10 | 11 0 0 | 12 0
7911: -------------------------------------
7912: 13 0 14 | 15 16 17 | 0 0
7913: Proc1 0 18 0 | 19 20 21 | 0 0
7914: 0 0 0 | 22 23 0 | 24 0
7915: -------------------------------------
7916: Proc2 25 26 27 | 0 0 28 | 29 0
7917: 30 0 0 | 31 32 33 | 0 34
7918: .ve
7920: Suppose isrow = [0 1 | 4 | 6 7] and iscol = [1 2 | 3 4 5 | 6]. The resulting submatrix is
7922: .vb
7923: 2 0 | 0 3 0 | 0
7924: Proc0 5 6 | 7 0 0 | 8
7925: -------------------------------
7926: Proc1 18 0 | 19 20 21 | 0
7927: -------------------------------
7928: Proc2 26 27 | 0 0 28 | 29
7929: 0 0 | 31 32 33 | 0
7930: .ve
7933: .seealso: MatCreateSubMatrices(), MatCreateSubMatricesMPI(), MatCreateSubMatrixVirtual(), MatSubMatrixVirtualUpdate()
7934: @*/
7935: PetscErrorCode MatCreateSubMatrix(Mat mat,IS isrow,IS iscol,MatReuse cll,Mat *newmat)
7936: {
7938: PetscMPIInt size;
7939: Mat *local;
7940: IS iscoltmp;
7941: PetscBool flg;
7950: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
7951: if (cll == MAT_IGNORE_MATRIX) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Cannot use MAT_IGNORE_MATRIX");
7953: MatCheckPreallocated(mat,1);
7954: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
7956: if (!iscol || isrow == iscol) {
7957: PetscBool stride;
7958: PetscMPIInt grabentirematrix = 0,grab;
7959: PetscObjectTypeCompare((PetscObject)isrow,ISSTRIDE,&stride);
7960: if (stride) {
7961: PetscInt first,step,n,rstart,rend;
7962: ISStrideGetInfo(isrow,&first,&step);
7963: if (step == 1) {
7964: MatGetOwnershipRange(mat,&rstart,&rend);
7965: if (rstart == first) {
7966: ISGetLocalSize(isrow,&n);
7967: if (n == rend-rstart) {
7968: grabentirematrix = 1;
7969: }
7970: }
7971: }
7972: }
7973: MPIU_Allreduce(&grabentirematrix,&grab,1,MPI_INT,MPI_MIN,PetscObjectComm((PetscObject)mat));
7974: if (grab) {
7975: PetscInfo(mat,"Getting entire matrix as submatrix\n");
7976: if (cll == MAT_INITIAL_MATRIX) {
7977: *newmat = mat;
7978: PetscObjectReference((PetscObject)mat);
7979: }
7980: return(0);
7981: }
7982: }
7984: if (!iscol) {
7985: ISCreateStride(PetscObjectComm((PetscObject)mat),mat->cmap->n,mat->cmap->rstart,1,&iscoltmp);
7986: } else {
7987: iscoltmp = iscol;
7988: }
7990: /* if original matrix is on just one processor then use submatrix generated */
7991: if (mat->ops->createsubmatrices && !mat->ops->createsubmatrix && size == 1 && cll == MAT_REUSE_MATRIX) {
7992: MatCreateSubMatrices(mat,1,&isrow,&iscoltmp,MAT_REUSE_MATRIX,&newmat);
7993: goto setproperties;
7994: } else if (mat->ops->createsubmatrices && !mat->ops->createsubmatrix && size == 1) {
7995: MatCreateSubMatrices(mat,1,&isrow,&iscoltmp,MAT_INITIAL_MATRIX,&local);
7996: *newmat = *local;
7997: PetscFree(local);
7998: goto setproperties;
7999: } else if (!mat->ops->createsubmatrix) {
8000: /* Create a new matrix type that implements the operation using the full matrix */
8001: PetscLogEventBegin(MAT_CreateSubMat,mat,0,0,0);
8002: switch (cll) {
8003: case MAT_INITIAL_MATRIX:
8004: MatCreateSubMatrixVirtual(mat,isrow,iscoltmp,newmat);
8005: break;
8006: case MAT_REUSE_MATRIX:
8007: MatSubMatrixVirtualUpdate(*newmat,mat,isrow,iscoltmp);
8008: break;
8009: default: SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Invalid MatReuse, must be either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX");
8010: }
8011: PetscLogEventEnd(MAT_CreateSubMat,mat,0,0,0);
8012: goto setproperties;
8013: }
8015: if (!mat->ops->createsubmatrix) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8016: PetscLogEventBegin(MAT_CreateSubMat,mat,0,0,0);
8017: (*mat->ops->createsubmatrix)(mat,isrow,iscoltmp,cll,newmat);
8018: PetscLogEventEnd(MAT_CreateSubMat,mat,0,0,0);
8020: setproperties:
8021: ISEqualUnsorted(isrow,iscoltmp,&flg);
8022: if (flg) {
8023: MatPropagateSymmetryOptions(mat,*newmat);
8024: }
8025: if (!iscol) {ISDestroy(&iscoltmp);}
8026: if (*newmat && cll == MAT_INITIAL_MATRIX) {PetscObjectStateIncrease((PetscObject)*newmat);}
8027: return(0);
8028: }
8030: /*@
8031: MatPropagateSymmetryOptions - Propagates symmetry options set on a matrix to another matrix
8033: Not Collective
8035: Input Parameters:
8036: + A - the matrix we wish to propagate options from
8037: - B - the matrix we wish to propagate options to
8039: Level: beginner
8041: Notes: Propagates the options associated to MAT_SYMMETRY_ETERNAL, MAT_STRUCTURALLY_SYMMETRIC, MAT_HERMITIAN, MAT_SPD and MAT_SYMMETRIC
8043: .seealso: MatSetOption()
8044: @*/
8045: PetscErrorCode MatPropagateSymmetryOptions(Mat A, Mat B)
8046: {
8052: if (A->symmetric_eternal) { /* symmetric_eternal does not have a corresponding *set flag */
8053: MatSetOption(B,MAT_SYMMETRY_ETERNAL,A->symmetric_eternal);
8054: }
8055: if (A->structurally_symmetric_set) {
8056: MatSetOption(B,MAT_STRUCTURALLY_SYMMETRIC,A->structurally_symmetric);
8057: }
8058: if (A->hermitian_set) {
8059: MatSetOption(B,MAT_HERMITIAN,A->hermitian);
8060: }
8061: if (A->spd_set) {
8062: MatSetOption(B,MAT_SPD,A->spd);
8063: }
8064: if (A->symmetric_set) {
8065: MatSetOption(B,MAT_SYMMETRIC,A->symmetric);
8066: }
8067: return(0);
8068: }
8070: /*@
8071: MatStashSetInitialSize - sets the sizes of the matrix stash, that is
8072: used during the assembly process to store values that belong to
8073: other processors.
8075: Not Collective
8077: Input Parameters:
8078: + mat - the matrix
8079: . size - the initial size of the stash.
8080: - bsize - the initial size of the block-stash(if used).
8082: Options Database Keys:
8083: + -matstash_initial_size <size> or <size0,size1,...sizep-1>
8084: - -matstash_block_initial_size <bsize> or <bsize0,bsize1,...bsizep-1>
8086: Level: intermediate
8088: Notes:
8089: The block-stash is used for values set with MatSetValuesBlocked() while
8090: the stash is used for values set with MatSetValues()
8092: Run with the option -info and look for output of the form
8093: MatAssemblyBegin_MPIXXX:Stash has MM entries, uses nn mallocs.
8094: to determine the appropriate value, MM, to use for size and
8095: MatAssemblyBegin_MPIXXX:Block-Stash has BMM entries, uses nn mallocs.
8096: to determine the value, BMM to use for bsize
8099: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashGetInfo()
8101: @*/
8102: PetscErrorCode MatStashSetInitialSize(Mat mat,PetscInt size, PetscInt bsize)
8103: {
8109: MatStashSetInitialSize_Private(&mat->stash,size);
8110: MatStashSetInitialSize_Private(&mat->bstash,bsize);
8111: return(0);
8112: }
8114: /*@
8115: MatInterpolateAdd - w = y + A*x or A'*x depending on the shape of
8116: the matrix
8118: Neighbor-wise Collective on Mat
8120: Input Parameters:
8121: + mat - the matrix
8122: . x,y - the vectors
8123: - w - where the result is stored
8125: Level: intermediate
8127: Notes:
8128: w may be the same vector as y.
8130: This allows one to use either the restriction or interpolation (its transpose)
8131: matrix to do the interpolation
8133: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict()
8135: @*/
8136: PetscErrorCode MatInterpolateAdd(Mat A,Vec x,Vec y,Vec w)
8137: {
8139: PetscInt M,N,Ny;
8147: MatCheckPreallocated(A,1);
8148: MatGetSize(A,&M,&N);
8149: VecGetSize(y,&Ny);
8150: if (M == Ny) {
8151: MatMultAdd(A,x,y,w);
8152: } else {
8153: MatMultTransposeAdd(A,x,y,w);
8154: }
8155: return(0);
8156: }
8158: /*@
8159: MatInterpolate - y = A*x or A'*x depending on the shape of
8160: the matrix
8162: Neighbor-wise Collective on Mat
8164: Input Parameters:
8165: + mat - the matrix
8166: - x,y - the vectors
8168: Level: intermediate
8170: Notes:
8171: This allows one to use either the restriction or interpolation (its transpose)
8172: matrix to do the interpolation
8174: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict()
8176: @*/
8177: PetscErrorCode MatInterpolate(Mat A,Vec x,Vec y)
8178: {
8180: PetscInt M,N,Ny;
8187: MatCheckPreallocated(A,1);
8188: MatGetSize(A,&M,&N);
8189: VecGetSize(y,&Ny);
8190: if (M == Ny) {
8191: MatMult(A,x,y);
8192: } else {
8193: MatMultTranspose(A,x,y);
8194: }
8195: return(0);
8196: }
8198: /*@
8199: MatRestrict - y = A*x or A'*x
8201: Neighbor-wise Collective on Mat
8203: Input Parameters:
8204: + mat - the matrix
8205: - x,y - the vectors
8207: Level: intermediate
8209: Notes:
8210: This allows one to use either the restriction or interpolation (its transpose)
8211: matrix to do the restriction
8213: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatInterpolate()
8215: @*/
8216: PetscErrorCode MatRestrict(Mat A,Vec x,Vec y)
8217: {
8219: PetscInt M,N,Ny;
8226: MatCheckPreallocated(A,1);
8228: MatGetSize(A,&M,&N);
8229: VecGetSize(y,&Ny);
8230: if (M == Ny) {
8231: MatMult(A,x,y);
8232: } else {
8233: MatMultTranspose(A,x,y);
8234: }
8235: return(0);
8236: }
8238: /*@
8239: MatGetNullSpace - retrieves the null space of a matrix.
8241: Logically Collective on Mat
8243: Input Parameters:
8244: + mat - the matrix
8245: - nullsp - the null space object
8247: Level: developer
8249: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatSetNullSpace()
8250: @*/
8251: PetscErrorCode MatGetNullSpace(Mat mat, MatNullSpace *nullsp)
8252: {
8256: *nullsp = (mat->symmetric_set && mat->symmetric && !mat->nullsp) ? mat->transnullsp : mat->nullsp;
8257: return(0);
8258: }
8260: /*@
8261: MatSetNullSpace - attaches a null space to a matrix.
8263: Logically Collective on Mat
8265: Input Parameters:
8266: + mat - the matrix
8267: - nullsp - the null space object
8269: Level: advanced
8271: Notes:
8272: This null space is used by the linear solvers. Overwrites any previous null space that may have been attached
8274: For inconsistent singular systems (linear systems where the right hand side is not in the range of the operator) you also likely should
8275: call MatSetTransposeNullSpace(). This allows the linear system to be solved in a least squares sense.
8277: You can remove the null space by calling this routine with an nullsp of NULL
8280: The fundamental theorem of linear algebra (Gilbert Strang, Introduction to Applied Mathematics, page 72) states that
8281: the domain of a matrix A (from R^n to R^m (m rows, n columns) R^n = the direct sum of the null space of A, n(A), + the range of A^T, R(A^T).
8282: Similarly R^m = direct sum n(A^T) + R(A). Hence the linear system A x = b has a solution only if b in R(A) (or correspondingly b is orthogonal to
8283: n(A^T)) and if x is a solution then x + alpha n(A) is a solution for any alpha. The minimum norm solution is orthogonal to n(A). For problems without a solution
8284: the solution that minimizes the norm of the residual (the least squares solution) can be obtained by solving A x = \hat{b} where \hat{b} is b orthogonalized to the n(A^T).
8286: Krylov solvers can produce the minimal norm solution to the least squares problem by utilizing MatNullSpaceRemove().
8288: If the matrix is known to be symmetric because it is an SBAIJ matrix or one as called MatSetOption(mat,MAT_SYMMETRIC or MAT_SYMMETRIC_ETERNAL,PETSC_TRUE); this
8289: routine also automatically calls MatSetTransposeNullSpace().
8291: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatGetNullSpace(), MatSetTransposeNullSpace(), MatGetTransposeNullSpace(), MatNullSpaceRemove()
8292: @*/
8293: PetscErrorCode MatSetNullSpace(Mat mat,MatNullSpace nullsp)
8294: {
8300: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8301: MatNullSpaceDestroy(&mat->nullsp);
8302: mat->nullsp = nullsp;
8303: if (mat->symmetric_set && mat->symmetric) {
8304: MatSetTransposeNullSpace(mat,nullsp);
8305: }
8306: return(0);
8307: }
8309: /*@
8310: MatGetTransposeNullSpace - retrieves the null space of the transpose of a matrix.
8312: Logically Collective on Mat
8314: Input Parameters:
8315: + mat - the matrix
8316: - nullsp - the null space object
8318: Level: developer
8320: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatSetTransposeNullSpace(), MatSetNullSpace(), MatGetNullSpace()
8321: @*/
8322: PetscErrorCode MatGetTransposeNullSpace(Mat mat, MatNullSpace *nullsp)
8323: {
8328: *nullsp = (mat->symmetric_set && mat->symmetric && !mat->transnullsp) ? mat->nullsp : mat->transnullsp;
8329: return(0);
8330: }
8332: /*@
8333: MatSetTransposeNullSpace - attaches a null space to a matrix.
8335: Logically Collective on Mat
8337: Input Parameters:
8338: + mat - the matrix
8339: - nullsp - the null space object
8341: Level: advanced
8343: Notes:
8344: For inconsistent singular systems (linear systems where the right hand side is not in the range of the operator) this allows the linear system to be solved in a least squares sense.
8345: You must also call MatSetNullSpace()
8348: The fundamental theorem of linear algebra (Gilbert Strang, Introduction to Applied Mathematics, page 72) states that
8349: the domain of a matrix A (from R^n to R^m (m rows, n columns) R^n = the direct sum of the null space of A, n(A), + the range of A^T, R(A^T).
8350: Similarly R^m = direct sum n(A^T) + R(A). Hence the linear system A x = b has a solution only if b in R(A) (or correspondingly b is orthogonal to
8351: n(A^T)) and if x is a solution then x + alpha n(A) is a solution for any alpha. The minimum norm solution is orthogonal to n(A). For problems without a solution
8352: the solution that minimizes the norm of the residual (the least squares solution) can be obtained by solving A x = \hat{b} where \hat{b} is b orthogonalized to the n(A^T).
8354: Krylov solvers can produce the minimal norm solution to the least squares problem by utilizing MatNullSpaceRemove().
8356: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatGetNullSpace(), MatSetNullSpace(), MatGetTransposeNullSpace(), MatNullSpaceRemove()
8357: @*/
8358: PetscErrorCode MatSetTransposeNullSpace(Mat mat,MatNullSpace nullsp)
8359: {
8365: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8366: MatNullSpaceDestroy(&mat->transnullsp);
8367: mat->transnullsp = nullsp;
8368: return(0);
8369: }
8371: /*@
8372: MatSetNearNullSpace - attaches a null space to a matrix, which is often the null space (rigid body modes) of the operator without boundary conditions
8373: This null space will be used to provide near null space vectors to a multigrid preconditioner built from this matrix.
8375: Logically Collective on Mat
8377: Input Parameters:
8378: + mat - the matrix
8379: - nullsp - the null space object
8381: Level: advanced
8383: Notes:
8384: Overwrites any previous near null space that may have been attached
8386: You can remove the null space by calling this routine with an nullsp of NULL
8388: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNullSpace(), MatNullSpaceCreateRigidBody(), MatGetNearNullSpace()
8389: @*/
8390: PetscErrorCode MatSetNearNullSpace(Mat mat,MatNullSpace nullsp)
8391: {
8398: MatCheckPreallocated(mat,1);
8399: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8400: MatNullSpaceDestroy(&mat->nearnullsp);
8401: mat->nearnullsp = nullsp;
8402: return(0);
8403: }
8405: /*@
8406: MatGetNearNullSpace - Get null space attached with MatSetNearNullSpace()
8408: Not Collective
8410: Input Parameter:
8411: . mat - the matrix
8413: Output Parameter:
8414: . nullsp - the null space object, NULL if not set
8416: Level: developer
8418: .seealso: MatSetNearNullSpace(), MatGetNullSpace(), MatNullSpaceCreate()
8419: @*/
8420: PetscErrorCode MatGetNearNullSpace(Mat mat,MatNullSpace *nullsp)
8421: {
8426: MatCheckPreallocated(mat,1);
8427: *nullsp = mat->nearnullsp;
8428: return(0);
8429: }
8431: /*@C
8432: MatICCFactor - Performs in-place incomplete Cholesky factorization of matrix.
8434: Collective on Mat
8436: Input Parameters:
8437: + mat - the matrix
8438: . row - row/column permutation
8439: . fill - expected fill factor >= 1.0
8440: - level - level of fill, for ICC(k)
8442: Notes:
8443: Probably really in-place only when level of fill is zero, otherwise allocates
8444: new space to store factored matrix and deletes previous memory.
8446: Most users should employ the simplified KSP interface for linear solvers
8447: instead of working directly with matrix algebra routines such as this.
8448: See, e.g., KSPCreate().
8450: Level: developer
8453: .seealso: MatICCFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor()
8455: Developer Note: fortran interface is not autogenerated as the f90
8456: interface defintion cannot be generated correctly [due to MatFactorInfo]
8458: @*/
8459: PetscErrorCode MatICCFactor(Mat mat,IS row,const MatFactorInfo *info)
8460: {
8468: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"matrix must be square");
8469: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
8470: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
8471: if (!mat->ops->iccfactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8472: MatCheckPreallocated(mat,1);
8473: (*mat->ops->iccfactor)(mat,row,info);
8474: PetscObjectStateIncrease((PetscObject)mat);
8475: return(0);
8476: }
8478: /*@
8479: MatDiagonalScaleLocal - Scales columns of a matrix given the scaling values including the
8480: ghosted ones.
8482: Not Collective
8484: Input Parameters:
8485: + mat - the matrix
8486: - diag = the diagonal values, including ghost ones
8488: Level: developer
8490: Notes:
8491: Works only for MPIAIJ and MPIBAIJ matrices
8493: .seealso: MatDiagonalScale()
8494: @*/
8495: PetscErrorCode MatDiagonalScaleLocal(Mat mat,Vec diag)
8496: {
8498: PetscMPIInt size;
8505: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled");
8506: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
8507: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
8508: if (size == 1) {
8509: PetscInt n,m;
8510: VecGetSize(diag,&n);
8511: MatGetSize(mat,NULL,&m);
8512: if (m == n) {
8513: MatDiagonalScale(mat,NULL,diag);
8514: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only supported for sequential matrices when no ghost points/periodic conditions");
8515: } else {
8516: PetscUseMethod(mat,"MatDiagonalScaleLocal_C",(Mat,Vec),(mat,diag));
8517: }
8518: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
8519: PetscObjectStateIncrease((PetscObject)mat);
8520: return(0);
8521: }
8523: /*@
8524: MatGetInertia - Gets the inertia from a factored matrix
8526: Collective on Mat
8528: Input Parameter:
8529: . mat - the matrix
8531: Output Parameters:
8532: + nneg - number of negative eigenvalues
8533: . nzero - number of zero eigenvalues
8534: - npos - number of positive eigenvalues
8536: Level: advanced
8538: Notes:
8539: Matrix must have been factored by MatCholeskyFactor()
8542: @*/
8543: PetscErrorCode MatGetInertia(Mat mat,PetscInt *nneg,PetscInt *nzero,PetscInt *npos)
8544: {
8550: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
8551: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Numeric factor mat is not assembled");
8552: if (!mat->ops->getinertia) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8553: (*mat->ops->getinertia)(mat,nneg,nzero,npos);
8554: return(0);
8555: }
8557: /* ----------------------------------------------------------------*/
8558: /*@C
8559: MatSolves - Solves A x = b, given a factored matrix, for a collection of vectors
8561: Neighbor-wise Collective on Mats
8563: Input Parameters:
8564: + mat - the factored matrix
8565: - b - the right-hand-side vectors
8567: Output Parameter:
8568: . x - the result vectors
8570: Notes:
8571: The vectors b and x cannot be the same. I.e., one cannot
8572: call MatSolves(A,x,x).
8574: Notes:
8575: Most users should employ the simplified KSP interface for linear solvers
8576: instead of working directly with matrix algebra routines such as this.
8577: See, e.g., KSPCreate().
8579: Level: developer
8581: .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd(), MatSolve()
8582: @*/
8583: PetscErrorCode MatSolves(Mat mat,Vecs b,Vecs x)
8584: {
8590: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
8591: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
8592: if (!mat->rmap->N && !mat->cmap->N) return(0);
8594: if (!mat->ops->solves) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8595: MatCheckPreallocated(mat,1);
8596: PetscLogEventBegin(MAT_Solves,mat,0,0,0);
8597: (*mat->ops->solves)(mat,b,x);
8598: PetscLogEventEnd(MAT_Solves,mat,0,0,0);
8599: return(0);
8600: }
8602: /*@
8603: MatIsSymmetric - Test whether a matrix is symmetric
8605: Collective on Mat
8607: Input Parameter:
8608: + A - the matrix to test
8609: - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact transpose)
8611: Output Parameters:
8612: . flg - the result
8614: Notes:
8615: For real numbers MatIsSymmetric() and MatIsHermitian() return identical results
8617: Level: intermediate
8619: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown()
8620: @*/
8621: PetscErrorCode MatIsSymmetric(Mat A,PetscReal tol,PetscBool *flg)
8622: {
8629: if (!A->symmetric_set) {
8630: if (!A->ops->issymmetric) {
8631: MatType mattype;
8632: MatGetType(A,&mattype);
8633: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type %s does not support checking for symmetric",mattype);
8634: }
8635: (*A->ops->issymmetric)(A,tol,flg);
8636: if (!tol) {
8637: MatSetOption(A,MAT_SYMMETRIC,*flg);
8638: }
8639: } else if (A->symmetric) {
8640: *flg = PETSC_TRUE;
8641: } else if (!tol) {
8642: *flg = PETSC_FALSE;
8643: } else {
8644: if (!A->ops->issymmetric) {
8645: MatType mattype;
8646: MatGetType(A,&mattype);
8647: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type %s does not support checking for symmetric",mattype);
8648: }
8649: (*A->ops->issymmetric)(A,tol,flg);
8650: }
8651: return(0);
8652: }
8654: /*@
8655: MatIsHermitian - Test whether a matrix is Hermitian
8657: Collective on Mat
8659: Input Parameter:
8660: + A - the matrix to test
8661: - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact Hermitian)
8663: Output Parameters:
8664: . flg - the result
8666: Level: intermediate
8668: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(),
8669: MatIsSymmetricKnown(), MatIsSymmetric()
8670: @*/
8671: PetscErrorCode MatIsHermitian(Mat A,PetscReal tol,PetscBool *flg)
8672: {
8679: if (!A->hermitian_set) {
8680: if (!A->ops->ishermitian) {
8681: MatType mattype;
8682: MatGetType(A,&mattype);
8683: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type %s does not support checking for hermitian",mattype);
8684: }
8685: (*A->ops->ishermitian)(A,tol,flg);
8686: if (!tol) {
8687: MatSetOption(A,MAT_HERMITIAN,*flg);
8688: }
8689: } else if (A->hermitian) {
8690: *flg = PETSC_TRUE;
8691: } else if (!tol) {
8692: *flg = PETSC_FALSE;
8693: } else {
8694: if (!A->ops->ishermitian) {
8695: MatType mattype;
8696: MatGetType(A,&mattype);
8697: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type %s does not support checking for hermitian",mattype);
8698: }
8699: (*A->ops->ishermitian)(A,tol,flg);
8700: }
8701: return(0);
8702: }
8704: /*@
8705: MatIsSymmetricKnown - Checks the flag on the matrix to see if it is symmetric.
8707: Not Collective
8709: Input Parameter:
8710: . A - the matrix to check
8712: Output Parameters:
8713: + set - if the symmetric flag is set (this tells you if the next flag is valid)
8714: - flg - the result
8716: Level: advanced
8718: Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsSymmetric()
8719: if you want it explicitly checked
8721: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric()
8722: @*/
8723: PetscErrorCode MatIsSymmetricKnown(Mat A,PetscBool *set,PetscBool *flg)
8724: {
8729: if (A->symmetric_set) {
8730: *set = PETSC_TRUE;
8731: *flg = A->symmetric;
8732: } else {
8733: *set = PETSC_FALSE;
8734: }
8735: return(0);
8736: }
8738: /*@
8739: MatIsHermitianKnown - Checks the flag on the matrix to see if it is hermitian.
8741: Not Collective
8743: Input Parameter:
8744: . A - the matrix to check
8746: Output Parameters:
8747: + set - if the hermitian flag is set (this tells you if the next flag is valid)
8748: - flg - the result
8750: Level: advanced
8752: Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsHermitian()
8753: if you want it explicitly checked
8755: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric()
8756: @*/
8757: PetscErrorCode MatIsHermitianKnown(Mat A,PetscBool *set,PetscBool *flg)
8758: {
8763: if (A->hermitian_set) {
8764: *set = PETSC_TRUE;
8765: *flg = A->hermitian;
8766: } else {
8767: *set = PETSC_FALSE;
8768: }
8769: return(0);
8770: }
8772: /*@
8773: MatIsStructurallySymmetric - Test whether a matrix is structurally symmetric
8775: Collective on Mat
8777: Input Parameter:
8778: . A - the matrix to test
8780: Output Parameters:
8781: . flg - the result
8783: Level: intermediate
8785: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsSymmetric(), MatSetOption()
8786: @*/
8787: PetscErrorCode MatIsStructurallySymmetric(Mat A,PetscBool *flg)
8788: {
8794: if (!A->structurally_symmetric_set) {
8795: if (!A->ops->isstructurallysymmetric) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Matrix of type %s does not support checking for structural symmetric",((PetscObject)A)->type_name);
8796: (*A->ops->isstructurallysymmetric)(A,flg);
8797: MatSetOption(A,MAT_STRUCTURALLY_SYMMETRIC,*flg);
8798: } else *flg = A->structurally_symmetric;
8799: return(0);
8800: }
8802: /*@
8803: MatStashGetInfo - Gets how many values are currently in the matrix stash, i.e. need
8804: to be communicated to other processors during the MatAssemblyBegin/End() process
8806: Not collective
8808: Input Parameter:
8809: . vec - the vector
8811: Output Parameters:
8812: + nstash - the size of the stash
8813: . reallocs - the number of additional mallocs incurred.
8814: . bnstash - the size of the block stash
8815: - breallocs - the number of additional mallocs incurred.in the block stash
8817: Level: advanced
8819: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashSetInitialSize()
8821: @*/
8822: PetscErrorCode MatStashGetInfo(Mat mat,PetscInt *nstash,PetscInt *reallocs,PetscInt *bnstash,PetscInt *breallocs)
8823: {
8827: MatStashGetInfo_Private(&mat->stash,nstash,reallocs);
8828: MatStashGetInfo_Private(&mat->bstash,bnstash,breallocs);
8829: return(0);
8830: }
8832: /*@C
8833: MatCreateVecs - Get vector(s) compatible with the matrix, i.e. with the same
8834: parallel layout
8836: Collective on Mat
8838: Input Parameter:
8839: . mat - the matrix
8841: Output Parameter:
8842: + right - (optional) vector that the matrix can be multiplied against
8843: - left - (optional) vector that the matrix vector product can be stored in
8845: Notes:
8846: The blocksize of the returned vectors is determined by the row and column block sizes set with MatSetBlockSizes() or the single blocksize (same for both) set by MatSetBlockSize().
8848: Notes:
8849: These are new vectors which are not owned by the Mat, they should be destroyed in VecDestroy() when no longer needed
8851: Level: advanced
8853: .seealso: MatCreate(), VecDestroy()
8854: @*/
8855: PetscErrorCode MatCreateVecs(Mat mat,Vec *right,Vec *left)
8856: {
8862: if (mat->ops->getvecs) {
8863: (*mat->ops->getvecs)(mat,right,left);
8864: } else {
8865: PetscInt rbs,cbs;
8866: MatGetBlockSizes(mat,&rbs,&cbs);
8867: if (right) {
8868: if (mat->cmap->n < 0) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"PetscLayout for columns not yet setup");
8869: VecCreate(PetscObjectComm((PetscObject)mat),right);
8870: VecSetSizes(*right,mat->cmap->n,PETSC_DETERMINE);
8871: VecSetBlockSize(*right,cbs);
8872: VecSetType(*right,mat->defaultvectype);
8873: PetscLayoutReference(mat->cmap,&(*right)->map);
8874: }
8875: if (left) {
8876: if (mat->rmap->n < 0) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"PetscLayout for rows not yet setup");
8877: VecCreate(PetscObjectComm((PetscObject)mat),left);
8878: VecSetSizes(*left,mat->rmap->n,PETSC_DETERMINE);
8879: VecSetBlockSize(*left,rbs);
8880: VecSetType(*left,mat->defaultvectype);
8881: PetscLayoutReference(mat->rmap,&(*left)->map);
8882: }
8883: }
8884: return(0);
8885: }
8887: /*@C
8888: MatFactorInfoInitialize - Initializes a MatFactorInfo data structure
8889: with default values.
8891: Not Collective
8893: Input Parameters:
8894: . info - the MatFactorInfo data structure
8897: Notes:
8898: The solvers are generally used through the KSP and PC objects, for example
8899: PCLU, PCILU, PCCHOLESKY, PCICC
8901: Level: developer
8903: .seealso: MatFactorInfo
8905: Developer Note: fortran interface is not autogenerated as the f90
8906: interface defintion cannot be generated correctly [due to MatFactorInfo]
8908: @*/
8910: PetscErrorCode MatFactorInfoInitialize(MatFactorInfo *info)
8911: {
8915: PetscMemzero(info,sizeof(MatFactorInfo));
8916: return(0);
8917: }
8919: /*@
8920: MatFactorSetSchurIS - Set indices corresponding to the Schur complement you wish to have computed
8922: Collective on Mat
8924: Input Parameters:
8925: + mat - the factored matrix
8926: - is - the index set defining the Schur indices (0-based)
8928: Notes:
8929: Call MatFactorSolveSchurComplement() or MatFactorSolveSchurComplementTranspose() after this call to solve a Schur complement system.
8931: You can call MatFactorGetSchurComplement() or MatFactorCreateSchurComplement() after this call.
8933: Level: developer
8935: .seealso: MatGetFactor(), MatFactorGetSchurComplement(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSolveSchurComplement(),
8936: MatFactorSolveSchurComplementTranspose(), MatFactorSolveSchurComplement()
8938: @*/
8939: PetscErrorCode MatFactorSetSchurIS(Mat mat,IS is)
8940: {
8941: PetscErrorCode ierr,(*f)(Mat,IS);
8949: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
8950: PetscObjectQueryFunction((PetscObject)mat,"MatFactorSetSchurIS_C",&f);
8951: if (!f) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"The selected MatSolverType does not support Schur complement computation. You should use MATSOLVERMUMPS or MATSOLVERMKL_PARDISO");
8952: MatDestroy(&mat->schur);
8953: (*f)(mat,is);
8954: if (!mat->schur) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_PLIB,"Schur complement has not been created");
8955: return(0);
8956: }
8958: /*@
8959: MatFactorCreateSchurComplement - Create a Schur complement matrix object using Schur data computed during the factorization step
8961: Logically Collective on Mat
8963: Input Parameters:
8964: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
8965: . S - location where to return the Schur complement, can be NULL
8966: - status - the status of the Schur complement matrix, can be NULL
8968: Notes:
8969: You must call MatFactorSetSchurIS() before calling this routine.
8971: The routine provides a copy of the Schur matrix stored within the solver data structures.
8972: The caller must destroy the object when it is no longer needed.
8973: If MatFactorInvertSchurComplement() has been called, the routine gets back the inverse.
8975: Use MatFactorGetSchurComplement() to get access to the Schur complement matrix inside the factored matrix instead of making a copy of it (which this function does)
8977: Developer Notes:
8978: The reason this routine exists is because the representation of the Schur complement within the factor matrix may be different than a standard PETSc
8979: matrix representation and we normally do not want to use the time or memory to make a copy as a regular PETSc matrix.
8981: See MatCreateSchurComplement() or MatGetSchurComplement() for ways to create virtual or approximate Schur complements.
8983: Level: advanced
8985: References:
8987: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorGetSchurComplement(), MatFactorSchurStatus
8988: @*/
8989: PetscErrorCode MatFactorCreateSchurComplement(Mat F,Mat* S,MatFactorSchurStatus* status)
8990: {
8997: if (S) {
8998: PetscErrorCode (*f)(Mat,Mat*);
9000: PetscObjectQueryFunction((PetscObject)F,"MatFactorCreateSchurComplement_C",&f);
9001: if (f) {
9002: (*f)(F,S);
9003: } else {
9004: MatDuplicate(F->schur,MAT_COPY_VALUES,S);
9005: }
9006: }
9007: if (status) *status = F->schur_status;
9008: return(0);
9009: }
9011: /*@
9012: MatFactorGetSchurComplement - Gets access to a Schur complement matrix using the current Schur data within a factored matrix
9014: Logically Collective on Mat
9016: Input Parameters:
9017: + F - the factored matrix obtained by calling MatGetFactor()
9018: . *S - location where to return the Schur complement, can be NULL
9019: - status - the status of the Schur complement matrix, can be NULL
9021: Notes:
9022: You must call MatFactorSetSchurIS() before calling this routine.
9024: Schur complement mode is currently implemented for sequential matrices.
9025: The routine returns a the Schur Complement stored within the data strutures of the solver.
9026: If MatFactorInvertSchurComplement() has previously been called, the returned matrix is actually the inverse of the Schur complement.
9027: The returned matrix should not be destroyed; the caller should call MatFactorRestoreSchurComplement() when the object is no longer needed.
9029: Use MatFactorCreateSchurComplement() to create a copy of the Schur complement matrix that is within a factored matrix
9031: See MatCreateSchurComplement() or MatGetSchurComplement() for ways to create virtual or approximate Schur complements.
9033: Level: advanced
9035: References:
9037: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSchurStatus
9038: @*/
9039: PetscErrorCode MatFactorGetSchurComplement(Mat F,Mat* S,MatFactorSchurStatus* status)
9040: {
9045: if (S) *S = F->schur;
9046: if (status) *status = F->schur_status;
9047: return(0);
9048: }
9050: /*@
9051: MatFactorRestoreSchurComplement - Restore the Schur complement matrix object obtained from a call to MatFactorGetSchurComplement
9053: Logically Collective on Mat
9055: Input Parameters:
9056: + F - the factored matrix obtained by calling MatGetFactor()
9057: . *S - location where the Schur complement is stored
9058: - status - the status of the Schur complement matrix (see MatFactorSchurStatus)
9060: Notes:
9062: Level: advanced
9064: References:
9066: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSchurStatus
9067: @*/
9068: PetscErrorCode MatFactorRestoreSchurComplement(Mat F,Mat* S,MatFactorSchurStatus status)
9069: {
9074: if (S) {
9076: *S = NULL;
9077: }
9078: F->schur_status = status;
9079: MatFactorUpdateSchurStatus_Private(F);
9080: return(0);
9081: }
9083: /*@
9084: MatFactorSolveSchurComplementTranspose - Solve the transpose of the Schur complement system computed during the factorization step
9086: Logically Collective on Mat
9088: Input Parameters:
9089: + F - the factored matrix obtained by calling MatGetFactor()
9090: . rhs - location where the right hand side of the Schur complement system is stored
9091: - sol - location where the solution of the Schur complement system has to be returned
9093: Notes:
9094: The sizes of the vectors should match the size of the Schur complement
9096: Must be called after MatFactorSetSchurIS()
9098: Level: advanced
9100: References:
9102: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorSolveSchurComplement()
9103: @*/
9104: PetscErrorCode MatFactorSolveSchurComplementTranspose(Mat F, Vec rhs, Vec sol)
9105: {
9117: MatFactorFactorizeSchurComplement(F);
9118: switch (F->schur_status) {
9119: case MAT_FACTOR_SCHUR_FACTORED:
9120: MatSolveTranspose(F->schur,rhs,sol);
9121: break;
9122: case MAT_FACTOR_SCHUR_INVERTED:
9123: MatMultTranspose(F->schur,rhs,sol);
9124: break;
9125: default:
9126: SETERRQ1(PetscObjectComm((PetscObject)F),PETSC_ERR_SUP,"Unhandled MatFactorSchurStatus %D",F->schur_status);
9127: break;
9128: }
9129: return(0);
9130: }
9132: /*@
9133: MatFactorSolveSchurComplement - Solve the Schur complement system computed during the factorization step
9135: Logically Collective on Mat
9137: Input Parameters:
9138: + F - the factored matrix obtained by calling MatGetFactor()
9139: . rhs - location where the right hand side of the Schur complement system is stored
9140: - sol - location where the solution of the Schur complement system has to be returned
9142: Notes:
9143: The sizes of the vectors should match the size of the Schur complement
9145: Must be called after MatFactorSetSchurIS()
9147: Level: advanced
9149: References:
9151: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorSolveSchurComplementTranspose()
9152: @*/
9153: PetscErrorCode MatFactorSolveSchurComplement(Mat F, Vec rhs, Vec sol)
9154: {
9166: MatFactorFactorizeSchurComplement(F);
9167: switch (F->schur_status) {
9168: case MAT_FACTOR_SCHUR_FACTORED:
9169: MatSolve(F->schur,rhs,sol);
9170: break;
9171: case MAT_FACTOR_SCHUR_INVERTED:
9172: MatMult(F->schur,rhs,sol);
9173: break;
9174: default:
9175: SETERRQ1(PetscObjectComm((PetscObject)F),PETSC_ERR_SUP,"Unhandled MatFactorSchurStatus %D",F->schur_status);
9176: break;
9177: }
9178: return(0);
9179: }
9181: /*@
9182: MatFactorInvertSchurComplement - Invert the Schur complement matrix computed during the factorization step
9184: Logically Collective on Mat
9186: Input Parameters:
9187: . F - the factored matrix obtained by calling MatGetFactor()
9189: Notes:
9190: Must be called after MatFactorSetSchurIS().
9192: Call MatFactorGetSchurComplement() or MatFactorCreateSchurComplement() AFTER this call to actually compute the inverse and get access to it.
9194: Level: advanced
9196: References:
9198: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorGetSchurComplement(), MatFactorCreateSchurComplement()
9199: @*/
9200: PetscErrorCode MatFactorInvertSchurComplement(Mat F)
9201: {
9207: if (F->schur_status == MAT_FACTOR_SCHUR_INVERTED) return(0);
9208: MatFactorFactorizeSchurComplement(F);
9209: MatFactorInvertSchurComplement_Private(F);
9210: F->schur_status = MAT_FACTOR_SCHUR_INVERTED;
9211: return(0);
9212: }
9214: /*@
9215: MatFactorFactorizeSchurComplement - Factorize the Schur complement matrix computed during the factorization step
9217: Logically Collective on Mat
9219: Input Parameters:
9220: . F - the factored matrix obtained by calling MatGetFactor()
9222: Notes:
9223: Must be called after MatFactorSetSchurIS().
9225: Level: advanced
9227: References:
9229: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorInvertSchurComplement()
9230: @*/
9231: PetscErrorCode MatFactorFactorizeSchurComplement(Mat F)
9232: {
9238: if (F->schur_status == MAT_FACTOR_SCHUR_INVERTED || F->schur_status == MAT_FACTOR_SCHUR_FACTORED) return(0);
9239: MatFactorFactorizeSchurComplement_Private(F);
9240: F->schur_status = MAT_FACTOR_SCHUR_FACTORED;
9241: return(0);
9242: }
9244: /*@
9245: MatPtAP - Creates the matrix product C = P^T * A * P
9247: Neighbor-wise Collective on Mat
9249: Input Parameters:
9250: + A - the matrix
9251: . P - the projection matrix
9252: . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
9253: - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(P)), use PETSC_DEFAULT if you do not have a good estimate
9254: if the result is a dense matrix this is irrelevent
9256: Output Parameters:
9257: . C - the product matrix
9259: Notes:
9260: C will be created and must be destroyed by the user with MatDestroy().
9262: For matrix types without special implementation the function fallbacks to MatMatMult() followed by MatTransposeMatMult().
9264: Level: intermediate
9266: .seealso: MatMatMult(), MatRARt()
9267: @*/
9268: PetscErrorCode MatPtAP(Mat A,Mat P,MatReuse scall,PetscReal fill,Mat *C)
9269: {
9273: if (scall == MAT_REUSE_MATRIX) MatCheckProduct(*C,5);
9274: if (scall == MAT_INPLACE_MATRIX) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Inplace product not supported");
9276: if (scall == MAT_INITIAL_MATRIX) {
9277: MatProductCreate(A,P,NULL,C);
9278: MatProductSetType(*C,MATPRODUCT_PtAP);
9279: MatProductSetAlgorithm(*C,"default");
9280: MatProductSetFill(*C,fill);
9282: (*C)->product->api_user = PETSC_TRUE;
9283: MatProductSetFromOptions(*C);
9284: if (!(*C)->ops->productsymbolic) SETERRQ3(PetscObjectComm((PetscObject)(*C)),PETSC_ERR_SUP,"MatProduct %s not supported for A %s and P %s",MatProductTypes[MATPRODUCT_PtAP],((PetscObject)A)->type_name,((PetscObject)P)->type_name);
9285: MatProductSymbolic(*C);
9286: } else { /* scall == MAT_REUSE_MATRIX */
9287: MatProductReplaceMats(A,P,NULL,*C);
9288: }
9290: MatProductNumeric(*C);
9291: if (A->symmetric_set && A->symmetric) {
9292: MatSetOption(*C,MAT_SYMMETRIC,PETSC_TRUE);
9293: }
9294: return(0);
9295: }
9297: /*@
9298: MatRARt - Creates the matrix product C = R * A * R^T
9300: Neighbor-wise Collective on Mat
9302: Input Parameters:
9303: + A - the matrix
9304: . R - the projection matrix
9305: . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
9306: - fill - expected fill as ratio of nnz(C)/nnz(A), use PETSC_DEFAULT if you do not have a good estimate
9307: if the result is a dense matrix this is irrelevent
9309: Output Parameters:
9310: . C - the product matrix
9312: Notes:
9313: C will be created and must be destroyed by the user with MatDestroy().
9315: This routine is currently only implemented for pairs of AIJ matrices and classes
9316: which inherit from AIJ. Due to PETSc sparse matrix block row distribution among processes,
9317: parallel MatRARt is implemented via explicit transpose of R, which could be very expensive.
9318: We recommend using MatPtAP().
9320: Level: intermediate
9322: .seealso: