Actual source code: aij.h
petsc-3.7.3 2016-08-01
5: #include <petsc/private/matimpl.h>
6: #include <petscctable.h>
8: /*
9: Struct header shared by SeqAIJ, SeqBAIJ and SeqSBAIJ matrix formats
10: */
11: #define SEQAIJHEADER(datatype) \
12: PetscBool roworiented; /* if true, row-oriented input, default */ \
13: PetscInt nonew; /* 1 don't add new nonzeros, -1 generate error on new */ \
14: PetscInt nounused; /* -1 generate error on unused space */ \
15: PetscBool singlemalloc; /* if true a, i, and j have been obtained with one big malloc */ \
16: PetscInt maxnz; /* allocated nonzeros */ \
17: PetscInt *imax; /* maximum space allocated for each row */ \
18: PetscInt *ilen; /* actual length of each row */ \
19: PetscBool free_imax_ilen; \
20: PetscInt reallocs; /* number of mallocs done during MatSetValues() \
21: as more values are set than were prealloced */\
22: PetscInt rmax; /* max nonzeros in any row */ \
23: PetscBool keepnonzeropattern; /* keeps matrix structure same in calls to MatZeroRows()*/ \
24: PetscBool ignorezeroentries; \
25: PetscBool free_ij; /* free the column indices j and row offsets i when the matrix is destroyed */ \
26: PetscBool free_a; /* free the numerical values when matrix is destroy */ \
27: Mat_CompressedRow compressedrow; /* use compressed row format */ \
28: PetscInt nz; /* nonzeros */ \
29: PetscInt *i; /* pointer to beginning of each row */ \
30: PetscInt *j; /* column values: j + i[k] - 1 is start of row k */ \
31: PetscInt *diag; /* pointers to diagonal elements */ \
32: PetscInt nonzerorowcnt; /* how many rows have nonzero entries */ \
33: PetscBool free_diag; \
34: datatype *a; /* nonzero elements */ \
35: PetscScalar *solve_work; /* work space used in MatSolve */ \
36: IS row, col, icol; /* index sets, used for reorderings */ \
37: PetscBool pivotinblocks; /* pivot inside factorization of each diagonal block */ \
38: Mat parent /* set if this matrix was formed with MatDuplicate(...,MAT_SHARE_NONZERO_PATTERN,....);
39: means that this shares some data structures with the parent including diag, ilen, imax, i, j */
41: typedef struct {
42: MatTransposeColoring matcoloring;
43: Mat Bt_den; /* dense matrix of B^T */
44: Mat ABt_den; /* dense matrix of A*B^T */
45: PetscBool usecoloring;
46: PetscErrorCode (*destroy)(Mat);
47: } Mat_MatMatTransMult;
49: typedef struct { /* for MatTransposeMatMult_SeqAIJ_SeqDense() */
50: Mat mA; /* maij matrix of A */
51: Vec bt,ct; /* vectors to hold locally transposed arrays of B and C */
52: PetscErrorCode (*destroy)(Mat);
53: } Mat_MatTransMatMult;
55: typedef struct {
56: PetscInt *api,*apj; /* symbolic structure of A*P */
57: PetscScalar *apa; /* temporary array for storing one row of A*P */
58: PetscErrorCode (*destroy)(Mat);
59: } Mat_PtAP;
61: typedef struct {
62: MatTransposeColoring matcoloring;
63: Mat Rt; /* sparse or dense matrix of R^T */
64: Mat RARt; /* dense matrix of R*A*R^T */
65: Mat ARt; /* A*R^T used for the case -matrart_color_art */
66: MatScalar *work; /* work array to store columns of A*R^T used in MatMatMatMultNumeric_SeqAIJ_SeqAIJ_SeqDense() */
67: PetscErrorCode (*destroy)(Mat);
68: } Mat_RARt;
70: typedef struct {
71: Mat BC; /* temp matrix for storing B*C */
72: PetscErrorCode (*destroy)(Mat);
73: } Mat_MatMatMatMult;
75: /*
76: MATSEQAIJ format - Compressed row storage (also called Yale sparse matrix
77: format) or compressed sparse row (CSR). The i[] and j[] arrays start at 0. For example,
78: j[i[k]+p] is the pth column in row k. Note that the diagonal
79: matrix elements are stored with the rest of the nonzeros (not separately).
80: */
82: /* Info about i-nodes (identical nodes) helper class for SeqAIJ */
83: typedef struct {
84: MatScalar *bdiag,*ibdiag,*ssor_work; /* diagonal blocks of matrix used for MatSOR_SeqAIJ_Inode() */
85: PetscInt bdiagsize; /* length of bdiag and ibdiag */
86: PetscBool ibdiagvalid; /* do ibdiag[] and bdiag[] contain the most recent values */
88: PetscBool use;
89: PetscInt node_count; /* number of inodes */
90: PetscInt *size; /* size of each inode */
91: PetscInt limit; /* inode limit */
92: PetscInt max_limit; /* maximum supported inode limit */
93: PetscBool checked; /* if inodes have been checked for */
94: PetscObjectState mat_nonzerostate; /* non-zero state when inodes were checked for */
95: } Mat_SeqAIJ_Inode;
97: PETSC_INTERN PetscErrorCode MatView_SeqAIJ_Inode(Mat,PetscViewer);
98: PETSC_INTERN PetscErrorCode MatAssemblyEnd_SeqAIJ_Inode(Mat,MatAssemblyType);
99: PETSC_INTERN PetscErrorCode MatDestroy_SeqAIJ_Inode(Mat);
100: PETSC_INTERN PetscErrorCode MatCreate_SeqAIJ_Inode(Mat);
101: PETSC_INTERN PetscErrorCode MatSetOption_SeqAIJ_Inode(Mat,MatOption,PetscBool);
102: PETSC_INTERN PetscErrorCode MatDuplicate_SeqAIJ_Inode(Mat,MatDuplicateOption,Mat*);
103: PETSC_INTERN PetscErrorCode MatDuplicateNoCreate_SeqAIJ(Mat,Mat,MatDuplicateOption,PetscBool);
104: PETSC_INTERN PetscErrorCode MatLUFactorNumeric_SeqAIJ_Inode_inplace(Mat,Mat,const MatFactorInfo*);
105: PETSC_INTERN PetscErrorCode MatLUFactorNumeric_SeqAIJ_Inode(Mat,Mat,const MatFactorInfo*);
107: typedef struct {
108: SEQAIJHEADER(MatScalar);
109: Mat_SeqAIJ_Inode inode;
110: MatScalar *saved_values; /* location for stashing nonzero values of matrix */
112: PetscScalar *idiag,*mdiag,*ssor_work; /* inverse of diagonal entries, diagonal values and workspace for Eisenstat trick */
113: PetscBool idiagvalid; /* current idiag[] and mdiag[] are valid */
114: PetscScalar *ibdiag; /* inverses of block diagonals */
115: PetscBool ibdiagvalid; /* inverses of block diagonals are valid. */
116: PetscScalar fshift,omega; /* last used omega and fshift */
118: ISColoring coloring; /* set with MatADSetColoring() used by MatADSetValues() */
120: PetscScalar *matmult_abdense; /* used by MatMatMult() */
121: Mat_PtAP *ptap; /* used by MatPtAP() */
122: Mat_MatMatMatMult *matmatmatmult; /* used by MatMatMatMult() */
123: Mat_RARt *rart; /* used by MatRARt() */
124: Mat_MatMatTransMult *abt; /* used by MatMatTransposeMult() */
125: } Mat_SeqAIJ;
127: /*
128: Frees the a, i, and j arrays from the XAIJ (AIJ, BAIJ, and SBAIJ) matrix types
129: */
132: PETSC_STATIC_INLINE PetscErrorCode MatSeqXAIJFreeAIJ(Mat AA,MatScalar **a,PetscInt **j,PetscInt **i)
133: {
135: Mat_SeqAIJ *A = (Mat_SeqAIJ*) AA->data;
136: if (A->singlemalloc) {
137: PetscFree3(*a,*j,*i);
138: } else {
139: if (A->free_a) {PetscFree(*a);}
140: if (A->free_ij) {PetscFree(*j);}
141: if (A->free_ij) {PetscFree(*i);}
142: }
143: return 0;
144: }
145: /*
146: Allocates larger a, i, and j arrays for the XAIJ (AIJ, BAIJ, and SBAIJ) matrix types
147: This is a macro because it takes the datatype as an argument which can be either a Mat or a MatScalar
148: */
149: #define MatSeqXAIJReallocateAIJ(Amat,AM,BS2,NROW,ROW,COL,RMAX,AA,AI,AJ,RP,AP,AIMAX,NONEW,datatype) \
150: if (NROW >= RMAX) { \
151: Mat_SeqAIJ *Ain = (Mat_SeqAIJ*)Amat->data; \
152: /* there is no extra room in row, therefore enlarge */ \
153: PetscInt CHUNKSIZE = 15,new_nz = AI[AM] + CHUNKSIZE,len,*new_i=0,*new_j=0; \
154: datatype *new_a; \
155: \
156: if (NONEW == -2) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"New nonzero at (%D,%D) caused a malloc\nUse MatSetOption(A, MAT_NEW_NONZERO_ALLOCATION_ERR, PETSC_FALSE) to turn off this check",ROW,COL); \
157: /* malloc new storage space */ \
158: PetscMalloc3(BS2*new_nz,&new_a,new_nz,&new_j,AM+1,&new_i); \
159: \
160: /* copy over old data into new slots */ \
161: for (ii=0; ii<ROW+1; ii++) {new_i[ii] = AI[ii];} \
162: for (ii=ROW+1; ii<AM+1; ii++) {new_i[ii] = AI[ii]+CHUNKSIZE;} \
163: PetscMemcpy(new_j,AJ,(AI[ROW]+NROW)*sizeof(PetscInt)); \
164: len = (new_nz - CHUNKSIZE - AI[ROW] - NROW); \
165: PetscMemcpy(new_j+AI[ROW]+NROW+CHUNKSIZE,AJ+AI[ROW]+NROW,len*sizeof(PetscInt)); \
166: PetscMemcpy(new_a,AA,BS2*(AI[ROW]+NROW)*sizeof(datatype)); \
167: PetscMemzero(new_a+BS2*(AI[ROW]+NROW),BS2*CHUNKSIZE*sizeof(datatype)); \
168: PetscMemcpy(new_a+BS2*(AI[ROW]+NROW+CHUNKSIZE),AA+BS2*(AI[ROW]+NROW),BS2*len*sizeof(datatype)); \
169: /* free up old matrix storage */ \
170: MatSeqXAIJFreeAIJ(A,&Ain->a,&Ain->j,&Ain->i); \
171: AA = new_a; \
172: Ain->a = (MatScalar*) new_a; \
173: AI = Ain->i = new_i; AJ = Ain->j = new_j; \
174: Ain->singlemalloc = PETSC_TRUE; \
175: \
176: RP = AJ + AI[ROW]; AP = AA + BS2*AI[ROW]; \
177: RMAX = AIMAX[ROW] = AIMAX[ROW] + CHUNKSIZE; \
178: Ain->maxnz += BS2*CHUNKSIZE; \
179: Ain->reallocs++; \
180: } \
183: PETSC_INTERN PetscErrorCode MatSeqAIJSetPreallocation_SeqAIJ(Mat,PetscInt,const PetscInt*);
184: PETSC_INTERN PetscErrorCode MatILUFactorSymbolic_SeqAIJ_inplace(Mat,Mat,IS,IS,const MatFactorInfo*);
185: PETSC_INTERN PetscErrorCode MatILUFactorSymbolic_SeqAIJ(Mat,Mat,IS,IS,const MatFactorInfo*);
186: PETSC_INTERN PetscErrorCode MatILUFactorSymbolic_SeqAIJ_ilu0(Mat,Mat,IS,IS,const MatFactorInfo*);
188: PETSC_INTERN PetscErrorCode MatICCFactorSymbolic_SeqAIJ_inplace(Mat,Mat,IS,const MatFactorInfo*);
189: PETSC_INTERN PetscErrorCode MatICCFactorSymbolic_SeqAIJ(Mat,Mat,IS,const MatFactorInfo*);
190: PETSC_INTERN PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ_inplace(Mat,Mat,IS,const MatFactorInfo*);
191: PETSC_INTERN PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ(Mat,Mat,IS,const MatFactorInfo*);
192: PETSC_INTERN PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ_inplace(Mat,Mat,const MatFactorInfo*);
193: PETSC_INTERN PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ(Mat,Mat,const MatFactorInfo*);
194: PETSC_INTERN PetscErrorCode MatDuplicate_SeqAIJ(Mat,MatDuplicateOption,Mat*);
195: PETSC_INTERN PetscErrorCode MatCopy_SeqAIJ(Mat,Mat,MatStructure);
196: PETSC_INTERN PetscErrorCode MatMissingDiagonal_SeqAIJ(Mat,PetscBool*,PetscInt*);
197: PETSC_INTERN PetscErrorCode MatMarkDiagonal_SeqAIJ(Mat);
198: PETSC_INTERN PetscErrorCode MatFindZeroDiagonals_SeqAIJ_Private(Mat,PetscInt*,PetscInt**);
200: PETSC_INTERN PetscErrorCode MatMult_SeqAIJ(Mat A,Vec,Vec);
201: PETSC_INTERN PetscErrorCode MatMultAdd_SeqAIJ(Mat A,Vec,Vec,Vec);
202: PETSC_INTERN PetscErrorCode MatMultTranspose_SeqAIJ(Mat A,Vec,Vec);
203: PETSC_INTERN PetscErrorCode MatMultTransposeAdd_SeqAIJ(Mat A,Vec,Vec,Vec);
204: PETSC_INTERN PetscErrorCode MatSOR_SeqAIJ(Mat,Vec,PetscReal,MatSORType,PetscReal,PetscInt,PetscInt,Vec);
206: PETSC_INTERN PetscErrorCode MatSetOption_SeqAIJ(Mat,MatOption,PetscBool);
207: PETSC_INTERN PetscErrorCode MatSetColoring_SeqAIJ(Mat,ISColoring);
208: PETSC_INTERN PetscErrorCode MatSetValuesAdifor_SeqAIJ(Mat,PetscInt,void*);
210: PETSC_INTERN PetscErrorCode MatGetSymbolicTranspose_SeqAIJ(Mat,PetscInt *[],PetscInt *[]);
211: PETSC_INTERN PetscErrorCode MatGetSymbolicTransposeReduced_SeqAIJ(Mat,PetscInt,PetscInt,PetscInt *[],PetscInt *[]);
212: PETSC_INTERN PetscErrorCode MatRestoreSymbolicTranspose_SeqAIJ(Mat,PetscInt *[],PetscInt *[]);
213: PETSC_INTERN PetscErrorCode MatTransposeSymbolic_SeqAIJ(Mat,Mat*);
214: PETSC_INTERN PetscErrorCode MatTranspose_SeqAIJ(Mat,MatReuse,Mat*);
215: PETSC_INTERN PetscErrorCode MatToSymmetricIJ_SeqAIJ(PetscInt,PetscInt*,PetscInt*,PetscBool,PetscInt,PetscInt,PetscInt**,PetscInt**);
216: PETSC_INTERN PetscErrorCode MatLUFactorSymbolic_SeqAIJ_inplace(Mat,Mat,IS,IS,const MatFactorInfo*);
217: PETSC_INTERN PetscErrorCode MatLUFactorSymbolic_SeqAIJ(Mat,Mat,IS,IS,const MatFactorInfo*);
218: PETSC_INTERN PetscErrorCode MatLUFactorNumeric_SeqAIJ_inplace(Mat,Mat,const MatFactorInfo*);
219: PETSC_INTERN PetscErrorCode MatLUFactorNumeric_SeqAIJ(Mat,Mat,const MatFactorInfo*);
220: PETSC_INTERN PetscErrorCode MatLUFactorNumeric_SeqAIJ_InplaceWithPerm(Mat,Mat,const MatFactorInfo*);
221: PETSC_INTERN PetscErrorCode MatLUFactor_SeqAIJ(Mat,IS,IS,const MatFactorInfo*);
222: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_inplace(Mat,Vec,Vec);
223: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ(Mat,Vec,Vec);
224: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_Inode_inplace(Mat,Vec,Vec);
225: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_Inode(Mat,Vec,Vec);
226: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering_inplace(Mat,Vec,Vec);
227: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering(Mat,Vec,Vec);
228: PETSC_INTERN PetscErrorCode MatSolve_SeqAIJ_InplaceWithPerm(Mat,Vec,Vec);
229: PETSC_INTERN PetscErrorCode MatSolveAdd_SeqAIJ_inplace(Mat,Vec,Vec,Vec);
230: PETSC_INTERN PetscErrorCode MatSolveAdd_SeqAIJ(Mat,Vec,Vec,Vec);
231: PETSC_INTERN PetscErrorCode MatSolveTranspose_SeqAIJ_inplace(Mat,Vec,Vec);
232: PETSC_INTERN PetscErrorCode MatSolveTranspose_SeqAIJ(Mat,Vec,Vec);
233: PETSC_INTERN PetscErrorCode MatSolveTransposeAdd_SeqAIJ_inplace(Mat,Vec,Vec,Vec);
234: PETSC_INTERN PetscErrorCode MatSolveTransposeAdd_SeqAIJ(Mat,Vec,Vec,Vec);
235: PETSC_INTERN PetscErrorCode MatMatSolve_SeqAIJ_inplace(Mat,Mat,Mat);
236: PETSC_INTERN PetscErrorCode MatMatSolve_SeqAIJ(Mat,Mat,Mat);
237: PETSC_INTERN PetscErrorCode MatEqual_SeqAIJ(Mat,Mat,PetscBool*);
238: PETSC_INTERN PetscErrorCode MatFDColoringCreate_SeqXAIJ(Mat,ISColoring,MatFDColoring);
239: PETSC_INTERN PetscErrorCode MatFDColoringSetUp_SeqXAIJ(Mat,ISColoring,MatFDColoring);
240: PETSC_INTERN PetscErrorCode MatFDColoringSetUpBlocked_AIJ_Private(Mat,MatFDColoring,PetscInt);
241: PETSC_INTERN PetscErrorCode MatFDColoringApply_AIJ(Mat,MatFDColoring,Vec,void*);
242: PETSC_INTERN PetscErrorCode MatLoad_SeqAIJ(Mat,PetscViewer);
243: PETSC_INTERN PetscErrorCode RegisterApplyPtAPRoutines_Private(Mat);
245: PETSC_INTERN PetscErrorCode MatMatMult_SeqAIJ_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
246: PETSC_INTERN PetscErrorCode MatMatMultSymbolic_SeqAIJ_SeqAIJ(Mat,Mat,PetscReal,Mat*);
247: PETSC_INTERN PetscErrorCode MatMatMultSymbolic_SeqAIJ_SeqAIJ_Scalable(Mat,Mat,PetscReal,Mat*);
248: PETSC_INTERN PetscErrorCode MatMatMultSymbolic_SeqAIJ_SeqAIJ_Scalable_fast(Mat,Mat,PetscReal,Mat*);
249: PETSC_INTERN PetscErrorCode MatMatMultSymbolic_SeqAIJ_SeqAIJ_Heap(Mat,Mat,PetscReal,Mat*);
250: PETSC_INTERN PetscErrorCode MatMatMultSymbolic_SeqAIJ_SeqAIJ_BTHeap(Mat,Mat,PetscReal,Mat*);
251: PETSC_INTERN PetscErrorCode MatMatMultNumeric_SeqAIJ_SeqAIJ(Mat,Mat,Mat);
252: PETSC_INTERN PetscErrorCode MatMatMultNumeric_SeqAIJ_SeqAIJ_Scalable(Mat,Mat,Mat);
254: PETSC_INTERN PetscErrorCode MatPtAP_SeqAIJ_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
255: PETSC_INTERN PetscErrorCode MatPtAPSymbolic_SeqAIJ_SeqAIJ_DenseAxpy(Mat,Mat,PetscReal,Mat*);
256: PETSC_INTERN PetscErrorCode MatPtAPSymbolic_SeqAIJ_SeqAIJ_SparseAxpy(Mat,Mat,PetscReal,Mat*);
257: PETSC_INTERN PetscErrorCode MatPtAPNumeric_SeqAIJ_SeqAIJ(Mat,Mat,Mat);
258: PETSC_INTERN PetscErrorCode MatPtAPNumeric_SeqAIJ_SeqAIJ_SparseAxpy(Mat,Mat,Mat);
260: PETSC_INTERN PetscErrorCode MatRARtSymbolic_SeqAIJ_SeqAIJ(Mat,Mat,PetscReal,Mat*);
261: PETSC_INTERN PetscErrorCode MatRARtSymbolic_SeqAIJ_SeqAIJ_matmattransposemult(Mat,Mat,PetscReal,Mat*);
262: PETSC_INTERN PetscErrorCode MatRARtSymbolic_SeqAIJ_SeqAIJ_colorrart(Mat,Mat,PetscReal,Mat*);
263: PETSC_INTERN PetscErrorCode MatRARtNumeric_SeqAIJ_SeqAIJ(Mat,Mat,Mat);
264: PETSC_INTERN PetscErrorCode MatRARtNumeric_SeqAIJ_SeqAIJ_matmattransposemult(Mat,Mat,Mat);
265: PETSC_INTERN PetscErrorCode MatRARtNumeric_SeqAIJ_SeqAIJ_colorrart(Mat,Mat,Mat);
267: PETSC_INTERN PetscErrorCode MatTransposeMatMult_SeqAIJ_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
268: PETSC_INTERN PetscErrorCode MatTransposeMatMultSymbolic_SeqAIJ_SeqAIJ(Mat,Mat,PetscReal,Mat*);
269: PETSC_INTERN PetscErrorCode MatTransposeMatMultNumeric_SeqAIJ_SeqAIJ(Mat,Mat,Mat);
271: PETSC_INTERN PetscErrorCode MatTransposeMatMult_SeqAIJ_SeqDense(Mat,Mat,MatReuse,PetscReal,Mat*);
272: PETSC_INTERN PetscErrorCode MatTransposeMatMultSymbolic_SeqAIJ_SeqDense(Mat,Mat,PetscReal,Mat*);
273: PETSC_INTERN PetscErrorCode MatTransposeMatMultNumeric_SeqAIJ_SeqDense(Mat,Mat,Mat);
275: PETSC_INTERN PetscErrorCode MatMatTransposeMult_SeqAIJ_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
276: PETSC_INTERN PetscErrorCode MatMatTransposeMultSymbolic_SeqAIJ_SeqAIJ(Mat,Mat,PetscReal,Mat*);
277: PETSC_INTERN PetscErrorCode MatMatTransposeMultNumeric_SeqAIJ_SeqAIJ(Mat,Mat,Mat);
278: PETSC_INTERN PetscErrorCode MatTransposeColoringCreate_SeqAIJ(Mat,ISColoring,MatTransposeColoring);
279: PETSC_INTERN PetscErrorCode MatTransColoringApplySpToDen_SeqAIJ(MatTransposeColoring,Mat,Mat);
280: PETSC_INTERN PetscErrorCode MatTransColoringApplyDenToSp_SeqAIJ(MatTransposeColoring,Mat,Mat);
282: PETSC_INTERN PetscErrorCode MatMatMatMult_SeqAIJ_SeqAIJ_SeqAIJ(Mat,Mat,Mat,MatReuse,PetscReal,Mat*);
283: PETSC_INTERN PetscErrorCode MatMatMatMultSymbolic_SeqAIJ_SeqAIJ_SeqAIJ(Mat,Mat,Mat,PetscReal,Mat*);
284: PETSC_INTERN PetscErrorCode MatMatMatMultNumeric_SeqAIJ_SeqAIJ_SeqAIJ(Mat,Mat,Mat,Mat);
286: PETSC_INTERN PetscErrorCode MatSetValues_SeqAIJ(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
287: PETSC_INTERN PetscErrorCode MatGetRow_SeqAIJ(Mat,PetscInt,PetscInt*,PetscInt**,PetscScalar**);
288: PETSC_INTERN PetscErrorCode MatRestoreRow_SeqAIJ(Mat,PetscInt,PetscInt*,PetscInt**,PetscScalar**);
289: PETSC_INTERN PetscErrorCode MatAXPY_SeqAIJ(Mat,PetscScalar,Mat,MatStructure);
290: PETSC_INTERN PetscErrorCode MatGetRowIJ_SeqAIJ(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool*);
291: PETSC_INTERN PetscErrorCode MatRestoreRowIJ_SeqAIJ(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool*);
292: PETSC_INTERN PetscErrorCode MatGetColumnIJ_SeqAIJ(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool*);
293: PETSC_INTERN PetscErrorCode MatRestoreColumnIJ_SeqAIJ(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool*);
294: PETSC_INTERN PetscErrorCode MatGetColumnIJ_SeqAIJ_Color(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscInt *[],PetscBool*);
295: PETSC_INTERN PetscErrorCode MatRestoreColumnIJ_SeqAIJ_Color(Mat,PetscInt,PetscBool,PetscBool,PetscInt*,const PetscInt *[],const PetscInt *[],PetscInt *[],PetscBool*);
296: PETSC_INTERN PetscErrorCode MatDestroy_SeqAIJ(Mat);
297: PETSC_INTERN PetscErrorCode MatSetUp_SeqAIJ(Mat);
298: PETSC_INTERN PetscErrorCode MatView_SeqAIJ(Mat,PetscViewer);
300: PETSC_INTERN PetscErrorCode MatSeqAIJInvalidateDiagonal(Mat);
301: PETSC_INTERN PetscErrorCode MatSeqAIJInvalidateDiagonal_Inode(Mat);
302: PETSC_INTERN PetscErrorCode MatSeqAIJCheckInode(Mat);
303: PETSC_INTERN PetscErrorCode MatSeqAIJCheckInode_FactorLU(Mat);
305: PETSC_INTERN PetscErrorCode MatAXPYGetPreallocation_SeqAIJ(Mat,Mat,PetscInt*);
307: PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqSBAIJ(Mat,MatType,MatReuse,Mat*);
308: PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqBAIJ(Mat,MatType,MatReuse,Mat*);
309: PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqAIJPERM(Mat,MatType,MatReuse,Mat*);
310: PETSC_INTERN PetscErrorCode MatReorderForNonzeroDiagonal_SeqAIJ(Mat,PetscReal,IS,IS);
311: PETSC_INTERN PetscErrorCode MatMatMult_SeqDense_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
312: PETSC_INTERN PetscErrorCode MatRARt_SeqAIJ_SeqAIJ(Mat,Mat,MatReuse,PetscReal,Mat*);
313: PETSC_EXTERN PetscErrorCode MatCreate_SeqAIJ(Mat);
314: PETSC_INTERN PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat,MatAssemblyType);
315: PETSC_INTERN PetscErrorCode MatDestroy_SeqAIJ(Mat);
317: PETSC_INTERN PetscErrorCode MatAXPYGetPreallocation_SeqX_private(PetscInt,const PetscInt*,const PetscInt*,const PetscInt*,const PetscInt*,PetscInt*);
318: PETSC_INTERN PetscErrorCode MatCreateMPIMatConcatenateSeqMat_SeqAIJ(MPI_Comm,Mat,PetscInt,MatReuse,Mat*);
319: PETSC_INTERN PetscErrorCode MatCreateMPIMatConcatenateSeqMat_MPIAIJ(MPI_Comm,Mat,PetscInt,MatReuse,Mat*);
321: PETSC_INTERN PetscErrorCode MatSetSeqMat_SeqAIJ(Mat,IS,IS,MatStructure,Mat);
323: /*
324: PetscSparseDenseMinusDot - The inner kernel of triangular solves and Gauss-Siedel smoothing. \sum_i xv[i] * r[xi[i]] for CSR storage
326: Input Parameters:
327: + nnz - the number of entries
328: . r - the array of vector values
329: . xv - the matrix values for the row
330: - xi - the column indices of the nonzeros in the row
332: Output Parameter:
333: . sum - negative the sum of results
335: PETSc compile flags:
336: + PETSC_KERNEL_USE_UNROLL_4 - don't use this; it changes nnz and hence is WRONG
337: - PETSC_KERNEL_USE_UNROLL_2 -
339: .seealso: PetscSparseDensePlusDot()
341: */
342: #if defined(PETSC_KERNEL_USE_UNROLL_4)
343: #define PetscSparseDenseMinusDot(sum,r,xv,xi,nnz) { \
344: if (nnz > 0) { \
345: switch (nnz & 0x3) { \
346: case 3: sum -= *xv++ *r[*xi++]; \
347: case 2: sum -= *xv++ *r[*xi++]; \
348: case 1: sum -= *xv++ *r[*xi++]; \
349: nnz -= 4;} \
350: while (nnz > 0) { \
351: sum -= xv[0] * r[xi[0]] - xv[1] * r[xi[1]] - \
352: xv[2] * r[xi[2]] - xv[3] * r[xi[3]]; \
353: xv += 4; xi += 4; nnz -= 4; }}}
355: #elif defined(PETSC_KERNEL_USE_UNROLL_2)
356: #define PetscSparseDenseMinusDot(sum,r,xv,xi,nnz) { \
357: PetscInt __i,__i1,__i2; \
358: for (__i=0; __i<nnz-1; __i+=2) {__i1 = xi[__i]; __i2=xi[__i+1]; \
359: sum -= (xv[__i]*r[__i1] + xv[__i+1]*r[__i2]);} \
360: if (nnz & 0x1) sum -= xv[__i] * r[xi[__i]];}
362: #else
363: #define PetscSparseDenseMinusDot(sum,r,xv,xi,nnz) { \
364: PetscInt __i; \
365: for (__i=0; __i<nnz; __i++) sum -= xv[__i] * r[xi[__i]];}
366: #endif
370: /*
371: PetscSparseDensePlusDot - The inner kernel of matrix-vector product \sum_i xv[i] * r[xi[i]] for CSR storage
373: Input Parameters:
374: + nnz - the number of entries
375: . r - the array of vector values
376: . xv - the matrix values for the row
377: - xi - the column indices of the nonzeros in the row
379: Output Parameter:
380: . sum - the sum of results
382: PETSc compile flags:
383: + PETSC_KERNEL_USE_UNROLL_4 - don't use this; it changes nnz and hence is WRONG
384: - PETSC_KERNEL_USE_UNROLL_2 -
386: .seealso: PetscSparseDenseMinusDot()
388: */
389: #if defined(PETSC_KERNEL_USE_UNROLL_4)
390: #define PetscSparseDensePlusDot(sum,r,xv,xi,nnz) { \
391: if (nnz > 0) { \
392: switch (nnz & 0x3) { \
393: case 3: sum += *xv++ *r[*xi++]; \
394: case 2: sum += *xv++ *r[*xi++]; \
395: case 1: sum += *xv++ *r[*xi++]; \
396: nnz -= 4;} \
397: while (nnz > 0) { \
398: sum += xv[0] * r[xi[0]] + xv[1] * r[xi[1]] + \
399: xv[2] * r[xi[2]] + xv[3] * r[xi[3]]; \
400: xv += 4; xi += 4; nnz -= 4; }}}
402: #elif defined(PETSC_KERNEL_USE_UNROLL_2)
403: #define PetscSparseDensePlusDot(sum,r,xv,xi,nnz) { \
404: PetscInt __i,__i1,__i2; \
405: for (__i=0; __i<nnz-1; __i+=2) {__i1 = xi[__i]; __i2=xi[__i+1]; \
406: sum += (xv[__i]*r[__i1] + xv[__i+1]*r[__i2]);} \
407: if (nnz & 0x1) sum += xv[__i] * r[xi[__i]];}
409: #else
410: #define PetscSparseDensePlusDot(sum,r,xv,xi,nnz) { \
411: PetscInt __i; \
412: for (__i=0; __i<nnz; __i++) sum += xv[__i] * r[xi[__i]];}
413: #endif
416: /*
417: PetscSparseDenseMaxDot - The inner kernel of a modified matrix-vector product \max_i xv[i] * r[xi[i]] for CSR storage
419: Input Parameters:
420: + nnz - the number of entries
421: . r - the array of vector values
422: . xv - the matrix values for the row
423: - xi - the column indices of the nonzeros in the row
425: Output Parameter:
426: . max - the max of results
428: .seealso: PetscSparseDensePlusDot(), PetscSparseDenseMinusDot()
430: */
431: #define PetscSparseDenseMaxDot(max,r,xv,xi,nnz) { \
432: PetscInt __i; \
433: for (__i=0; __i<nnz; __i++) max = PetscMax(PetscRealPart(max), PetscRealPart(xv[__i] * r[xi[__i]]));}
435: /*
436: Add column indices into table for counting the max nonzeros of merged rows
437: */
438: #define MatRowMergeMax_SeqAIJ(mat,nrows,ta) { \
439: PetscInt _j,_row,_nz,*_col; \
440: for (_row=0; _row<nrows; _row++) {\
441: _nz = mat->i[_row+1] - mat->i[_row]; \
442: for (_j=0; _j<_nz; _j++) { \
443: _col = _j + mat->j + mat->i[_row]; \
444: PetscTableAdd(ta,*_col+1,1,INSERT_VALUES); \
445: } \
446: } \
447: }
449: /*
450: Add column indices into table for counting the nonzeros of merged rows
451: */
452: #define MatMergeRows_SeqAIJ(mat,nrows,rows,ta) { \
453: PetscInt _j,_row,_nz,*_col,_i; \
454: for (_i=0; _i<nrows; _i++) {\
455: _row = rows[_i]; \
456: _nz = mat->i[_row+1] - mat->i[_row]; \
457: for (_j=0; _j<_nz; _j++) { \
458: _col = _j + mat->j + mat->i[_row]; \
459: PetscTableAdd(ta,*_col+1,1,INSERT_VALUES); \
460: } \
461: } \
462: }
464: #endif