Actual source code: matrix.c
petsc-master 2019-08-22
2: /*
3: This is where the abstract matrix operations are defined
4: */
6: #include <petsc/private/matimpl.h>
7: #include <petsc/private/isimpl.h>
8: #include <petsc/private/vecimpl.h>
10: /* Logging support */
11: PetscClassId MAT_CLASSID;
12: PetscClassId MAT_COLORING_CLASSID;
13: PetscClassId MAT_FDCOLORING_CLASSID;
14: PetscClassId MAT_TRANSPOSECOLORING_CLASSID;
16: PetscLogEvent MAT_Mult, MAT_Mults, MAT_MultConstrained, MAT_MultAdd, MAT_MultTranspose;
17: PetscLogEvent MAT_MultTransposeConstrained, MAT_MultTransposeAdd, MAT_Solve, MAT_Solves, MAT_SolveAdd, MAT_SolveTranspose, MAT_MatSolve,MAT_MatTrSolve;
18: PetscLogEvent MAT_SolveTransposeAdd, MAT_SOR, MAT_ForwardSolve, MAT_BackwardSolve, MAT_LUFactor, MAT_LUFactorSymbolic;
19: PetscLogEvent MAT_LUFactorNumeric, MAT_CholeskyFactor, MAT_CholeskyFactorSymbolic, MAT_CholeskyFactorNumeric, MAT_ILUFactor;
20: PetscLogEvent MAT_ILUFactorSymbolic, MAT_ICCFactorSymbolic, MAT_Copy, MAT_Convert, MAT_Scale, MAT_AssemblyBegin;
21: PetscLogEvent MAT_AssemblyEnd, MAT_SetValues, MAT_GetValues, MAT_GetRow, MAT_GetRowIJ, MAT_CreateSubMats, MAT_GetOrdering, MAT_RedundantMat, MAT_GetSeqNonzeroStructure;
22: PetscLogEvent MAT_IncreaseOverlap, MAT_Partitioning, MAT_PartitioningND, MAT_Coarsen, MAT_ZeroEntries, MAT_Load, MAT_View, MAT_AXPY, MAT_FDColoringCreate;
23: PetscLogEvent MAT_FDColoringSetUp, MAT_FDColoringApply,MAT_Transpose,MAT_FDColoringFunction, MAT_CreateSubMat;
24: PetscLogEvent MAT_TransposeColoringCreate;
25: PetscLogEvent MAT_MatMult, MAT_MatMultSymbolic, MAT_MatMultNumeric;
26: PetscLogEvent MAT_PtAP, MAT_PtAPSymbolic, MAT_PtAPNumeric,MAT_RARt, MAT_RARtSymbolic, MAT_RARtNumeric;
27: PetscLogEvent MAT_MatTransposeMult, MAT_MatTransposeMultSymbolic, MAT_MatTransposeMultNumeric;
28: PetscLogEvent MAT_TransposeMatMult, MAT_TransposeMatMultSymbolic, MAT_TransposeMatMultNumeric;
29: PetscLogEvent MAT_MatMatMult, MAT_MatMatMultSymbolic, MAT_MatMatMultNumeric;
30: PetscLogEvent MAT_MultHermitianTranspose,MAT_MultHermitianTransposeAdd;
31: PetscLogEvent MAT_Getsymtranspose, MAT_Getsymtransreduced, MAT_GetBrowsOfAcols;
32: PetscLogEvent MAT_GetBrowsOfAocols, MAT_Getlocalmat, MAT_Getlocalmatcondensed, MAT_Seqstompi, MAT_Seqstompinum, MAT_Seqstompisym;
33: PetscLogEvent MAT_Applypapt, MAT_Applypapt_numeric, MAT_Applypapt_symbolic, MAT_GetSequentialNonzeroStructure;
34: PetscLogEvent MAT_GetMultiProcBlock;
35: PetscLogEvent MAT_CUSPARSECopyToGPU, MAT_SetValuesBatch;
36: PetscLogEvent MAT_ViennaCLCopyToGPU;
37: PetscLogEvent MAT_DenseCopyToGPU, MAT_DenseCopyFromGPU;
38: PetscLogEvent MAT_Merge,MAT_Residual,MAT_SetRandom;
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_",0};
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
46: Logically Collective on Mat
48: Input Parameters:
49: + x - the matrix
50: - rctx - the random number context, formed by PetscRandomCreate(), or NULL and
51: it will create one internally.
53: Output Parameter:
54: . x - the matrix
56: Example of Usage:
57: .vb
58: PetscRandomCreate(PETSC_COMM_WORLD,&rctx);
59: MatSetRandom(x,rctx);
60: PetscRandomDestroy(rctx);
61: .ve
63: Level: intermediate
66: .seealso: MatZeroEntries(), MatSetValues(), PetscRandomCreate(), PetscRandomDestroy()
67: @*/
68: PetscErrorCode MatSetRandom(Mat x,PetscRandom rctx)
69: {
71: PetscRandom randObj = NULL;
78: if (!x->ops->setrandom) SETERRQ1(PetscObjectComm((PetscObject)x),PETSC_ERR_SUP,"Mat type %s",((PetscObject)x)->type_name);
80: if (!rctx) {
81: MPI_Comm comm;
82: PetscObjectGetComm((PetscObject)x,&comm);
83: PetscRandomCreate(comm,&randObj);
84: PetscRandomSetFromOptions(randObj);
85: rctx = randObj;
86: }
88: PetscLogEventBegin(MAT_SetRandom,x,rctx,0,0);
89: (*x->ops->setrandom)(x,rctx);
90: PetscLogEventEnd(MAT_SetRandom,x,rctx,0,0);
92: MatAssemblyBegin(x, MAT_FINAL_ASSEMBLY);
93: MatAssemblyEnd(x, MAT_FINAL_ASSEMBLY);
94: PetscRandomDestroy(&randObj);
95: return(0);
96: }
98: /*@
99: MatFactorGetErrorZeroPivot - returns the pivot value that was determined to be zero and the row it occurred in
101: Logically Collective on Mat
103: Input Parameters:
104: . mat - the factored matrix
106: Output Parameter:
107: + pivot - the pivot value computed
108: - row - the row that the zero pivot occurred. Note that this row must be interpreted carefully due to row reorderings and which processes
109: the share the matrix
111: Level: advanced
113: Notes:
114: This routine does not work for factorizations done with external packages.
115: This routine should only be called if MatGetFactorError() returns a value of MAT_FACTOR_NUMERIC_ZEROPIVOT
117: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
119: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatFactorSymbolic(), MatFactorClearError(), MatFactorGetErrorZeroPivot()
120: @*/
121: PetscErrorCode MatFactorGetErrorZeroPivot(Mat mat,PetscReal *pivot,PetscInt *row)
122: {
125: *pivot = mat->factorerror_zeropivot_value;
126: *row = mat->factorerror_zeropivot_row;
127: return(0);
128: }
130: /*@
131: MatFactorGetError - gets the error code from a factorization
133: Logically Collective on Mat
135: Input Parameters:
136: . mat - the factored matrix
138: Output Parameter:
139: . err - the error code
141: Level: advanced
143: Notes:
144: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
146: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatFactorSymbolic(), MatFactorClearError(), MatFactorGetErrorZeroPivot()
147: @*/
148: PetscErrorCode MatFactorGetError(Mat mat,MatFactorError *err)
149: {
152: *err = mat->factorerrortype;
153: return(0);
154: }
156: /*@
157: MatFactorClearError - clears the error code in a factorization
159: Logically Collective on Mat
161: Input Parameter:
162: . mat - the factored matrix
164: Level: developer
166: Notes:
167: This can be called on non-factored matrices that come from, for example, matrices used in SOR.
169: .seealso: MatZeroEntries(), MatFactor(), MatGetFactor(), MatFactorSymbolic(), MatFactorGetError(), MatFactorGetErrorZeroPivot()
170: @*/
171: PetscErrorCode MatFactorClearError(Mat mat)
172: {
175: mat->factorerrortype = MAT_FACTOR_NOERROR;
176: mat->factorerror_zeropivot_value = 0.0;
177: mat->factorerror_zeropivot_row = 0;
178: return(0);
179: }
181: PETSC_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat mat,PetscBool cols,PetscReal tol,IS *nonzero)
182: {
183: PetscErrorCode ierr;
184: Vec r,l;
185: const PetscScalar *al;
186: PetscInt i,nz,gnz,N,n;
189: MatCreateVecs(mat,&r,&l);
190: if (!cols) { /* nonzero rows */
191: MatGetSize(mat,&N,NULL);
192: MatGetLocalSize(mat,&n,NULL);
193: VecSet(l,0.0);
194: VecSetRandom(r,NULL);
195: MatMult(mat,r,l);
196: VecGetArrayRead(l,&al);
197: } else { /* nonzero columns */
198: MatGetSize(mat,NULL,&N);
199: MatGetLocalSize(mat,NULL,&n);
200: VecSet(r,0.0);
201: VecSetRandom(l,NULL);
202: MatMultTranspose(mat,l,r);
203: VecGetArrayRead(r,&al);
204: }
205: if (tol <= 0.0) { for (i=0,nz=0;i<n;i++) if (al[i] != 0.0) nz++; }
206: else { for (i=0,nz=0;i<n;i++) if (PetscAbsScalar(al[i]) > tol) nz++; }
207: MPIU_Allreduce(&nz,&gnz,1,MPIU_INT,MPI_SUM,PetscObjectComm((PetscObject)mat));
208: if (gnz != N) {
209: PetscInt *nzr;
210: PetscMalloc1(nz,&nzr);
211: if (nz) {
212: if (tol < 0) { for (i=0,nz=0;i<n;i++) if (al[i] != 0.0) nzr[nz++] = i; }
213: else { for (i=0,nz=0;i<n;i++) if (PetscAbsScalar(al[i]) > tol) nzr[nz++] = i; }
214: }
215: ISCreateGeneral(PetscObjectComm((PetscObject)mat),nz,nzr,PETSC_OWN_POINTER,nonzero);
216: } else *nonzero = NULL;
217: if (!cols) { /* nonzero rows */
218: VecRestoreArrayRead(l,&al);
219: } else {
220: VecRestoreArrayRead(r,&al);
221: }
222: VecDestroy(&l);
223: VecDestroy(&r);
224: return(0);
225: }
227: /*@
228: MatFindNonzeroRows - Locate all rows that are not completely zero in the matrix
230: Input Parameter:
231: . A - the matrix
233: Output Parameter:
234: . keptrows - the rows that are not completely zero
236: Notes:
237: keptrows is set to NULL if all rows are nonzero.
239: Level: intermediate
241: @*/
242: PetscErrorCode MatFindNonzeroRows(Mat mat,IS *keptrows)
243: {
250: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
251: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
252: if (!mat->ops->findnonzerorows) {
253: MatFindNonzeroRowsOrCols_Basic(mat,PETSC_FALSE,0.0,keptrows);
254: } else {
255: (*mat->ops->findnonzerorows)(mat,keptrows);
256: }
257: return(0);
258: }
260: /*@
261: MatFindZeroRows - Locate all rows that are completely zero in the matrix
263: Input Parameter:
264: . A - the matrix
266: Output Parameter:
267: . zerorows - the rows that are completely zero
269: Notes:
270: zerorows is set to NULL if no rows are zero.
272: Level: intermediate
274: @*/
275: PetscErrorCode MatFindZeroRows(Mat mat,IS *zerorows)
276: {
278: IS keptrows;
279: PetscInt m, n;
284: MatFindNonzeroRows(mat, &keptrows);
285: /* MatFindNonzeroRows sets keptrows to NULL if there are no zero rows.
286: In keeping with this convention, we set zerorows to NULL if there are no zero
287: rows. */
288: if (keptrows == NULL) {
289: *zerorows = NULL;
290: } else {
291: MatGetOwnershipRange(mat,&m,&n);
292: ISComplement(keptrows,m,n,zerorows);
293: ISDestroy(&keptrows);
294: }
295: return(0);
296: }
298: /*@
299: MatGetDiagonalBlock - Returns the part of the matrix associated with the on-process coupling
301: Not Collective
303: Input Parameters:
304: . A - the matrix
306: Output Parameters:
307: . a - the diagonal part (which is a SEQUENTIAL matrix)
309: Notes:
310: see the manual page for MatCreateAIJ() for more information on the "diagonal part" of the matrix.
311: Use caution, as the reference count on the returned matrix is not incremented and it is used as
312: part of the containing MPI Mat's normal operation.
314: Level: advanced
316: @*/
317: PetscErrorCode MatGetDiagonalBlock(Mat A,Mat *a)
318: {
325: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
326: if (!A->ops->getdiagonalblock) {
327: PetscMPIInt size;
328: MPI_Comm_size(PetscObjectComm((PetscObject)A),&size);
329: if (size == 1) {
330: *a = A;
331: return(0);
332: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Not coded for this matrix type");
333: }
334: (*A->ops->getdiagonalblock)(A,a);
335: return(0);
336: }
338: /*@
339: MatGetTrace - Gets the trace of a matrix. The sum of the diagonal entries.
341: Collective on Mat
343: Input Parameters:
344: . mat - the matrix
346: Output Parameter:
347: . trace - the sum of the diagonal entries
349: Level: advanced
351: @*/
352: PetscErrorCode MatGetTrace(Mat mat,PetscScalar *trace)
353: {
355: Vec diag;
358: MatCreateVecs(mat,&diag,NULL);
359: MatGetDiagonal(mat,diag);
360: VecSum(diag,trace);
361: VecDestroy(&diag);
362: return(0);
363: }
365: /*@
366: MatRealPart - Zeros out the imaginary part of the matrix
368: Logically Collective on Mat
370: Input Parameters:
371: . mat - the matrix
373: Level: advanced
376: .seealso: MatImaginaryPart()
377: @*/
378: PetscErrorCode MatRealPart(Mat mat)
379: {
385: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
386: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
387: if (!mat->ops->realpart) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
388: MatCheckPreallocated(mat,1);
389: (*mat->ops->realpart)(mat);
390: return(0);
391: }
393: /*@C
394: MatGetGhosts - Get the global index of all ghost nodes defined by the sparse matrix
396: Collective on Mat
398: Input Parameter:
399: . mat - the matrix
401: Output Parameters:
402: + nghosts - number of ghosts (note for BAIJ matrices there is one ghost for each block)
403: - ghosts - the global indices of the ghost points
405: Notes:
406: the nghosts and ghosts are suitable to pass into VecCreateGhost()
408: Level: advanced
410: @*/
411: PetscErrorCode MatGetGhosts(Mat mat,PetscInt *nghosts,const PetscInt *ghosts[])
412: {
418: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
419: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
420: if (!mat->ops->getghosts) {
421: if (nghosts) *nghosts = 0;
422: if (ghosts) *ghosts = 0;
423: } else {
424: (*mat->ops->getghosts)(mat,nghosts,ghosts);
425: }
426: return(0);
427: }
430: /*@
431: MatImaginaryPart - Moves the imaginary part of the matrix to the real part and zeros the imaginary part
433: Logically Collective on Mat
435: Input Parameters:
436: . mat - the matrix
438: Level: advanced
441: .seealso: MatRealPart()
442: @*/
443: PetscErrorCode MatImaginaryPart(Mat mat)
444: {
450: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
451: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
452: if (!mat->ops->imaginarypart) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
453: MatCheckPreallocated(mat,1);
454: (*mat->ops->imaginarypart)(mat);
455: return(0);
456: }
458: /*@
459: MatMissingDiagonal - Determine if sparse matrix is missing a diagonal entry (or block entry for BAIJ matrices)
461: Not Collective
463: Input Parameter:
464: . mat - the matrix
466: Output Parameters:
467: + missing - is any diagonal missing
468: - dd - first diagonal entry that is missing (optional) on this process
470: Level: advanced
473: .seealso: MatRealPart()
474: @*/
475: PetscErrorCode MatMissingDiagonal(Mat mat,PetscBool *missing,PetscInt *dd)
476: {
482: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
483: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
484: if (!mat->ops->missingdiagonal) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
485: (*mat->ops->missingdiagonal)(mat,missing,dd);
486: return(0);
487: }
489: /*@C
490: MatGetRow - Gets a row of a matrix. You MUST call MatRestoreRow()
491: for each row that you get to ensure that your application does
492: not bleed memory.
494: Not Collective
496: Input Parameters:
497: + mat - the matrix
498: - row - the row to get
500: Output Parameters:
501: + ncols - if not NULL, the number of nonzeros in the row
502: . cols - if not NULL, the column numbers
503: - vals - if not NULL, the values
505: Notes:
506: This routine is provided for people who need to have direct access
507: to the structure of a matrix. We hope that we provide enough
508: high-level matrix routines that few users will need it.
510: MatGetRow() always returns 0-based column indices, regardless of
511: whether the internal representation is 0-based (default) or 1-based.
513: For better efficiency, set cols and/or vals to NULL if you do
514: not wish to extract these quantities.
516: The user can only examine the values extracted with MatGetRow();
517: the values cannot be altered. To change the matrix entries, one
518: must use MatSetValues().
520: You can only have one call to MatGetRow() outstanding for a particular
521: matrix at a time, per processor. MatGetRow() can only obtain rows
522: associated with the given processor, it cannot get rows from the
523: other processors; for that we suggest using MatCreateSubMatrices(), then
524: MatGetRow() on the submatrix. The row index passed to MatGetRow()
525: is in the global number of rows.
527: Fortran Notes:
528: The calling sequence from Fortran is
529: .vb
530: MatGetRow(matrix,row,ncols,cols,values,ierr)
531: Mat matrix (input)
532: integer row (input)
533: integer ncols (output)
534: integer cols(maxcols) (output)
535: double precision (or double complex) values(maxcols) output
536: .ve
537: where maxcols >= maximum nonzeros in any row of the matrix.
540: Caution:
541: Do not try to change the contents of the output arrays (cols and vals).
542: In some cases, this may corrupt the matrix.
544: Level: advanced
546: .seealso: MatRestoreRow(), MatSetValues(), MatGetValues(), MatCreateSubMatrices(), MatGetDiagonal()
547: @*/
548: PetscErrorCode MatGetRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[])
549: {
551: PetscInt incols;
556: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
557: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
558: if (!mat->ops->getrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
559: MatCheckPreallocated(mat,1);
560: PetscLogEventBegin(MAT_GetRow,mat,0,0,0);
561: (*mat->ops->getrow)(mat,row,&incols,(PetscInt**)cols,(PetscScalar**)vals);
562: if (ncols) *ncols = incols;
563: PetscLogEventEnd(MAT_GetRow,mat,0,0,0);
564: return(0);
565: }
567: /*@
568: MatConjugate - replaces the matrix values with their complex conjugates
570: Logically Collective on Mat
572: Input Parameters:
573: . mat - the matrix
575: Level: advanced
577: .seealso: VecConjugate()
578: @*/
579: PetscErrorCode MatConjugate(Mat mat)
580: {
581: #if defined(PETSC_USE_COMPLEX)
586: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
587: if (!mat->ops->conjugate) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not provided for this matrix format, send email to petsc-maint@mcs.anl.gov");
588: (*mat->ops->conjugate)(mat);
589: #else
591: #endif
592: return(0);
593: }
595: /*@C
596: MatRestoreRow - Frees any temporary space allocated by MatGetRow().
598: Not Collective
600: Input Parameters:
601: + mat - the matrix
602: . row - the row to get
603: . ncols, cols - the number of nonzeros and their columns
604: - vals - if nonzero the column values
606: Notes:
607: This routine should be called after you have finished examining the entries.
609: This routine zeros out ncols, cols, and vals. This is to prevent accidental
610: us of the array after it has been restored. If you pass NULL, it will
611: not zero the pointers. Use of cols or vals after MatRestoreRow is invalid.
613: Fortran Notes:
614: The calling sequence from Fortran is
615: .vb
616: MatRestoreRow(matrix,row,ncols,cols,values,ierr)
617: Mat matrix (input)
618: integer row (input)
619: integer ncols (output)
620: integer cols(maxcols) (output)
621: double precision (or double complex) values(maxcols) output
622: .ve
623: Where maxcols >= maximum nonzeros in any row of the matrix.
625: In Fortran MatRestoreRow() MUST be called after MatGetRow()
626: before another call to MatGetRow() can be made.
628: Level: advanced
630: .seealso: MatGetRow()
631: @*/
632: PetscErrorCode MatRestoreRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[])
633: {
639: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
640: if (!mat->ops->restorerow) return(0);
641: (*mat->ops->restorerow)(mat,row,ncols,(PetscInt **)cols,(PetscScalar **)vals);
642: if (ncols) *ncols = 0;
643: if (cols) *cols = NULL;
644: if (vals) *vals = NULL;
645: return(0);
646: }
648: /*@
649: MatGetRowUpperTriangular - Sets a flag to enable calls to MatGetRow() for matrix in MATSBAIJ format.
650: You should call MatRestoreRowUpperTriangular() after calling MatGetRow/MatRestoreRow() to disable the flag.
652: Not Collective
654: Input Parameters:
655: . mat - the matrix
657: Notes:
658: 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.
660: Level: advanced
662: .seealso: MatRestoreRowUpperTriangular()
663: @*/
664: PetscErrorCode MatGetRowUpperTriangular(Mat mat)
665: {
671: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
672: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
673: MatCheckPreallocated(mat,1);
674: if (!mat->ops->getrowuppertriangular) return(0);
675: (*mat->ops->getrowuppertriangular)(mat);
676: return(0);
677: }
679: /*@
680: MatRestoreRowUpperTriangular - Disable calls to MatGetRow() for matrix in MATSBAIJ format.
682: Not Collective
684: Input Parameters:
685: . mat - the matrix
687: Notes:
688: This routine should be called after you have finished MatGetRow/MatRestoreRow().
691: Level: advanced
693: .seealso: MatGetRowUpperTriangular()
694: @*/
695: PetscErrorCode MatRestoreRowUpperTriangular(Mat mat)
696: {
702: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
703: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
704: MatCheckPreallocated(mat,1);
705: if (!mat->ops->restorerowuppertriangular) return(0);
706: (*mat->ops->restorerowuppertriangular)(mat);
707: return(0);
708: }
710: /*@C
711: MatSetOptionsPrefix - Sets the prefix used for searching for all
712: Mat options in the database.
714: Logically Collective on Mat
716: Input Parameter:
717: + A - the Mat context
718: - prefix - the prefix to prepend to all option names
720: Notes:
721: A hyphen (-) must NOT be given at the beginning of the prefix name.
722: The first character of all runtime options is AUTOMATICALLY the hyphen.
724: Level: advanced
726: .seealso: MatSetFromOptions()
727: @*/
728: PetscErrorCode MatSetOptionsPrefix(Mat A,const char prefix[])
729: {
734: PetscObjectSetOptionsPrefix((PetscObject)A,prefix);
735: return(0);
736: }
738: /*@C
739: MatAppendOptionsPrefix - Appends to the prefix used for searching for all
740: Mat options in the database.
742: Logically Collective on Mat
744: Input Parameters:
745: + A - the Mat context
746: - prefix - the prefix to prepend to all option names
748: Notes:
749: A hyphen (-) must NOT be given at the beginning of the prefix name.
750: The first character of all runtime options is AUTOMATICALLY the hyphen.
752: Level: advanced
754: .seealso: MatGetOptionsPrefix()
755: @*/
756: PetscErrorCode MatAppendOptionsPrefix(Mat A,const char prefix[])
757: {
762: PetscObjectAppendOptionsPrefix((PetscObject)A,prefix);
763: return(0);
764: }
766: /*@C
767: MatGetOptionsPrefix - Sets the prefix used for searching for all
768: Mat options in the database.
770: Not Collective
772: Input Parameter:
773: . A - the Mat context
775: Output Parameter:
776: . prefix - pointer to the prefix string used
778: Notes:
779: On the fortran side, the user should pass in a string 'prefix' of
780: sufficient length to hold the prefix.
782: Level: advanced
784: .seealso: MatAppendOptionsPrefix()
785: @*/
786: PetscErrorCode MatGetOptionsPrefix(Mat A,const char *prefix[])
787: {
792: PetscObjectGetOptionsPrefix((PetscObject)A,prefix);
793: return(0);
794: }
796: /*@
797: MatResetPreallocation - Reset mat to use the original nonzero pattern provided by users.
799: Collective on Mat
801: Input Parameters:
802: . A - the Mat context
804: Notes:
805: The allocated memory will be shrunk after calling MatAssembly with MAT_FINAL_ASSEMBLY. Users can reset the preallocation to access the original memory.
806: Currently support MPIAIJ and SEQAIJ.
808: Level: beginner
810: .seealso: MatSeqAIJSetPreallocation(), MatMPIAIJSetPreallocation(), MatXAIJSetPreallocation()
811: @*/
812: PetscErrorCode MatResetPreallocation(Mat A)
813: {
819: PetscUseMethod(A,"MatResetPreallocation_C",(Mat),(A));
820: return(0);
821: }
824: /*@
825: MatSetUp - Sets up the internal matrix data structures for the later use.
827: Collective on Mat
829: Input Parameters:
830: . A - the Mat context
832: Notes:
833: If the user has not set preallocation for this matrix then a default preallocation that is likely to be inefficient is used.
835: If a suitable preallocation routine is used, this function does not need to be called.
837: See the Performance chapter of the PETSc users manual for how to preallocate matrices
839: Level: beginner
841: .seealso: MatCreate(), MatDestroy()
842: @*/
843: PetscErrorCode MatSetUp(Mat A)
844: {
845: PetscMPIInt size;
850: if (!((PetscObject)A)->type_name) {
851: MPI_Comm_size(PetscObjectComm((PetscObject)A), &size);
852: if (size == 1) {
853: MatSetType(A, MATSEQAIJ);
854: } else {
855: MatSetType(A, MATMPIAIJ);
856: }
857: }
858: if (!A->preallocated && A->ops->setup) {
859: PetscInfo(A,"Warning not preallocating matrix storage\n");
860: (*A->ops->setup)(A);
861: }
862: PetscLayoutSetUp(A->rmap);
863: PetscLayoutSetUp(A->cmap);
864: A->preallocated = PETSC_TRUE;
865: return(0);
866: }
868: #if defined(PETSC_HAVE_SAWS)
869: #include <petscviewersaws.h>
870: #endif
871: /*@C
872: MatView - Visualizes a matrix object.
874: Collective on Mat
876: Input Parameters:
877: + mat - the matrix
878: - viewer - visualization context
880: Notes:
881: The available visualization contexts include
882: + PETSC_VIEWER_STDOUT_SELF - for sequential matrices
883: . PETSC_VIEWER_STDOUT_WORLD - for parallel matrices created on PETSC_COMM_WORLD
884: . PETSC_VIEWER_STDOUT_(comm) - for matrices created on MPI communicator comm
885: - PETSC_VIEWER_DRAW_WORLD - graphical display of nonzero structure
887: The user can open alternative visualization contexts with
888: + PetscViewerASCIIOpen() - Outputs matrix to a specified file
889: . PetscViewerBinaryOpen() - Outputs matrix in binary to a
890: specified file; corresponding input uses MatLoad()
891: . PetscViewerDrawOpen() - Outputs nonzero matrix structure to
892: an X window display
893: - PetscViewerSocketOpen() - Outputs matrix to Socket viewer.
894: Currently only the sequential dense and AIJ
895: matrix types support the Socket viewer.
897: The user can call PetscViewerPushFormat() to specify the output
898: format of ASCII printed objects (when using PETSC_VIEWER_STDOUT_SELF,
899: PETSC_VIEWER_STDOUT_WORLD and PetscViewerASCIIOpen). Available formats include
900: + PETSC_VIEWER_DEFAULT - default, prints matrix contents
901: . PETSC_VIEWER_ASCII_MATLAB - prints matrix contents in Matlab format
902: . PETSC_VIEWER_ASCII_DENSE - prints entire matrix including zeros
903: . PETSC_VIEWER_ASCII_COMMON - prints matrix contents, using a sparse
904: format common among all matrix types
905: . PETSC_VIEWER_ASCII_IMPL - prints matrix contents, using an implementation-specific
906: format (which is in many cases the same as the default)
907: . PETSC_VIEWER_ASCII_INFO - prints basic information about the matrix
908: size and structure (not the matrix entries)
909: - PETSC_VIEWER_ASCII_INFO_DETAIL - prints more detailed information about
910: the matrix structure
912: Options Database Keys:
913: + -mat_view ::ascii_info - Prints info on matrix at conclusion of MatAssemblyEnd()
914: . -mat_view ::ascii_info_detail - Prints more detailed info
915: . -mat_view - Prints matrix in ASCII format
916: . -mat_view ::ascii_matlab - Prints matrix in Matlab format
917: . -mat_view draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX().
918: . -display <name> - Sets display name (default is host)
919: . -draw_pause <sec> - Sets number of seconds to pause after display
920: . -mat_view socket - Sends matrix to socket, can be accessed from Matlab (see Users-Manual: Chapter 12 Using MATLAB with PETSc for details)
921: . -viewer_socket_machine <machine> -
922: . -viewer_socket_port <port> -
923: . -mat_view binary - save matrix to file in binary format
924: - -viewer_binary_filename <name> -
925: Level: beginner
927: Notes:
928: The ASCII viewers are only recommended for small matrices on at most a moderate number of processes,
929: the program will seemingly hang and take hours for larger matrices, for larger matrices one should use the binary format.
931: See the manual page for MatLoad() for the exact format of the binary file when the binary
932: viewer is used.
934: See share/petsc/matlab/PetscBinaryRead.m for a Matlab code that can read in the binary file when the binary
935: viewer is used.
937: One can use '-mat_view draw -draw_pause -1' to pause the graphical display of matrix nonzero structure,
938: and then use the following mouse functions.
939: + left mouse: zoom in
940: . middle mouse: zoom out
941: - right mouse: continue with the simulation
943: .seealso: PetscViewerPushFormat(), PetscViewerASCIIOpen(), PetscViewerDrawOpen(),
944: PetscViewerSocketOpen(), PetscViewerBinaryOpen(), MatLoad()
945: @*/
946: PetscErrorCode MatView(Mat mat,PetscViewer viewer)
947: {
948: PetscErrorCode ierr;
949: PetscInt rows,cols,rbs,cbs;
950: PetscBool iascii,ibinary,isstring;
951: PetscViewerFormat format;
952: PetscMPIInt size;
953: #if defined(PETSC_HAVE_SAWS)
954: PetscBool issaws;
955: #endif
960: if (!viewer) {
961: PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)mat),&viewer);
962: }
965: MatCheckPreallocated(mat,1);
966: PetscViewerGetFormat(viewer,&format);
967: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
968: if (size == 1 && format == PETSC_VIEWER_LOAD_BALANCE) return(0);
969: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&ibinary);
970: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSTRING,&isstring);
971: if (ibinary) {
972: PetscBool mpiio;
973: PetscViewerBinaryGetUseMPIIO(viewer,&mpiio);
974: if (mpiio) SETERRQ(PetscObjectComm((PetscObject)viewer),PETSC_ERR_SUP,"PETSc matrix viewers do not support using MPI-IO, turn off that flag");
975: }
977: PetscLogEventBegin(MAT_View,mat,viewer,0,0);
978: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);
979: if ((!iascii || (format != PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL)) && mat->factortype) {
980: SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"No viewers for factored matrix except ASCII info or info_detailed");
981: }
983: #if defined(PETSC_HAVE_SAWS)
984: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSAWS,&issaws);
985: #endif
986: if (iascii) {
987: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ORDER,"Must call MatAssemblyBegin/End() before viewing matrix");
988: PetscObjectPrintClassNamePrefixType((PetscObject)mat,viewer);
989: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
990: MatNullSpace nullsp,transnullsp;
992: PetscViewerASCIIPushTab(viewer);
993: MatGetSize(mat,&rows,&cols);
994: MatGetBlockSizes(mat,&rbs,&cbs);
995: if (rbs != 1 || cbs != 1) {
996: if (rbs != cbs) {PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D, rbs=%D, cbs = %D\n",rows,cols,rbs,cbs);}
997: else {PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D, bs=%D\n",rows,cols,rbs);}
998: } else {
999: PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D\n",rows,cols);
1000: }
1001: if (mat->factortype) {
1002: MatSolverType solver;
1003: MatFactorGetSolverType(mat,&solver);
1004: PetscViewerASCIIPrintf(viewer,"package used to perform factorization: %s\n",solver);
1005: }
1006: if (mat->ops->getinfo) {
1007: MatInfo info;
1008: MatGetInfo(mat,MAT_GLOBAL_SUM,&info);
1009: PetscViewerASCIIPrintf(viewer,"total: nonzeros=%.f, allocated nonzeros=%.f\n",info.nz_used,info.nz_allocated);
1010: PetscViewerASCIIPrintf(viewer,"total number of mallocs used during MatSetValues calls =%D\n",(PetscInt)info.mallocs);
1011: }
1012: MatGetNullSpace(mat,&nullsp);
1013: MatGetTransposeNullSpace(mat,&transnullsp);
1014: if (nullsp) {PetscViewerASCIIPrintf(viewer," has attached null space\n");}
1015: if (transnullsp && transnullsp != nullsp) {PetscViewerASCIIPrintf(viewer," has attached transposed null space\n");}
1016: MatGetNearNullSpace(mat,&nullsp);
1017: if (nullsp) {PetscViewerASCIIPrintf(viewer," has attached near null space\n");}
1018: }
1019: #if defined(PETSC_HAVE_SAWS)
1020: } else if (issaws) {
1021: PetscMPIInt rank;
1023: PetscObjectName((PetscObject)mat);
1024: MPI_Comm_rank(PETSC_COMM_WORLD,&rank);
1025: if (!((PetscObject)mat)->amsmem && !rank) {
1026: PetscObjectViewSAWs((PetscObject)mat,viewer);
1027: }
1028: #endif
1029: } else if (isstring) {
1030: const char *type;
1031: MatGetType(mat,&type);
1032: PetscViewerStringSPrintf(viewer," MatType: %-7.7s",type);
1033: if (mat->ops->view) {(*mat->ops->view)(mat,viewer);}
1034: }
1035: if ((format == PETSC_VIEWER_NATIVE || format == PETSC_VIEWER_LOAD_BALANCE) && mat->ops->viewnative) {
1036: PetscViewerASCIIPushTab(viewer);
1037: (*mat->ops->viewnative)(mat,viewer);
1038: PetscViewerASCIIPopTab(viewer);
1039: } else if (mat->ops->view) {
1040: PetscViewerASCIIPushTab(viewer);
1041: (*mat->ops->view)(mat,viewer);
1042: PetscViewerASCIIPopTab(viewer);
1043: }
1044: if (iascii) {
1045: PetscViewerGetFormat(viewer,&format);
1046: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
1047: PetscViewerASCIIPopTab(viewer);
1048: }
1049: }
1050: PetscLogEventEnd(MAT_View,mat,viewer,0,0);
1051: return(0);
1052: }
1054: #if defined(PETSC_USE_DEBUG)
1055: #include <../src/sys/totalview/tv_data_display.h>
1056: PETSC_UNUSED static int TV_display_type(const struct _p_Mat *mat)
1057: {
1058: TV_add_row("Local rows", "int", &mat->rmap->n);
1059: TV_add_row("Local columns", "int", &mat->cmap->n);
1060: TV_add_row("Global rows", "int", &mat->rmap->N);
1061: TV_add_row("Global columns", "int", &mat->cmap->N);
1062: TV_add_row("Typename", TV_ascii_string_type, ((PetscObject)mat)->type_name);
1063: return TV_format_OK;
1064: }
1065: #endif
1067: /*@C
1068: MatLoad - Loads a matrix that has been stored in binary/HDF5 format
1069: with MatView(). The matrix format is determined from the options database.
1070: Generates a parallel MPI matrix if the communicator has more than one
1071: processor. The default matrix type is AIJ.
1073: Collective on PetscViewer
1075: Input Parameters:
1076: + newmat - the newly loaded matrix, this needs to have been created with MatCreate()
1077: or some related function before a call to MatLoad()
1078: - viewer - binary/HDF5 file viewer
1080: Options Database Keys:
1081: Used with block matrix formats (MATSEQBAIJ, ...) to specify
1082: block size
1083: . -matload_block_size <bs>
1085: Level: beginner
1087: Notes:
1088: If the Mat type has not yet been given then MATAIJ is used, call MatSetFromOptions() on the
1089: Mat before calling this routine if you wish to set it from the options database.
1091: MatLoad() automatically loads into the options database any options
1092: given in the file filename.info where filename is the name of the file
1093: that was passed to the PetscViewerBinaryOpen(). The options in the info
1094: file will be ignored if you use the -viewer_binary_skip_info option.
1096: If the type or size of newmat is not set before a call to MatLoad, PETSc
1097: sets the default matrix type AIJ and sets the local and global sizes.
1098: If type and/or size is already set, then the same are used.
1100: In parallel, each processor can load a subset of rows (or the
1101: entire matrix). This routine is especially useful when a large
1102: matrix is stored on disk and only part of it is desired on each
1103: processor. For example, a parallel solver may access only some of
1104: the rows from each processor. The algorithm used here reads
1105: relatively small blocks of data rather than reading the entire
1106: matrix and then subsetting it.
1108: Viewer's PetscViewerType must be either PETSCVIEWERBINARY or PETSCVIEWERHDF5.
1109: Such viewer can be created using PetscViewerBinaryOpen()/PetscViewerHDF5Open(),
1110: or the sequence like
1111: $ PetscViewer v;
1112: $ PetscViewerCreate(PETSC_COMM_WORLD,&v);
1113: $ PetscViewerSetType(v,PETSCVIEWERBINARY);
1114: $ PetscViewerSetFromOptions(v);
1115: $ PetscViewerFileSetMode(v,FILE_MODE_READ);
1116: $ PetscViewerFileSetName(v,"datafile");
1117: The optional PetscViewerSetFromOptions() call allows to override PetscViewerSetType() using option
1118: $ -viewer_type {binary,hdf5}
1120: See the example src/ksp/ksp/examples/tutorials/ex27.c with the first approach,
1121: and src/mat/examples/tutorials/ex10.c with the second approach.
1123: Notes about the PETSc binary format:
1124: In case of PETSCVIEWERBINARY, a native PETSc binary format is used. Each of the blocks
1125: is read onto rank 0 and then shipped to its destination rank, one after another.
1126: Multiple objects, both matrices and vectors, can be stored within the same file.
1127: Their PetscObject name is ignored; they are loaded in the order of their storage.
1129: Most users should not need to know the details of the binary storage
1130: format, since MatLoad() and MatView() completely hide these details.
1131: But for anyone who's interested, the standard binary matrix storage
1132: format is
1134: $ int MAT_FILE_CLASSID
1135: $ int number of rows
1136: $ int number of columns
1137: $ int total number of nonzeros
1138: $ int *number nonzeros in each row
1139: $ int *column indices of all nonzeros (starting index is zero)
1140: $ PetscScalar *values of all nonzeros
1142: PETSc automatically does the byte swapping for
1143: machines that store the bytes reversed, e.g. DEC alpha, freebsd,
1144: linux, Windows and the paragon; thus if you write your own binary
1145: read/write routines you have to swap the bytes; see PetscBinaryRead()
1146: and PetscBinaryWrite() to see how this may be done.
1148: Notes about the HDF5 (MATLAB MAT-File Version 7.3) format:
1149: In case of PETSCVIEWERHDF5, a parallel HDF5 reader is used.
1150: Each processor's chunk is loaded independently by its owning rank.
1151: Multiple objects, both matrices and vectors, can be stored within the same file.
1152: They are looked up by their PetscObject name.
1154: As the MATLAB MAT-File Version 7.3 format is also a HDF5 flavor, we decided to use
1155: by default the same structure and naming of the AIJ arrays and column count
1156: within the HDF5 file. This means that a MAT file saved with -v7.3 flag, e.g.
1157: $ save example.mat A b -v7.3
1158: can be directly read by this routine (see Reference 1 for details).
1159: Note that depending on your MATLAB version, this format might be a default,
1160: otherwise you can set it as default in Preferences.
1162: Unless -nocompression flag is used to save the file in MATLAB,
1163: PETSc must be configured with ZLIB package.
1165: See also examples src/mat/examples/tutorials/ex10.c and src/ksp/ksp/examples/tutorials/ex27.c
1167: Current HDF5 (MAT-File) limitations:
1168: This reader currently supports only real MATSEQAIJ, MATMPIAIJ, MATSEQDENSE and MATMPIDENSE matrices.
1170: Corresponding MatView() is not yet implemented.
1172: The loaded matrix is actually a transpose of the original one in MATLAB,
1173: unless you push PETSC_VIEWER_HDF5_MAT format (see examples above).
1174: With this format, matrix is automatically transposed by PETSc,
1175: unless the matrix is marked as SPD or symmetric
1176: (see MatSetOption(), MAT_SPD, MAT_SYMMETRIC).
1178: References:
1179: 1. MATLAB(R) Documentation, manual page of save(), https://www.mathworks.com/help/matlab/ref/save.html#btox10b-1-version
1181: .seealso: PetscViewerBinaryOpen(), PetscViewerSetType(), MatView(), VecLoad()
1183: @*/
1184: PetscErrorCode MatLoad(Mat newmat,PetscViewer viewer)
1185: {
1187: PetscBool flg;
1193: if (!((PetscObject)newmat)->type_name) {
1194: MatSetType(newmat,MATAIJ);
1195: }
1197: flg = PETSC_FALSE;
1198: PetscOptionsGetBool(((PetscObject)newmat)->options,((PetscObject)newmat)->prefix,"-matload_symmetric",&flg,NULL);
1199: if (flg) {
1200: MatSetOption(newmat,MAT_SYMMETRIC,PETSC_TRUE);
1201: MatSetOption(newmat,MAT_SYMMETRY_ETERNAL,PETSC_TRUE);
1202: }
1203: flg = PETSC_FALSE;
1204: PetscOptionsGetBool(((PetscObject)newmat)->options,((PetscObject)newmat)->prefix,"-matload_spd",&flg,NULL);
1205: if (flg) {
1206: MatSetOption(newmat,MAT_SPD,PETSC_TRUE);
1207: }
1209: if (!newmat->ops->load) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatLoad is not supported for type");
1210: PetscLogEventBegin(MAT_Load,viewer,0,0,0);
1211: (*newmat->ops->load)(newmat,viewer);
1212: PetscLogEventEnd(MAT_Load,viewer,0,0,0);
1213: return(0);
1214: }
1216: PetscErrorCode MatDestroy_Redundant(Mat_Redundant **redundant)
1217: {
1219: Mat_Redundant *redund = *redundant;
1220: PetscInt i;
1223: if (redund){
1224: if (redund->matseq) { /* via MatCreateSubMatrices() */
1225: ISDestroy(&redund->isrow);
1226: ISDestroy(&redund->iscol);
1227: MatDestroySubMatrices(1,&redund->matseq);
1228: } else {
1229: PetscFree2(redund->send_rank,redund->recv_rank);
1230: PetscFree(redund->sbuf_j);
1231: PetscFree(redund->sbuf_a);
1232: for (i=0; i<redund->nrecvs; i++) {
1233: PetscFree(redund->rbuf_j[i]);
1234: PetscFree(redund->rbuf_a[i]);
1235: }
1236: PetscFree4(redund->sbuf_nz,redund->rbuf_nz,redund->rbuf_j,redund->rbuf_a);
1237: }
1239: if (redund->subcomm) {
1240: PetscCommDestroy(&redund->subcomm);
1241: }
1242: PetscFree(redund);
1243: }
1244: return(0);
1245: }
1247: /*@
1248: MatDestroy - Frees space taken by a matrix.
1250: Collective on Mat
1252: Input Parameter:
1253: . A - the matrix
1255: Level: beginner
1257: @*/
1258: PetscErrorCode MatDestroy(Mat *A)
1259: {
1263: if (!*A) return(0);
1265: if (--((PetscObject)(*A))->refct > 0) {*A = NULL; return(0);}
1267: /* if memory was published with SAWs then destroy it */
1268: PetscObjectSAWsViewOff((PetscObject)*A);
1269: if ((*A)->ops->destroy) {
1270: (*(*A)->ops->destroy)(*A);
1271: }
1273: PetscFree((*A)->defaultvectype);
1274: PetscFree((*A)->bsizes);
1275: PetscFree((*A)->solvertype);
1276: MatDestroy_Redundant(&(*A)->redundant);
1277: MatNullSpaceDestroy(&(*A)->nullsp);
1278: MatNullSpaceDestroy(&(*A)->transnullsp);
1279: MatNullSpaceDestroy(&(*A)->nearnullsp);
1280: MatDestroy(&(*A)->schur);
1281: PetscLayoutDestroy(&(*A)->rmap);
1282: PetscLayoutDestroy(&(*A)->cmap);
1283: PetscHeaderDestroy(A);
1284: return(0);
1285: }
1287: /*@C
1288: MatSetValues - Inserts or adds a block of values into a matrix.
1289: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
1290: MUST be called after all calls to MatSetValues() have been completed.
1292: Not Collective
1294: Input Parameters:
1295: + mat - the matrix
1296: . v - a logically two-dimensional array of values
1297: . m, idxm - the number of rows and their global indices
1298: . n, idxn - the number of columns and their global indices
1299: - addv - either ADD_VALUES or INSERT_VALUES, where
1300: ADD_VALUES adds values to any existing entries, and
1301: INSERT_VALUES replaces existing entries with new values
1303: Notes:
1304: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
1305: MatSetUp() before using this routine
1307: By default the values, v, are row-oriented. See MatSetOption() for other options.
1309: Calls to MatSetValues() with the INSERT_VALUES and ADD_VALUES
1310: options cannot be mixed without intervening calls to the assembly
1311: routines.
1313: MatSetValues() uses 0-based row and column numbers in Fortran
1314: as well as in C.
1316: Negative indices may be passed in idxm and idxn, these rows and columns are
1317: simply ignored. This allows easily inserting element stiffness matrices
1318: with homogeneous Dirchlet boundary conditions that you don't want represented
1319: in the matrix.
1321: Efficiency Alert:
1322: The routine MatSetValuesBlocked() may offer much better efficiency
1323: for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ).
1325: Level: beginner
1327: Developer Notes:
1328: This is labeled with C so does not automatically generate Fortran stubs and interfaces
1329: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
1331: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1332: InsertMode, INSERT_VALUES, ADD_VALUES
1333: @*/
1334: PetscErrorCode MatSetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv)
1335: {
1337: #if defined(PETSC_USE_DEBUG)
1338: PetscInt i,j;
1339: #endif
1344: if (!m || !n) return(0); /* no values to insert */
1347: MatCheckPreallocated(mat,1);
1349: if (mat->insertmode == NOT_SET_VALUES) {
1350: mat->insertmode = addv;
1351: }
1352: #if defined(PETSC_USE_DEBUG)
1353: else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
1354: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1355: if (!mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1357: for (i=0; i<m; i++) {
1358: for (j=0; j<n; j++) {
1359: if (mat->erroriffailure && PetscIsInfOrNanScalar(v[i*n+j]))
1360: #if defined(PETSC_USE_COMPLEX)
1361: 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]);
1362: #else
1363: SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_FP,"Inserting %g at matrix entry (%D,%D)",(double)v[i*n+j],idxm[i],idxn[j]);
1364: #endif
1365: }
1366: }
1367: #endif
1369: if (mat->assembled) {
1370: mat->was_assembled = PETSC_TRUE;
1371: mat->assembled = PETSC_FALSE;
1372: }
1373: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1374: (*mat->ops->setvalues)(mat,m,idxm,n,idxn,v,addv);
1375: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1376: return(0);
1377: }
1380: /*@
1381: MatSetValuesRowLocal - Inserts a row (block row for BAIJ matrices) of nonzero
1382: values into a matrix
1384: Not Collective
1386: Input Parameters:
1387: + mat - the matrix
1388: . row - the (block) row to set
1389: - v - a logically two-dimensional array of values
1391: Notes:
1392: By the values, v, are column-oriented (for the block version) and sorted
1394: All the nonzeros in the row must be provided
1396: The matrix must have previously had its column indices set
1398: The row must belong to this process
1400: Level: intermediate
1402: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1403: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues(), MatSetValuesRow(), MatSetLocalToGlobalMapping()
1404: @*/
1405: PetscErrorCode MatSetValuesRowLocal(Mat mat,PetscInt row,const PetscScalar v[])
1406: {
1408: PetscInt globalrow;
1414: ISLocalToGlobalMappingApply(mat->rmap->mapping,1,&row,&globalrow);
1415: MatSetValuesRow(mat,globalrow,v);
1416: return(0);
1417: }
1419: /*@
1420: MatSetValuesRow - Inserts a row (block row for BAIJ matrices) of nonzero
1421: values into a matrix
1423: Not Collective
1425: Input Parameters:
1426: + mat - the matrix
1427: . row - the (block) row to set
1428: - 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
1430: Notes:
1431: The values, v, are column-oriented for the block version.
1433: All the nonzeros in the row must be provided
1435: THE MATRIX MUST HAVE PREVIOUSLY HAD ITS COLUMN INDICES SET. IT IS RARE THAT THIS ROUTINE IS USED, usually MatSetValues() is used.
1437: The row must belong to this process
1439: Level: advanced
1441: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1442: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues()
1443: @*/
1444: PetscErrorCode MatSetValuesRow(Mat mat,PetscInt row,const PetscScalar v[])
1445: {
1451: MatCheckPreallocated(mat,1);
1453: #if defined(PETSC_USE_DEBUG)
1454: if (mat->insertmode == ADD_VALUES) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add and insert values");
1455: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1456: #endif
1457: mat->insertmode = INSERT_VALUES;
1459: if (mat->assembled) {
1460: mat->was_assembled = PETSC_TRUE;
1461: mat->assembled = PETSC_FALSE;
1462: }
1463: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1464: if (!mat->ops->setvaluesrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1465: (*mat->ops->setvaluesrow)(mat,row,v);
1466: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1467: return(0);
1468: }
1470: /*@
1471: MatSetValuesStencil - Inserts or adds a block of values into a matrix.
1472: Using structured grid indexing
1474: Not Collective
1476: Input Parameters:
1477: + mat - the matrix
1478: . m - number of rows being entered
1479: . idxm - grid coordinates (and component number when dof > 1) for matrix rows being entered
1480: . n - number of columns being entered
1481: . idxn - grid coordinates (and component number when dof > 1) for matrix columns being entered
1482: . v - a logically two-dimensional array of values
1483: - addv - either ADD_VALUES or INSERT_VALUES, where
1484: ADD_VALUES adds values to any existing entries, and
1485: INSERT_VALUES replaces existing entries with new values
1487: Notes:
1488: By default the values, v, are row-oriented. See MatSetOption() for other options.
1490: Calls to MatSetValuesStencil() with the INSERT_VALUES and ADD_VALUES
1491: options cannot be mixed without intervening calls to the assembly
1492: routines.
1494: The grid coordinates are across the entire grid, not just the local portion
1496: MatSetValuesStencil() uses 0-based row and column numbers in Fortran
1497: as well as in C.
1499: For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine
1501: In order to use this routine you must either obtain the matrix with DMCreateMatrix()
1502: or call MatSetLocalToGlobalMapping() and MatSetStencil() first.
1504: The columns and rows in the stencil passed in MUST be contained within the
1505: ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example,
1506: if you create a DMDA with an overlap of one grid level and on a particular process its first
1507: local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the
1508: first i index you can use in your column and row indices in MatSetStencil() is 5.
1510: In Fortran idxm and idxn should be declared as
1511: $ MatStencil idxm(4,m),idxn(4,n)
1512: and the values inserted using
1513: $ idxm(MatStencil_i,1) = i
1514: $ idxm(MatStencil_j,1) = j
1515: $ idxm(MatStencil_k,1) = k
1516: $ idxm(MatStencil_c,1) = c
1517: etc
1519: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
1520: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
1521: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
1522: DM_BOUNDARY_PERIODIC boundary type.
1524: 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
1525: a single value per point) you can skip filling those indices.
1527: Inspired by the structured grid interface to the HYPRE package
1528: (https://computation.llnl.gov/projects/hypre-scalable-linear-solvers-multigrid-methods)
1530: Efficiency Alert:
1531: The routine MatSetValuesBlockedStencil() may offer much better efficiency
1532: for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ).
1534: Level: beginner
1536: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1537: MatSetValues(), MatSetValuesBlockedStencil(), MatSetStencil(), DMCreateMatrix(), DMDAVecGetArray(), MatStencil
1538: @*/
1539: PetscErrorCode MatSetValuesStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv)
1540: {
1542: PetscInt buf[8192],*bufm=0,*bufn=0,*jdxm,*jdxn;
1543: PetscInt j,i,dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp;
1544: PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc);
1547: if (!m || !n) return(0); /* no values to insert */
1553: if ((m+n) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1554: jdxm = buf; jdxn = buf+m;
1555: } else {
1556: PetscMalloc2(m,&bufm,n,&bufn);
1557: jdxm = bufm; jdxn = bufn;
1558: }
1559: for (i=0; i<m; i++) {
1560: for (j=0; j<3-sdim; j++) dxm++;
1561: tmp = *dxm++ - starts[0];
1562: for (j=0; j<dim-1; j++) {
1563: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1564: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
1565: }
1566: if (mat->stencil.noc) dxm++;
1567: jdxm[i] = tmp;
1568: }
1569: for (i=0; i<n; i++) {
1570: for (j=0; j<3-sdim; j++) dxn++;
1571: tmp = *dxn++ - starts[0];
1572: for (j=0; j<dim-1; j++) {
1573: if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1574: else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1];
1575: }
1576: if (mat->stencil.noc) dxn++;
1577: jdxn[i] = tmp;
1578: }
1579: MatSetValuesLocal(mat,m,jdxm,n,jdxn,v,addv);
1580: PetscFree2(bufm,bufn);
1581: return(0);
1582: }
1584: /*@
1585: MatSetValuesBlockedStencil - Inserts or adds a block of values into a matrix.
1586: Using structured grid indexing
1588: Not Collective
1590: Input Parameters:
1591: + mat - the matrix
1592: . m - number of rows being entered
1593: . idxm - grid coordinates for matrix rows being entered
1594: . n - number of columns being entered
1595: . idxn - grid coordinates for matrix columns being entered
1596: . v - a logically two-dimensional array of values
1597: - addv - either ADD_VALUES or INSERT_VALUES, where
1598: ADD_VALUES adds values to any existing entries, and
1599: INSERT_VALUES replaces existing entries with new values
1601: Notes:
1602: By default the values, v, are row-oriented and unsorted.
1603: See MatSetOption() for other options.
1605: Calls to MatSetValuesBlockedStencil() with the INSERT_VALUES and ADD_VALUES
1606: options cannot be mixed without intervening calls to the assembly
1607: routines.
1609: The grid coordinates are across the entire grid, not just the local portion
1611: MatSetValuesBlockedStencil() uses 0-based row and column numbers in Fortran
1612: as well as in C.
1614: For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine
1616: In order to use this routine you must either obtain the matrix with DMCreateMatrix()
1617: or call MatSetBlockSize(), MatSetLocalToGlobalMapping() and MatSetStencil() first.
1619: The columns and rows in the stencil passed in MUST be contained within the
1620: ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example,
1621: if you create a DMDA with an overlap of one grid level and on a particular process its first
1622: local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the
1623: first i index you can use in your column and row indices in MatSetStencil() is 5.
1625: In Fortran idxm and idxn should be declared as
1626: $ MatStencil idxm(4,m),idxn(4,n)
1627: and the values inserted using
1628: $ idxm(MatStencil_i,1) = i
1629: $ idxm(MatStencil_j,1) = j
1630: $ idxm(MatStencil_k,1) = k
1631: etc
1633: Negative indices may be passed in idxm and idxn, these rows and columns are
1634: simply ignored. This allows easily inserting element stiffness matrices
1635: with homogeneous Dirchlet boundary conditions that you don't want represented
1636: in the matrix.
1638: Inspired by the structured grid interface to the HYPRE package
1639: (https://computation.llnl.gov/projects/hypre-scalable-linear-solvers-multigrid-methods)
1641: Level: beginner
1643: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1644: MatSetValues(), MatSetValuesStencil(), MatSetStencil(), DMCreateMatrix(), DMDAVecGetArray(), MatStencil,
1645: MatSetBlockSize(), MatSetLocalToGlobalMapping()
1646: @*/
1647: PetscErrorCode MatSetValuesBlockedStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv)
1648: {
1650: PetscInt buf[8192],*bufm=0,*bufn=0,*jdxm,*jdxn;
1651: PetscInt j,i,dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp;
1652: PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc);
1655: if (!m || !n) return(0); /* no values to insert */
1662: if ((m+n) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1663: jdxm = buf; jdxn = buf+m;
1664: } else {
1665: PetscMalloc2(m,&bufm,n,&bufn);
1666: jdxm = bufm; jdxn = bufn;
1667: }
1668: for (i=0; i<m; i++) {
1669: for (j=0; j<3-sdim; j++) dxm++;
1670: tmp = *dxm++ - starts[0];
1671: for (j=0; j<sdim-1; j++) {
1672: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1673: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
1674: }
1675: dxm++;
1676: jdxm[i] = tmp;
1677: }
1678: for (i=0; i<n; i++) {
1679: for (j=0; j<3-sdim; j++) dxn++;
1680: tmp = *dxn++ - starts[0];
1681: for (j=0; j<sdim-1; j++) {
1682: if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = -1;
1683: else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1];
1684: }
1685: dxn++;
1686: jdxn[i] = tmp;
1687: }
1688: MatSetValuesBlockedLocal(mat,m,jdxm,n,jdxn,v,addv);
1689: PetscFree2(bufm,bufn);
1690: return(0);
1691: }
1693: /*@
1694: MatSetStencil - Sets the grid information for setting values into a matrix via
1695: MatSetValuesStencil()
1697: Not Collective
1699: Input Parameters:
1700: + mat - the matrix
1701: . dim - dimension of the grid 1, 2, or 3
1702: . dims - number of grid points in x, y, and z direction, including ghost points on your processor
1703: . starts - starting point of ghost nodes on your processor in x, y, and z direction
1704: - dof - number of degrees of freedom per node
1707: Inspired by the structured grid interface to the HYPRE package
1708: (www.llnl.gov/CASC/hyper)
1710: For matrices generated with DMCreateMatrix() this routine is automatically called and so not needed by the
1711: user.
1713: Level: beginner
1715: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal()
1716: MatSetValues(), MatSetValuesBlockedStencil(), MatSetValuesStencil()
1717: @*/
1718: PetscErrorCode MatSetStencil(Mat mat,PetscInt dim,const PetscInt dims[],const PetscInt starts[],PetscInt dof)
1719: {
1720: PetscInt i;
1727: mat->stencil.dim = dim + (dof > 1);
1728: for (i=0; i<dim; i++) {
1729: mat->stencil.dims[i] = dims[dim-i-1]; /* copy the values in backwards */
1730: mat->stencil.starts[i] = starts[dim-i-1];
1731: }
1732: mat->stencil.dims[dim] = dof;
1733: mat->stencil.starts[dim] = 0;
1734: mat->stencil.noc = (PetscBool)(dof == 1);
1735: return(0);
1736: }
1738: /*@C
1739: MatSetValuesBlocked - Inserts or adds a block of values into a matrix.
1741: Not Collective
1743: Input Parameters:
1744: + mat - the matrix
1745: . v - a logically two-dimensional array of values
1746: . m, idxm - the number of block rows and their global block indices
1747: . n, idxn - the number of block columns and their global block indices
1748: - addv - either ADD_VALUES or INSERT_VALUES, where
1749: ADD_VALUES adds values to any existing entries, and
1750: INSERT_VALUES replaces existing entries with new values
1752: Notes:
1753: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call
1754: MatXXXXSetPreallocation() or MatSetUp() before using this routine.
1756: The m and n count the NUMBER of blocks in the row direction and column direction,
1757: NOT the total number of rows/columns; for example, if the block size is 2 and
1758: you are passing in values for rows 2,3,4,5 then m would be 2 (not 4).
1759: The values in idxm would be 1 2; that is the first index for each block divided by
1760: the block size.
1762: Note that you must call MatSetBlockSize() when constructing this matrix (before
1763: preallocating it).
1765: By default the values, v, are row-oriented, so the layout of
1766: v is the same as for MatSetValues(). See MatSetOption() for other options.
1768: Calls to MatSetValuesBlocked() with the INSERT_VALUES and ADD_VALUES
1769: options cannot be mixed without intervening calls to the assembly
1770: routines.
1772: MatSetValuesBlocked() uses 0-based row and column numbers in Fortran
1773: as well as in C.
1775: Negative indices may be passed in idxm and idxn, these rows and columns are
1776: simply ignored. This allows easily inserting element stiffness matrices
1777: with homogeneous Dirchlet boundary conditions that you don't want represented
1778: in the matrix.
1780: Each time an entry is set within a sparse matrix via MatSetValues(),
1781: internal searching must be done to determine where to place the
1782: data in the matrix storage space. By instead inserting blocks of
1783: entries via MatSetValuesBlocked(), the overhead of matrix assembly is
1784: reduced.
1786: Example:
1787: $ Suppose m=n=2 and block size(bs) = 2 The array is
1788: $
1789: $ 1 2 | 3 4
1790: $ 5 6 | 7 8
1791: $ - - - | - - -
1792: $ 9 10 | 11 12
1793: $ 13 14 | 15 16
1794: $
1795: $ v[] should be passed in like
1796: $ v[] = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16]
1797: $
1798: $ If you are not using row oriented storage of v (that is you called MatSetOption(mat,MAT_ROW_ORIENTED,PETSC_FALSE)) then
1799: $ v[] = [1,5,9,13,2,6,10,14,3,7,11,15,4,8,12,16]
1801: Level: intermediate
1803: .seealso: MatSetBlockSize(), MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal()
1804: @*/
1805: PetscErrorCode MatSetValuesBlocked(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv)
1806: {
1812: if (!m || !n) return(0); /* no values to insert */
1816: MatCheckPreallocated(mat,1);
1817: if (mat->insertmode == NOT_SET_VALUES) {
1818: mat->insertmode = addv;
1819: }
1820: #if defined(PETSC_USE_DEBUG)
1821: else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
1822: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1823: if (!mat->ops->setvaluesblocked && !mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1824: #endif
1826: if (mat->assembled) {
1827: mat->was_assembled = PETSC_TRUE;
1828: mat->assembled = PETSC_FALSE;
1829: }
1830: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
1831: if (mat->ops->setvaluesblocked) {
1832: (*mat->ops->setvaluesblocked)(mat,m,idxm,n,idxn,v,addv);
1833: } else {
1834: PetscInt buf[8192],*bufr=0,*bufc=0,*iidxm,*iidxn;
1835: PetscInt i,j,bs,cbs;
1836: MatGetBlockSizes(mat,&bs,&cbs);
1837: if (m*bs+n*cbs <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
1838: iidxm = buf; iidxn = buf + m*bs;
1839: } else {
1840: PetscMalloc2(m*bs,&bufr,n*cbs,&bufc);
1841: iidxm = bufr; iidxn = bufc;
1842: }
1843: for (i=0; i<m; i++) {
1844: for (j=0; j<bs; j++) {
1845: iidxm[i*bs+j] = bs*idxm[i] + j;
1846: }
1847: }
1848: for (i=0; i<n; i++) {
1849: for (j=0; j<cbs; j++) {
1850: iidxn[i*cbs+j] = cbs*idxn[i] + j;
1851: }
1852: }
1853: MatSetValues(mat,m*bs,iidxm,n*cbs,iidxn,v,addv);
1854: PetscFree2(bufr,bufc);
1855: }
1856: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
1857: return(0);
1858: }
1860: /*@
1861: MatGetValues - Gets a block of values from a matrix.
1863: Not Collective; currently only returns a local block
1865: Input Parameters:
1866: + mat - the matrix
1867: . v - a logically two-dimensional array for storing the values
1868: . m, idxm - the number of rows and their global indices
1869: - n, idxn - the number of columns and their global indices
1871: Notes:
1872: The user must allocate space (m*n PetscScalars) for the values, v.
1873: The values, v, are then returned in a row-oriented format,
1874: analogous to that used by default in MatSetValues().
1876: MatGetValues() uses 0-based row and column numbers in
1877: Fortran as well as in C.
1879: MatGetValues() requires that the matrix has been assembled
1880: with MatAssemblyBegin()/MatAssemblyEnd(). Thus, calls to
1881: MatSetValues() and MatGetValues() CANNOT be made in succession
1882: without intermediate matrix assembly.
1884: Negative row or column indices will be ignored and those locations in v[] will be
1885: left unchanged.
1887: Level: advanced
1889: .seealso: MatGetRow(), MatCreateSubMatrices(), MatSetValues()
1890: @*/
1891: PetscErrorCode MatGetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[])
1892: {
1898: if (!m || !n) return(0);
1902: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
1903: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1904: if (!mat->ops->getvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
1905: MatCheckPreallocated(mat,1);
1907: PetscLogEventBegin(MAT_GetValues,mat,0,0,0);
1908: (*mat->ops->getvalues)(mat,m,idxm,n,idxn,v);
1909: PetscLogEventEnd(MAT_GetValues,mat,0,0,0);
1910: return(0);
1911: }
1913: /*@
1914: MatSetValuesBatch - Adds (ADD_VALUES) many blocks of values into a matrix at once. The blocks must all be square and
1915: the same size. Currently, this can only be called once and creates the given matrix.
1917: Not Collective
1919: Input Parameters:
1920: + mat - the matrix
1921: . nb - the number of blocks
1922: . bs - the number of rows (and columns) in each block
1923: . rows - a concatenation of the rows for each block
1924: - v - a concatenation of logically two-dimensional arrays of values
1926: Notes:
1927: In the future, we will extend this routine to handle rectangular blocks, and to allow multiple calls for a given matrix.
1929: Level: advanced
1931: .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(),
1932: InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues()
1933: @*/
1934: PetscErrorCode MatSetValuesBatch(Mat mat, PetscInt nb, PetscInt bs, PetscInt rows[], const PetscScalar v[])
1935: {
1943: #if defined(PETSC_USE_DEBUG)
1944: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
1945: #endif
1947: PetscLogEventBegin(MAT_SetValuesBatch,mat,0,0,0);
1948: if (mat->ops->setvaluesbatch) {
1949: (*mat->ops->setvaluesbatch)(mat,nb,bs,rows,v);
1950: } else {
1951: PetscInt b;
1952: for (b = 0; b < nb; ++b) {
1953: MatSetValues(mat, bs, &rows[b*bs], bs, &rows[b*bs], &v[b*bs*bs], ADD_VALUES);
1954: }
1955: }
1956: PetscLogEventEnd(MAT_SetValuesBatch,mat,0,0,0);
1957: return(0);
1958: }
1960: /*@
1961: MatSetLocalToGlobalMapping - Sets a local-to-global numbering for use by
1962: the routine MatSetValuesLocal() to allow users to insert matrix entries
1963: using a local (per-processor) numbering.
1965: Not Collective
1967: Input Parameters:
1968: + x - the matrix
1969: . rmapping - row mapping created with ISLocalToGlobalMappingCreate() or ISLocalToGlobalMappingCreateIS()
1970: - cmapping - column mapping
1972: Level: intermediate
1975: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesLocal()
1976: @*/
1977: PetscErrorCode MatSetLocalToGlobalMapping(Mat x,ISLocalToGlobalMapping rmapping,ISLocalToGlobalMapping cmapping)
1978: {
1987: if (x->ops->setlocaltoglobalmapping) {
1988: (*x->ops->setlocaltoglobalmapping)(x,rmapping,cmapping);
1989: } else {
1990: PetscLayoutSetISLocalToGlobalMapping(x->rmap,rmapping);
1991: PetscLayoutSetISLocalToGlobalMapping(x->cmap,cmapping);
1992: }
1993: return(0);
1994: }
1997: /*@
1998: MatGetLocalToGlobalMapping - Gets the local-to-global numbering set by MatSetLocalToGlobalMapping()
2000: Not Collective
2002: Input Parameters:
2003: . A - the matrix
2005: Output Parameters:
2006: + rmapping - row mapping
2007: - cmapping - column mapping
2009: Level: advanced
2012: .seealso: MatSetValuesLocal()
2013: @*/
2014: PetscErrorCode MatGetLocalToGlobalMapping(Mat A,ISLocalToGlobalMapping *rmapping,ISLocalToGlobalMapping *cmapping)
2015: {
2021: if (rmapping) *rmapping = A->rmap->mapping;
2022: if (cmapping) *cmapping = A->cmap->mapping;
2023: return(0);
2024: }
2026: /*@
2027: MatGetLayouts - Gets the PetscLayout objects for rows and columns
2029: Not Collective
2031: Input Parameters:
2032: . A - the matrix
2034: Output Parameters:
2035: + rmap - row layout
2036: - cmap - column layout
2038: Level: advanced
2040: .seealso: MatCreateVecs(), MatGetLocalToGlobalMapping()
2041: @*/
2042: PetscErrorCode MatGetLayouts(Mat A,PetscLayout *rmap,PetscLayout *cmap)
2043: {
2049: if (rmap) *rmap = A->rmap;
2050: if (cmap) *cmap = A->cmap;
2051: return(0);
2052: }
2054: /*@C
2055: MatSetValuesLocal - Inserts or adds values into certain locations of a matrix,
2056: using a local ordering of the nodes.
2058: Not Collective
2060: Input Parameters:
2061: + mat - the matrix
2062: . nrow, irow - number of rows and their local indices
2063: . ncol, icol - number of columns and their local indices
2064: . y - a logically two-dimensional array of values
2065: - addv - either INSERT_VALUES or ADD_VALUES, where
2066: ADD_VALUES adds values to any existing entries, and
2067: INSERT_VALUES replaces existing entries with new values
2069: Notes:
2070: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
2071: MatSetUp() before using this routine
2073: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatSetLocalToGlobalMapping() before using this routine
2075: Calls to MatSetValuesLocal() with the INSERT_VALUES and ADD_VALUES
2076: options cannot be mixed without intervening calls to the assembly
2077: routines.
2079: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
2080: MUST be called after all calls to MatSetValuesLocal() have been completed.
2082: Level: intermediate
2084: Developer Notes:
2085: This is labeled with C so does not automatically generate Fortran stubs and interfaces
2086: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
2088: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetLocalToGlobalMapping(),
2089: MatSetValueLocal()
2090: @*/
2091: PetscErrorCode MatSetValuesLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv)
2092: {
2098: MatCheckPreallocated(mat,1);
2099: if (!nrow || !ncol) return(0); /* no values to insert */
2102: if (mat->insertmode == NOT_SET_VALUES) {
2103: mat->insertmode = addv;
2104: }
2105: #if defined(PETSC_USE_DEBUG)
2106: else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
2107: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2108: if (!mat->ops->setvalueslocal && !mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2109: #endif
2111: if (mat->assembled) {
2112: mat->was_assembled = PETSC_TRUE;
2113: mat->assembled = PETSC_FALSE;
2114: }
2115: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
2116: if (mat->ops->setvalueslocal) {
2117: (*mat->ops->setvalueslocal)(mat,nrow,irow,ncol,icol,y,addv);
2118: } else {
2119: PetscInt buf[8192],*bufr=0,*bufc=0,*irowm,*icolm;
2120: if ((nrow+ncol) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
2121: irowm = buf; icolm = buf+nrow;
2122: } else {
2123: PetscMalloc2(nrow,&bufr,ncol,&bufc);
2124: irowm = bufr; icolm = bufc;
2125: }
2126: ISLocalToGlobalMappingApply(mat->rmap->mapping,nrow,irow,irowm);
2127: ISLocalToGlobalMappingApply(mat->cmap->mapping,ncol,icol,icolm);
2128: MatSetValues(mat,nrow,irowm,ncol,icolm,y,addv);
2129: PetscFree2(bufr,bufc);
2130: }
2131: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
2132: return(0);
2133: }
2135: /*@C
2136: MatSetValuesBlockedLocal - Inserts or adds values into certain locations of a matrix,
2137: using a local ordering of the nodes a block at a time.
2139: Not Collective
2141: Input Parameters:
2142: + x - the matrix
2143: . nrow, irow - number of rows and their local indices
2144: . ncol, icol - number of columns and their local indices
2145: . y - a logically two-dimensional array of values
2146: - addv - either INSERT_VALUES or ADD_VALUES, where
2147: ADD_VALUES adds values to any existing entries, and
2148: INSERT_VALUES replaces existing entries with new values
2150: Notes:
2151: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatXXXXSetPreallocation() or
2152: MatSetUp() before using this routine
2154: If you create the matrix yourself (that is not with a call to DMCreateMatrix()) then you MUST call MatSetBlockSize() and MatSetLocalToGlobalMapping()
2155: before using this routineBefore calling MatSetValuesLocal(), the user must first set the
2157: Calls to MatSetValuesBlockedLocal() with the INSERT_VALUES and ADD_VALUES
2158: options cannot be mixed without intervening calls to the assembly
2159: routines.
2161: These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd()
2162: MUST be called after all calls to MatSetValuesBlockedLocal() have been completed.
2164: Level: intermediate
2166: Developer Notes:
2167: This is labeled with C so does not automatically generate Fortran stubs and interfaces
2168: because it requires multiple Fortran interfaces depending on which arguments are scalar or arrays.
2170: .seealso: MatSetBlockSize(), MatSetLocalToGlobalMapping(), MatAssemblyBegin(), MatAssemblyEnd(),
2171: MatSetValuesLocal(), MatSetValuesBlocked()
2172: @*/
2173: PetscErrorCode MatSetValuesBlockedLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv)
2174: {
2180: MatCheckPreallocated(mat,1);
2181: if (!nrow || !ncol) return(0); /* no values to insert */
2185: if (mat->insertmode == NOT_SET_VALUES) {
2186: mat->insertmode = addv;
2187: }
2188: #if defined(PETSC_USE_DEBUG)
2189: else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values");
2190: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2191: 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);
2192: #endif
2194: if (mat->assembled) {
2195: mat->was_assembled = PETSC_TRUE;
2196: mat->assembled = PETSC_FALSE;
2197: }
2198: #if defined(PETSC_USE_DEBUG)
2199: /* Condition on the mapping existing, because MatSetValuesBlockedLocal_IS does not require it to be set. */
2200: if (mat->rmap->mapping) {
2201: PetscInt irbs, rbs;
2202: MatGetBlockSizes(mat, &rbs, NULL);
2203: ISLocalToGlobalMappingGetBlockSize(mat->rmap->mapping,&irbs);
2204: if (rbs != irbs) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Different row block sizes! mat %D, row l2g map %D",rbs,irbs);
2205: }
2206: if (mat->cmap->mapping) {
2207: PetscInt icbs, cbs;
2208: MatGetBlockSizes(mat,NULL,&cbs);
2209: ISLocalToGlobalMappingGetBlockSize(mat->cmap->mapping,&icbs);
2210: if (cbs != icbs) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Different col block sizes! mat %D, col l2g map %D",cbs,icbs);
2211: }
2212: #endif
2213: PetscLogEventBegin(MAT_SetValues,mat,0,0,0);
2214: if (mat->ops->setvaluesblockedlocal) {
2215: (*mat->ops->setvaluesblockedlocal)(mat,nrow,irow,ncol,icol,y,addv);
2216: } else {
2217: PetscInt buf[8192],*bufr=0,*bufc=0,*irowm,*icolm;
2218: if ((nrow+ncol) <= (PetscInt)(sizeof(buf)/sizeof(PetscInt))) {
2219: irowm = buf; icolm = buf + nrow;
2220: } else {
2221: PetscMalloc2(nrow,&bufr,ncol,&bufc);
2222: irowm = bufr; icolm = bufc;
2223: }
2224: ISLocalToGlobalMappingApplyBlock(mat->rmap->mapping,nrow,irow,irowm);
2225: ISLocalToGlobalMappingApplyBlock(mat->cmap->mapping,ncol,icol,icolm);
2226: MatSetValuesBlocked(mat,nrow,irowm,ncol,icolm,y,addv);
2227: PetscFree2(bufr,bufc);
2228: }
2229: PetscLogEventEnd(MAT_SetValues,mat,0,0,0);
2230: return(0);
2231: }
2233: /*@
2234: MatMultDiagonalBlock - Computes the matrix-vector product, y = Dx. Where D is defined by the inode or block structure of the diagonal
2236: Collective on Mat
2238: Input Parameters:
2239: + mat - the matrix
2240: - x - the vector to be multiplied
2242: Output Parameters:
2243: . y - the result
2245: Notes:
2246: The vectors x and y cannot be the same. I.e., one cannot
2247: call MatMult(A,y,y).
2249: Level: developer
2251: .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2252: @*/
2253: PetscErrorCode MatMultDiagonalBlock(Mat mat,Vec x,Vec y)
2254: {
2263: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2264: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2265: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2266: MatCheckPreallocated(mat,1);
2268: if (!mat->ops->multdiagonalblock) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"This matrix type does not have a multiply defined");
2269: (*mat->ops->multdiagonalblock)(mat,x,y);
2270: PetscObjectStateIncrease((PetscObject)y);
2271: return(0);
2272: }
2274: /* --------------------------------------------------------*/
2275: /*@
2276: MatMult - Computes the matrix-vector product, y = Ax.
2278: Neighbor-wise Collective on Mat
2280: Input Parameters:
2281: + mat - the matrix
2282: - x - the vector to be multiplied
2284: Output Parameters:
2285: . y - the result
2287: Notes:
2288: The vectors x and y cannot be the same. I.e., one cannot
2289: call MatMult(A,y,y).
2291: Level: beginner
2293: .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2294: @*/
2295: PetscErrorCode MatMult(Mat mat,Vec x,Vec y)
2296: {
2304: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2305: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2306: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2307: #if !defined(PETSC_HAVE_CONSTRAINTS)
2308: 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);
2309: 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);
2310: 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);
2311: #endif
2312: VecSetErrorIfLocked(y,3);
2313: if (mat->erroriffailure) {VecValidValues(x,2,PETSC_TRUE);}
2314: MatCheckPreallocated(mat,1);
2316: VecLockReadPush(x);
2317: if (!mat->ops->mult) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"This matrix type does not have a multiply defined");
2318: PetscLogEventBegin(MAT_Mult,mat,x,y,0);
2319: (*mat->ops->mult)(mat,x,y);
2320: PetscLogEventEnd(MAT_Mult,mat,x,y,0);
2321: if (mat->erroriffailure) {VecValidValues(y,3,PETSC_FALSE);}
2322: VecLockReadPop(x);
2323: return(0);
2324: }
2326: /*@
2327: MatMultTranspose - Computes matrix transpose times a vector y = A^T * x.
2329: Neighbor-wise Collective on Mat
2331: Input Parameters:
2332: + mat - the matrix
2333: - x - the vector to be multiplied
2335: Output Parameters:
2336: . y - the result
2338: Notes:
2339: The vectors x and y cannot be the same. I.e., one cannot
2340: call MatMultTranspose(A,y,y).
2342: For complex numbers this does NOT compute the Hermitian (complex conjugate) transpose multiple,
2343: use MatMultHermitianTranspose()
2345: Level: beginner
2347: .seealso: MatMult(), MatMultAdd(), MatMultTransposeAdd(), MatMultHermitianTranspose(), MatTranspose()
2348: @*/
2349: PetscErrorCode MatMultTranspose(Mat mat,Vec x,Vec y)
2350: {
2359: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2360: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2361: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2362: #if !defined(PETSC_HAVE_CONSTRAINTS)
2363: 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);
2364: 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);
2365: #endif
2366: if (mat->erroriffailure) {VecValidValues(x,2,PETSC_TRUE);}
2367: MatCheckPreallocated(mat,1);
2369: if (!mat->ops->multtranspose) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"This matrix type does not have a multiply transpose defined");
2370: PetscLogEventBegin(MAT_MultTranspose,mat,x,y,0);
2371: VecLockReadPush(x);
2372: (*mat->ops->multtranspose)(mat,x,y);
2373: VecLockReadPop(x);
2374: PetscLogEventEnd(MAT_MultTranspose,mat,x,y,0);
2375: PetscObjectStateIncrease((PetscObject)y);
2376: if (mat->erroriffailure) {VecValidValues(y,3,PETSC_FALSE);}
2377: return(0);
2378: }
2380: /*@
2381: MatMultHermitianTranspose - Computes matrix Hermitian transpose times a vector.
2383: Neighbor-wise Collective on Mat
2385: Input Parameters:
2386: + mat - the matrix
2387: - x - the vector to be multilplied
2389: Output Parameters:
2390: . y - the result
2392: Notes:
2393: The vectors x and y cannot be the same. I.e., one cannot
2394: call MatMultHermitianTranspose(A,y,y).
2396: Also called the conjugate transpose, complex conjugate transpose, or adjoint.
2398: For real numbers MatMultTranspose() and MatMultHermitianTranspose() are identical.
2400: Level: beginner
2402: .seealso: MatMult(), MatMultAdd(), MatMultHermitianTransposeAdd(), MatMultTranspose()
2403: @*/
2404: PetscErrorCode MatMultHermitianTranspose(Mat mat,Vec x,Vec y)
2405: {
2407: Vec w;
2415: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2416: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2417: if (x == y) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2418: #if !defined(PETSC_HAVE_CONSTRAINTS)
2419: 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);
2420: 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);
2421: #endif
2422: MatCheckPreallocated(mat,1);
2424: PetscLogEventBegin(MAT_MultHermitianTranspose,mat,x,y,0);
2425: if (mat->ops->multhermitiantranspose) {
2426: VecLockReadPush(x);
2427: (*mat->ops->multhermitiantranspose)(mat,x,y);
2428: VecLockReadPop(x);
2429: } else {
2430: VecDuplicate(x,&w);
2431: VecCopy(x,w);
2432: VecConjugate(w);
2433: MatMultTranspose(mat,w,y);
2434: VecDestroy(&w);
2435: VecConjugate(y);
2436: }
2437: PetscLogEventEnd(MAT_MultHermitianTranspose,mat,x,y,0);
2438: PetscObjectStateIncrease((PetscObject)y);
2439: return(0);
2440: }
2442: /*@
2443: MatMultAdd - Computes v3 = v2 + A * v1.
2445: Neighbor-wise Collective on Mat
2447: Input Parameters:
2448: + mat - the matrix
2449: - v1, v2 - the vectors
2451: Output Parameters:
2452: . v3 - the result
2454: Notes:
2455: The vectors v1 and v3 cannot be the same. I.e., one cannot
2456: call MatMultAdd(A,v1,v2,v1).
2458: Level: beginner
2460: .seealso: MatMultTranspose(), MatMult(), MatMultTransposeAdd()
2461: @*/
2462: PetscErrorCode MatMultAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2463: {
2473: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2474: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2475: 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);
2476: /* 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);
2477: 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); */
2478: 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);
2479: 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);
2480: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2481: MatCheckPreallocated(mat,1);
2483: if (!mat->ops->multadd) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"No MatMultAdd() for matrix type '%s'",((PetscObject)mat)->type_name);
2484: PetscLogEventBegin(MAT_MultAdd,mat,v1,v2,v3);
2485: VecLockReadPush(v1);
2486: (*mat->ops->multadd)(mat,v1,v2,v3);
2487: VecLockReadPop(v1);
2488: PetscLogEventEnd(MAT_MultAdd,mat,v1,v2,v3);
2489: PetscObjectStateIncrease((PetscObject)v3);
2490: return(0);
2491: }
2493: /*@
2494: MatMultTransposeAdd - Computes v3 = v2 + A' * v1.
2496: Neighbor-wise Collective on Mat
2498: Input Parameters:
2499: + mat - the matrix
2500: - v1, v2 - the vectors
2502: Output Parameters:
2503: . v3 - the result
2505: Notes:
2506: The vectors v1 and v3 cannot be the same. I.e., one cannot
2507: call MatMultTransposeAdd(A,v1,v2,v1).
2509: Level: beginner
2511: .seealso: MatMultTranspose(), MatMultAdd(), MatMult()
2512: @*/
2513: PetscErrorCode MatMultTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2514: {
2524: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2525: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2526: if (!mat->ops->multtransposeadd) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2527: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2528: 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);
2529: 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);
2530: 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);
2531: MatCheckPreallocated(mat,1);
2533: PetscLogEventBegin(MAT_MultTransposeAdd,mat,v1,v2,v3);
2534: VecLockReadPush(v1);
2535: (*mat->ops->multtransposeadd)(mat,v1,v2,v3);
2536: VecLockReadPop(v1);
2537: PetscLogEventEnd(MAT_MultTransposeAdd,mat,v1,v2,v3);
2538: PetscObjectStateIncrease((PetscObject)v3);
2539: return(0);
2540: }
2542: /*@
2543: MatMultHermitianTransposeAdd - Computes v3 = v2 + A^H * v1.
2545: Neighbor-wise Collective on Mat
2547: Input Parameters:
2548: + mat - the matrix
2549: - v1, v2 - the vectors
2551: Output Parameters:
2552: . v3 - the result
2554: Notes:
2555: The vectors v1 and v3 cannot be the same. I.e., one cannot
2556: call MatMultHermitianTransposeAdd(A,v1,v2,v1).
2558: Level: beginner
2560: .seealso: MatMultHermitianTranspose(), MatMultTranspose(), MatMultAdd(), MatMult()
2561: @*/
2562: PetscErrorCode MatMultHermitianTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3)
2563: {
2573: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2574: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2575: if (v1 == v3) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors");
2576: 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);
2577: 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);
2578: 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);
2579: MatCheckPreallocated(mat,1);
2581: PetscLogEventBegin(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);
2582: VecLockReadPush(v1);
2583: if (mat->ops->multhermitiantransposeadd) {
2584: (*mat->ops->multhermitiantransposeadd)(mat,v1,v2,v3);
2585: } else {
2586: Vec w,z;
2587: VecDuplicate(v1,&w);
2588: VecCopy(v1,w);
2589: VecConjugate(w);
2590: VecDuplicate(v3,&z);
2591: MatMultTranspose(mat,w,z);
2592: VecDestroy(&w);
2593: VecConjugate(z);
2594: if (v2 != v3) {
2595: VecWAXPY(v3,1.0,v2,z);
2596: } else {
2597: VecAXPY(v3,1.0,z);
2598: }
2599: VecDestroy(&z);
2600: }
2601: VecLockReadPop(v1);
2602: PetscLogEventEnd(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);
2603: PetscObjectStateIncrease((PetscObject)v3);
2604: return(0);
2605: }
2607: /*@
2608: MatMultConstrained - The inner multiplication routine for a
2609: constrained matrix P^T A P.
2611: Neighbor-wise Collective on Mat
2613: Input Parameters:
2614: + mat - the matrix
2615: - x - the vector to be multilplied
2617: Output Parameters:
2618: . y - the result
2620: Notes:
2621: The vectors x and y cannot be the same. I.e., one cannot
2622: call MatMult(A,y,y).
2624: Level: beginner
2626: .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2627: @*/
2628: PetscErrorCode MatMultConstrained(Mat mat,Vec x,Vec y)
2629: {
2636: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2637: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2638: if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2639: 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);
2640: 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);
2641: 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);
2643: PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);
2644: VecLockReadPush(x);
2645: (*mat->ops->multconstrained)(mat,x,y);
2646: VecLockReadPop(x);
2647: PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);
2648: PetscObjectStateIncrease((PetscObject)y);
2649: return(0);
2650: }
2652: /*@
2653: MatMultTransposeConstrained - The inner multiplication routine for a
2654: constrained matrix P^T A^T P.
2656: Neighbor-wise Collective on Mat
2658: Input Parameters:
2659: + mat - the matrix
2660: - x - the vector to be multilplied
2662: Output Parameters:
2663: . y - the result
2665: Notes:
2666: The vectors x and y cannot be the same. I.e., one cannot
2667: call MatMult(A,y,y).
2669: Level: beginner
2671: .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd()
2672: @*/
2673: PetscErrorCode MatMultTransposeConstrained(Mat mat,Vec x,Vec y)
2674: {
2681: if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2682: if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2683: if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors");
2684: 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);
2685: 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);
2687: PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);
2688: (*mat->ops->multtransposeconstrained)(mat,x,y);
2689: PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);
2690: PetscObjectStateIncrease((PetscObject)y);
2691: return(0);
2692: }
2694: /*@C
2695: MatGetFactorType - gets the type of factorization it is
2697: Not Collective
2699: Input Parameters:
2700: . mat - the matrix
2702: Output Parameters:
2703: . t - the type, one of MAT_FACTOR_NONE, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ILU, MAT_FACTOR_ICC,MAT_FACTOR_ILUDT
2705: Level: intermediate
2707: .seealso: MatFactorType, MatGetFactor(), MatSetFactorType()
2708: @*/
2709: PetscErrorCode MatGetFactorType(Mat mat,MatFactorType *t)
2710: {
2715: *t = mat->factortype;
2716: return(0);
2717: }
2719: /*@C
2720: MatSetFactorType - sets the type of factorization it is
2722: Logically Collective on Mat
2724: Input Parameters:
2725: + mat - the matrix
2726: - t - the type, one of MAT_FACTOR_NONE, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ILU, MAT_FACTOR_ICC,MAT_FACTOR_ILUDT
2728: Level: intermediate
2730: .seealso: MatFactorType, MatGetFactor(), MatGetFactorType()
2731: @*/
2732: PetscErrorCode MatSetFactorType(Mat mat, MatFactorType t)
2733: {
2737: mat->factortype = t;
2738: return(0);
2739: }
2741: /* ------------------------------------------------------------*/
2742: /*@C
2743: MatGetInfo - Returns information about matrix storage (number of
2744: nonzeros, memory, etc.).
2746: Collective on Mat if MAT_GLOBAL_MAX or MAT_GLOBAL_SUM is used as the flag
2748: Input Parameters:
2749: . mat - the matrix
2751: Output Parameters:
2752: + flag - flag indicating the type of parameters to be returned
2753: (MAT_LOCAL - local matrix, MAT_GLOBAL_MAX - maximum over all processors,
2754: MAT_GLOBAL_SUM - sum over all processors)
2755: - info - matrix information context
2757: Notes:
2758: The MatInfo context contains a variety of matrix data, including
2759: number of nonzeros allocated and used, number of mallocs during
2760: matrix assembly, etc. Additional information for factored matrices
2761: is provided (such as the fill ratio, number of mallocs during
2762: factorization, etc.). Much of this info is printed to PETSC_STDOUT
2763: when using the runtime options
2764: $ -info -mat_view ::ascii_info
2766: Example for C/C++ Users:
2767: See the file ${PETSC_DIR}/include/petscmat.h for a complete list of
2768: data within the MatInfo context. For example,
2769: .vb
2770: MatInfo info;
2771: Mat A;
2772: double mal, nz_a, nz_u;
2774: MatGetInfo(A,MAT_LOCAL,&info);
2775: mal = info.mallocs;
2776: nz_a = info.nz_allocated;
2777: .ve
2779: Example for Fortran Users:
2780: Fortran users should declare info as a double precision
2781: array of dimension MAT_INFO_SIZE, and then extract the parameters
2782: of interest. See the file ${PETSC_DIR}/include/petsc/finclude/petscmat.h
2783: a complete list of parameter names.
2784: .vb
2785: double precision info(MAT_INFO_SIZE)
2786: double precision mal, nz_a
2787: Mat A
2788: integer ierr
2790: call MatGetInfo(A,MAT_LOCAL,info,ierr)
2791: mal = info(MAT_INFO_MALLOCS)
2792: nz_a = info(MAT_INFO_NZ_ALLOCATED)
2793: .ve
2795: Level: intermediate
2797: Developer Note: fortran interface is not autogenerated as the f90
2798: interface defintion cannot be generated correctly [due to MatInfo]
2800: .seealso: MatStashGetInfo()
2802: @*/
2803: PetscErrorCode MatGetInfo(Mat mat,MatInfoType flag,MatInfo *info)
2804: {
2811: if (!mat->ops->getinfo) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2812: MatCheckPreallocated(mat,1);
2813: (*mat->ops->getinfo)(mat,flag,info);
2814: return(0);
2815: }
2817: /*
2818: This is used by external packages where it is not easy to get the info from the actual
2819: matrix factorization.
2820: */
2821: PetscErrorCode MatGetInfo_External(Mat A,MatInfoType flag,MatInfo *info)
2822: {
2826: PetscMemzero(info,sizeof(MatInfo));
2827: return(0);
2828: }
2830: /* ----------------------------------------------------------*/
2832: /*@C
2833: MatLUFactor - Performs in-place LU factorization of matrix.
2835: Collective on Mat
2837: Input Parameters:
2838: + mat - the matrix
2839: . row - row permutation
2840: . col - column permutation
2841: - info - options for factorization, includes
2842: $ fill - expected fill as ratio of original fill.
2843: $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting)
2844: $ Run with the option -info to determine an optimal value to use
2846: Notes:
2847: Most users should employ the simplified KSP interface for linear solvers
2848: instead of working directly with matrix algebra routines such as this.
2849: See, e.g., KSPCreate().
2851: This changes the state of the matrix to a factored matrix; it cannot be used
2852: for example with MatSetValues() unless one first calls MatSetUnfactored().
2854: Level: developer
2856: .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(),
2857: MatGetOrdering(), MatSetUnfactored(), MatFactorInfo, MatGetFactor()
2859: Developer Note: fortran interface is not autogenerated as the f90
2860: interface defintion cannot be generated correctly [due to MatFactorInfo]
2862: @*/
2863: PetscErrorCode MatLUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info)
2864: {
2866: MatFactorInfo tinfo;
2874: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2875: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2876: if (!mat->ops->lufactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2877: MatCheckPreallocated(mat,1);
2878: if (!info) {
2879: MatFactorInfoInitialize(&tinfo);
2880: info = &tinfo;
2881: }
2883: PetscLogEventBegin(MAT_LUFactor,mat,row,col,0);
2884: (*mat->ops->lufactor)(mat,row,col,info);
2885: PetscLogEventEnd(MAT_LUFactor,mat,row,col,0);
2886: PetscObjectStateIncrease((PetscObject)mat);
2887: return(0);
2888: }
2890: /*@C
2891: MatILUFactor - Performs in-place ILU factorization of matrix.
2893: Collective on Mat
2895: Input Parameters:
2896: + mat - the matrix
2897: . row - row permutation
2898: . col - column permutation
2899: - info - structure containing
2900: $ levels - number of levels of fill.
2901: $ expected fill - as ratio of original fill.
2902: $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices
2903: missing diagonal entries)
2905: Notes:
2906: Probably really in-place only when level of fill is zero, otherwise allocates
2907: new space to store factored matrix and deletes previous memory.
2909: Most users should employ the simplified KSP interface for linear solvers
2910: instead of working directly with matrix algebra routines such as this.
2911: See, e.g., KSPCreate().
2913: Level: developer
2915: .seealso: MatILUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo
2917: Developer Note: fortran interface is not autogenerated as the f90
2918: interface defintion cannot be generated correctly [due to MatFactorInfo]
2920: @*/
2921: PetscErrorCode MatILUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info)
2922: {
2931: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"matrix must be square");
2932: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
2933: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2934: if (!mat->ops->ilufactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
2935: MatCheckPreallocated(mat,1);
2937: PetscLogEventBegin(MAT_ILUFactor,mat,row,col,0);
2938: (*mat->ops->ilufactor)(mat,row,col,info);
2939: PetscLogEventEnd(MAT_ILUFactor,mat,row,col,0);
2940: PetscObjectStateIncrease((PetscObject)mat);
2941: return(0);
2942: }
2944: /*@C
2945: MatLUFactorSymbolic - Performs symbolic LU factorization of matrix.
2946: Call this routine before calling MatLUFactorNumeric().
2948: Collective on Mat
2950: Input Parameters:
2951: + fact - the factor matrix obtained with MatGetFactor()
2952: . mat - the matrix
2953: . row, col - row and column permutations
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
2960: Notes:
2961: See Users-Manual: ch_mat for additional information about choosing the fill factor for better efficiency.
2963: Most users should employ the simplified KSP interface for linear solvers
2964: instead of working directly with matrix algebra routines such as this.
2965: See, e.g., KSPCreate().
2967: Level: developer
2969: .seealso: MatLUFactor(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo, MatFactorInfoInitialize()
2971: Developer Note: fortran interface is not autogenerated as the f90
2972: interface defintion cannot be generated correctly [due to MatFactorInfo]
2974: @*/
2975: PetscErrorCode MatLUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info)
2976: {
2978: 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 (!(fact)->ops->lufactorsymbolic) {
2990: MatSolverType spackage;
2991: MatFactorGetSolverType(fact,&spackage);
2992: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic LU using solver package %s",((PetscObject)mat)->type_name,spackage);
2993: }
2994: MatCheckPreallocated(mat,2);
2995: if (!info) {
2996: MatFactorInfoInitialize(&tinfo);
2997: info = &tinfo;
2998: }
3000: PetscLogEventBegin(MAT_LUFactorSymbolic,mat,row,col,0);
3001: (fact->ops->lufactorsymbolic)(fact,mat,row,col,info);
3002: PetscLogEventEnd(MAT_LUFactorSymbolic,mat,row,col,0);
3003: PetscObjectStateIncrease((PetscObject)fact);
3004: return(0);
3005: }
3007: /*@C
3008: MatLUFactorNumeric - Performs numeric LU factorization of a matrix.
3009: Call this routine after first calling MatLUFactorSymbolic().
3011: Collective on Mat
3013: Input Parameters:
3014: + fact - the factor matrix obtained with MatGetFactor()
3015: . mat - the matrix
3016: - info - options for factorization
3018: Notes:
3019: See MatLUFactor() for in-place factorization. See
3020: MatCholeskyFactorNumeric() for the symmetric, positive definite case.
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: MatLUFactorSymbolic(), MatLUFactor(), MatCholeskyFactor()
3030: Developer Note: fortran interface is not autogenerated as the f90
3031: interface defintion cannot be generated correctly [due to MatFactorInfo]
3033: @*/
3034: PetscErrorCode MatLUFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info)
3035: {
3036: MatFactorInfo tinfo;
3044: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3045: 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);
3047: if (!(fact)->ops->lufactornumeric) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s numeric LU",((PetscObject)mat)->type_name);
3048: MatCheckPreallocated(mat,2);
3049: if (!info) {
3050: MatFactorInfoInitialize(&tinfo);
3051: info = &tinfo;
3052: }
3054: PetscLogEventBegin(MAT_LUFactorNumeric,mat,fact,0,0);
3055: (fact->ops->lufactornumeric)(fact,mat,info);
3056: PetscLogEventEnd(MAT_LUFactorNumeric,mat,fact,0,0);
3057: MatViewFromOptions(fact,NULL,"-mat_factor_view");
3058: PetscObjectStateIncrease((PetscObject)fact);
3059: return(0);
3060: }
3062: /*@C
3063: MatCholeskyFactor - Performs in-place Cholesky factorization of a
3064: symmetric matrix.
3066: Collective on Mat
3068: Input Parameters:
3069: + mat - the matrix
3070: . perm - row and column permutations
3071: - f - expected fill as ratio of original fill
3073: Notes:
3074: See MatLUFactor() for the nonsymmetric case. See also
3075: MatCholeskyFactorSymbolic(), and MatCholeskyFactorNumeric().
3077: Most users should employ the simplified KSP interface for linear solvers
3078: instead of working directly with matrix algebra routines such as this.
3079: See, e.g., KSPCreate().
3081: Level: developer
3083: .seealso: MatLUFactor(), MatCholeskyFactorSymbolic(), MatCholeskyFactorNumeric()
3084: MatGetOrdering()
3086: Developer Note: fortran interface is not autogenerated as the f90
3087: interface defintion cannot be generated correctly [due to MatFactorInfo]
3089: @*/
3090: PetscErrorCode MatCholeskyFactor(Mat mat,IS perm,const MatFactorInfo *info)
3091: {
3093: MatFactorInfo tinfo;
3100: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"Matrix must be square");
3101: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3102: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3103: 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);
3104: MatCheckPreallocated(mat,1);
3105: if (!info) {
3106: MatFactorInfoInitialize(&tinfo);
3107: info = &tinfo;
3108: }
3110: PetscLogEventBegin(MAT_CholeskyFactor,mat,perm,0,0);
3111: (*mat->ops->choleskyfactor)(mat,perm,info);
3112: PetscLogEventEnd(MAT_CholeskyFactor,mat,perm,0,0);
3113: PetscObjectStateIncrease((PetscObject)mat);
3114: return(0);
3115: }
3117: /*@C
3118: MatCholeskyFactorSymbolic - Performs symbolic Cholesky factorization
3119: of a symmetric matrix.
3121: Collective on Mat
3123: Input Parameters:
3124: + fact - the factor matrix obtained with MatGetFactor()
3125: . mat - the matrix
3126: . perm - row and column permutations
3127: - info - options for factorization, includes
3128: $ fill - expected fill as ratio of original fill.
3129: $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting)
3130: $ Run with the option -info to determine an optimal value to use
3132: Notes:
3133: See MatLUFactorSymbolic() for the nonsymmetric case. See also
3134: MatCholeskyFactor() and MatCholeskyFactorNumeric().
3136: Most users should employ the simplified KSP interface for linear solvers
3137: instead of working directly with matrix algebra routines such as this.
3138: See, e.g., KSPCreate().
3140: Level: developer
3142: .seealso: MatLUFactorSymbolic(), MatCholeskyFactor(), MatCholeskyFactorNumeric()
3143: MatGetOrdering()
3145: Developer Note: fortran interface is not autogenerated as the f90
3146: interface defintion cannot be generated correctly [due to MatFactorInfo]
3148: @*/
3149: PetscErrorCode MatCholeskyFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info)
3150: {
3152: MatFactorInfo tinfo;
3160: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"Matrix must be square");
3161: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3162: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3163: if (!(fact)->ops->choleskyfactorsymbolic) {
3164: MatSolverType spackage;
3165: MatFactorGetSolverType(fact,&spackage);
3166: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s symbolic factor Cholesky using solver package %s",((PetscObject)mat)->type_name,spackage);
3167: }
3168: MatCheckPreallocated(mat,2);
3169: if (!info) {
3170: MatFactorInfoInitialize(&tinfo);
3171: info = &tinfo;
3172: }
3174: PetscLogEventBegin(MAT_CholeskyFactorSymbolic,mat,perm,0,0);
3175: (fact->ops->choleskyfactorsymbolic)(fact,mat,perm,info);
3176: PetscLogEventEnd(MAT_CholeskyFactorSymbolic,mat,perm,0,0);
3177: PetscObjectStateIncrease((PetscObject)fact);
3178: return(0);
3179: }
3181: /*@C
3182: MatCholeskyFactorNumeric - Performs numeric Cholesky factorization
3183: of a symmetric matrix. Call this routine after first calling
3184: MatCholeskyFactorSymbolic().
3186: Collective on Mat
3188: Input Parameters:
3189: + fact - the factor matrix obtained with MatGetFactor()
3190: . mat - the initial matrix
3191: . info - options for factorization
3192: - fact - the symbolic factor of mat
3195: Notes:
3196: Most users should employ the simplified KSP interface for linear solvers
3197: instead of working directly with matrix algebra routines such as this.
3198: See, e.g., KSPCreate().
3200: Level: developer
3202: .seealso: MatCholeskyFactorSymbolic(), MatCholeskyFactor(), MatLUFactorNumeric()
3204: Developer Note: fortran interface is not autogenerated as the f90
3205: interface defintion cannot be generated correctly [due to MatFactorInfo]
3207: @*/
3208: PetscErrorCode MatCholeskyFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info)
3209: {
3210: MatFactorInfo tinfo;
3218: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3219: if (!(fact)->ops->choleskyfactornumeric) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s numeric factor Cholesky",((PetscObject)mat)->type_name);
3220: 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);
3221: MatCheckPreallocated(mat,2);
3222: if (!info) {
3223: MatFactorInfoInitialize(&tinfo);
3224: info = &tinfo;
3225: }
3227: PetscLogEventBegin(MAT_CholeskyFactorNumeric,mat,fact,0,0);
3228: (fact->ops->choleskyfactornumeric)(fact,mat,info);
3229: PetscLogEventEnd(MAT_CholeskyFactorNumeric,mat,fact,0,0);
3230: MatViewFromOptions(fact,NULL,"-mat_factor_view");
3231: PetscObjectStateIncrease((PetscObject)fact);
3232: return(0);
3233: }
3235: /* ----------------------------------------------------------------*/
3236: /*@
3237: MatSolve - Solves A x = b, given a factored matrix.
3239: Neighbor-wise Collective on Mat
3241: Input Parameters:
3242: + mat - the factored matrix
3243: - b - the right-hand-side vector
3245: Output Parameter:
3246: . x - the result vector
3248: Notes:
3249: The vectors b and x cannot be the same. I.e., one cannot
3250: call MatSolve(A,x,x).
3252: Notes:
3253: Most users should employ the simplified KSP interface for linear solvers
3254: instead of working directly with matrix algebra routines such as this.
3255: See, e.g., KSPCreate().
3257: Level: developer
3259: .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd()
3260: @*/
3261: PetscErrorCode MatSolve(Mat mat,Vec b,Vec x)
3262: {
3272: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3273: 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);
3274: 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);
3275: 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);
3276: if (!mat->rmap->N && !mat->cmap->N) return(0);
3277: if (!mat->ops->solve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3278: MatCheckPreallocated(mat,1);
3280: PetscLogEventBegin(MAT_Solve,mat,b,x,0);
3281: if (mat->factorerrortype) {
3282: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3283: VecSetInf(x);
3284: } else {
3285: if (!mat->ops->solve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3286: (*mat->ops->solve)(mat,b,x);
3287: }
3288: PetscLogEventEnd(MAT_Solve,mat,b,x,0);
3289: PetscObjectStateIncrease((PetscObject)x);
3290: return(0);
3291: }
3293: static PetscErrorCode MatMatSolve_Basic(Mat A,Mat B,Mat X, PetscBool trans)
3294: {
3296: Vec b,x;
3297: PetscInt m,N,i;
3298: PetscScalar *bb,*xx;
3301: MatDenseGetArray(B,&bb);
3302: MatDenseGetArray(X,&xx);
3303: MatGetLocalSize(B,&m,NULL); /* number local rows */
3304: MatGetSize(B,NULL,&N); /* total columns in dense matrix */
3305: MatCreateVecs(A,&x,&b);
3306: for (i=0; i<N; i++) {
3307: VecPlaceArray(b,bb + i*m);
3308: VecPlaceArray(x,xx + i*m);
3309: if (trans) {
3310: MatSolveTranspose(A,b,x);
3311: } else {
3312: MatSolve(A,b,x);
3313: }
3314: VecResetArray(x);
3315: VecResetArray(b);
3316: }
3317: VecDestroy(&b);
3318: VecDestroy(&x);
3319: MatDenseRestoreArray(B,&bb);
3320: MatDenseRestoreArray(X,&xx);
3321: return(0);
3322: }
3324: /*@
3325: MatMatSolve - Solves A X = B, given a factored matrix.
3327: Neighbor-wise Collective on Mat
3329: Input Parameters:
3330: + A - the factored matrix
3331: - B - the right-hand-side matrix (dense matrix)
3333: Output Parameter:
3334: . X - the result matrix (dense matrix)
3336: Notes:
3337: The matrices b and x cannot be the same. I.e., one cannot
3338: call MatMatSolve(A,x,x).
3340: Notes:
3341: Most users should usually employ the simplified KSP interface for linear solvers
3342: instead of working directly with matrix algebra routines such as this.
3343: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3344: at a time.
3346: When using SuperLU_Dist as a parallel solver PETSc will use the SuperLU_Dist functionality to solve multiple right hand sides simultaneously. For MUMPS
3347: it calls a separate solve for each right hand side since MUMPS does not yet support distributed right hand sides.
3349: Since the resulting matrix X must always be dense we do not support sparse representation of the matrix B.
3351: Level: developer
3353: .seealso: MatMatSolveTranspose(), MatLUFactor(), MatCholeskyFactor()
3354: @*/
3355: PetscErrorCode MatMatSolve(Mat A,Mat B,Mat X)
3356: {
3366: if (X == B) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_IDN,"X and B must be different matrices");
3367: 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);
3368: 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);
3369: 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");
3370: if (!A->rmap->N && !A->cmap->N) return(0);
3371: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3372: MatCheckPreallocated(A,1);
3374: PetscLogEventBegin(MAT_MatSolve,A,B,X,0);
3375: if (!A->ops->matsolve) {
3376: PetscInfo1(A,"Mat type %s using basic MatMatSolve\n",((PetscObject)A)->type_name);
3377: MatMatSolve_Basic(A,B,X,PETSC_FALSE);
3378: } else {
3379: (*A->ops->matsolve)(A,B,X);
3380: }
3381: PetscLogEventEnd(MAT_MatSolve,A,B,X,0);
3382: PetscObjectStateIncrease((PetscObject)X);
3383: return(0);
3384: }
3386: /*@
3387: MatMatSolveTranspose - Solves A^T X = B, given a factored matrix.
3389: Neighbor-wise Collective on Mat
3391: Input Parameters:
3392: + A - the factored matrix
3393: - B - the right-hand-side matrix (dense matrix)
3395: Output Parameter:
3396: . X - the result matrix (dense matrix)
3398: Notes:
3399: The matrices B and X cannot be the same. I.e., one cannot
3400: call MatMatSolveTranspose(A,X,X).
3402: Notes:
3403: Most users should usually employ the simplified KSP interface for linear solvers
3404: instead of working directly with matrix algebra routines such as this.
3405: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3406: at a time.
3408: When using SuperLU_Dist or MUMPS as a parallel solver, PETSc will use their functionality to solve multiple right hand sides simultaneously.
3410: Level: developer
3412: .seealso: MatMatSolve(), MatLUFactor(), MatCholeskyFactor()
3413: @*/
3414: PetscErrorCode MatMatSolveTranspose(Mat A,Mat B,Mat X)
3415: {
3425: if (X == B) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_IDN,"X and B must be different matrices");
3426: 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);
3427: 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);
3428: 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);
3429: 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");
3430: if (!A->rmap->N && !A->cmap->N) return(0);
3431: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3432: MatCheckPreallocated(A,1);
3434: PetscLogEventBegin(MAT_MatSolve,A,B,X,0);
3435: if (!A->ops->matsolvetranspose) {
3436: PetscInfo1(A,"Mat type %s using basic MatMatSolveTranspose\n",((PetscObject)A)->type_name);
3437: MatMatSolve_Basic(A,B,X,PETSC_TRUE);
3438: } else {
3439: (*A->ops->matsolvetranspose)(A,B,X);
3440: }
3441: PetscLogEventEnd(MAT_MatSolve,A,B,X,0);
3442: PetscObjectStateIncrease((PetscObject)X);
3443: return(0);
3444: }
3446: /*@
3447: MatMatTransposeSolve - Solves A X = B^T, given a factored matrix.
3449: Neighbor-wise Collective on Mat
3451: Input Parameters:
3452: + A - the factored matrix
3453: - Bt - the transpose of right-hand-side matrix
3455: Output Parameter:
3456: . X - the result matrix (dense matrix)
3458: Notes:
3459: Most users should usually employ the simplified KSP interface for linear solvers
3460: instead of working directly with matrix algebra routines such as this.
3461: See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X)
3462: at a time.
3464: 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().
3466: Level: developer
3468: .seealso: MatMatSolve(), MatMatSolveTranspose(), MatLUFactor(), MatCholeskyFactor()
3469: @*/
3470: PetscErrorCode MatMatTransposeSolve(Mat A,Mat Bt,Mat X)
3471: {
3482: if (X == Bt) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_IDN,"X and B must be different matrices");
3483: 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);
3484: 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);
3485: 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");
3486: if (!A->rmap->N && !A->cmap->N) return(0);
3487: if (!A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3488: MatCheckPreallocated(A,1);
3490: if (!A->ops->mattransposesolve) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)A)->type_name);
3491: PetscLogEventBegin(MAT_MatTrSolve,A,Bt,X,0);
3492: (*A->ops->mattransposesolve)(A,Bt,X);
3493: PetscLogEventEnd(MAT_MatTrSolve,A,Bt,X,0);
3494: PetscObjectStateIncrease((PetscObject)X);
3495: return(0);
3496: }
3498: /*@
3499: MatForwardSolve - Solves L x = b, given a factored matrix, A = LU, or
3500: U^T*D^(1/2) x = b, given a factored symmetric matrix, A = U^T*D*U,
3502: Neighbor-wise Collective on Mat
3504: Input Parameters:
3505: + mat - the factored matrix
3506: - b - the right-hand-side vector
3508: Output Parameter:
3509: . x - the result vector
3511: Notes:
3512: MatSolve() should be used for most applications, as it performs
3513: a forward solve followed by a backward solve.
3515: The vectors b and x cannot be the same, i.e., one cannot
3516: call MatForwardSolve(A,x,x).
3518: For matrix in seqsbaij format with block size larger than 1,
3519: the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet.
3520: MatForwardSolve() solves U^T*D y = b, and
3521: MatBackwardSolve() solves U x = y.
3522: Thus they do not provide a symmetric preconditioner.
3524: Most users should employ the simplified KSP interface for linear solvers
3525: instead of working directly with matrix algebra routines such as this.
3526: See, e.g., KSPCreate().
3528: Level: developer
3530: .seealso: MatSolve(), MatBackwardSolve()
3531: @*/
3532: PetscErrorCode MatForwardSolve(Mat mat,Vec b,Vec x)
3533: {
3543: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3544: 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);
3545: 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);
3546: 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);
3547: if (!mat->rmap->N && !mat->cmap->N) return(0);
3548: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3549: MatCheckPreallocated(mat,1);
3551: if (!mat->ops->forwardsolve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3552: PetscLogEventBegin(MAT_ForwardSolve,mat,b,x,0);
3553: (*mat->ops->forwardsolve)(mat,b,x);
3554: PetscLogEventEnd(MAT_ForwardSolve,mat,b,x,0);
3555: PetscObjectStateIncrease((PetscObject)x);
3556: return(0);
3557: }
3559: /*@
3560: MatBackwardSolve - Solves U x = b, given a factored matrix, A = LU.
3561: D^(1/2) U x = b, given a factored symmetric matrix, A = U^T*D*U,
3563: Neighbor-wise Collective on Mat
3565: Input Parameters:
3566: + mat - the factored matrix
3567: - b - the right-hand-side vector
3569: Output Parameter:
3570: . x - the result vector
3572: Notes:
3573: MatSolve() should be used for most applications, as it performs
3574: a forward solve followed by a backward solve.
3576: The vectors b and x cannot be the same. I.e., one cannot
3577: call MatBackwardSolve(A,x,x).
3579: For matrix in seqsbaij format with block size larger than 1,
3580: the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet.
3581: MatForwardSolve() solves U^T*D y = b, and
3582: MatBackwardSolve() solves U x = y.
3583: Thus they do not provide a symmetric preconditioner.
3585: Most users should employ the simplified KSP interface for linear solvers
3586: instead of working directly with matrix algebra routines such as this.
3587: See, e.g., KSPCreate().
3589: Level: developer
3591: .seealso: MatSolve(), MatForwardSolve()
3592: @*/
3593: PetscErrorCode MatBackwardSolve(Mat mat,Vec b,Vec x)
3594: {
3604: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3605: 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);
3606: 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);
3607: 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);
3608: if (!mat->rmap->N && !mat->cmap->N) return(0);
3609: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3610: MatCheckPreallocated(mat,1);
3612: if (!mat->ops->backwardsolve) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3613: PetscLogEventBegin(MAT_BackwardSolve,mat,b,x,0);
3614: (*mat->ops->backwardsolve)(mat,b,x);
3615: PetscLogEventEnd(MAT_BackwardSolve,mat,b,x,0);
3616: PetscObjectStateIncrease((PetscObject)x);
3617: return(0);
3618: }
3620: /*@
3621: MatSolveAdd - Computes x = y + inv(A)*b, given a factored matrix.
3623: Neighbor-wise Collective on Mat
3625: Input Parameters:
3626: + mat - the factored matrix
3627: . b - the right-hand-side vector
3628: - y - the vector to be added to
3630: Output Parameter:
3631: . x - the result vector
3633: Notes:
3634: The vectors b and x cannot be the same. I.e., one cannot
3635: call MatSolveAdd(A,x,y,x).
3637: Most users should employ the simplified KSP interface for linear solvers
3638: instead of working directly with matrix algebra routines such as this.
3639: See, e.g., KSPCreate().
3641: Level: developer
3643: .seealso: MatSolve(), MatSolveTranspose(), MatSolveTransposeAdd()
3644: @*/
3645: PetscErrorCode MatSolveAdd(Mat mat,Vec b,Vec y,Vec x)
3646: {
3647: PetscScalar one = 1.0;
3648: Vec tmp;
3660: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3661: 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);
3662: 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);
3663: 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);
3664: 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);
3665: 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);
3666: if (!mat->rmap->N && !mat->cmap->N) return(0);
3667: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3668: MatCheckPreallocated(mat,1);
3670: PetscLogEventBegin(MAT_SolveAdd,mat,b,x,y);
3671: if (mat->ops->solveadd) {
3672: (*mat->ops->solveadd)(mat,b,y,x);
3673: } else {
3674: /* do the solve then the add manually */
3675: if (x != y) {
3676: MatSolve(mat,b,x);
3677: VecAXPY(x,one,y);
3678: } else {
3679: VecDuplicate(x,&tmp);
3680: PetscLogObjectParent((PetscObject)mat,(PetscObject)tmp);
3681: VecCopy(x,tmp);
3682: MatSolve(mat,b,x);
3683: VecAXPY(x,one,tmp);
3684: VecDestroy(&tmp);
3685: }
3686: }
3687: PetscLogEventEnd(MAT_SolveAdd,mat,b,x,y);
3688: PetscObjectStateIncrease((PetscObject)x);
3689: return(0);
3690: }
3692: /*@
3693: MatSolveTranspose - Solves A' x = b, given a factored matrix.
3695: Neighbor-wise Collective on Mat
3697: Input Parameters:
3698: + mat - the factored matrix
3699: - b - the right-hand-side vector
3701: Output Parameter:
3702: . x - the result vector
3704: Notes:
3705: The vectors b and x cannot be the same. I.e., one cannot
3706: call MatSolveTranspose(A,x,x).
3708: Most users should employ the simplified KSP interface for linear solvers
3709: instead of working directly with matrix algebra routines such as this.
3710: See, e.g., KSPCreate().
3712: Level: developer
3714: .seealso: MatSolve(), MatSolveAdd(), MatSolveTransposeAdd()
3715: @*/
3716: PetscErrorCode MatSolveTranspose(Mat mat,Vec b,Vec x)
3717: {
3727: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3728: 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);
3729: 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);
3730: if (!mat->rmap->N && !mat->cmap->N) return(0);
3731: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3732: MatCheckPreallocated(mat,1);
3733: PetscLogEventBegin(MAT_SolveTranspose,mat,b,x,0);
3734: if (mat->factorerrortype) {
3735: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3736: VecSetInf(x);
3737: } else {
3738: if (!mat->ops->solvetranspose) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s",((PetscObject)mat)->type_name);
3739: (*mat->ops->solvetranspose)(mat,b,x);
3740: }
3741: PetscLogEventEnd(MAT_SolveTranspose,mat,b,x,0);
3742: PetscObjectStateIncrease((PetscObject)x);
3743: return(0);
3744: }
3746: /*@
3747: MatSolveTransposeAdd - Computes x = y + inv(Transpose(A)) b, given a
3748: factored matrix.
3750: Neighbor-wise Collective on Mat
3752: Input Parameters:
3753: + mat - the factored matrix
3754: . b - the right-hand-side vector
3755: - y - the vector to be added to
3757: Output Parameter:
3758: . x - the result vector
3760: Notes:
3761: The vectors b and x cannot be the same. I.e., one cannot
3762: call MatSolveTransposeAdd(A,x,y,x).
3764: Most users should employ the simplified KSP interface for linear solvers
3765: instead of working directly with matrix algebra routines such as this.
3766: See, e.g., KSPCreate().
3768: Level: developer
3770: .seealso: MatSolve(), MatSolveAdd(), MatSolveTranspose()
3771: @*/
3772: PetscErrorCode MatSolveTransposeAdd(Mat mat,Vec b,Vec y,Vec x)
3773: {
3774: PetscScalar one = 1.0;
3776: Vec tmp;
3787: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
3788: 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);
3789: 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);
3790: 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);
3791: 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);
3792: if (!mat->rmap->N && !mat->cmap->N) return(0);
3793: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
3794: MatCheckPreallocated(mat,1);
3796: PetscLogEventBegin(MAT_SolveTransposeAdd,mat,b,x,y);
3797: if (mat->ops->solvetransposeadd) {
3798: if (mat->factorerrortype) {
3799: PetscInfo1(mat,"MatFactorError %D\n",mat->factorerrortype);
3800: VecSetInf(x);
3801: } else {
3802: (*mat->ops->solvetransposeadd)(mat,b,y,x);
3803: }
3804: } else {
3805: /* do the solve then the add manually */
3806: if (x != y) {
3807: MatSolveTranspose(mat,b,x);
3808: VecAXPY(x,one,y);
3809: } else {
3810: VecDuplicate(x,&tmp);
3811: PetscLogObjectParent((PetscObject)mat,(PetscObject)tmp);
3812: VecCopy(x,tmp);
3813: MatSolveTranspose(mat,b,x);
3814: VecAXPY(x,one,tmp);
3815: VecDestroy(&tmp);
3816: }
3817: }
3818: PetscLogEventEnd(MAT_SolveTransposeAdd,mat,b,x,y);
3819: PetscObjectStateIncrease((PetscObject)x);
3820: return(0);
3821: }
3822: /* ----------------------------------------------------------------*/
3824: /*@
3825: MatSOR - Computes relaxation (SOR, Gauss-Seidel) sweeps.
3827: Neighbor-wise Collective on Mat
3829: Input Parameters:
3830: + mat - the matrix
3831: . b - the right hand side
3832: . omega - the relaxation factor
3833: . flag - flag indicating the type of SOR (see below)
3834: . shift - diagonal shift
3835: . its - the number of iterations
3836: - lits - the number of local iterations
3838: Output Parameters:
3839: . x - the solution (can contain an initial guess, use option SOR_ZERO_INITIAL_GUESS to indicate no guess)
3841: SOR Flags:
3842: + SOR_FORWARD_SWEEP - forward SOR
3843: . SOR_BACKWARD_SWEEP - backward SOR
3844: . SOR_SYMMETRIC_SWEEP - SSOR (symmetric SOR)
3845: . SOR_LOCAL_FORWARD_SWEEP - local forward SOR
3846: . SOR_LOCAL_BACKWARD_SWEEP - local forward SOR
3847: . SOR_LOCAL_SYMMETRIC_SWEEP - local SSOR
3848: . SOR_APPLY_UPPER, SOR_APPLY_LOWER - applies
3849: upper/lower triangular part of matrix to
3850: vector (with omega)
3851: - SOR_ZERO_INITIAL_GUESS - zero initial guess
3853: Notes:
3854: SOR_LOCAL_FORWARD_SWEEP, SOR_LOCAL_BACKWARD_SWEEP, and
3855: SOR_LOCAL_SYMMETRIC_SWEEP perform separate independent smoothings
3856: on each processor.
3858: Application programmers will not generally use MatSOR() directly,
3859: but instead will employ the KSP/PC interface.
3861: Notes:
3862: for BAIJ, SBAIJ, and AIJ matrices with Inodes this does a block SOR smoothing, otherwise it does a pointwise smoothing
3864: Notes for Advanced Users:
3865: The flags are implemented as bitwise inclusive or operations.
3866: For example, use (SOR_ZERO_INITIAL_GUESS | SOR_SYMMETRIC_SWEEP)
3867: to specify a zero initial guess for SSOR.
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: Vectors x and b CANNOT be the same
3875: Developer Note: We should add block SOR support for AIJ matrices with block size set to great than one and no inodes
3877: Level: developer
3879: @*/
3880: PetscErrorCode MatSOR(Mat mat,Vec b,PetscReal omega,MatSORType flag,PetscReal shift,PetscInt its,PetscInt lits,Vec x)
3881: {
3891: if (!mat->ops->sor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
3892: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3893: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3894: 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);
3895: 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);
3896: 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);
3897: if (its <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D positive",its);
3898: if (lits <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires local its %D positive",lits);
3899: if (b == x) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_IDN,"b and x vector cannot be the same");
3901: MatCheckPreallocated(mat,1);
3902: PetscLogEventBegin(MAT_SOR,mat,b,x,0);
3903: ierr =(*mat->ops->sor)(mat,b,omega,flag,shift,its,lits,x);
3904: PetscLogEventEnd(MAT_SOR,mat,b,x,0);
3905: PetscObjectStateIncrease((PetscObject)x);
3906: return(0);
3907: }
3909: /*
3910: Default matrix copy routine.
3911: */
3912: PetscErrorCode MatCopy_Basic(Mat A,Mat B,MatStructure str)
3913: {
3914: PetscErrorCode ierr;
3915: PetscInt i,rstart = 0,rend = 0,nz;
3916: const PetscInt *cwork;
3917: const PetscScalar *vwork;
3920: if (B->assembled) {
3921: MatZeroEntries(B);
3922: }
3923: if (str == SAME_NONZERO_PATTERN) {
3924: MatGetOwnershipRange(A,&rstart,&rend);
3925: for (i=rstart; i<rend; i++) {
3926: MatGetRow(A,i,&nz,&cwork,&vwork);
3927: MatSetValues(B,1,&i,nz,cwork,vwork,INSERT_VALUES);
3928: MatRestoreRow(A,i,&nz,&cwork,&vwork);
3929: }
3930: } else {
3931: MatAYPX(B,0.0,A,str);
3932: }
3933: MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
3934: MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);
3935: return(0);
3936: }
3938: /*@
3939: MatCopy - Copies a matrix to another matrix.
3941: Collective on Mat
3943: Input Parameters:
3944: + A - the matrix
3945: - str - SAME_NONZERO_PATTERN or DIFFERENT_NONZERO_PATTERN
3947: Output Parameter:
3948: . B - where the copy is put
3950: Notes:
3951: If you use SAME_NONZERO_PATTERN then the two matrices had better have the
3952: same nonzero pattern or the routine will crash.
3954: MatCopy() copies the matrix entries of a matrix to another existing
3955: matrix (after first zeroing the second matrix). A related routine is
3956: MatConvert(), which first creates a new matrix and then copies the data.
3958: Level: intermediate
3960: .seealso: MatConvert(), MatDuplicate()
3962: @*/
3963: PetscErrorCode MatCopy(Mat A,Mat B,MatStructure str)
3964: {
3966: PetscInt i;
3974: MatCheckPreallocated(B,2);
3975: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
3976: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
3977: 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);
3978: MatCheckPreallocated(A,1);
3979: if (A == B) return(0);
3981: PetscLogEventBegin(MAT_Copy,A,B,0,0);
3982: if (A->ops->copy) {
3983: (*A->ops->copy)(A,B,str);
3984: } else { /* generic conversion */
3985: MatCopy_Basic(A,B,str);
3986: }
3988: B->stencil.dim = A->stencil.dim;
3989: B->stencil.noc = A->stencil.noc;
3990: for (i=0; i<=A->stencil.dim; i++) {
3991: B->stencil.dims[i] = A->stencil.dims[i];
3992: B->stencil.starts[i] = A->stencil.starts[i];
3993: }
3995: PetscLogEventEnd(MAT_Copy,A,B,0,0);
3996: PetscObjectStateIncrease((PetscObject)B);
3997: return(0);
3998: }
4000: /*@C
4001: MatConvert - Converts a matrix to another matrix, either of the same
4002: or different type.
4004: Collective on Mat
4006: Input Parameters:
4007: + mat - the matrix
4008: . newtype - new matrix type. Use MATSAME to create a new matrix of the
4009: same type as the original matrix.
4010: - reuse - denotes if the destination matrix is to be created or reused.
4011: 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
4012: 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).
4014: Output Parameter:
4015: . M - pointer to place new matrix
4017: Notes:
4018: MatConvert() first creates a new matrix and then copies the data from
4019: the first matrix. A related routine is MatCopy(), which copies the matrix
4020: entries of one matrix to another already existing matrix context.
4022: Cannot be used to convert a sequential matrix to parallel or parallel to sequential,
4023: the MPI communicator of the generated matrix is always the same as the communicator
4024: of the input matrix.
4026: Level: intermediate
4028: .seealso: MatCopy(), MatDuplicate()
4029: @*/
4030: PetscErrorCode MatConvert(Mat mat, MatType newtype,MatReuse reuse,Mat *M)
4031: {
4033: PetscBool sametype,issame,flg;
4034: char convname[256],mtype[256];
4035: Mat B;
4041: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4042: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4043: MatCheckPreallocated(mat,1);
4045: PetscOptionsGetString(((PetscObject)mat)->options,((PetscObject)mat)->prefix,"-matconvert_type",mtype,256,&flg);
4046: if (flg) {
4047: newtype = mtype;
4048: }
4049: PetscObjectTypeCompare((PetscObject)mat,newtype,&sametype);
4050: PetscStrcmp(newtype,"same",&issame);
4051: if ((reuse == MAT_INPLACE_MATRIX) && (mat != *M)) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"MAT_INPLACE_MATRIX requires same input and output matrix");
4052: 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");
4054: if ((reuse == MAT_INPLACE_MATRIX) && (issame || sametype)) return(0);
4056: if ((sametype || issame) && (reuse==MAT_INITIAL_MATRIX) && mat->ops->duplicate) {
4057: (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);
4058: } else {
4059: PetscErrorCode (*conv)(Mat, MatType,MatReuse,Mat*)=NULL;
4060: const char *prefix[3] = {"seq","mpi",""};
4061: PetscInt i;
4062: /*
4063: Order of precedence:
4064: 0) See if newtype is a superclass of the current matrix.
4065: 1) See if a specialized converter is known to the current matrix.
4066: 2) See if a specialized converter is known to the desired matrix class.
4067: 3) See if a good general converter is registered for the desired class
4068: (as of 6/27/03 only MATMPIADJ falls into this category).
4069: 4) See if a good general converter is known for the current matrix.
4070: 5) Use a really basic converter.
4071: */
4073: /* 0) See if newtype is a superclass of the current matrix.
4074: i.e mat is mpiaij and newtype is aij */
4075: for (i=0; i<2; i++) {
4076: PetscStrncpy(convname,prefix[i],sizeof(convname));
4077: PetscStrlcat(convname,newtype,sizeof(convname));
4078: PetscStrcmp(convname,((PetscObject)mat)->type_name,&flg);
4079: PetscInfo3(mat,"Check superclass %s %s -> %d\n",convname,((PetscObject)mat)->type_name,flg);
4080: if (flg) {
4081: if (reuse == MAT_INPLACE_MATRIX) {
4082: return(0);
4083: } else if (reuse == MAT_INITIAL_MATRIX && mat->ops->duplicate) {
4084: (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);
4085: return(0);
4086: } else if (reuse == MAT_REUSE_MATRIX && mat->ops->copy) {
4087: MatCopy(mat,*M,SAME_NONZERO_PATTERN);
4088: return(0);
4089: }
4090: }
4091: }
4092: /* 1) See if a specialized converter is known to the current matrix and the desired class */
4093: for (i=0; i<3; i++) {
4094: PetscStrncpy(convname,"MatConvert_",sizeof(convname));
4095: PetscStrlcat(convname,((PetscObject)mat)->type_name,sizeof(convname));
4096: PetscStrlcat(convname,"_",sizeof(convname));
4097: PetscStrlcat(convname,prefix[i],sizeof(convname));
4098: PetscStrlcat(convname,issame ? ((PetscObject)mat)->type_name : newtype,sizeof(convname));
4099: PetscStrlcat(convname,"_C",sizeof(convname));
4100: PetscObjectQueryFunction((PetscObject)mat,convname,&conv);
4101: PetscInfo3(mat,"Check specialized (1) %s (%s) -> %d\n",convname,((PetscObject)mat)->type_name,!!conv);
4102: if (conv) goto foundconv;
4103: }
4105: /* 2) See if a specialized converter is known to the desired matrix class. */
4106: MatCreate(PetscObjectComm((PetscObject)mat),&B);
4107: MatSetSizes(B,mat->rmap->n,mat->cmap->n,mat->rmap->N,mat->cmap->N);
4108: MatSetType(B,newtype);
4109: for (i=0; i<3; i++) {
4110: PetscStrncpy(convname,"MatConvert_",sizeof(convname));
4111: PetscStrlcat(convname,((PetscObject)mat)->type_name,sizeof(convname));
4112: PetscStrlcat(convname,"_",sizeof(convname));
4113: PetscStrlcat(convname,prefix[i],sizeof(convname));
4114: PetscStrlcat(convname,newtype,sizeof(convname));
4115: PetscStrlcat(convname,"_C",sizeof(convname));
4116: PetscObjectQueryFunction((PetscObject)B,convname,&conv);
4117: PetscInfo3(mat,"Check specialized (2) %s (%s) -> %d\n",convname,((PetscObject)B)->type_name,!!conv);
4118: if (conv) {
4119: MatDestroy(&B);
4120: goto foundconv;
4121: }
4122: }
4124: /* 3) See if a good general converter is registered for the desired class */
4125: conv = B->ops->convertfrom;
4126: PetscInfo2(mat,"Check convertfrom (%s) -> %d\n",((PetscObject)B)->type_name,!!conv);
4127: MatDestroy(&B);
4128: if (conv) goto foundconv;
4130: /* 4) See if a good general converter is known for the current matrix */
4131: if (mat->ops->convert) {
4132: conv = mat->ops->convert;
4133: }
4134: PetscInfo2(mat,"Check general convert (%s) -> %d\n",((PetscObject)mat)->type_name,!!conv);
4135: if (conv) goto foundconv;
4137: /* 5) Use a really basic converter. */
4138: PetscInfo(mat,"Using MatConvert_Basic\n");
4139: conv = MatConvert_Basic;
4141: foundconv:
4142: PetscLogEventBegin(MAT_Convert,mat,0,0,0);
4143: (*conv)(mat,newtype,reuse,M);
4144: if (mat->rmap->mapping && mat->cmap->mapping && !(*M)->rmap->mapping && !(*M)->cmap->mapping) {
4145: /* the block sizes must be same if the mappings are copied over */
4146: (*M)->rmap->bs = mat->rmap->bs;
4147: (*M)->cmap->bs = mat->cmap->bs;
4148: PetscObjectReference((PetscObject)mat->rmap->mapping);
4149: PetscObjectReference((PetscObject)mat->cmap->mapping);
4150: (*M)->rmap->mapping = mat->rmap->mapping;
4151: (*M)->cmap->mapping = mat->cmap->mapping;
4152: }
4153: (*M)->stencil.dim = mat->stencil.dim;
4154: (*M)->stencil.noc = mat->stencil.noc;
4155: for (i=0; i<=mat->stencil.dim; i++) {
4156: (*M)->stencil.dims[i] = mat->stencil.dims[i];
4157: (*M)->stencil.starts[i] = mat->stencil.starts[i];
4158: }
4159: PetscLogEventEnd(MAT_Convert,mat,0,0,0);
4160: }
4161: PetscObjectStateIncrease((PetscObject)*M);
4163: /* Copy Mat options */
4164: if (mat->symmetric) {MatSetOption(*M,MAT_SYMMETRIC,PETSC_TRUE);}
4165: if (mat->hermitian) {MatSetOption(*M,MAT_HERMITIAN,PETSC_TRUE);}
4166: return(0);
4167: }
4169: /*@C
4170: MatFactorGetSolverType - Returns name of the package providing the factorization routines
4172: Not Collective
4174: Input Parameter:
4175: . mat - the matrix, must be a factored matrix
4177: Output Parameter:
4178: . type - the string name of the package (do not free this string)
4180: Notes:
4181: In Fortran you pass in a empty string and the package name will be copied into it.
4182: (Make sure the string is long enough)
4184: Level: intermediate
4186: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatGetFactor()
4187: @*/
4188: PetscErrorCode MatFactorGetSolverType(Mat mat, MatSolverType *type)
4189: {
4190: PetscErrorCode ierr, (*conv)(Mat,MatSolverType*);
4195: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
4196: PetscObjectQueryFunction((PetscObject)mat,"MatFactorGetSolverType_C",&conv);
4197: if (!conv) {
4198: *type = MATSOLVERPETSC;
4199: } else {
4200: (*conv)(mat,type);
4201: }
4202: return(0);
4203: }
4205: typedef struct _MatSolverTypeForSpecifcType* MatSolverTypeForSpecifcType;
4206: struct _MatSolverTypeForSpecifcType {
4207: MatType mtype;
4208: PetscErrorCode (*getfactor[4])(Mat,MatFactorType,Mat*);
4209: MatSolverTypeForSpecifcType next;
4210: };
4212: typedef struct _MatSolverTypeHolder* MatSolverTypeHolder;
4213: struct _MatSolverTypeHolder {
4214: char *name;
4215: MatSolverTypeForSpecifcType handlers;
4216: MatSolverTypeHolder next;
4217: };
4219: static MatSolverTypeHolder MatSolverTypeHolders = NULL;
4221: /*@C
4222: MatSolvePackageRegister - Registers a MatSolverType that works for a particular matrix type
4224: Input Parameters:
4225: + package - name of the package, for example petsc or superlu
4226: . mtype - the matrix type that works with this package
4227: . ftype - the type of factorization supported by the package
4228: - getfactor - routine that will create the factored matrix ready to be used
4230: Level: intermediate
4232: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable()
4233: @*/
4234: PetscErrorCode MatSolverTypeRegister(MatSolverType package,MatType mtype,MatFactorType ftype,PetscErrorCode (*getfactor)(Mat,MatFactorType,Mat*))
4235: {
4236: PetscErrorCode ierr;
4237: MatSolverTypeHolder next = MatSolverTypeHolders,prev = NULL;
4238: PetscBool flg;
4239: MatSolverTypeForSpecifcType inext,iprev = NULL;
4242: MatInitializePackage();
4243: if (!next) {
4244: PetscNew(&MatSolverTypeHolders);
4245: PetscStrallocpy(package,&MatSolverTypeHolders->name);
4246: PetscNew(&MatSolverTypeHolders->handlers);
4247: PetscStrallocpy(mtype,(char **)&MatSolverTypeHolders->handlers->mtype);
4248: MatSolverTypeHolders->handlers->getfactor[(int)ftype-1] = getfactor;
4249: return(0);
4250: }
4251: while (next) {
4252: PetscStrcasecmp(package,next->name,&flg);
4253: if (flg) {
4254: if (!next->handlers) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"MatSolverTypeHolder is missing handlers");
4255: inext = next->handlers;
4256: while (inext) {
4257: PetscStrcasecmp(mtype,inext->mtype,&flg);
4258: if (flg) {
4259: inext->getfactor[(int)ftype-1] = getfactor;
4260: return(0);
4261: }
4262: iprev = inext;
4263: inext = inext->next;
4264: }
4265: PetscNew(&iprev->next);
4266: PetscStrallocpy(mtype,(char **)&iprev->next->mtype);
4267: iprev->next->getfactor[(int)ftype-1] = getfactor;
4268: return(0);
4269: }
4270: prev = next;
4271: next = next->next;
4272: }
4273: PetscNew(&prev->next);
4274: PetscStrallocpy(package,&prev->next->name);
4275: PetscNew(&prev->next->handlers);
4276: PetscStrallocpy(mtype,(char **)&prev->next->handlers->mtype);
4277: prev->next->handlers->getfactor[(int)ftype-1] = getfactor;
4278: return(0);
4279: }
4281: /*@C
4282: MatSolvePackageGet - Get's the function that creates the factor matrix if it exist
4284: Input Parameters:
4285: + package - name of the package, for example petsc or superlu
4286: . ftype - the type of factorization supported by the package
4287: - mtype - the matrix type that works with this package
4289: Output Parameters:
4290: + foundpackage - PETSC_TRUE if the package was registered
4291: . foundmtype - PETSC_TRUE if the package supports the requested mtype
4292: - getfactor - routine that will create the factored matrix ready to be used or NULL if not found
4294: Level: intermediate
4296: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable()
4297: @*/
4298: PetscErrorCode MatSolverTypeGet(MatSolverType package,MatType mtype,MatFactorType ftype,PetscBool *foundpackage,PetscBool *foundmtype,PetscErrorCode (**getfactor)(Mat,MatFactorType,Mat*))
4299: {
4300: PetscErrorCode ierr;
4301: MatSolverTypeHolder next = MatSolverTypeHolders;
4302: PetscBool flg;
4303: MatSolverTypeForSpecifcType inext;
4306: if (foundpackage) *foundpackage = PETSC_FALSE;
4307: if (foundmtype) *foundmtype = PETSC_FALSE;
4308: if (getfactor) *getfactor = NULL;
4310: if (package) {
4311: while (next) {
4312: PetscStrcasecmp(package,next->name,&flg);
4313: if (flg) {
4314: if (foundpackage) *foundpackage = PETSC_TRUE;
4315: inext = next->handlers;
4316: while (inext) {
4317: PetscStrbeginswith(mtype,inext->mtype,&flg);
4318: if (flg) {
4319: if (foundmtype) *foundmtype = PETSC_TRUE;
4320: if (getfactor) *getfactor = inext->getfactor[(int)ftype-1];
4321: return(0);
4322: }
4323: inext = inext->next;
4324: }
4325: }
4326: next = next->next;
4327: }
4328: } else {
4329: while (next) {
4330: inext = next->handlers;
4331: while (inext) {
4332: PetscStrbeginswith(mtype,inext->mtype,&flg);
4333: if (flg && inext->getfactor[(int)ftype-1]) {
4334: if (foundpackage) *foundpackage = PETSC_TRUE;
4335: if (foundmtype) *foundmtype = PETSC_TRUE;
4336: if (getfactor) *getfactor = inext->getfactor[(int)ftype-1];
4337: return(0);
4338: }
4339: inext = inext->next;
4340: }
4341: next = next->next;
4342: }
4343: }
4344: return(0);
4345: }
4347: PetscErrorCode MatSolverTypeDestroy(void)
4348: {
4349: PetscErrorCode ierr;
4350: MatSolverTypeHolder next = MatSolverTypeHolders,prev;
4351: MatSolverTypeForSpecifcType inext,iprev;
4354: while (next) {
4355: PetscFree(next->name);
4356: inext = next->handlers;
4357: while (inext) {
4358: PetscFree(inext->mtype);
4359: iprev = inext;
4360: inext = inext->next;
4361: PetscFree(iprev);
4362: }
4363: prev = next;
4364: next = next->next;
4365: PetscFree(prev);
4366: }
4367: MatSolverTypeHolders = NULL;
4368: return(0);
4369: }
4371: /*@C
4372: MatGetFactor - Returns a matrix suitable to calls to MatXXFactorSymbolic()
4374: Collective on Mat
4376: Input Parameters:
4377: + mat - the matrix
4378: . type - name of solver type, for example, superlu, petsc (to use PETSc's default)
4379: - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU,
4381: Output Parameters:
4382: . f - the factor matrix used with MatXXFactorSymbolic() calls
4384: Notes:
4385: Some PETSc matrix formats have alternative solvers available that are contained in alternative packages
4386: such as pastix, superlu, mumps etc.
4388: PETSc must have been ./configure to use the external solver, using the option --download-package
4390: Level: intermediate
4392: .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable()
4393: @*/
4394: PetscErrorCode MatGetFactor(Mat mat, MatSolverType type,MatFactorType ftype,Mat *f)
4395: {
4396: PetscErrorCode ierr,(*conv)(Mat,MatFactorType,Mat*);
4397: PetscBool foundpackage,foundmtype;
4403: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4404: MatCheckPreallocated(mat,1);
4406: MatSolverTypeGet(type,((PetscObject)mat)->type_name,ftype,&foundpackage,&foundmtype,&conv);
4407: if (!foundpackage) {
4408: if (type) {
4409: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"Could not locate solver package %s. Perhaps you must ./configure with --download-%s",type,type);
4410: } else {
4411: SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"Could not locate a solver package. Perhaps you must ./configure with --download-<package>");
4412: }
4413: }
4415: if (!foundmtype) SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_MISSING_FACTOR,"MatSolverType %s does not support matrix type %s",type,((PetscObject)mat)->type_name);
4416: 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);
4418: #if defined(PETSC_USE_COMPLEX)
4419: if (mat->hermitian && !mat->symmetric && (ftype == MAT_FACTOR_CHOLESKY||ftype == MAT_FACTOR_ICC)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Hermitian CHOLESKY or ICC Factor is not supported");
4420: #endif
4422: (*conv)(mat,ftype,f);
4423: return(0);
4424: }
4426: /*@C
4427: MatGetFactorAvailable - Returns a a flag if matrix supports particular package and factor type
4429: Not Collective
4431: Input Parameters:
4432: + mat - the matrix
4433: . type - name of solver type, for example, superlu, petsc (to use PETSc's default)
4434: - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU,
4436: Output Parameter:
4437: . flg - PETSC_TRUE if the factorization is available
4439: Notes:
4440: Some PETSc matrix formats have alternative solvers available that are contained in alternative packages
4441: such as pastix, superlu, mumps etc.
4443: PETSc must have been ./configure to use the external solver, using the option --download-package
4445: Level: intermediate
4447: .seealso: MatCopy(), MatDuplicate(), MatGetFactor()
4448: @*/
4449: PetscErrorCode MatGetFactorAvailable(Mat mat, MatSolverType type,MatFactorType ftype,PetscBool *flg)
4450: {
4451: PetscErrorCode ierr, (*gconv)(Mat,MatFactorType,Mat*);
4457: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4458: MatCheckPreallocated(mat,1);
4460: *flg = PETSC_FALSE;
4461: MatSolverTypeGet(type,((PetscObject)mat)->type_name,ftype,NULL,NULL,&gconv);
4462: if (gconv) {
4463: *flg = PETSC_TRUE;
4464: }
4465: return(0);
4466: }
4468: #include <petscdmtypes.h>
4470: /*@
4471: MatDuplicate - Duplicates a matrix including the non-zero structure.
4473: Collective on Mat
4475: Input Parameters:
4476: + mat - the matrix
4477: - op - One of MAT_DO_NOT_COPY_VALUES, MAT_COPY_VALUES, or MAT_SHARE_NONZERO_PATTERN.
4478: See the manual page for MatDuplicateOption for an explanation of these options.
4480: Output Parameter:
4481: . M - pointer to place new matrix
4483: Level: intermediate
4485: Notes:
4486: You cannot change the nonzero pattern for the parent or child matrix if you use MAT_SHARE_NONZERO_PATTERN.
4487: 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.
4489: .seealso: MatCopy(), MatConvert(), MatDuplicateOption
4490: @*/
4491: PetscErrorCode MatDuplicate(Mat mat,MatDuplicateOption op,Mat *M)
4492: {
4494: Mat B;
4495: PetscInt i;
4496: DM dm;
4497: void (*viewf)(void);
4503: if (op == MAT_COPY_VALUES && !mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"MAT_COPY_VALUES not allowed for unassembled matrix");
4504: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4505: MatCheckPreallocated(mat,1);
4507: *M = 0;
4508: if (!mat->ops->duplicate) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not written for this matrix type");
4509: PetscLogEventBegin(MAT_Convert,mat,0,0,0);
4510: (*mat->ops->duplicate)(mat,op,M);
4511: B = *M;
4513: MatGetOperation(mat,MATOP_VIEW,&viewf);
4514: if (viewf) {
4515: MatSetOperation(B,MATOP_VIEW,viewf);
4516: }
4518: B->stencil.dim = mat->stencil.dim;
4519: B->stencil.noc = mat->stencil.noc;
4520: for (i=0; i<=mat->stencil.dim; i++) {
4521: B->stencil.dims[i] = mat->stencil.dims[i];
4522: B->stencil.starts[i] = mat->stencil.starts[i];
4523: }
4525: B->nooffproczerorows = mat->nooffproczerorows;
4526: B->nooffprocentries = mat->nooffprocentries;
4528: PetscObjectQuery((PetscObject) mat, "__PETSc_dm", (PetscObject*) &dm);
4529: if (dm) {
4530: PetscObjectCompose((PetscObject) B, "__PETSc_dm", (PetscObject) dm);
4531: }
4532: PetscLogEventEnd(MAT_Convert,mat,0,0,0);
4533: PetscObjectStateIncrease((PetscObject)B);
4534: return(0);
4535: }
4537: /*@
4538: MatGetDiagonal - Gets the diagonal of a matrix.
4540: Logically Collective on Mat
4542: Input Parameters:
4543: + mat - the matrix
4544: - v - the vector for storing the diagonal
4546: Output Parameter:
4547: . v - the diagonal of the matrix
4549: Level: intermediate
4551: Note:
4552: Currently only correct in parallel for square matrices.
4554: .seealso: MatGetRow(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs()
4555: @*/
4556: PetscErrorCode MatGetDiagonal(Mat mat,Vec v)
4557: {
4564: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4565: if (!mat->ops->getdiagonal) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4566: MatCheckPreallocated(mat,1);
4568: (*mat->ops->getdiagonal)(mat,v);
4569: PetscObjectStateIncrease((PetscObject)v);
4570: return(0);
4571: }
4573: /*@C
4574: MatGetRowMin - Gets the minimum value (of the real part) of each
4575: row of the matrix
4577: Logically Collective on Mat
4579: Input Parameters:
4580: . mat - the matrix
4582: Output Parameter:
4583: + v - the vector for storing the maximums
4584: - idx - the indices of the column found for each row (optional)
4586: Level: intermediate
4588: Notes:
4589: The result of this call are the same as if one converted the matrix to dense format
4590: and found the minimum value in each row (i.e. the implicit zeros are counted as zeros).
4592: This code is only implemented for a couple of matrix formats.
4594: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs(),
4595: MatGetRowMax()
4596: @*/
4597: PetscErrorCode MatGetRowMin(Mat mat,Vec v,PetscInt idx[])
4598: {
4605: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4606: if (!mat->ops->getrowmax) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4607: MatCheckPreallocated(mat,1);
4609: (*mat->ops->getrowmin)(mat,v,idx);
4610: PetscObjectStateIncrease((PetscObject)v);
4611: return(0);
4612: }
4614: /*@C
4615: MatGetRowMinAbs - Gets the minimum value (in absolute value) of each
4616: row of the matrix
4618: Logically Collective on Mat
4620: Input Parameters:
4621: . mat - the matrix
4623: Output Parameter:
4624: + v - the vector for storing the minimums
4625: - idx - the indices of the column found for each row (or NULL if not needed)
4627: Level: intermediate
4629: Notes:
4630: if a row is completely empty or has only 0.0 values then the idx[] value for that
4631: row is 0 (the first column).
4633: This code is only implemented for a couple of matrix formats.
4635: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMaxAbs(), MatGetRowMin()
4636: @*/
4637: PetscErrorCode MatGetRowMinAbs(Mat mat,Vec v,PetscInt idx[])
4638: {
4645: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4646: if (!mat->ops->getrowminabs) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4647: MatCheckPreallocated(mat,1);
4648: if (idx) {PetscArrayzero(idx,mat->rmap->n);}
4650: (*mat->ops->getrowminabs)(mat,v,idx);
4651: PetscObjectStateIncrease((PetscObject)v);
4652: return(0);
4653: }
4655: /*@C
4656: MatGetRowMax - Gets the maximum value (of the real part) of each
4657: row of the matrix
4659: Logically Collective on Mat
4661: Input Parameters:
4662: . mat - the matrix
4664: Output Parameter:
4665: + v - the vector for storing the maximums
4666: - idx - the indices of the column found for each row (optional)
4668: Level: intermediate
4670: Notes:
4671: The result of this call are the same as if one converted the matrix to dense format
4672: and found the minimum value in each row (i.e. the implicit zeros are counted as zeros).
4674: This code is only implemented for a couple of matrix formats.
4676: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMaxAbs(), MatGetRowMin()
4677: @*/
4678: PetscErrorCode MatGetRowMax(Mat mat,Vec v,PetscInt idx[])
4679: {
4686: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4687: if (!mat->ops->getrowmax) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4688: MatCheckPreallocated(mat,1);
4690: (*mat->ops->getrowmax)(mat,v,idx);
4691: PetscObjectStateIncrease((PetscObject)v);
4692: return(0);
4693: }
4695: /*@C
4696: MatGetRowMaxAbs - Gets the maximum value (in absolute value) of each
4697: row of the matrix
4699: Logically Collective on Mat
4701: Input Parameters:
4702: . mat - the matrix
4704: Output Parameter:
4705: + v - the vector for storing the maximums
4706: - idx - the indices of the column found for each row (or NULL if not needed)
4708: Level: intermediate
4710: Notes:
4711: if a row is completely empty or has only 0.0 values then the idx[] value for that
4712: row is 0 (the first column).
4714: This code is only implemented for a couple of matrix formats.
4716: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMin()
4717: @*/
4718: PetscErrorCode MatGetRowMaxAbs(Mat mat,Vec v,PetscInt idx[])
4719: {
4726: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4727: if (!mat->ops->getrowmaxabs) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4728: MatCheckPreallocated(mat,1);
4729: if (idx) {PetscArrayzero(idx,mat->rmap->n);}
4731: (*mat->ops->getrowmaxabs)(mat,v,idx);
4732: PetscObjectStateIncrease((PetscObject)v);
4733: return(0);
4734: }
4736: /*@
4737: MatGetRowSum - Gets the sum of each row of the matrix
4739: Logically or Neighborhood Collective on Mat
4741: Input Parameters:
4742: . mat - the matrix
4744: Output Parameter:
4745: . v - the vector for storing the sum of rows
4747: Level: intermediate
4749: Notes:
4750: This code is slow since it is not currently specialized for different formats
4752: .seealso: MatGetDiagonal(), MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRowMax(), MatGetRowMin()
4753: @*/
4754: PetscErrorCode MatGetRowSum(Mat mat, Vec v)
4755: {
4756: Vec ones;
4763: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4764: MatCheckPreallocated(mat,1);
4765: MatCreateVecs(mat,&ones,NULL);
4766: VecSet(ones,1.);
4767: MatMult(mat,ones,v);
4768: VecDestroy(&ones);
4769: return(0);
4770: }
4772: /*@
4773: MatTranspose - Computes an in-place or out-of-place transpose of a matrix.
4775: Collective on Mat
4777: Input Parameter:
4778: + mat - the matrix to transpose
4779: - reuse - either MAT_INITIAL_MATRIX, MAT_REUSE_MATRIX, or MAT_INPLACE_MATRIX
4781: Output Parameters:
4782: . B - the transpose
4784: Notes:
4785: If you use MAT_INPLACE_MATRIX then you must pass in &mat for B
4787: MAT_REUSE_MATRIX causes the B matrix from a previous call to this function with MAT_INITIAL_MATRIX to be used
4789: Consider using MatCreateTranspose() instead if you only need a matrix that behaves like the transpose, but don't need the storage to be changed.
4791: Level: intermediate
4793: .seealso: MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse
4794: @*/
4795: PetscErrorCode MatTranspose(Mat mat,MatReuse reuse,Mat *B)
4796: {
4802: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4803: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4804: if (!mat->ops->transpose) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
4805: if (reuse == MAT_INPLACE_MATRIX && mat != *B) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"MAT_INPLACE_MATRIX requires last matrix to match first");
4806: if (reuse == MAT_REUSE_MATRIX && mat == *B) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Perhaps you mean MAT_INPLACE_MATRIX");
4807: MatCheckPreallocated(mat,1);
4809: PetscLogEventBegin(MAT_Transpose,mat,0,0,0);
4810: (*mat->ops->transpose)(mat,reuse,B);
4811: PetscLogEventEnd(MAT_Transpose,mat,0,0,0);
4812: if (B) {PetscObjectStateIncrease((PetscObject)*B);}
4813: return(0);
4814: }
4816: /*@
4817: MatIsTranspose - Test whether a matrix is another one's transpose,
4818: or its own, in which case it tests symmetry.
4820: Collective on Mat
4822: Input Parameter:
4823: + A - the matrix to test
4824: - B - the matrix to test against, this can equal the first parameter
4826: Output Parameters:
4827: . flg - the result
4829: Notes:
4830: Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm
4831: has a running time of the order of the number of nonzeros; the parallel
4832: test involves parallel copies of the block-offdiagonal parts of the matrix.
4834: Level: intermediate
4836: .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian()
4837: @*/
4838: PetscErrorCode MatIsTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg)
4839: {
4840: PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool*),(*g)(Mat,Mat,PetscReal,PetscBool*);
4846: PetscObjectQueryFunction((PetscObject)A,"MatIsTranspose_C",&f);
4847: PetscObjectQueryFunction((PetscObject)B,"MatIsTranspose_C",&g);
4848: *flg = PETSC_FALSE;
4849: if (f && g) {
4850: if (f == g) {
4851: (*f)(A,B,tol,flg);
4852: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for symmetry test");
4853: } else {
4854: MatType mattype;
4855: if (!f) {
4856: MatGetType(A,&mattype);
4857: } else {
4858: MatGetType(B,&mattype);
4859: }
4860: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for transpose",mattype);
4861: }
4862: return(0);
4863: }
4865: /*@
4866: MatHermitianTranspose - Computes an in-place or out-of-place transpose of a matrix in complex conjugate.
4868: Collective on Mat
4870: Input Parameter:
4871: + mat - the matrix to transpose and complex conjugate
4872: - reuse - MAT_INITIAL_MATRIX to create a new matrix, MAT_INPLACE_MATRIX to reuse the first argument to store the transpose
4874: Output Parameters:
4875: . B - the Hermitian
4877: Level: intermediate
4879: .seealso: MatTranspose(), MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse
4880: @*/
4881: PetscErrorCode MatHermitianTranspose(Mat mat,MatReuse reuse,Mat *B)
4882: {
4886: MatTranspose(mat,reuse,B);
4887: #if defined(PETSC_USE_COMPLEX)
4888: MatConjugate(*B);
4889: #endif
4890: return(0);
4891: }
4893: /*@
4894: MatIsHermitianTranspose - Test whether a matrix is another one's Hermitian transpose,
4896: Collective on Mat
4898: Input Parameter:
4899: + A - the matrix to test
4900: - B - the matrix to test against, this can equal the first parameter
4902: Output Parameters:
4903: . flg - the result
4905: Notes:
4906: Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm
4907: has a running time of the order of the number of nonzeros; the parallel
4908: test involves parallel copies of the block-offdiagonal parts of the matrix.
4910: Level: intermediate
4912: .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian(), MatIsTranspose()
4913: @*/
4914: PetscErrorCode MatIsHermitianTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg)
4915: {
4916: PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool*),(*g)(Mat,Mat,PetscReal,PetscBool*);
4922: PetscObjectQueryFunction((PetscObject)A,"MatIsHermitianTranspose_C",&f);
4923: PetscObjectQueryFunction((PetscObject)B,"MatIsHermitianTranspose_C",&g);
4924: if (f && g) {
4925: if (f==g) {
4926: (*f)(A,B,tol,flg);
4927: } else SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for Hermitian test");
4928: }
4929: return(0);
4930: }
4932: /*@
4933: MatPermute - Creates a new matrix with rows and columns permuted from the
4934: original.
4936: Collective on Mat
4938: Input Parameters:
4939: + mat - the matrix to permute
4940: . row - row permutation, each processor supplies only the permutation for its rows
4941: - col - column permutation, each processor supplies only the permutation for its columns
4943: Output Parameters:
4944: . B - the permuted matrix
4946: Level: advanced
4948: Note:
4949: The index sets map from row/col of permuted matrix to row/col of original matrix.
4950: The index sets should be on the same communicator as Mat and have the same local sizes.
4952: .seealso: MatGetOrdering(), ISAllGather()
4954: @*/
4955: PetscErrorCode MatPermute(Mat mat,IS row,IS col,Mat *B)
4956: {
4965: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
4966: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
4967: if (!mat->ops->permute) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatPermute not available for Mat type %s",((PetscObject)mat)->type_name);
4968: MatCheckPreallocated(mat,1);
4970: (*mat->ops->permute)(mat,row,col,B);
4971: PetscObjectStateIncrease((PetscObject)*B);
4972: return(0);
4973: }
4975: /*@
4976: MatEqual - Compares two matrices.
4978: Collective on Mat
4980: Input Parameters:
4981: + A - the first matrix
4982: - B - the second matrix
4984: Output Parameter:
4985: . flg - PETSC_TRUE if the matrices are equal; PETSC_FALSE otherwise.
4987: Level: intermediate
4989: @*/
4990: PetscErrorCode MatEqual(Mat A,Mat B,PetscBool *flg)
4991: {
5001: MatCheckPreallocated(B,2);
5002: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5003: if (!B->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5004: 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);
5005: if (!A->ops->equal) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)A)->type_name);
5006: if (!B->ops->equal) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Mat type %s",((PetscObject)B)->type_name);
5007: 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);
5008: MatCheckPreallocated(A,1);
5010: (*A->ops->equal)(A,B,flg);
5011: return(0);
5012: }
5014: /*@
5015: MatDiagonalScale - Scales a matrix on the left and right by diagonal
5016: matrices that are stored as vectors. Either of the two scaling
5017: matrices can be NULL.
5019: Collective on Mat
5021: Input Parameters:
5022: + mat - the matrix to be scaled
5023: . l - the left scaling vector (or NULL)
5024: - r - the right scaling vector (or NULL)
5026: Notes:
5027: MatDiagonalScale() computes A = LAR, where
5028: L = a diagonal matrix (stored as a vector), R = a diagonal matrix (stored as a vector)
5029: The L scales the rows of the matrix, the R scales the columns of the matrix.
5031: Level: intermediate
5034: .seealso: MatScale(), MatShift(), MatDiagonalSet()
5035: @*/
5036: PetscErrorCode MatDiagonalScale(Mat mat,Vec l,Vec r)
5037: {
5043: if (!mat->ops->diagonalscale) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5046: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5047: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5048: MatCheckPreallocated(mat,1);
5050: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
5051: (*mat->ops->diagonalscale)(mat,l,r);
5052: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
5053: PetscObjectStateIncrease((PetscObject)mat);
5054: return(0);
5055: }
5057: /*@
5058: MatScale - Scales all elements of a matrix by a given number.
5060: Logically Collective on Mat
5062: Input Parameters:
5063: + mat - the matrix to be scaled
5064: - a - the scaling value
5066: Output Parameter:
5067: . mat - the scaled matrix
5069: Level: intermediate
5071: .seealso: MatDiagonalScale()
5072: @*/
5073: PetscErrorCode MatScale(Mat mat,PetscScalar a)
5074: {
5080: if (a != (PetscScalar)1.0 && !mat->ops->scale) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5081: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5082: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5084: MatCheckPreallocated(mat,1);
5086: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
5087: if (a != (PetscScalar)1.0) {
5088: (*mat->ops->scale)(mat,a);
5089: PetscObjectStateIncrease((PetscObject)mat);
5090: }
5091: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
5092: return(0);
5093: }
5095: /*@
5096: MatNorm - Calculates various norms of a matrix.
5098: Collective on Mat
5100: Input Parameters:
5101: + mat - the matrix
5102: - type - the type of norm, NORM_1, NORM_FROBENIUS, NORM_INFINITY
5104: Output Parameters:
5105: . nrm - the resulting norm
5107: Level: intermediate
5109: @*/
5110: PetscErrorCode MatNorm(Mat mat,NormType type,PetscReal *nrm)
5111: {
5119: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5120: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5121: if (!mat->ops->norm) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5122: MatCheckPreallocated(mat,1);
5124: (*mat->ops->norm)(mat,type,nrm);
5125: return(0);
5126: }
5128: /*
5129: This variable is used to prevent counting of MatAssemblyBegin() that
5130: are called from within a MatAssemblyEnd().
5131: */
5132: static PetscInt MatAssemblyEnd_InUse = 0;
5133: /*@
5134: MatAssemblyBegin - Begins assembling the matrix. This routine should
5135: be called after completing all calls to MatSetValues().
5137: Collective on Mat
5139: Input Parameters:
5140: + mat - the matrix
5141: - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY
5143: Notes:
5144: MatSetValues() generally caches the values. The matrix is ready to
5145: use only after MatAssemblyBegin() and MatAssemblyEnd() have been called.
5146: Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES
5147: in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before
5148: using the matrix.
5150: ALL processes that share a matrix MUST call MatAssemblyBegin() and MatAssemblyEnd() the SAME NUMBER of times, and each time with the
5151: 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
5152: a global collective operation requring all processes that share the matrix.
5154: Space for preallocated nonzeros that is not filled by a call to MatSetValues() or a related routine are compressed
5155: out by assembly. If you intend to use that extra space on a subsequent assembly, be sure to insert explicit zeros
5156: before MAT_FINAL_ASSEMBLY so the space is not compressed out.
5158: Level: beginner
5160: .seealso: MatAssemblyEnd(), MatSetValues(), MatAssembled()
5161: @*/
5162: PetscErrorCode MatAssemblyBegin(Mat mat,MatAssemblyType type)
5163: {
5169: MatCheckPreallocated(mat,1);
5170: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix.\nDid you forget to call MatSetUnfactored()?");
5171: if (mat->assembled) {
5172: mat->was_assembled = PETSC_TRUE;
5173: mat->assembled = PETSC_FALSE;
5174: }
5176: if (!MatAssemblyEnd_InUse) {
5177: PetscLogEventBegin(MAT_AssemblyBegin,mat,0,0,0);
5178: if (mat->ops->assemblybegin) {(*mat->ops->assemblybegin)(mat,type);}
5179: PetscLogEventEnd(MAT_AssemblyBegin,mat,0,0,0);
5180: } else if (mat->ops->assemblybegin) {
5181: (*mat->ops->assemblybegin)(mat,type);
5182: }
5183: return(0);
5184: }
5186: /*@
5187: MatAssembled - Indicates if a matrix has been assembled and is ready for
5188: use; for example, in matrix-vector product.
5190: Not Collective
5192: Input Parameter:
5193: . mat - the matrix
5195: Output Parameter:
5196: . assembled - PETSC_TRUE or PETSC_FALSE
5198: Level: advanced
5200: .seealso: MatAssemblyEnd(), MatSetValues(), MatAssemblyBegin()
5201: @*/
5202: PetscErrorCode MatAssembled(Mat mat,PetscBool *assembled)
5203: {
5207: *assembled = mat->assembled;
5208: return(0);
5209: }
5211: /*@
5212: MatAssemblyEnd - Completes assembling the matrix. This routine should
5213: be called after MatAssemblyBegin().
5215: Collective on Mat
5217: Input Parameters:
5218: + mat - the matrix
5219: - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY
5221: Options Database Keys:
5222: + -mat_view ::ascii_info - Prints info on matrix at conclusion of MatEndAssembly()
5223: . -mat_view ::ascii_info_detail - Prints more detailed info
5224: . -mat_view - Prints matrix in ASCII format
5225: . -mat_view ::ascii_matlab - Prints matrix in Matlab format
5226: . -mat_view draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX().
5227: . -display <name> - Sets display name (default is host)
5228: . -draw_pause <sec> - Sets number of seconds to pause after display
5229: . -mat_view socket - Sends matrix to socket, can be accessed from Matlab (See Users-Manual: Chapter 12 Using MATLAB with PETSc )
5230: . -viewer_socket_machine <machine> - Machine to use for socket
5231: . -viewer_socket_port <port> - Port number to use for socket
5232: - -mat_view binary:filename[:append] - Save matrix to file in binary format
5234: Notes:
5235: MatSetValues() generally caches the values. The matrix is ready to
5236: use only after MatAssemblyBegin() and MatAssemblyEnd() have been called.
5237: Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES
5238: in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before
5239: using the matrix.
5241: Space for preallocated nonzeros that is not filled by a call to MatSetValues() or a related routine are compressed
5242: out by assembly. If you intend to use that extra space on a subsequent assembly, be sure to insert explicit zeros
5243: before MAT_FINAL_ASSEMBLY so the space is not compressed out.
5245: Level: beginner
5247: .seealso: MatAssemblyBegin(), MatSetValues(), PetscDrawOpenX(), PetscDrawCreate(), MatView(), MatAssembled(), PetscViewerSocketOpen()
5248: @*/
5249: PetscErrorCode MatAssemblyEnd(Mat mat,MatAssemblyType type)
5250: {
5251: PetscErrorCode ierr;
5252: static PetscInt inassm = 0;
5253: PetscBool flg = PETSC_FALSE;
5259: inassm++;
5260: MatAssemblyEnd_InUse++;
5261: if (MatAssemblyEnd_InUse == 1) { /* Do the logging only the first time through */
5262: PetscLogEventBegin(MAT_AssemblyEnd,mat,0,0,0);
5263: if (mat->ops->assemblyend) {
5264: (*mat->ops->assemblyend)(mat,type);
5265: }
5266: PetscLogEventEnd(MAT_AssemblyEnd,mat,0,0,0);
5267: } else if (mat->ops->assemblyend) {
5268: (*mat->ops->assemblyend)(mat,type);
5269: }
5271: /* Flush assembly is not a true assembly */
5272: if (type != MAT_FLUSH_ASSEMBLY) {
5273: mat->num_ass++;
5274: mat->assembled = PETSC_TRUE;
5275: mat->ass_nonzerostate = mat->nonzerostate;
5276: }
5278: mat->insertmode = NOT_SET_VALUES;
5279: MatAssemblyEnd_InUse--;
5280: PetscObjectStateIncrease((PetscObject)mat);
5281: if (!mat->symmetric_eternal) {
5282: mat->symmetric_set = PETSC_FALSE;
5283: mat->hermitian_set = PETSC_FALSE;
5284: mat->structurally_symmetric_set = PETSC_FALSE;
5285: }
5286: if (inassm == 1 && type != MAT_FLUSH_ASSEMBLY) {
5287: MatViewFromOptions(mat,NULL,"-mat_view");
5289: if (mat->checksymmetryonassembly) {
5290: MatIsSymmetric(mat,mat->checksymmetrytol,&flg);
5291: if (flg) {
5292: PetscPrintf(PetscObjectComm((PetscObject)mat),"Matrix is symmetric (tolerance %g)\n",(double)mat->checksymmetrytol);
5293: } else {
5294: PetscPrintf(PetscObjectComm((PetscObject)mat),"Matrix is not symmetric (tolerance %g)\n",(double)mat->checksymmetrytol);
5295: }
5296: }
5297: if (mat->nullsp && mat->checknullspaceonassembly) {
5298: MatNullSpaceTest(mat->nullsp,mat,NULL);
5299: }
5300: }
5301: inassm--;
5302: return(0);
5303: }
5305: /*@
5306: MatSetOption - Sets a parameter option for a matrix. Some options
5307: may be specific to certain storage formats. Some options
5308: determine how values will be inserted (or added). Sorted,
5309: row-oriented input will generally assemble the fastest. The default
5310: is row-oriented.
5312: Logically Collective on Mat for certain operations, such as MAT_SPD, not collective for MAT_ROW_ORIENTED, see MatOption
5314: Input Parameters:
5315: + mat - the matrix
5316: . option - the option, one of those listed below (and possibly others),
5317: - flg - turn the option on (PETSC_TRUE) or off (PETSC_FALSE)
5319: Options Describing Matrix Structure:
5320: + MAT_SPD - symmetric positive definite
5321: . MAT_SYMMETRIC - symmetric in terms of both structure and value
5322: . MAT_HERMITIAN - transpose is the complex conjugation
5323: . MAT_STRUCTURALLY_SYMMETRIC - symmetric nonzero structure
5324: - MAT_SYMMETRY_ETERNAL - if you would like the symmetry/Hermitian flag
5325: you set to be kept with all future use of the matrix
5326: including after MatAssemblyBegin/End() which could
5327: potentially change the symmetry structure, i.e. you
5328: KNOW the matrix will ALWAYS have the property you set.
5331: Options For Use with MatSetValues():
5332: Insert a logically dense subblock, which can be
5333: . MAT_ROW_ORIENTED - row-oriented (default)
5335: Note these options reflect the data you pass in with MatSetValues(); it has
5336: nothing to do with how the data is stored internally in the matrix
5337: data structure.
5339: When (re)assembling a matrix, we can restrict the input for
5340: efficiency/debugging purposes. These options include:
5341: + MAT_NEW_NONZERO_LOCATIONS - additional insertions will be allowed if they generate a new nonzero (slow)
5342: . MAT_NEW_DIAGONALS - new diagonals will be allowed (for block diagonal format only)
5343: . MAT_IGNORE_OFF_PROC_ENTRIES - drops off-processor entries
5344: . MAT_NEW_NONZERO_LOCATION_ERR - generates an error for new matrix entry
5345: . MAT_USE_HASH_TABLE - uses a hash table to speed up matrix assembly
5346: . MAT_NO_OFF_PROC_ENTRIES - you know each process will only set values for its own rows, will generate an error if
5347: any process sets values for another process. This avoids all reductions in the MatAssembly routines and thus improves
5348: performance for very large process counts.
5349: - MAT_SUBSET_OFF_PROC_ENTRIES - you know that the first assembly after setting this flag will set a superset
5350: of the off-process entries required for all subsequent assemblies. This avoids a rendezvous step in the MatAssembly
5351: functions, instead sending only neighbor messages.
5353: Notes:
5354: Except for MAT_UNUSED_NONZERO_LOCATION_ERR and MAT_ROW_ORIENTED all processes that share the matrix must pass the same value in flg!
5356: Some options are relevant only for particular matrix types and
5357: are thus ignored by others. Other options are not supported by
5358: certain matrix types and will generate an error message if set.
5360: If using a Fortran 77 module to compute a matrix, one may need to
5361: use the column-oriented option (or convert to the row-oriented
5362: format).
5364: MAT_NEW_NONZERO_LOCATIONS set to PETSC_FALSE indicates that any add or insertion
5365: that would generate a new entry in the nonzero structure is instead
5366: ignored. Thus, if memory has not alredy been allocated for this particular
5367: data, then the insertion is ignored. For dense matrices, in which
5368: the entire array is allocated, no entries are ever ignored.
5369: Set after the first MatAssemblyEnd(). If this option is set then the MatAssemblyBegin/End() processes has one less global reduction
5371: MAT_NEW_NONZERO_LOCATION_ERR set to PETSC_TRUE indicates that any add or insertion
5372: that would generate a new entry in the nonzero structure instead produces
5373: 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
5375: MAT_NEW_NONZERO_ALLOCATION_ERR set to PETSC_TRUE indicates that any add or insertion
5376: that would generate a new entry that has not been preallocated will
5377: instead produce an error. (Currently supported for AIJ and BAIJ formats
5378: only.) This is a useful flag when debugging matrix memory preallocation.
5379: If this option is set then the MatAssemblyBegin/End() processes has one less global reduction
5381: MAT_IGNORE_OFF_PROC_ENTRIES set to PETSC_TRUE indicates entries destined for
5382: other processors should be dropped, rather than stashed.
5383: This is useful if you know that the "owning" processor is also
5384: always generating the correct matrix entries, so that PETSc need
5385: not transfer duplicate entries generated on another processor.
5387: MAT_USE_HASH_TABLE indicates that a hash table be used to improve the
5388: searches during matrix assembly. When this flag is set, the hash table
5389: is created during the first Matrix Assembly. This hash table is
5390: used the next time through, during MatSetVaules()/MatSetVaulesBlocked()
5391: to improve the searching of indices. MAT_NEW_NONZERO_LOCATIONS flag
5392: should be used with MAT_USE_HASH_TABLE flag. This option is currently
5393: supported by MATMPIBAIJ format only.
5395: MAT_KEEP_NONZERO_PATTERN indicates when MatZeroRows() is called the zeroed entries
5396: are kept in the nonzero structure
5398: MAT_IGNORE_ZERO_ENTRIES - for AIJ/IS matrices this will stop zero values from creating
5399: a zero location in the matrix
5401: MAT_USE_INODES - indicates using inode version of the code - works with AIJ matrix types
5403: MAT_NO_OFF_PROC_ZERO_ROWS - you know each process will only zero its own rows. This avoids all reductions in the
5404: zero row routines and thus improves performance for very large process counts.
5406: MAT_IGNORE_LOWER_TRIANGULAR - For SBAIJ matrices will ignore any insertions you make in the lower triangular
5407: part of the matrix (since they should match the upper triangular part).
5409: MAT_SORTED_FULL - each process provides exactly its local rows; all column indices for a given row are passed in a
5410: single call to MatSetValues(), preallocation is perfect, row oriented, INSERT_VALUES is used. Common
5411: with finite difference schemes with non-periodic boundary conditions.
5412: Notes:
5413: Can only be called after MatSetSizes() and MatSetType() have been set.
5415: Level: intermediate
5417: .seealso: MatOption, Mat
5419: @*/
5420: PetscErrorCode MatSetOption(Mat mat,MatOption op,PetscBool flg)
5421: {
5427: if (op > 0) {
5430: }
5432: 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);
5433: 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()");
5435: switch (op) {
5436: case MAT_NO_OFF_PROC_ENTRIES:
5437: mat->nooffprocentries = flg;
5438: return(0);
5439: break;
5440: case MAT_SUBSET_OFF_PROC_ENTRIES:
5441: mat->assembly_subset = flg;
5442: if (!mat->assembly_subset) { /* See the same logic in VecAssembly wrt VEC_SUBSET_OFF_PROC_ENTRIES */
5443: #if !defined(PETSC_HAVE_MPIUNI)
5444: MatStashScatterDestroy_BTS(&mat->stash);
5445: #endif
5446: mat->stash.first_assembly_done = PETSC_FALSE;
5447: }
5448: return(0);
5449: case MAT_NO_OFF_PROC_ZERO_ROWS:
5450: mat->nooffproczerorows = flg;
5451: return(0);
5452: break;
5453: case MAT_SPD:
5454: mat->spd_set = PETSC_TRUE;
5455: mat->spd = flg;
5456: if (flg) {
5457: mat->symmetric = PETSC_TRUE;
5458: mat->structurally_symmetric = PETSC_TRUE;
5459: mat->symmetric_set = PETSC_TRUE;
5460: mat->structurally_symmetric_set = PETSC_TRUE;
5461: }
5462: break;
5463: case MAT_SYMMETRIC:
5464: mat->symmetric = flg;
5465: if (flg) mat->structurally_symmetric = PETSC_TRUE;
5466: mat->symmetric_set = PETSC_TRUE;
5467: mat->structurally_symmetric_set = flg;
5468: #if !defined(PETSC_USE_COMPLEX)
5469: mat->hermitian = flg;
5470: mat->hermitian_set = PETSC_TRUE;
5471: #endif
5472: break;
5473: case MAT_HERMITIAN:
5474: mat->hermitian = flg;
5475: if (flg) mat->structurally_symmetric = PETSC_TRUE;
5476: mat->hermitian_set = PETSC_TRUE;
5477: mat->structurally_symmetric_set = flg;
5478: #if !defined(PETSC_USE_COMPLEX)
5479: mat->symmetric = flg;
5480: mat->symmetric_set = PETSC_TRUE;
5481: #endif
5482: break;
5483: case MAT_STRUCTURALLY_SYMMETRIC:
5484: mat->structurally_symmetric = flg;
5485: mat->structurally_symmetric_set = PETSC_TRUE;
5486: break;
5487: case MAT_SYMMETRY_ETERNAL:
5488: mat->symmetric_eternal = flg;
5489: break;
5490: case MAT_STRUCTURE_ONLY:
5491: mat->structure_only = flg;
5492: break;
5493: case MAT_SORTED_FULL:
5494: mat->sortedfull = flg;
5495: break;
5496: default:
5497: break;
5498: }
5499: if (mat->ops->setoption) {
5500: (*mat->ops->setoption)(mat,op,flg);
5501: }
5502: return(0);
5503: }
5505: /*@
5506: MatGetOption - Gets a parameter option that has been set for a matrix.
5508: Logically Collective on Mat for certain operations, such as MAT_SPD, not collective for MAT_ROW_ORIENTED, see MatOption
5510: Input Parameters:
5511: + mat - the matrix
5512: - option - the option, this only responds to certain options, check the code for which ones
5514: Output Parameter:
5515: . flg - turn the option on (PETSC_TRUE) or off (PETSC_FALSE)
5517: Notes:
5518: Can only be called after MatSetSizes() and MatSetType() have been set.
5520: Level: intermediate
5522: .seealso: MatOption, MatSetOption()
5524: @*/
5525: PetscErrorCode MatGetOption(Mat mat,MatOption op,PetscBool *flg)
5526: {
5531: 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);
5532: 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()");
5534: switch (op) {
5535: case MAT_NO_OFF_PROC_ENTRIES:
5536: *flg = mat->nooffprocentries;
5537: break;
5538: case MAT_NO_OFF_PROC_ZERO_ROWS:
5539: *flg = mat->nooffproczerorows;
5540: break;
5541: case MAT_SYMMETRIC:
5542: *flg = mat->symmetric;
5543: break;
5544: case MAT_HERMITIAN:
5545: *flg = mat->hermitian;
5546: break;
5547: case MAT_STRUCTURALLY_SYMMETRIC:
5548: *flg = mat->structurally_symmetric;
5549: break;
5550: case MAT_SYMMETRY_ETERNAL:
5551: *flg = mat->symmetric_eternal;
5552: break;
5553: case MAT_SPD:
5554: *flg = mat->spd;
5555: break;
5556: default:
5557: break;
5558: }
5559: return(0);
5560: }
5562: /*@
5563: MatZeroEntries - Zeros all entries of a matrix. For sparse matrices
5564: this routine retains the old nonzero structure.
5566: Logically Collective on Mat
5568: Input Parameters:
5569: . mat - the matrix
5571: Level: intermediate
5573: Notes:
5574: 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.
5575: See the Performance chapter of the users manual for information on preallocating matrices.
5577: .seealso: MatZeroRows()
5578: @*/
5579: PetscErrorCode MatZeroEntries(Mat mat)
5580: {
5586: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5587: 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");
5588: if (!mat->ops->zeroentries) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5589: MatCheckPreallocated(mat,1);
5591: PetscLogEventBegin(MAT_ZeroEntries,mat,0,0,0);
5592: (*mat->ops->zeroentries)(mat);
5593: PetscLogEventEnd(MAT_ZeroEntries,mat,0,0,0);
5594: PetscObjectStateIncrease((PetscObject)mat);
5595: return(0);
5596: }
5598: /*@
5599: MatZeroRowsColumns - Zeros all entries (except possibly the main diagonal)
5600: of a set of rows and columns of a matrix.
5602: Collective on Mat
5604: Input Parameters:
5605: + mat - the matrix
5606: . numRows - the number of rows to remove
5607: . rows - the global row indices
5608: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5609: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5610: - b - optional vector of right hand side, that will be adjusted by provided solution
5612: Notes:
5613: This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix.
5615: The user can set a value in the diagonal entry (or for the AIJ and
5616: row formats can optionally remove the main diagonal entry from the
5617: nonzero structure as well, by passing 0.0 as the final argument).
5619: For the parallel case, all processes that share the matrix (i.e.,
5620: those in the communicator used for matrix creation) MUST call this
5621: routine, regardless of whether any rows being zeroed are owned by
5622: them.
5624: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5625: list only rows local to itself).
5627: The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine.
5629: Level: intermediate
5631: .seealso: MatZeroRowsIS(), MatZeroRows(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5632: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5633: @*/
5634: PetscErrorCode MatZeroRowsColumns(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
5635: {
5642: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5643: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5644: if (!mat->ops->zerorowscolumns) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5645: MatCheckPreallocated(mat,1);
5647: (*mat->ops->zerorowscolumns)(mat,numRows,rows,diag,x,b);
5648: MatViewFromOptions(mat,NULL,"-mat_view");
5649: PetscObjectStateIncrease((PetscObject)mat);
5650: return(0);
5651: }
5653: /*@
5654: MatZeroRowsColumnsIS - Zeros all entries (except possibly the main diagonal)
5655: of a set of rows and columns of a matrix.
5657: Collective on Mat
5659: Input Parameters:
5660: + mat - the matrix
5661: . is - the rows to zero
5662: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5663: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5664: - b - optional vector of right hand side, that will be adjusted by provided solution
5666: Notes:
5667: This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix.
5669: The user can set a value in the diagonal entry (or for the AIJ and
5670: row formats can optionally remove the main diagonal entry from the
5671: nonzero structure as well, by passing 0.0 as the final argument).
5673: For the parallel case, all processes that share the matrix (i.e.,
5674: those in the communicator used for matrix creation) MUST call this
5675: routine, regardless of whether any rows being zeroed are owned by
5676: them.
5678: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5679: list only rows local to itself).
5681: The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine.
5683: Level: intermediate
5685: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5686: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRows(), MatZeroRowsColumnsStencil()
5687: @*/
5688: PetscErrorCode MatZeroRowsColumnsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
5689: {
5691: PetscInt numRows;
5692: const PetscInt *rows;
5699: ISGetLocalSize(is,&numRows);
5700: ISGetIndices(is,&rows);
5701: MatZeroRowsColumns(mat,numRows,rows,diag,x,b);
5702: ISRestoreIndices(is,&rows);
5703: return(0);
5704: }
5706: /*@
5707: MatZeroRows - Zeros all entries (except possibly the main diagonal)
5708: of a set of rows of a matrix.
5710: Collective on Mat
5712: Input Parameters:
5713: + mat - the matrix
5714: . numRows - the number of rows to remove
5715: . rows - the global row indices
5716: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5717: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5718: - b - optional vector of right hand side, that will be adjusted by provided solution
5720: Notes:
5721: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5722: but does not release memory. For the dense and block diagonal
5723: formats this does not alter the nonzero structure.
5725: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5726: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5727: merely zeroed.
5729: The user can set a value in the diagonal entry (or for the AIJ and
5730: row formats can optionally remove the main diagonal entry from the
5731: nonzero structure as well, by passing 0.0 as the final argument).
5733: For the parallel case, all processes that share the matrix (i.e.,
5734: those in the communicator used for matrix creation) MUST call this
5735: routine, regardless of whether any rows being zeroed are owned by
5736: them.
5738: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5739: list only rows local to itself).
5741: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
5742: owns that are to be zeroed. This saves a global synchronization in the implementation.
5744: Level: intermediate
5746: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5747: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5748: @*/
5749: PetscErrorCode MatZeroRows(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
5750: {
5757: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
5758: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
5759: if (!mat->ops->zerorows) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
5760: MatCheckPreallocated(mat,1);
5762: (*mat->ops->zerorows)(mat,numRows,rows,diag,x,b);
5763: MatViewFromOptions(mat,NULL,"-mat_view");
5764: PetscObjectStateIncrease((PetscObject)mat);
5765: return(0);
5766: }
5768: /*@
5769: MatZeroRowsIS - Zeros all entries (except possibly the main diagonal)
5770: of a set of rows of a matrix.
5772: Collective on Mat
5774: Input Parameters:
5775: + mat - the matrix
5776: . is - index set of rows to remove
5777: . diag - value put in all diagonals of eliminated rows
5778: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5779: - b - optional vector of right hand side, that will be adjusted by provided solution
5781: Notes:
5782: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5783: but does not release memory. For the dense and block diagonal
5784: formats this does not alter the nonzero structure.
5786: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5787: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5788: merely zeroed.
5790: The user can set a value in the diagonal entry (or for the AIJ and
5791: row formats can optionally remove the main diagonal entry from the
5792: nonzero structure as well, by passing 0.0 as the final argument).
5794: For the parallel case, all processes that share the matrix (i.e.,
5795: those in the communicator used for matrix creation) MUST call this
5796: routine, regardless of whether any rows being zeroed are owned by
5797: them.
5799: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5800: list only rows local to itself).
5802: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
5803: owns that are to be zeroed. This saves a global synchronization in the implementation.
5805: Level: intermediate
5807: .seealso: MatZeroRows(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5808: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5809: @*/
5810: PetscErrorCode MatZeroRowsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
5811: {
5812: PetscInt numRows;
5813: const PetscInt *rows;
5820: ISGetLocalSize(is,&numRows);
5821: ISGetIndices(is,&rows);
5822: MatZeroRows(mat,numRows,rows,diag,x,b);
5823: ISRestoreIndices(is,&rows);
5824: return(0);
5825: }
5827: /*@
5828: MatZeroRowsStencil - Zeros all entries (except possibly the main diagonal)
5829: of a set of rows of a matrix. These rows must be local to the process.
5831: Collective on Mat
5833: Input Parameters:
5834: + mat - the matrix
5835: . numRows - the number of rows to remove
5836: . rows - the grid coordinates (and component number when dof > 1) for matrix rows
5837: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5838: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5839: - b - optional vector of right hand side, that will be adjusted by provided solution
5841: Notes:
5842: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5843: but does not release memory. For the dense and block diagonal
5844: formats this does not alter the nonzero structure.
5846: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5847: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5848: merely zeroed.
5850: The user can set a value in the diagonal entry (or for the AIJ and
5851: row formats can optionally remove the main diagonal entry from the
5852: nonzero structure as well, by passing 0.0 as the final argument).
5854: For the parallel case, all processes that share the matrix (i.e.,
5855: those in the communicator used for matrix creation) MUST call this
5856: routine, regardless of whether any rows being zeroed are owned by
5857: them.
5859: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5860: list only rows local to itself).
5862: The grid coordinates are across the entire grid, not just the local portion
5864: In Fortran idxm and idxn should be declared as
5865: $ MatStencil idxm(4,m)
5866: and the values inserted using
5867: $ idxm(MatStencil_i,1) = i
5868: $ idxm(MatStencil_j,1) = j
5869: $ idxm(MatStencil_k,1) = k
5870: $ idxm(MatStencil_c,1) = c
5871: etc
5873: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
5874: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
5875: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
5876: DM_BOUNDARY_PERIODIC boundary type.
5878: 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
5879: a single value per point) you can skip filling those indices.
5881: Level: intermediate
5883: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsl(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5884: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
5885: @*/
5886: PetscErrorCode MatZeroRowsStencil(Mat mat,PetscInt numRows,const MatStencil rows[],PetscScalar diag,Vec x,Vec b)
5887: {
5888: PetscInt dim = mat->stencil.dim;
5889: PetscInt sdim = dim - (1 - (PetscInt) mat->stencil.noc);
5890: PetscInt *dims = mat->stencil.dims+1;
5891: PetscInt *starts = mat->stencil.starts;
5892: PetscInt *dxm = (PetscInt*) rows;
5893: PetscInt *jdxm, i, j, tmp, numNewRows = 0;
5901: PetscMalloc1(numRows, &jdxm);
5902: for (i = 0; i < numRows; ++i) {
5903: /* Skip unused dimensions (they are ordered k, j, i, c) */
5904: for (j = 0; j < 3-sdim; ++j) dxm++;
5905: /* Local index in X dir */
5906: tmp = *dxm++ - starts[0];
5907: /* Loop over remaining dimensions */
5908: for (j = 0; j < dim-1; ++j) {
5909: /* If nonlocal, set index to be negative */
5910: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT;
5911: /* Update local index */
5912: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
5913: }
5914: /* Skip component slot if necessary */
5915: if (mat->stencil.noc) dxm++;
5916: /* Local row number */
5917: if (tmp >= 0) {
5918: jdxm[numNewRows++] = tmp;
5919: }
5920: }
5921: MatZeroRowsLocal(mat,numNewRows,jdxm,diag,x,b);
5922: PetscFree(jdxm);
5923: return(0);
5924: }
5926: /*@
5927: MatZeroRowsColumnsStencil - Zeros all row and column entries (except possibly the main diagonal)
5928: of a set of rows and columns of a matrix.
5930: Collective on Mat
5932: Input Parameters:
5933: + mat - the matrix
5934: . numRows - the number of rows/columns to remove
5935: . rows - the grid coordinates (and component number when dof > 1) for matrix rows
5936: . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry)
5937: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
5938: - b - optional vector of right hand side, that will be adjusted by provided solution
5940: Notes:
5941: For the AIJ and BAIJ matrix formats this removes the old nonzero structure,
5942: but does not release memory. For the dense and block diagonal
5943: formats this does not alter the nonzero structure.
5945: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
5946: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
5947: merely zeroed.
5949: The user can set a value in the diagonal entry (or for the AIJ and
5950: row formats can optionally remove the main diagonal entry from the
5951: nonzero structure as well, by passing 0.0 as the final argument).
5953: For the parallel case, all processes that share the matrix (i.e.,
5954: those in the communicator used for matrix creation) MUST call this
5955: routine, regardless of whether any rows being zeroed are owned by
5956: them.
5958: Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to
5959: list only rows local to itself, but the row/column numbers are given in local numbering).
5961: The grid coordinates are across the entire grid, not just the local portion
5963: In Fortran idxm and idxn should be declared as
5964: $ MatStencil idxm(4,m)
5965: and the values inserted using
5966: $ idxm(MatStencil_i,1) = i
5967: $ idxm(MatStencil_j,1) = j
5968: $ idxm(MatStencil_k,1) = k
5969: $ idxm(MatStencil_c,1) = c
5970: etc
5972: For periodic boundary conditions use negative indices for values to the left (below 0; that are to be
5973: obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one
5974: etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the
5975: DM_BOUNDARY_PERIODIC boundary type.
5977: 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
5978: a single value per point) you can skip filling those indices.
5980: Level: intermediate
5982: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
5983: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRows()
5984: @*/
5985: PetscErrorCode MatZeroRowsColumnsStencil(Mat mat,PetscInt numRows,const MatStencil rows[],PetscScalar diag,Vec x,Vec b)
5986: {
5987: PetscInt dim = mat->stencil.dim;
5988: PetscInt sdim = dim - (1 - (PetscInt) mat->stencil.noc);
5989: PetscInt *dims = mat->stencil.dims+1;
5990: PetscInt *starts = mat->stencil.starts;
5991: PetscInt *dxm = (PetscInt*) rows;
5992: PetscInt *jdxm, i, j, tmp, numNewRows = 0;
6000: PetscMalloc1(numRows, &jdxm);
6001: for (i = 0; i < numRows; ++i) {
6002: /* Skip unused dimensions (they are ordered k, j, i, c) */
6003: for (j = 0; j < 3-sdim; ++j) dxm++;
6004: /* Local index in X dir */
6005: tmp = *dxm++ - starts[0];
6006: /* Loop over remaining dimensions */
6007: for (j = 0; j < dim-1; ++j) {
6008: /* If nonlocal, set index to be negative */
6009: if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT;
6010: /* Update local index */
6011: else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1];
6012: }
6013: /* Skip component slot if necessary */
6014: if (mat->stencil.noc) dxm++;
6015: /* Local row number */
6016: if (tmp >= 0) {
6017: jdxm[numNewRows++] = tmp;
6018: }
6019: }
6020: MatZeroRowsColumnsLocal(mat,numNewRows,jdxm,diag,x,b);
6021: PetscFree(jdxm);
6022: return(0);
6023: }
6025: /*@C
6026: MatZeroRowsLocal - Zeros all entries (except possibly the main diagonal)
6027: of a set of rows of a matrix; using local numbering of rows.
6029: Collective on Mat
6031: Input Parameters:
6032: + mat - the matrix
6033: . numRows - the number of rows to remove
6034: . rows - the global row indices
6035: . diag - value put in all diagonals of eliminated rows
6036: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6037: - b - optional vector of right hand side, that will be adjusted by provided solution
6039: Notes:
6040: Before calling MatZeroRowsLocal(), the user must first set the
6041: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6043: For the AIJ matrix formats this removes the old nonzero structure,
6044: but does not release memory. For the dense and block diagonal
6045: formats this does not alter the nonzero structure.
6047: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6048: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6049: merely zeroed.
6051: The user can set a value in the diagonal entry (or for the AIJ and
6052: row formats can optionally remove the main diagonal entry from the
6053: nonzero structure as well, by passing 0.0 as the final argument).
6055: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
6056: owns that are to be zeroed. This saves a global synchronization in the implementation.
6058: Level: intermediate
6060: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRows(), MatSetOption(),
6061: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6062: @*/
6063: PetscErrorCode MatZeroRowsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
6064: {
6071: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6072: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6073: MatCheckPreallocated(mat,1);
6075: if (mat->ops->zerorowslocal) {
6076: (*mat->ops->zerorowslocal)(mat,numRows,rows,diag,x,b);
6077: } else {
6078: IS is, newis;
6079: const PetscInt *newRows;
6081: if (!mat->rmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first");
6082: ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);
6083: ISLocalToGlobalMappingApplyIS(mat->rmap->mapping,is,&newis);
6084: ISGetIndices(newis,&newRows);
6085: (*mat->ops->zerorows)(mat,numRows,newRows,diag,x,b);
6086: ISRestoreIndices(newis,&newRows);
6087: ISDestroy(&newis);
6088: ISDestroy(&is);
6089: }
6090: PetscObjectStateIncrease((PetscObject)mat);
6091: return(0);
6092: }
6094: /*@
6095: MatZeroRowsLocalIS - Zeros all entries (except possibly the main diagonal)
6096: of a set of rows of a matrix; using local numbering of rows.
6098: Collective on Mat
6100: Input Parameters:
6101: + mat - the matrix
6102: . is - index set of rows to remove
6103: . diag - value put in all diagonals of eliminated rows
6104: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6105: - b - optional vector of right hand side, that will be adjusted by provided solution
6107: Notes:
6108: Before calling MatZeroRowsLocalIS(), the user must first set the
6109: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6111: For the AIJ matrix formats this removes the old nonzero structure,
6112: but does not release memory. For the dense and block diagonal
6113: formats this does not alter the nonzero structure.
6115: If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure
6116: of the matrix is not changed (even for AIJ and BAIJ matrices) the values are
6117: merely zeroed.
6119: The user can set a value in the diagonal entry (or for the AIJ and
6120: row formats can optionally remove the main diagonal entry from the
6121: nonzero structure as well, by passing 0.0 as the final argument).
6123: You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it
6124: owns that are to be zeroed. This saves a global synchronization in the implementation.
6126: Level: intermediate
6128: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRows(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6129: MatZeroRowsColumnsLocal(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6130: @*/
6131: PetscErrorCode MatZeroRowsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
6132: {
6134: PetscInt numRows;
6135: const PetscInt *rows;
6141: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6142: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6143: MatCheckPreallocated(mat,1);
6145: ISGetLocalSize(is,&numRows);
6146: ISGetIndices(is,&rows);
6147: MatZeroRowsLocal(mat,numRows,rows,diag,x,b);
6148: ISRestoreIndices(is,&rows);
6149: return(0);
6150: }
6152: /*@
6153: MatZeroRowsColumnsLocal - Zeros all entries (except possibly the main diagonal)
6154: of a set of rows and columns of a matrix; using local numbering of rows.
6156: Collective on Mat
6158: Input Parameters:
6159: + mat - the matrix
6160: . numRows - the number of rows to remove
6161: . rows - the global row indices
6162: . diag - value put in all diagonals of eliminated rows
6163: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6164: - b - optional vector of right hand side, that will be adjusted by provided solution
6166: Notes:
6167: Before calling MatZeroRowsColumnsLocal(), the user must first set the
6168: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6170: The user can set a value in the diagonal entry (or for the AIJ and
6171: row formats can optionally remove the main diagonal entry from the
6172: nonzero structure as well, by passing 0.0 as the final argument).
6174: Level: intermediate
6176: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6177: MatZeroRows(), MatZeroRowsColumnsLocalIS(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6178: @*/
6179: PetscErrorCode MatZeroRowsColumnsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b)
6180: {
6182: IS is, newis;
6183: const PetscInt *newRows;
6189: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6190: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6191: MatCheckPreallocated(mat,1);
6193: if (!mat->cmap->mapping) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first");
6194: ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);
6195: ISLocalToGlobalMappingApplyIS(mat->cmap->mapping,is,&newis);
6196: ISGetIndices(newis,&newRows);
6197: (*mat->ops->zerorowscolumns)(mat,numRows,newRows,diag,x,b);
6198: ISRestoreIndices(newis,&newRows);
6199: ISDestroy(&newis);
6200: ISDestroy(&is);
6201: PetscObjectStateIncrease((PetscObject)mat);
6202: return(0);
6203: }
6205: /*@
6206: MatZeroRowsColumnsLocalIS - Zeros all entries (except possibly the main diagonal)
6207: of a set of rows and columns of a matrix; using local numbering of rows.
6209: Collective on Mat
6211: Input Parameters:
6212: + mat - the matrix
6213: . is - index set of rows to remove
6214: . diag - value put in all diagonals of eliminated rows
6215: . x - optional vector of solutions for zeroed rows (other entries in vector are not used)
6216: - b - optional vector of right hand side, that will be adjusted by provided solution
6218: Notes:
6219: Before calling MatZeroRowsColumnsLocalIS(), the user must first set the
6220: local-to-global mapping by calling MatSetLocalToGlobalMapping().
6222: The user can set a value in the diagonal entry (or for the AIJ and
6223: row formats can optionally remove the main diagonal entry from the
6224: nonzero structure as well, by passing 0.0 as the final argument).
6226: Level: intermediate
6228: .seealso: MatZeroRowsIS(), MatZeroRowsColumns(), MatZeroRowsLocalIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(),
6229: MatZeroRowsColumnsLocal(), MatZeroRows(), MatZeroRowsColumnsIS(), MatZeroRowsColumnsStencil()
6230: @*/
6231: PetscErrorCode MatZeroRowsColumnsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b)
6232: {
6234: PetscInt numRows;
6235: const PetscInt *rows;
6241: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6242: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6243: MatCheckPreallocated(mat,1);
6245: ISGetLocalSize(is,&numRows);
6246: ISGetIndices(is,&rows);
6247: MatZeroRowsColumnsLocal(mat,numRows,rows,diag,x,b);
6248: ISRestoreIndices(is,&rows);
6249: return(0);
6250: }
6252: /*@C
6253: MatGetSize - Returns the numbers of rows and columns in a matrix.
6255: Not Collective
6257: Input Parameter:
6258: . mat - the matrix
6260: Output Parameters:
6261: + m - the number of global rows
6262: - n - the number of global columns
6264: Note: both output parameters can be NULL on input.
6266: Level: beginner
6268: .seealso: MatGetLocalSize()
6269: @*/
6270: PetscErrorCode MatGetSize(Mat mat,PetscInt *m,PetscInt *n)
6271: {
6274: if (m) *m = mat->rmap->N;
6275: if (n) *n = mat->cmap->N;
6276: return(0);
6277: }
6279: /*@C
6280: MatGetLocalSize - Returns the number of rows and columns in a matrix
6281: stored locally. This information may be implementation dependent, so
6282: use with care.
6284: Not Collective
6286: Input Parameters:
6287: . mat - the matrix
6289: Output Parameters:
6290: + m - the number of local rows
6291: - n - the number of local columns
6293: Note: both output parameters can be NULL on input.
6295: Level: beginner
6297: .seealso: MatGetSize()
6298: @*/
6299: PetscErrorCode MatGetLocalSize(Mat mat,PetscInt *m,PetscInt *n)
6300: {
6305: if (m) *m = mat->rmap->n;
6306: if (n) *n = mat->cmap->n;
6307: return(0);
6308: }
6310: /*@C
6311: MatGetOwnershipRangeColumn - Returns the range of matrix columns associated with rows of a vector one multiplies by that owned by
6312: this processor. (The columns of the "diagonal block")
6314: Not Collective, unless matrix has not been allocated, then collective on Mat
6316: Input Parameters:
6317: . mat - the matrix
6319: Output Parameters:
6320: + m - the global index of the first local column
6321: - n - one more than the global index of the last local column
6323: Notes:
6324: both output parameters can be NULL on input.
6326: Level: developer
6328: .seealso: MatGetOwnershipRange(), MatGetOwnershipRanges(), MatGetOwnershipRangesColumn()
6330: @*/
6331: PetscErrorCode MatGetOwnershipRangeColumn(Mat mat,PetscInt *m,PetscInt *n)
6332: {
6338: MatCheckPreallocated(mat,1);
6339: if (m) *m = mat->cmap->rstart;
6340: if (n) *n = mat->cmap->rend;
6341: return(0);
6342: }
6344: /*@C
6345: MatGetOwnershipRange - Returns the range of matrix rows owned by
6346: this processor, assuming that the matrix is laid out with the first
6347: n1 rows on the first processor, the next n2 rows on the second, etc.
6348: For certain parallel layouts this range may not be well defined.
6350: Not Collective
6352: Input Parameters:
6353: . mat - the matrix
6355: Output Parameters:
6356: + m - the global index of the first local row
6357: - n - one more than the global index of the last local row
6359: Note: Both output parameters can be NULL on input.
6360: $ This function requires that the matrix be preallocated. If you have not preallocated, consider using
6361: $ PetscSplitOwnership(MPI_Comm comm, PetscInt *n, PetscInt *N)
6362: $ and then MPI_Scan() to calculate prefix sums of the local sizes.
6364: Level: beginner
6366: .seealso: MatGetOwnershipRanges(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn(), PetscSplitOwnership(), PetscSplitOwnershipBlock()
6368: @*/
6369: PetscErrorCode MatGetOwnershipRange(Mat mat,PetscInt *m,PetscInt *n)
6370: {
6376: MatCheckPreallocated(mat,1);
6377: if (m) *m = mat->rmap->rstart;
6378: if (n) *n = mat->rmap->rend;
6379: return(0);
6380: }
6382: /*@C
6383: MatGetOwnershipRanges - Returns the range of matrix rows owned by
6384: each process
6386: Not Collective, unless matrix has not been allocated, then collective on Mat
6388: Input Parameters:
6389: . mat - the matrix
6391: Output Parameters:
6392: . ranges - start of each processors portion plus one more than the total length at the end
6394: Level: beginner
6396: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn()
6398: @*/
6399: PetscErrorCode MatGetOwnershipRanges(Mat mat,const PetscInt **ranges)
6400: {
6406: MatCheckPreallocated(mat,1);
6407: PetscLayoutGetRanges(mat->rmap,ranges);
6408: return(0);
6409: }
6411: /*@C
6412: MatGetOwnershipRangesColumn - Returns the range of matrix columns associated with rows of a vector one multiplies by that owned by
6413: this processor. (The columns of the "diagonal blocks" for each process)
6415: Not Collective, unless matrix has not been allocated, then collective on Mat
6417: Input Parameters:
6418: . mat - the matrix
6420: Output Parameters:
6421: . ranges - start of each processors portion plus one more then the total length at the end
6423: Level: beginner
6425: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRanges()
6427: @*/
6428: PetscErrorCode MatGetOwnershipRangesColumn(Mat mat,const PetscInt **ranges)
6429: {
6435: MatCheckPreallocated(mat,1);
6436: PetscLayoutGetRanges(mat->cmap,ranges);
6437: return(0);
6438: }
6440: /*@C
6441: MatGetOwnershipIS - Get row and column ownership as index sets
6443: Not Collective
6445: Input Arguments:
6446: . A - matrix of type Elemental
6448: Output Arguments:
6449: + rows - rows in which this process owns elements
6450: - cols - columns in which this process owns elements
6452: Level: intermediate
6454: .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatSetValues(), MATELEMENTAL
6455: @*/
6456: PetscErrorCode MatGetOwnershipIS(Mat A,IS *rows,IS *cols)
6457: {
6458: PetscErrorCode ierr,(*f)(Mat,IS*,IS*);
6461: MatCheckPreallocated(A,1);
6462: PetscObjectQueryFunction((PetscObject)A,"MatGetOwnershipIS_C",&f);
6463: if (f) {
6464: (*f)(A,rows,cols);
6465: } else { /* Create a standard row-based partition, each process is responsible for ALL columns in their row block */
6466: if (rows) {ISCreateStride(PETSC_COMM_SELF,A->rmap->n,A->rmap->rstart,1,rows);}
6467: if (cols) {ISCreateStride(PETSC_COMM_SELF,A->cmap->N,0,1,cols);}
6468: }
6469: return(0);
6470: }
6472: /*@C
6473: MatILUFactorSymbolic - Performs symbolic ILU factorization of a matrix.
6474: Uses levels of fill only, not drop tolerance. Use MatLUFactorNumeric()
6475: to complete the factorization.
6477: Collective on Mat
6479: Input Parameters:
6480: + mat - the matrix
6481: . row - row permutation
6482: . column - column permutation
6483: - info - structure containing
6484: $ levels - number of levels of fill.
6485: $ expected fill - as ratio of original fill.
6486: $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices
6487: missing diagonal entries)
6489: Output Parameters:
6490: . fact - new matrix that has been symbolically factored
6492: Notes:
6493: See Users-Manual: ch_mat for additional information about choosing the fill factor for better efficiency.
6495: Most users should employ the simplified KSP interface for linear solvers
6496: instead of working directly with matrix algebra routines such as this.
6497: See, e.g., KSPCreate().
6499: Level: developer
6501: .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor()
6502: MatGetOrdering(), MatFactorInfo
6504: Note: this uses the definition of level of fill as in Y. Saad, 2003
6506: Developer Note: fortran interface is not autogenerated as the f90
6507: interface defintion cannot be generated correctly [due to MatFactorInfo]
6509: References:
6510: Y. Saad, Iterative methods for sparse linear systems Philadelphia: Society for Industrial and Applied Mathematics, 2003
6511: @*/
6512: PetscErrorCode MatILUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info)
6513: {
6523: if (info->levels < 0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Levels of fill negative %D",(PetscInt)info->levels);
6524: if (info->fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %g",(double)info->fill);
6525: if (!(fact)->ops->ilufactorsymbolic) {
6526: MatSolverType spackage;
6527: MatFactorGetSolverType(fact,&spackage);
6528: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic ILU using solver package %s",((PetscObject)mat)->type_name,spackage);
6529: }
6530: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6531: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6532: MatCheckPreallocated(mat,2);
6534: PetscLogEventBegin(MAT_ILUFactorSymbolic,mat,row,col,0);
6535: (fact->ops->ilufactorsymbolic)(fact,mat,row,col,info);
6536: PetscLogEventEnd(MAT_ILUFactorSymbolic,mat,row,col,0);
6537: return(0);
6538: }
6540: /*@C
6541: MatICCFactorSymbolic - Performs symbolic incomplete
6542: Cholesky factorization for a symmetric matrix. Use
6543: MatCholeskyFactorNumeric() to complete the factorization.
6545: Collective on Mat
6547: Input Parameters:
6548: + mat - the matrix
6549: . perm - row and column permutation
6550: - info - structure containing
6551: $ levels - number of levels of fill.
6552: $ expected fill - as ratio of original fill.
6554: Output Parameter:
6555: . fact - the factored matrix
6557: Notes:
6558: Most users should employ the KSP interface for linear solvers
6559: instead of working directly with matrix algebra routines such as this.
6560: See, e.g., KSPCreate().
6562: Level: developer
6564: .seealso: MatCholeskyFactorNumeric(), MatCholeskyFactor(), MatFactorInfo
6566: Note: this uses the definition of level of fill as in Y. Saad, 2003
6568: Developer Note: fortran interface is not autogenerated as the f90
6569: interface defintion cannot be generated correctly [due to MatFactorInfo]
6571: References:
6572: Y. Saad, Iterative methods for sparse linear systems Philadelphia: Society for Industrial and Applied Mathematics, 2003
6573: @*/
6574: PetscErrorCode MatICCFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info)
6575: {
6584: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6585: if (info->levels < 0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Levels negative %D",(PetscInt) info->levels);
6586: if (info->fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %g",(double)info->fill);
6587: if (!(fact)->ops->iccfactorsymbolic) {
6588: MatSolverType spackage;
6589: MatFactorGetSolverType(fact,&spackage);
6590: SETERRQ2(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Matrix type %s symbolic ICC using solver package %s",((PetscObject)mat)->type_name,spackage);
6591: }
6592: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6593: MatCheckPreallocated(mat,2);
6595: PetscLogEventBegin(MAT_ICCFactorSymbolic,mat,perm,0,0);
6596: (fact->ops->iccfactorsymbolic)(fact,mat,perm,info);
6597: PetscLogEventEnd(MAT_ICCFactorSymbolic,mat,perm,0,0);
6598: return(0);
6599: }
6601: /*@C
6602: MatCreateSubMatrices - Extracts several submatrices from a matrix. If submat
6603: points to an array of valid matrices, they may be reused to store the new
6604: submatrices.
6606: Collective on Mat
6608: Input Parameters:
6609: + mat - the matrix
6610: . n - the number of submatrixes to be extracted (on this processor, may be zero)
6611: . irow, icol - index sets of rows and columns to extract
6612: - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
6614: Output Parameter:
6615: . submat - the array of submatrices
6617: Notes:
6618: MatCreateSubMatrices() can extract ONLY sequential submatrices
6619: (from both sequential and parallel matrices). Use MatCreateSubMatrix()
6620: to extract a parallel submatrix.
6622: Some matrix types place restrictions on the row and column
6623: indices, such as that they be sorted or that they be equal to each other.
6625: The index sets may not have duplicate entries.
6627: When extracting submatrices from a parallel matrix, each processor can
6628: form a different submatrix by setting the rows and columns of its
6629: individual index sets according to the local submatrix desired.
6631: When finished using the submatrices, the user should destroy
6632: them with MatDestroySubMatrices().
6634: MAT_REUSE_MATRIX can only be used when the nonzero structure of the
6635: original matrix has not changed from that last call to MatCreateSubMatrices().
6637: This routine creates the matrices in submat; you should NOT create them before
6638: calling it. It also allocates the array of matrix pointers submat.
6640: For BAIJ matrices the index sets must respect the block structure, that is if they
6641: request one row/column in a block, they must request all rows/columns that are in
6642: that block. For example, if the block size is 2 you cannot request just row 0 and
6643: column 0.
6645: Fortran Note:
6646: The Fortran interface is slightly different from that given below; it
6647: requires one to pass in as submat a Mat (integer) array of size at least n+1.
6649: Level: advanced
6652: .seealso: MatDestroySubMatrices(), MatCreateSubMatrix(), MatGetRow(), MatGetDiagonal(), MatReuse
6653: @*/
6654: PetscErrorCode MatCreateSubMatrices(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[])
6655: {
6657: PetscInt i;
6658: PetscBool eq;
6663: if (n) {
6668: }
6670: if (n && scall == MAT_REUSE_MATRIX) {
6673: }
6674: if (!mat->ops->createsubmatrices) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
6675: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6676: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6677: MatCheckPreallocated(mat,1);
6679: PetscLogEventBegin(MAT_CreateSubMats,mat,0,0,0);
6680: (*mat->ops->createsubmatrices)(mat,n,irow,icol,scall,submat);
6681: PetscLogEventEnd(MAT_CreateSubMats,mat,0,0,0);
6682: for (i=0; i<n; i++) {
6683: (*submat)[i]->factortype = MAT_FACTOR_NONE; /* in case in place factorization was previously done on submatrix */
6684: if (mat->symmetric || mat->structurally_symmetric || mat->hermitian) {
6685: ISEqual(irow[i],icol[i],&eq);
6686: if (eq) {
6687: if (mat->symmetric) {
6688: MatSetOption((*submat)[i],MAT_SYMMETRIC,PETSC_TRUE);
6689: } else if (mat->hermitian) {
6690: MatSetOption((*submat)[i],MAT_HERMITIAN,PETSC_TRUE);
6691: } else if (mat->structurally_symmetric) {
6692: MatSetOption((*submat)[i],MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);
6693: }
6694: }
6695: }
6696: }
6697: return(0);
6698: }
6700: /*@C
6701: MatCreateSubMatricesMPI - Extracts MPI submatrices across a sub communicator of mat (by pairs of IS that may live on subcomms).
6703: Collective on Mat
6705: Input Parameters:
6706: + mat - the matrix
6707: . n - the number of submatrixes to be extracted
6708: . irow, icol - index sets of rows and columns to extract
6709: - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
6711: Output Parameter:
6712: . submat - the array of submatrices
6714: Level: advanced
6717: .seealso: MatCreateSubMatrices(), MatCreateSubMatrix(), MatGetRow(), MatGetDiagonal(), MatReuse
6718: @*/
6719: PetscErrorCode MatCreateSubMatricesMPI(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[])
6720: {
6722: PetscInt i;
6723: PetscBool eq;
6728: if (n) {
6733: }
6735: if (n && scall == MAT_REUSE_MATRIX) {
6738: }
6739: if (!mat->ops->createsubmatricesmpi) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
6740: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6741: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6742: MatCheckPreallocated(mat,1);
6744: PetscLogEventBegin(MAT_CreateSubMats,mat,0,0,0);
6745: (*mat->ops->createsubmatricesmpi)(mat,n,irow,icol,scall,submat);
6746: PetscLogEventEnd(MAT_CreateSubMats,mat,0,0,0);
6747: for (i=0; i<n; i++) {
6748: if (mat->symmetric || mat->structurally_symmetric || mat->hermitian) {
6749: ISEqual(irow[i],icol[i],&eq);
6750: if (eq) {
6751: if (mat->symmetric) {
6752: MatSetOption((*submat)[i],MAT_SYMMETRIC,PETSC_TRUE);
6753: } else if (mat->hermitian) {
6754: MatSetOption((*submat)[i],MAT_HERMITIAN,PETSC_TRUE);
6755: } else if (mat->structurally_symmetric) {
6756: MatSetOption((*submat)[i],MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);
6757: }
6758: }
6759: }
6760: }
6761: return(0);
6762: }
6764: /*@C
6765: MatDestroyMatrices - Destroys an array of matrices.
6767: Collective on Mat
6769: Input Parameters:
6770: + n - the number of local matrices
6771: - mat - the matrices (note that this is a pointer to the array of matrices)
6773: Level: advanced
6775: Notes:
6776: Frees not only the matrices, but also the array that contains the matrices
6777: In Fortran will not free the array.
6779: .seealso: MatCreateSubMatrices() MatDestroySubMatrices()
6780: @*/
6781: PetscErrorCode MatDestroyMatrices(PetscInt n,Mat *mat[])
6782: {
6784: PetscInt i;
6787: if (!*mat) return(0);
6788: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n);
6791: for (i=0; i<n; i++) {
6792: MatDestroy(&(*mat)[i]);
6793: }
6795: /* memory is allocated even if n = 0 */
6796: PetscFree(*mat);
6797: return(0);
6798: }
6800: /*@C
6801: MatDestroySubMatrices - Destroys a set of matrices obtained with MatCreateSubMatrices().
6803: Collective on Mat
6805: Input Parameters:
6806: + n - the number of local matrices
6807: - mat - the matrices (note that this is a pointer to the array of matrices, just to match the calling
6808: sequence of MatCreateSubMatrices())
6810: Level: advanced
6812: Notes:
6813: Frees not only the matrices, but also the array that contains the matrices
6814: In Fortran will not free the array.
6816: .seealso: MatCreateSubMatrices()
6817: @*/
6818: PetscErrorCode MatDestroySubMatrices(PetscInt n,Mat *mat[])
6819: {
6821: Mat mat0;
6824: if (!*mat) return(0);
6825: /* mat[] is an array of length n+1, see MatCreateSubMatrices_xxx() */
6826: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n);
6829: mat0 = (*mat)[0];
6830: if (mat0 && mat0->ops->destroysubmatrices) {
6831: (mat0->ops->destroysubmatrices)(n,mat);
6832: } else {
6833: MatDestroyMatrices(n,mat);
6834: }
6835: return(0);
6836: }
6838: /*@C
6839: MatGetSeqNonzeroStructure - Extracts the sequential nonzero structure from a matrix.
6841: Collective on Mat
6843: Input Parameters:
6844: . mat - the matrix
6846: Output Parameter:
6847: . matstruct - the sequential matrix with the nonzero structure of mat
6849: Level: intermediate
6851: .seealso: MatDestroySeqNonzeroStructure(), MatCreateSubMatrices(), MatDestroyMatrices()
6852: @*/
6853: PetscErrorCode MatGetSeqNonzeroStructure(Mat mat,Mat *matstruct)
6854: {
6862: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6863: MatCheckPreallocated(mat,1);
6865: if (!mat->ops->getseqnonzerostructure) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Not for matrix type %s\n",((PetscObject)mat)->type_name);
6866: PetscLogEventBegin(MAT_GetSeqNonzeroStructure,mat,0,0,0);
6867: (*mat->ops->getseqnonzerostructure)(mat,matstruct);
6868: PetscLogEventEnd(MAT_GetSeqNonzeroStructure,mat,0,0,0);
6869: return(0);
6870: }
6872: /*@C
6873: MatDestroySeqNonzeroStructure - Destroys matrix obtained with MatGetSeqNonzeroStructure().
6875: Collective on Mat
6877: Input Parameters:
6878: . mat - the matrix (note that this is a pointer to the array of matrices, just to match the calling
6879: sequence of MatGetSequentialNonzeroStructure())
6881: Level: advanced
6883: Notes:
6884: Frees not only the matrices, but also the array that contains the matrices
6886: .seealso: MatGetSeqNonzeroStructure()
6887: @*/
6888: PetscErrorCode MatDestroySeqNonzeroStructure(Mat *mat)
6889: {
6894: MatDestroy(mat);
6895: return(0);
6896: }
6898: /*@
6899: MatIncreaseOverlap - Given a set of submatrices indicated by index sets,
6900: replaces the index sets by larger ones that represent submatrices with
6901: additional overlap.
6903: Collective on Mat
6905: Input Parameters:
6906: + mat - the matrix
6907: . n - the number of index sets
6908: . is - the array of index sets (these index sets will changed during the call)
6909: - ov - the additional overlap requested
6911: Options Database:
6912: . -mat_increase_overlap_scalable - use a scalable algorithm to compute the overlap (supported by MPIAIJ matrix)
6914: Level: developer
6917: .seealso: MatCreateSubMatrices()
6918: @*/
6919: PetscErrorCode MatIncreaseOverlap(Mat mat,PetscInt n,IS is[],PetscInt ov)
6920: {
6926: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n);
6927: if (n) {
6930: }
6931: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6932: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6933: MatCheckPreallocated(mat,1);
6935: if (!ov) return(0);
6936: if (!mat->ops->increaseoverlap) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
6937: PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);
6938: (*mat->ops->increaseoverlap)(mat,n,is,ov);
6939: PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);
6940: return(0);
6941: }
6944: PetscErrorCode MatIncreaseOverlapSplit_Single(Mat,IS*,PetscInt);
6946: /*@
6947: MatIncreaseOverlapSplit - Given a set of submatrices indicated by index sets across
6948: a sub communicator, replaces the index sets by larger ones that represent submatrices with
6949: additional overlap.
6951: Collective on Mat
6953: Input Parameters:
6954: + mat - the matrix
6955: . n - the number of index sets
6956: . is - the array of index sets (these index sets will changed during the call)
6957: - ov - the additional overlap requested
6959: Options Database:
6960: . -mat_increase_overlap_scalable - use a scalable algorithm to compute the overlap (supported by MPIAIJ matrix)
6962: Level: developer
6965: .seealso: MatCreateSubMatrices()
6966: @*/
6967: PetscErrorCode MatIncreaseOverlapSplit(Mat mat,PetscInt n,IS is[],PetscInt ov)
6968: {
6969: PetscInt i;
6975: if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n);
6976: if (n) {
6979: }
6980: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
6981: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
6982: MatCheckPreallocated(mat,1);
6983: if (!ov) return(0);
6984: PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);
6985: for(i=0; i<n; i++){
6986: MatIncreaseOverlapSplit_Single(mat,&is[i],ov);
6987: }
6988: PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);
6989: return(0);
6990: }
6995: /*@
6996: MatGetBlockSize - Returns the matrix block size.
6998: Not Collective
7000: Input Parameter:
7001: . mat - the matrix
7003: Output Parameter:
7004: . bs - block size
7006: Notes:
7007: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7009: If the block size has not been set yet this routine returns 1.
7011: Level: intermediate
7013: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSizes()
7014: @*/
7015: PetscErrorCode MatGetBlockSize(Mat mat,PetscInt *bs)
7016: {
7020: *bs = PetscAbs(mat->rmap->bs);
7021: return(0);
7022: }
7024: /*@
7025: MatGetBlockSizes - Returns the matrix block row and column sizes.
7027: Not Collective
7029: Input Parameter:
7030: . mat - the matrix
7032: Output Parameter:
7033: + rbs - row block size
7034: - cbs - column block size
7036: Notes:
7037: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7038: If you pass a different block size for the columns than the rows, the row block size determines the square block storage.
7040: If a block size has not been set yet this routine returns 1.
7042: Level: intermediate
7044: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSize(), MatSetBlockSizes()
7045: @*/
7046: PetscErrorCode MatGetBlockSizes(Mat mat,PetscInt *rbs, PetscInt *cbs)
7047: {
7052: if (rbs) *rbs = PetscAbs(mat->rmap->bs);
7053: if (cbs) *cbs = PetscAbs(mat->cmap->bs);
7054: return(0);
7055: }
7057: /*@
7058: MatSetBlockSize - Sets the matrix block size.
7060: Logically Collective on Mat
7062: Input Parameters:
7063: + mat - the matrix
7064: - bs - block size
7066: Notes:
7067: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7068: This must be called before MatSetUp() or MatXXXSetPreallocation() (or will default to 1) and the block size cannot be changed later.
7070: For MATMPIAIJ and MATSEQAIJ matrix formats, this function can be called at a later stage, provided that the specified block size
7071: is compatible with the matrix local sizes.
7073: Level: intermediate
7075: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes()
7076: @*/
7077: PetscErrorCode MatSetBlockSize(Mat mat,PetscInt bs)
7078: {
7084: MatSetBlockSizes(mat,bs,bs);
7085: return(0);
7086: }
7088: /*@
7089: MatSetVariableBlockSizes - Sets a diagonal blocks of the matrix that need not be of the same size
7091: Logically Collective on Mat
7093: Input Parameters:
7094: + mat - the matrix
7095: . nblocks - the number of blocks on this process
7096: - bsizes - the block sizes
7098: Notes:
7099: Currently used by PCVPBJACOBI for SeqAIJ matrices
7101: Level: intermediate
7103: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes(), MatGetVariableBlockSizes()
7104: @*/
7105: PetscErrorCode MatSetVariableBlockSizes(Mat mat,PetscInt nblocks,PetscInt *bsizes)
7106: {
7108: PetscInt i,ncnt = 0, nlocal;
7112: if (nblocks < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Number of local blocks must be great than or equal to zero");
7113: MatGetLocalSize(mat,&nlocal,NULL);
7114: for (i=0; i<nblocks; i++) ncnt += bsizes[i];
7115: 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);
7116: PetscFree(mat->bsizes);
7117: mat->nblocks = nblocks;
7118: PetscMalloc1(nblocks,&mat->bsizes);
7119: PetscArraycpy(mat->bsizes,bsizes,nblocks);
7120: return(0);
7121: }
7123: /*@C
7124: MatGetVariableBlockSizes - Gets a diagonal blocks of the matrix that need not be of the same size
7126: Logically Collective on Mat
7128: Input Parameters:
7129: . mat - the matrix
7131: Output Parameters:
7132: + nblocks - the number of blocks on this process
7133: - bsizes - the block sizes
7135: Notes: Currently not supported from Fortran
7137: Level: intermediate
7139: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes(), MatGetBlockSizes(), MatSetVariableBlockSizes()
7140: @*/
7141: PetscErrorCode MatGetVariableBlockSizes(Mat mat,PetscInt *nblocks,const PetscInt **bsizes)
7142: {
7145: *nblocks = mat->nblocks;
7146: *bsizes = mat->bsizes;
7147: return(0);
7148: }
7150: /*@
7151: MatSetBlockSizes - Sets the matrix block row and column sizes.
7153: Logically Collective on Mat
7155: Input Parameters:
7156: + mat - the matrix
7157: - rbs - row block size
7158: - cbs - column block size
7160: Notes:
7161: Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ. These formats ALWAYS have square block storage in the matrix.
7162: If you pass a different block size for the columns than the rows, the row block size determines the square block storage.
7163: This must be called before MatSetUp() or MatXXXSetPreallocation() (or will default to 1) and the block size cannot be changed later
7165: For MATMPIAIJ and MATSEQAIJ matrix formats, this function can be called at a later stage, provided that the specified block sizes
7166: are compatible with the matrix local sizes.
7168: The row and column block size determine the blocksize of the "row" and "column" vectors returned by MatCreateVecs().
7170: Level: intermediate
7172: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSize(), MatGetBlockSizes()
7173: @*/
7174: PetscErrorCode MatSetBlockSizes(Mat mat,PetscInt rbs,PetscInt cbs)
7175: {
7182: if (mat->ops->setblocksizes) {
7183: (*mat->ops->setblocksizes)(mat,rbs,cbs);
7184: }
7185: if (mat->rmap->refcnt) {
7186: ISLocalToGlobalMapping l2g = NULL;
7187: PetscLayout nmap = NULL;
7189: PetscLayoutDuplicate(mat->rmap,&nmap);
7190: if (mat->rmap->mapping) {
7191: ISLocalToGlobalMappingDuplicate(mat->rmap->mapping,&l2g);
7192: }
7193: PetscLayoutDestroy(&mat->rmap);
7194: mat->rmap = nmap;
7195: mat->rmap->mapping = l2g;
7196: }
7197: if (mat->cmap->refcnt) {
7198: ISLocalToGlobalMapping l2g = NULL;
7199: PetscLayout nmap = NULL;
7201: PetscLayoutDuplicate(mat->cmap,&nmap);
7202: if (mat->cmap->mapping) {
7203: ISLocalToGlobalMappingDuplicate(mat->cmap->mapping,&l2g);
7204: }
7205: PetscLayoutDestroy(&mat->cmap);
7206: mat->cmap = nmap;
7207: mat->cmap->mapping = l2g;
7208: }
7209: PetscLayoutSetBlockSize(mat->rmap,rbs);
7210: PetscLayoutSetBlockSize(mat->cmap,cbs);
7211: return(0);
7212: }
7214: /*@
7215: MatSetBlockSizesFromMats - Sets the matrix block row and column sizes to match a pair of matrices
7217: Logically Collective on Mat
7219: Input Parameters:
7220: + mat - the matrix
7221: . fromRow - matrix from which to copy row block size
7222: - fromCol - matrix from which to copy column block size (can be same as fromRow)
7224: Level: developer
7226: .seealso: MatCreateSeqBAIJ(), MatCreateBAIJ(), MatGetBlockSize(), MatSetBlockSizes()
7227: @*/
7228: PetscErrorCode MatSetBlockSizesFromMats(Mat mat,Mat fromRow,Mat fromCol)
7229: {
7236: if (fromRow->rmap->bs > 0) {PetscLayoutSetBlockSize(mat->rmap,fromRow->rmap->bs);}
7237: if (fromCol->cmap->bs > 0) {PetscLayoutSetBlockSize(mat->cmap,fromCol->cmap->bs);}
7238: return(0);
7239: }
7241: /*@
7242: MatResidual - Default routine to calculate the residual.
7244: Collective on Mat
7246: Input Parameters:
7247: + mat - the matrix
7248: . b - the right-hand-side
7249: - x - the approximate solution
7251: Output Parameter:
7252: . r - location to store the residual
7254: Level: developer
7256: .seealso: PCMGSetResidual()
7257: @*/
7258: PetscErrorCode MatResidual(Mat mat,Vec b,Vec x,Vec r)
7259: {
7268: MatCheckPreallocated(mat,1);
7269: PetscLogEventBegin(MAT_Residual,mat,0,0,0);
7270: if (!mat->ops->residual) {
7271: MatMult(mat,x,r);
7272: VecAYPX(r,-1.0,b);
7273: } else {
7274: (*mat->ops->residual)(mat,b,x,r);
7275: }
7276: PetscLogEventEnd(MAT_Residual,mat,0,0,0);
7277: return(0);
7278: }
7280: /*@C
7281: MatGetRowIJ - Returns the compressed row storage i and j indices for sequential matrices.
7283: Collective on Mat
7285: Input Parameters:
7286: + mat - the matrix
7287: . shift - 0 or 1 indicating we want the indices starting at 0 or 1
7288: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be symmetrized
7289: - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7290: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7291: always used.
7293: Output Parameters:
7294: + n - number of rows in the (possibly compressed) matrix
7295: . ia - the row pointers; that is ia[0] = 0, ia[row] = ia[row-1] + number of elements in that row of the matrix
7296: . ja - the column indices
7297: - done - indicates if the routine actually worked and returned appropriate ia[] and ja[] arrays; callers
7298: are responsible for handling the case when done == PETSC_FALSE and ia and ja are not set
7300: Level: developer
7302: Notes:
7303: You CANNOT change any of the ia[] or ja[] values.
7305: Use MatRestoreRowIJ() when you are finished accessing the ia[] and ja[] values.
7307: Fortran Notes:
7308: In Fortran use
7309: $
7310: $ PetscInt ia(1), ja(1)
7311: $ PetscOffset iia, jja
7312: $ call MatGetRowIJ(mat,shift,symmetric,inodecompressed,n,ia,iia,ja,jja,done,ierr)
7313: $ ! Access the ith and jth entries via ia(iia + i) and ja(jja + j)
7315: or
7316: $
7317: $ PetscInt, pointer :: ia(:),ja(:)
7318: $ call MatGetRowIJF90(mat,shift,symmetric,inodecompressed,n,ia,ja,done,ierr)
7319: $ ! Access the ith and jth entries via ia(i) and ja(j)
7321: .seealso: MatGetColumnIJ(), MatRestoreRowIJ(), MatSeqAIJGetArray()
7322: @*/
7323: PetscErrorCode MatGetRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7324: {
7334: MatCheckPreallocated(mat,1);
7335: if (!mat->ops->getrowij) *done = PETSC_FALSE;
7336: else {
7337: *done = PETSC_TRUE;
7338: PetscLogEventBegin(MAT_GetRowIJ,mat,0,0,0);
7339: (*mat->ops->getrowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7340: PetscLogEventEnd(MAT_GetRowIJ,mat,0,0,0);
7341: }
7342: return(0);
7343: }
7345: /*@C
7346: MatGetColumnIJ - Returns the compressed column storage i and j indices for sequential matrices.
7348: Collective on Mat
7350: Input Parameters:
7351: + mat - the matrix
7352: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7353: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7354: symmetrized
7355: . inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7356: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7357: always used.
7358: . n - number of columns in the (possibly compressed) matrix
7359: . ia - the column pointers; that is ia[0] = 0, ia[col] = i[col-1] + number of elements in that col of the matrix
7360: - ja - the row indices
7362: Output Parameters:
7363: . done - PETSC_TRUE or PETSC_FALSE, indicating whether the values have been returned
7365: Level: developer
7367: .seealso: MatGetRowIJ(), MatRestoreColumnIJ()
7368: @*/
7369: PetscErrorCode MatGetColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7370: {
7380: MatCheckPreallocated(mat,1);
7381: if (!mat->ops->getcolumnij) *done = PETSC_FALSE;
7382: else {
7383: *done = PETSC_TRUE;
7384: (*mat->ops->getcolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7385: }
7386: return(0);
7387: }
7389: /*@C
7390: MatRestoreRowIJ - Call after you are completed with the ia,ja indices obtained with
7391: MatGetRowIJ().
7393: Collective on Mat
7395: Input Parameters:
7396: + mat - the matrix
7397: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7398: . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7399: symmetrized
7400: . inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7401: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7402: always used.
7403: . n - size of (possibly compressed) matrix
7404: . ia - the row pointers
7405: - ja - the column indices
7407: Output Parameters:
7408: . done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned
7410: Note:
7411: This routine zeros out n, ia, and ja. This is to prevent accidental
7412: us of the array after it has been restored. If you pass NULL, it will
7413: not zero the pointers. Use of ia or ja after MatRestoreRowIJ() is invalid.
7415: Level: developer
7417: .seealso: MatGetRowIJ(), MatRestoreColumnIJ()
7418: @*/
7419: PetscErrorCode MatRestoreRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7420: {
7429: MatCheckPreallocated(mat,1);
7431: if (!mat->ops->restorerowij) *done = PETSC_FALSE;
7432: else {
7433: *done = PETSC_TRUE;
7434: (*mat->ops->restorerowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7435: if (n) *n = 0;
7436: if (ia) *ia = NULL;
7437: if (ja) *ja = NULL;
7438: }
7439: return(0);
7440: }
7442: /*@C
7443: MatRestoreColumnIJ - Call after you are completed with the ia,ja indices obtained with
7444: MatGetColumnIJ().
7446: Collective on Mat
7448: Input Parameters:
7449: + mat - the matrix
7450: . shift - 1 or zero indicating we want the indices starting at 0 or 1
7451: - symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be
7452: symmetrized
7453: - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the
7454: inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is
7455: always used.
7457: Output Parameters:
7458: + n - size of (possibly compressed) matrix
7459: . ia - the column pointers
7460: . ja - the row indices
7461: - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned
7463: Level: developer
7465: .seealso: MatGetColumnIJ(), MatRestoreRowIJ()
7466: @*/
7467: PetscErrorCode MatRestoreColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
7468: {
7477: MatCheckPreallocated(mat,1);
7479: if (!mat->ops->restorecolumnij) *done = PETSC_FALSE;
7480: else {
7481: *done = PETSC_TRUE;
7482: (*mat->ops->restorecolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);
7483: if (n) *n = 0;
7484: if (ia) *ia = NULL;
7485: if (ja) *ja = NULL;
7486: }
7487: return(0);
7488: }
7490: /*@C
7491: MatColoringPatch -Used inside matrix coloring routines that
7492: use MatGetRowIJ() and/or MatGetColumnIJ().
7494: Collective on Mat
7496: Input Parameters:
7497: + mat - the matrix
7498: . ncolors - max color value
7499: . n - number of entries in colorarray
7500: - colorarray - array indicating color for each column
7502: Output Parameters:
7503: . iscoloring - coloring generated using colorarray information
7505: Level: developer
7507: .seealso: MatGetRowIJ(), MatGetColumnIJ()
7509: @*/
7510: PetscErrorCode MatColoringPatch(Mat mat,PetscInt ncolors,PetscInt n,ISColoringValue colorarray[],ISColoring *iscoloring)
7511: {
7519: MatCheckPreallocated(mat,1);
7521: if (!mat->ops->coloringpatch) {
7522: ISColoringCreate(PetscObjectComm((PetscObject)mat),ncolors,n,colorarray,PETSC_OWN_POINTER,iscoloring);
7523: } else {
7524: (*mat->ops->coloringpatch)(mat,ncolors,n,colorarray,iscoloring);
7525: }
7526: return(0);
7527: }
7530: /*@
7531: MatSetUnfactored - Resets a factored matrix to be treated as unfactored.
7533: Logically Collective on Mat
7535: Input Parameter:
7536: . mat - the factored matrix to be reset
7538: Notes:
7539: This routine should be used only with factored matrices formed by in-place
7540: factorization via ILU(0) (or by in-place LU factorization for the MATSEQDENSE
7541: format). This option can save memory, for example, when solving nonlinear
7542: systems with a matrix-free Newton-Krylov method and a matrix-based, in-place
7543: ILU(0) preconditioner.
7545: Note that one can specify in-place ILU(0) factorization by calling
7546: .vb
7547: PCType(pc,PCILU);
7548: PCFactorSeUseInPlace(pc);
7549: .ve
7550: or by using the options -pc_type ilu -pc_factor_in_place
7552: In-place factorization ILU(0) can also be used as a local
7553: solver for the blocks within the block Jacobi or additive Schwarz
7554: methods (runtime option: -sub_pc_factor_in_place). See Users-Manual: ch_pc
7555: for details on setting local solver options.
7557: Most users should employ the simplified KSP interface for linear solvers
7558: instead of working directly with matrix algebra routines such as this.
7559: See, e.g., KSPCreate().
7561: Level: developer
7563: .seealso: PCFactorSetUseInPlace(), PCFactorGetUseInPlace()
7565: @*/
7566: PetscErrorCode MatSetUnfactored(Mat mat)
7567: {
7573: MatCheckPreallocated(mat,1);
7574: mat->factortype = MAT_FACTOR_NONE;
7575: if (!mat->ops->setunfactored) return(0);
7576: (*mat->ops->setunfactored)(mat);
7577: return(0);
7578: }
7580: /*MC
7581: MatDenseGetArrayF90 - Accesses a matrix array from Fortran90.
7583: Synopsis:
7584: MatDenseGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:,:)},integer ierr)
7586: Not collective
7588: Input Parameter:
7589: . x - matrix
7591: Output Parameters:
7592: + xx_v - the Fortran90 pointer to the array
7593: - ierr - error code
7595: Example of Usage:
7596: .vb
7597: PetscScalar, pointer xx_v(:,:)
7598: ....
7599: call MatDenseGetArrayF90(x,xx_v,ierr)
7600: a = xx_v(3)
7601: call MatDenseRestoreArrayF90(x,xx_v,ierr)
7602: .ve
7604: Level: advanced
7606: .seealso: MatDenseRestoreArrayF90(), MatDenseGetArray(), MatDenseRestoreArray(), MatSeqAIJGetArrayF90()
7608: M*/
7610: /*MC
7611: MatDenseRestoreArrayF90 - Restores a matrix array that has been
7612: accessed with MatDenseGetArrayF90().
7614: Synopsis:
7615: MatDenseRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:,:)},integer ierr)
7617: Not collective
7619: Input Parameters:
7620: + x - matrix
7621: - xx_v - the Fortran90 pointer to the array
7623: Output Parameter:
7624: . ierr - error code
7626: Example of Usage:
7627: .vb
7628: PetscScalar, pointer xx_v(:,:)
7629: ....
7630: call MatDenseGetArrayF90(x,xx_v,ierr)
7631: a = xx_v(3)
7632: call MatDenseRestoreArrayF90(x,xx_v,ierr)
7633: .ve
7635: Level: advanced
7637: .seealso: MatDenseGetArrayF90(), MatDenseGetArray(), MatDenseRestoreArray(), MatSeqAIJRestoreArrayF90()
7639: M*/
7642: /*MC
7643: MatSeqAIJGetArrayF90 - Accesses a matrix array from Fortran90.
7645: Synopsis:
7646: MatSeqAIJGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr)
7648: Not collective
7650: Input Parameter:
7651: . x - matrix
7653: Output Parameters:
7654: + xx_v - the Fortran90 pointer to the array
7655: - ierr - error code
7657: Example of Usage:
7658: .vb
7659: PetscScalar, pointer xx_v(:)
7660: ....
7661: call MatSeqAIJGetArrayF90(x,xx_v,ierr)
7662: a = xx_v(3)
7663: call MatSeqAIJRestoreArrayF90(x,xx_v,ierr)
7664: .ve
7666: Level: advanced
7668: .seealso: MatSeqAIJRestoreArrayF90(), MatSeqAIJGetArray(), MatSeqAIJRestoreArray(), MatDenseGetArrayF90()
7670: M*/
7672: /*MC
7673: MatSeqAIJRestoreArrayF90 - Restores a matrix array that has been
7674: accessed with MatSeqAIJGetArrayF90().
7676: Synopsis:
7677: MatSeqAIJRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr)
7679: Not collective
7681: Input Parameters:
7682: + x - matrix
7683: - xx_v - the Fortran90 pointer to the array
7685: Output Parameter:
7686: . ierr - error code
7688: Example of Usage:
7689: .vb
7690: PetscScalar, pointer xx_v(:)
7691: ....
7692: call MatSeqAIJGetArrayF90(x,xx_v,ierr)
7693: a = xx_v(3)
7694: call MatSeqAIJRestoreArrayF90(x,xx_v,ierr)
7695: .ve
7697: Level: advanced
7699: .seealso: MatSeqAIJGetArrayF90(), MatSeqAIJGetArray(), MatSeqAIJRestoreArray(), MatDenseRestoreArrayF90()
7701: M*/
7704: /*@
7705: MatCreateSubMatrix - Gets a single submatrix on the same number of processors
7706: as the original matrix.
7708: Collective on Mat
7710: Input Parameters:
7711: + mat - the original matrix
7712: . isrow - parallel IS containing the rows this processor should obtain
7713: . 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.
7714: - cll - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
7716: Output Parameter:
7717: . newmat - the new submatrix, of the same type as the old
7719: Level: advanced
7721: Notes:
7722: The submatrix will be able to be multiplied with vectors using the same layout as iscol.
7724: Some matrix types place restrictions on the row and column indices, such
7725: as that they be sorted or that they be equal to each other.
7727: The index sets may not have duplicate entries.
7729: The first time this is called you should use a cll of MAT_INITIAL_MATRIX,
7730: the MatCreateSubMatrix() routine will create the newmat for you. Any additional calls
7731: to this routine with a mat of the same nonzero structure and with a call of MAT_REUSE_MATRIX
7732: will reuse the matrix generated the first time. You should call MatDestroy() on newmat when
7733: you are finished using it.
7735: The communicator of the newly obtained matrix is ALWAYS the same as the communicator of
7736: the input matrix.
7738: If iscol is NULL then all columns are obtained (not supported in Fortran).
7740: Example usage:
7741: Consider the following 8x8 matrix with 34 non-zero values, that is
7742: assembled across 3 processors. Let's assume that proc0 owns 3 rows,
7743: proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown
7744: as follows:
7746: .vb
7747: 1 2 0 | 0 3 0 | 0 4
7748: Proc0 0 5 6 | 7 0 0 | 8 0
7749: 9 0 10 | 11 0 0 | 12 0
7750: -------------------------------------
7751: 13 0 14 | 15 16 17 | 0 0
7752: Proc1 0 18 0 | 19 20 21 | 0 0
7753: 0 0 0 | 22 23 0 | 24 0
7754: -------------------------------------
7755: Proc2 25 26 27 | 0 0 28 | 29 0
7756: 30 0 0 | 31 32 33 | 0 34
7757: .ve
7759: Suppose isrow = [0 1 | 4 | 6 7] and iscol = [1 2 | 3 4 5 | 6]. The resulting submatrix is
7761: .vb
7762: 2 0 | 0 3 0 | 0
7763: Proc0 5 6 | 7 0 0 | 8
7764: -------------------------------
7765: Proc1 18 0 | 19 20 21 | 0
7766: -------------------------------
7767: Proc2 26 27 | 0 0 28 | 29
7768: 0 0 | 31 32 33 | 0
7769: .ve
7772: .seealso: MatCreateSubMatrices()
7773: @*/
7774: PetscErrorCode MatCreateSubMatrix(Mat mat,IS isrow,IS iscol,MatReuse cll,Mat *newmat)
7775: {
7777: PetscMPIInt size;
7778: Mat *local;
7779: IS iscoltmp;
7788: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
7789: if (cll == MAT_IGNORE_MATRIX) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Cannot use MAT_IGNORE_MATRIX");
7791: MatCheckPreallocated(mat,1);
7792: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
7794: if (!iscol || isrow == iscol) {
7795: PetscBool stride;
7796: PetscMPIInt grabentirematrix = 0,grab;
7797: PetscObjectTypeCompare((PetscObject)isrow,ISSTRIDE,&stride);
7798: if (stride) {
7799: PetscInt first,step,n,rstart,rend;
7800: ISStrideGetInfo(isrow,&first,&step);
7801: if (step == 1) {
7802: MatGetOwnershipRange(mat,&rstart,&rend);
7803: if (rstart == first) {
7804: ISGetLocalSize(isrow,&n);
7805: if (n == rend-rstart) {
7806: grabentirematrix = 1;
7807: }
7808: }
7809: }
7810: }
7811: MPIU_Allreduce(&grabentirematrix,&grab,1,MPI_INT,MPI_MIN,PetscObjectComm((PetscObject)mat));
7812: if (grab) {
7813: PetscInfo(mat,"Getting entire matrix as submatrix\n");
7814: if (cll == MAT_INITIAL_MATRIX) {
7815: *newmat = mat;
7816: PetscObjectReference((PetscObject)mat);
7817: }
7818: return(0);
7819: }
7820: }
7822: if (!iscol) {
7823: ISCreateStride(PetscObjectComm((PetscObject)mat),mat->cmap->n,mat->cmap->rstart,1,&iscoltmp);
7824: } else {
7825: iscoltmp = iscol;
7826: }
7828: /* if original matrix is on just one processor then use submatrix generated */
7829: if (mat->ops->createsubmatrices && !mat->ops->createsubmatrix && size == 1 && cll == MAT_REUSE_MATRIX) {
7830: MatCreateSubMatrices(mat,1,&isrow,&iscoltmp,MAT_REUSE_MATRIX,&newmat);
7831: goto setproperties;
7832: } else if (mat->ops->createsubmatrices && !mat->ops->createsubmatrix && size == 1) {
7833: MatCreateSubMatrices(mat,1,&isrow,&iscoltmp,MAT_INITIAL_MATRIX,&local);
7834: *newmat = *local;
7835: PetscFree(local);
7836: goto setproperties;
7837: } else if (!mat->ops->createsubmatrix) {
7838: /* Create a new matrix type that implements the operation using the full matrix */
7839: PetscLogEventBegin(MAT_CreateSubMat,mat,0,0,0);
7840: switch (cll) {
7841: case MAT_INITIAL_MATRIX:
7842: MatCreateSubMatrixVirtual(mat,isrow,iscoltmp,newmat);
7843: break;
7844: case MAT_REUSE_MATRIX:
7845: MatSubMatrixVirtualUpdate(*newmat,mat,isrow,iscoltmp);
7846: break;
7847: default: SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_OUTOFRANGE,"Invalid MatReuse, must be either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX");
7848: }
7849: PetscLogEventEnd(MAT_CreateSubMat,mat,0,0,0);
7850: goto setproperties;
7851: }
7853: if (!mat->ops->createsubmatrix) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
7854: PetscLogEventBegin(MAT_CreateSubMat,mat,0,0,0);
7855: (*mat->ops->createsubmatrix)(mat,isrow,iscoltmp,cll,newmat);
7856: PetscLogEventEnd(MAT_CreateSubMat,mat,0,0,0);
7858: /* Propagate symmetry information for diagonal blocks */
7859: setproperties:
7860: if (isrow == iscoltmp) {
7861: if (mat->symmetric_set && mat->symmetric) {
7862: MatSetOption(*newmat,MAT_SYMMETRIC,PETSC_TRUE);
7863: }
7864: if (mat->structurally_symmetric_set && mat->structurally_symmetric) {
7865: MatSetOption(*newmat,MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);
7866: }
7867: if (mat->hermitian_set && mat->hermitian) {
7868: MatSetOption(*newmat,MAT_HERMITIAN,PETSC_TRUE);
7869: }
7870: if (mat->spd_set && mat->spd) {
7871: MatSetOption(*newmat,MAT_SPD,PETSC_TRUE);
7872: }
7873: }
7875: if (!iscol) {ISDestroy(&iscoltmp);}
7876: if (*newmat && cll == MAT_INITIAL_MATRIX) {PetscObjectStateIncrease((PetscObject)*newmat);}
7877: return(0);
7878: }
7880: /*@
7881: MatStashSetInitialSize - sets the sizes of the matrix stash, that is
7882: used during the assembly process to store values that belong to
7883: other processors.
7885: Not Collective
7887: Input Parameters:
7888: + mat - the matrix
7889: . size - the initial size of the stash.
7890: - bsize - the initial size of the block-stash(if used).
7892: Options Database Keys:
7893: + -matstash_initial_size <size> or <size0,size1,...sizep-1>
7894: - -matstash_block_initial_size <bsize> or <bsize0,bsize1,...bsizep-1>
7896: Level: intermediate
7898: Notes:
7899: The block-stash is used for values set with MatSetValuesBlocked() while
7900: the stash is used for values set with MatSetValues()
7902: Run with the option -info and look for output of the form
7903: MatAssemblyBegin_MPIXXX:Stash has MM entries, uses nn mallocs.
7904: to determine the appropriate value, MM, to use for size and
7905: MatAssemblyBegin_MPIXXX:Block-Stash has BMM entries, uses nn mallocs.
7906: to determine the value, BMM to use for bsize
7909: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashGetInfo()
7911: @*/
7912: PetscErrorCode MatStashSetInitialSize(Mat mat,PetscInt size, PetscInt bsize)
7913: {
7919: MatStashSetInitialSize_Private(&mat->stash,size);
7920: MatStashSetInitialSize_Private(&mat->bstash,bsize);
7921: return(0);
7922: }
7924: /*@
7925: MatInterpolateAdd - w = y + A*x or A'*x depending on the shape of
7926: the matrix
7928: Neighbor-wise Collective on Mat
7930: Input Parameters:
7931: + mat - the matrix
7932: . x,y - the vectors
7933: - w - where the result is stored
7935: Level: intermediate
7937: Notes:
7938: w may be the same vector as y.
7940: This allows one to use either the restriction or interpolation (its transpose)
7941: matrix to do the interpolation
7943: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict()
7945: @*/
7946: PetscErrorCode MatInterpolateAdd(Mat A,Vec x,Vec y,Vec w)
7947: {
7949: PetscInt M,N,Ny;
7957: MatCheckPreallocated(A,1);
7958: MatGetSize(A,&M,&N);
7959: VecGetSize(y,&Ny);
7960: if (M == Ny) {
7961: MatMultAdd(A,x,y,w);
7962: } else {
7963: MatMultTransposeAdd(A,x,y,w);
7964: }
7965: return(0);
7966: }
7968: /*@
7969: MatInterpolate - y = A*x or A'*x depending on the shape of
7970: the matrix
7972: Neighbor-wise Collective on Mat
7974: Input Parameters:
7975: + mat - the matrix
7976: - x,y - the vectors
7978: Level: intermediate
7980: Notes:
7981: This allows one to use either the restriction or interpolation (its transpose)
7982: matrix to do the interpolation
7984: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict()
7986: @*/
7987: PetscErrorCode MatInterpolate(Mat A,Vec x,Vec y)
7988: {
7990: PetscInt M,N,Ny;
7997: MatCheckPreallocated(A,1);
7998: MatGetSize(A,&M,&N);
7999: VecGetSize(y,&Ny);
8000: if (M == Ny) {
8001: MatMult(A,x,y);
8002: } else {
8003: MatMultTranspose(A,x,y);
8004: }
8005: return(0);
8006: }
8008: /*@
8009: MatRestrict - y = A*x or A'*x
8011: Neighbor-wise Collective on Mat
8013: Input Parameters:
8014: + mat - the matrix
8015: - x,y - the vectors
8017: Level: intermediate
8019: Notes:
8020: This allows one to use either the restriction or interpolation (its transpose)
8021: matrix to do the restriction
8023: .seealso: MatMultAdd(), MatMultTransposeAdd(), MatInterpolate()
8025: @*/
8026: PetscErrorCode MatRestrict(Mat A,Vec x,Vec y)
8027: {
8029: PetscInt M,N,Ny;
8036: MatCheckPreallocated(A,1);
8038: MatGetSize(A,&M,&N);
8039: VecGetSize(y,&Ny);
8040: if (M == Ny) {
8041: MatMult(A,x,y);
8042: } else {
8043: MatMultTranspose(A,x,y);
8044: }
8045: return(0);
8046: }
8048: /*@
8049: MatGetNullSpace - retrieves the null space of a matrix.
8051: Logically Collective on Mat
8053: Input Parameters:
8054: + mat - the matrix
8055: - nullsp - the null space object
8057: Level: developer
8059: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatSetNullSpace()
8060: @*/
8061: PetscErrorCode MatGetNullSpace(Mat mat, MatNullSpace *nullsp)
8062: {
8066: *nullsp = (mat->symmetric_set && mat->symmetric && !mat->nullsp) ? mat->transnullsp : mat->nullsp;
8067: return(0);
8068: }
8070: /*@
8071: MatSetNullSpace - attaches a null space to a matrix.
8073: Logically Collective on Mat
8075: Input Parameters:
8076: + mat - the matrix
8077: - nullsp - the null space object
8079: Level: advanced
8081: Notes:
8082: This null space is used by the linear solvers. Overwrites any previous null space that may have been attached
8084: For inconsistent singular systems (linear systems where the right hand side is not in the range of the operator) you also likely should
8085: call MatSetTransposeNullSpace(). This allows the linear system to be solved in a least squares sense.
8087: You can remove the null space by calling this routine with an nullsp of NULL
8090: The fundamental theorem of linear algebra (Gilbert Strang, Introduction to Applied Mathematics, page 72) states that
8091: 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).
8092: 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
8093: 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
8094: 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).
8096: Krylov solvers can produce the minimal norm solution to the least squares problem by utilizing MatNullSpaceRemove().
8098: 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
8099: routine also automatically calls MatSetTransposeNullSpace().
8101: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatGetNullSpace(), MatSetTransposeNullSpace(), MatGetTransposeNullSpace(), MatNullSpaceRemove()
8102: @*/
8103: PetscErrorCode MatSetNullSpace(Mat mat,MatNullSpace nullsp)
8104: {
8110: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8111: MatNullSpaceDestroy(&mat->nullsp);
8112: mat->nullsp = nullsp;
8113: if (mat->symmetric_set && mat->symmetric) {
8114: MatSetTransposeNullSpace(mat,nullsp);
8115: }
8116: return(0);
8117: }
8119: /*@
8120: MatGetTransposeNullSpace - retrieves the null space of the transpose of a matrix.
8122: Logically Collective on Mat
8124: Input Parameters:
8125: + mat - the matrix
8126: - nullsp - the null space object
8128: Level: developer
8130: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatSetTransposeNullSpace(), MatSetNullSpace(), MatGetNullSpace()
8131: @*/
8132: PetscErrorCode MatGetTransposeNullSpace(Mat mat, MatNullSpace *nullsp)
8133: {
8138: *nullsp = (mat->symmetric_set && mat->symmetric && !mat->transnullsp) ? mat->nullsp : mat->transnullsp;
8139: return(0);
8140: }
8142: /*@
8143: MatSetTransposeNullSpace - attaches a null space to a matrix.
8145: Logically Collective on Mat
8147: Input Parameters:
8148: + mat - the matrix
8149: - nullsp - the null space object
8151: Level: advanced
8153: Notes:
8154: 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.
8155: You must also call MatSetNullSpace()
8158: The fundamental theorem of linear algebra (Gilbert Strang, Introduction to Applied Mathematics, page 72) states that
8159: 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).
8160: 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
8161: 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
8162: 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).
8164: Krylov solvers can produce the minimal norm solution to the least squares problem by utilizing MatNullSpaceRemove().
8166: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNearNullSpace(), MatGetNullSpace(), MatSetNullSpace(), MatGetTransposeNullSpace(), MatNullSpaceRemove()
8167: @*/
8168: PetscErrorCode MatSetTransposeNullSpace(Mat mat,MatNullSpace nullsp)
8169: {
8175: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8176: MatNullSpaceDestroy(&mat->transnullsp);
8177: mat->transnullsp = nullsp;
8178: return(0);
8179: }
8181: /*@
8182: MatSetNearNullSpace - attaches a null space to a matrix, which is often the null space (rigid body modes) of the operator without boundary conditions
8183: This null space will be used to provide near null space vectors to a multigrid preconditioner built from this matrix.
8185: Logically Collective on Mat
8187: Input Parameters:
8188: + mat - the matrix
8189: - nullsp - the null space object
8191: Level: advanced
8193: Notes:
8194: Overwrites any previous near null space that may have been attached
8196: You can remove the null space by calling this routine with an nullsp of NULL
8198: .seealso: MatCreate(), MatNullSpaceCreate(), MatSetNullSpace(), MatNullSpaceCreateRigidBody(), MatGetNearNullSpace()
8199: @*/
8200: PetscErrorCode MatSetNearNullSpace(Mat mat,MatNullSpace nullsp)
8201: {
8208: MatCheckPreallocated(mat,1);
8209: if (nullsp) {PetscObjectReference((PetscObject)nullsp);}
8210: MatNullSpaceDestroy(&mat->nearnullsp);
8211: mat->nearnullsp = nullsp;
8212: return(0);
8213: }
8215: /*@
8216: MatGetNearNullSpace -Get null space attached with MatSetNearNullSpace()
8218: Not Collective
8220: Input Parameters:
8221: . mat - the matrix
8223: Output Parameters:
8224: . nullsp - the null space object, NULL if not set
8226: Level: developer
8228: .seealso: MatSetNearNullSpace(), MatGetNullSpace(), MatNullSpaceCreate()
8229: @*/
8230: PetscErrorCode MatGetNearNullSpace(Mat mat,MatNullSpace *nullsp)
8231: {
8236: MatCheckPreallocated(mat,1);
8237: *nullsp = mat->nearnullsp;
8238: return(0);
8239: }
8241: /*@C
8242: MatICCFactor - Performs in-place incomplete Cholesky factorization of matrix.
8244: Collective on Mat
8246: Input Parameters:
8247: + mat - the matrix
8248: . row - row/column permutation
8249: . fill - expected fill factor >= 1.0
8250: - level - level of fill, for ICC(k)
8252: Notes:
8253: Probably really in-place only when level of fill is zero, otherwise allocates
8254: new space to store factored matrix and deletes previous memory.
8256: Most users should employ the simplified KSP interface for linear solvers
8257: instead of working directly with matrix algebra routines such as this.
8258: See, e.g., KSPCreate().
8260: Level: developer
8263: .seealso: MatICCFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor()
8265: Developer Note: fortran interface is not autogenerated as the f90
8266: interface defintion cannot be generated correctly [due to MatFactorInfo]
8268: @*/
8269: PetscErrorCode MatICCFactor(Mat mat,IS row,const MatFactorInfo *info)
8270: {
8278: if (mat->rmap->N != mat->cmap->N) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONG,"matrix must be square");
8279: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
8280: if (mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
8281: if (!mat->ops->iccfactor) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8282: MatCheckPreallocated(mat,1);
8283: (*mat->ops->iccfactor)(mat,row,info);
8284: PetscObjectStateIncrease((PetscObject)mat);
8285: return(0);
8286: }
8288: /*@
8289: MatDiagonalScaleLocal - Scales columns of a matrix given the scaling values including the
8290: ghosted ones.
8292: Not Collective
8294: Input Parameters:
8295: + mat - the matrix
8296: - diag = the diagonal values, including ghost ones
8298: Level: developer
8300: Notes:
8301: Works only for MPIAIJ and MPIBAIJ matrices
8303: .seealso: MatDiagonalScale()
8304: @*/
8305: PetscErrorCode MatDiagonalScaleLocal(Mat mat,Vec diag)
8306: {
8308: PetscMPIInt size;
8315: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled");
8316: PetscLogEventBegin(MAT_Scale,mat,0,0,0);
8317: MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);
8318: if (size == 1) {
8319: PetscInt n,m;
8320: VecGetSize(diag,&n);
8321: MatGetSize(mat,0,&m);
8322: if (m == n) {
8323: MatDiagonalScale(mat,0,diag);
8324: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only supported for sequential matrices when no ghost points/periodic conditions");
8325: } else {
8326: PetscUseMethod(mat,"MatDiagonalScaleLocal_C",(Mat,Vec),(mat,diag));
8327: }
8328: PetscLogEventEnd(MAT_Scale,mat,0,0,0);
8329: PetscObjectStateIncrease((PetscObject)mat);
8330: return(0);
8331: }
8333: /*@
8334: MatGetInertia - Gets the inertia from a factored matrix
8336: Collective on Mat
8338: Input Parameter:
8339: . mat - the matrix
8341: Output Parameters:
8342: + nneg - number of negative eigenvalues
8343: . nzero - number of zero eigenvalues
8344: - npos - number of positive eigenvalues
8346: Level: advanced
8348: Notes:
8349: Matrix must have been factored by MatCholeskyFactor()
8352: @*/
8353: PetscErrorCode MatGetInertia(Mat mat,PetscInt *nneg,PetscInt *nzero,PetscInt *npos)
8354: {
8360: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
8361: if (!mat->assembled) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Numeric factor mat is not assembled");
8362: if (!mat->ops->getinertia) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8363: (*mat->ops->getinertia)(mat,nneg,nzero,npos);
8364: return(0);
8365: }
8367: /* ----------------------------------------------------------------*/
8368: /*@C
8369: MatSolves - Solves A x = b, given a factored matrix, for a collection of vectors
8371: Neighbor-wise Collective on Mats
8373: Input Parameters:
8374: + mat - the factored matrix
8375: - b - the right-hand-side vectors
8377: Output Parameter:
8378: . x - the result vectors
8380: Notes:
8381: The vectors b and x cannot be the same. I.e., one cannot
8382: call MatSolves(A,x,x).
8384: Notes:
8385: Most users should employ the simplified KSP interface for linear solvers
8386: instead of working directly with matrix algebra routines such as this.
8387: See, e.g., KSPCreate().
8389: Level: developer
8391: .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd(), MatSolve()
8392: @*/
8393: PetscErrorCode MatSolves(Mat mat,Vecs b,Vecs x)
8394: {
8400: if (x == b) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_IDN,"x and b must be different vectors");
8401: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix");
8402: if (!mat->rmap->N && !mat->cmap->N) return(0);
8404: if (!mat->ops->solves) SETERRQ1(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name);
8405: MatCheckPreallocated(mat,1);
8406: PetscLogEventBegin(MAT_Solves,mat,0,0,0);
8407: (*mat->ops->solves)(mat,b,x);
8408: PetscLogEventEnd(MAT_Solves,mat,0,0,0);
8409: return(0);
8410: }
8412: /*@
8413: MatIsSymmetric - Test whether a matrix is symmetric
8415: Collective on Mat
8417: Input Parameter:
8418: + A - the matrix to test
8419: - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact transpose)
8421: Output Parameters:
8422: . flg - the result
8424: Notes:
8425: For real numbers MatIsSymmetric() and MatIsHermitian() return identical results
8427: Level: intermediate
8429: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown()
8430: @*/
8431: PetscErrorCode MatIsSymmetric(Mat A,PetscReal tol,PetscBool *flg)
8432: {
8439: if (!A->symmetric_set) {
8440: if (!A->ops->issymmetric) {
8441: MatType mattype;
8442: MatGetType(A,&mattype);
8443: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for symmetric",mattype);
8444: }
8445: (*A->ops->issymmetric)(A,tol,flg);
8446: if (!tol) {
8447: A->symmetric_set = PETSC_TRUE;
8448: A->symmetric = *flg;
8449: if (A->symmetric) {
8450: A->structurally_symmetric_set = PETSC_TRUE;
8451: A->structurally_symmetric = PETSC_TRUE;
8452: }
8453: }
8454: } else if (A->symmetric) {
8455: *flg = PETSC_TRUE;
8456: } else if (!tol) {
8457: *flg = PETSC_FALSE;
8458: } else {
8459: if (!A->ops->issymmetric) {
8460: MatType mattype;
8461: MatGetType(A,&mattype);
8462: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for symmetric",mattype);
8463: }
8464: (*A->ops->issymmetric)(A,tol,flg);
8465: }
8466: return(0);
8467: }
8469: /*@
8470: MatIsHermitian - Test whether a matrix is Hermitian
8472: Collective on Mat
8474: Input Parameter:
8475: + A - the matrix to test
8476: - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact Hermitian)
8478: Output Parameters:
8479: . flg - the result
8481: Level: intermediate
8483: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(),
8484: MatIsSymmetricKnown(), MatIsSymmetric()
8485: @*/
8486: PetscErrorCode MatIsHermitian(Mat A,PetscReal tol,PetscBool *flg)
8487: {
8494: if (!A->hermitian_set) {
8495: if (!A->ops->ishermitian) {
8496: MatType mattype;
8497: MatGetType(A,&mattype);
8498: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for hermitian",mattype);
8499: }
8500: (*A->ops->ishermitian)(A,tol,flg);
8501: if (!tol) {
8502: A->hermitian_set = PETSC_TRUE;
8503: A->hermitian = *flg;
8504: if (A->hermitian) {
8505: A->structurally_symmetric_set = PETSC_TRUE;
8506: A->structurally_symmetric = PETSC_TRUE;
8507: }
8508: }
8509: } else if (A->hermitian) {
8510: *flg = PETSC_TRUE;
8511: } else if (!tol) {
8512: *flg = PETSC_FALSE;
8513: } else {
8514: if (!A->ops->ishermitian) {
8515: MatType mattype;
8516: MatGetType(A,&mattype);
8517: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for hermitian",mattype);
8518: }
8519: (*A->ops->ishermitian)(A,tol,flg);
8520: }
8521: return(0);
8522: }
8524: /*@
8525: MatIsSymmetricKnown - Checks the flag on the matrix to see if it is symmetric.
8527: Not Collective
8529: Input Parameter:
8530: . A - the matrix to check
8532: Output Parameters:
8533: + set - if the symmetric flag is set (this tells you if the next flag is valid)
8534: - flg - the result
8536: Level: advanced
8538: Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsSymmetric()
8539: if you want it explicitly checked
8541: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric()
8542: @*/
8543: PetscErrorCode MatIsSymmetricKnown(Mat A,PetscBool *set,PetscBool *flg)
8544: {
8549: if (A->symmetric_set) {
8550: *set = PETSC_TRUE;
8551: *flg = A->symmetric;
8552: } else {
8553: *set = PETSC_FALSE;
8554: }
8555: return(0);
8556: }
8558: /*@
8559: MatIsHermitianKnown - Checks the flag on the matrix to see if it is hermitian.
8561: Not Collective
8563: Input Parameter:
8564: . A - the matrix to check
8566: Output Parameters:
8567: + set - if the hermitian flag is set (this tells you if the next flag is valid)
8568: - flg - the result
8570: Level: advanced
8572: Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsHermitian()
8573: if you want it explicitly checked
8575: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric()
8576: @*/
8577: PetscErrorCode MatIsHermitianKnown(Mat A,PetscBool *set,PetscBool *flg)
8578: {
8583: if (A->hermitian_set) {
8584: *set = PETSC_TRUE;
8585: *flg = A->hermitian;
8586: } else {
8587: *set = PETSC_FALSE;
8588: }
8589: return(0);
8590: }
8592: /*@
8593: MatIsStructurallySymmetric - Test whether a matrix is structurally symmetric
8595: Collective on Mat
8597: Input Parameter:
8598: . A - the matrix to test
8600: Output Parameters:
8601: . flg - the result
8603: Level: intermediate
8605: .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsSymmetric(), MatSetOption()
8606: @*/
8607: PetscErrorCode MatIsStructurallySymmetric(Mat A,PetscBool *flg)
8608: {
8614: if (!A->structurally_symmetric_set) {
8615: if (!A->ops->isstructurallysymmetric) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Matrix does not support checking for structural symmetric");
8616: (*A->ops->isstructurallysymmetric)(A,&A->structurally_symmetric);
8618: A->structurally_symmetric_set = PETSC_TRUE;
8619: }
8620: *flg = A->structurally_symmetric;
8621: return(0);
8622: }
8624: /*@
8625: MatStashGetInfo - Gets how many values are currently in the matrix stash, i.e. need
8626: to be communicated to other processors during the MatAssemblyBegin/End() process
8628: Not collective
8630: Input Parameter:
8631: . vec - the vector
8633: Output Parameters:
8634: + nstash - the size of the stash
8635: . reallocs - the number of additional mallocs incurred.
8636: . bnstash - the size of the block stash
8637: - breallocs - the number of additional mallocs incurred.in the block stash
8639: Level: advanced
8641: .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashSetInitialSize()
8643: @*/
8644: PetscErrorCode MatStashGetInfo(Mat mat,PetscInt *nstash,PetscInt *reallocs,PetscInt *bnstash,PetscInt *breallocs)
8645: {
8649: MatStashGetInfo_Private(&mat->stash,nstash,reallocs);
8650: MatStashGetInfo_Private(&mat->bstash,bnstash,breallocs);
8651: return(0);
8652: }
8654: /*@C
8655: MatCreateVecs - Get vector(s) compatible with the matrix, i.e. with the same
8656: parallel layout
8658: Collective on Mat
8660: Input Parameter:
8661: . mat - the matrix
8663: Output Parameter:
8664: + right - (optional) vector that the matrix can be multiplied against
8665: - left - (optional) vector that the matrix vector product can be stored in
8667: Notes:
8668: 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().
8670: Notes:
8671: These are new vectors which are not owned by the Mat, they should be destroyed in VecDestroy() when no longer needed
8673: Level: advanced
8675: .seealso: MatCreate(), VecDestroy()
8676: @*/
8677: PetscErrorCode MatCreateVecs(Mat mat,Vec *right,Vec *left)
8678: {
8684: if (mat->ops->getvecs) {
8685: (*mat->ops->getvecs)(mat,right,left);
8686: } else {
8687: PetscInt rbs,cbs;
8688: MatGetBlockSizes(mat,&rbs,&cbs);
8689: if (right) {
8690: if (mat->cmap->n < 0) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"PetscLayout for columns not yet setup");
8691: VecCreate(PetscObjectComm((PetscObject)mat),right);
8692: VecSetSizes(*right,mat->cmap->n,PETSC_DETERMINE);
8693: VecSetBlockSize(*right,cbs);
8694: VecSetType(*right,mat->defaultvectype);
8695: PetscLayoutReference(mat->cmap,&(*right)->map);
8696: }
8697: if (left) {
8698: if (mat->rmap->n < 0) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"PetscLayout for rows not yet setup");
8699: VecCreate(PetscObjectComm((PetscObject)mat),left);
8700: VecSetSizes(*left,mat->rmap->n,PETSC_DETERMINE);
8701: VecSetBlockSize(*left,rbs);
8702: VecSetType(*left,mat->defaultvectype);
8703: PetscLayoutReference(mat->rmap,&(*left)->map);
8704: }
8705: }
8706: return(0);
8707: }
8709: /*@C
8710: MatFactorInfoInitialize - Initializes a MatFactorInfo data structure
8711: with default values.
8713: Not Collective
8715: Input Parameters:
8716: . info - the MatFactorInfo data structure
8719: Notes:
8720: The solvers are generally used through the KSP and PC objects, for example
8721: PCLU, PCILU, PCCHOLESKY, PCICC
8723: Level: developer
8725: .seealso: MatFactorInfo
8727: Developer Note: fortran interface is not autogenerated as the f90
8728: interface defintion cannot be generated correctly [due to MatFactorInfo]
8730: @*/
8732: PetscErrorCode MatFactorInfoInitialize(MatFactorInfo *info)
8733: {
8737: PetscMemzero(info,sizeof(MatFactorInfo));
8738: return(0);
8739: }
8741: /*@
8742: MatFactorSetSchurIS - Set indices corresponding to the Schur complement you wish to have computed
8744: Collective on Mat
8746: Input Parameters:
8747: + mat - the factored matrix
8748: - is - the index set defining the Schur indices (0-based)
8750: Notes:
8751: Call MatFactorSolveSchurComplement() or MatFactorSolveSchurComplementTranspose() after this call to solve a Schur complement system.
8753: You can call MatFactorGetSchurComplement() or MatFactorCreateSchurComplement() after this call.
8755: Level: developer
8757: .seealso: MatGetFactor(), MatFactorGetSchurComplement(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSolveSchurComplement(),
8758: MatFactorSolveSchurComplementTranspose(), MatFactorSolveSchurComplement()
8760: @*/
8761: PetscErrorCode MatFactorSetSchurIS(Mat mat,IS is)
8762: {
8763: PetscErrorCode ierr,(*f)(Mat,IS);
8771: if (!mat->factortype) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix");
8772: PetscObjectQueryFunction((PetscObject)mat,"MatFactorSetSchurIS_C",&f);
8773: 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");
8774: if (mat->schur) {
8775: MatDestroy(&mat->schur);
8776: }
8777: (*f)(mat,is);
8778: if (!mat->schur) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_PLIB,"Schur complement has not been created");
8779: MatFactorSetUpInPlaceSchur_Private(mat);
8780: return(0);
8781: }
8783: /*@
8784: MatFactorCreateSchurComplement - Create a Schur complement matrix object using Schur data computed during the factorization step
8786: Logically Collective on Mat
8788: Input Parameters:
8789: + F - the factored matrix obtained by calling MatGetFactor() from PETSc-MUMPS interface
8790: . S - location where to return the Schur complement, can be NULL
8791: - status - the status of the Schur complement matrix, can be NULL
8793: Notes:
8794: You must call MatFactorSetSchurIS() before calling this routine.
8796: The routine provides a copy of the Schur matrix stored within the solver data structures.
8797: The caller must destroy the object when it is no longer needed.
8798: If MatFactorInvertSchurComplement() has been called, the routine gets back the inverse.
8800: 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)
8802: Developer Notes:
8803: The reason this routine exists is because the representation of the Schur complement within the factor matrix may be different than a standard PETSc
8804: matrix representation and we normally do not want to use the time or memory to make a copy as a regular PETSc matrix.
8806: See MatCreateSchurComplement() or MatGetSchurComplement() for ways to create virtual or approximate Schur complements.
8808: Level: advanced
8810: References:
8812: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorGetSchurComplement(), MatFactorSchurStatus
8813: @*/
8814: PetscErrorCode MatFactorCreateSchurComplement(Mat F,Mat* S,MatFactorSchurStatus* status)
8815: {
8822: if (S) {
8823: PetscErrorCode (*f)(Mat,Mat*);
8825: PetscObjectQueryFunction((PetscObject)F,"MatFactorCreateSchurComplement_C",&f);
8826: if (f) {
8827: (*f)(F,S);
8828: } else {
8829: MatDuplicate(F->schur,MAT_COPY_VALUES,S);
8830: }
8831: }
8832: if (status) *status = F->schur_status;
8833: return(0);
8834: }
8836: /*@
8837: MatFactorGetSchurComplement - Gets access to a Schur complement matrix using the current Schur data within a factored matrix
8839: Logically Collective on Mat
8841: Input Parameters:
8842: + F - the factored matrix obtained by calling MatGetFactor()
8843: . *S - location where to return the Schur complement, can be NULL
8844: - status - the status of the Schur complement matrix, can be NULL
8846: Notes:
8847: You must call MatFactorSetSchurIS() before calling this routine.
8849: Schur complement mode is currently implemented for sequential matrices.
8850: The routine returns a the Schur Complement stored within the data strutures of the solver.
8851: If MatFactorInvertSchurComplement() has previously been called, the returned matrix is actually the inverse of the Schur complement.
8852: The returned matrix should not be destroyed; the caller should call MatFactorRestoreSchurComplement() when the object is no longer needed.
8854: Use MatFactorCreateSchurComplement() to create a copy of the Schur complement matrix that is within a factored matrix
8856: See MatCreateSchurComplement() or MatGetSchurComplement() for ways to create virtual or approximate Schur complements.
8858: Level: advanced
8860: References:
8862: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSchurStatus
8863: @*/
8864: PetscErrorCode MatFactorGetSchurComplement(Mat F,Mat* S,MatFactorSchurStatus* status)
8865: {
8870: if (S) *S = F->schur;
8871: if (status) *status = F->schur_status;
8872: return(0);
8873: }
8875: /*@
8876: MatFactorRestoreSchurComplement - Restore the Schur complement matrix object obtained from a call to MatFactorGetSchurComplement
8878: Logically Collective on Mat
8880: Input Parameters:
8881: + F - the factored matrix obtained by calling MatGetFactor()
8882: . *S - location where the Schur complement is stored
8883: - status - the status of the Schur complement matrix (see MatFactorSchurStatus)
8885: Notes:
8887: Level: advanced
8889: References:
8891: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorRestoreSchurComplement(), MatFactorCreateSchurComplement(), MatFactorSchurStatus
8892: @*/
8893: PetscErrorCode MatFactorRestoreSchurComplement(Mat F,Mat* S,MatFactorSchurStatus status)
8894: {
8899: if (S) {
8901: *S = NULL;
8902: }
8903: F->schur_status = status;
8904: MatFactorUpdateSchurStatus_Private(F);
8905: return(0);
8906: }
8908: /*@
8909: MatFactorSolveSchurComplementTranspose - Solve the transpose of the Schur complement system computed during the factorization step
8911: Logically Collective on Mat
8913: Input Parameters:
8914: + F - the factored matrix obtained by calling MatGetFactor()
8915: . rhs - location where the right hand side of the Schur complement system is stored
8916: - sol - location where the solution of the Schur complement system has to be returned
8918: Notes:
8919: The sizes of the vectors should match the size of the Schur complement
8921: Must be called after MatFactorSetSchurIS()
8923: Level: advanced
8925: References:
8927: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorSolveSchurComplement()
8928: @*/
8929: PetscErrorCode MatFactorSolveSchurComplementTranspose(Mat F, Vec rhs, Vec sol)
8930: {
8942: MatFactorFactorizeSchurComplement(F);
8943: switch (F->schur_status) {
8944: case MAT_FACTOR_SCHUR_FACTORED:
8945: MatSolveTranspose(F->schur,rhs,sol);
8946: break;
8947: case MAT_FACTOR_SCHUR_INVERTED:
8948: MatMultTranspose(F->schur,rhs,sol);
8949: break;
8950: default:
8951: SETERRQ1(PetscObjectComm((PetscObject)F),PETSC_ERR_SUP,"Unhandled MatFactorSchurStatus %D",F->schur_status);
8952: break;
8953: }
8954: return(0);
8955: }
8957: /*@
8958: MatFactorSolveSchurComplement - Solve the Schur complement system computed during the factorization step
8960: Logically Collective on Mat
8962: Input Parameters:
8963: + F - the factored matrix obtained by calling MatGetFactor()
8964: . rhs - location where the right hand side of the Schur complement system is stored
8965: - sol - location where the solution of the Schur complement system has to be returned
8967: Notes:
8968: The sizes of the vectors should match the size of the Schur complement
8970: Must be called after MatFactorSetSchurIS()
8972: Level: advanced
8974: References:
8976: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorSolveSchurComplementTranspose()
8977: @*/
8978: PetscErrorCode MatFactorSolveSchurComplement(Mat F, Vec rhs, Vec sol)
8979: {
8991: MatFactorFactorizeSchurComplement(F);
8992: switch (F->schur_status) {
8993: case MAT_FACTOR_SCHUR_FACTORED:
8994: MatSolve(F->schur,rhs,sol);
8995: break;
8996: case MAT_FACTOR_SCHUR_INVERTED:
8997: MatMult(F->schur,rhs,sol);
8998: break;
8999: default:
9000: SETERRQ1(PetscObjectComm((PetscObject)F),PETSC_ERR_SUP,"Unhandled MatFactorSchurStatus %D",F->schur_status);
9001: break;
9002: }
9003: return(0);
9004: }
9006: /*@
9007: MatFactorInvertSchurComplement - Invert the Schur complement matrix computed during the factorization step
9009: Logically Collective on Mat
9011: Input Parameters:
9012: . F - the factored matrix obtained by calling MatGetFactor()
9014: Notes:
9015: Must be called after MatFactorSetSchurIS().
9017: Call MatFactorGetSchurComplement() or MatFactorCreateSchurComplement() AFTER this call to actually compute the inverse and get access to it.
9019: Level: advanced
9021: References:
9023: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorGetSchurComplement(), MatFactorCreateSchurComplement()
9024: @*/
9025: PetscErrorCode MatFactorInvertSchurComplement(Mat F)
9026: {
9032: if (F->schur_status == MAT_FACTOR_SCHUR_INVERTED) return(0);
9033: MatFactorFactorizeSchurComplement(F);
9034: MatFactorInvertSchurComplement_Private(F);
9035: F->schur_status = MAT_FACTOR_SCHUR_INVERTED;
9036: return(0);
9037: }
9039: /*@
9040: MatFactorFactorizeSchurComplement - Factorize the Schur complement matrix computed during the factorization step
9042: Logically Collective on Mat
9044: Input Parameters:
9045: . F - the factored matrix obtained by calling MatGetFactor()
9047: Notes:
9048: Must be called after MatFactorSetSchurIS().
9050: Level: advanced
9052: References:
9054: .seealso: MatGetFactor(), MatFactorSetSchurIS(), MatFactorInvertSchurComplement()
9055: @*/
9056: PetscErrorCode MatFactorFactorizeSchurComplement(Mat F)
9057: {
9063: if (F->schur_status == MAT_FACTOR_SCHUR_INVERTED || F->schur_status == MAT_FACTOR_SCHUR_FACTORED) return(0);
9064: MatFactorFactorizeSchurComplement_Private(F);
9065: F->schur_status = MAT_FACTOR_SCHUR_FACTORED;
9066: return(0);
9067: }
9069: PetscErrorCode MatPtAP_Basic(Mat A,Mat P,MatReuse scall,PetscReal fill,Mat *C)
9070: {
9071: Mat AP;
9075: PetscInfo2(A,"Mat types %s and %s using basic PtAP\n",((PetscObject)A)->type_name,((PetscObject)P)->type_name);
9076: MatMatMult(A,P,MAT_INITIAL_MATRIX,PETSC_DEFAULT,&AP);
9077: MatTransposeMatMult(P,AP,scall,fill,C);
9078: MatDestroy(&AP);
9079: return(0);
9080: }
9082: /*@
9083: MatPtAP - Creates the matrix product C = P^T * A * P
9085: Neighbor-wise Collective on Mat
9087: Input Parameters:
9088: + A - the matrix
9089: . P - the projection matrix
9090: . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
9091: - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(P)), use PETSC_DEFAULT if you do not have a good estimate
9092: if the result is a dense matrix this is irrelevent
9094: Output Parameters:
9095: . C - the product matrix
9097: Notes:
9098: C will be created and must be destroyed by the user with MatDestroy().
9100: For matrix types without special implementation the function fallbacks to MatMatMult() followed by MatTransposeMatMult().
9102: Level: intermediate
9104: .seealso: MatPtAPSymbolic(), MatPtAPNumeric(), MatMatMult(), MatRARt()
9105: @*/
9106: PetscErrorCode MatPtAP(Mat A,Mat P,MatReuse scall,PetscReal fill,Mat *C)
9107: {
9109: PetscErrorCode (*fA)(Mat,Mat,MatReuse,PetscReal,Mat*);
9110: PetscErrorCode (*fP)(Mat,Mat,MatReuse,PetscReal,Mat*);
9111: PetscErrorCode (*ptap)(Mat,Mat,MatReuse,PetscReal,Mat*)=NULL;
9112: PetscBool sametype;
9117: MatCheckPreallocated(A,1);
9118: if (scall == MAT_INPLACE_MATRIX) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"Inplace product not supported");
9119: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9120: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9123: MatCheckPreallocated(P,2);
9124: if (!P->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9125: if (P->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9127: if (A->rmap->N != A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix A must be square, %D != %D",A->rmap->N,A->cmap->N);
9128: if (P->rmap->N != A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N);
9129: if (fill == PETSC_DEFAULT || fill == PETSC_DECIDE) fill = 2.0;
9130: if (fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Expected fill=%g must be >= 1.0",(double)fill);
9132: if (scall == MAT_REUSE_MATRIX) {
9136: PetscLogEventBegin(MAT_PtAP,A,P,0,0);
9137: PetscLogEventBegin(MAT_PtAPNumeric,A,P,0,0);
9138: if ((*C)->ops->ptapnumeric) {
9139: (*(*C)->ops->ptapnumeric)(A,P,*C);
9140: } else {
9141: MatPtAP_Basic(A,P,scall,fill,C);
9142: }
9143: PetscLogEventEnd(MAT_PtAPNumeric,A,P,0,0);
9144: PetscLogEventEnd(MAT_PtAP,A,P,0,0);
9145: return(0);
9146: }
9148: if (fill == PETSC_DEFAULT || fill == PETSC_DECIDE) fill = 2.0;
9149: if (fill < 1.0) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Expected fill=%g must be >= 1.0",(double)fill);
9151: fA = A->ops->ptap;
9152: fP = P->ops->ptap;
9153: PetscStrcmp(((PetscObject)A)->type_name,((PetscObject)P)->type_name,&sametype);
9154: if (fP == fA && sametype) {
9155: ptap = fA;
9156: } else {
9157: /* dispatch based on the type of A and P from their PetscObject's PetscFunctionLists. */
9158: char ptapname[256];
9159: PetscStrncpy(ptapname,"MatPtAP_",sizeof(ptapname));
9160: PetscStrlcat(ptapname,((PetscObject)A)->type_name,sizeof(ptapname));
9161: PetscStrlcat(ptapname,"_",sizeof(ptapname));
9162: PetscStrlcat(ptapname,((PetscObject)P)->type_name,sizeof(ptapname));
9163: PetscStrlcat(ptapname,"_C",sizeof(ptapname)); /* e.g., ptapname = "MatPtAP_seqdense_seqaij_C" */
9164: PetscObjectQueryFunction((PetscObject)P,ptapname,&ptap);
9165: }
9167: if (!ptap) ptap = MatPtAP_Basic;
9168: PetscLogEventBegin(MAT_PtAP,A,P,0,0);
9169: (*ptap)(A,P,scall,fill,C);
9170: PetscLogEventEnd(MAT_PtAP,A,P,0,0);
9171: if (A->symmetric_set && A->symmetric) {
9172: MatSetOption(*C,MAT_SYMMETRIC,PETSC_TRUE);
9173: }
9174: return(0);
9175: }
9177: /*@
9178: MatPtAPNumeric - Computes the matrix product C = P^T * A * P
9180: Neighbor-wise Collective on Mat
9182: Input Parameters:
9183: + A - the matrix
9184: - P - the projection matrix
9186: Output Parameters:
9187: . C - the product matrix
9189: Notes:
9190: C must have been created by calling MatPtAPSymbolic and must be destroyed by
9191: the user using MatDeatroy().
9193: This routine is currently only implemented for pairs of AIJ matrices and classes
9194: which inherit from AIJ. C will be of type MATAIJ.
9196: Level: intermediate
9198: .seealso: MatPtAP(), MatPtAPSymbolic(), MatMatMultNumeric()
9199: @*/
9200: PetscErrorCode MatPtAPNumeric(Mat A,Mat P,Mat C)
9201: {
9207: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9208: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9211: MatCheckPreallocated(P,2);
9212: if (!P->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9213: if (P->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9216: MatCheckPreallocated(C,3);
9217: if (C->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9218: if (P->cmap->N!=C->rmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap->N,C->rmap->N);
9219: if (P->rmap->N!=A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N);
9220: if (A->rmap->N!=A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap->N,A->cmap->N);
9221: if (P->cmap->N!=C->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap->N,C->cmap->N);
9222: MatCheckPreallocated(A,1);
9224: if (!C->ops->ptapnumeric) SETERRQ(PetscObjectComm((PetscObject)C),PETSC_ERR_ARG_WRONGSTATE,"MatPtAPNumeric implementation is missing. You should call MatPtAPSymbolic first");
9225: PetscLogEventBegin(MAT_PtAPNumeric,A,P,0,0);
9226: (*C->ops->ptapnumeric)(A,P,C);
9227: PetscLogEventEnd(MAT_PtAPNumeric,A,P,0,0);
9228: return(0);
9229: }
9231: /*@
9232: MatPtAPSymbolic - Creates the (i,j) structure of the matrix product C = P^T * A * P
9234: Neighbor-wise Collective on Mat
9236: Input Parameters:
9237: + A - the matrix
9238: - P - the projection matrix
9240: Output Parameters:
9241: . C - the (i,j) structure of the product matrix
9243: Notes:
9244: C will be created and must be destroyed by the user with MatDestroy().
9246: This routine is currently only implemented for pairs of SeqAIJ matrices and classes
9247: which inherit from SeqAIJ. C will be of type MATSEQAIJ. The product is computed using
9248: this (i,j) structure by calling MatPtAPNumeric().
9250: Level: intermediate
9252: .seealso: MatPtAP(), MatPtAPNumeric(), MatMatMultSymbolic()
9253: @*/
9254: PetscErrorCode MatPtAPSymbolic(Mat A,Mat P,PetscReal fill,Mat *C)
9255: {
9261: if (!A->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9262: if (A->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9263: if (fill <1.0) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Expected fill=%g must be >= 1.0",(double)fill);
9266: MatCheckPreallocated(P,2);
9267: if (!P->assembled) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix");
9268: if (P->factortype) SETERRQ(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
9271: if (P->rmap->N!=A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N);
9272: if (A->rmap->N!=A->cmap->N) SETERRQ2(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap->N,A->cmap->N);
9273: MatCheckPreallocated(A,1);
9275: if (!A->ops->ptapsymbolic) SETERRQ1(PetscObjectComm((PetscObject)A),PETSC_ERR_SUP,"MatType %s",((PetscObject)A)->type_name);
9276: PetscLogEventBegin(MAT_PtAPSymbolic,A,P,0,0);
9277: (*A->ops->ptapsymbolic)(A,P,fill,C);
9278: PetscLogEventEnd(MAT_PtAPSymbolic,A,P,0,0);
9280: /* MatSetBlockSize(*C,A->rmap->bs); NO! this is not always true -ma */
9281: return(0);
9282: }
9284: /*@
9285: MatRARt - Creates the matrix product C = R * A * R^T
9287: Neighbor-wise Collective on Mat
9289: Input Parameters:
9290: + A - the matrix
9291: . R - the projection matrix
9292: . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
9293: - fill - expected fill as ratio of nnz(C)/nnz(A), use PETSC_DEFAULT if you do not have a good estimate
9294: if the result is a dense matrix this is irrelevent
9296: Output Parameters:
9297: . C - the product matrix
9299: Notes: