1 2 /* 3 This is where the abstract matrix operations are defined 4 */ 5 6 #include "private/matimpl.h" /*I "petscmat.h" I*/ 7 #include "private/vecimpl.h" 8 9 /* Logging support */ 10 PetscClassId MAT_CLASSID; 11 PetscClassId MAT_FDCOLORING_CLASSID; 12 13 PetscLogEvent MAT_Mult, MAT_Mults, MAT_MultConstrained, MAT_MultAdd, MAT_MultTranspose; 14 PetscLogEvent MAT_MultTransposeConstrained, MAT_MultTransposeAdd, MAT_Solve, MAT_Solves, MAT_SolveAdd, MAT_SolveTranspose, MAT_MatSolve; 15 PetscLogEvent MAT_SolveTransposeAdd, MAT_SOR, MAT_ForwardSolve, MAT_BackwardSolve, MAT_LUFactor, MAT_LUFactorSymbolic; 16 PetscLogEvent MAT_LUFactorNumeric, MAT_CholeskyFactor, MAT_CholeskyFactorSymbolic, MAT_CholeskyFactorNumeric, MAT_ILUFactor; 17 PetscLogEvent MAT_ILUFactorSymbolic, MAT_ICCFactorSymbolic, MAT_Copy, MAT_Convert, MAT_Scale, MAT_AssemblyBegin; 18 PetscLogEvent MAT_AssemblyEnd, MAT_SetValues, MAT_GetValues, MAT_GetRow, MAT_GetRowIJ, MAT_GetSubMatrices, MAT_GetColoring, MAT_GetOrdering, MAT_GetRedundantMatrix, MAT_GetSeqNonzeroStructure; 19 PetscLogEvent MAT_IncreaseOverlap, MAT_Partitioning, MAT_ZeroEntries, MAT_Load, MAT_View, MAT_AXPY, MAT_FDColoringCreate; 20 PetscLogEvent MAT_FDColoringApply,MAT_Transpose,MAT_FDColoringFunction; 21 PetscLogEvent MAT_MatMult, MAT_MatMultSymbolic, MAT_MatMultNumeric; 22 PetscLogEvent MAT_PtAP, MAT_PtAPSymbolic, MAT_PtAPNumeric; 23 PetscLogEvent MAT_MatMultTranspose, MAT_MatMultTransposeSymbolic, MAT_MatMultTransposeNumeric; 24 PetscLogEvent MAT_MultHermitianTranspose,MAT_MultHermitianTransposeAdd; 25 PetscLogEvent MAT_Getsymtranspose, MAT_Getsymtransreduced, MAT_Transpose_SeqAIJ, MAT_GetBrowsOfAcols; 26 PetscLogEvent MAT_GetBrowsOfAocols, MAT_Getlocalmat, MAT_Getlocalmatcondensed, MAT_Seqstompi, MAT_Seqstompinum, MAT_Seqstompisym; 27 PetscLogEvent MAT_Applypapt, MAT_Applypapt_numeric, MAT_Applypapt_symbolic, MAT_GetSequentialNonzeroStructure; 28 PetscLogEvent MAT_GetMultiProcBlock; 29 PetscLogEvent MAT_CUSPCopyToGPU; 30 31 /* nasty global values for MatSetValue() */ 32 PetscInt MatSetValue_Row = 0; 33 PetscInt MatSetValue_Column = 0; 34 PetscScalar MatSetValue_Value = 0.0; 35 36 const char *const MatFactorTypes[] = {"NONE","LU","CHOLESKY","ILU","ICC","ILUDT","MatFactorType","MAT_FACTOR_",0}; 37 38 #undef __FUNCT__ 39 #define __FUNCT__ "MatFindNonzeroRows" 40 /*@C 41 MatFindNonzeroRows - Locate all rows that are not completely zero in the matrix 42 43 Input Parameter: 44 . A - the matrix 45 46 Output Parameter: 47 . keptrows - the rows that are not completely zero 48 49 @*/ 50 PetscErrorCode MatFindNonzeroRows(Mat mat,IS *keptrows) 51 { 52 PetscErrorCode ierr; 53 54 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 55 PetscValidType(mat,1); 56 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 57 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 58 if (!mat->ops->findnonzerorows) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Not coded for this matrix type"); 59 ierr = (*mat->ops->findnonzerorows)(mat,keptrows);CHKERRQ(ierr); 60 PetscFunctionReturn(0); 61 } 62 63 #undef __FUNCT__ 64 #define __FUNCT__ "MatGetDiagonalBlock" 65 /*@ 66 MatGetDiagonalBlock - Returns the part of the matrix associated with the on-process coupling 67 68 Not Collective 69 70 Input Parameters: 71 + mat - the matrix 72 - reuse - indicates you are passing in the a matrix and want it reused 73 74 Output Parameters: 75 + iscopy - indicates a copy of the diagonal matrix was created and you should use MatDestroy() on it 76 - a - the diagonal part (which is a SEQUENTIAL matrix) 77 78 Notes: see the manual page for MatCreateMPIAIJ() for more information on the "diagonal part" of the matrix 79 80 Level: advanced 81 82 @*/ 83 PetscErrorCode MatGetDiagonalBlock(Mat A,PetscBool *iscopy,MatReuse reuse,Mat *a) 84 { 85 PetscErrorCode ierr,(*f)(Mat,PetscBool *,MatReuse,Mat*); 86 PetscMPIInt size; 87 88 PetscFunctionBegin; 89 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 90 PetscValidType(A,1); 91 PetscValidPointer(iscopy,2); 92 PetscValidPointer(a,3); 93 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 94 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 95 ierr = MPI_Comm_size(((PetscObject)A)->comm,&size);CHKERRQ(ierr); 96 ierr = PetscObjectQueryFunction((PetscObject)A,"MatGetDiagonalBlock_C",(void (**)(void))&f);CHKERRQ(ierr); 97 if (f) { 98 ierr = (*f)(A,iscopy,reuse,a);CHKERRQ(ierr); 99 } else if (size == 1) { 100 *a = A; 101 *iscopy = PETSC_FALSE; 102 } else SETERRQ(((PetscObject)A)->comm,PETSC_ERR_SUP,"Cannot get diagonal part for this matrix"); 103 PetscFunctionReturn(0); 104 } 105 106 #undef __FUNCT__ 107 #define __FUNCT__ "MatGetTrace" 108 /*@ 109 MatGetTrace - Gets the trace of a matrix. The sum of the diagonal entries. 110 111 Collective on Mat 112 113 Input Parameters: 114 . mat - the matrix 115 116 Output Parameter: 117 . trace - the sum of the diagonal entries 118 119 Level: advanced 120 121 @*/ 122 PetscErrorCode MatGetTrace(Mat mat,PetscScalar *trace) 123 { 124 PetscErrorCode ierr; 125 Vec diag; 126 127 PetscFunctionBegin; 128 ierr = MatGetVecs(mat,&diag,PETSC_NULL);CHKERRQ(ierr); 129 ierr = MatGetDiagonal(mat,diag);CHKERRQ(ierr); 130 ierr = VecSum(diag,trace);CHKERRQ(ierr); 131 ierr = VecDestroy(diag);CHKERRQ(ierr); 132 PetscFunctionReturn(0); 133 } 134 135 #undef __FUNCT__ 136 #define __FUNCT__ "MatRealPart" 137 /*@ 138 MatRealPart - Zeros out the imaginary part of the matrix 139 140 Logically Collective on Mat 141 142 Input Parameters: 143 . mat - the matrix 144 145 Level: advanced 146 147 148 .seealso: MatImaginaryPart() 149 @*/ 150 PetscErrorCode MatRealPart(Mat mat) 151 { 152 PetscErrorCode ierr; 153 154 PetscFunctionBegin; 155 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 156 PetscValidType(mat,1); 157 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 158 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 159 if (!mat->ops->realpart) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 160 ierr = MatPreallocated(mat);CHKERRQ(ierr); 161 ierr = (*mat->ops->realpart)(mat);CHKERRQ(ierr); 162 #if defined(PETSC_HAVE_CUSP) 163 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 164 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 165 } 166 #endif 167 PetscFunctionReturn(0); 168 } 169 170 #undef __FUNCT__ 171 #define __FUNCT__ "MatGetGhosts" 172 /*@C 173 MatGetGhosts - Get the global index of all ghost nodes defined by the sparse matrix 174 175 Collective on Mat 176 177 Input Parameter: 178 . mat - the matrix 179 180 Output Parameters: 181 + nghosts - number of ghosts (note for BAIJ matrices there is one ghost for each block) 182 - ghosts - the global indices of the ghost points 183 184 Notes: the nghosts and ghosts are suitable to pass into VecCreateGhost() 185 186 Level: advanced 187 188 @*/ 189 PetscErrorCode MatGetGhosts(Mat mat,PetscInt *nghosts,const PetscInt *ghosts[]) 190 { 191 PetscErrorCode ierr; 192 193 PetscFunctionBegin; 194 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 195 PetscValidType(mat,1); 196 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 197 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 198 if (!mat->ops->getghosts) { 199 if (nghosts) *nghosts = 0; 200 if (ghosts) *ghosts = 0; 201 } else { 202 ierr = (*mat->ops->getghosts)(mat,nghosts,ghosts);CHKERRQ(ierr); 203 } 204 PetscFunctionReturn(0); 205 } 206 207 208 #undef __FUNCT__ 209 #define __FUNCT__ "MatImaginaryPart" 210 /*@ 211 MatImaginaryPart - Moves the imaginary part of the matrix to the real part and zeros the imaginary part 212 213 Logically Collective on Mat 214 215 Input Parameters: 216 . mat - the matrix 217 218 Level: advanced 219 220 221 .seealso: MatRealPart() 222 @*/ 223 PetscErrorCode MatImaginaryPart(Mat mat) 224 { 225 PetscErrorCode ierr; 226 227 PetscFunctionBegin; 228 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 229 PetscValidType(mat,1); 230 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 231 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 232 if (!mat->ops->imaginarypart) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 233 ierr = MatPreallocated(mat);CHKERRQ(ierr); 234 ierr = (*mat->ops->imaginarypart)(mat);CHKERRQ(ierr); 235 #if defined(PETSC_HAVE_CUSP) 236 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 237 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 238 } 239 #endif 240 PetscFunctionReturn(0); 241 } 242 243 #undef __FUNCT__ 244 #define __FUNCT__ "MatMissingDiagonal" 245 /*@ 246 MatMissingDiagonal - Determine if sparse matrix is missing a diagonal entry (or block entry for BAIJ matrices) 247 248 Collective on Mat 249 250 Input Parameter: 251 . mat - the matrix 252 253 Output Parameters: 254 + missing - is any diagonal missing 255 - dd - first diagonal entry that is missing (optional) 256 257 Level: advanced 258 259 260 .seealso: MatRealPart() 261 @*/ 262 PetscErrorCode MatMissingDiagonal(Mat mat,PetscBool *missing,PetscInt *dd) 263 { 264 PetscErrorCode ierr; 265 266 PetscFunctionBegin; 267 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 268 PetscValidType(mat,1); 269 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 270 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 271 if (!mat->ops->missingdiagonal) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 272 ierr = (*mat->ops->missingdiagonal)(mat,missing,dd);CHKERRQ(ierr); 273 PetscFunctionReturn(0); 274 } 275 276 #undef __FUNCT__ 277 #define __FUNCT__ "MatGetRow" 278 /*@C 279 MatGetRow - Gets a row of a matrix. You MUST call MatRestoreRow() 280 for each row that you get to ensure that your application does 281 not bleed memory. 282 283 Not Collective 284 285 Input Parameters: 286 + mat - the matrix 287 - row - the row to get 288 289 Output Parameters: 290 + ncols - if not NULL, the number of nonzeros in the row 291 . cols - if not NULL, the column numbers 292 - vals - if not NULL, the values 293 294 Notes: 295 This routine is provided for people who need to have direct access 296 to the structure of a matrix. We hope that we provide enough 297 high-level matrix routines that few users will need it. 298 299 MatGetRow() always returns 0-based column indices, regardless of 300 whether the internal representation is 0-based (default) or 1-based. 301 302 For better efficiency, set cols and/or vals to PETSC_NULL if you do 303 not wish to extract these quantities. 304 305 The user can only examine the values extracted with MatGetRow(); 306 the values cannot be altered. To change the matrix entries, one 307 must use MatSetValues(). 308 309 You can only have one call to MatGetRow() outstanding for a particular 310 matrix at a time, per processor. MatGetRow() can only obtain rows 311 associated with the given processor, it cannot get rows from the 312 other processors; for that we suggest using MatGetSubMatrices(), then 313 MatGetRow() on the submatrix. The row indix passed to MatGetRows() 314 is in the global number of rows. 315 316 Fortran Notes: 317 The calling sequence from Fortran is 318 .vb 319 MatGetRow(matrix,row,ncols,cols,values,ierr) 320 Mat matrix (input) 321 integer row (input) 322 integer ncols (output) 323 integer cols(maxcols) (output) 324 double precision (or double complex) values(maxcols) output 325 .ve 326 where maxcols >= maximum nonzeros in any row of the matrix. 327 328 329 Caution: 330 Do not try to change the contents of the output arrays (cols and vals). 331 In some cases, this may corrupt the matrix. 332 333 Level: advanced 334 335 Concepts: matrices^row access 336 337 .seealso: MatRestoreRow(), MatSetValues(), MatGetValues(), MatGetSubMatrices(), MatGetDiagonal() 338 @*/ 339 PetscErrorCode MatGetRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[]) 340 { 341 PetscErrorCode ierr; 342 PetscInt incols; 343 344 PetscFunctionBegin; 345 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 346 PetscValidType(mat,1); 347 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 348 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 349 if (!mat->ops->getrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 350 ierr = MatPreallocated(mat);CHKERRQ(ierr); 351 ierr = PetscLogEventBegin(MAT_GetRow,mat,0,0,0);CHKERRQ(ierr); 352 ierr = (*mat->ops->getrow)(mat,row,&incols,(PetscInt **)cols,(PetscScalar **)vals);CHKERRQ(ierr); 353 if (ncols) *ncols = incols; 354 ierr = PetscLogEventEnd(MAT_GetRow,mat,0,0,0);CHKERRQ(ierr); 355 PetscFunctionReturn(0); 356 } 357 358 #undef __FUNCT__ 359 #define __FUNCT__ "MatConjugate" 360 /*@ 361 MatConjugate - replaces the matrix values with their complex conjugates 362 363 Logically Collective on Mat 364 365 Input Parameters: 366 . mat - the matrix 367 368 Level: advanced 369 370 .seealso: VecConjugate() 371 @*/ 372 PetscErrorCode MatConjugate(Mat mat) 373 { 374 PetscErrorCode ierr; 375 376 PetscFunctionBegin; 377 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 378 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 379 if (!mat->ops->conjugate) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Not provided for this matrix format, send email to petsc-maint@mcs.anl.gov"); 380 ierr = (*mat->ops->conjugate)(mat);CHKERRQ(ierr); 381 #if defined(PETSC_HAVE_CUSP) 382 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 383 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 384 } 385 #endif 386 PetscFunctionReturn(0); 387 } 388 389 #undef __FUNCT__ 390 #define __FUNCT__ "MatRestoreRow" 391 /*@C 392 MatRestoreRow - Frees any temporary space allocated by MatGetRow(). 393 394 Not Collective 395 396 Input Parameters: 397 + mat - the matrix 398 . row - the row to get 399 . ncols, cols - the number of nonzeros and their columns 400 - vals - if nonzero the column values 401 402 Notes: 403 This routine should be called after you have finished examining the entries. 404 405 Fortran Notes: 406 The calling sequence from Fortran is 407 .vb 408 MatRestoreRow(matrix,row,ncols,cols,values,ierr) 409 Mat matrix (input) 410 integer row (input) 411 integer ncols (output) 412 integer cols(maxcols) (output) 413 double precision (or double complex) values(maxcols) output 414 .ve 415 Where maxcols >= maximum nonzeros in any row of the matrix. 416 417 In Fortran MatRestoreRow() MUST be called after MatGetRow() 418 before another call to MatGetRow() can be made. 419 420 Level: advanced 421 422 .seealso: MatGetRow() 423 @*/ 424 PetscErrorCode MatRestoreRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[]) 425 { 426 PetscErrorCode ierr; 427 428 PetscFunctionBegin; 429 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 430 PetscValidIntPointer(ncols,3); 431 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 432 if (!mat->ops->restorerow) PetscFunctionReturn(0); 433 ierr = (*mat->ops->restorerow)(mat,row,ncols,(PetscInt **)cols,(PetscScalar **)vals);CHKERRQ(ierr); 434 PetscFunctionReturn(0); 435 } 436 437 #undef __FUNCT__ 438 #define __FUNCT__ "MatGetRowUpperTriangular" 439 /*@ 440 MatGetRowUpperTriangular - Sets a flag to enable calls to MatGetRow() for matrix in MATSBAIJ format. 441 You should call MatRestoreRowUpperTriangular() after calling MatGetRow/MatRestoreRow() to disable the flag. 442 443 Not Collective 444 445 Input Parameters: 446 + mat - the matrix 447 448 Notes: 449 The flag is to ensure that users are aware of MatGetRow() only provides the upper trianglular part of the row for the matrices in MATSBAIJ format. 450 451 Level: advanced 452 453 Concepts: matrices^row access 454 455 .seealso: MatRestoreRowRowUpperTriangular() 456 @*/ 457 PetscErrorCode MatGetRowUpperTriangular(Mat mat) 458 { 459 PetscErrorCode ierr; 460 461 PetscFunctionBegin; 462 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 463 PetscValidType(mat,1); 464 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 465 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 466 if (!mat->ops->getrowuppertriangular) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 467 ierr = MatPreallocated(mat);CHKERRQ(ierr); 468 ierr = (*mat->ops->getrowuppertriangular)(mat);CHKERRQ(ierr); 469 PetscFunctionReturn(0); 470 } 471 472 #undef __FUNCT__ 473 #define __FUNCT__ "MatRestoreRowUpperTriangular" 474 /*@ 475 MatRestoreRowUpperTriangular - Disable calls to MatGetRow() for matrix in MATSBAIJ format. 476 477 Not Collective 478 479 Input Parameters: 480 + mat - the matrix 481 482 Notes: 483 This routine should be called after you have finished MatGetRow/MatRestoreRow(). 484 485 486 Level: advanced 487 488 .seealso: MatGetRowUpperTriangular() 489 @*/ 490 PetscErrorCode MatRestoreRowUpperTriangular(Mat mat) 491 { 492 PetscErrorCode ierr; 493 494 PetscFunctionBegin; 495 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 496 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 497 if (!mat->ops->restorerowuppertriangular) PetscFunctionReturn(0); 498 ierr = (*mat->ops->restorerowuppertriangular)(mat);CHKERRQ(ierr); 499 PetscFunctionReturn(0); 500 } 501 502 #undef __FUNCT__ 503 #define __FUNCT__ "MatSetOptionsPrefix" 504 /*@C 505 MatSetOptionsPrefix - Sets the prefix used for searching for all 506 Mat options in the database. 507 508 Logically Collective on Mat 509 510 Input Parameter: 511 + A - the Mat context 512 - prefix - the prefix to prepend to all option names 513 514 Notes: 515 A hyphen (-) must NOT be given at the beginning of the prefix name. 516 The first character of all runtime options is AUTOMATICALLY the hyphen. 517 518 Level: advanced 519 520 .keywords: Mat, set, options, prefix, database 521 522 .seealso: MatSetFromOptions() 523 @*/ 524 PetscErrorCode MatSetOptionsPrefix(Mat A,const char prefix[]) 525 { 526 PetscErrorCode ierr; 527 528 PetscFunctionBegin; 529 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 530 ierr = PetscObjectSetOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 531 PetscFunctionReturn(0); 532 } 533 534 #undef __FUNCT__ 535 #define __FUNCT__ "MatAppendOptionsPrefix" 536 /*@C 537 MatAppendOptionsPrefix - Appends to the prefix used for searching for all 538 Mat options in the database. 539 540 Logically Collective on Mat 541 542 Input Parameters: 543 + A - the Mat context 544 - prefix - the prefix to prepend to all option names 545 546 Notes: 547 A hyphen (-) must NOT be given at the beginning of the prefix name. 548 The first character of all runtime options is AUTOMATICALLY the hyphen. 549 550 Level: advanced 551 552 .keywords: Mat, append, options, prefix, database 553 554 .seealso: MatGetOptionsPrefix() 555 @*/ 556 PetscErrorCode MatAppendOptionsPrefix(Mat A,const char prefix[]) 557 { 558 PetscErrorCode ierr; 559 560 PetscFunctionBegin; 561 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 562 ierr = PetscObjectAppendOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 563 PetscFunctionReturn(0); 564 } 565 566 #undef __FUNCT__ 567 #define __FUNCT__ "MatGetOptionsPrefix" 568 /*@C 569 MatGetOptionsPrefix - Sets the prefix used for searching for all 570 Mat options in the database. 571 572 Not Collective 573 574 Input Parameter: 575 . A - the Mat context 576 577 Output Parameter: 578 . prefix - pointer to the prefix string used 579 580 Notes: On the fortran side, the user should pass in a string 'prefix' of 581 sufficient length to hold the prefix. 582 583 Level: advanced 584 585 .keywords: Mat, get, options, prefix, database 586 587 .seealso: MatAppendOptionsPrefix() 588 @*/ 589 PetscErrorCode MatGetOptionsPrefix(Mat A,const char *prefix[]) 590 { 591 PetscErrorCode ierr; 592 593 PetscFunctionBegin; 594 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 595 ierr = PetscObjectGetOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 596 PetscFunctionReturn(0); 597 } 598 599 #undef __FUNCT__ 600 #define __FUNCT__ "MatSetUp" 601 /*@ 602 MatSetUp - Sets up the internal matrix data structures for the later use. 603 604 Collective on Mat 605 606 Input Parameters: 607 . A - the Mat context 608 609 Notes: 610 For basic use of the Mat classes the user need not explicitly call 611 MatSetUp(), since these actions will happen automatically. 612 613 Level: advanced 614 615 .keywords: Mat, setup 616 617 .seealso: MatCreate(), MatDestroy() 618 @*/ 619 PetscErrorCode MatSetUp(Mat A) 620 { 621 PetscMPIInt size; 622 PetscErrorCode ierr; 623 624 PetscFunctionBegin; 625 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 626 if (!((PetscObject)A)->type_name) { 627 ierr = MPI_Comm_size(((PetscObject)A)->comm, &size);CHKERRQ(ierr); 628 if (size == 1) { 629 ierr = MatSetType(A, MATSEQAIJ);CHKERRQ(ierr); 630 } else { 631 ierr = MatSetType(A, MATMPIAIJ);CHKERRQ(ierr); 632 } 633 } 634 ierr = MatSetUpPreallocation(A);CHKERRQ(ierr); 635 PetscFunctionReturn(0); 636 } 637 638 639 #undef __FUNCT__ 640 #define __FUNCT__ "MatView" 641 /*@C 642 MatView - Visualizes a matrix object. 643 644 Collective on Mat 645 646 Input Parameters: 647 + mat - the matrix 648 - viewer - visualization context 649 650 Notes: 651 The available visualization contexts include 652 + PETSC_VIEWER_STDOUT_SELF - standard output (default) 653 . PETSC_VIEWER_STDOUT_WORLD - synchronized standard 654 output where only the first processor opens 655 the file. All other processors send their 656 data to the first processor to print. 657 - PETSC_VIEWER_DRAW_WORLD - graphical display of nonzero structure 658 659 The user can open alternative visualization contexts with 660 + PetscViewerASCIIOpen() - Outputs matrix to a specified file 661 . PetscViewerBinaryOpen() - Outputs matrix in binary to a 662 specified file; corresponding input uses MatLoad() 663 . PetscViewerDrawOpen() - Outputs nonzero matrix structure to 664 an X window display 665 - PetscViewerSocketOpen() - Outputs matrix to Socket viewer. 666 Currently only the sequential dense and AIJ 667 matrix types support the Socket viewer. 668 669 The user can call PetscViewerSetFormat() to specify the output 670 format of ASCII printed objects (when using PETSC_VIEWER_STDOUT_SELF, 671 PETSC_VIEWER_STDOUT_WORLD and PetscViewerASCIIOpen). Available formats include 672 + PETSC_VIEWER_DEFAULT - default, prints matrix contents 673 . PETSC_VIEWER_ASCII_MATLAB - prints matrix contents in Matlab format 674 . PETSC_VIEWER_ASCII_DENSE - prints entire matrix including zeros 675 . PETSC_VIEWER_ASCII_COMMON - prints matrix contents, using a sparse 676 format common among all matrix types 677 . PETSC_VIEWER_ASCII_IMPL - prints matrix contents, using an implementation-specific 678 format (which is in many cases the same as the default) 679 . PETSC_VIEWER_ASCII_INFO - prints basic information about the matrix 680 size and structure (not the matrix entries) 681 . PETSC_VIEWER_ASCII_INFO_DETAIL - prints more detailed information about 682 the matrix structure 683 684 Options Database Keys: 685 + -mat_view_info - Prints info on matrix at conclusion of MatEndAssembly() 686 . -mat_view_info_detailed - Prints more detailed info 687 . -mat_view - Prints matrix in ASCII format 688 . -mat_view_matlab - Prints matrix in Matlab format 689 . -mat_view_draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX(). 690 . -display <name> - Sets display name (default is host) 691 . -draw_pause <sec> - Sets number of seconds to pause after display 692 . -mat_view_socket - Sends matrix to socket, can be accessed from Matlab (see the <a href="../../docs/manual.pdf">users manual</a> for details). 693 . -viewer_socket_machine <machine> 694 . -viewer_socket_port <port> 695 . -mat_view_binary - save matrix to file in binary format 696 - -viewer_binary_filename <name> 697 Level: beginner 698 699 Notes: see the manual page for MatLoad() for the exact format of the binary file when the binary 700 viewer is used. 701 702 See bin/matlab/PetscBinaryRead.m for a Matlab code that can read in the binary file when the binary 703 viewer is used. 704 705 Concepts: matrices^viewing 706 Concepts: matrices^plotting 707 Concepts: matrices^printing 708 709 .seealso: PetscViewerSetFormat(), PetscViewerASCIIOpen(), PetscViewerDrawOpen(), 710 PetscViewerSocketOpen(), PetscViewerBinaryOpen(), MatLoad() 711 @*/ 712 PetscErrorCode MatView(Mat mat,PetscViewer viewer) 713 { 714 PetscErrorCode ierr; 715 PetscInt rows,cols; 716 PetscBool iascii; 717 PetscViewerFormat format; 718 719 PetscFunctionBegin; 720 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 721 PetscValidType(mat,1); 722 if (!viewer) { 723 ierr = PetscViewerASCIIGetStdout(((PetscObject)mat)->comm,&viewer);CHKERRQ(ierr); 724 } 725 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2); 726 PetscCheckSameComm(mat,1,viewer,2); 727 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ORDER,"Must call MatAssemblyBegin/End() before viewing matrix"); 728 ierr = MatPreallocated(mat);CHKERRQ(ierr); 729 730 ierr = PetscLogEventBegin(MAT_View,mat,viewer,0,0);CHKERRQ(ierr); 731 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);CHKERRQ(ierr); 732 if (iascii) { 733 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 734 if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 735 ierr = PetscObjectPrintClassNamePrefixType((PetscObject)mat,viewer,"Matrix Object");CHKERRQ(ierr); 736 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 737 ierr = MatGetSize(mat,&rows,&cols);CHKERRQ(ierr); 738 ierr = PetscViewerASCIIPrintf(viewer,"rows=%D, cols=%D\n",rows,cols);CHKERRQ(ierr); 739 if (mat->factortype) { 740 const MatSolverPackage solver; 741 ierr = MatFactorGetSolverPackage(mat,&solver);CHKERRQ(ierr); 742 ierr = PetscViewerASCIIPrintf(viewer,"package used to perform factorization: %s\n",solver);CHKERRQ(ierr); 743 } 744 if (mat->ops->getinfo) { 745 MatInfo info; 746 ierr = MatGetInfo(mat,MAT_GLOBAL_SUM,&info);CHKERRQ(ierr); 747 ierr = PetscViewerASCIIPrintf(viewer,"total: nonzeros=%D, allocated nonzeros=%D\n",(PetscInt)info.nz_used,(PetscInt)info.nz_allocated);CHKERRQ(ierr); 748 ierr = PetscViewerASCIIPrintf(viewer,"total number of mallocs used during MatSetValues calls =%D\n",(PetscInt)info.mallocs);CHKERRQ(ierr); 749 } 750 } 751 } 752 if (mat->ops->view) { 753 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 754 ierr = (*mat->ops->view)(mat,viewer);CHKERRQ(ierr); 755 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 756 } else if (!iascii) { 757 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Viewer type %s not supported",((PetscObject)viewer)->type_name); 758 } 759 if (iascii) { 760 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 761 if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 762 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 763 } 764 } 765 ierr = PetscLogEventEnd(MAT_View,mat,viewer,0,0);CHKERRQ(ierr); 766 PetscFunctionReturn(0); 767 } 768 769 #if defined(PETSC_USE_DEBUG) 770 #include "../src/sys/totalview/tv_data_display.h" 771 PETSC_UNUSED static int TV_display_type(const struct _p_Mat *mat) 772 { 773 TV_add_row("Local rows", "int", &mat->rmap->n); 774 TV_add_row("Local columns", "int", &mat->cmap->n); 775 TV_add_row("Global rows", "int", &mat->rmap->N); 776 TV_add_row("Global columns", "int", &mat->cmap->N); 777 TV_add_row("Typename", TV_ascii_string_type, ((PetscObject)mat)->type_name); 778 return TV_format_OK; 779 } 780 #endif 781 782 #undef __FUNCT__ 783 #define __FUNCT__ "MatLoad" 784 /*@C 785 MatLoad - Loads a matrix that has been stored in binary format 786 with MatView(). The matrix format is determined from the options database. 787 Generates a parallel MPI matrix if the communicator has more than one 788 processor. The default matrix type is AIJ. 789 790 Collective on PetscViewer 791 792 Input Parameters: 793 + newmat - the newly loaded matrix, this needs to have been created with MatCreate() 794 or some related function before a call to MatLoad() 795 - viewer - binary file viewer, created with PetscViewerBinaryOpen() 796 797 Options Database Keys: 798 Used with block matrix formats (MATSEQBAIJ, ...) to specify 799 block size 800 . -matload_block_size <bs> 801 802 Level: beginner 803 804 Notes: 805 If the Mat type has not yet been given then MATAIJ is used, call MatSetFromOptions() on the 806 Mat before calling this routine if you wish to set it from the options database. 807 808 MatLoad() automatically loads into the options database any options 809 given in the file filename.info where filename is the name of the file 810 that was passed to the PetscViewerBinaryOpen(). The options in the info 811 file will be ignored if you use the -viewer_binary_skip_info option. 812 813 If the type or size of newmat is not set before a call to MatLoad, PETSc 814 sets the default matrix type AIJ and sets the local and global sizes. 815 If type and/or size is already set, then the same are used. 816 817 In parallel, each processor can load a subset of rows (or the 818 entire matrix). This routine is especially useful when a large 819 matrix is stored on disk and only part of it is desired on each 820 processor. For example, a parallel solver may access only some of 821 the rows from each processor. The algorithm used here reads 822 relatively small blocks of data rather than reading the entire 823 matrix and then subsetting it. 824 825 Notes for advanced users: 826 Most users should not need to know the details of the binary storage 827 format, since MatLoad() and MatView() completely hide these details. 828 But for anyone who's interested, the standard binary matrix storage 829 format is 830 831 $ int MAT_FILE_CLASSID 832 $ int number of rows 833 $ int number of columns 834 $ int total number of nonzeros 835 $ int *number nonzeros in each row 836 $ int *column indices of all nonzeros (starting index is zero) 837 $ PetscScalar *values of all nonzeros 838 839 PETSc automatically does the byte swapping for 840 machines that store the bytes reversed, e.g. DEC alpha, freebsd, 841 linux, Windows and the paragon; thus if you write your own binary 842 read/write routines you have to swap the bytes; see PetscBinaryRead() 843 and PetscBinaryWrite() to see how this may be done. 844 845 .keywords: matrix, load, binary, input 846 847 .seealso: PetscViewerBinaryOpen(), MatView(), VecLoad() 848 849 @*/ 850 PetscErrorCode MatLoad(Mat newmat,PetscViewer viewer) 851 { 852 PetscErrorCode ierr; 853 PetscBool isbinary,flg; 854 855 PetscFunctionBegin; 856 PetscValidHeaderSpecific(newmat,MAT_CLASSID,1); 857 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2); 858 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 859 if (!isbinary) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Invalid viewer; open viewer with PetscViewerBinaryOpen()"); 860 861 if (!((PetscObject)newmat)->type_name) { 862 ierr = MatSetType(newmat,MATAIJ);CHKERRQ(ierr); 863 } 864 865 if (!newmat->ops->load) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatLoad is not supported for type"); 866 ierr = PetscLogEventBegin(MAT_Load,viewer,0,0,0);CHKERRQ(ierr); 867 ierr = (*newmat->ops->load)(newmat,viewer);CHKERRQ(ierr); 868 ierr = PetscLogEventEnd(MAT_Load,viewer,0,0,0);CHKERRQ(ierr); 869 870 flg = PETSC_FALSE; 871 ierr = PetscOptionsGetBool(((PetscObject)newmat)->prefix,"-matload_symmetric",&flg,PETSC_NULL);CHKERRQ(ierr); 872 if (flg) { 873 ierr = MatSetOption(newmat,MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 874 ierr = MatSetOption(newmat,MAT_SYMMETRY_ETERNAL,PETSC_TRUE);CHKERRQ(ierr); 875 } 876 flg = PETSC_FALSE; 877 ierr = PetscOptionsGetBool(((PetscObject)newmat)->prefix,"-matload_spd",&flg,PETSC_NULL);CHKERRQ(ierr); 878 if (flg) { 879 ierr = MatSetOption(newmat,MAT_SPD,PETSC_TRUE);CHKERRQ(ierr); 880 } 881 PetscFunctionReturn(0); 882 } 883 884 #undef __FUNCT__ 885 #define __FUNCT__ "MatScaleSystem" 886 /*@ 887 MatScaleSystem - Scale a vector solution and right hand side to 888 match the scaling of a scaled matrix. 889 890 Collective on Mat 891 892 Input Parameter: 893 + mat - the matrix 894 . b - right hand side vector (or PETSC_NULL) 895 - x - solution vector (or PETSC_NULL) 896 897 898 Notes: 899 For AIJ, and BAIJ matrix formats, the matrices are not 900 internally scaled, so this does nothing. 901 902 The KSP methods automatically call this routine when required 903 (via PCPreSolve()) so it is rarely used directly. 904 905 Level: Developer 906 907 Concepts: matrices^scaling 908 909 .seealso: MatUseScaledForm(), MatUnScaleSystem() 910 @*/ 911 PetscErrorCode MatScaleSystem(Mat mat,Vec b,Vec x) 912 { 913 PetscErrorCode ierr; 914 915 PetscFunctionBegin; 916 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 917 PetscValidType(mat,1); 918 ierr = MatPreallocated(mat);CHKERRQ(ierr); 919 if (x) {PetscValidHeaderSpecific(x,VEC_CLASSID,3);PetscCheckSameComm(mat,1,x,3);} 920 if (b) {PetscValidHeaderSpecific(b,VEC_CLASSID,2);PetscCheckSameComm(mat,1,b,2);} 921 922 if (mat->ops->scalesystem) { 923 ierr = (*mat->ops->scalesystem)(mat,b,x);CHKERRQ(ierr); 924 } 925 PetscFunctionReturn(0); 926 } 927 928 #undef __FUNCT__ 929 #define __FUNCT__ "MatUnScaleSystem" 930 /*@ 931 MatUnScaleSystem - Unscales a vector solution and right hand side to 932 match the original scaling of a scaled matrix. 933 934 Collective on Mat 935 936 Input Parameter: 937 + mat - the matrix 938 . b - right hand side vector (or PETSC_NULL) 939 - x - solution vector (or PETSC_NULL) 940 941 942 Notes: 943 For AIJ and BAIJ matrix formats, the matrices are not 944 internally scaled, so this does nothing. 945 946 The KSP methods automatically call this routine when required 947 (via PCPreSolve()) so it is rarely used directly. 948 949 Level: Developer 950 951 .seealso: MatUseScaledForm(), MatScaleSystem() 952 @*/ 953 PetscErrorCode MatUnScaleSystem(Mat mat,Vec b,Vec x) 954 { 955 PetscErrorCode ierr; 956 957 PetscFunctionBegin; 958 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 959 PetscValidType(mat,1); 960 ierr = MatPreallocated(mat);CHKERRQ(ierr); 961 if (x) {PetscValidHeaderSpecific(x,VEC_CLASSID,3);PetscCheckSameComm(mat,1,x,3);} 962 if (b) {PetscValidHeaderSpecific(b,VEC_CLASSID,2);PetscCheckSameComm(mat,1,b,2);} 963 if (mat->ops->unscalesystem) { 964 ierr = (*mat->ops->unscalesystem)(mat,b,x);CHKERRQ(ierr); 965 } 966 PetscFunctionReturn(0); 967 } 968 969 #undef __FUNCT__ 970 #define __FUNCT__ "MatUseScaledForm" 971 /*@ 972 MatUseScaledForm - For matrix storage formats that scale the 973 matrix indicates matrix operations (MatMult() etc) are 974 applied using the scaled matrix. 975 976 Logically Collective on Mat 977 978 Input Parameter: 979 + mat - the matrix 980 - scaled - PETSC_TRUE for applying the scaled, PETSC_FALSE for 981 applying the original matrix 982 983 Notes: 984 For scaled matrix formats, applying the original, unscaled matrix 985 will be slightly more expensive 986 987 Level: Developer 988 989 .seealso: MatScaleSystem(), MatUnScaleSystem() 990 @*/ 991 PetscErrorCode MatUseScaledForm(Mat mat,PetscBool scaled) 992 { 993 PetscErrorCode ierr; 994 995 PetscFunctionBegin; 996 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 997 PetscValidType(mat,1); 998 PetscValidLogicalCollectiveBool(mat,scaled,2); 999 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1000 if (mat->ops->usescaledform) { 1001 ierr = (*mat->ops->usescaledform)(mat,scaled);CHKERRQ(ierr); 1002 } 1003 PetscFunctionReturn(0); 1004 } 1005 1006 #undef __FUNCT__ 1007 #define __FUNCT__ "MatDestroy_" 1008 /*@C 1009 MatDestroy_ - Frees space taken by a matrix. 1010 1011 Collective on Mat 1012 1013 Input Parameter: 1014 . A - the matrix 1015 1016 Level: beginner 1017 1018 @*/ 1019 PetscErrorCode MatDestroy_(Mat A) 1020 { 1021 PetscErrorCode ierr; 1022 PetscFunctionBegin; 1023 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 1024 if (--((PetscObject)A)->refct > 0) PetscFunctionReturn(0); 1025 /* if no sizes were ever set in matrix then MatPreallocated() may generate an error, so set the sizes */ 1026 if (A->rmap->n == -1 && A->rmap->N == -1) A->rmap->n = A->rmap->N = 0; 1027 if (A->cmap->n == -1 && A->cmap->N == -1) A->cmap->n = A->cmap->N = 0; 1028 ierr = MatPreallocated(A);CHKERRQ(ierr); 1029 /* if memory was published with AMS then destroy it */ 1030 ierr = PetscObjectDepublish(A);CHKERRQ(ierr); 1031 if (A->ops->destroy) { 1032 ierr = (*A->ops->destroy)(A);CHKERRQ(ierr); 1033 } 1034 if (A->rmapping) {ierr = ISLocalToGlobalMappingDestroy(A->rmapping);CHKERRQ(ierr);} 1035 if (A->cmapping) {ierr = ISLocalToGlobalMappingDestroy(A->cmapping);CHKERRQ(ierr);} 1036 if (A->rbmapping) {ierr = ISLocalToGlobalMappingDestroy(A->rbmapping);CHKERRQ(ierr);} 1037 if (A->cbmapping) {ierr = ISLocalToGlobalMappingDestroy(A->cbmapping);CHKERRQ(ierr);} 1038 1039 ierr = PetscFree(A->spptr);CHKERRQ(ierr); 1040 ierr = PetscLayoutDestroy(A->rmap);CHKERRQ(ierr); 1041 ierr = PetscLayoutDestroy(A->cmap);CHKERRQ(ierr); 1042 ierr = PetscHeaderDestroy(A);CHKERRQ(ierr); 1043 PetscFunctionReturn(0); 1044 } 1045 1046 #undef __FUNCT__ 1047 #define __FUNCT__ "MatSetValues" 1048 /*@ 1049 MatSetValues - Inserts or adds a block of values into a matrix. 1050 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 1051 MUST be called after all calls to MatSetValues() have been completed. 1052 1053 Not Collective 1054 1055 Input Parameters: 1056 + mat - the matrix 1057 . v - a logically two-dimensional array of values 1058 . m, idxm - the number of rows and their global indices 1059 . n, idxn - the number of columns and their global indices 1060 - addv - either ADD_VALUES or INSERT_VALUES, where 1061 ADD_VALUES adds values to any existing entries, and 1062 INSERT_VALUES replaces existing entries with new values 1063 1064 Notes: 1065 By default the values, v, are row-oriented. See MatSetOption() for other options. 1066 1067 Calls to MatSetValues() with the INSERT_VALUES and ADD_VALUES 1068 options cannot be mixed without intervening calls to the assembly 1069 routines. 1070 1071 MatSetValues() uses 0-based row and column numbers in Fortran 1072 as well as in C. 1073 1074 Negative indices may be passed in idxm and idxn, these rows and columns are 1075 simply ignored. This allows easily inserting element stiffness matrices 1076 with homogeneous Dirchlet boundary conditions that you don't want represented 1077 in the matrix. 1078 1079 Efficiency Alert: 1080 The routine MatSetValuesBlocked() may offer much better efficiency 1081 for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ). 1082 1083 Level: beginner 1084 1085 Concepts: matrices^putting entries in 1086 1087 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 1088 InsertMode, INSERT_VALUES, ADD_VALUES 1089 @*/ 1090 PetscErrorCode MatSetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv) 1091 { 1092 PetscErrorCode ierr; 1093 1094 PetscFunctionBegin; 1095 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1096 PetscValidType(mat,1); 1097 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1098 PetscValidIntPointer(idxm,3); 1099 PetscValidIntPointer(idxn,5); 1100 if (v) PetscValidDoublePointer(v,6); 1101 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1102 if (mat->insertmode == NOT_SET_VALUES) { 1103 mat->insertmode = addv; 1104 } 1105 #if defined(PETSC_USE_DEBUG) 1106 else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1107 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1108 #endif 1109 1110 if (mat->assembled) { 1111 mat->was_assembled = PETSC_TRUE; 1112 mat->assembled = PETSC_FALSE; 1113 } 1114 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1115 if (!mat->ops->setvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 1116 ierr = (*mat->ops->setvalues)(mat,m,idxm,n,idxn,v,addv);CHKERRQ(ierr); 1117 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1118 #if defined(PETSC_HAVE_CUSP) 1119 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1120 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1121 } 1122 #endif 1123 PetscFunctionReturn(0); 1124 } 1125 1126 1127 #undef __FUNCT__ 1128 #define __FUNCT__ "MatSetValuesRowLocal" 1129 /*@ 1130 MatSetValuesRowLocal - Inserts a row (block row for BAIJ matrices) of nonzero 1131 values into a matrix 1132 1133 Not Collective 1134 1135 Input Parameters: 1136 + mat - the matrix 1137 . row - the (block) row to set 1138 - v - a logically two-dimensional array of values 1139 1140 Notes: 1141 By the values, v, are column-oriented (for the block version) and sorted 1142 1143 All the nonzeros in the row must be provided 1144 1145 The matrix must have previously had its column indices set 1146 1147 The row must belong to this process 1148 1149 Level: intermediate 1150 1151 Concepts: matrices^putting entries in 1152 1153 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 1154 InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues(), MatSetValuesRow(), MatSetLocalToGlobalMapping() 1155 @*/ 1156 PetscErrorCode MatSetValuesRowLocal(Mat mat,PetscInt row,const PetscScalar v[]) 1157 { 1158 PetscErrorCode ierr; 1159 1160 PetscFunctionBegin; 1161 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1162 PetscValidType(mat,1); 1163 PetscValidScalarPointer(v,2); 1164 ierr = MatSetValuesRow(mat, mat->rmapping->indices[row],v);CHKERRQ(ierr); 1165 #if defined(PETSC_HAVE_CUSP) 1166 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1167 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1168 } 1169 #endif 1170 PetscFunctionReturn(0); 1171 } 1172 1173 #undef __FUNCT__ 1174 #define __FUNCT__ "MatSetValuesRow" 1175 /*@ 1176 MatSetValuesRow - Inserts a row (block row for BAIJ matrices) of nonzero 1177 values into a matrix 1178 1179 Not Collective 1180 1181 Input Parameters: 1182 + mat - the matrix 1183 . row - the (block) row to set 1184 - v - a logically two-dimensional array of values 1185 1186 Notes: 1187 The values, v, are column-oriented for the block version. 1188 1189 All the nonzeros in the row must be provided 1190 1191 THE MATRIX MUSAT HAVE PREVIOUSLY HAD ITS COLUMN INDICES SET. IT IS RARE THAT THIS ROUTINE IS USED, usually MatSetValues() is used. 1192 1193 The row must belong to this process 1194 1195 Level: advanced 1196 1197 Concepts: matrices^putting entries in 1198 1199 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 1200 InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues() 1201 @*/ 1202 PetscErrorCode MatSetValuesRow(Mat mat,PetscInt row,const PetscScalar v[]) 1203 { 1204 PetscErrorCode ierr; 1205 1206 PetscFunctionBegin; 1207 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1208 PetscValidType(mat,1); 1209 PetscValidScalarPointer(v,2); 1210 #if defined(PETSC_USE_DEBUG) 1211 if (mat->insertmode == ADD_VALUES) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add and insert values"); 1212 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1213 #endif 1214 mat->insertmode = INSERT_VALUES; 1215 1216 if (mat->assembled) { 1217 mat->was_assembled = PETSC_TRUE; 1218 mat->assembled = PETSC_FALSE; 1219 } 1220 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1221 if (!mat->ops->setvaluesrow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 1222 ierr = (*mat->ops->setvaluesrow)(mat,row,v);CHKERRQ(ierr); 1223 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1224 #if defined(PETSC_HAVE_CUSP) 1225 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1226 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1227 } 1228 #endif 1229 PetscFunctionReturn(0); 1230 } 1231 1232 #undef __FUNCT__ 1233 #define __FUNCT__ "MatSetValuesStencil" 1234 /*@ 1235 MatSetValuesStencil - Inserts or adds a block of values into a matrix. 1236 Using structured grid indexing 1237 1238 Not Collective 1239 1240 Input Parameters: 1241 + mat - the matrix 1242 . m - number of rows being entered 1243 . idxm - grid coordinates (and component number when dof > 1) for matrix rows being entered 1244 . n - number of columns being entered 1245 . idxn - grid coordinates (and component number when dof > 1) for matrix columns being entered 1246 . v - a logically two-dimensional array of values 1247 - addv - either ADD_VALUES or INSERT_VALUES, where 1248 ADD_VALUES adds values to any existing entries, and 1249 INSERT_VALUES replaces existing entries with new values 1250 1251 Notes: 1252 By default the values, v, are row-oriented. See MatSetOption() for other options. 1253 1254 Calls to MatSetValuesStencil() with the INSERT_VALUES and ADD_VALUES 1255 options cannot be mixed without intervening calls to the assembly 1256 routines. 1257 1258 The grid coordinates are across the entire grid, not just the local portion 1259 1260 MatSetValuesStencil() uses 0-based row and column numbers in Fortran 1261 as well as in C. 1262 1263 For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine 1264 1265 In order to use this routine you must either obtain the matrix with DMGetMatrix() 1266 or call MatSetLocalToGlobalMapping() and MatSetStencil() first. 1267 1268 The columns and rows in the stencil passed in MUST be contained within the 1269 ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example, 1270 if you create a DMDA with an overlap of one grid level and on a particular process its first 1271 local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the 1272 first i index you can use in your column and row indices in MatSetStencil() is 5. 1273 1274 In Fortran idxm and idxn should be declared as 1275 $ MatStencil idxm(4,m),idxn(4,n) 1276 and the values inserted using 1277 $ idxm(MatStencil_i,1) = i 1278 $ idxm(MatStencil_j,1) = j 1279 $ idxm(MatStencil_k,1) = k 1280 $ idxm(MatStencil_c,1) = c 1281 etc 1282 1283 For periodic boundary conditions use negative indices for values to the left (below 0; that are to be 1284 obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one 1285 etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the 1286 DMDA_BOUNDARY_PERIODIC boundary type. 1287 1288 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 1289 a single value per point) you can skip filling those indices. 1290 1291 Inspired by the structured grid interface to the HYPRE package 1292 (http://www.llnl.gov/CASC/hypre) 1293 1294 Efficiency Alert: 1295 The routine MatSetValuesBlockedStencil() may offer much better efficiency 1296 for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ). 1297 1298 Level: beginner 1299 1300 Concepts: matrices^putting entries in 1301 1302 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1303 MatSetValues(), MatSetValuesBlockedStencil(), MatSetStencil(), DMGetMatrix(), DMDAVecGetArray(), MatStencil 1304 @*/ 1305 PetscErrorCode MatSetValuesStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv) 1306 { 1307 PetscErrorCode ierr; 1308 PetscInt j,i,jdxm[128],jdxn[256],dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp; 1309 PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc); 1310 1311 PetscFunctionBegin; 1312 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1313 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1314 PetscValidType(mat,1); 1315 PetscValidIntPointer(idxm,3); 1316 PetscValidIntPointer(idxn,5); 1317 PetscValidScalarPointer(v,6); 1318 1319 if (m > 128) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Can only set 128 rows at a time; trying to set %D",m); 1320 if (n > 256) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Can only set 256 columns at a time; trying to set %D",n); 1321 1322 for (i=0; i<m; i++) { 1323 for (j=0; j<3-sdim; j++) dxm++; 1324 tmp = *dxm++ - starts[0]; 1325 for (j=0; j<dim-1; j++) { 1326 if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1327 else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1]; 1328 } 1329 if (mat->stencil.noc) dxm++; 1330 jdxm[i] = tmp; 1331 } 1332 for (i=0; i<n; i++) { 1333 for (j=0; j<3-sdim; j++) dxn++; 1334 tmp = *dxn++ - starts[0]; 1335 for (j=0; j<dim-1; j++) { 1336 if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1337 else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1]; 1338 } 1339 if (mat->stencil.noc) dxn++; 1340 jdxn[i] = tmp; 1341 } 1342 ierr = MatSetValuesLocal(mat,m,jdxm,n,jdxn,v,addv);CHKERRQ(ierr); 1343 PetscFunctionReturn(0); 1344 } 1345 1346 #undef __FUNCT__ 1347 #define __FUNCT__ "MatSetValuesBlockedStencil" 1348 /*@C 1349 MatSetValuesBlockedStencil - Inserts or adds a block of values into a matrix. 1350 Using structured grid indexing 1351 1352 Not Collective 1353 1354 Input Parameters: 1355 + mat - the matrix 1356 . m - number of rows being entered 1357 . idxm - grid coordinates for matrix rows being entered 1358 . n - number of columns being entered 1359 . idxn - grid coordinates for matrix columns being entered 1360 . v - a logically two-dimensional array of values 1361 - addv - either ADD_VALUES or INSERT_VALUES, where 1362 ADD_VALUES adds values to any existing entries, and 1363 INSERT_VALUES replaces existing entries with new values 1364 1365 Notes: 1366 By default the values, v, are row-oriented and unsorted. 1367 See MatSetOption() for other options. 1368 1369 Calls to MatSetValuesBlockedStencil() with the INSERT_VALUES and ADD_VALUES 1370 options cannot be mixed without intervening calls to the assembly 1371 routines. 1372 1373 The grid coordinates are across the entire grid, not just the local portion 1374 1375 MatSetValuesBlockedStencil() uses 0-based row and column numbers in Fortran 1376 as well as in C. 1377 1378 For setting/accessing vector values via array coordinates you can use the DMDAVecGetArray() routine 1379 1380 In order to use this routine you must either obtain the matrix with DMGetMatrix() 1381 or call MatSetBlockSize(), MatSetLocalToGlobalMapping() and MatSetStencil() first. 1382 1383 The columns and rows in the stencil passed in MUST be contained within the 1384 ghost region of the given process as set with DMDACreateXXX() or MatSetStencil(). For example, 1385 if you create a DMDA with an overlap of one grid level and on a particular process its first 1386 local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the 1387 first i index you can use in your column and row indices in MatSetStencil() is 5. 1388 1389 In Fortran idxm and idxn should be declared as 1390 $ MatStencil idxm(4,m),idxn(4,n) 1391 and the values inserted using 1392 $ idxm(MatStencil_i,1) = i 1393 $ idxm(MatStencil_j,1) = j 1394 $ idxm(MatStencil_k,1) = k 1395 etc 1396 1397 Negative indices may be passed in idxm and idxn, these rows and columns are 1398 simply ignored. This allows easily inserting element stiffness matrices 1399 with homogeneous Dirchlet boundary conditions that you don't want represented 1400 in the matrix. 1401 1402 Inspired by the structured grid interface to the HYPRE package 1403 (http://www.llnl.gov/CASC/hypre) 1404 1405 Level: beginner 1406 1407 Concepts: matrices^putting entries in 1408 1409 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1410 MatSetValues(), MatSetValuesStencil(), MatSetStencil(), DMGetMatrix(), DMDAVecGetArray(), MatStencil, 1411 MatSetBlockSize(), MatSetLocalToGlobalMapping() 1412 @*/ 1413 PetscErrorCode MatSetValuesBlockedStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv) 1414 { 1415 PetscErrorCode ierr; 1416 PetscInt j,i,jdxm[128],jdxn[256],dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp; 1417 PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc); 1418 1419 PetscFunctionBegin; 1420 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1421 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1422 PetscValidType(mat,1); 1423 PetscValidIntPointer(idxm,3); 1424 PetscValidIntPointer(idxn,5); 1425 PetscValidScalarPointer(v,6); 1426 1427 if (m > 128) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Can only set 128 rows at a time; trying to set %D",m); 1428 if (n > 128) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Can only set 256 columns at a time; trying to set %D",n); 1429 1430 for (i=0; i<m; i++) { 1431 for (j=0; j<3-sdim; j++) dxm++; 1432 tmp = *dxm++ - starts[0]; 1433 for (j=0; j<sdim-1; j++) { 1434 if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1435 else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1]; 1436 } 1437 dxm++; 1438 jdxm[i] = tmp; 1439 } 1440 for (i=0; i<n; i++) { 1441 for (j=0; j<3-sdim; j++) dxn++; 1442 tmp = *dxn++ - starts[0]; 1443 for (j=0; j<sdim-1; j++) { 1444 if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1445 else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1]; 1446 } 1447 dxn++; 1448 jdxn[i] = tmp; 1449 } 1450 ierr = MatSetValuesBlockedLocal(mat,m,jdxm,n,jdxn,v,addv);CHKERRQ(ierr); 1451 #if defined(PETSC_HAVE_CUSP) 1452 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1453 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1454 } 1455 #endif 1456 PetscFunctionReturn(0); 1457 } 1458 1459 #undef __FUNCT__ 1460 #define __FUNCT__ "MatSetStencil" 1461 /*@ 1462 MatSetStencil - Sets the grid information for setting values into a matrix via 1463 MatSetValuesStencil() 1464 1465 Not Collective 1466 1467 Input Parameters: 1468 + mat - the matrix 1469 . dim - dimension of the grid 1, 2, or 3 1470 . dims - number of grid points in x, y, and z direction, including ghost points on your processor 1471 . starts - starting point of ghost nodes on your processor in x, y, and z direction 1472 - dof - number of degrees of freedom per node 1473 1474 1475 Inspired by the structured grid interface to the HYPRE package 1476 (www.llnl.gov/CASC/hyper) 1477 1478 For matrices generated with DMGetMatrix() this routine is automatically called and so not needed by the 1479 user. 1480 1481 Level: beginner 1482 1483 Concepts: matrices^putting entries in 1484 1485 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1486 MatSetValues(), MatSetValuesBlockedStencil(), MatSetValuesStencil() 1487 @*/ 1488 PetscErrorCode MatSetStencil(Mat mat,PetscInt dim,const PetscInt dims[],const PetscInt starts[],PetscInt dof) 1489 { 1490 PetscInt i; 1491 1492 PetscFunctionBegin; 1493 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1494 PetscValidIntPointer(dims,3); 1495 PetscValidIntPointer(starts,4); 1496 1497 mat->stencil.dim = dim + (dof > 1); 1498 for (i=0; i<dim; i++) { 1499 mat->stencil.dims[i] = dims[dim-i-1]; /* copy the values in backwards */ 1500 mat->stencil.starts[i] = starts[dim-i-1]; 1501 } 1502 mat->stencil.dims[dim] = dof; 1503 mat->stencil.starts[dim] = 0; 1504 mat->stencil.noc = (PetscBool)(dof == 1); 1505 PetscFunctionReturn(0); 1506 } 1507 1508 #undef __FUNCT__ 1509 #define __FUNCT__ "MatSetValuesBlocked" 1510 /*@ 1511 MatSetValuesBlocked - Inserts or adds a block of values into a matrix. 1512 1513 Not Collective 1514 1515 Input Parameters: 1516 + mat - the matrix 1517 . v - a logically two-dimensional array of values 1518 . m, idxm - the number of block rows and their global block indices 1519 . n, idxn - the number of block columns and their global block indices 1520 - addv - either ADD_VALUES or INSERT_VALUES, where 1521 ADD_VALUES adds values to any existing entries, and 1522 INSERT_VALUES replaces existing entries with new values 1523 1524 Notes: 1525 The m and n count the NUMBER of blocks in the row direction and column direction, 1526 NOT the total number of rows/columns; for example, if the block size is 2 and 1527 you are passing in values for rows 2,3,4,5 then m would be 2 (not 4). 1528 The values in idxm would be 1 2; that is the first index for each block divided by 1529 the block size. 1530 1531 Note that you must call MatSetBlockSize() when constructing this matrix (after 1532 preallocating it). 1533 1534 By default the values, v, are row-oriented, so the layout of 1535 v is the same as for MatSetValues(). See MatSetOption() for other options. 1536 1537 Calls to MatSetValuesBlocked() with the INSERT_VALUES and ADD_VALUES 1538 options cannot be mixed without intervening calls to the assembly 1539 routines. 1540 1541 MatSetValuesBlocked() uses 0-based row and column numbers in Fortran 1542 as well as in C. 1543 1544 Negative indices may be passed in idxm and idxn, these rows and columns are 1545 simply ignored. This allows easily inserting element stiffness matrices 1546 with homogeneous Dirchlet boundary conditions that you don't want represented 1547 in the matrix. 1548 1549 Each time an entry is set within a sparse matrix via MatSetValues(), 1550 internal searching must be done to determine where to place the the 1551 data in the matrix storage space. By instead inserting blocks of 1552 entries via MatSetValuesBlocked(), the overhead of matrix assembly is 1553 reduced. 1554 1555 Example: 1556 $ Suppose m=n=2 and block size(bs) = 2 The array is 1557 $ 1558 $ 1 2 | 3 4 1559 $ 5 6 | 7 8 1560 $ - - - | - - - 1561 $ 9 10 | 11 12 1562 $ 13 14 | 15 16 1563 $ 1564 $ v[] should be passed in like 1565 $ v[] = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16] 1566 $ 1567 $ If you are not using row oriented storage of v (that is you called MatSetOption(mat,MAT_ROW_ORIENTED,PETSC_FALSE)) then 1568 $ v[] = [1,5,9,13,2,6,10,14,3,7,11,15,4,8,12,16] 1569 1570 Level: intermediate 1571 1572 Concepts: matrices^putting entries in blocked 1573 1574 .seealso: MatSetBlockSize(), MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal() 1575 @*/ 1576 PetscErrorCode MatSetValuesBlocked(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv) 1577 { 1578 PetscErrorCode ierr; 1579 1580 PetscFunctionBegin; 1581 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1582 PetscValidType(mat,1); 1583 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1584 PetscValidIntPointer(idxm,3); 1585 PetscValidIntPointer(idxn,5); 1586 PetscValidScalarPointer(v,6); 1587 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1588 if (mat->insertmode == NOT_SET_VALUES) { 1589 mat->insertmode = addv; 1590 } 1591 #if defined(PETSC_USE_DEBUG) 1592 else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1593 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1594 #endif 1595 1596 if (mat->assembled) { 1597 mat->was_assembled = PETSC_TRUE; 1598 mat->assembled = PETSC_FALSE; 1599 } 1600 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1601 if (mat->ops->setvaluesblocked) { 1602 ierr = (*mat->ops->setvaluesblocked)(mat,m,idxm,n,idxn,v,addv);CHKERRQ(ierr); 1603 } else { 1604 PetscInt buf[4096],*ibufm=0,*ibufn=0; 1605 PetscInt i,j,*iidxm,*iidxn,bs=mat->rmap->bs; 1606 if ((m+n)*bs <= 4096) { 1607 iidxm = buf; iidxn = buf + m*bs; 1608 } else { 1609 ierr = PetscMalloc2(m*bs,PetscInt,&ibufm,n*bs,PetscInt,&ibufn);CHKERRQ(ierr); 1610 iidxm = ibufm; iidxn = ibufn; 1611 } 1612 for (i=0; i<m; i++) { 1613 for (j=0; j<bs; j++) { 1614 iidxm[i*bs+j] = bs*idxm[i] + j; 1615 } 1616 } 1617 for (i=0; i<n; i++) { 1618 for (j=0; j<bs; j++) { 1619 iidxn[i*bs+j] = bs*idxn[i] + j; 1620 } 1621 } 1622 ierr = MatSetValues(mat,bs*m,iidxm,bs*n,iidxn,v,addv);CHKERRQ(ierr); 1623 ierr = PetscFree2(ibufm,ibufn);CHKERRQ(ierr); 1624 } 1625 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1626 #if defined(PETSC_HAVE_CUSP) 1627 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1628 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1629 } 1630 #endif 1631 PetscFunctionReturn(0); 1632 } 1633 1634 #undef __FUNCT__ 1635 #define __FUNCT__ "MatGetValues" 1636 /*@ 1637 MatGetValues - Gets a block of values from a matrix. 1638 1639 Not Collective; currently only returns a local block 1640 1641 Input Parameters: 1642 + mat - the matrix 1643 . v - a logically two-dimensional array for storing the values 1644 . m, idxm - the number of rows and their global indices 1645 - n, idxn - the number of columns and their global indices 1646 1647 Notes: 1648 The user must allocate space (m*n PetscScalars) for the values, v. 1649 The values, v, are then returned in a row-oriented format, 1650 analogous to that used by default in MatSetValues(). 1651 1652 MatGetValues() uses 0-based row and column numbers in 1653 Fortran as well as in C. 1654 1655 MatGetValues() requires that the matrix has been assembled 1656 with MatAssemblyBegin()/MatAssemblyEnd(). Thus, calls to 1657 MatSetValues() and MatGetValues() CANNOT be made in succession 1658 without intermediate matrix assembly. 1659 1660 Negative row or column indices will be ignored and those locations in v[] will be 1661 left unchanged. 1662 1663 Level: advanced 1664 1665 Concepts: matrices^accessing values 1666 1667 .seealso: MatGetRow(), MatGetSubMatrices(), MatSetValues() 1668 @*/ 1669 PetscErrorCode MatGetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[]) 1670 { 1671 PetscErrorCode ierr; 1672 1673 PetscFunctionBegin; 1674 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1675 PetscValidType(mat,1); 1676 if (!m || !n) PetscFunctionReturn(0); 1677 PetscValidIntPointer(idxm,3); 1678 PetscValidIntPointer(idxn,5); 1679 PetscValidScalarPointer(v,6); 1680 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1681 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1682 if (!mat->ops->getvalues) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 1683 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1684 1685 ierr = PetscLogEventBegin(MAT_GetValues,mat,0,0,0);CHKERRQ(ierr); 1686 ierr = (*mat->ops->getvalues)(mat,m,idxm,n,idxn,v);CHKERRQ(ierr); 1687 ierr = PetscLogEventEnd(MAT_GetValues,mat,0,0,0);CHKERRQ(ierr); 1688 PetscFunctionReturn(0); 1689 } 1690 1691 #undef __FUNCT__ 1692 #define __FUNCT__ "MatSetLocalToGlobalMapping" 1693 /*@ 1694 MatSetLocalToGlobalMapping - Sets a local-to-global numbering for use by 1695 the routine MatSetValuesLocal() to allow users to insert matrix entries 1696 using a local (per-processor) numbering. 1697 1698 Not Collective 1699 1700 Input Parameters: 1701 + x - the matrix 1702 . rmapping - row mapping created with ISLocalToGlobalMappingCreate() 1703 or ISLocalToGlobalMappingCreateIS() 1704 - cmapping - column mapping 1705 1706 Level: intermediate 1707 1708 Concepts: matrices^local to global mapping 1709 Concepts: local to global mapping^for matrices 1710 1711 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesLocal() 1712 @*/ 1713 PetscErrorCode MatSetLocalToGlobalMapping(Mat x,ISLocalToGlobalMapping rmapping,ISLocalToGlobalMapping cmapping) 1714 { 1715 PetscErrorCode ierr; 1716 PetscFunctionBegin; 1717 PetscValidHeaderSpecific(x,MAT_CLASSID,1); 1718 PetscValidType(x,1); 1719 PetscValidHeaderSpecific(rmapping,IS_LTOGM_CLASSID,2); 1720 PetscValidHeaderSpecific(cmapping,IS_LTOGM_CLASSID,3); 1721 if (x->rmapping) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Mapping already set for matrix"); 1722 ierr = MatPreallocated(x);CHKERRQ(ierr); 1723 1724 if (x->ops->setlocaltoglobalmapping) { 1725 ierr = (*x->ops->setlocaltoglobalmapping)(x,rmapping,cmapping);CHKERRQ(ierr); 1726 } else { 1727 ierr = PetscObjectReference((PetscObject)rmapping);CHKERRQ(ierr); 1728 if (x->rmapping) { ierr = ISLocalToGlobalMappingDestroy(x->rmapping);CHKERRQ(ierr); } 1729 x->rmapping = rmapping; 1730 ierr = PetscObjectReference((PetscObject)cmapping);CHKERRQ(ierr); 1731 if (x->cmapping) { ierr = ISLocalToGlobalMappingDestroy(x->cmapping);CHKERRQ(ierr); } 1732 x->cmapping = cmapping; 1733 } 1734 PetscFunctionReturn(0); 1735 } 1736 1737 #undef __FUNCT__ 1738 #define __FUNCT__ "MatSetLocalToGlobalMappingBlock" 1739 /*@ 1740 MatSetLocalToGlobalMappingBlock - Sets a local-to-global numbering for use 1741 by the routine MatSetValuesBlockedLocal() to allow users to insert matrix 1742 entries using a local (per-processor) numbering. 1743 1744 Not Collective 1745 1746 Input Parameters: 1747 + x - the matrix 1748 . rmapping - row mapping created with ISLocalToGlobalMappingCreate() or 1749 ISLocalToGlobalMappingCreateIS() 1750 - cmapping - column mapping 1751 1752 Level: intermediate 1753 1754 Concepts: matrices^local to global mapping blocked 1755 Concepts: local to global mapping^for matrices, blocked 1756 1757 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal(), 1758 MatSetValuesBlocked(), MatSetValuesLocal() 1759 @*/ 1760 PetscErrorCode MatSetLocalToGlobalMappingBlock(Mat x,ISLocalToGlobalMapping rmapping,ISLocalToGlobalMapping cmapping) 1761 { 1762 PetscErrorCode ierr; 1763 PetscFunctionBegin; 1764 PetscValidHeaderSpecific(x,MAT_CLASSID,1); 1765 PetscValidType(x,1); 1766 PetscValidHeaderSpecific(rmapping,IS_LTOGM_CLASSID,2); 1767 PetscValidHeaderSpecific(cmapping,IS_LTOGM_CLASSID,3); 1768 if (x->rbmapping) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Mapping already set for matrix"); 1769 ierr = PetscObjectReference((PetscObject)rmapping);CHKERRQ(ierr); 1770 if (x->rbmapping) { ierr = ISLocalToGlobalMappingDestroy(x->rbmapping);CHKERRQ(ierr); } 1771 x->rbmapping = rmapping; 1772 ierr = PetscObjectReference((PetscObject)cmapping);CHKERRQ(ierr); 1773 if (x->cbmapping) { ierr = ISLocalToGlobalMappingDestroy(x->cbmapping);CHKERRQ(ierr); } 1774 x->cbmapping = cmapping; 1775 PetscFunctionReturn(0); 1776 } 1777 1778 #undef __FUNCT__ 1779 #define __FUNCT__ "MatGetLocalToGlobalMapping" 1780 /*@ 1781 MatGetLocalToGlobalMapping - Gets the local-to-global numbering set by MatSetLocalToGlobalMapping() 1782 1783 Not Collective 1784 1785 Input Parameters: 1786 . A - the matrix 1787 1788 Output Parameters: 1789 + rmapping - row mapping 1790 - cmapping - column mapping 1791 1792 Level: advanced 1793 1794 Concepts: matrices^local to global mapping 1795 Concepts: local to global mapping^for matrices 1796 1797 .seealso: MatSetValuesLocal(), MatGetLocalToGlobalMappingBlock() 1798 @*/ 1799 PetscErrorCode MatGetLocalToGlobalMapping(Mat A,ISLocalToGlobalMapping *rmapping,ISLocalToGlobalMapping *cmapping) 1800 { 1801 PetscFunctionBegin; 1802 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 1803 PetscValidType(A,1); 1804 if (rmapping) PetscValidPointer(rmapping,2); 1805 if (cmapping) PetscValidPointer(cmapping,3); 1806 if (rmapping) *rmapping = A->rmapping; 1807 if (cmapping) *cmapping = A->cmapping; 1808 PetscFunctionReturn(0); 1809 } 1810 1811 #undef __FUNCT__ 1812 #define __FUNCT__ "MatGetLocalToGlobalMappingBlock" 1813 /*@ 1814 MatGetLocalToGlobalMappingBlock - Gets the local-to-global numbering set by MatSetLocalToGlobalMappingBlock() 1815 1816 Not Collective 1817 1818 Input Parameters: 1819 . A - the matrix 1820 1821 Output Parameters: 1822 + rmapping - row mapping 1823 - cmapping - column mapping 1824 1825 Level: advanced 1826 1827 Concepts: matrices^local to global mapping blocked 1828 Concepts: local to global mapping^for matrices, blocked 1829 1830 .seealso: MatSetValuesBlockedLocal(), MatGetLocalToGlobalMapping() 1831 @*/ 1832 PetscErrorCode MatGetLocalToGlobalMappingBlock(Mat A,ISLocalToGlobalMapping *rmapping,ISLocalToGlobalMapping *cmapping) 1833 { 1834 PetscFunctionBegin; 1835 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 1836 PetscValidType(A,1); 1837 if (rmapping) PetscValidPointer(rmapping,2); 1838 if (cmapping) PetscValidPointer(cmapping,3); 1839 if (rmapping) *rmapping = A->rbmapping; 1840 if (cmapping) *cmapping = A->cbmapping; 1841 PetscFunctionReturn(0); 1842 } 1843 1844 #undef __FUNCT__ 1845 #define __FUNCT__ "MatSetValuesLocal" 1846 /*@ 1847 MatSetValuesLocal - Inserts or adds values into certain locations of a matrix, 1848 using a local ordering of the nodes. 1849 1850 Not Collective 1851 1852 Input Parameters: 1853 + x - the matrix 1854 . nrow, irow - number of rows and their local indices 1855 . ncol, icol - number of columns and their local indices 1856 . y - a logically two-dimensional array of values 1857 - addv - either INSERT_VALUES or ADD_VALUES, where 1858 ADD_VALUES adds values to any existing entries, and 1859 INSERT_VALUES replaces existing entries with new values 1860 1861 Notes: 1862 Before calling MatSetValuesLocal(), the user must first set the 1863 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 1864 1865 Calls to MatSetValuesLocal() with the INSERT_VALUES and ADD_VALUES 1866 options cannot be mixed without intervening calls to the assembly 1867 routines. 1868 1869 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 1870 MUST be called after all calls to MatSetValuesLocal() have been completed. 1871 1872 Level: intermediate 1873 1874 Concepts: matrices^putting entries in with local numbering 1875 1876 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetLocalToGlobalMapping(), 1877 MatSetValueLocal() 1878 @*/ 1879 PetscErrorCode MatSetValuesLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv) 1880 { 1881 PetscErrorCode ierr; 1882 PetscInt irowm[2048],icolm[2048]; 1883 1884 PetscFunctionBegin; 1885 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1886 PetscValidType(mat,1); 1887 if (!nrow || !ncol) PetscFunctionReturn(0); /* no values to insert */ 1888 PetscValidIntPointer(irow,3); 1889 PetscValidIntPointer(icol,5); 1890 PetscValidScalarPointer(y,6); 1891 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1892 if (mat->insertmode == NOT_SET_VALUES) { 1893 mat->insertmode = addv; 1894 } 1895 #if defined(PETSC_USE_DEBUG) 1896 else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1897 if (!mat->ops->setvalueslocal && (nrow > 2048 || ncol > 2048)) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_SUP,"Number column/row indices must be <= 2048: are %D %D",nrow,ncol); 1898 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1899 #endif 1900 1901 if (mat->assembled) { 1902 mat->was_assembled = PETSC_TRUE; 1903 mat->assembled = PETSC_FALSE; 1904 } 1905 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1906 if (!mat->ops->setvalueslocal) { 1907 ierr = ISLocalToGlobalMappingApply(mat->rmapping,nrow,irow,irowm);CHKERRQ(ierr); 1908 ierr = ISLocalToGlobalMappingApply(mat->cmapping,ncol,icol,icolm);CHKERRQ(ierr); 1909 ierr = (*mat->ops->setvalues)(mat,nrow,irowm,ncol,icolm,y,addv);CHKERRQ(ierr); 1910 } else { 1911 ierr = (*mat->ops->setvalueslocal)(mat,nrow,irow,ncol,icol,y,addv);CHKERRQ(ierr); 1912 } 1913 mat->same_nonzero = PETSC_FALSE; 1914 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1915 #if defined(PETSC_HAVE_CUSP) 1916 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 1917 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 1918 } 1919 #endif 1920 PetscFunctionReturn(0); 1921 } 1922 1923 #undef __FUNCT__ 1924 #define __FUNCT__ "MatSetValuesBlockedLocal" 1925 /*@ 1926 MatSetValuesBlockedLocal - Inserts or adds values into certain locations of a matrix, 1927 using a local ordering of the nodes a block at a time. 1928 1929 Not Collective 1930 1931 Input Parameters: 1932 + x - the matrix 1933 . nrow, irow - number of rows and their local indices 1934 . ncol, icol - number of columns and their local indices 1935 . y - a logically two-dimensional array of values 1936 - addv - either INSERT_VALUES or ADD_VALUES, where 1937 ADD_VALUES adds values to any existing entries, and 1938 INSERT_VALUES replaces existing entries with new values 1939 1940 Notes: 1941 Before calling MatSetValuesBlockedLocal(), the user must first set the 1942 block size using MatSetBlockSize(), and the local-to-global mapping by 1943 calling MatSetLocalToGlobalMappingBlock(), where the mapping MUST be 1944 set for matrix blocks, not for matrix elements. 1945 1946 Calls to MatSetValuesBlockedLocal() with the INSERT_VALUES and ADD_VALUES 1947 options cannot be mixed without intervening calls to the assembly 1948 routines. 1949 1950 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 1951 MUST be called after all calls to MatSetValuesBlockedLocal() have been completed. 1952 1953 Level: intermediate 1954 1955 Concepts: matrices^putting blocked values in with local numbering 1956 1957 .seealso: MatSetBlockSize(), MatSetLocalToGlobalMappingBlock(), MatAssemblyBegin(), MatAssemblyEnd(), 1958 MatSetValuesLocal(), MatSetLocalToGlobalMappingBlock(), MatSetValuesBlocked() 1959 @*/ 1960 PetscErrorCode MatSetValuesBlockedLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv) 1961 { 1962 PetscErrorCode ierr; 1963 PetscInt irowm[2048],icolm[2048]; 1964 1965 PetscFunctionBegin; 1966 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 1967 PetscValidType(mat,1); 1968 if (!nrow || !ncol) PetscFunctionReturn(0); /* no values to insert */ 1969 PetscValidIntPointer(irow,3); 1970 PetscValidIntPointer(icol,5); 1971 PetscValidScalarPointer(y,6); 1972 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1973 if (mat->insertmode == NOT_SET_VALUES) { 1974 mat->insertmode = addv; 1975 } 1976 #if defined(PETSC_USE_DEBUG) 1977 else if (mat->insertmode != addv) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1978 if (nrow > 2048 || ncol > 2048) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_SUP,"Number column/row indices must be <= 2048: are %D %D",nrow,ncol); 1979 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1980 #endif 1981 1982 if (mat->assembled) { 1983 mat->was_assembled = PETSC_TRUE; 1984 mat->assembled = PETSC_FALSE; 1985 } 1986 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1987 if (mat->ops->setvaluesblockedlocal) { 1988 ierr = (*mat->ops->setvaluesblockedlocal)(mat,nrow,irow,ncol,icol,y,addv);CHKERRQ(ierr); 1989 } else if (mat->rbmapping && mat->cbmapping) { 1990 ierr = ISLocalToGlobalMappingApply(mat->rbmapping,nrow,irow,irowm);CHKERRQ(ierr); 1991 ierr = ISLocalToGlobalMappingApply(mat->cbmapping,ncol,icol,icolm);CHKERRQ(ierr); 1992 if (mat->ops->setvaluesblocked) { 1993 ierr = (*mat->ops->setvaluesblocked)(mat,nrow,irowm,ncol,icolm,y,addv);CHKERRQ(ierr); 1994 } else { 1995 PetscInt buf[4096],*ibufm=0,*ibufn=0; 1996 PetscInt i,j,*iirowm,*iicolm,bs=mat->rmap->bs; 1997 if ((nrow+ncol)*bs <= 4096) { 1998 iirowm = buf; iicolm = buf + nrow*bs; 1999 } else { 2000 ierr = PetscMalloc2(nrow*bs,PetscInt,&ibufm,ncol*bs,PetscInt,&ibufn);CHKERRQ(ierr); 2001 iirowm = ibufm; iicolm = ibufn; 2002 } 2003 for (i=0; i<nrow; i++) { 2004 for (j=0; j<bs; j++) { 2005 iirowm[i*bs+j] = bs*irowm[i] + j; 2006 } 2007 } 2008 for (i=0; i<ncol; i++) { 2009 for (j=0; j<bs; j++) { 2010 iicolm[i*bs+j] = bs*icolm[i] + j; 2011 } 2012 } 2013 ierr = MatSetValues(mat,bs*nrow,iirowm,bs*ncol,iicolm,y,addv);CHKERRQ(ierr); 2014 ierr = PetscFree2(ibufm,ibufn);CHKERRQ(ierr); 2015 } 2016 } else { 2017 PetscInt buf[4096],*ibufm=0,*ibufn=0; 2018 PetscInt i,j,*iirowm,*iicolm,bs=mat->rmap->bs; 2019 if ((nrow+ncol)*bs <= 4096) { 2020 iirowm = buf; iicolm = buf + nrow*bs; 2021 } else { 2022 ierr = PetscMalloc2(nrow*bs,PetscInt,&ibufm,ncol*bs,PetscInt,&ibufn);CHKERRQ(ierr); 2023 iirowm = ibufm; iicolm = ibufn; 2024 } 2025 for (i=0; i<nrow; i++) { 2026 for (j=0; j<bs; j++) iirowm[i*bs+j] = irow[i]*bs+j; 2027 } 2028 for (i=0; i<ncol; i++) { 2029 for (j=0; j<bs; j++) iicolm[i*bs+j] = icol[i]*bs+j; 2030 } 2031 ierr = MatSetValuesLocal(mat,nrow*bs,iirowm,ncol*bs,iicolm,y,addv);CHKERRQ(ierr); 2032 ierr = PetscFree2(ibufm,ibufn);CHKERRQ(ierr); 2033 } 2034 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 2035 #if defined(PETSC_HAVE_CUSP) 2036 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 2037 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 2038 } 2039 #endif 2040 PetscFunctionReturn(0); 2041 } 2042 2043 #undef __FUNCT__ 2044 #define __FUNCT__ "MatMultDiagonalBlock" 2045 /*@ 2046 MatMultDiagonalBlock - Computes the matrix-vector product, y = Dx. Where D is defined by the inode or block structure of the diagonal 2047 2048 Collective on Mat and Vec 2049 2050 Input Parameters: 2051 + mat - the matrix 2052 - x - the vector to be multiplied 2053 2054 Output Parameters: 2055 . y - the result 2056 2057 Notes: 2058 The vectors x and y cannot be the same. I.e., one cannot 2059 call MatMult(A,y,y). 2060 2061 Level: developer 2062 2063 Concepts: matrix-vector product 2064 2065 .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 2066 @*/ 2067 PetscErrorCode MatMultDiagonalBlock(Mat mat,Vec x,Vec y) 2068 { 2069 PetscErrorCode ierr; 2070 2071 PetscFunctionBegin; 2072 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2073 PetscValidType(mat,1); 2074 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2075 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2076 2077 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2078 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2079 if (x == y) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2080 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2081 2082 if (!mat->ops->multdiagonalblock) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"This matrix type does not have a multiply defined"); 2083 ierr = (*mat->ops->multdiagonalblock)(mat,x,y);CHKERRQ(ierr); 2084 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2085 PetscFunctionReturn(0); 2086 } 2087 2088 /* --------------------------------------------------------*/ 2089 #undef __FUNCT__ 2090 #define __FUNCT__ "MatMult" 2091 /*@ 2092 MatMult - Computes the matrix-vector product, y = Ax. 2093 2094 Neighbor-wise Collective on Mat and Vec 2095 2096 Input Parameters: 2097 + mat - the matrix 2098 - x - the vector to be multiplied 2099 2100 Output Parameters: 2101 . y - the result 2102 2103 Notes: 2104 The vectors x and y cannot be the same. I.e., one cannot 2105 call MatMult(A,y,y). 2106 2107 Level: beginner 2108 2109 Concepts: matrix-vector product 2110 2111 .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 2112 @*/ 2113 PetscErrorCode MatMult(Mat mat,Vec x,Vec y) 2114 { 2115 PetscErrorCode ierr; 2116 2117 PetscFunctionBegin; 2118 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2119 PetscValidType(mat,1); 2120 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2121 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2122 2123 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2124 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2125 if (x == y) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2126 #ifndef PETSC_HAVE_CONSTRAINTS 2127 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); 2128 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); 2129 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); 2130 #endif 2131 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2132 2133 if (mat->nullsp) { 2134 ierr = MatNullSpaceRemove(mat->nullsp,x,&x);CHKERRQ(ierr); 2135 } 2136 2137 if (!mat->ops->mult) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"This matrix type does not have a multiply defined"); 2138 ierr = PetscLogEventBegin(MAT_Mult,mat,x,y,0);CHKERRQ(ierr); 2139 ierr = (*mat->ops->mult)(mat,x,y);CHKERRQ(ierr); 2140 ierr = PetscLogEventEnd(MAT_Mult,mat,x,y,0);CHKERRQ(ierr); 2141 2142 if (mat->nullsp) { 2143 ierr = MatNullSpaceRemove(mat->nullsp,y,PETSC_NULL);CHKERRQ(ierr); 2144 } 2145 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2146 PetscFunctionReturn(0); 2147 } 2148 2149 #undef __FUNCT__ 2150 #define __FUNCT__ "MatMultTranspose" 2151 /*@ 2152 MatMultTranspose - Computes matrix transpose times a vector. 2153 2154 Neighbor-wise Collective on Mat and Vec 2155 2156 Input Parameters: 2157 + mat - the matrix 2158 - x - the vector to be multilplied 2159 2160 Output Parameters: 2161 . y - the result 2162 2163 Notes: 2164 The vectors x and y cannot be the same. I.e., one cannot 2165 call MatMultTranspose(A,y,y). 2166 2167 Level: beginner 2168 2169 Concepts: matrix vector product^transpose 2170 2171 .seealso: MatMult(), MatMultAdd(), MatMultTransposeAdd() 2172 @*/ 2173 PetscErrorCode MatMultTranspose(Mat mat,Vec x,Vec y) 2174 { 2175 PetscErrorCode ierr; 2176 2177 PetscFunctionBegin; 2178 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2179 PetscValidType(mat,1); 2180 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2181 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2182 2183 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2184 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2185 if (x == y) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2186 #ifndef PETSC_HAVE_CONSTRAINTS 2187 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); 2188 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); 2189 #endif 2190 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2191 2192 if (!mat->ops->multtranspose) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"This matrix type does not have a multiply tranpose defined"); 2193 ierr = PetscLogEventBegin(MAT_MultTranspose,mat,x,y,0);CHKERRQ(ierr); 2194 ierr = (*mat->ops->multtranspose)(mat,x,y);CHKERRQ(ierr); 2195 ierr = PetscLogEventEnd(MAT_MultTranspose,mat,x,y,0);CHKERRQ(ierr); 2196 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2197 PetscFunctionReturn(0); 2198 } 2199 2200 #undef __FUNCT__ 2201 #define __FUNCT__ "MatMultHermitianTranspose" 2202 /*@ 2203 MatMultHermitianTranspose - Computes matrix Hermitian transpose times a vector. 2204 2205 Neighbor-wise Collective on Mat and Vec 2206 2207 Input Parameters: 2208 + mat - the matrix 2209 - x - the vector to be multilplied 2210 2211 Output Parameters: 2212 . y - the result 2213 2214 Notes: 2215 The vectors x and y cannot be the same. I.e., one cannot 2216 call MatMultHermitianTranspose(A,y,y). 2217 2218 Level: beginner 2219 2220 Concepts: matrix vector product^transpose 2221 2222 .seealso: MatMult(), MatMultAdd(), MatMultHermitianTransposeAdd(), MatMultTranspose() 2223 @*/ 2224 PetscErrorCode MatMultHermitianTranspose(Mat mat,Vec x,Vec y) 2225 { 2226 PetscErrorCode ierr; 2227 2228 PetscFunctionBegin; 2229 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2230 PetscValidType(mat,1); 2231 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2232 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2233 2234 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2235 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2236 if (x == y) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2237 #ifndef PETSC_HAVE_CONSTRAINTS 2238 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); 2239 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); 2240 #endif 2241 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2242 2243 if (!mat->ops->multhermitiantranspose) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2244 ierr = PetscLogEventBegin(MAT_MultHermitianTranspose,mat,x,y,0);CHKERRQ(ierr); 2245 ierr = (*mat->ops->multhermitiantranspose)(mat,x,y);CHKERRQ(ierr); 2246 ierr = PetscLogEventEnd(MAT_MultHermitianTranspose,mat,x,y,0);CHKERRQ(ierr); 2247 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2248 PetscFunctionReturn(0); 2249 } 2250 2251 #undef __FUNCT__ 2252 #define __FUNCT__ "MatMultAdd" 2253 /*@ 2254 MatMultAdd - Computes v3 = v2 + A * v1. 2255 2256 Neighbor-wise Collective on Mat and Vec 2257 2258 Input Parameters: 2259 + mat - the matrix 2260 - v1, v2 - the vectors 2261 2262 Output Parameters: 2263 . v3 - the result 2264 2265 Notes: 2266 The vectors v1 and v3 cannot be the same. I.e., one cannot 2267 call MatMultAdd(A,v1,v2,v1). 2268 2269 Level: beginner 2270 2271 Concepts: matrix vector product^addition 2272 2273 .seealso: MatMultTranspose(), MatMult(), MatMultTransposeAdd() 2274 @*/ 2275 PetscErrorCode MatMultAdd(Mat mat,Vec v1,Vec v2,Vec v3) 2276 { 2277 PetscErrorCode ierr; 2278 2279 PetscFunctionBegin; 2280 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2281 PetscValidType(mat,1); 2282 PetscValidHeaderSpecific(v1,VEC_CLASSID,2); 2283 PetscValidHeaderSpecific(v2,VEC_CLASSID,3); 2284 PetscValidHeaderSpecific(v3,VEC_CLASSID,4); 2285 2286 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2287 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2288 if (mat->cmap->N != v1->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->cmap->N,v1->map->N); 2289 /* 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); 2290 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); */ 2291 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); 2292 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); 2293 if (v1 == v3) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors"); 2294 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2295 2296 if (!mat->ops->multadd) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"No MatMultAdd() for this matrix type"); 2297 ierr = PetscLogEventBegin(MAT_MultAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2298 ierr = (*mat->ops->multadd)(mat,v1,v2,v3);CHKERRQ(ierr); 2299 ierr = PetscLogEventEnd(MAT_MultAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2300 ierr = PetscObjectStateIncrease((PetscObject)v3);CHKERRQ(ierr); 2301 PetscFunctionReturn(0); 2302 } 2303 2304 #undef __FUNCT__ 2305 #define __FUNCT__ "MatMultTransposeAdd" 2306 /*@ 2307 MatMultTransposeAdd - Computes v3 = v2 + A' * v1. 2308 2309 Neighbor-wise Collective on Mat and Vec 2310 2311 Input Parameters: 2312 + mat - the matrix 2313 - v1, v2 - the vectors 2314 2315 Output Parameters: 2316 . v3 - the result 2317 2318 Notes: 2319 The vectors v1 and v3 cannot be the same. I.e., one cannot 2320 call MatMultTransposeAdd(A,v1,v2,v1). 2321 2322 Level: beginner 2323 2324 Concepts: matrix vector product^transpose and addition 2325 2326 .seealso: MatMultTranspose(), MatMultAdd(), MatMult() 2327 @*/ 2328 PetscErrorCode MatMultTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3) 2329 { 2330 PetscErrorCode ierr; 2331 2332 PetscFunctionBegin; 2333 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2334 PetscValidType(mat,1); 2335 PetscValidHeaderSpecific(v1,VEC_CLASSID,2); 2336 PetscValidHeaderSpecific(v2,VEC_CLASSID,3); 2337 PetscValidHeaderSpecific(v3,VEC_CLASSID,4); 2338 2339 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2340 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2341 if (!mat->ops->multtransposeadd) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2342 if (v1 == v3) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors"); 2343 if (mat->rmap->N != v1->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->rmap->N,v1->map->N); 2344 if (mat->cmap->N != v2->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->cmap->N,v2->map->N); 2345 if (mat->cmap->N != v3->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->cmap->N,v3->map->N); 2346 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2347 2348 ierr = PetscLogEventBegin(MAT_MultTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2349 ierr = (*mat->ops->multtransposeadd)(mat,v1,v2,v3);CHKERRQ(ierr); 2350 ierr = PetscLogEventEnd(MAT_MultTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2351 ierr = PetscObjectStateIncrease((PetscObject)v3);CHKERRQ(ierr); 2352 PetscFunctionReturn(0); 2353 } 2354 2355 #undef __FUNCT__ 2356 #define __FUNCT__ "MatMultHermitianTransposeAdd" 2357 /*@ 2358 MatMultHermitianTransposeAdd - Computes v3 = v2 + A^H * v1. 2359 2360 Neighbor-wise Collective on Mat and Vec 2361 2362 Input Parameters: 2363 + mat - the matrix 2364 - v1, v2 - the vectors 2365 2366 Output Parameters: 2367 . v3 - the result 2368 2369 Notes: 2370 The vectors v1 and v3 cannot be the same. I.e., one cannot 2371 call MatMultHermitianTransposeAdd(A,v1,v2,v1). 2372 2373 Level: beginner 2374 2375 Concepts: matrix vector product^transpose and addition 2376 2377 .seealso: MatMultHermitianTranspose(), MatMultTranspose(), MatMultAdd(), MatMult() 2378 @*/ 2379 PetscErrorCode MatMultHermitianTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3) 2380 { 2381 PetscErrorCode ierr; 2382 2383 PetscFunctionBegin; 2384 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2385 PetscValidType(mat,1); 2386 PetscValidHeaderSpecific(v1,VEC_CLASSID,2); 2387 PetscValidHeaderSpecific(v2,VEC_CLASSID,3); 2388 PetscValidHeaderSpecific(v3,VEC_CLASSID,4); 2389 2390 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2391 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2392 if (!mat->ops->multhermitiantransposeadd) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2393 if (v1 == v3) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors"); 2394 if (mat->rmap->N != v1->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->rmap->N,v1->map->N); 2395 if (mat->cmap->N != v2->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->cmap->N,v2->map->N); 2396 if (mat->cmap->N != v3->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->cmap->N,v3->map->N); 2397 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2398 2399 ierr = PetscLogEventBegin(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2400 ierr = (*mat->ops->multhermitiantransposeadd)(mat,v1,v2,v3);CHKERRQ(ierr); 2401 ierr = PetscLogEventEnd(MAT_MultHermitianTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 2402 ierr = PetscObjectStateIncrease((PetscObject)v3);CHKERRQ(ierr); 2403 PetscFunctionReturn(0); 2404 } 2405 2406 #undef __FUNCT__ 2407 #define __FUNCT__ "MatMultConstrained" 2408 /*@ 2409 MatMultConstrained - The inner multiplication routine for a 2410 constrained matrix P^T A P. 2411 2412 Neighbor-wise Collective on Mat and Vec 2413 2414 Input Parameters: 2415 + mat - the matrix 2416 - x - the vector to be multilplied 2417 2418 Output Parameters: 2419 . y - the result 2420 2421 Notes: 2422 The vectors x and y cannot be the same. I.e., one cannot 2423 call MatMult(A,y,y). 2424 2425 Level: beginner 2426 2427 .keywords: matrix, multiply, matrix-vector product, constraint 2428 .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 2429 @*/ 2430 PetscErrorCode MatMultConstrained(Mat mat,Vec x,Vec y) 2431 { 2432 PetscErrorCode ierr; 2433 2434 PetscFunctionBegin; 2435 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2436 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2437 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2438 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2439 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2440 if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2441 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); 2442 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); 2443 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); 2444 2445 ierr = PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 2446 ierr = (*mat->ops->multconstrained)(mat,x,y);CHKERRQ(ierr); 2447 ierr = PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 2448 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2449 2450 PetscFunctionReturn(0); 2451 } 2452 2453 #undef __FUNCT__ 2454 #define __FUNCT__ "MatMultTransposeConstrained" 2455 /*@ 2456 MatMultTransposeConstrained - The inner multiplication routine for a 2457 constrained matrix P^T A^T P. 2458 2459 Neighbor-wise Collective on Mat and Vec 2460 2461 Input Parameters: 2462 + mat - the matrix 2463 - x - the vector to be multilplied 2464 2465 Output Parameters: 2466 . y - the result 2467 2468 Notes: 2469 The vectors x and y cannot be the same. I.e., one cannot 2470 call MatMult(A,y,y). 2471 2472 Level: beginner 2473 2474 .keywords: matrix, multiply, matrix-vector product, constraint 2475 .seealso: MatMult(), MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 2476 @*/ 2477 PetscErrorCode MatMultTransposeConstrained(Mat mat,Vec x,Vec y) 2478 { 2479 PetscErrorCode ierr; 2480 2481 PetscFunctionBegin; 2482 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2483 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 2484 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2485 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2486 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2487 if (x == y) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 2488 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); 2489 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); 2490 2491 ierr = PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 2492 ierr = (*mat->ops->multtransposeconstrained)(mat,x,y);CHKERRQ(ierr); 2493 ierr = PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 2494 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 2495 2496 PetscFunctionReturn(0); 2497 } 2498 2499 #undef __FUNCT__ 2500 #define __FUNCT__ "MatGetFactorType" 2501 /*@C 2502 MatGetFactorType - gets the type of factorization it is 2503 2504 Note Collective 2505 as the flag 2506 2507 Input Parameters: 2508 . mat - the matrix 2509 2510 Output Parameters: 2511 . t - the type, one of MAT_FACTOR_NONE, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ILU, MAT_FACTOR_ICC,MAT_FACTOR_ILUDT 2512 2513 Level: intermediate 2514 2515 .seealso: MatFactorType, MatGetFactor() 2516 @*/ 2517 PetscErrorCode MatGetFactorType(Mat mat,MatFactorType *t) 2518 { 2519 PetscFunctionBegin; 2520 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2521 PetscValidType(mat,1); 2522 *t = mat->factortype; 2523 PetscFunctionReturn(0); 2524 } 2525 2526 /* ------------------------------------------------------------*/ 2527 #undef __FUNCT__ 2528 #define __FUNCT__ "MatGetInfo" 2529 /*@C 2530 MatGetInfo - Returns information about matrix storage (number of 2531 nonzeros, memory, etc.). 2532 2533 Collective on Mat if MAT_GLOBAL_MAX or MAT_GLOBAL_SUM is used as the flag 2534 2535 Input Parameters: 2536 . mat - the matrix 2537 2538 Output Parameters: 2539 + flag - flag indicating the type of parameters to be returned 2540 (MAT_LOCAL - local matrix, MAT_GLOBAL_MAX - maximum over all processors, 2541 MAT_GLOBAL_SUM - sum over all processors) 2542 - info - matrix information context 2543 2544 Notes: 2545 The MatInfo context contains a variety of matrix data, including 2546 number of nonzeros allocated and used, number of mallocs during 2547 matrix assembly, etc. Additional information for factored matrices 2548 is provided (such as the fill ratio, number of mallocs during 2549 factorization, etc.). Much of this info is printed to PETSC_STDOUT 2550 when using the runtime options 2551 $ -info -mat_view_info 2552 2553 Example for C/C++ Users: 2554 See the file ${PETSC_DIR}/include/petscmat.h for a complete list of 2555 data within the MatInfo context. For example, 2556 .vb 2557 MatInfo info; 2558 Mat A; 2559 double mal, nz_a, nz_u; 2560 2561 MatGetInfo(A,MAT_LOCAL,&info); 2562 mal = info.mallocs; 2563 nz_a = info.nz_allocated; 2564 .ve 2565 2566 Example for Fortran Users: 2567 Fortran users should declare info as a double precision 2568 array of dimension MAT_INFO_SIZE, and then extract the parameters 2569 of interest. See the file ${PETSC_DIR}/include/finclude/petscmat.h 2570 a complete list of parameter names. 2571 .vb 2572 double precision info(MAT_INFO_SIZE) 2573 double precision mal, nz_a 2574 Mat A 2575 integer ierr 2576 2577 call MatGetInfo(A,MAT_LOCAL,info,ierr) 2578 mal = info(MAT_INFO_MALLOCS) 2579 nz_a = info(MAT_INFO_NZ_ALLOCATED) 2580 .ve 2581 2582 Level: intermediate 2583 2584 Concepts: matrices^getting information on 2585 2586 Developer Note: fortran interface is not autogenerated as the f90 2587 interface defintion cannot be generated correctly [due to MatInfo] 2588 2589 @*/ 2590 PetscErrorCode MatGetInfo(Mat mat,MatInfoType flag,MatInfo *info) 2591 { 2592 PetscErrorCode ierr; 2593 2594 PetscFunctionBegin; 2595 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2596 PetscValidType(mat,1); 2597 PetscValidPointer(info,3); 2598 if (!mat->ops->getinfo) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2599 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2600 ierr = (*mat->ops->getinfo)(mat,flag,info);CHKERRQ(ierr); 2601 PetscFunctionReturn(0); 2602 } 2603 2604 /* ----------------------------------------------------------*/ 2605 2606 #undef __FUNCT__ 2607 #define __FUNCT__ "MatLUFactor" 2608 /*@C 2609 MatLUFactor - Performs in-place LU factorization of matrix. 2610 2611 Collective on Mat 2612 2613 Input Parameters: 2614 + mat - the matrix 2615 . row - row permutation 2616 . col - column permutation 2617 - info - options for factorization, includes 2618 $ fill - expected fill as ratio of original fill. 2619 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2620 $ Run with the option -info to determine an optimal value to use 2621 2622 Notes: 2623 Most users should employ the simplified KSP interface for linear solvers 2624 instead of working directly with matrix algebra routines such as this. 2625 See, e.g., KSPCreate(). 2626 2627 This changes the state of the matrix to a factored matrix; it cannot be used 2628 for example with MatSetValues() unless one first calls MatSetUnfactored(). 2629 2630 Level: developer 2631 2632 Concepts: matrices^LU factorization 2633 2634 .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), 2635 MatGetOrdering(), MatSetUnfactored(), MatFactorInfo 2636 2637 Developer Note: fortran interface is not autogenerated as the f90 2638 interface defintion cannot be generated correctly [due to MatFactorInfo] 2639 2640 @*/ 2641 PetscErrorCode MatLUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info) 2642 { 2643 PetscErrorCode ierr; 2644 2645 PetscFunctionBegin; 2646 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2647 if (row) PetscValidHeaderSpecific(row,IS_CLASSID,2); 2648 if (col) PetscValidHeaderSpecific(col,IS_CLASSID,3); 2649 if (info) PetscValidPointer(info,4); 2650 PetscValidType(mat,1); 2651 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2652 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2653 if (!mat->ops->lufactor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2654 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2655 2656 ierr = PetscLogEventBegin(MAT_LUFactor,mat,row,col,0);CHKERRQ(ierr); 2657 ierr = (*mat->ops->lufactor)(mat,row,col,info);CHKERRQ(ierr); 2658 ierr = PetscLogEventEnd(MAT_LUFactor,mat,row,col,0);CHKERRQ(ierr); 2659 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2660 PetscFunctionReturn(0); 2661 } 2662 2663 #undef __FUNCT__ 2664 #define __FUNCT__ "MatILUFactor" 2665 /*@C 2666 MatILUFactor - Performs in-place ILU factorization of matrix. 2667 2668 Collective on Mat 2669 2670 Input Parameters: 2671 + mat - the matrix 2672 . row - row permutation 2673 . col - column permutation 2674 - info - structure containing 2675 $ levels - number of levels of fill. 2676 $ expected fill - as ratio of original fill. 2677 $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices 2678 missing diagonal entries) 2679 2680 Notes: 2681 Probably really in-place only when level of fill is zero, otherwise allocates 2682 new space to store factored matrix and deletes previous memory. 2683 2684 Most users should employ the simplified KSP interface for linear solvers 2685 instead of working directly with matrix algebra routines such as this. 2686 See, e.g., KSPCreate(). 2687 2688 Level: developer 2689 2690 Concepts: matrices^ILU factorization 2691 2692 .seealso: MatILUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 2693 2694 Developer Note: fortran interface is not autogenerated as the f90 2695 interface defintion cannot be generated correctly [due to MatFactorInfo] 2696 2697 @*/ 2698 PetscErrorCode MatILUFactor(Mat mat,IS row,IS col,const MatFactorInfo *info) 2699 { 2700 PetscErrorCode ierr; 2701 2702 PetscFunctionBegin; 2703 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2704 if (row) PetscValidHeaderSpecific(row,IS_CLASSID,2); 2705 if (col) PetscValidHeaderSpecific(col,IS_CLASSID,3); 2706 PetscValidPointer(info,4); 2707 PetscValidType(mat,1); 2708 if (mat->rmap->N != mat->cmap->N) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONG,"matrix must be square"); 2709 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2710 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2711 if (!mat->ops->ilufactor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2712 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2713 2714 ierr = PetscLogEventBegin(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2715 ierr = (*mat->ops->ilufactor)(mat,row,col,info);CHKERRQ(ierr); 2716 ierr = PetscLogEventEnd(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2717 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2718 PetscFunctionReturn(0); 2719 } 2720 2721 #undef __FUNCT__ 2722 #define __FUNCT__ "MatLUFactorSymbolic" 2723 /*@C 2724 MatLUFactorSymbolic - Performs symbolic LU factorization of matrix. 2725 Call this routine before calling MatLUFactorNumeric(). 2726 2727 Collective on Mat 2728 2729 Input Parameters: 2730 + fact - the factor matrix obtained with MatGetFactor() 2731 . mat - the matrix 2732 . row, col - row and column permutations 2733 - info - options for factorization, includes 2734 $ fill - expected fill as ratio of original fill. 2735 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2736 $ Run with the option -info to determine an optimal value to use 2737 2738 2739 Notes: 2740 See the <a href="../../docs/manual.pdf">users manual</a> for additional information about 2741 choosing the fill factor for better efficiency. 2742 2743 Most users should employ the simplified KSP interface for linear solvers 2744 instead of working directly with matrix algebra routines such as this. 2745 See, e.g., KSPCreate(). 2746 2747 Level: developer 2748 2749 Concepts: matrices^LU symbolic factorization 2750 2751 .seealso: MatLUFactor(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 2752 2753 Developer Note: fortran interface is not autogenerated as the f90 2754 interface defintion cannot be generated correctly [due to MatFactorInfo] 2755 2756 @*/ 2757 PetscErrorCode MatLUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info) 2758 { 2759 PetscErrorCode ierr; 2760 2761 PetscFunctionBegin; 2762 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2763 if (row) PetscValidHeaderSpecific(row,IS_CLASSID,2); 2764 if (col) PetscValidHeaderSpecific(col,IS_CLASSID,3); 2765 if (info) PetscValidPointer(info,4); 2766 PetscValidType(mat,1); 2767 PetscValidPointer(fact,5); 2768 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2769 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2770 if (!(fact)->ops->lufactorsymbolic) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Matrix type %s symbolic LU",((PetscObject)mat)->type_name); 2771 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2772 2773 ierr = PetscLogEventBegin(MAT_LUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 2774 ierr = (fact->ops->lufactorsymbolic)(fact,mat,row,col,info);CHKERRQ(ierr); 2775 ierr = PetscLogEventEnd(MAT_LUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 2776 ierr = PetscObjectStateIncrease((PetscObject)fact);CHKERRQ(ierr); 2777 PetscFunctionReturn(0); 2778 } 2779 2780 #undef __FUNCT__ 2781 #define __FUNCT__ "MatLUFactorNumeric" 2782 /*@C 2783 MatLUFactorNumeric - Performs numeric LU factorization of a matrix. 2784 Call this routine after first calling MatLUFactorSymbolic(). 2785 2786 Collective on Mat 2787 2788 Input Parameters: 2789 + fact - the factor matrix obtained with MatGetFactor() 2790 . mat - the matrix 2791 - info - options for factorization 2792 2793 Notes: 2794 See MatLUFactor() for in-place factorization. See 2795 MatCholeskyFactorNumeric() for the symmetric, positive definite case. 2796 2797 Most users should employ the simplified KSP interface for linear solvers 2798 instead of working directly with matrix algebra routines such as this. 2799 See, e.g., KSPCreate(). 2800 2801 Level: developer 2802 2803 Concepts: matrices^LU numeric factorization 2804 2805 .seealso: MatLUFactorSymbolic(), MatLUFactor(), MatCholeskyFactor() 2806 2807 Developer Note: fortran interface is not autogenerated as the f90 2808 interface defintion cannot be generated correctly [due to MatFactorInfo] 2809 2810 @*/ 2811 PetscErrorCode MatLUFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info) 2812 { 2813 PetscErrorCode ierr; 2814 2815 PetscFunctionBegin; 2816 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2817 PetscValidType(mat,1); 2818 PetscValidPointer(fact,2); 2819 PetscValidHeaderSpecific(fact,MAT_CLASSID,2); 2820 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2821 if (mat->rmap->N != (fact)->rmap->N || mat->cmap->N != (fact)->cmap->N) { 2822 SETERRQ4(((PetscObject)mat)->comm,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); 2823 } 2824 if (!(fact)->ops->lufactornumeric) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s numeric LU",((PetscObject)mat)->type_name); 2825 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2826 ierr = PetscLogEventBegin(MAT_LUFactorNumeric,mat,fact,0,0);CHKERRQ(ierr); 2827 ierr = (fact->ops->lufactornumeric)(fact,mat,info);CHKERRQ(ierr); 2828 ierr = PetscLogEventEnd(MAT_LUFactorNumeric,mat,fact,0,0);CHKERRQ(ierr); 2829 2830 ierr = MatView_Private(fact);CHKERRQ(ierr); 2831 ierr = PetscObjectStateIncrease((PetscObject)fact);CHKERRQ(ierr); 2832 PetscFunctionReturn(0); 2833 } 2834 2835 #undef __FUNCT__ 2836 #define __FUNCT__ "MatCholeskyFactor" 2837 /*@C 2838 MatCholeskyFactor - Performs in-place Cholesky factorization of a 2839 symmetric matrix. 2840 2841 Collective on Mat 2842 2843 Input Parameters: 2844 + mat - the matrix 2845 . perm - row and column permutations 2846 - f - expected fill as ratio of original fill 2847 2848 Notes: 2849 See MatLUFactor() for the nonsymmetric case. See also 2850 MatCholeskyFactorSymbolic(), and MatCholeskyFactorNumeric(). 2851 2852 Most users should employ the simplified KSP interface for linear solvers 2853 instead of working directly with matrix algebra routines such as this. 2854 See, e.g., KSPCreate(). 2855 2856 Level: developer 2857 2858 Concepts: matrices^Cholesky factorization 2859 2860 .seealso: MatLUFactor(), MatCholeskyFactorSymbolic(), MatCholeskyFactorNumeric() 2861 MatGetOrdering() 2862 2863 Developer Note: fortran interface is not autogenerated as the f90 2864 interface defintion cannot be generated correctly [due to MatFactorInfo] 2865 2866 @*/ 2867 PetscErrorCode MatCholeskyFactor(Mat mat,IS perm,const MatFactorInfo *info) 2868 { 2869 PetscErrorCode ierr; 2870 2871 PetscFunctionBegin; 2872 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2873 PetscValidType(mat,1); 2874 if (perm) PetscValidHeaderSpecific(perm,IS_CLASSID,2); 2875 if (info) PetscValidPointer(info,3); 2876 if (mat->rmap->N != mat->cmap->N) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONG,"Matrix must be square"); 2877 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2878 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2879 if (!mat->ops->choleskyfactor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 2880 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2881 2882 ierr = PetscLogEventBegin(MAT_CholeskyFactor,mat,perm,0,0);CHKERRQ(ierr); 2883 ierr = (*mat->ops->choleskyfactor)(mat,perm,info);CHKERRQ(ierr); 2884 ierr = PetscLogEventEnd(MAT_CholeskyFactor,mat,perm,0,0);CHKERRQ(ierr); 2885 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2886 PetscFunctionReturn(0); 2887 } 2888 2889 #undef __FUNCT__ 2890 #define __FUNCT__ "MatCholeskyFactorSymbolic" 2891 /*@C 2892 MatCholeskyFactorSymbolic - Performs symbolic Cholesky factorization 2893 of a symmetric matrix. 2894 2895 Collective on Mat 2896 2897 Input Parameters: 2898 + fact - the factor matrix obtained with MatGetFactor() 2899 . mat - the matrix 2900 . perm - row and column permutations 2901 - info - options for factorization, includes 2902 $ fill - expected fill as ratio of original fill. 2903 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2904 $ Run with the option -info to determine an optimal value to use 2905 2906 Notes: 2907 See MatLUFactorSymbolic() for the nonsymmetric case. See also 2908 MatCholeskyFactor() and MatCholeskyFactorNumeric(). 2909 2910 Most users should employ the simplified KSP interface for linear solvers 2911 instead of working directly with matrix algebra routines such as this. 2912 See, e.g., KSPCreate(). 2913 2914 Level: developer 2915 2916 Concepts: matrices^Cholesky symbolic factorization 2917 2918 .seealso: MatLUFactorSymbolic(), MatCholeskyFactor(), MatCholeskyFactorNumeric() 2919 MatGetOrdering() 2920 2921 Developer Note: fortran interface is not autogenerated as the f90 2922 interface defintion cannot be generated correctly [due to MatFactorInfo] 2923 2924 @*/ 2925 PetscErrorCode MatCholeskyFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info) 2926 { 2927 PetscErrorCode ierr; 2928 2929 PetscFunctionBegin; 2930 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2931 PetscValidType(mat,1); 2932 if (perm) PetscValidHeaderSpecific(perm,IS_CLASSID,2); 2933 if (info) PetscValidPointer(info,3); 2934 PetscValidPointer(fact,4); 2935 if (mat->rmap->N != mat->cmap->N) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONG,"Matrix must be square"); 2936 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2937 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2938 if (!(fact)->ops->choleskyfactorsymbolic) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s symbolic factor Cholesky",((PetscObject)mat)->type_name); 2939 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2940 2941 ierr = PetscLogEventBegin(MAT_CholeskyFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 2942 ierr = (fact->ops->choleskyfactorsymbolic)(fact,mat,perm,info);CHKERRQ(ierr); 2943 ierr = PetscLogEventEnd(MAT_CholeskyFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 2944 ierr = PetscObjectStateIncrease((PetscObject)fact);CHKERRQ(ierr); 2945 PetscFunctionReturn(0); 2946 } 2947 2948 #undef __FUNCT__ 2949 #define __FUNCT__ "MatCholeskyFactorNumeric" 2950 /*@C 2951 MatCholeskyFactorNumeric - Performs numeric Cholesky factorization 2952 of a symmetric matrix. Call this routine after first calling 2953 MatCholeskyFactorSymbolic(). 2954 2955 Collective on Mat 2956 2957 Input Parameters: 2958 + fact - the factor matrix obtained with MatGetFactor() 2959 . mat - the initial matrix 2960 . info - options for factorization 2961 - fact - the symbolic factor of mat 2962 2963 2964 Notes: 2965 Most users should employ the simplified KSP interface for linear solvers 2966 instead of working directly with matrix algebra routines such as this. 2967 See, e.g., KSPCreate(). 2968 2969 Level: developer 2970 2971 Concepts: matrices^Cholesky numeric factorization 2972 2973 .seealso: MatCholeskyFactorSymbolic(), MatCholeskyFactor(), MatLUFactorNumeric() 2974 2975 Developer Note: fortran interface is not autogenerated as the f90 2976 interface defintion cannot be generated correctly [due to MatFactorInfo] 2977 2978 @*/ 2979 PetscErrorCode MatCholeskyFactorNumeric(Mat fact,Mat mat,const MatFactorInfo *info) 2980 { 2981 PetscErrorCode ierr; 2982 2983 PetscFunctionBegin; 2984 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 2985 PetscValidType(mat,1); 2986 PetscValidPointer(fact,2); 2987 PetscValidHeaderSpecific(fact,MAT_CLASSID,2); 2988 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2989 if (!(fact)->ops->choleskyfactornumeric) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s numeric factor Cholesky",((PetscObject)mat)->type_name); 2990 if (mat->rmap->N != (fact)->rmap->N || mat->cmap->N != (fact)->cmap->N) { 2991 SETERRQ4(((PetscObject)mat)->comm,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); 2992 } 2993 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2994 2995 ierr = PetscLogEventBegin(MAT_CholeskyFactorNumeric,mat,fact,0,0);CHKERRQ(ierr); 2996 ierr = (fact->ops->choleskyfactornumeric)(fact,mat,info);CHKERRQ(ierr); 2997 ierr = PetscLogEventEnd(MAT_CholeskyFactorNumeric,mat,fact,0,0);CHKERRQ(ierr); 2998 2999 ierr = MatView_Private(fact);CHKERRQ(ierr); 3000 ierr = PetscObjectStateIncrease((PetscObject)fact);CHKERRQ(ierr); 3001 PetscFunctionReturn(0); 3002 } 3003 3004 /* ----------------------------------------------------------------*/ 3005 #undef __FUNCT__ 3006 #define __FUNCT__ "MatSolve" 3007 /*@ 3008 MatSolve - Solves A x = b, given a factored matrix. 3009 3010 Neighbor-wise Collective on Mat and Vec 3011 3012 Input Parameters: 3013 + mat - the factored matrix 3014 - b - the right-hand-side vector 3015 3016 Output Parameter: 3017 . x - the result vector 3018 3019 Notes: 3020 The vectors b and x cannot be the same. I.e., one cannot 3021 call MatSolve(A,x,x). 3022 3023 Notes: 3024 Most users should employ the simplified KSP interface for linear solvers 3025 instead of working directly with matrix algebra routines such as this. 3026 See, e.g., KSPCreate(). 3027 3028 Level: developer 3029 3030 Concepts: matrices^triangular solves 3031 3032 .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd() 3033 @*/ 3034 PetscErrorCode MatSolve(Mat mat,Vec b,Vec x) 3035 { 3036 PetscErrorCode ierr; 3037 3038 PetscFunctionBegin; 3039 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3040 PetscValidType(mat,1); 3041 PetscValidHeaderSpecific(b,VEC_CLASSID,2); 3042 PetscValidHeaderSpecific(x,VEC_CLASSID,3); 3043 PetscCheckSameComm(mat,1,b,2); 3044 PetscCheckSameComm(mat,1,x,3); 3045 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3046 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3047 if (mat->cmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N); 3048 if (mat->rmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N); 3049 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); 3050 if (!mat->rmap->N && !mat->cmap->N) PetscFunctionReturn(0); 3051 if (!mat->ops->solve) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 3052 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3053 3054 ierr = PetscLogEventBegin(MAT_Solve,mat,b,x,0);CHKERRQ(ierr); 3055 ierr = (*mat->ops->solve)(mat,b,x);CHKERRQ(ierr); 3056 ierr = PetscLogEventEnd(MAT_Solve,mat,b,x,0);CHKERRQ(ierr); 3057 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3058 PetscFunctionReturn(0); 3059 } 3060 3061 #undef __FUNCT__ 3062 #define __FUNCT__ "MatMatSolve_Basic" 3063 PetscErrorCode MatMatSolve_Basic(Mat A,Mat B,Mat X) 3064 { 3065 PetscErrorCode ierr; 3066 Vec b,x; 3067 PetscInt m,N,i; 3068 PetscScalar *bb,*xx; 3069 3070 PetscFunctionBegin; 3071 ierr = MatGetArray(B,&bb);CHKERRQ(ierr); 3072 ierr = MatGetArray(X,&xx);CHKERRQ(ierr); 3073 ierr = MatGetLocalSize(B,&m,PETSC_NULL);CHKERRQ(ierr); /* number local rows */ 3074 ierr = MatGetSize(B,PETSC_NULL,&N);CHKERRQ(ierr); /* total columns in dense matrix */ 3075 ierr = MatGetVecs(A,&x,&b);CHKERRQ(ierr); 3076 for (i=0; i<N; i++) { 3077 ierr = VecPlaceArray(b,bb + i*m);CHKERRQ(ierr); 3078 ierr = VecPlaceArray(x,xx + i*m);CHKERRQ(ierr); 3079 ierr = MatSolve(A,b,x);CHKERRQ(ierr); 3080 ierr = VecResetArray(x);CHKERRQ(ierr); 3081 ierr = VecResetArray(b);CHKERRQ(ierr); 3082 } 3083 ierr = VecDestroy(b);CHKERRQ(ierr); 3084 ierr = VecDestroy(x);CHKERRQ(ierr); 3085 ierr = MatRestoreArray(B,&bb);CHKERRQ(ierr); 3086 ierr = MatRestoreArray(X,&xx);CHKERRQ(ierr); 3087 PetscFunctionReturn(0); 3088 } 3089 3090 #undef __FUNCT__ 3091 #define __FUNCT__ "MatMatSolve" 3092 /*@ 3093 MatMatSolve - Solves A X = B, given a factored matrix. 3094 3095 Neighbor-wise Collective on Mat 3096 3097 Input Parameters: 3098 + mat - the factored matrix 3099 - B - the right-hand-side matrix (dense matrix) 3100 3101 Output Parameter: 3102 . X - the result matrix (dense matrix) 3103 3104 Notes: 3105 The matrices b and x cannot be the same. I.e., one cannot 3106 call MatMatSolve(A,x,x). 3107 3108 Notes: 3109 Most users should usually employ the simplified KSP interface for linear solvers 3110 instead of working directly with matrix algebra routines such as this. 3111 See, e.g., KSPCreate(). However KSP can only solve for one vector (column of X) 3112 at a time. 3113 3114 Level: developer 3115 3116 Concepts: matrices^triangular solves 3117 3118 .seealso: MatMatSolveAdd(), MatMatSolveTranspose(), MatMatSolveTransposeAdd(), MatLUFactor(), MatCholeskyFactor() 3119 @*/ 3120 PetscErrorCode MatMatSolve(Mat A,Mat B,Mat X) 3121 { 3122 PetscErrorCode ierr; 3123 3124 PetscFunctionBegin; 3125 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 3126 PetscValidType(A,1); 3127 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 3128 PetscValidHeaderSpecific(X,MAT_CLASSID,3); 3129 PetscCheckSameComm(A,1,B,2); 3130 PetscCheckSameComm(A,1,X,3); 3131 if (X == B) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_IDN,"X and B must be different matrices"); 3132 if (!A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3133 if (A->cmap->N != X->rmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Mat A,Mat X: global dim %D %D",A->cmap->N,X->rmap->N); 3134 if (A->rmap->N != B->rmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim %D %D",A->rmap->N,B->rmap->N); 3135 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); 3136 if (!A->rmap->N && !A->cmap->N) PetscFunctionReturn(0); 3137 ierr = MatPreallocated(A);CHKERRQ(ierr); 3138 3139 ierr = PetscLogEventBegin(MAT_MatSolve,A,B,X,0);CHKERRQ(ierr); 3140 if (!A->ops->matsolve) { 3141 ierr = PetscInfo1(A,"Mat type %s using basic MatMatSolve",((PetscObject)A)->type_name);CHKERRQ(ierr); 3142 ierr = MatMatSolve_Basic(A,B,X);CHKERRQ(ierr); 3143 } else { 3144 ierr = (*A->ops->matsolve)(A,B,X);CHKERRQ(ierr); 3145 } 3146 ierr = PetscLogEventEnd(MAT_MatSolve,A,B,X,0);CHKERRQ(ierr); 3147 ierr = PetscObjectStateIncrease((PetscObject)X);CHKERRQ(ierr); 3148 PetscFunctionReturn(0); 3149 } 3150 3151 3152 #undef __FUNCT__ 3153 #define __FUNCT__ "MatForwardSolve" 3154 /*@ 3155 MatForwardSolve - Solves L x = b, given a factored matrix, A = LU, or 3156 U^T*D^(1/2) x = b, given a factored symmetric matrix, A = U^T*D*U, 3157 3158 Neighbor-wise Collective on Mat and Vec 3159 3160 Input Parameters: 3161 + mat - the factored matrix 3162 - b - the right-hand-side vector 3163 3164 Output Parameter: 3165 . x - the result vector 3166 3167 Notes: 3168 MatSolve() should be used for most applications, as it performs 3169 a forward solve followed by a backward solve. 3170 3171 The vectors b and x cannot be the same, i.e., one cannot 3172 call MatForwardSolve(A,x,x). 3173 3174 For matrix in seqsbaij format with block size larger than 1, 3175 the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet. 3176 MatForwardSolve() solves U^T*D y = b, and 3177 MatBackwardSolve() solves U x = y. 3178 Thus they do not provide a symmetric preconditioner. 3179 3180 Most users should employ the simplified KSP interface for linear solvers 3181 instead of working directly with matrix algebra routines such as this. 3182 See, e.g., KSPCreate(). 3183 3184 Level: developer 3185 3186 Concepts: matrices^forward solves 3187 3188 .seealso: MatSolve(), MatBackwardSolve() 3189 @*/ 3190 PetscErrorCode MatForwardSolve(Mat mat,Vec b,Vec x) 3191 { 3192 PetscErrorCode ierr; 3193 3194 PetscFunctionBegin; 3195 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3196 PetscValidType(mat,1); 3197 PetscValidHeaderSpecific(b,VEC_CLASSID,2); 3198 PetscValidHeaderSpecific(x,VEC_CLASSID,3); 3199 PetscCheckSameComm(mat,1,b,2); 3200 PetscCheckSameComm(mat,1,x,3); 3201 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3202 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3203 if (!mat->ops->forwardsolve) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 3204 if (mat->cmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N); 3205 if (mat->rmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N); 3206 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); 3207 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3208 ierr = PetscLogEventBegin(MAT_ForwardSolve,mat,b,x,0);CHKERRQ(ierr); 3209 ierr = (*mat->ops->forwardsolve)(mat,b,x);CHKERRQ(ierr); 3210 ierr = PetscLogEventEnd(MAT_ForwardSolve,mat,b,x,0);CHKERRQ(ierr); 3211 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3212 PetscFunctionReturn(0); 3213 } 3214 3215 #undef __FUNCT__ 3216 #define __FUNCT__ "MatBackwardSolve" 3217 /*@ 3218 MatBackwardSolve - Solves U x = b, given a factored matrix, A = LU. 3219 D^(1/2) U x = b, given a factored symmetric matrix, A = U^T*D*U, 3220 3221 Neighbor-wise Collective on Mat and Vec 3222 3223 Input Parameters: 3224 + mat - the factored matrix 3225 - b - the right-hand-side vector 3226 3227 Output Parameter: 3228 . x - the result vector 3229 3230 Notes: 3231 MatSolve() should be used for most applications, as it performs 3232 a forward solve followed by a backward solve. 3233 3234 The vectors b and x cannot be the same. I.e., one cannot 3235 call MatBackwardSolve(A,x,x). 3236 3237 For matrix in seqsbaij format with block size larger than 1, 3238 the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet. 3239 MatForwardSolve() solves U^T*D y = b, and 3240 MatBackwardSolve() solves U x = y. 3241 Thus they do not provide a symmetric preconditioner. 3242 3243 Most users should employ the simplified KSP interface for linear solvers 3244 instead of working directly with matrix algebra routines such as this. 3245 See, e.g., KSPCreate(). 3246 3247 Level: developer 3248 3249 Concepts: matrices^backward solves 3250 3251 .seealso: MatSolve(), MatForwardSolve() 3252 @*/ 3253 PetscErrorCode MatBackwardSolve(Mat mat,Vec b,Vec x) 3254 { 3255 PetscErrorCode ierr; 3256 3257 PetscFunctionBegin; 3258 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3259 PetscValidType(mat,1); 3260 PetscValidHeaderSpecific(b,VEC_CLASSID,2); 3261 PetscValidHeaderSpecific(x,VEC_CLASSID,3); 3262 PetscCheckSameComm(mat,1,b,2); 3263 PetscCheckSameComm(mat,1,x,3); 3264 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3265 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3266 if (!mat->ops->backwardsolve) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 3267 if (mat->cmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N); 3268 if (mat->rmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N); 3269 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); 3270 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3271 3272 ierr = PetscLogEventBegin(MAT_BackwardSolve,mat,b,x,0);CHKERRQ(ierr); 3273 ierr = (*mat->ops->backwardsolve)(mat,b,x);CHKERRQ(ierr); 3274 ierr = PetscLogEventEnd(MAT_BackwardSolve,mat,b,x,0);CHKERRQ(ierr); 3275 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3276 PetscFunctionReturn(0); 3277 } 3278 3279 #undef __FUNCT__ 3280 #define __FUNCT__ "MatSolveAdd" 3281 /*@ 3282 MatSolveAdd - Computes x = y + inv(A)*b, given a factored matrix. 3283 3284 Neighbor-wise Collective on Mat and Vec 3285 3286 Input Parameters: 3287 + mat - the factored matrix 3288 . b - the right-hand-side vector 3289 - y - the vector to be added to 3290 3291 Output Parameter: 3292 . x - the result vector 3293 3294 Notes: 3295 The vectors b and x cannot be the same. I.e., one cannot 3296 call MatSolveAdd(A,x,y,x). 3297 3298 Most users should employ the simplified KSP interface for linear solvers 3299 instead of working directly with matrix algebra routines such as this. 3300 See, e.g., KSPCreate(). 3301 3302 Level: developer 3303 3304 Concepts: matrices^triangular solves 3305 3306 .seealso: MatSolve(), MatSolveTranspose(), MatSolveTransposeAdd() 3307 @*/ 3308 PetscErrorCode MatSolveAdd(Mat mat,Vec b,Vec y,Vec x) 3309 { 3310 PetscScalar one = 1.0; 3311 Vec tmp; 3312 PetscErrorCode ierr; 3313 3314 PetscFunctionBegin; 3315 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3316 PetscValidType(mat,1); 3317 PetscValidHeaderSpecific(y,VEC_CLASSID,2); 3318 PetscValidHeaderSpecific(b,VEC_CLASSID,3); 3319 PetscValidHeaderSpecific(x,VEC_CLASSID,4); 3320 PetscCheckSameComm(mat,1,b,2); 3321 PetscCheckSameComm(mat,1,y,2); 3322 PetscCheckSameComm(mat,1,x,3); 3323 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3324 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3325 if (mat->cmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N); 3326 if (mat->rmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N); 3327 if (mat->rmap->N != y->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap->N,y->map->N); 3328 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); 3329 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); 3330 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3331 3332 ierr = PetscLogEventBegin(MAT_SolveAdd,mat,b,x,y);CHKERRQ(ierr); 3333 if (mat->ops->solveadd) { 3334 ierr = (*mat->ops->solveadd)(mat,b,y,x);CHKERRQ(ierr); 3335 } else { 3336 /* do the solve then the add manually */ 3337 if (x != y) { 3338 ierr = MatSolve(mat,b,x);CHKERRQ(ierr); 3339 ierr = VecAXPY(x,one,y);CHKERRQ(ierr); 3340 } else { 3341 ierr = VecDuplicate(x,&tmp);CHKERRQ(ierr); 3342 ierr = PetscLogObjectParent(mat,tmp);CHKERRQ(ierr); 3343 ierr = VecCopy(x,tmp);CHKERRQ(ierr); 3344 ierr = MatSolve(mat,b,x);CHKERRQ(ierr); 3345 ierr = VecAXPY(x,one,tmp);CHKERRQ(ierr); 3346 ierr = VecDestroy(tmp);CHKERRQ(ierr); 3347 } 3348 } 3349 ierr = PetscLogEventEnd(MAT_SolveAdd,mat,b,x,y);CHKERRQ(ierr); 3350 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3351 PetscFunctionReturn(0); 3352 } 3353 3354 #undef __FUNCT__ 3355 #define __FUNCT__ "MatSolveTranspose" 3356 /*@ 3357 MatSolveTranspose - Solves A' x = b, given a factored matrix. 3358 3359 Neighbor-wise Collective on Mat and Vec 3360 3361 Input Parameters: 3362 + mat - the factored matrix 3363 - b - the right-hand-side vector 3364 3365 Output Parameter: 3366 . x - the result vector 3367 3368 Notes: 3369 The vectors b and x cannot be the same. I.e., one cannot 3370 call MatSolveTranspose(A,x,x). 3371 3372 Most users should employ the simplified KSP interface for linear solvers 3373 instead of working directly with matrix algebra routines such as this. 3374 See, e.g., KSPCreate(). 3375 3376 Level: developer 3377 3378 Concepts: matrices^triangular solves 3379 3380 .seealso: MatSolve(), MatSolveAdd(), MatSolveTransposeAdd() 3381 @*/ 3382 PetscErrorCode MatSolveTranspose(Mat mat,Vec b,Vec x) 3383 { 3384 PetscErrorCode ierr; 3385 3386 PetscFunctionBegin; 3387 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3388 PetscValidType(mat,1); 3389 PetscValidHeaderSpecific(b,VEC_CLASSID,2); 3390 PetscValidHeaderSpecific(x,VEC_CLASSID,3); 3391 PetscCheckSameComm(mat,1,b,2); 3392 PetscCheckSameComm(mat,1,x,3); 3393 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3394 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3395 if (!mat->ops->solvetranspose) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Matrix type %s",((PetscObject)mat)->type_name); 3396 if (mat->rmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N); 3397 if (mat->cmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap->N,b->map->N); 3398 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3399 ierr = PetscLogEventBegin(MAT_SolveTranspose,mat,b,x,0);CHKERRQ(ierr); 3400 ierr = (*mat->ops->solvetranspose)(mat,b,x);CHKERRQ(ierr); 3401 ierr = PetscLogEventEnd(MAT_SolveTranspose,mat,b,x,0);CHKERRQ(ierr); 3402 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3403 PetscFunctionReturn(0); 3404 } 3405 3406 #undef __FUNCT__ 3407 #define __FUNCT__ "MatSolveTransposeAdd" 3408 /*@ 3409 MatSolveTransposeAdd - Computes x = y + inv(Transpose(A)) b, given a 3410 factored matrix. 3411 3412 Neighbor-wise Collective on Mat and Vec 3413 3414 Input Parameters: 3415 + mat - the factored matrix 3416 . b - the right-hand-side vector 3417 - y - the vector to be added to 3418 3419 Output Parameter: 3420 . x - the result vector 3421 3422 Notes: 3423 The vectors b and x cannot be the same. I.e., one cannot 3424 call MatSolveTransposeAdd(A,x,y,x). 3425 3426 Most users should employ the simplified KSP interface for linear solvers 3427 instead of working directly with matrix algebra routines such as this. 3428 See, e.g., KSPCreate(). 3429 3430 Level: developer 3431 3432 Concepts: matrices^triangular solves 3433 3434 .seealso: MatSolve(), MatSolveAdd(), MatSolveTranspose() 3435 @*/ 3436 PetscErrorCode MatSolveTransposeAdd(Mat mat,Vec b,Vec y,Vec x) 3437 { 3438 PetscScalar one = 1.0; 3439 PetscErrorCode ierr; 3440 Vec tmp; 3441 3442 PetscFunctionBegin; 3443 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3444 PetscValidType(mat,1); 3445 PetscValidHeaderSpecific(y,VEC_CLASSID,2); 3446 PetscValidHeaderSpecific(b,VEC_CLASSID,3); 3447 PetscValidHeaderSpecific(x,VEC_CLASSID,4); 3448 PetscCheckSameComm(mat,1,b,2); 3449 PetscCheckSameComm(mat,1,y,3); 3450 PetscCheckSameComm(mat,1,x,4); 3451 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 3452 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 3453 if (mat->rmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap->N,x->map->N); 3454 if (mat->cmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap->N,b->map->N); 3455 if (mat->cmap->N != y->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap->N,y->map->N); 3456 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); 3457 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3458 3459 ierr = PetscLogEventBegin(MAT_SolveTransposeAdd,mat,b,x,y);CHKERRQ(ierr); 3460 if (mat->ops->solvetransposeadd) { 3461 ierr = (*mat->ops->solvetransposeadd)(mat,b,y,x);CHKERRQ(ierr); 3462 } else { 3463 /* do the solve then the add manually */ 3464 if (x != y) { 3465 ierr = MatSolveTranspose(mat,b,x);CHKERRQ(ierr); 3466 ierr = VecAXPY(x,one,y);CHKERRQ(ierr); 3467 } else { 3468 ierr = VecDuplicate(x,&tmp);CHKERRQ(ierr); 3469 ierr = PetscLogObjectParent(mat,tmp);CHKERRQ(ierr); 3470 ierr = VecCopy(x,tmp);CHKERRQ(ierr); 3471 ierr = MatSolveTranspose(mat,b,x);CHKERRQ(ierr); 3472 ierr = VecAXPY(x,one,tmp);CHKERRQ(ierr); 3473 ierr = VecDestroy(tmp);CHKERRQ(ierr); 3474 } 3475 } 3476 ierr = PetscLogEventEnd(MAT_SolveTransposeAdd,mat,b,x,y);CHKERRQ(ierr); 3477 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3478 PetscFunctionReturn(0); 3479 } 3480 /* ----------------------------------------------------------------*/ 3481 3482 #undef __FUNCT__ 3483 #define __FUNCT__ "MatSOR" 3484 /*@ 3485 MatSOR - Computes relaxation (SOR, Gauss-Seidel) sweeps. 3486 3487 Neighbor-wise Collective on Mat and Vec 3488 3489 Input Parameters: 3490 + mat - the matrix 3491 . b - the right hand side 3492 . omega - the relaxation factor 3493 . flag - flag indicating the type of SOR (see below) 3494 . shift - diagonal shift 3495 . its - the number of iterations 3496 - lits - the number of local iterations 3497 3498 Output Parameters: 3499 . x - the solution (can contain an initial guess, use option SOR_ZERO_INITIAL_GUESS to indicate no guess) 3500 3501 SOR Flags: 3502 . SOR_FORWARD_SWEEP - forward SOR 3503 . SOR_BACKWARD_SWEEP - backward SOR 3504 . SOR_SYMMETRIC_SWEEP - SSOR (symmetric SOR) 3505 . SOR_LOCAL_FORWARD_SWEEP - local forward SOR 3506 . SOR_LOCAL_BACKWARD_SWEEP - local forward SOR 3507 . SOR_LOCAL_SYMMETRIC_SWEEP - local SSOR 3508 . SOR_APPLY_UPPER, SOR_APPLY_LOWER - applies 3509 upper/lower triangular part of matrix to 3510 vector (with omega) 3511 . SOR_ZERO_INITIAL_GUESS - zero initial guess 3512 3513 Notes: 3514 SOR_LOCAL_FORWARD_SWEEP, SOR_LOCAL_BACKWARD_SWEEP, and 3515 SOR_LOCAL_SYMMETRIC_SWEEP perform separate independent smoothings 3516 on each processor. 3517 3518 Application programmers will not generally use MatSOR() directly, 3519 but instead will employ the KSP/PC interface. 3520 3521 Notes: for BAIJ, SBAIJ, and AIJ matrices with Inodes this does a block SOR smoothing, otherwise it does a pointwise smoothing 3522 3523 Notes for Advanced Users: 3524 The flags are implemented as bitwise inclusive or operations. 3525 For example, use (SOR_ZERO_INITIAL_GUESS | SOR_SYMMETRIC_SWEEP) 3526 to specify a zero initial guess for SSOR. 3527 3528 Most users should employ the simplified KSP interface for linear solvers 3529 instead of working directly with matrix algebra routines such as this. 3530 See, e.g., KSPCreate(). 3531 3532 Vectors x and b CANNOT be the same 3533 3534 Level: developer 3535 3536 Concepts: matrices^relaxation 3537 Concepts: matrices^SOR 3538 Concepts: matrices^Gauss-Seidel 3539 3540 @*/ 3541 PetscErrorCode MatSOR(Mat mat,Vec b,PetscReal omega,MatSORType flag,PetscReal shift,PetscInt its,PetscInt lits,Vec x) 3542 { 3543 PetscErrorCode ierr; 3544 3545 PetscFunctionBegin; 3546 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3547 PetscValidType(mat,1); 3548 PetscValidHeaderSpecific(b,VEC_CLASSID,2); 3549 PetscValidHeaderSpecific(x,VEC_CLASSID,8); 3550 PetscCheckSameComm(mat,1,b,2); 3551 PetscCheckSameComm(mat,1,x,8); 3552 if (!mat->ops->sor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 3553 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3554 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3555 if (mat->cmap->N != x->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap->N,x->map->N); 3556 if (mat->rmap->N != b->map->N) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap->N,b->map->N); 3557 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); 3558 if (its <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D positive",its); 3559 if (lits <= 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires local its %D positive",lits); 3560 if (b == x) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_IDN,"b and x vector cannot be the same"); 3561 3562 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3563 ierr = PetscLogEventBegin(MAT_SOR,mat,b,x,0);CHKERRQ(ierr); 3564 ierr =(*mat->ops->sor)(mat,b,omega,flag,shift,its,lits,x);CHKERRQ(ierr); 3565 ierr = PetscLogEventEnd(MAT_SOR,mat,b,x,0);CHKERRQ(ierr); 3566 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 3567 PetscFunctionReturn(0); 3568 } 3569 3570 #undef __FUNCT__ 3571 #define __FUNCT__ "MatCopy_Basic" 3572 /* 3573 Default matrix copy routine. 3574 */ 3575 PetscErrorCode MatCopy_Basic(Mat A,Mat B,MatStructure str) 3576 { 3577 PetscErrorCode ierr; 3578 PetscInt i,rstart = 0,rend = 0,nz; 3579 const PetscInt *cwork; 3580 const PetscScalar *vwork; 3581 3582 PetscFunctionBegin; 3583 if (B->assembled) { 3584 ierr = MatZeroEntries(B);CHKERRQ(ierr); 3585 } 3586 ierr = MatGetOwnershipRange(A,&rstart,&rend);CHKERRQ(ierr); 3587 for (i=rstart; i<rend; i++) { 3588 ierr = MatGetRow(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 3589 ierr = MatSetValues(B,1,&i,nz,cwork,vwork,INSERT_VALUES);CHKERRQ(ierr); 3590 ierr = MatRestoreRow(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 3591 } 3592 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3593 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3594 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 3595 PetscFunctionReturn(0); 3596 } 3597 3598 #undef __FUNCT__ 3599 #define __FUNCT__ "MatCopy" 3600 /*@ 3601 MatCopy - Copys a matrix to another matrix. 3602 3603 Collective on Mat 3604 3605 Input Parameters: 3606 + A - the matrix 3607 - str - SAME_NONZERO_PATTERN or DIFFERENT_NONZERO_PATTERN 3608 3609 Output Parameter: 3610 . B - where the copy is put 3611 3612 Notes: 3613 If you use SAME_NONZERO_PATTERN then the two matrices had better have the 3614 same nonzero pattern or the routine will crash. 3615 3616 MatCopy() copies the matrix entries of a matrix to another existing 3617 matrix (after first zeroing the second matrix). A related routine is 3618 MatConvert(), which first creates a new matrix and then copies the data. 3619 3620 Level: intermediate 3621 3622 Concepts: matrices^copying 3623 3624 .seealso: MatConvert(), MatDuplicate() 3625 3626 @*/ 3627 PetscErrorCode MatCopy(Mat A,Mat B,MatStructure str) 3628 { 3629 PetscErrorCode ierr; 3630 PetscInt i; 3631 3632 PetscFunctionBegin; 3633 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 3634 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 3635 PetscValidType(A,1); 3636 PetscValidType(B,2); 3637 PetscCheckSameComm(A,1,B,2); 3638 ierr = MatPreallocated(B);CHKERRQ(ierr); 3639 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3640 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3641 if (A->rmap->N != B->rmap->N || A->cmap->N != B->cmap->N) SETERRQ4(((PetscObject)A)->comm,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); 3642 ierr = MatPreallocated(A);CHKERRQ(ierr); 3643 3644 ierr = PetscLogEventBegin(MAT_Copy,A,B,0,0);CHKERRQ(ierr); 3645 if (A->ops->copy) { 3646 ierr = (*A->ops->copy)(A,B,str);CHKERRQ(ierr); 3647 } else { /* generic conversion */ 3648 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 3649 } 3650 if (A->rmapping) { 3651 if (B->rmapping) { 3652 ierr = ISLocalToGlobalMappingDestroy(B->rmapping);CHKERRQ(ierr);B->rmapping = 0; 3653 ierr = ISLocalToGlobalMappingDestroy(B->cmapping);CHKERRQ(ierr);B->cmapping = 0; 3654 } 3655 ierr = MatSetLocalToGlobalMapping(B,A->rmapping,A->cmapping);CHKERRQ(ierr); 3656 } 3657 if (A->rbmapping) { 3658 if (B->rbmapping) { 3659 ierr = ISLocalToGlobalMappingDestroy(B->rbmapping);CHKERRQ(ierr);B->rbmapping = 0; 3660 ierr = ISLocalToGlobalMappingDestroy(B->cbmapping);CHKERRQ(ierr);B->cbmapping = 0; 3661 } 3662 ierr = MatSetLocalToGlobalMappingBlock(B,A->rbmapping,A->cbmapping);CHKERRQ(ierr); 3663 } 3664 3665 B->stencil.dim = A->stencil.dim; 3666 B->stencil.noc = A->stencil.noc; 3667 for (i=0; i<=A->stencil.dim; i++) { 3668 B->stencil.dims[i] = A->stencil.dims[i]; 3669 B->stencil.starts[i] = A->stencil.starts[i]; 3670 } 3671 3672 ierr = PetscLogEventEnd(MAT_Copy,A,B,0,0);CHKERRQ(ierr); 3673 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 3674 PetscFunctionReturn(0); 3675 } 3676 3677 #undef __FUNCT__ 3678 #define __FUNCT__ "MatConvert" 3679 /*@C 3680 MatConvert - Converts a matrix to another matrix, either of the same 3681 or different type. 3682 3683 Collective on Mat 3684 3685 Input Parameters: 3686 + mat - the matrix 3687 . newtype - new matrix type. Use MATSAME to create a new matrix of the 3688 same type as the original matrix. 3689 - reuse - denotes if the destination matrix is to be created or reused. Currently 3690 MAT_REUSE_MATRIX is only supported for inplace conversion, otherwise use 3691 MAT_INITIAL_MATRIX. 3692 3693 Output Parameter: 3694 . M - pointer to place new matrix 3695 3696 Notes: 3697 MatConvert() first creates a new matrix and then copies the data from 3698 the first matrix. A related routine is MatCopy(), which copies the matrix 3699 entries of one matrix to another already existing matrix context. 3700 3701 Cannot be used to convert a sequential matrix to parallel or parallel to sequential, 3702 the MPI communicator of the generated matrix is always the same as the communicator 3703 of the input matrix. 3704 3705 Level: intermediate 3706 3707 Concepts: matrices^converting between storage formats 3708 3709 .seealso: MatCopy(), MatDuplicate() 3710 @*/ 3711 PetscErrorCode MatConvert(Mat mat, const MatType newtype,MatReuse reuse,Mat *M) 3712 { 3713 PetscErrorCode ierr; 3714 PetscBool sametype,issame,flg; 3715 char convname[256],mtype[256]; 3716 Mat B; 3717 3718 PetscFunctionBegin; 3719 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3720 PetscValidType(mat,1); 3721 PetscValidPointer(M,3); 3722 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3723 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3724 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3725 3726 ierr = PetscOptionsGetString(((PetscObject)mat)->prefix,"-matconvert_type",mtype,256,&flg);CHKERRQ(ierr); 3727 if (flg) { 3728 newtype = mtype; 3729 } 3730 ierr = PetscTypeCompare((PetscObject)mat,newtype,&sametype);CHKERRQ(ierr); 3731 ierr = PetscStrcmp(newtype,"same",&issame);CHKERRQ(ierr); 3732 if ((reuse == MAT_REUSE_MATRIX) && (mat != *M)) { 3733 SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"MAT_REUSE_MATRIX only supported for in-place conversion currently"); 3734 } 3735 3736 if ((reuse == MAT_REUSE_MATRIX) && (issame || sametype)) PetscFunctionReturn(0); 3737 3738 if ((sametype || issame) && (reuse==MAT_INITIAL_MATRIX) && mat->ops->duplicate) { 3739 ierr = (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);CHKERRQ(ierr); 3740 } else { 3741 PetscErrorCode (*conv)(Mat, const MatType,MatReuse,Mat*)=PETSC_NULL; 3742 const char *prefix[3] = {"seq","mpi",""}; 3743 PetscInt i; 3744 /* 3745 Order of precedence: 3746 1) See if a specialized converter is known to the current matrix. 3747 2) See if a specialized converter is known to the desired matrix class. 3748 3) See if a good general converter is registered for the desired class 3749 (as of 6/27/03 only MATMPIADJ falls into this category). 3750 4) See if a good general converter is known for the current matrix. 3751 5) Use a really basic converter. 3752 */ 3753 3754 /* 1) See if a specialized converter is known to the current matrix and the desired class */ 3755 for (i=0; i<3; i++) { 3756 ierr = PetscStrcpy(convname,"MatConvert_");CHKERRQ(ierr); 3757 ierr = PetscStrcat(convname,((PetscObject)mat)->type_name);CHKERRQ(ierr); 3758 ierr = PetscStrcat(convname,"_");CHKERRQ(ierr); 3759 ierr = PetscStrcat(convname,prefix[i]);CHKERRQ(ierr); 3760 ierr = PetscStrcat(convname,newtype);CHKERRQ(ierr); 3761 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3762 ierr = PetscObjectQueryFunction((PetscObject)mat,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3763 if (conv) goto foundconv; 3764 } 3765 3766 /* 2) See if a specialized converter is known to the desired matrix class. */ 3767 ierr = MatCreate(((PetscObject)mat)->comm,&B);CHKERRQ(ierr); 3768 ierr = MatSetSizes(B,mat->rmap->n,mat->cmap->n,mat->rmap->N,mat->cmap->N);CHKERRQ(ierr); 3769 ierr = MatSetType(B,newtype);CHKERRQ(ierr); 3770 for (i=0; i<3; i++) { 3771 ierr = PetscStrcpy(convname,"MatConvert_");CHKERRQ(ierr); 3772 ierr = PetscStrcat(convname,((PetscObject)mat)->type_name);CHKERRQ(ierr); 3773 ierr = PetscStrcat(convname,"_");CHKERRQ(ierr); 3774 ierr = PetscStrcat(convname,prefix[i]);CHKERRQ(ierr); 3775 ierr = PetscStrcat(convname,newtype);CHKERRQ(ierr); 3776 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3777 ierr = PetscObjectQueryFunction((PetscObject)B,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3778 if (conv) { 3779 ierr = MatDestroy(B);CHKERRQ(ierr); 3780 goto foundconv; 3781 } 3782 } 3783 3784 /* 3) See if a good general converter is registered for the desired class */ 3785 conv = B->ops->convertfrom; 3786 ierr = MatDestroy(B);CHKERRQ(ierr); 3787 if (conv) goto foundconv; 3788 3789 /* 4) See if a good general converter is known for the current matrix */ 3790 if (mat->ops->convert) { 3791 conv = mat->ops->convert; 3792 } 3793 if (conv) goto foundconv; 3794 3795 /* 5) Use a really basic converter. */ 3796 conv = MatConvert_Basic; 3797 3798 foundconv: 3799 ierr = PetscLogEventBegin(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3800 ierr = (*conv)(mat,newtype,reuse,M);CHKERRQ(ierr); 3801 ierr = PetscLogEventEnd(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3802 } 3803 ierr = PetscObjectStateIncrease((PetscObject)*M);CHKERRQ(ierr); 3804 PetscFunctionReturn(0); 3805 } 3806 3807 #undef __FUNCT__ 3808 #define __FUNCT__ "MatFactorGetSolverPackage" 3809 /*@C 3810 MatFactorGetSolverPackage - Returns name of the package providing the factorization routines 3811 3812 Not Collective 3813 3814 Input Parameter: 3815 . mat - the matrix, must be a factored matrix 3816 3817 Output Parameter: 3818 . type - the string name of the package (do not free this string) 3819 3820 Notes: 3821 In Fortran you pass in a empty string and the package name will be copied into it. 3822 (Make sure the string is long enough) 3823 3824 Level: intermediate 3825 3826 .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable(), MatGetFactor() 3827 @*/ 3828 PetscErrorCode MatFactorGetSolverPackage(Mat mat, const MatSolverPackage *type) 3829 { 3830 PetscErrorCode ierr; 3831 PetscErrorCode (*conv)(Mat,const MatSolverPackage*); 3832 3833 PetscFunctionBegin; 3834 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3835 PetscValidType(mat,1); 3836 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Only for factored matrix"); 3837 ierr = PetscObjectQueryFunction((PetscObject)mat,"MatFactorGetSolverPackage_C",(void (**)(void))&conv);CHKERRQ(ierr); 3838 if (!conv) { 3839 *type = MATSOLVERPETSC; 3840 } else { 3841 ierr = (*conv)(mat,type);CHKERRQ(ierr); 3842 } 3843 PetscFunctionReturn(0); 3844 } 3845 3846 #undef __FUNCT__ 3847 #define __FUNCT__ "MatGetFactor" 3848 /*@C 3849 MatGetFactor - Returns a matrix suitable to calls to MatXXFactorSymbolic() 3850 3851 Collective on Mat 3852 3853 Input Parameters: 3854 + mat - the matrix 3855 . type - name of solver type, for example, spooles, superlu, plapack, petsc (to use PETSc's default) 3856 - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU, 3857 3858 Output Parameters: 3859 . f - the factor matrix used with MatXXFactorSymbolic() calls 3860 3861 Notes: 3862 Some PETSc matrix formats have alternative solvers available that are contained in alternative packages 3863 such as pastix, superlu, mumps, spooles etc. 3864 3865 PETSc must have been ./configure to use the external solver, using the option --download-package 3866 3867 Level: intermediate 3868 3869 .seealso: MatCopy(), MatDuplicate(), MatGetFactorAvailable() 3870 @*/ 3871 PetscErrorCode MatGetFactor(Mat mat, const MatSolverPackage type,MatFactorType ftype,Mat *f) 3872 { 3873 PetscErrorCode ierr,(*conv)(Mat,MatFactorType,Mat*); 3874 char convname[256]; 3875 3876 PetscFunctionBegin; 3877 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3878 PetscValidType(mat,1); 3879 3880 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3881 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3882 3883 ierr = PetscStrcpy(convname,"MatGetFactor_");CHKERRQ(ierr); 3884 ierr = PetscStrcat(convname,type);CHKERRQ(ierr); 3885 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3886 ierr = PetscObjectQueryFunction((PetscObject)mat,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3887 if (!conv) { 3888 PetscBool flag; 3889 MPI_Comm comm; 3890 3891 ierr = PetscObjectGetComm((PetscObject)mat,&comm);CHKERRQ(ierr); 3892 ierr = PetscStrcasecmp(MATSOLVERPETSC,type,&flag);CHKERRQ(ierr); 3893 if (flag) { 3894 SETERRQ2(comm,PETSC_ERR_SUP,"Matrix format %s does not have a built-in PETSc %s",((PetscObject)mat)->type_name,MatFactorTypes[ftype]); 3895 } else { 3896 SETERRQ4(comm,PETSC_ERR_SUP,"Matrix format %s does not have a solver package %s for %s. Perhaps you must ./configure with --download-%s",((PetscObject)mat)->type_name,type,MatFactorTypes[ftype],type); 3897 } 3898 } 3899 ierr = (*conv)(mat,ftype,f);CHKERRQ(ierr); 3900 PetscFunctionReturn(0); 3901 } 3902 3903 #undef __FUNCT__ 3904 #define __FUNCT__ "MatGetFactorAvailable" 3905 /*@C 3906 MatGetFactorAvailable - Returns a a flag if matrix supports particular package and factor type 3907 3908 Not Collective 3909 3910 Input Parameters: 3911 + mat - the matrix 3912 . type - name of solver type, for example, spooles, superlu, plapack, petsc (to use PETSc's default) 3913 - ftype - factor type, MAT_FACTOR_LU, MAT_FACTOR_CHOLESKY, MAT_FACTOR_ICC, MAT_FACTOR_ILU, 3914 3915 Output Parameter: 3916 . flg - PETSC_TRUE if the factorization is available 3917 3918 Notes: 3919 Some PETSc matrix formats have alternative solvers available that are contained in alternative packages 3920 such as pastix, superlu, mumps, spooles etc. 3921 3922 PETSc must have been ./configure to use the external solver, using the option --download-package 3923 3924 Level: intermediate 3925 3926 .seealso: MatCopy(), MatDuplicate(), MatGetFactor() 3927 @*/ 3928 PetscErrorCode MatGetFactorAvailable(Mat mat, const MatSolverPackage type,MatFactorType ftype,PetscBool *flg) 3929 { 3930 PetscErrorCode ierr; 3931 char convname[256]; 3932 PetscErrorCode (*conv)(Mat,MatFactorType,PetscBool *); 3933 3934 PetscFunctionBegin; 3935 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3936 PetscValidType(mat,1); 3937 3938 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3939 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3940 3941 ierr = PetscStrcpy(convname,"MatGetFactorAvailable_");CHKERRQ(ierr); 3942 ierr = PetscStrcat(convname,type);CHKERRQ(ierr); 3943 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3944 ierr = PetscObjectQueryFunction((PetscObject)mat,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3945 if (!conv) { 3946 *flg = PETSC_FALSE; 3947 } else { 3948 ierr = (*conv)(mat,ftype,flg);CHKERRQ(ierr); 3949 } 3950 PetscFunctionReturn(0); 3951 } 3952 3953 3954 #undef __FUNCT__ 3955 #define __FUNCT__ "MatDuplicate" 3956 /*@ 3957 MatDuplicate - Duplicates a matrix including the non-zero structure. 3958 3959 Collective on Mat 3960 3961 Input Parameters: 3962 + mat - the matrix 3963 - op - either MAT_DO_NOT_COPY_VALUES or MAT_COPY_VALUES, cause it to copy the numerical values in the matrix 3964 MAT_SHARE_NONZERO_PATTERN to share the nonzero patterns with the previous matrix and not copy them. 3965 3966 Output Parameter: 3967 . M - pointer to place new matrix 3968 3969 Level: intermediate 3970 3971 Concepts: matrices^duplicating 3972 3973 Notes: You cannot change the nonzero pattern for the parent or child matrix if you use MAT_SHARE_NONZERO_PATTERN. 3974 3975 .seealso: MatCopy(), MatConvert() 3976 @*/ 3977 PetscErrorCode MatDuplicate(Mat mat,MatDuplicateOption op,Mat *M) 3978 { 3979 PetscErrorCode ierr; 3980 Mat B; 3981 PetscInt i; 3982 3983 PetscFunctionBegin; 3984 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3985 PetscValidType(mat,1); 3986 PetscValidPointer(M,3); 3987 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3988 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3989 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3990 3991 *M = 0; 3992 if (!mat->ops->duplicate) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Not written for this matrix type"); 3993 ierr = PetscLogEventBegin(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3994 ierr = (*mat->ops->duplicate)(mat,op,M);CHKERRQ(ierr); 3995 B = *M; 3996 if (mat->rmapping) { 3997 ierr = MatSetLocalToGlobalMapping(B,mat->rmapping,mat->cmapping);CHKERRQ(ierr); 3998 } 3999 if (mat->rbmapping) { 4000 ierr = MatSetLocalToGlobalMappingBlock(B,mat->rbmapping,mat->cbmapping);CHKERRQ(ierr); 4001 } 4002 ierr = PetscLayoutCopy(mat->rmap,&B->rmap);CHKERRQ(ierr); 4003 ierr = PetscLayoutCopy(mat->cmap,&B->cmap);CHKERRQ(ierr); 4004 4005 B->stencil.dim = mat->stencil.dim; 4006 B->stencil.noc = mat->stencil.noc; 4007 for (i=0; i<=mat->stencil.dim; i++) { 4008 B->stencil.dims[i] = mat->stencil.dims[i]; 4009 B->stencil.starts[i] = mat->stencil.starts[i]; 4010 } 4011 4012 B->nooffproczerorows = mat->nooffproczerorows; 4013 B->nooffprocentries = mat->nooffprocentries; 4014 ierr = PetscLogEventEnd(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 4015 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 4016 PetscFunctionReturn(0); 4017 } 4018 4019 #undef __FUNCT__ 4020 #define __FUNCT__ "MatGetDiagonal" 4021 /*@ 4022 MatGetDiagonal - Gets the diagonal of a matrix. 4023 4024 Logically Collective on Mat and Vec 4025 4026 Input Parameters: 4027 + mat - the matrix 4028 - v - the vector for storing the diagonal 4029 4030 Output Parameter: 4031 . v - the diagonal of the matrix 4032 4033 Level: intermediate 4034 4035 Concepts: matrices^accessing diagonals 4036 4037 .seealso: MatGetRow(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs() 4038 @*/ 4039 PetscErrorCode MatGetDiagonal(Mat mat,Vec v) 4040 { 4041 PetscErrorCode ierr; 4042 4043 PetscFunctionBegin; 4044 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4045 PetscValidType(mat,1); 4046 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4047 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4048 if (!mat->ops->getdiagonal) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4049 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4050 4051 ierr = (*mat->ops->getdiagonal)(mat,v);CHKERRQ(ierr); 4052 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 4053 PetscFunctionReturn(0); 4054 } 4055 4056 #undef __FUNCT__ 4057 #define __FUNCT__ "MatGetRowMin" 4058 /*@ 4059 MatGetRowMin - Gets the minimum value (of the real part) of each 4060 row of the matrix 4061 4062 Logically Collective on Mat and Vec 4063 4064 Input Parameters: 4065 . mat - the matrix 4066 4067 Output Parameter: 4068 + v - the vector for storing the maximums 4069 - idx - the indices of the column found for each row (optional) 4070 4071 Level: intermediate 4072 4073 Notes: The result of this call are the same as if one converted the matrix to dense format 4074 and found the minimum value in each row (i.e. the implicit zeros are counted as zeros). 4075 4076 This code is only implemented for a couple of matrix formats. 4077 4078 Concepts: matrices^getting row maximums 4079 4080 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs(), 4081 MatGetRowMax() 4082 @*/ 4083 PetscErrorCode MatGetRowMin(Mat mat,Vec v,PetscInt idx[]) 4084 { 4085 PetscErrorCode ierr; 4086 4087 PetscFunctionBegin; 4088 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4089 PetscValidType(mat,1); 4090 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4091 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4092 if (!mat->ops->getrowmax) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4093 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4094 4095 ierr = (*mat->ops->getrowmin)(mat,v,idx);CHKERRQ(ierr); 4096 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 4097 PetscFunctionReturn(0); 4098 } 4099 4100 #undef __FUNCT__ 4101 #define __FUNCT__ "MatGetRowMinAbs" 4102 /*@ 4103 MatGetRowMinAbs - Gets the minimum value (in absolute value) of each 4104 row of the matrix 4105 4106 Logically Collective on Mat and Vec 4107 4108 Input Parameters: 4109 . mat - the matrix 4110 4111 Output Parameter: 4112 + v - the vector for storing the minimums 4113 - idx - the indices of the column found for each row (optional) 4114 4115 Level: intermediate 4116 4117 Notes: if a row is completely empty or has only 0.0 values then the idx[] value for that 4118 row is 0 (the first column). 4119 4120 This code is only implemented for a couple of matrix formats. 4121 4122 Concepts: matrices^getting row maximums 4123 4124 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMax(), MatGetRowMaxAbs(), MatGetRowMin() 4125 @*/ 4126 PetscErrorCode MatGetRowMinAbs(Mat mat,Vec v,PetscInt idx[]) 4127 { 4128 PetscErrorCode ierr; 4129 4130 PetscFunctionBegin; 4131 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4132 PetscValidType(mat,1); 4133 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4134 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4135 if (!mat->ops->getrowminabs) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4136 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4137 if (idx) {ierr = PetscMemzero(idx,mat->rmap->n*sizeof(PetscInt));CHKERRQ(ierr);} 4138 4139 ierr = (*mat->ops->getrowminabs)(mat,v,idx);CHKERRQ(ierr); 4140 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 4141 PetscFunctionReturn(0); 4142 } 4143 4144 #undef __FUNCT__ 4145 #define __FUNCT__ "MatGetRowMax" 4146 /*@ 4147 MatGetRowMax - Gets the maximum value (of the real part) of each 4148 row of the matrix 4149 4150 Logically Collective on Mat and Vec 4151 4152 Input Parameters: 4153 . mat - the matrix 4154 4155 Output Parameter: 4156 + v - the vector for storing the maximums 4157 - idx - the indices of the column found for each row (optional) 4158 4159 Level: intermediate 4160 4161 Notes: The result of this call are the same as if one converted the matrix to dense format 4162 and found the minimum value in each row (i.e. the implicit zeros are counted as zeros). 4163 4164 This code is only implemented for a couple of matrix formats. 4165 4166 Concepts: matrices^getting row maximums 4167 4168 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs(), MatGetRowMin() 4169 @*/ 4170 PetscErrorCode MatGetRowMax(Mat mat,Vec v,PetscInt idx[]) 4171 { 4172 PetscErrorCode ierr; 4173 4174 PetscFunctionBegin; 4175 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4176 PetscValidType(mat,1); 4177 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4178 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4179 if (!mat->ops->getrowmax) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4180 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4181 4182 ierr = (*mat->ops->getrowmax)(mat,v,idx);CHKERRQ(ierr); 4183 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 4184 PetscFunctionReturn(0); 4185 } 4186 4187 #undef __FUNCT__ 4188 #define __FUNCT__ "MatGetRowMaxAbs" 4189 /*@ 4190 MatGetRowMaxAbs - Gets the maximum value (in absolute value) of each 4191 row of the matrix 4192 4193 Logically Collective on Mat and Vec 4194 4195 Input Parameters: 4196 . mat - the matrix 4197 4198 Output Parameter: 4199 + v - the vector for storing the maximums 4200 - idx - the indices of the column found for each row (optional) 4201 4202 Level: intermediate 4203 4204 Notes: if a row is completely empty or has only 0.0 values then the idx[] value for that 4205 row is 0 (the first column). 4206 4207 This code is only implemented for a couple of matrix formats. 4208 4209 Concepts: matrices^getting row maximums 4210 4211 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMax(), MatGetRowMin() 4212 @*/ 4213 PetscErrorCode MatGetRowMaxAbs(Mat mat,Vec v,PetscInt idx[]) 4214 { 4215 PetscErrorCode ierr; 4216 4217 PetscFunctionBegin; 4218 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4219 PetscValidType(mat,1); 4220 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4221 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4222 if (!mat->ops->getrowmaxabs) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4223 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4224 if (idx) {ierr = PetscMemzero(idx,mat->rmap->n*sizeof(PetscInt));CHKERRQ(ierr);} 4225 4226 ierr = (*mat->ops->getrowmaxabs)(mat,v,idx);CHKERRQ(ierr); 4227 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 4228 PetscFunctionReturn(0); 4229 } 4230 4231 #undef __FUNCT__ 4232 #define __FUNCT__ "MatGetRowSum" 4233 /*@ 4234 MatGetRowSum - Gets the sum of each row of the matrix 4235 4236 Logically Collective on Mat and Vec 4237 4238 Input Parameters: 4239 . mat - the matrix 4240 4241 Output Parameter: 4242 . v - the vector for storing the sum of rows 4243 4244 Level: intermediate 4245 4246 Notes: This code is slow since it is not currently specialized for different formats 4247 4248 Concepts: matrices^getting row sums 4249 4250 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMax(), MatGetRowMin() 4251 @*/ 4252 PetscErrorCode MatGetRowSum(Mat mat, Vec v) 4253 { 4254 PetscInt start = 0, end = 0, row; 4255 PetscScalar *array; 4256 PetscErrorCode ierr; 4257 4258 PetscFunctionBegin; 4259 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4260 PetscValidType(mat,1); 4261 PetscValidHeaderSpecific(v,VEC_CLASSID,2); 4262 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4263 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4264 ierr = MatGetOwnershipRange(mat, &start, &end);CHKERRQ(ierr); 4265 ierr = VecGetArray(v, &array);CHKERRQ(ierr); 4266 for(row = start; row < end; ++row) { 4267 PetscInt ncols, col; 4268 const PetscInt *cols; 4269 const PetscScalar *vals; 4270 4271 array[row - start] = 0.0; 4272 ierr = MatGetRow(mat, row, &ncols, &cols, &vals);CHKERRQ(ierr); 4273 for(col = 0; col < ncols; col++) { 4274 array[row - start] += vals[col]; 4275 } 4276 ierr = MatRestoreRow(mat, row, &ncols, &cols, &vals);CHKERRQ(ierr); 4277 } 4278 ierr = VecRestoreArray(v, &array);CHKERRQ(ierr); 4279 ierr = PetscObjectStateIncrease((PetscObject) v);CHKERRQ(ierr); 4280 PetscFunctionReturn(0); 4281 } 4282 4283 #undef __FUNCT__ 4284 #define __FUNCT__ "MatTranspose" 4285 /*@ 4286 MatTranspose - Computes an in-place or out-of-place transpose of a matrix. 4287 4288 Collective on Mat 4289 4290 Input Parameter: 4291 + mat - the matrix to transpose 4292 - reuse - store the transpose matrix in the provided B 4293 4294 Output Parameters: 4295 . B - the transpose 4296 4297 Notes: 4298 If you pass in &mat for B the transpose will be done in place, for example MatTranspose(mat,MAT_REUSE_MATRIX,&mat); 4299 4300 Level: intermediate 4301 4302 Concepts: matrices^transposing 4303 4304 .seealso: MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse 4305 @*/ 4306 PetscErrorCode MatTranspose(Mat mat,MatReuse reuse,Mat *B) 4307 { 4308 PetscErrorCode ierr; 4309 4310 PetscFunctionBegin; 4311 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4312 PetscValidType(mat,1); 4313 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4314 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4315 if (!mat->ops->transpose) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4316 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4317 4318 ierr = PetscLogEventBegin(MAT_Transpose,mat,0,0,0);CHKERRQ(ierr); 4319 ierr = (*mat->ops->transpose)(mat,reuse,B);CHKERRQ(ierr); 4320 ierr = PetscLogEventEnd(MAT_Transpose,mat,0,0,0);CHKERRQ(ierr); 4321 if (B) {ierr = PetscObjectStateIncrease((PetscObject)*B);CHKERRQ(ierr);} 4322 PetscFunctionReturn(0); 4323 } 4324 4325 #undef __FUNCT__ 4326 #define __FUNCT__ "MatIsTranspose" 4327 /*@ 4328 MatIsTranspose - Test whether a matrix is another one's transpose, 4329 or its own, in which case it tests symmetry. 4330 4331 Collective on Mat 4332 4333 Input Parameter: 4334 + A - the matrix to test 4335 - B - the matrix to test against, this can equal the first parameter 4336 4337 Output Parameters: 4338 . flg - the result 4339 4340 Notes: 4341 Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm 4342 has a running time of the order of the number of nonzeros; the parallel 4343 test involves parallel copies of the block-offdiagonal parts of the matrix. 4344 4345 Level: intermediate 4346 4347 Concepts: matrices^transposing, matrix^symmetry 4348 4349 .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian() 4350 @*/ 4351 PetscErrorCode MatIsTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg) 4352 { 4353 PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool *),(*g)(Mat,Mat,PetscReal,PetscBool *); 4354 4355 PetscFunctionBegin; 4356 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 4357 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 4358 PetscValidPointer(flg,3); 4359 ierr = PetscObjectQueryFunction((PetscObject)A,"MatIsTranspose_C",(void (**)(void))&f);CHKERRQ(ierr); 4360 ierr = PetscObjectQueryFunction((PetscObject)B,"MatIsTranspose_C",(void (**)(void))&g);CHKERRQ(ierr); 4361 if (f && g) { 4362 if (f==g) { 4363 ierr = (*f)(A,B,tol,flg);CHKERRQ(ierr); 4364 } else { 4365 SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for symmetry test"); 4366 } 4367 } 4368 PetscFunctionReturn(0); 4369 } 4370 4371 #undef __FUNCT__ 4372 #define __FUNCT__ "MatHermitianTranspose" 4373 /*@ 4374 MatHermitianTranspose - Computes an in-place or out-of-place transpose of a matrix in complex conjugate. 4375 4376 Collective on Mat 4377 4378 Input Parameter: 4379 + mat - the matrix to transpose and complex conjugate 4380 - reuse - store the transpose matrix in the provided B 4381 4382 Output Parameters: 4383 . B - the Hermitian 4384 4385 Notes: 4386 If you pass in &mat for B the Hermitian will be done in place 4387 4388 Level: intermediate 4389 4390 Concepts: matrices^transposing, complex conjugatex 4391 4392 .seealso: MatTranspose(), MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose(), MatReuse 4393 @*/ 4394 PetscErrorCode MatHermitianTranspose(Mat mat,MatReuse reuse,Mat *B) 4395 { 4396 PetscErrorCode ierr; 4397 4398 PetscFunctionBegin; 4399 ierr = MatTranspose(mat,reuse,B);CHKERRQ(ierr); 4400 #if defined(PETSC_USE_COMPLEX) 4401 ierr = MatConjugate(*B);CHKERRQ(ierr); 4402 #endif 4403 PetscFunctionReturn(0); 4404 } 4405 4406 #undef __FUNCT__ 4407 #define __FUNCT__ "MatIsHermitianTranspose" 4408 /*@ 4409 MatIsHermitianTranspose - Test whether a matrix is another one's Hermitian transpose, 4410 4411 Collective on Mat 4412 4413 Input Parameter: 4414 + A - the matrix to test 4415 - B - the matrix to test against, this can equal the first parameter 4416 4417 Output Parameters: 4418 . flg - the result 4419 4420 Notes: 4421 Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm 4422 has a running time of the order of the number of nonzeros; the parallel 4423 test involves parallel copies of the block-offdiagonal parts of the matrix. 4424 4425 Level: intermediate 4426 4427 Concepts: matrices^transposing, matrix^symmetry 4428 4429 .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian(), MatIsTranspose() 4430 @*/ 4431 PetscErrorCode MatIsHermitianTranspose(Mat A,Mat B,PetscReal tol,PetscBool *flg) 4432 { 4433 PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscBool *),(*g)(Mat,Mat,PetscReal,PetscBool *); 4434 4435 PetscFunctionBegin; 4436 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 4437 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 4438 PetscValidPointer(flg,3); 4439 ierr = PetscObjectQueryFunction((PetscObject)A,"MatIsHermitianTranspose_C",(void (**)(void))&f);CHKERRQ(ierr); 4440 ierr = PetscObjectQueryFunction((PetscObject)B,"MatIsHermitianTranspose_C",(void (**)(void))&g);CHKERRQ(ierr); 4441 if (f && g) { 4442 if (f==g) { 4443 ierr = (*f)(A,B,tol,flg);CHKERRQ(ierr); 4444 } else { 4445 SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for Hermitian test"); 4446 } 4447 } 4448 PetscFunctionReturn(0); 4449 } 4450 4451 #undef __FUNCT__ 4452 #define __FUNCT__ "MatPermute" 4453 /*@ 4454 MatPermute - Creates a new matrix with rows and columns permuted from the 4455 original. 4456 4457 Collective on Mat 4458 4459 Input Parameters: 4460 + mat - the matrix to permute 4461 . row - row permutation, each processor supplies only the permutation for its rows 4462 - col - column permutation, each processor needs the entire column permutation, that is 4463 this is the same size as the total number of columns in the matrix. It can often 4464 be obtained with ISAllGather() on the row permutation 4465 4466 Output Parameters: 4467 . B - the permuted matrix 4468 4469 Level: advanced 4470 4471 Concepts: matrices^permuting 4472 4473 .seealso: MatGetOrdering(), ISAllGather() 4474 4475 @*/ 4476 PetscErrorCode MatPermute(Mat mat,IS row,IS col,Mat *B) 4477 { 4478 PetscErrorCode ierr; 4479 4480 PetscFunctionBegin; 4481 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4482 PetscValidType(mat,1); 4483 PetscValidHeaderSpecific(row,IS_CLASSID,2); 4484 PetscValidHeaderSpecific(col,IS_CLASSID,3); 4485 PetscValidPointer(B,4); 4486 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4487 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4488 if (!mat->ops->permute) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatPermute not available for Mat type %s",((PetscObject)mat)->type_name); 4489 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4490 4491 ierr = (*mat->ops->permute)(mat,row,col,B);CHKERRQ(ierr); 4492 ierr = PetscObjectStateIncrease((PetscObject)*B);CHKERRQ(ierr); 4493 PetscFunctionReturn(0); 4494 } 4495 4496 #undef __FUNCT__ 4497 #define __FUNCT__ "MatEqual" 4498 /*@ 4499 MatEqual - Compares two matrices. 4500 4501 Collective on Mat 4502 4503 Input Parameters: 4504 + A - the first matrix 4505 - B - the second matrix 4506 4507 Output Parameter: 4508 . flg - PETSC_TRUE if the matrices are equal; PETSC_FALSE otherwise. 4509 4510 Level: intermediate 4511 4512 Concepts: matrices^equality between 4513 @*/ 4514 PetscErrorCode MatEqual(Mat A,Mat B,PetscBool *flg) 4515 { 4516 PetscErrorCode ierr; 4517 4518 PetscFunctionBegin; 4519 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 4520 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 4521 PetscValidType(A,1); 4522 PetscValidType(B,2); 4523 PetscValidIntPointer(flg,3); 4524 PetscCheckSameComm(A,1,B,2); 4525 ierr = MatPreallocated(B);CHKERRQ(ierr); 4526 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4527 if (!B->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4528 if (A->rmap->N != B->rmap->N || A->cmap->N != B->cmap->N) SETERRQ4(((PetscObject)A)->comm,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); 4529 if (!A->ops->equal) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)A)->type_name); 4530 if (!B->ops->equal) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)B)->type_name); 4531 if (A->ops->equal != B->ops->equal) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_INCOMP,"A is type: %s\nB is type: %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name); 4532 ierr = MatPreallocated(A);CHKERRQ(ierr); 4533 4534 ierr = (*A->ops->equal)(A,B,flg);CHKERRQ(ierr); 4535 PetscFunctionReturn(0); 4536 } 4537 4538 #undef __FUNCT__ 4539 #define __FUNCT__ "MatDiagonalScale" 4540 /*@ 4541 MatDiagonalScale - Scales a matrix on the left and right by diagonal 4542 matrices that are stored as vectors. Either of the two scaling 4543 matrices can be PETSC_NULL. 4544 4545 Collective on Mat 4546 4547 Input Parameters: 4548 + mat - the matrix to be scaled 4549 . l - the left scaling vector (or PETSC_NULL) 4550 - r - the right scaling vector (or PETSC_NULL) 4551 4552 Notes: 4553 MatDiagonalScale() computes A = LAR, where 4554 L = a diagonal matrix (stored as a vector), R = a diagonal matrix (stored as a vector) 4555 The L scales the rows of the matrix, the R scales the columns of the matrix. 4556 4557 Level: intermediate 4558 4559 Concepts: matrices^diagonal scaling 4560 Concepts: diagonal scaling of matrices 4561 4562 .seealso: MatScale() 4563 @*/ 4564 PetscErrorCode MatDiagonalScale(Mat mat,Vec l,Vec r) 4565 { 4566 PetscErrorCode ierr; 4567 4568 PetscFunctionBegin; 4569 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4570 PetscValidType(mat,1); 4571 if (!mat->ops->diagonalscale) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4572 if (l) {PetscValidHeaderSpecific(l,VEC_CLASSID,2);PetscCheckSameComm(mat,1,l,2);} 4573 if (r) {PetscValidHeaderSpecific(r,VEC_CLASSID,3);PetscCheckSameComm(mat,1,r,3);} 4574 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4575 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4576 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4577 4578 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 4579 ierr = (*mat->ops->diagonalscale)(mat,l,r);CHKERRQ(ierr); 4580 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 4581 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4582 #if defined(PETSC_HAVE_CUSP) 4583 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 4584 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 4585 } 4586 #endif 4587 PetscFunctionReturn(0); 4588 } 4589 4590 #undef __FUNCT__ 4591 #define __FUNCT__ "MatScale" 4592 /*@ 4593 MatScale - Scales all elements of a matrix by a given number. 4594 4595 Logically Collective on Mat 4596 4597 Input Parameters: 4598 + mat - the matrix to be scaled 4599 - a - the scaling value 4600 4601 Output Parameter: 4602 . mat - the scaled matrix 4603 4604 Level: intermediate 4605 4606 Concepts: matrices^scaling all entries 4607 4608 .seealso: MatDiagonalScale() 4609 @*/ 4610 PetscErrorCode MatScale(Mat mat,PetscScalar a) 4611 { 4612 PetscErrorCode ierr; 4613 4614 PetscFunctionBegin; 4615 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4616 PetscValidType(mat,1); 4617 if (a != 1.0 && !mat->ops->scale) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4618 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4619 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4620 PetscValidLogicalCollectiveScalar(mat,a,2); 4621 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4622 4623 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 4624 if (a != 1.0) { 4625 ierr = (*mat->ops->scale)(mat,a);CHKERRQ(ierr); 4626 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4627 } 4628 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 4629 #if defined(PETSC_HAVE_CUSP) 4630 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 4631 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 4632 } 4633 #endif 4634 PetscFunctionReturn(0); 4635 } 4636 4637 #undef __FUNCT__ 4638 #define __FUNCT__ "MatNorm" 4639 /*@ 4640 MatNorm - Calculates various norms of a matrix. 4641 4642 Collective on Mat 4643 4644 Input Parameters: 4645 + mat - the matrix 4646 - type - the type of norm, NORM_1, NORM_FROBENIUS, NORM_INFINITY 4647 4648 Output Parameters: 4649 . nrm - the resulting norm 4650 4651 Level: intermediate 4652 4653 Concepts: matrices^norm 4654 Concepts: norm^of matrix 4655 @*/ 4656 PetscErrorCode MatNorm(Mat mat,NormType type,PetscReal *nrm) 4657 { 4658 PetscErrorCode ierr; 4659 4660 PetscFunctionBegin; 4661 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4662 PetscValidType(mat,1); 4663 PetscValidScalarPointer(nrm,3); 4664 4665 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4666 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4667 if (!mat->ops->norm) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 4668 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4669 4670 ierr = (*mat->ops->norm)(mat,type,nrm);CHKERRQ(ierr); 4671 PetscFunctionReturn(0); 4672 } 4673 4674 /* 4675 This variable is used to prevent counting of MatAssemblyBegin() that 4676 are called from within a MatAssemblyEnd(). 4677 */ 4678 static PetscInt MatAssemblyEnd_InUse = 0; 4679 #undef __FUNCT__ 4680 #define __FUNCT__ "MatAssemblyBegin" 4681 /*@ 4682 MatAssemblyBegin - Begins assembling the matrix. This routine should 4683 be called after completing all calls to MatSetValues(). 4684 4685 Collective on Mat 4686 4687 Input Parameters: 4688 + mat - the matrix 4689 - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY 4690 4691 Notes: 4692 MatSetValues() generally caches the values. The matrix is ready to 4693 use only after MatAssemblyBegin() and MatAssemblyEnd() have been called. 4694 Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES 4695 in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before 4696 using the matrix. 4697 4698 Level: beginner 4699 4700 Concepts: matrices^assembling 4701 4702 .seealso: MatAssemblyEnd(), MatSetValues(), MatAssembled() 4703 @*/ 4704 PetscErrorCode MatAssemblyBegin(Mat mat,MatAssemblyType type) 4705 { 4706 PetscErrorCode ierr; 4707 4708 PetscFunctionBegin; 4709 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4710 PetscValidType(mat,1); 4711 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4712 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix.\nDid you forget to call MatSetUnfactored()?"); 4713 if (mat->assembled) { 4714 mat->was_assembled = PETSC_TRUE; 4715 mat->assembled = PETSC_FALSE; 4716 } 4717 if (!MatAssemblyEnd_InUse) { 4718 ierr = PetscLogEventBegin(MAT_AssemblyBegin,mat,0,0,0);CHKERRQ(ierr); 4719 if (mat->ops->assemblybegin){ierr = (*mat->ops->assemblybegin)(mat,type);CHKERRQ(ierr);} 4720 ierr = PetscLogEventEnd(MAT_AssemblyBegin,mat,0,0,0);CHKERRQ(ierr); 4721 } else { 4722 if (mat->ops->assemblybegin){ierr = (*mat->ops->assemblybegin)(mat,type);CHKERRQ(ierr);} 4723 } 4724 PetscFunctionReturn(0); 4725 } 4726 4727 #undef __FUNCT__ 4728 #define __FUNCT__ "MatAssembled" 4729 /*@ 4730 MatAssembled - Indicates if a matrix has been assembled and is ready for 4731 use; for example, in matrix-vector product. 4732 4733 Not Collective 4734 4735 Input Parameter: 4736 . mat - the matrix 4737 4738 Output Parameter: 4739 . assembled - PETSC_TRUE or PETSC_FALSE 4740 4741 Level: advanced 4742 4743 Concepts: matrices^assembled? 4744 4745 .seealso: MatAssemblyEnd(), MatSetValues(), MatAssemblyBegin() 4746 @*/ 4747 PetscErrorCode MatAssembled(Mat mat,PetscBool *assembled) 4748 { 4749 PetscFunctionBegin; 4750 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4751 PetscValidType(mat,1); 4752 PetscValidPointer(assembled,2); 4753 *assembled = mat->assembled; 4754 PetscFunctionReturn(0); 4755 } 4756 4757 #undef __FUNCT__ 4758 #define __FUNCT__ "MatView_Private" 4759 /* 4760 Processes command line options to determine if/how a matrix 4761 is to be viewed. Called by MatAssemblyEnd() and MatLoad(). 4762 */ 4763 PetscErrorCode MatView_Private(Mat mat) 4764 { 4765 PetscErrorCode ierr; 4766 PetscBool flg1 = PETSC_FALSE,flg2 = PETSC_FALSE,flg3 = PETSC_FALSE,flg4 = PETSC_FALSE,flg6 = PETSC_FALSE,flg7 = PETSC_FALSE,flg8 = PETSC_FALSE; 4767 static PetscBool incall = PETSC_FALSE; 4768 #if defined(PETSC_USE_SOCKET_VIEWER) 4769 PetscBool flg5 = PETSC_FALSE; 4770 #endif 4771 4772 PetscFunctionBegin; 4773 if (incall) PetscFunctionReturn(0); 4774 incall = PETSC_TRUE; 4775 ierr = PetscOptionsBegin(((PetscObject)mat)->comm,((PetscObject)mat)->prefix,"Matrix Options","Mat");CHKERRQ(ierr); 4776 ierr = PetscOptionsBool("-mat_view_info","Information on matrix size","MatView",flg1,&flg1,PETSC_NULL);CHKERRQ(ierr); 4777 ierr = PetscOptionsBool("-mat_view_info_detailed","Nonzeros in the matrix","MatView",flg2,&flg2,PETSC_NULL);CHKERRQ(ierr); 4778 ierr = PetscOptionsBool("-mat_view","Print matrix to stdout","MatView",flg3,&flg3,PETSC_NULL);CHKERRQ(ierr); 4779 ierr = PetscOptionsBool("-mat_view_matlab","Print matrix to stdout in a format Matlab can read","MatView",flg4,&flg4,PETSC_NULL);CHKERRQ(ierr); 4780 #if defined(PETSC_USE_SOCKET_VIEWER) 4781 ierr = PetscOptionsBool("-mat_view_socket","Send matrix to socket (can be read from matlab)","MatView",flg5,&flg5,PETSC_NULL);CHKERRQ(ierr); 4782 #endif 4783 ierr = PetscOptionsBool("-mat_view_binary","Save matrix to file in binary format","MatView",flg6,&flg6,PETSC_NULL);CHKERRQ(ierr); 4784 ierr = PetscOptionsBool("-mat_view_draw","Draw the matrix nonzero structure","MatView",flg7,&flg7,PETSC_NULL);CHKERRQ(ierr); 4785 ierr = PetscOptionsEnd();CHKERRQ(ierr); 4786 4787 if (flg1) { 4788 PetscViewer viewer; 4789 4790 ierr = PetscViewerASCIIGetStdout(((PetscObject)mat)->comm,&viewer);CHKERRQ(ierr); 4791 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_INFO);CHKERRQ(ierr); 4792 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4793 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4794 } 4795 if (flg2) { 4796 PetscViewer viewer; 4797 4798 ierr = PetscViewerASCIIGetStdout(((PetscObject)mat)->comm,&viewer);CHKERRQ(ierr); 4799 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_INFO_DETAIL);CHKERRQ(ierr); 4800 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4801 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4802 } 4803 if (flg3) { 4804 PetscViewer viewer; 4805 4806 ierr = PetscViewerASCIIGetStdout(((PetscObject)mat)->comm,&viewer);CHKERRQ(ierr); 4807 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4808 } 4809 if (flg4) { 4810 PetscViewer viewer; 4811 4812 ierr = PetscViewerASCIIGetStdout(((PetscObject)mat)->comm,&viewer);CHKERRQ(ierr); 4813 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_MATLAB);CHKERRQ(ierr); 4814 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4815 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4816 } 4817 #if defined(PETSC_USE_SOCKET_VIEWER) 4818 if (flg5) { 4819 ierr = MatView(mat,PETSC_VIEWER_SOCKET_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4820 ierr = PetscViewerFlush(PETSC_VIEWER_SOCKET_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4821 } 4822 #endif 4823 if (flg6) { 4824 ierr = MatView(mat,PETSC_VIEWER_BINARY_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4825 ierr = PetscViewerFlush(PETSC_VIEWER_BINARY_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4826 } 4827 if (flg7) { 4828 ierr = PetscOptionsGetBool(((PetscObject)mat)->prefix,"-mat_view_contour",&flg8,PETSC_NULL);CHKERRQ(ierr); 4829 if (flg8) { 4830 PetscViewerPushFormat(PETSC_VIEWER_DRAW_(((PetscObject)mat)->comm),PETSC_VIEWER_DRAW_CONTOUR);CHKERRQ(ierr); 4831 } 4832 ierr = MatView(mat,PETSC_VIEWER_DRAW_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4833 ierr = PetscViewerFlush(PETSC_VIEWER_DRAW_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4834 if (flg8) { 4835 PetscViewerPopFormat(PETSC_VIEWER_DRAW_(((PetscObject)mat)->comm));CHKERRQ(ierr); 4836 } 4837 } 4838 incall = PETSC_FALSE; 4839 PetscFunctionReturn(0); 4840 } 4841 4842 #undef __FUNCT__ 4843 #define __FUNCT__ "MatAssemblyEnd" 4844 /*@ 4845 MatAssemblyEnd - Completes assembling the matrix. This routine should 4846 be called after MatAssemblyBegin(). 4847 4848 Collective on Mat 4849 4850 Input Parameters: 4851 + mat - the matrix 4852 - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY 4853 4854 Options Database Keys: 4855 + -mat_view_info - Prints info on matrix at conclusion of MatEndAssembly() 4856 . -mat_view_info_detailed - Prints more detailed info 4857 . -mat_view - Prints matrix in ASCII format 4858 . -mat_view_matlab - Prints matrix in Matlab format 4859 . -mat_view_draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX(). 4860 . -display <name> - Sets display name (default is host) 4861 . -draw_pause <sec> - Sets number of seconds to pause after display 4862 . -mat_view_socket - Sends matrix to socket, can be accessed from Matlab (See the <a href="../../docs/manual.pdf">users manual</a>) 4863 . -viewer_socket_machine <machine> 4864 . -viewer_socket_port <port> 4865 . -mat_view_binary - save matrix to file in binary format 4866 - -viewer_binary_filename <name> 4867 4868 Notes: 4869 MatSetValues() generally caches the values. The matrix is ready to 4870 use only after MatAssemblyBegin() and MatAssemblyEnd() have been called. 4871 Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES 4872 in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before 4873 using the matrix. 4874 4875 Level: beginner 4876 4877 .seealso: MatAssemblyBegin(), MatSetValues(), PetscDrawOpenX(), MatView(), MatAssembled(), PetscViewerSocketOpen() 4878 @*/ 4879 PetscErrorCode MatAssemblyEnd(Mat mat,MatAssemblyType type) 4880 { 4881 PetscErrorCode ierr; 4882 static PetscInt inassm = 0; 4883 PetscBool flg = PETSC_FALSE; 4884 4885 PetscFunctionBegin; 4886 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4887 PetscValidType(mat,1); 4888 4889 inassm++; 4890 MatAssemblyEnd_InUse++; 4891 if (MatAssemblyEnd_InUse == 1) { /* Do the logging only the first time through */ 4892 ierr = PetscLogEventBegin(MAT_AssemblyEnd,mat,0,0,0);CHKERRQ(ierr); 4893 if (mat->ops->assemblyend) { 4894 ierr = (*mat->ops->assemblyend)(mat,type);CHKERRQ(ierr); 4895 } 4896 ierr = PetscLogEventEnd(MAT_AssemblyEnd,mat,0,0,0);CHKERRQ(ierr); 4897 } else { 4898 if (mat->ops->assemblyend) { 4899 ierr = (*mat->ops->assemblyend)(mat,type);CHKERRQ(ierr); 4900 } 4901 } 4902 4903 /* Flush assembly is not a true assembly */ 4904 if (type != MAT_FLUSH_ASSEMBLY) { 4905 mat->assembled = PETSC_TRUE; mat->num_ass++; 4906 } 4907 mat->insertmode = NOT_SET_VALUES; 4908 MatAssemblyEnd_InUse--; 4909 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4910 if (!mat->symmetric_eternal) { 4911 mat->symmetric_set = PETSC_FALSE; 4912 mat->hermitian_set = PETSC_FALSE; 4913 mat->structurally_symmetric_set = PETSC_FALSE; 4914 } 4915 if (inassm == 1 && type != MAT_FLUSH_ASSEMBLY) { 4916 ierr = MatView_Private(mat);CHKERRQ(ierr); 4917 ierr = PetscOptionsGetBool(((PetscObject)mat)->prefix,"-mat_is_symmetric",&flg,PETSC_NULL);CHKERRQ(ierr); 4918 if (flg) { 4919 PetscReal tol = 0.0; 4920 ierr = PetscOptionsGetReal(((PetscObject)mat)->prefix,"-mat_is_symmetric",&tol,PETSC_NULL);CHKERRQ(ierr); 4921 ierr = MatIsSymmetric(mat,tol,&flg);CHKERRQ(ierr); 4922 if (flg) { 4923 ierr = PetscPrintf(((PetscObject)mat)->comm,"Matrix is symmetric (tolerance %G)\n",tol);CHKERRQ(ierr); 4924 } else { 4925 ierr = PetscPrintf(((PetscObject)mat)->comm,"Matrix is not symmetric (tolerance %G)\n",tol);CHKERRQ(ierr); 4926 } 4927 } 4928 } 4929 inassm--; 4930 #if defined(PETSC_HAVE_CUSP) 4931 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 4932 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 4933 } 4934 #endif 4935 PetscFunctionReturn(0); 4936 } 4937 4938 #undef __FUNCT__ 4939 #define __FUNCT__ "MatSetOption" 4940 /*@ 4941 MatSetOption - Sets a parameter option for a matrix. Some options 4942 may be specific to certain storage formats. Some options 4943 determine how values will be inserted (or added). Sorted, 4944 row-oriented input will generally assemble the fastest. The default 4945 is row-oriented, nonsorted input. 4946 4947 Logically Collective on Mat 4948 4949 Input Parameters: 4950 + mat - the matrix 4951 . option - the option, one of those listed below (and possibly others), 4952 - flg - turn the option on (PETSC_TRUE) or off (PETSC_FALSE) 4953 4954 Options Describing Matrix Structure: 4955 + MAT_SPD - symmetric positive definite 4956 - MAT_SYMMETRIC - symmetric in terms of both structure and value 4957 . MAT_HERMITIAN - transpose is the complex conjugation 4958 . MAT_STRUCTURALLY_SYMMETRIC - symmetric nonzero structure 4959 - MAT_SYMMETRY_ETERNAL - if you would like the symmetry/Hermitian flag 4960 you set to be kept with all future use of the matrix 4961 including after MatAssemblyBegin/End() which could 4962 potentially change the symmetry structure, i.e. you 4963 KNOW the matrix will ALWAYS have the property you set. 4964 4965 4966 Options For Use with MatSetValues(): 4967 Insert a logically dense subblock, which can be 4968 . MAT_ROW_ORIENTED - row-oriented (default) 4969 4970 Note these options reflect the data you pass in with MatSetValues(); it has 4971 nothing to do with how the data is stored internally in the matrix 4972 data structure. 4973 4974 When (re)assembling a matrix, we can restrict the input for 4975 efficiency/debugging purposes. These options include 4976 + MAT_NEW_NONZERO_LOCATIONS - additional insertions will be 4977 allowed if they generate a new nonzero 4978 . MAT_NEW_DIAGONALS - new diagonals will be allowed (for block diagonal format only) 4979 . MAT_IGNORE_OFF_PROC_ENTRIES - drops off-processor entries 4980 . MAT_NEW_NONZERO_LOCATION_ERR - generates an error for new matrix entry 4981 . MAT_USE_HASH_TABLE - uses a hash table to speed up matrix assembly 4982 + MAT_NO_OFF_PROC_ENTRIES - you know each process will only set values for its own rows, will generate an error if 4983 any process sets values for another process. This avoids all reductions in the MatAssembly routines and thus improves 4984 performance for very large process counts. 4985 4986 Notes: 4987 Some options are relevant only for particular matrix types and 4988 are thus ignored by others. Other options are not supported by 4989 certain matrix types and will generate an error message if set. 4990 4991 If using a Fortran 77 module to compute a matrix, one may need to 4992 use the column-oriented option (or convert to the row-oriented 4993 format). 4994 4995 MAT_NEW_NONZERO_LOCATIONS set to PETSC_FALSE indicates that any add or insertion 4996 that would generate a new entry in the nonzero structure is instead 4997 ignored. Thus, if memory has not alredy been allocated for this particular 4998 data, then the insertion is ignored. For dense matrices, in which 4999 the entire array is allocated, no entries are ever ignored. 5000 Set after the first MatAssemblyEnd() 5001 5002 MAT_NEW_NONZERO_LOCATION_ERR indicates that any add or insertion 5003 that would generate a new entry in the nonzero structure instead produces 5004 an error. (Currently supported for AIJ and BAIJ formats only.) 5005 This is a useful flag when using SAME_NONZERO_PATTERN in calling 5006 KSPSetOperators() to ensure that the nonzero pattern truely does 5007 remain unchanged. Set after the first MatAssemblyEnd() 5008 5009 MAT_NEW_NONZERO_ALLOCATION_ERR indicates that any add or insertion 5010 that would generate a new entry that has not been preallocated will 5011 instead produce an error. (Currently supported for AIJ and BAIJ formats 5012 only.) This is a useful flag when debugging matrix memory preallocation. 5013 5014 MAT_IGNORE_OFF_PROC_ENTRIES indicates entries destined for 5015 other processors should be dropped, rather than stashed. 5016 This is useful if you know that the "owning" processor is also 5017 always generating the correct matrix entries, so that PETSc need 5018 not transfer duplicate entries generated on another processor. 5019 5020 MAT_USE_HASH_TABLE indicates that a hash table be used to improve the 5021 searches during matrix assembly. When this flag is set, the hash table 5022 is created during the first Matrix Assembly. This hash table is 5023 used the next time through, during MatSetVaules()/MatSetVaulesBlocked() 5024 to improve the searching of indices. MAT_NEW_NONZERO_LOCATIONS flag 5025 should be used with MAT_USE_HASH_TABLE flag. This option is currently 5026 supported by MATMPIBAIJ format only. 5027 5028 MAT_KEEP_NONZERO_PATTERN indicates when MatZeroRows() is called the zeroed entries 5029 are kept in the nonzero structure 5030 5031 MAT_IGNORE_ZERO_ENTRIES - for AIJ/IS matrices this will stop zero values from creating 5032 a zero location in the matrix 5033 5034 MAT_USE_INODES - indicates using inode version of the code - works with AIJ and 5035 ROWBS matrix types 5036 5037 MAT_NO_OFF_PROC_ZERO_ROWS - you know each process will only zero its own rows. This avoids all reductions in the 5038 zero row routines and thus improves performance for very large process counts. 5039 5040 Level: intermediate 5041 5042 Concepts: matrices^setting options 5043 5044 @*/ 5045 PetscErrorCode MatSetOption(Mat mat,MatOption op,PetscBool flg) 5046 { 5047 PetscErrorCode ierr; 5048 5049 PetscFunctionBegin; 5050 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5051 PetscValidType(mat,1); 5052 PetscValidLogicalCollectiveEnum(mat,op,2); 5053 PetscValidLogicalCollectiveBool(mat,flg,3); 5054 5055 if (((int) op) < 0 || ((int) op) >= NUM_MAT_OPTIONS) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Options %d is out of range",(int)op); 5056 if (!((PetscObject)mat)->type_name) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_TYPENOTSET,"Cannot set options until type and size have been set, see MatSetType() and MatSetSizes()"); 5057 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5058 switch (op) { 5059 case MAT_NO_OFF_PROC_ENTRIES: 5060 mat->nooffprocentries = flg; 5061 PetscFunctionReturn(0); 5062 break; 5063 case MAT_NO_OFF_PROC_ZERO_ROWS: 5064 mat->nooffproczerorows = flg; 5065 PetscFunctionReturn(0); 5066 break; 5067 case MAT_SPD: 5068 mat->spd_set = PETSC_TRUE; 5069 mat->spd = flg; 5070 if (flg) { 5071 mat->symmetric = PETSC_TRUE; 5072 mat->structurally_symmetric = PETSC_TRUE; 5073 mat->symmetric_set = PETSC_TRUE; 5074 mat->structurally_symmetric_set = PETSC_TRUE; 5075 } 5076 break; 5077 case MAT_SYMMETRIC: 5078 mat->symmetric = flg; 5079 if (flg) mat->structurally_symmetric = PETSC_TRUE; 5080 mat->symmetric_set = PETSC_TRUE; 5081 mat->structurally_symmetric_set = flg; 5082 break; 5083 case MAT_HERMITIAN: 5084 mat->hermitian = flg; 5085 if (flg) mat->structurally_symmetric = PETSC_TRUE; 5086 mat->hermitian_set = PETSC_TRUE; 5087 mat->structurally_symmetric_set = flg; 5088 break; 5089 case MAT_STRUCTURALLY_SYMMETRIC: 5090 mat->structurally_symmetric = flg; 5091 mat->structurally_symmetric_set = PETSC_TRUE; 5092 break; 5093 case MAT_SYMMETRY_ETERNAL: 5094 mat->symmetric_eternal = flg; 5095 break; 5096 default: 5097 break; 5098 } 5099 if (mat->ops->setoption) { 5100 ierr = (*mat->ops->setoption)(mat,op,flg);CHKERRQ(ierr); 5101 } 5102 PetscFunctionReturn(0); 5103 } 5104 5105 #undef __FUNCT__ 5106 #define __FUNCT__ "MatZeroEntries" 5107 /*@ 5108 MatZeroEntries - Zeros all entries of a matrix. For sparse matrices 5109 this routine retains the old nonzero structure. 5110 5111 Logically Collective on Mat 5112 5113 Input Parameters: 5114 . mat - the matrix 5115 5116 Level: intermediate 5117 5118 Concepts: matrices^zeroing 5119 5120 .seealso: MatZeroRows() 5121 @*/ 5122 PetscErrorCode MatZeroEntries(Mat mat) 5123 { 5124 PetscErrorCode ierr; 5125 5126 PetscFunctionBegin; 5127 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5128 PetscValidType(mat,1); 5129 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5130 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"); 5131 if (!mat->ops->zeroentries) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 5132 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5133 5134 ierr = PetscLogEventBegin(MAT_ZeroEntries,mat,0,0,0);CHKERRQ(ierr); 5135 ierr = (*mat->ops->zeroentries)(mat);CHKERRQ(ierr); 5136 ierr = PetscLogEventEnd(MAT_ZeroEntries,mat,0,0,0);CHKERRQ(ierr); 5137 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5138 #if defined(PETSC_HAVE_CUSP) 5139 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 5140 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 5141 } 5142 #endif 5143 PetscFunctionReturn(0); 5144 } 5145 5146 #undef __FUNCT__ 5147 #define __FUNCT__ "MatZeroRowsColumns" 5148 /*@C 5149 MatZeroRowsColumns - Zeros all entries (except possibly the main diagonal) 5150 of a set of rows and columns of a matrix. 5151 5152 Collective on Mat 5153 5154 Input Parameters: 5155 + mat - the matrix 5156 . numRows - the number of rows to remove 5157 . rows - the global row indices 5158 . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry) 5159 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5160 - b - optional vector of right hand side, that will be adjusted by provided solution 5161 5162 Notes: 5163 This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix. 5164 5165 The user can set a value in the diagonal entry (or for the AIJ and 5166 row formats can optionally remove the main diagonal entry from the 5167 nonzero structure as well, by passing 0.0 as the final argument). 5168 5169 For the parallel case, all processes that share the matrix (i.e., 5170 those in the communicator used for matrix creation) MUST call this 5171 routine, regardless of whether any rows being zeroed are owned by 5172 them. 5173 5174 Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to 5175 list only rows local to itself). 5176 5177 The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine. 5178 5179 Level: intermediate 5180 5181 Concepts: matrices^zeroing rows 5182 5183 .seealso: MatZeroRowsIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(), MatZeroRowsColumnsIS() 5184 @*/ 5185 PetscErrorCode MatZeroRowsColumns(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 5186 { 5187 PetscErrorCode ierr; 5188 5189 PetscFunctionBegin; 5190 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5191 PetscValidType(mat,1); 5192 if (numRows) PetscValidIntPointer(rows,3); 5193 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5194 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5195 if (!mat->ops->zerorowscolumns) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 5196 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5197 5198 ierr = (*mat->ops->zerorowscolumns)(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5199 ierr = MatView_Private(mat);CHKERRQ(ierr); 5200 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5201 #if defined(PETSC_HAVE_CUSP) 5202 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 5203 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 5204 } 5205 #endif 5206 PetscFunctionReturn(0); 5207 } 5208 5209 #undef __FUNCT__ 5210 #define __FUNCT__ "MatZeroRowsColumnsIS" 5211 /*@C 5212 MatZeroRowsColumnsIS - Zeros all entries (except possibly the main diagonal) 5213 of a set of rows and columns of a matrix. 5214 5215 Collective on Mat 5216 5217 Input Parameters: 5218 + mat - the matrix 5219 . is - the rows to zero 5220 . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry) 5221 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5222 - b - optional vector of right hand side, that will be adjusted by provided solution 5223 5224 Notes: 5225 This does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix. 5226 5227 The user can set a value in the diagonal entry (or for the AIJ and 5228 row formats can optionally remove the main diagonal entry from the 5229 nonzero structure as well, by passing 0.0 as the final argument). 5230 5231 For the parallel case, all processes that share the matrix (i.e., 5232 those in the communicator used for matrix creation) MUST call this 5233 routine, regardless of whether any rows being zeroed are owned by 5234 them. 5235 5236 Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to 5237 list only rows local to itself). 5238 5239 The option MAT_NO_OFF_PROC_ZERO_ROWS does not apply to this routine. 5240 5241 Level: intermediate 5242 5243 Concepts: matrices^zeroing rows 5244 5245 .seealso: MatZeroRowsIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption(), MatZeroRowsColumns() 5246 @*/ 5247 PetscErrorCode MatZeroRowsColumnsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b) 5248 { 5249 PetscErrorCode ierr; 5250 PetscInt numRows; 5251 const PetscInt *rows; 5252 5253 PetscFunctionBegin; 5254 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5255 PetscValidHeaderSpecific(is,IS_CLASSID,2); 5256 PetscValidType(mat,1); 5257 PetscValidType(is,2); 5258 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 5259 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 5260 ierr = MatZeroRowsColumns(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5261 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 5262 PetscFunctionReturn(0); 5263 } 5264 5265 #undef __FUNCT__ 5266 #define __FUNCT__ "MatZeroRows" 5267 /*@C 5268 MatZeroRows - Zeros all entries (except possibly the main diagonal) 5269 of a set of rows of a matrix. 5270 5271 Collective on Mat 5272 5273 Input Parameters: 5274 + mat - the matrix 5275 . numRows - the number of rows to remove 5276 . rows - the global row indices 5277 . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry) 5278 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5279 - b - optional vector of right hand side, that will be adjusted by provided solution 5280 5281 Notes: 5282 For the AIJ and BAIJ matrix formats this removes the old nonzero structure, 5283 but does not release memory. For the dense and block diagonal 5284 formats this does not alter the nonzero structure. 5285 5286 If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure 5287 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 5288 merely zeroed. 5289 5290 The user can set a value in the diagonal entry (or for the AIJ and 5291 row formats can optionally remove the main diagonal entry from the 5292 nonzero structure as well, by passing 0.0 as the final argument). 5293 5294 For the parallel case, all processes that share the matrix (i.e., 5295 those in the communicator used for matrix creation) MUST call this 5296 routine, regardless of whether any rows being zeroed are owned by 5297 them. 5298 5299 Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to 5300 list only rows local to itself). 5301 5302 You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it 5303 owns that are to be zeroed. This saves a global synchronization in the implementation. 5304 5305 Level: intermediate 5306 5307 Concepts: matrices^zeroing rows 5308 5309 .seealso: MatZeroRowsIS(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption() 5310 @*/ 5311 PetscErrorCode MatZeroRows(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 5312 { 5313 PetscErrorCode ierr; 5314 5315 PetscFunctionBegin; 5316 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5317 PetscValidType(mat,1); 5318 if (numRows) PetscValidIntPointer(rows,3); 5319 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5320 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5321 if (!mat->ops->zerorows) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 5322 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5323 5324 ierr = (*mat->ops->zerorows)(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5325 ierr = MatView_Private(mat);CHKERRQ(ierr); 5326 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5327 #if defined(PETSC_HAVE_CUSP) 5328 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 5329 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 5330 } 5331 #endif 5332 PetscFunctionReturn(0); 5333 } 5334 5335 #undef __FUNCT__ 5336 #define __FUNCT__ "MatZeroRowsIS" 5337 /*@C 5338 MatZeroRowsIS - Zeros all entries (except possibly the main diagonal) 5339 of a set of rows of a matrix. 5340 5341 Collective on Mat 5342 5343 Input Parameters: 5344 + mat - the matrix 5345 . is - index set of rows to remove 5346 . diag - value put in all diagonals of eliminated rows 5347 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5348 - b - optional vector of right hand side, that will be adjusted by provided solution 5349 5350 Notes: 5351 For the AIJ and BAIJ matrix formats this removes the old nonzero structure, 5352 but does not release memory. For the dense and block diagonal 5353 formats this does not alter the nonzero structure. 5354 5355 If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure 5356 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 5357 merely zeroed. 5358 5359 The user can set a value in the diagonal entry (or for the AIJ and 5360 row formats can optionally remove the main diagonal entry from the 5361 nonzero structure as well, by passing 0.0 as the final argument). 5362 5363 For the parallel case, all processes that share the matrix (i.e., 5364 those in the communicator used for matrix creation) MUST call this 5365 routine, regardless of whether any rows being zeroed are owned by 5366 them. 5367 5368 Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to 5369 list only rows local to itself). 5370 5371 You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it 5372 owns that are to be zeroed. This saves a global synchronization in the implementation. 5373 5374 Level: intermediate 5375 5376 Concepts: matrices^zeroing rows 5377 5378 .seealso: MatZeroRows(), MatZeroRowsStencil(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption() 5379 @*/ 5380 PetscErrorCode MatZeroRowsIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b) 5381 { 5382 PetscInt numRows; 5383 const PetscInt *rows; 5384 PetscErrorCode ierr; 5385 5386 PetscFunctionBegin; 5387 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5388 PetscValidType(mat,1); 5389 PetscValidHeaderSpecific(is,IS_CLASSID,2); 5390 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 5391 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 5392 ierr = MatZeroRows(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5393 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 5394 PetscFunctionReturn(0); 5395 } 5396 5397 #undef __FUNCT__ 5398 #define __FUNCT__ "MatZeroRowsStencil" 5399 /*@C 5400 MatZeroRowsStencil - Zeros all entries (except possibly the main diagonal) 5401 of a set of rows of a matrix. These rows must be local to the process. 5402 5403 Collective on Mat 5404 5405 Input Parameters: 5406 + mat - the matrix 5407 . numRows - the number of rows to remove 5408 . rows - the grid coordinates (and component number when dof > 1) for matrix rows 5409 . diag - value put in all diagonals of eliminated rows (0.0 will even eliminate diagonal entry) 5410 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5411 - b - optional vector of right hand side, that will be adjusted by provided solution 5412 5413 Notes: 5414 For the AIJ and BAIJ matrix formats this removes the old nonzero structure, 5415 but does not release memory. For the dense and block diagonal 5416 formats this does not alter the nonzero structure. 5417 5418 If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure 5419 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 5420 merely zeroed. 5421 5422 The user can set a value in the diagonal entry (or for the AIJ and 5423 row formats can optionally remove the main diagonal entry from the 5424 nonzero structure as well, by passing 0.0 as the final argument). 5425 5426 For the parallel case, all processes that share the matrix (i.e., 5427 those in the communicator used for matrix creation) MUST call this 5428 routine, regardless of whether any rows being zeroed are owned by 5429 them. 5430 5431 Each processor can indicate any rows in the entire matrix to be zeroed (i.e. each process does NOT have to 5432 list only rows local to itself). 5433 5434 The grid coordinates are across the entire grid, not just the local portion 5435 5436 In Fortran idxm and idxn should be declared as 5437 $ MatStencil idxm(4,m) 5438 and the values inserted using 5439 $ idxm(MatStencil_i,1) = i 5440 $ idxm(MatStencil_j,1) = j 5441 $ idxm(MatStencil_k,1) = k 5442 $ idxm(MatStencil_c,1) = c 5443 etc 5444 5445 For periodic boundary conditions use negative indices for values to the left (below 0; that are to be 5446 obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one 5447 etc to obtain values that obtained by wrapping the values from the left edge. This does not work for anything but the 5448 DMDA_BOUNDARY_PERIODIC boundary type. 5449 5450 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 5451 a single value per point) you can skip filling those indices. 5452 5453 Level: intermediate 5454 5455 Concepts: matrices^zeroing rows 5456 5457 .seealso: MatZeroRows(), MatZeroRowsIS(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption() 5458 @*/ 5459 PetscErrorCode MatZeroRowsStencil(Mat mat,PetscInt numRows,const MatStencil rows[],PetscScalar diag,Vec x,Vec b) 5460 { 5461 PetscInt dim = mat->stencil.dim; 5462 PetscInt sdim = dim - (1 - (PetscInt) mat->stencil.noc); 5463 PetscInt *dims = mat->stencil.dims+1; 5464 PetscInt *starts = mat->stencil.starts; 5465 PetscInt *dxm = (PetscInt *) rows; 5466 PetscInt *jdxm, i, j, tmp, numNewRows = 0; 5467 PetscErrorCode ierr; 5468 5469 PetscFunctionBegin; 5470 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5471 PetscValidType(mat,1); 5472 if (numRows) PetscValidIntPointer(rows,3); 5473 5474 ierr = PetscMalloc(numRows * sizeof(PetscInt), &jdxm);CHKERRQ(ierr); 5475 for(i = 0; i < numRows; ++i) { 5476 /* Skip unused dimensions (they are ordered k, j, i, c) */ 5477 for(j = 0; j < 3-sdim; ++j) dxm++; 5478 /* Local index in X dir */ 5479 tmp = *dxm++ - starts[0]; 5480 /* Loop over remaining dimensions */ 5481 for(j = 0; j < dim-1; ++j) { 5482 /* If nonlocal, set index to be negative */ 5483 if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 5484 /* Update local index */ 5485 else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1]; 5486 } 5487 /* Skip component slot if necessary */ 5488 if (mat->stencil.noc) dxm++; 5489 /* Local row number */ 5490 if (tmp >= 0) { 5491 jdxm[numNewRows++] = tmp; 5492 } 5493 } 5494 ierr = MatZeroRowsLocal(mat,numNewRows,jdxm,diag,x,b);CHKERRQ(ierr); 5495 ierr = PetscFree(jdxm);CHKERRQ(ierr); 5496 PetscFunctionReturn(0); 5497 } 5498 5499 #undef __FUNCT__ 5500 #define __FUNCT__ "MatZeroRowsLocal" 5501 /*@C 5502 MatZeroRowsLocal - Zeros all entries (except possibly the main diagonal) 5503 of a set of rows of a matrix; using local numbering of rows. 5504 5505 Collective on Mat 5506 5507 Input Parameters: 5508 + mat - the matrix 5509 . numRows - the number of rows to remove 5510 . rows - the global row indices 5511 . diag - value put in all diagonals of eliminated rows 5512 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5513 - b - optional vector of right hand side, that will be adjusted by provided solution 5514 5515 Notes: 5516 Before calling MatZeroRowsLocal(), the user must first set the 5517 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 5518 5519 For the AIJ matrix formats this removes the old nonzero structure, 5520 but does not release memory. For the dense and block diagonal 5521 formats this does not alter the nonzero structure. 5522 5523 If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure 5524 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 5525 merely zeroed. 5526 5527 The user can set a value in the diagonal entry (or for the AIJ and 5528 row formats can optionally remove the main diagonal entry from the 5529 nonzero structure as well, by passing 0.0 as the final argument). 5530 5531 You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it 5532 owns that are to be zeroed. This saves a global synchronization in the implementation. 5533 5534 Level: intermediate 5535 5536 Concepts: matrices^zeroing 5537 5538 .seealso: MatZeroRows(), MatZeroRowsLocalIS(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 5539 @*/ 5540 PetscErrorCode MatZeroRowsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 5541 { 5542 PetscErrorCode ierr; 5543 PetscMPIInt size; 5544 5545 PetscFunctionBegin; 5546 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5547 PetscValidType(mat,1); 5548 if (numRows) PetscValidIntPointer(rows,3); 5549 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5550 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5551 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5552 5553 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 5554 if (mat->ops->zerorowslocal) { 5555 ierr = (*mat->ops->zerorowslocal)(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5556 } else if (size == 1) { 5557 ierr = (*mat->ops->zerorows)(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5558 } else { 5559 IS is, newis; 5560 const PetscInt *newRows; 5561 5562 if (!mat->rmapping) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first"); 5563 ierr = ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);CHKERRQ(ierr); 5564 ierr = ISLocalToGlobalMappingApplyIS(mat->rmapping,is,&newis);CHKERRQ(ierr); 5565 ierr = ISGetIndices(newis,&newRows);CHKERRQ(ierr); 5566 ierr = (*mat->ops->zerorows)(mat,numRows,newRows,diag,x,b);CHKERRQ(ierr); 5567 ierr = ISRestoreIndices(newis,&newRows);CHKERRQ(ierr); 5568 ierr = ISDestroy(newis);CHKERRQ(ierr); 5569 ierr = ISDestroy(is);CHKERRQ(ierr); 5570 } 5571 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5572 #if defined(PETSC_HAVE_CUSP) 5573 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 5574 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 5575 } 5576 #endif 5577 PetscFunctionReturn(0); 5578 } 5579 5580 #undef __FUNCT__ 5581 #define __FUNCT__ "MatZeroRowsLocalIS" 5582 /*@C 5583 MatZeroRowsLocalIS - Zeros all entries (except possibly the main diagonal) 5584 of a set of rows of a matrix; using local numbering of rows. 5585 5586 Collective on Mat 5587 5588 Input Parameters: 5589 + mat - the matrix 5590 . is - index set of rows to remove 5591 . diag - value put in all diagonals of eliminated rows 5592 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5593 - b - optional vector of right hand side, that will be adjusted by provided solution 5594 5595 Notes: 5596 Before calling MatZeroRowsLocalIS(), the user must first set the 5597 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 5598 5599 For the AIJ matrix formats this removes the old nonzero structure, 5600 but does not release memory. For the dense and block diagonal 5601 formats this does not alter the nonzero structure. 5602 5603 If the option MatSetOption(mat,MAT_KEEP_NONZERO_PATTERN,PETSC_TRUE) the nonzero structure 5604 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 5605 merely zeroed. 5606 5607 The user can set a value in the diagonal entry (or for the AIJ and 5608 row formats can optionally remove the main diagonal entry from the 5609 nonzero structure as well, by passing 0.0 as the final argument). 5610 5611 You can call MatSetOption(mat,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) if each process indicates only rows it 5612 owns that are to be zeroed. This saves a global synchronization in the implementation. 5613 5614 Level: intermediate 5615 5616 Concepts: matrices^zeroing 5617 5618 .seealso: MatZeroRows(), MatZeroRowsLocal(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 5619 @*/ 5620 PetscErrorCode MatZeroRowsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b) 5621 { 5622 PetscErrorCode ierr; 5623 PetscInt numRows; 5624 const PetscInt *rows; 5625 5626 PetscFunctionBegin; 5627 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5628 PetscValidType(mat,1); 5629 PetscValidHeaderSpecific(is,IS_CLASSID,2); 5630 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5631 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5632 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5633 5634 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 5635 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 5636 ierr = MatZeroRowsLocal(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5637 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 5638 PetscFunctionReturn(0); 5639 } 5640 5641 #undef __FUNCT__ 5642 #define __FUNCT__ "MatZeroRowsColumnsLocal" 5643 /*@C 5644 MatZeroRowsColumnsLocal - Zeros all entries (except possibly the main diagonal) 5645 of a set of rows and columns of a matrix; using local numbering of rows. 5646 5647 Collective on Mat 5648 5649 Input Parameters: 5650 + mat - the matrix 5651 . numRows - the number of rows to remove 5652 . rows - the global row indices 5653 . diag - value put in all diagonals of eliminated rows 5654 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5655 - b - optional vector of right hand side, that will be adjusted by provided solution 5656 5657 Notes: 5658 Before calling MatZeroRowsColumnsLocal(), the user must first set the 5659 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 5660 5661 The user can set a value in the diagonal entry (or for the AIJ and 5662 row formats can optionally remove the main diagonal entry from the 5663 nonzero structure as well, by passing 0.0 as the final argument). 5664 5665 Level: intermediate 5666 5667 Concepts: matrices^zeroing 5668 5669 .seealso: MatZeroRows(), MatZeroRowsLocalIS(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 5670 @*/ 5671 PetscErrorCode MatZeroRowsColumnsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 5672 { 5673 PetscErrorCode ierr; 5674 PetscMPIInt size; 5675 5676 PetscFunctionBegin; 5677 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5678 PetscValidType(mat,1); 5679 if (numRows) PetscValidIntPointer(rows,3); 5680 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5681 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5682 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5683 5684 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 5685 if (size == 1) { 5686 ierr = (*mat->ops->zerorowscolumns)(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5687 } else { 5688 IS is, newis; 5689 const PetscInt *newRows; 5690 5691 if (!mat->cmapping) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first"); 5692 ierr = ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,PETSC_COPY_VALUES,&is);CHKERRQ(ierr); 5693 ierr = ISLocalToGlobalMappingApplyIS(mat->cmapping,is,&newis);CHKERRQ(ierr); 5694 ierr = ISGetIndices(newis,&newRows);CHKERRQ(ierr); 5695 ierr = (*mat->ops->zerorowscolumns)(mat,numRows,newRows,diag,x,b);CHKERRQ(ierr); 5696 ierr = ISRestoreIndices(newis,&newRows);CHKERRQ(ierr); 5697 ierr = ISDestroy(newis);CHKERRQ(ierr); 5698 ierr = ISDestroy(is);CHKERRQ(ierr); 5699 } 5700 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5701 #if defined(PETSC_HAVE_CUSP) 5702 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 5703 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 5704 } 5705 #endif 5706 PetscFunctionReturn(0); 5707 } 5708 5709 #undef __FUNCT__ 5710 #define __FUNCT__ "MatZeroRowsColumnsLocalIS" 5711 /*@C 5712 MatZeroRowsColumnsLocalIS - Zeros all entries (except possibly the main diagonal) 5713 of a set of rows and columns of a matrix; using local numbering of rows. 5714 5715 Collective on Mat 5716 5717 Input Parameters: 5718 + mat - the matrix 5719 . is - index set of rows to remove 5720 . diag - value put in all diagonals of eliminated rows 5721 . x - optional vector of solutions for zeroed rows (other entries in vector are not used) 5722 - b - optional vector of right hand side, that will be adjusted by provided solution 5723 5724 Notes: 5725 Before calling MatZeroRowsColumnsLocalIS(), the user must first set the 5726 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 5727 5728 The user can set a value in the diagonal entry (or for the AIJ and 5729 row formats can optionally remove the main diagonal entry from the 5730 nonzero structure as well, by passing 0.0 as the final argument). 5731 5732 Level: intermediate 5733 5734 Concepts: matrices^zeroing 5735 5736 .seealso: MatZeroRows(), MatZeroRowsLocal(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 5737 @*/ 5738 PetscErrorCode MatZeroRowsColumnsLocalIS(Mat mat,IS is,PetscScalar diag,Vec x,Vec b) 5739 { 5740 PetscErrorCode ierr; 5741 PetscInt numRows; 5742 const PetscInt *rows; 5743 5744 PetscFunctionBegin; 5745 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5746 PetscValidType(mat,1); 5747 PetscValidHeaderSpecific(is,IS_CLASSID,2); 5748 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5749 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5750 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5751 5752 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 5753 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 5754 ierr = MatZeroRowsColumnsLocal(mat,numRows,rows,diag,x,b);CHKERRQ(ierr); 5755 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 5756 PetscFunctionReturn(0); 5757 } 5758 5759 #undef __FUNCT__ 5760 #define __FUNCT__ "MatGetSize" 5761 /*@ 5762 MatGetSize - Returns the numbers of rows and columns in a matrix. 5763 5764 Not Collective 5765 5766 Input Parameter: 5767 . mat - the matrix 5768 5769 Output Parameters: 5770 + m - the number of global rows 5771 - n - the number of global columns 5772 5773 Note: both output parameters can be PETSC_NULL on input. 5774 5775 Level: beginner 5776 5777 Concepts: matrices^size 5778 5779 .seealso: MatGetLocalSize() 5780 @*/ 5781 PetscErrorCode MatGetSize(Mat mat,PetscInt *m,PetscInt* n) 5782 { 5783 PetscFunctionBegin; 5784 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5785 if (m) *m = mat->rmap->N; 5786 if (n) *n = mat->cmap->N; 5787 PetscFunctionReturn(0); 5788 } 5789 5790 #undef __FUNCT__ 5791 #define __FUNCT__ "MatGetLocalSize" 5792 /*@ 5793 MatGetLocalSize - Returns the number of rows and columns in a matrix 5794 stored locally. This information may be implementation dependent, so 5795 use with care. 5796 5797 Not Collective 5798 5799 Input Parameters: 5800 . mat - the matrix 5801 5802 Output Parameters: 5803 + m - the number of local rows 5804 - n - the number of local columns 5805 5806 Note: both output parameters can be PETSC_NULL on input. 5807 5808 Level: beginner 5809 5810 Concepts: matrices^local size 5811 5812 .seealso: MatGetSize() 5813 @*/ 5814 PetscErrorCode MatGetLocalSize(Mat mat,PetscInt *m,PetscInt* n) 5815 { 5816 PetscFunctionBegin; 5817 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5818 if (m) PetscValidIntPointer(m,2); 5819 if (n) PetscValidIntPointer(n,3); 5820 if (m) *m = mat->rmap->n; 5821 if (n) *n = mat->cmap->n; 5822 PetscFunctionReturn(0); 5823 } 5824 5825 #undef __FUNCT__ 5826 #define __FUNCT__ "MatGetOwnershipRangeColumn" 5827 /*@ 5828 MatGetOwnershipRangeColumn - Returns the range of matrix columns owned by 5829 this processor. 5830 5831 Not Collective, unless matrix has not been allocated, then collective on Mat 5832 5833 Input Parameters: 5834 . mat - the matrix 5835 5836 Output Parameters: 5837 + m - the global index of the first local column 5838 - n - one more than the global index of the last local column 5839 5840 Notes: both output parameters can be PETSC_NULL on input. 5841 5842 Level: developer 5843 5844 Concepts: matrices^column ownership 5845 5846 .seealso: MatGetOwnershipRange(), MatGetOwnershipRanges(), MatGetOwnershipRangesColumn() 5847 5848 @*/ 5849 PetscErrorCode MatGetOwnershipRangeColumn(Mat mat,PetscInt *m,PetscInt* n) 5850 { 5851 PetscErrorCode ierr; 5852 5853 PetscFunctionBegin; 5854 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5855 PetscValidType(mat,1); 5856 if (m) PetscValidIntPointer(m,2); 5857 if (n) PetscValidIntPointer(n,3); 5858 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5859 if (m) *m = mat->cmap->rstart; 5860 if (n) *n = mat->cmap->rend; 5861 PetscFunctionReturn(0); 5862 } 5863 5864 #undef __FUNCT__ 5865 #define __FUNCT__ "MatGetOwnershipRange" 5866 /*@ 5867 MatGetOwnershipRange - Returns the range of matrix rows owned by 5868 this processor, assuming that the matrix is laid out with the first 5869 n1 rows on the first processor, the next n2 rows on the second, etc. 5870 For certain parallel layouts this range may not be well defined. 5871 5872 Not Collective, unless matrix has not been allocated, then collective on Mat 5873 5874 Input Parameters: 5875 . mat - the matrix 5876 5877 Output Parameters: 5878 + m - the global index of the first local row 5879 - n - one more than the global index of the last local row 5880 5881 Note: both output parameters can be PETSC_NULL on input. 5882 5883 Level: beginner 5884 5885 Concepts: matrices^row ownership 5886 5887 .seealso: MatGetOwnershipRanges(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn() 5888 5889 @*/ 5890 PetscErrorCode MatGetOwnershipRange(Mat mat,PetscInt *m,PetscInt* n) 5891 { 5892 PetscErrorCode ierr; 5893 5894 PetscFunctionBegin; 5895 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5896 PetscValidType(mat,1); 5897 if (m) PetscValidIntPointer(m,2); 5898 if (n) PetscValidIntPointer(n,3); 5899 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5900 if (m) *m = mat->rmap->rstart; 5901 if (n) *n = mat->rmap->rend; 5902 PetscFunctionReturn(0); 5903 } 5904 5905 #undef __FUNCT__ 5906 #define __FUNCT__ "MatGetOwnershipRanges" 5907 /*@C 5908 MatGetOwnershipRanges - Returns the range of matrix rows owned by 5909 each process 5910 5911 Not Collective, unless matrix has not been allocated, then collective on Mat 5912 5913 Input Parameters: 5914 . mat - the matrix 5915 5916 Output Parameters: 5917 . ranges - start of each processors portion plus one more then the total length at the end 5918 5919 Level: beginner 5920 5921 Concepts: matrices^row ownership 5922 5923 .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRangesColumn() 5924 5925 @*/ 5926 PetscErrorCode MatGetOwnershipRanges(Mat mat,const PetscInt **ranges) 5927 { 5928 PetscErrorCode ierr; 5929 5930 PetscFunctionBegin; 5931 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5932 PetscValidType(mat,1); 5933 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5934 ierr = PetscLayoutGetRanges(mat->rmap,ranges);CHKERRQ(ierr); 5935 PetscFunctionReturn(0); 5936 } 5937 5938 #undef __FUNCT__ 5939 #define __FUNCT__ "MatGetOwnershipRangesColumn" 5940 /*@C 5941 MatGetOwnershipRangesColumn - Returns the range of local columns for each process 5942 5943 Not Collective, unless matrix has not been allocated, then collective on Mat 5944 5945 Input Parameters: 5946 . mat - the matrix 5947 5948 Output Parameters: 5949 . ranges - start of each processors portion plus one more then the total length at the end 5950 5951 Level: beginner 5952 5953 Concepts: matrices^column ownership 5954 5955 .seealso: MatGetOwnershipRange(), MatGetOwnershipRangeColumn(), MatGetOwnershipRanges() 5956 5957 @*/ 5958 PetscErrorCode MatGetOwnershipRangesColumn(Mat mat,const PetscInt **ranges) 5959 { 5960 PetscErrorCode ierr; 5961 5962 PetscFunctionBegin; 5963 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 5964 PetscValidType(mat,1); 5965 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5966 ierr = PetscLayoutGetRanges(mat->cmap,ranges);CHKERRQ(ierr); 5967 PetscFunctionReturn(0); 5968 } 5969 5970 #undef __FUNCT__ 5971 #define __FUNCT__ "MatILUFactorSymbolic" 5972 /*@C 5973 MatILUFactorSymbolic - Performs symbolic ILU factorization of a matrix. 5974 Uses levels of fill only, not drop tolerance. Use MatLUFactorNumeric() 5975 to complete the factorization. 5976 5977 Collective on Mat 5978 5979 Input Parameters: 5980 + mat - the matrix 5981 . row - row permutation 5982 . column - column permutation 5983 - info - structure containing 5984 $ levels - number of levels of fill. 5985 $ expected fill - as ratio of original fill. 5986 $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices 5987 missing diagonal entries) 5988 5989 Output Parameters: 5990 . fact - new matrix that has been symbolically factored 5991 5992 Notes: 5993 See the <a href="../../docs/manual.pdf">users manual</a> for additional information about 5994 choosing the fill factor for better efficiency. 5995 5996 Most users should employ the simplified KSP interface for linear solvers 5997 instead of working directly with matrix algebra routines such as this. 5998 See, e.g., KSPCreate(). 5999 6000 Level: developer 6001 6002 Concepts: matrices^symbolic LU factorization 6003 Concepts: matrices^factorization 6004 Concepts: LU^symbolic factorization 6005 6006 .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor() 6007 MatGetOrdering(), MatFactorInfo 6008 6009 Developer Note: fortran interface is not autogenerated as the f90 6010 interface defintion cannot be generated correctly [due to MatFactorInfo] 6011 6012 @*/ 6013 PetscErrorCode MatILUFactorSymbolic(Mat fact,Mat mat,IS row,IS col,const MatFactorInfo *info) 6014 { 6015 PetscErrorCode ierr; 6016 6017 PetscFunctionBegin; 6018 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6019 PetscValidType(mat,1); 6020 PetscValidHeaderSpecific(row,IS_CLASSID,2); 6021 PetscValidHeaderSpecific(col,IS_CLASSID,3); 6022 PetscValidPointer(info,4); 6023 PetscValidPointer(fact,5); 6024 if (info->levels < 0) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Levels of fill negative %D",(PetscInt)info->levels); 6025 if (info->fill < 1.0) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %G",info->fill); 6026 if (!(fact)->ops->ilufactorsymbolic) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Matrix type %s symbolic ILU",((PetscObject)mat)->type_name); 6027 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6028 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6029 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6030 6031 ierr = PetscLogEventBegin(MAT_ILUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 6032 ierr = (fact->ops->ilufactorsymbolic)(fact,mat,row,col,info);CHKERRQ(ierr); 6033 ierr = PetscLogEventEnd(MAT_ILUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 6034 PetscFunctionReturn(0); 6035 } 6036 6037 #undef __FUNCT__ 6038 #define __FUNCT__ "MatICCFactorSymbolic" 6039 /*@C 6040 MatICCFactorSymbolic - Performs symbolic incomplete 6041 Cholesky factorization for a symmetric matrix. Use 6042 MatCholeskyFactorNumeric() to complete the factorization. 6043 6044 Collective on Mat 6045 6046 Input Parameters: 6047 + mat - the matrix 6048 . perm - row and column permutation 6049 - info - structure containing 6050 $ levels - number of levels of fill. 6051 $ expected fill - as ratio of original fill. 6052 6053 Output Parameter: 6054 . fact - the factored matrix 6055 6056 Notes: 6057 Most users should employ the KSP interface for linear solvers 6058 instead of working directly with matrix algebra routines such as this. 6059 See, e.g., KSPCreate(). 6060 6061 Level: developer 6062 6063 Concepts: matrices^symbolic incomplete Cholesky factorization 6064 Concepts: matrices^factorization 6065 Concepts: Cholsky^symbolic factorization 6066 6067 .seealso: MatCholeskyFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 6068 6069 Developer Note: fortran interface is not autogenerated as the f90 6070 interface defintion cannot be generated correctly [due to MatFactorInfo] 6071 6072 @*/ 6073 PetscErrorCode MatICCFactorSymbolic(Mat fact,Mat mat,IS perm,const MatFactorInfo *info) 6074 { 6075 PetscErrorCode ierr; 6076 6077 PetscFunctionBegin; 6078 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6079 PetscValidType(mat,1); 6080 PetscValidHeaderSpecific(perm,IS_CLASSID,2); 6081 PetscValidPointer(info,3); 6082 PetscValidPointer(fact,4); 6083 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6084 if (info->levels < 0) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Levels negative %D",(PetscInt) info->levels); 6085 if (info->fill < 1.0) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %G",info->fill); 6086 if (!(fact)->ops->iccfactorsymbolic) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Matrix type %s symbolic ICC",((PetscObject)mat)->type_name); 6087 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6088 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6089 6090 ierr = PetscLogEventBegin(MAT_ICCFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 6091 ierr = (fact->ops->iccfactorsymbolic)(fact,mat,perm,info);CHKERRQ(ierr); 6092 ierr = PetscLogEventEnd(MAT_ICCFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 6093 PetscFunctionReturn(0); 6094 } 6095 6096 #undef __FUNCT__ 6097 #define __FUNCT__ "MatGetArray" 6098 /*@C 6099 MatGetArray - Returns a pointer to the element values in the matrix. 6100 The result of this routine is dependent on the underlying matrix data 6101 structure, and may not even work for certain matrix types. You MUST 6102 call MatRestoreArray() when you no longer need to access the array. 6103 6104 Not Collective 6105 6106 Input Parameter: 6107 . mat - the matrix 6108 6109 Output Parameter: 6110 . v - the location of the values 6111 6112 6113 Fortran Note: 6114 This routine is used differently from Fortran, e.g., 6115 .vb 6116 Mat mat 6117 PetscScalar mat_array(1) 6118 PetscOffset i_mat 6119 PetscErrorCode ierr 6120 call MatGetArray(mat,mat_array,i_mat,ierr) 6121 6122 C Access first local entry in matrix; note that array is 6123 C treated as one dimensional 6124 value = mat_array(i_mat + 1) 6125 6126 [... other code ...] 6127 call MatRestoreArray(mat,mat_array,i_mat,ierr) 6128 .ve 6129 6130 See the <a href="../../docs/manual.pdf#ch_fortran">Fortran chapter of the users manual</a> and 6131 src/mat/examples/tests for details. 6132 6133 Level: advanced 6134 6135 Concepts: matrices^access array 6136 6137 .seealso: MatRestoreArray(), MatGetArrayF90(), MatGetRowIJ() 6138 @*/ 6139 PetscErrorCode MatGetArray(Mat mat,PetscScalar *v[]) 6140 { 6141 PetscErrorCode ierr; 6142 6143 PetscFunctionBegin; 6144 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6145 PetscValidType(mat,1); 6146 PetscValidPointer(v,2); 6147 if (!mat->ops->getarray) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 6148 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6149 ierr = (*mat->ops->getarray)(mat,v);CHKERRQ(ierr); 6150 CHKMEMQ; 6151 PetscFunctionReturn(0); 6152 } 6153 6154 #undef __FUNCT__ 6155 #define __FUNCT__ "MatRestoreArray" 6156 /*@C 6157 MatRestoreArray - Restores the matrix after MatGetArray() has been called. 6158 6159 Not Collective 6160 6161 Input Parameter: 6162 + mat - the matrix 6163 - v - the location of the values 6164 6165 Fortran Note: 6166 This routine is used differently from Fortran, e.g., 6167 .vb 6168 Mat mat 6169 PetscScalar mat_array(1) 6170 PetscOffset i_mat 6171 PetscErrorCode ierr 6172 call MatGetArray(mat,mat_array,i_mat,ierr) 6173 6174 C Access first local entry in matrix; note that array is 6175 C treated as one dimensional 6176 value = mat_array(i_mat + 1) 6177 6178 [... other code ...] 6179 call MatRestoreArray(mat,mat_array,i_mat,ierr) 6180 .ve 6181 6182 See the <a href="../../docs/manual.pdf#ch_fortran">Fortran chapter of the users manual</a> 6183 src/mat/examples/tests for details 6184 6185 Level: advanced 6186 6187 .seealso: MatGetArray(), MatRestoreArrayF90() 6188 @*/ 6189 PetscErrorCode MatRestoreArray(Mat mat,PetscScalar *v[]) 6190 { 6191 PetscErrorCode ierr; 6192 6193 PetscFunctionBegin; 6194 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6195 PetscValidType(mat,1); 6196 PetscValidPointer(v,2); 6197 #if defined(PETSC_USE_DEBUG) 6198 CHKMEMQ; 6199 #endif 6200 if (!mat->ops->restorearray) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 6201 ierr = (*mat->ops->restorearray)(mat,v);CHKERRQ(ierr); 6202 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 6203 #if defined(PETSC_HAVE_CUSP) 6204 if (mat->valid_GPU_matrix != PETSC_CUSP_UNALLOCATED) { 6205 mat->valid_GPU_matrix = PETSC_CUSP_CPU; 6206 } 6207 #endif 6208 PetscFunctionReturn(0); 6209 } 6210 6211 #undef __FUNCT__ 6212 #define __FUNCT__ "MatGetSubMatrices" 6213 /*@C 6214 MatGetSubMatrices - Extracts several submatrices from a matrix. If submat 6215 points to an array of valid matrices, they may be reused to store the new 6216 submatrices. 6217 6218 Collective on Mat 6219 6220 Input Parameters: 6221 + mat - the matrix 6222 . n - the number of submatrixes to be extracted (on this processor, may be zero) 6223 . irow, icol - index sets of rows and columns to extract 6224 - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 6225 6226 Output Parameter: 6227 . submat - the array of submatrices 6228 6229 Notes: 6230 MatGetSubMatrices() can extract ONLY sequential submatrices 6231 (from both sequential and parallel matrices). Use MatGetSubMatrix() 6232 to extract a parallel submatrix. 6233 6234 When extracting submatrices from a parallel matrix, each processor can 6235 form a different submatrix by setting the rows and columns of its 6236 individual index sets according to the local submatrix desired. 6237 6238 When finished using the submatrices, the user should destroy 6239 them with MatDestroyMatrices(). 6240 6241 MAT_REUSE_MATRIX can only be used when the nonzero structure of the 6242 original matrix has not changed from that last call to MatGetSubMatrices(). 6243 6244 This routine creates the matrices in submat; you should NOT create them before 6245 calling it. It also allocates the array of matrix pointers submat. 6246 6247 For BAIJ matrices the index sets must respect the block structure, that is if they 6248 request one row/column in a block, they must request all rows/columns that are in 6249 that block. For example, if the block size is 2 you cannot request just row 0 and 6250 column 0. 6251 6252 Fortran Note: 6253 The Fortran interface is slightly different from that given below; it 6254 requires one to pass in as submat a Mat (integer) array of size at least m. 6255 6256 Level: advanced 6257 6258 Concepts: matrices^accessing submatrices 6259 Concepts: submatrices 6260 6261 .seealso: MatDestroyMatrices(), MatGetSubMatrix(), MatGetRow(), MatGetDiagonal(), MatReuse 6262 @*/ 6263 PetscErrorCode MatGetSubMatrices(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[]) 6264 { 6265 PetscErrorCode ierr; 6266 PetscInt i; 6267 PetscBool eq; 6268 6269 PetscFunctionBegin; 6270 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6271 PetscValidType(mat,1); 6272 if (n) { 6273 PetscValidPointer(irow,3); 6274 PetscValidHeaderSpecific(*irow,IS_CLASSID,3); 6275 PetscValidPointer(icol,4); 6276 PetscValidHeaderSpecific(*icol,IS_CLASSID,4); 6277 } 6278 PetscValidPointer(submat,6); 6279 if (n && scall == MAT_REUSE_MATRIX) { 6280 PetscValidPointer(*submat,6); 6281 PetscValidHeaderSpecific(**submat,MAT_CLASSID,6); 6282 } 6283 if (!mat->ops->getsubmatrices) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 6284 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6285 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6286 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6287 6288 ierr = PetscLogEventBegin(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 6289 ierr = (*mat->ops->getsubmatrices)(mat,n,irow,icol,scall,submat);CHKERRQ(ierr); 6290 ierr = PetscLogEventEnd(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 6291 for (i=0; i<n; i++) { 6292 if (mat->symmetric || mat->structurally_symmetric || mat->hermitian) { 6293 ierr = ISEqual(irow[i],icol[i],&eq);CHKERRQ(ierr); 6294 if (eq) { 6295 if (mat->symmetric){ 6296 ierr = MatSetOption((*submat)[i],MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 6297 } else if (mat->hermitian) { 6298 ierr = MatSetOption((*submat)[i],MAT_HERMITIAN,PETSC_TRUE);CHKERRQ(ierr); 6299 } else if (mat->structurally_symmetric) { 6300 ierr = MatSetOption((*submat)[i],MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 6301 } 6302 } 6303 } 6304 } 6305 PetscFunctionReturn(0); 6306 } 6307 6308 #undef __FUNCT__ 6309 #define __FUNCT__ "MatGetSubMatricesParallel" 6310 PetscErrorCode MatGetSubMatricesParallel(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[]) 6311 { 6312 PetscErrorCode ierr; 6313 PetscInt i; 6314 PetscBool eq; 6315 6316 PetscFunctionBegin; 6317 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6318 PetscValidType(mat,1); 6319 if (n) { 6320 PetscValidPointer(irow,3); 6321 PetscValidHeaderSpecific(*irow,IS_CLASSID,3); 6322 PetscValidPointer(icol,4); 6323 PetscValidHeaderSpecific(*icol,IS_CLASSID,4); 6324 } 6325 PetscValidPointer(submat,6); 6326 if (n && scall == MAT_REUSE_MATRIX) { 6327 PetscValidPointer(*submat,6); 6328 PetscValidHeaderSpecific(**submat,MAT_CLASSID,6); 6329 } 6330 if (!mat->ops->getsubmatricesparallel) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 6331 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6332 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6333 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6334 6335 ierr = PetscLogEventBegin(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 6336 ierr = (*mat->ops->getsubmatricesparallel)(mat,n,irow,icol,scall,submat);CHKERRQ(ierr); 6337 ierr = PetscLogEventEnd(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 6338 for (i=0; i<n; i++) { 6339 if (mat->symmetric || mat->structurally_symmetric || mat->hermitian) { 6340 ierr = ISEqual(irow[i],icol[i],&eq);CHKERRQ(ierr); 6341 if (eq) { 6342 if (mat->symmetric){ 6343 ierr = MatSetOption((*submat)[i],MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 6344 } else if (mat->hermitian) { 6345 ierr = MatSetOption((*submat)[i],MAT_HERMITIAN,PETSC_TRUE);CHKERRQ(ierr); 6346 } else if (mat->structurally_symmetric) { 6347 ierr = MatSetOption((*submat)[i],MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 6348 } 6349 } 6350 } 6351 } 6352 PetscFunctionReturn(0); 6353 } 6354 6355 #undef __FUNCT__ 6356 #define __FUNCT__ "MatDestroyMatrices" 6357 /*@C 6358 MatDestroyMatrices - Destroys a set of matrices obtained with MatGetSubMatrices(). 6359 6360 Collective on Mat 6361 6362 Input Parameters: 6363 + n - the number of local matrices 6364 - mat - the matrices (note that this is a pointer to the array of matrices, just to match the calling 6365 sequence of MatGetSubMatrices()) 6366 6367 Level: advanced 6368 6369 Notes: Frees not only the matrices, but also the array that contains the matrices 6370 In Fortran will not free the array. 6371 6372 .seealso: MatGetSubMatrices() 6373 @*/ 6374 PetscErrorCode MatDestroyMatrices(PetscInt n,Mat *mat[]) 6375 { 6376 PetscErrorCode ierr; 6377 PetscInt i; 6378 6379 PetscFunctionBegin; 6380 if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n); 6381 PetscValidPointer(mat,2); 6382 for (i=0; i<n; i++) { 6383 ierr = MatDestroy((*mat)[i]);CHKERRQ(ierr); 6384 } 6385 /* memory is allocated even if n = 0 */ 6386 ierr = PetscFree(*mat);CHKERRQ(ierr); 6387 *mat = PETSC_NULL; 6388 PetscFunctionReturn(0); 6389 } 6390 6391 #undef __FUNCT__ 6392 #define __FUNCT__ "MatGetSeqNonzeroStructure" 6393 /*@C 6394 MatGetSeqNonzeroStructure - Extracts the sequential nonzero structure from a matrix. 6395 6396 Collective on Mat 6397 6398 Input Parameters: 6399 . mat - the matrix 6400 6401 Output Parameter: 6402 . matstruct - the sequential matrix with the nonzero structure of mat 6403 6404 Level: intermediate 6405 6406 .seealso: MatDestroySeqNonzeroStructure(), MatGetSubMatrices(), MatDestroyMatrices() 6407 @*/ 6408 PetscErrorCode MatGetSeqNonzeroStructure(Mat mat,Mat *matstruct) 6409 { 6410 PetscErrorCode ierr; 6411 6412 PetscFunctionBegin; 6413 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6414 PetscValidPointer(matstruct,2); 6415 6416 PetscValidType(mat,1); 6417 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6418 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6419 6420 if (!mat->ops->getseqnonzerostructure) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Not for matrix type %s\n",((PetscObject)mat)->type_name); 6421 ierr = PetscLogEventBegin(MAT_GetSeqNonzeroStructure,mat,0,0,0);CHKERRQ(ierr); 6422 ierr = (*mat->ops->getseqnonzerostructure)(mat,matstruct);CHKERRQ(ierr); 6423 ierr = PetscLogEventEnd(MAT_GetSeqNonzeroStructure,mat,0,0,0);CHKERRQ(ierr); 6424 PetscFunctionReturn(0); 6425 } 6426 6427 #undef __FUNCT__ 6428 #define __FUNCT__ "MatDestroySeqNonzeroStructure" 6429 /*@C 6430 MatDestroySeqNonzeroStructure - Destroys matrix obtained with MatGetSeqNonzeroStructure(). 6431 6432 Collective on Mat 6433 6434 Input Parameters: 6435 . mat - the matrix (note that this is a pointer to the array of matrices, just to match the calling 6436 sequence of MatGetSequentialNonzeroStructure()) 6437 6438 Level: advanced 6439 6440 Notes: Frees not only the matrices, but also the array that contains the matrices 6441 6442 .seealso: MatGetSeqNonzeroStructure() 6443 @*/ 6444 PetscErrorCode MatDestroySeqNonzeroStructure(Mat *mat) 6445 { 6446 PetscErrorCode ierr; 6447 6448 PetscFunctionBegin; 6449 PetscValidPointer(mat,1); 6450 ierr = MatDestroy(*mat);CHKERRQ(ierr); 6451 PetscFunctionReturn(0); 6452 } 6453 6454 #undef __FUNCT__ 6455 #define __FUNCT__ "MatIncreaseOverlap" 6456 /*@ 6457 MatIncreaseOverlap - Given a set of submatrices indicated by index sets, 6458 replaces the index sets by larger ones that represent submatrices with 6459 additional overlap. 6460 6461 Collective on Mat 6462 6463 Input Parameters: 6464 + mat - the matrix 6465 . n - the number of index sets 6466 . is - the array of index sets (these index sets will changed during the call) 6467 - ov - the additional overlap requested 6468 6469 Level: developer 6470 6471 Concepts: overlap 6472 Concepts: ASM^computing overlap 6473 6474 .seealso: MatGetSubMatrices() 6475 @*/ 6476 PetscErrorCode MatIncreaseOverlap(Mat mat,PetscInt n,IS is[],PetscInt ov) 6477 { 6478 PetscErrorCode ierr; 6479 6480 PetscFunctionBegin; 6481 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6482 PetscValidType(mat,1); 6483 if (n < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n); 6484 if (n) { 6485 PetscValidPointer(is,3); 6486 PetscValidHeaderSpecific(*is,IS_CLASSID,3); 6487 } 6488 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6489 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6490 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6491 6492 if (!ov) PetscFunctionReturn(0); 6493 if (!mat->ops->increaseoverlap) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 6494 ierr = PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);CHKERRQ(ierr); 6495 ierr = (*mat->ops->increaseoverlap)(mat,n,is,ov);CHKERRQ(ierr); 6496 ierr = PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);CHKERRQ(ierr); 6497 PetscFunctionReturn(0); 6498 } 6499 6500 #undef __FUNCT__ 6501 #define __FUNCT__ "MatGetBlockSize" 6502 /*@ 6503 MatGetBlockSize - Returns the matrix block size; useful especially for the 6504 block row and block diagonal formats. 6505 6506 Not Collective 6507 6508 Input Parameter: 6509 . mat - the matrix 6510 6511 Output Parameter: 6512 . bs - block size 6513 6514 Notes: 6515 Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ 6516 6517 Level: intermediate 6518 6519 Concepts: matrices^block size 6520 6521 .seealso: MatCreateSeqBAIJ(), MatCreateMPIBAIJ() 6522 @*/ 6523 PetscErrorCode MatGetBlockSize(Mat mat,PetscInt *bs) 6524 { 6525 PetscErrorCode ierr; 6526 6527 PetscFunctionBegin; 6528 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6529 PetscValidType(mat,1); 6530 PetscValidIntPointer(bs,2); 6531 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6532 *bs = mat->rmap->bs; 6533 PetscFunctionReturn(0); 6534 } 6535 6536 #undef __FUNCT__ 6537 #define __FUNCT__ "MatSetBlockSize" 6538 /*@ 6539 MatSetBlockSize - Sets the matrix block size; for many matrix types you 6540 cannot use this and MUST set the blocksize when you preallocate the matrix 6541 6542 Logically Collective on Mat 6543 6544 Input Parameters: 6545 + mat - the matrix 6546 - bs - block size 6547 6548 Notes: 6549 For BAIJ matrices, this just checks that the block size agrees with the BAIJ size, 6550 it is not possible to change BAIJ block sizes after preallocation. 6551 6552 Level: intermediate 6553 6554 Concepts: matrices^block size 6555 6556 .seealso: MatCreateSeqBAIJ(), MatCreateMPIBAIJ(), MatGetBlockSize() 6557 @*/ 6558 PetscErrorCode MatSetBlockSize(Mat mat,PetscInt bs) 6559 { 6560 PetscErrorCode ierr; 6561 6562 PetscFunctionBegin; 6563 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6564 PetscValidType(mat,1); 6565 PetscValidLogicalCollectiveInt(mat,bs,2); 6566 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6567 if (bs < 1) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Block size %D, must be positive",bs); 6568 if (mat->ops->setblocksize) { 6569 ierr = (*mat->ops->setblocksize)(mat,bs);CHKERRQ(ierr); 6570 } else { 6571 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Cannot set the blocksize for matrix type %s",((PetscObject)mat)->type_name); 6572 } 6573 PetscFunctionReturn(0); 6574 } 6575 6576 #undef __FUNCT__ 6577 #define __FUNCT__ "MatGetRowIJ" 6578 /*@C 6579 MatGetRowIJ - Returns the compressed row storage i and j indices for sequential matrices. 6580 6581 Collective on Mat 6582 6583 Input Parameters: 6584 + mat - the matrix 6585 . shift - 0 or 1 indicating we want the indices starting at 0 or 1 6586 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 6587 symmetrized 6588 - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 6589 inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is 6590 always used. 6591 6592 Output Parameters: 6593 + n - number of rows in the (possibly compressed) matrix 6594 . ia - the row pointers [of length n+1] 6595 . ja - the column indices 6596 - done - indicates if the routine actually worked and returned appropriate ia[] and ja[] arrays; callers 6597 are responsible for handling the case when done == PETSC_FALSE and ia and ja are not set 6598 6599 Level: developer 6600 6601 Notes: You CANNOT change any of the ia[] or ja[] values. 6602 6603 Use MatRestoreRowIJ() when you are finished accessing the ia[] and ja[] values 6604 6605 Fortran Node 6606 6607 In Fortran use 6608 $ PetscInt ia(1), ja(1) 6609 $ PetscOffset iia, jja 6610 $ call MatGetRowIJ(mat,shift,symmetric,inodecompressed,n,ia,iia,ja,jja,done,ierr) 6611 $ 6612 $ or 6613 $ 6614 $ PetscScalar, pointer :: xx_v(:) 6615 $ call MatGetRowIJF90(mat,shift,symmetric,inodecompressed,n,ia,ja,done,ierr) 6616 6617 6618 Acess the ith and jth entries via ia(iia + i) and ja(jja + j) 6619 6620 .seealso: MatGetColumnIJ(), MatRestoreRowIJ(), MatGetArray() 6621 @*/ 6622 PetscErrorCode MatGetRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscBool *done) 6623 { 6624 PetscErrorCode ierr; 6625 6626 PetscFunctionBegin; 6627 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6628 PetscValidType(mat,1); 6629 PetscValidIntPointer(n,4); 6630 if (ia) PetscValidIntPointer(ia,5); 6631 if (ja) PetscValidIntPointer(ja,6); 6632 PetscValidIntPointer(done,7); 6633 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6634 if (!mat->ops->getrowij) *done = PETSC_FALSE; 6635 else { 6636 *done = PETSC_TRUE; 6637 ierr = PetscLogEventBegin(MAT_GetRowIJ,mat,0,0,0);CHKERRQ(ierr); 6638 ierr = (*mat->ops->getrowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);CHKERRQ(ierr); 6639 ierr = PetscLogEventEnd(MAT_GetRowIJ,mat,0,0,0);CHKERRQ(ierr); 6640 } 6641 PetscFunctionReturn(0); 6642 } 6643 6644 #undef __FUNCT__ 6645 #define __FUNCT__ "MatGetColumnIJ" 6646 /*@C 6647 MatGetColumnIJ - Returns the compressed column storage i and j indices for sequential matrices. 6648 6649 Collective on Mat 6650 6651 Input Parameters: 6652 + mat - the matrix 6653 . shift - 1 or zero indicating we want the indices starting at 0 or 1 6654 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 6655 symmetrized 6656 - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 6657 inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is 6658 always used. 6659 6660 Output Parameters: 6661 + n - number of columns in the (possibly compressed) matrix 6662 . ia - the column pointers 6663 . ja - the row indices 6664 - done - PETSC_TRUE or PETSC_FALSE, indicating whether the values have been returned 6665 6666 Level: developer 6667 6668 .seealso: MatGetRowIJ(), MatRestoreColumnIJ() 6669 @*/ 6670 PetscErrorCode MatGetColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscBool *done) 6671 { 6672 PetscErrorCode ierr; 6673 6674 PetscFunctionBegin; 6675 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6676 PetscValidType(mat,1); 6677 PetscValidIntPointer(n,4); 6678 if (ia) PetscValidIntPointer(ia,5); 6679 if (ja) PetscValidIntPointer(ja,6); 6680 PetscValidIntPointer(done,7); 6681 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6682 if (!mat->ops->getcolumnij) *done = PETSC_FALSE; 6683 else { 6684 *done = PETSC_TRUE; 6685 ierr = (*mat->ops->getcolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);CHKERRQ(ierr); 6686 } 6687 PetscFunctionReturn(0); 6688 } 6689 6690 #undef __FUNCT__ 6691 #define __FUNCT__ "MatRestoreRowIJ" 6692 /*@C 6693 MatRestoreRowIJ - Call after you are completed with the ia,ja indices obtained with 6694 MatGetRowIJ(). 6695 6696 Collective on Mat 6697 6698 Input Parameters: 6699 + mat - the matrix 6700 . shift - 1 or zero indicating we want the indices starting at 0 or 1 6701 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 6702 symmetrized 6703 - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 6704 inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is 6705 always used. 6706 6707 Output Parameters: 6708 + n - size of (possibly compressed) matrix 6709 . ia - the row pointers 6710 . ja - the column indices 6711 - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned 6712 6713 Level: developer 6714 6715 .seealso: MatGetRowIJ(), MatRestoreColumnIJ() 6716 @*/ 6717 PetscErrorCode MatRestoreRowIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscBool *done) 6718 { 6719 PetscErrorCode ierr; 6720 6721 PetscFunctionBegin; 6722 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6723 PetscValidType(mat,1); 6724 if (ia) PetscValidIntPointer(ia,5); 6725 if (ja) PetscValidIntPointer(ja,6); 6726 PetscValidIntPointer(done,7); 6727 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6728 6729 if (!mat->ops->restorerowij) *done = PETSC_FALSE; 6730 else { 6731 *done = PETSC_TRUE; 6732 ierr = (*mat->ops->restorerowij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);CHKERRQ(ierr); 6733 } 6734 PetscFunctionReturn(0); 6735 } 6736 6737 #undef __FUNCT__ 6738 #define __FUNCT__ "MatRestoreColumnIJ" 6739 /*@C 6740 MatRestoreColumnIJ - Call after you are completed with the ia,ja indices obtained with 6741 MatGetColumnIJ(). 6742 6743 Collective on Mat 6744 6745 Input Parameters: 6746 + mat - the matrix 6747 . shift - 1 or zero indicating we want the indices starting at 0 or 1 6748 - symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 6749 symmetrized 6750 - inodecompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 6751 inodes or the nonzero elements is wanted. For BAIJ matrices the compressed version is 6752 always used. 6753 6754 Output Parameters: 6755 + n - size of (possibly compressed) matrix 6756 . ia - the column pointers 6757 . ja - the row indices 6758 - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned 6759 6760 Level: developer 6761 6762 .seealso: MatGetColumnIJ(), MatRestoreRowIJ() 6763 @*/ 6764 PetscErrorCode MatRestoreColumnIJ(Mat mat,PetscInt shift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscBool *done) 6765 { 6766 PetscErrorCode ierr; 6767 6768 PetscFunctionBegin; 6769 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6770 PetscValidType(mat,1); 6771 if (ia) PetscValidIntPointer(ia,5); 6772 if (ja) PetscValidIntPointer(ja,6); 6773 PetscValidIntPointer(done,7); 6774 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6775 6776 if (!mat->ops->restorecolumnij) *done = PETSC_FALSE; 6777 else { 6778 *done = PETSC_TRUE; 6779 ierr = (*mat->ops->restorecolumnij)(mat,shift,symmetric,inodecompressed,n,ia,ja,done);CHKERRQ(ierr); 6780 } 6781 PetscFunctionReturn(0); 6782 } 6783 6784 #undef __FUNCT__ 6785 #define __FUNCT__ "MatColoringPatch" 6786 /*@C 6787 MatColoringPatch -Used inside matrix coloring routines that 6788 use MatGetRowIJ() and/or MatGetColumnIJ(). 6789 6790 Collective on Mat 6791 6792 Input Parameters: 6793 + mat - the matrix 6794 . ncolors - max color value 6795 . n - number of entries in colorarray 6796 - colorarray - array indicating color for each column 6797 6798 Output Parameters: 6799 . iscoloring - coloring generated using colorarray information 6800 6801 Level: developer 6802 6803 .seealso: MatGetRowIJ(), MatGetColumnIJ() 6804 6805 @*/ 6806 PetscErrorCode MatColoringPatch(Mat mat,PetscInt ncolors,PetscInt n,ISColoringValue colorarray[],ISColoring *iscoloring) 6807 { 6808 PetscErrorCode ierr; 6809 6810 PetscFunctionBegin; 6811 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6812 PetscValidType(mat,1); 6813 PetscValidIntPointer(colorarray,4); 6814 PetscValidPointer(iscoloring,5); 6815 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6816 6817 if (!mat->ops->coloringpatch){ 6818 ierr = ISColoringCreate(((PetscObject)mat)->comm,ncolors,n,colorarray,iscoloring);CHKERRQ(ierr); 6819 } else { 6820 ierr = (*mat->ops->coloringpatch)(mat,ncolors,n,colorarray,iscoloring);CHKERRQ(ierr); 6821 } 6822 PetscFunctionReturn(0); 6823 } 6824 6825 6826 #undef __FUNCT__ 6827 #define __FUNCT__ "MatSetUnfactored" 6828 /*@ 6829 MatSetUnfactored - Resets a factored matrix to be treated as unfactored. 6830 6831 Logically Collective on Mat 6832 6833 Input Parameter: 6834 . mat - the factored matrix to be reset 6835 6836 Notes: 6837 This routine should be used only with factored matrices formed by in-place 6838 factorization via ILU(0) (or by in-place LU factorization for the MATSEQDENSE 6839 format). This option can save memory, for example, when solving nonlinear 6840 systems with a matrix-free Newton-Krylov method and a matrix-based, in-place 6841 ILU(0) preconditioner. 6842 6843 Note that one can specify in-place ILU(0) factorization by calling 6844 .vb 6845 PCType(pc,PCILU); 6846 PCFactorSeUseInPlace(pc); 6847 .ve 6848 or by using the options -pc_type ilu -pc_factor_in_place 6849 6850 In-place factorization ILU(0) can also be used as a local 6851 solver for the blocks within the block Jacobi or additive Schwarz 6852 methods (runtime option: -sub_pc_factor_in_place). See the discussion 6853 of these preconditioners in the <a href="../../docs/manual.pdf#ch_pc">PC chapter of the users manual</a> for details on setting 6854 local solver options. 6855 6856 Most users should employ the simplified KSP interface for linear solvers 6857 instead of working directly with matrix algebra routines such as this. 6858 See, e.g., KSPCreate(). 6859 6860 Level: developer 6861 6862 .seealso: PCFactorSetUseInPlace() 6863 6864 Concepts: matrices^unfactored 6865 6866 @*/ 6867 PetscErrorCode MatSetUnfactored(Mat mat) 6868 { 6869 PetscErrorCode ierr; 6870 6871 PetscFunctionBegin; 6872 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 6873 PetscValidType(mat,1); 6874 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6875 mat->factortype = MAT_FACTOR_NONE; 6876 if (!mat->ops->setunfactored) PetscFunctionReturn(0); 6877 ierr = (*mat->ops->setunfactored)(mat);CHKERRQ(ierr); 6878 PetscFunctionReturn(0); 6879 } 6880 6881 /*MC 6882 MatGetArrayF90 - Accesses a matrix array from Fortran90. 6883 6884 Synopsis: 6885 MatGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr) 6886 6887 Not collective 6888 6889 Input Parameter: 6890 . x - matrix 6891 6892 Output Parameters: 6893 + xx_v - the Fortran90 pointer to the array 6894 - ierr - error code 6895 6896 Example of Usage: 6897 .vb 6898 PetscScalar, pointer xx_v(:) 6899 .... 6900 call MatGetArrayF90(x,xx_v,ierr) 6901 a = xx_v(3) 6902 call MatRestoreArrayF90(x,xx_v,ierr) 6903 .ve 6904 6905 Notes: 6906 Not yet supported for all F90 compilers 6907 6908 Level: advanced 6909 6910 .seealso: MatRestoreArrayF90(), MatGetArray(), MatRestoreArray() 6911 6912 Concepts: matrices^accessing array 6913 6914 M*/ 6915 6916 /*MC 6917 MatRestoreArrayF90 - Restores a matrix array that has been 6918 accessed with MatGetArrayF90(). 6919 6920 Synopsis: 6921 MatRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr) 6922 6923 Not collective 6924 6925 Input Parameters: 6926 + x - matrix 6927 - xx_v - the Fortran90 pointer to the array 6928 6929 Output Parameter: 6930 . ierr - error code 6931 6932 Example of Usage: 6933 .vb 6934 PetscScalar, pointer xx_v(:) 6935 .... 6936 call MatGetArrayF90(x,xx_v,ierr) 6937 a = xx_v(3) 6938 call MatRestoreArrayF90(x,xx_v,ierr) 6939 .ve 6940 6941 Notes: 6942 Not yet supported for all F90 compilers 6943 6944 Level: advanced 6945 6946 .seealso: MatGetArrayF90(), MatGetArray(), MatRestoreArray() 6947 6948 M*/ 6949 6950 6951 #undef __FUNCT__ 6952 #define __FUNCT__ "MatGetSubMatrix" 6953 /*@ 6954 MatGetSubMatrix - Gets a single submatrix on the same number of processors 6955 as the original matrix. 6956 6957 Collective on Mat 6958 6959 Input Parameters: 6960 + mat - the original matrix 6961 . isrow - parallel IS containing the rows this processor should obtain 6962 . 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. 6963 - cll - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 6964 6965 Output Parameter: 6966 . newmat - the new submatrix, of the same type as the old 6967 6968 Level: advanced 6969 6970 Notes: 6971 The submatrix will be able to be multiplied with vectors using the same layout as iscol. 6972 6973 The rows in isrow will be sorted into the same order as the original matrix on each process. 6974 6975 The first time this is called you should use a cll of MAT_INITIAL_MATRIX, 6976 the MatGetSubMatrix() routine will create the newmat for you. Any additional calls 6977 to this routine with a mat of the same nonzero structure and with a call of MAT_REUSE_MATRIX 6978 will reuse the matrix generated the first time. You should call MatDestroy() on newmat when 6979 you are finished using it. 6980 6981 The communicator of the newly obtained matrix is ALWAYS the same as the communicator of 6982 the input matrix. 6983 6984 If iscol is PETSC_NULL then all columns are obtained (not supported in Fortran). 6985 6986 Example usage: 6987 Consider the following 8x8 matrix with 34 non-zero values, that is 6988 assembled across 3 processors. Let's assume that proc0 owns 3 rows, 6989 proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown 6990 as follows: 6991 6992 .vb 6993 1 2 0 | 0 3 0 | 0 4 6994 Proc0 0 5 6 | 7 0 0 | 8 0 6995 9 0 10 | 11 0 0 | 12 0 6996 ------------------------------------- 6997 13 0 14 | 15 16 17 | 0 0 6998 Proc1 0 18 0 | 19 20 21 | 0 0 6999 0 0 0 | 22 23 0 | 24 0 7000 ------------------------------------- 7001 Proc2 25 26 27 | 0 0 28 | 29 0 7002 30 0 0 | 31 32 33 | 0 34 7003 .ve 7004 7005 Suppose isrow = [0 1 | 4 | 6 7] and iscol = [1 2 | 3 4 5 | 6]. The resulting submatrix is 7006 7007 .vb 7008 2 0 | 0 3 0 | 0 7009 Proc0 5 6 | 7 0 0 | 8 7010 ------------------------------- 7011 Proc1 18 0 | 19 20 21 | 0 7012 ------------------------------- 7013 Proc2 26 27 | 0 0 28 | 29 7014 0 0 | 31 32 33 | 0 7015 .ve 7016 7017 7018 Concepts: matrices^submatrices 7019 7020 .seealso: MatGetSubMatrices() 7021 @*/ 7022 PetscErrorCode MatGetSubMatrix(Mat mat,IS isrow,IS iscol,MatReuse cll,Mat *newmat) 7023 { 7024 PetscErrorCode ierr; 7025 PetscMPIInt size; 7026 Mat *local; 7027 IS iscoltmp; 7028 7029 PetscFunctionBegin; 7030 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7031 PetscValidHeaderSpecific(isrow,IS_CLASSID,2); 7032 if (iscol) PetscValidHeaderSpecific(iscol,IS_CLASSID,3); 7033 PetscValidPointer(newmat,5); 7034 if (cll == MAT_REUSE_MATRIX) PetscValidHeaderSpecific(*newmat,MAT_CLASSID,5); 7035 PetscValidType(mat,1); 7036 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7037 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7038 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 7039 7040 if (!iscol) { 7041 ierr = ISCreateStride(((PetscObject)mat)->comm,mat->cmap->n,mat->cmap->rstart,1,&iscoltmp);CHKERRQ(ierr); 7042 } else { 7043 iscoltmp = iscol; 7044 } 7045 7046 /* if original matrix is on just one processor then use submatrix generated */ 7047 if (mat->ops->getsubmatrices && !mat->ops->getsubmatrix && size == 1 && cll == MAT_REUSE_MATRIX) { 7048 ierr = MatGetSubMatrices(mat,1,&isrow,&iscoltmp,MAT_REUSE_MATRIX,&newmat);CHKERRQ(ierr); 7049 if (!iscol) {ierr = ISDestroy(iscoltmp);CHKERRQ(ierr);} 7050 PetscFunctionReturn(0); 7051 } else if (mat->ops->getsubmatrices && !mat->ops->getsubmatrix && size == 1) { 7052 ierr = MatGetSubMatrices(mat,1,&isrow,&iscoltmp,MAT_INITIAL_MATRIX,&local);CHKERRQ(ierr); 7053 *newmat = *local; 7054 ierr = PetscFree(local);CHKERRQ(ierr); 7055 if (!iscol) {ierr = ISDestroy(iscoltmp);CHKERRQ(ierr);} 7056 PetscFunctionReturn(0); 7057 } else if (!mat->ops->getsubmatrix) { 7058 /* Create a new matrix type that implements the operation using the full matrix */ 7059 switch (cll) { 7060 case MAT_INITIAL_MATRIX: 7061 ierr = MatCreateSubMatrix(mat,isrow,iscoltmp,newmat);CHKERRQ(ierr); 7062 break; 7063 case MAT_REUSE_MATRIX: 7064 ierr = MatSubMatrixUpdate(*newmat,mat,isrow,iscoltmp);CHKERRQ(ierr); 7065 break; 7066 default: SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Invalid MatReuse, must be either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX"); 7067 } 7068 if (!iscol) {ierr = ISDestroy(iscoltmp);CHKERRQ(ierr);} 7069 PetscFunctionReturn(0); 7070 } 7071 7072 if (!mat->ops->getsubmatrix) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7073 ierr = (*mat->ops->getsubmatrix)(mat,isrow,iscoltmp,cll,newmat);CHKERRQ(ierr); 7074 if (!iscol) {ierr = ISDestroy(iscoltmp);CHKERRQ(ierr);} 7075 ierr = PetscObjectStateIncrease((PetscObject)*newmat);CHKERRQ(ierr); 7076 PetscFunctionReturn(0); 7077 } 7078 7079 #undef __FUNCT__ 7080 #define __FUNCT__ "MatStashSetInitialSize" 7081 /*@ 7082 MatStashSetInitialSize - sets the sizes of the matrix stash, that is 7083 used during the assembly process to store values that belong to 7084 other processors. 7085 7086 Not Collective 7087 7088 Input Parameters: 7089 + mat - the matrix 7090 . size - the initial size of the stash. 7091 - bsize - the initial size of the block-stash(if used). 7092 7093 Options Database Keys: 7094 + -matstash_initial_size <size> or <size0,size1,...sizep-1> 7095 - -matstash_block_initial_size <bsize> or <bsize0,bsize1,...bsizep-1> 7096 7097 Level: intermediate 7098 7099 Notes: 7100 The block-stash is used for values set with MatSetValuesBlocked() while 7101 the stash is used for values set with MatSetValues() 7102 7103 Run with the option -info and look for output of the form 7104 MatAssemblyBegin_MPIXXX:Stash has MM entries, uses nn mallocs. 7105 to determine the appropriate value, MM, to use for size and 7106 MatAssemblyBegin_MPIXXX:Block-Stash has BMM entries, uses nn mallocs. 7107 to determine the value, BMM to use for bsize 7108 7109 Concepts: stash^setting matrix size 7110 Concepts: matrices^stash 7111 7112 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashGetInfo() 7113 7114 @*/ 7115 PetscErrorCode MatStashSetInitialSize(Mat mat,PetscInt size, PetscInt bsize) 7116 { 7117 PetscErrorCode ierr; 7118 7119 PetscFunctionBegin; 7120 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7121 PetscValidType(mat,1); 7122 ierr = MatStashSetInitialSize_Private(&mat->stash,size);CHKERRQ(ierr); 7123 ierr = MatStashSetInitialSize_Private(&mat->bstash,bsize);CHKERRQ(ierr); 7124 PetscFunctionReturn(0); 7125 } 7126 7127 #undef __FUNCT__ 7128 #define __FUNCT__ "MatInterpolateAdd" 7129 /*@ 7130 MatInterpolateAdd - w = y + A*x or A'*x depending on the shape of 7131 the matrix 7132 7133 Neighbor-wise Collective on Mat 7134 7135 Input Parameters: 7136 + mat - the matrix 7137 . x,y - the vectors 7138 - w - where the result is stored 7139 7140 Level: intermediate 7141 7142 Notes: 7143 w may be the same vector as y. 7144 7145 This allows one to use either the restriction or interpolation (its transpose) 7146 matrix to do the interpolation 7147 7148 Concepts: interpolation 7149 7150 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict() 7151 7152 @*/ 7153 PetscErrorCode MatInterpolateAdd(Mat A,Vec x,Vec y,Vec w) 7154 { 7155 PetscErrorCode ierr; 7156 PetscInt M,N,Ny; 7157 7158 PetscFunctionBegin; 7159 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7160 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 7161 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 7162 PetscValidHeaderSpecific(w,VEC_CLASSID,4); 7163 PetscValidType(A,1); 7164 ierr = MatPreallocated(A);CHKERRQ(ierr); 7165 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 7166 ierr = VecGetSize(y,&Ny);CHKERRQ(ierr); 7167 if (M == Ny) { 7168 ierr = MatMultAdd(A,x,y,w);CHKERRQ(ierr); 7169 } else { 7170 ierr = MatMultTransposeAdd(A,x,y,w);CHKERRQ(ierr); 7171 } 7172 PetscFunctionReturn(0); 7173 } 7174 7175 #undef __FUNCT__ 7176 #define __FUNCT__ "MatInterpolate" 7177 /*@ 7178 MatInterpolate - y = A*x or A'*x depending on the shape of 7179 the matrix 7180 7181 Neighbor-wise Collective on Mat 7182 7183 Input Parameters: 7184 + mat - the matrix 7185 - x,y - the vectors 7186 7187 Level: intermediate 7188 7189 Notes: 7190 This allows one to use either the restriction or interpolation (its transpose) 7191 matrix to do the interpolation 7192 7193 Concepts: matrices^interpolation 7194 7195 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict() 7196 7197 @*/ 7198 PetscErrorCode MatInterpolate(Mat A,Vec x,Vec y) 7199 { 7200 PetscErrorCode ierr; 7201 PetscInt M,N,Ny; 7202 7203 PetscFunctionBegin; 7204 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7205 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 7206 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 7207 PetscValidType(A,1); 7208 ierr = MatPreallocated(A);CHKERRQ(ierr); 7209 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 7210 ierr = VecGetSize(y,&Ny);CHKERRQ(ierr); 7211 if (M == Ny) { 7212 ierr = MatMult(A,x,y);CHKERRQ(ierr); 7213 } else { 7214 ierr = MatMultTranspose(A,x,y);CHKERRQ(ierr); 7215 } 7216 PetscFunctionReturn(0); 7217 } 7218 7219 #undef __FUNCT__ 7220 #define __FUNCT__ "MatRestrict" 7221 /*@ 7222 MatRestrict - y = A*x or A'*x 7223 7224 Neighbor-wise Collective on Mat 7225 7226 Input Parameters: 7227 + mat - the matrix 7228 - x,y - the vectors 7229 7230 Level: intermediate 7231 7232 Notes: 7233 This allows one to use either the restriction or interpolation (its transpose) 7234 matrix to do the restriction 7235 7236 Concepts: matrices^restriction 7237 7238 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatInterpolate() 7239 7240 @*/ 7241 PetscErrorCode MatRestrict(Mat A,Vec x,Vec y) 7242 { 7243 PetscErrorCode ierr; 7244 PetscInt M,N,Ny; 7245 7246 PetscFunctionBegin; 7247 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7248 PetscValidHeaderSpecific(x,VEC_CLASSID,2); 7249 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 7250 PetscValidType(A,1); 7251 ierr = MatPreallocated(A);CHKERRQ(ierr); 7252 7253 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 7254 ierr = VecGetSize(y,&Ny);CHKERRQ(ierr); 7255 if (M == Ny) { 7256 ierr = MatMult(A,x,y);CHKERRQ(ierr); 7257 } else { 7258 ierr = MatMultTranspose(A,x,y);CHKERRQ(ierr); 7259 } 7260 PetscFunctionReturn(0); 7261 } 7262 7263 #undef __FUNCT__ 7264 #define __FUNCT__ "MatNullSpaceAttach" 7265 /*@ 7266 MatNullSpaceAttach - attaches a null space to a matrix. 7267 This null space will be removed from the resulting vector whenever 7268 MatMult() is called 7269 7270 Logically Collective on Mat and MatNullSpace 7271 7272 Input Parameters: 7273 + mat - the matrix 7274 - nullsp - the null space object 7275 7276 Level: developer 7277 7278 Notes: 7279 Overwrites any previous null space that may have been attached 7280 7281 Concepts: null space^attaching to matrix 7282 7283 .seealso: MatCreate(), MatNullSpaceCreate() 7284 @*/ 7285 PetscErrorCode MatNullSpaceAttach(Mat mat,MatNullSpace nullsp) 7286 { 7287 PetscErrorCode ierr; 7288 7289 PetscFunctionBegin; 7290 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7291 PetscValidType(mat,1); 7292 PetscValidHeaderSpecific(nullsp,MAT_NULLSPACE_CLASSID,2); 7293 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7294 ierr = PetscObjectReference((PetscObject)nullsp);CHKERRQ(ierr); 7295 if (mat->nullsp) { ierr = MatNullSpaceDestroy(mat->nullsp);CHKERRQ(ierr); } 7296 mat->nullsp = nullsp; 7297 PetscFunctionReturn(0); 7298 } 7299 7300 #undef __FUNCT__ 7301 #define __FUNCT__ "MatICCFactor" 7302 /*@C 7303 MatICCFactor - Performs in-place incomplete Cholesky factorization of matrix. 7304 7305 Collective on Mat 7306 7307 Input Parameters: 7308 + mat - the matrix 7309 . row - row/column permutation 7310 . fill - expected fill factor >= 1.0 7311 - level - level of fill, for ICC(k) 7312 7313 Notes: 7314 Probably really in-place only when level of fill is zero, otherwise allocates 7315 new space to store factored matrix and deletes previous memory. 7316 7317 Most users should employ the simplified KSP interface for linear solvers 7318 instead of working directly with matrix algebra routines such as this. 7319 See, e.g., KSPCreate(). 7320 7321 Level: developer 7322 7323 Concepts: matrices^incomplete Cholesky factorization 7324 Concepts: Cholesky factorization 7325 7326 .seealso: MatICCFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor() 7327 7328 Developer Note: fortran interface is not autogenerated as the f90 7329 interface defintion cannot be generated correctly [due to MatFactorInfo] 7330 7331 @*/ 7332 PetscErrorCode MatICCFactor(Mat mat,IS row,const MatFactorInfo* info) 7333 { 7334 PetscErrorCode ierr; 7335 7336 PetscFunctionBegin; 7337 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7338 PetscValidType(mat,1); 7339 if (row) PetscValidHeaderSpecific(row,IS_CLASSID,2); 7340 PetscValidPointer(info,3); 7341 if (mat->rmap->N != mat->cmap->N) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONG,"matrix must be square"); 7342 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7343 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7344 if (!mat->ops->iccfactor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7345 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7346 ierr = (*mat->ops->iccfactor)(mat,row,info);CHKERRQ(ierr); 7347 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 7348 PetscFunctionReturn(0); 7349 } 7350 7351 #undef __FUNCT__ 7352 #define __FUNCT__ "MatSetValuesAdic" 7353 /*@ 7354 MatSetValuesAdic - Sets values computed with ADIC automatic differentiation into a matrix. 7355 7356 Not Collective 7357 7358 Input Parameters: 7359 + mat - the matrix 7360 - v - the values compute with ADIC 7361 7362 Level: developer 7363 7364 Notes: 7365 Must call MatSetColoring() before using this routine. Also this matrix must already 7366 have its nonzero pattern determined. 7367 7368 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 7369 MatSetValues(), MatSetColoring(), MatSetValuesAdifor() 7370 @*/ 7371 PetscErrorCode MatSetValuesAdic(Mat mat,void *v) 7372 { 7373 PetscErrorCode ierr; 7374 7375 PetscFunctionBegin; 7376 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7377 PetscValidType(mat,1); 7378 PetscValidPointer(mat,2); 7379 7380 if (!mat->assembled) { 7381 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 7382 } 7383 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 7384 if (!mat->ops->setvaluesadic) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7385 ierr = (*mat->ops->setvaluesadic)(mat,v);CHKERRQ(ierr); 7386 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 7387 ierr = MatView_Private(mat);CHKERRQ(ierr); 7388 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 7389 PetscFunctionReturn(0); 7390 } 7391 7392 7393 #undef __FUNCT__ 7394 #define __FUNCT__ "MatSetColoring" 7395 /*@ 7396 MatSetColoring - Sets a coloring used by calls to MatSetValuesAdic() 7397 7398 Not Collective 7399 7400 Input Parameters: 7401 + mat - the matrix 7402 - coloring - the coloring 7403 7404 Level: developer 7405 7406 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 7407 MatSetValues(), MatSetValuesAdic() 7408 @*/ 7409 PetscErrorCode MatSetColoring(Mat mat,ISColoring coloring) 7410 { 7411 PetscErrorCode ierr; 7412 7413 PetscFunctionBegin; 7414 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7415 PetscValidType(mat,1); 7416 PetscValidPointer(coloring,2); 7417 7418 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 7419 if (!mat->ops->setcoloring) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7420 ierr = (*mat->ops->setcoloring)(mat,coloring);CHKERRQ(ierr); 7421 PetscFunctionReturn(0); 7422 } 7423 7424 #undef __FUNCT__ 7425 #define __FUNCT__ "MatSetValuesAdifor" 7426 /*@ 7427 MatSetValuesAdifor - Sets values computed with automatic differentiation into a matrix. 7428 7429 Not Collective 7430 7431 Input Parameters: 7432 + mat - the matrix 7433 . nl - leading dimension of v 7434 - v - the values compute with ADIFOR 7435 7436 Level: developer 7437 7438 Notes: 7439 Must call MatSetColoring() before using this routine. Also this matrix must already 7440 have its nonzero pattern determined. 7441 7442 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 7443 MatSetValues(), MatSetColoring() 7444 @*/ 7445 PetscErrorCode MatSetValuesAdifor(Mat mat,PetscInt nl,void *v) 7446 { 7447 PetscErrorCode ierr; 7448 7449 PetscFunctionBegin; 7450 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7451 PetscValidType(mat,1); 7452 PetscValidPointer(v,3); 7453 7454 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 7455 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 7456 if (!mat->ops->setvaluesadifor) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7457 ierr = (*mat->ops->setvaluesadifor)(mat,nl,v);CHKERRQ(ierr); 7458 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 7459 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 7460 PetscFunctionReturn(0); 7461 } 7462 7463 #undef __FUNCT__ 7464 #define __FUNCT__ "MatDiagonalScaleLocal" 7465 /*@ 7466 MatDiagonalScaleLocal - Scales columns of a matrix given the scaling values including the 7467 ghosted ones. 7468 7469 Not Collective 7470 7471 Input Parameters: 7472 + mat - the matrix 7473 - diag = the diagonal values, including ghost ones 7474 7475 Level: developer 7476 7477 Notes: Works only for MPIAIJ and MPIBAIJ matrices 7478 7479 .seealso: MatDiagonalScale() 7480 @*/ 7481 PetscErrorCode MatDiagonalScaleLocal(Mat mat,Vec diag) 7482 { 7483 PetscErrorCode ierr; 7484 PetscMPIInt size; 7485 7486 PetscFunctionBegin; 7487 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7488 PetscValidHeaderSpecific(diag,VEC_CLASSID,2); 7489 PetscValidType(mat,1); 7490 7491 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 7492 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 7493 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 7494 if (size == 1) { 7495 PetscInt n,m; 7496 ierr = VecGetSize(diag,&n);CHKERRQ(ierr); 7497 ierr = MatGetSize(mat,0,&m);CHKERRQ(ierr); 7498 if (m == n) { 7499 ierr = MatDiagonalScale(mat,0,diag);CHKERRQ(ierr); 7500 } else { 7501 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only supported for sequential matrices when no ghost points/periodic conditions"); 7502 } 7503 } else { 7504 ierr = PetscUseMethod(mat,"MatDiagonalScaleLocal_C",(Mat,Vec),(mat,diag));CHKERRQ(ierr); 7505 } 7506 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 7507 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 7508 PetscFunctionReturn(0); 7509 } 7510 7511 #undef __FUNCT__ 7512 #define __FUNCT__ "MatGetInertia" 7513 /*@ 7514 MatGetInertia - Gets the inertia from a factored matrix 7515 7516 Collective on Mat 7517 7518 Input Parameter: 7519 . mat - the matrix 7520 7521 Output Parameters: 7522 + nneg - number of negative eigenvalues 7523 . nzero - number of zero eigenvalues 7524 - npos - number of positive eigenvalues 7525 7526 Level: advanced 7527 7528 Notes: Matrix must have been factored by MatCholeskyFactor() 7529 7530 7531 @*/ 7532 PetscErrorCode MatGetInertia(Mat mat,PetscInt *nneg,PetscInt *nzero,PetscInt *npos) 7533 { 7534 PetscErrorCode ierr; 7535 7536 PetscFunctionBegin; 7537 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7538 PetscValidType(mat,1); 7539 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 7540 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Numeric factor mat is not assembled"); 7541 if (!mat->ops->getinertia) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7542 ierr = (*mat->ops->getinertia)(mat,nneg,nzero,npos);CHKERRQ(ierr); 7543 PetscFunctionReturn(0); 7544 } 7545 7546 /* ----------------------------------------------------------------*/ 7547 #undef __FUNCT__ 7548 #define __FUNCT__ "MatSolves" 7549 /*@C 7550 MatSolves - Solves A x = b, given a factored matrix, for a collection of vectors 7551 7552 Neighbor-wise Collective on Mat and Vecs 7553 7554 Input Parameters: 7555 + mat - the factored matrix 7556 - b - the right-hand-side vectors 7557 7558 Output Parameter: 7559 . x - the result vectors 7560 7561 Notes: 7562 The vectors b and x cannot be the same. I.e., one cannot 7563 call MatSolves(A,x,x). 7564 7565 Notes: 7566 Most users should employ the simplified KSP interface for linear solvers 7567 instead of working directly with matrix algebra routines such as this. 7568 See, e.g., KSPCreate(). 7569 7570 Level: developer 7571 7572 Concepts: matrices^triangular solves 7573 7574 .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd(), MatSolve() 7575 @*/ 7576 PetscErrorCode MatSolves(Mat mat,Vecs b,Vecs x) 7577 { 7578 PetscErrorCode ierr; 7579 7580 PetscFunctionBegin; 7581 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7582 PetscValidType(mat,1); 7583 if (x == b) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 7584 if (!mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 7585 if (!mat->rmap->N && !mat->cmap->N) PetscFunctionReturn(0); 7586 7587 if (!mat->ops->solves) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 7588 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7589 ierr = PetscLogEventBegin(MAT_Solves,mat,0,0,0);CHKERRQ(ierr); 7590 ierr = (*mat->ops->solves)(mat,b,x);CHKERRQ(ierr); 7591 ierr = PetscLogEventEnd(MAT_Solves,mat,0,0,0);CHKERRQ(ierr); 7592 PetscFunctionReturn(0); 7593 } 7594 7595 #undef __FUNCT__ 7596 #define __FUNCT__ "MatIsSymmetric" 7597 /*@ 7598 MatIsSymmetric - Test whether a matrix is symmetric 7599 7600 Collective on Mat 7601 7602 Input Parameter: 7603 + A - the matrix to test 7604 - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact transpose) 7605 7606 Output Parameters: 7607 . flg - the result 7608 7609 Level: intermediate 7610 7611 Concepts: matrix^symmetry 7612 7613 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown() 7614 @*/ 7615 PetscErrorCode MatIsSymmetric(Mat A,PetscReal tol,PetscBool *flg) 7616 { 7617 PetscErrorCode ierr; 7618 7619 PetscFunctionBegin; 7620 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7621 PetscValidPointer(flg,2); 7622 7623 if (!A->symmetric_set) { 7624 if (!A->ops->issymmetric) { 7625 const MatType mattype; 7626 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 7627 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for symmetric",mattype); 7628 } 7629 ierr = (*A->ops->issymmetric)(A,tol,flg);CHKERRQ(ierr); 7630 if (!tol) { 7631 A->symmetric_set = PETSC_TRUE; 7632 A->symmetric = *flg; 7633 if (A->symmetric) { 7634 A->structurally_symmetric_set = PETSC_TRUE; 7635 A->structurally_symmetric = PETSC_TRUE; 7636 } 7637 } 7638 } else if (A->symmetric) { 7639 *flg = PETSC_TRUE; 7640 } else if (!tol) { 7641 *flg = PETSC_FALSE; 7642 } else { 7643 if (!A->ops->issymmetric) { 7644 const MatType mattype; 7645 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 7646 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for symmetric",mattype); 7647 } 7648 ierr = (*A->ops->issymmetric)(A,tol,flg);CHKERRQ(ierr); 7649 } 7650 PetscFunctionReturn(0); 7651 } 7652 7653 #undef __FUNCT__ 7654 #define __FUNCT__ "MatIsHermitian" 7655 /*@ 7656 MatIsHermitian - Test whether a matrix is Hermitian 7657 7658 Collective on Mat 7659 7660 Input Parameter: 7661 + A - the matrix to test 7662 - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact Hermitian) 7663 7664 Output Parameters: 7665 . flg - the result 7666 7667 Level: intermediate 7668 7669 Concepts: matrix^symmetry 7670 7671 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown() 7672 @*/ 7673 PetscErrorCode MatIsHermitian(Mat A,PetscReal tol,PetscBool *flg) 7674 { 7675 PetscErrorCode ierr; 7676 7677 PetscFunctionBegin; 7678 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7679 PetscValidPointer(flg,2); 7680 7681 if (!A->hermitian_set) { 7682 if (!A->ops->ishermitian) { 7683 const MatType mattype; 7684 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 7685 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for hermitian",mattype); 7686 } 7687 ierr = (*A->ops->ishermitian)(A,tol,flg);CHKERRQ(ierr); 7688 if (!tol) { 7689 A->hermitian_set = PETSC_TRUE; 7690 A->hermitian = *flg; 7691 if (A->hermitian) { 7692 A->structurally_symmetric_set = PETSC_TRUE; 7693 A->structurally_symmetric = PETSC_TRUE; 7694 } 7695 } 7696 } else if (A->hermitian) { 7697 *flg = PETSC_TRUE; 7698 } else if (!tol) { 7699 *flg = PETSC_FALSE; 7700 } else { 7701 if (!A->ops->ishermitian) { 7702 const MatType mattype; 7703 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 7704 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for hermitian",mattype); 7705 } 7706 ierr = (*A->ops->ishermitian)(A,tol,flg);CHKERRQ(ierr); 7707 } 7708 PetscFunctionReturn(0); 7709 } 7710 7711 #undef __FUNCT__ 7712 #define __FUNCT__ "MatIsSymmetricKnown" 7713 /*@ 7714 MatIsSymmetricKnown - Checks the flag on the matrix to see if it is symmetric. 7715 7716 Not Collective 7717 7718 Input Parameter: 7719 . A - the matrix to check 7720 7721 Output Parameters: 7722 + set - if the symmetric flag is set (this tells you if the next flag is valid) 7723 - flg - the result 7724 7725 Level: advanced 7726 7727 Concepts: matrix^symmetry 7728 7729 Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsSymmetric() 7730 if you want it explicitly checked 7731 7732 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric() 7733 @*/ 7734 PetscErrorCode MatIsSymmetricKnown(Mat A,PetscBool *set,PetscBool *flg) 7735 { 7736 PetscFunctionBegin; 7737 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7738 PetscValidPointer(set,2); 7739 PetscValidPointer(flg,3); 7740 if (A->symmetric_set) { 7741 *set = PETSC_TRUE; 7742 *flg = A->symmetric; 7743 } else { 7744 *set = PETSC_FALSE; 7745 } 7746 PetscFunctionReturn(0); 7747 } 7748 7749 #undef __FUNCT__ 7750 #define __FUNCT__ "MatIsHermitianKnown" 7751 /*@ 7752 MatIsHermitianKnown - Checks the flag on the matrix to see if it is hermitian. 7753 7754 Not Collective 7755 7756 Input Parameter: 7757 . A - the matrix to check 7758 7759 Output Parameters: 7760 + set - if the hermitian flag is set (this tells you if the next flag is valid) 7761 - flg - the result 7762 7763 Level: advanced 7764 7765 Concepts: matrix^symmetry 7766 7767 Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsHermitian() 7768 if you want it explicitly checked 7769 7770 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric() 7771 @*/ 7772 PetscErrorCode MatIsHermitianKnown(Mat A,PetscBool *set,PetscBool *flg) 7773 { 7774 PetscFunctionBegin; 7775 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7776 PetscValidPointer(set,2); 7777 PetscValidPointer(flg,3); 7778 if (A->hermitian_set) { 7779 *set = PETSC_TRUE; 7780 *flg = A->hermitian; 7781 } else { 7782 *set = PETSC_FALSE; 7783 } 7784 PetscFunctionReturn(0); 7785 } 7786 7787 #undef __FUNCT__ 7788 #define __FUNCT__ "MatIsStructurallySymmetric" 7789 /*@ 7790 MatIsStructurallySymmetric - Test whether a matrix is structurally symmetric 7791 7792 Collective on Mat 7793 7794 Input Parameter: 7795 . A - the matrix to test 7796 7797 Output Parameters: 7798 . flg - the result 7799 7800 Level: intermediate 7801 7802 Concepts: matrix^symmetry 7803 7804 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsSymmetric(), MatSetOption() 7805 @*/ 7806 PetscErrorCode MatIsStructurallySymmetric(Mat A,PetscBool *flg) 7807 { 7808 PetscErrorCode ierr; 7809 7810 PetscFunctionBegin; 7811 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7812 PetscValidPointer(flg,2); 7813 if (!A->structurally_symmetric_set) { 7814 if (!A->ops->isstructurallysymmetric) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_SUP,"Matrix does not support checking for structural symmetric"); 7815 ierr = (*A->ops->isstructurallysymmetric)(A,&A->structurally_symmetric);CHKERRQ(ierr); 7816 A->structurally_symmetric_set = PETSC_TRUE; 7817 } 7818 *flg = A->structurally_symmetric; 7819 PetscFunctionReturn(0); 7820 } 7821 7822 #undef __FUNCT__ 7823 #define __FUNCT__ "MatStashGetInfo" 7824 extern PetscErrorCode MatStashGetInfo_Private(MatStash*,PetscInt*,PetscInt*); 7825 /*@ 7826 MatStashGetInfo - Gets how many values are currently in the vector stash, i.e. need 7827 to be communicated to other processors during the MatAssemblyBegin/End() process 7828 7829 Not collective 7830 7831 Input Parameter: 7832 . vec - the vector 7833 7834 Output Parameters: 7835 + nstash - the size of the stash 7836 . reallocs - the number of additional mallocs incurred. 7837 . bnstash - the size of the block stash 7838 - breallocs - the number of additional mallocs incurred.in the block stash 7839 7840 Level: advanced 7841 7842 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashSetInitialSize() 7843 7844 @*/ 7845 PetscErrorCode MatStashGetInfo(Mat mat,PetscInt *nstash,PetscInt *reallocs,PetscInt *bnstash,PetscInt *breallocs) 7846 { 7847 PetscErrorCode ierr; 7848 PetscFunctionBegin; 7849 ierr = MatStashGetInfo_Private(&mat->stash,nstash,reallocs);CHKERRQ(ierr); 7850 ierr = MatStashGetInfo_Private(&mat->bstash,bnstash,breallocs);CHKERRQ(ierr); 7851 PetscFunctionReturn(0); 7852 } 7853 7854 #undef __FUNCT__ 7855 #define __FUNCT__ "MatGetVecs" 7856 /*@C 7857 MatGetVecs - Get vector(s) compatible with the matrix, i.e. with the same 7858 parallel layout 7859 7860 Collective on Mat 7861 7862 Input Parameter: 7863 . mat - the matrix 7864 7865 Output Parameter: 7866 + right - (optional) vector that the matrix can be multiplied against 7867 - left - (optional) vector that the matrix vector product can be stored in 7868 7869 Level: advanced 7870 7871 .seealso: MatCreate() 7872 @*/ 7873 PetscErrorCode MatGetVecs(Mat mat,Vec *right,Vec *left) 7874 { 7875 PetscErrorCode ierr; 7876 7877 PetscFunctionBegin; 7878 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 7879 PetscValidType(mat,1); 7880 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7881 if (mat->ops->getvecs) { 7882 ierr = (*mat->ops->getvecs)(mat,right,left);CHKERRQ(ierr); 7883 } else { 7884 PetscMPIInt size; 7885 ierr = MPI_Comm_size(((PetscObject)mat)->comm, &size);CHKERRQ(ierr); 7886 if (right) { 7887 ierr = VecCreate(((PetscObject)mat)->comm,right);CHKERRQ(ierr); 7888 ierr = VecSetSizes(*right,mat->cmap->n,PETSC_DETERMINE);CHKERRQ(ierr); 7889 ierr = VecSetBlockSize(*right,mat->rmap->bs);CHKERRQ(ierr); 7890 if (size > 1) { 7891 /* New vectors uses Mat cmap and does not create a new one */ 7892 ierr = PetscLayoutDestroy((*right)->map);CHKERRQ(ierr); 7893 (*right)->map = mat->cmap; 7894 mat->cmap->refcnt++; 7895 7896 ierr = VecSetType(*right,VECMPI);CHKERRQ(ierr); 7897 } else {ierr = VecSetType(*right,VECSEQ);CHKERRQ(ierr);} 7898 } 7899 if (left) { 7900 ierr = VecCreate(((PetscObject)mat)->comm,left);CHKERRQ(ierr); 7901 ierr = VecSetSizes(*left,mat->rmap->n,PETSC_DETERMINE);CHKERRQ(ierr); 7902 ierr = VecSetBlockSize(*left,mat->rmap->bs);CHKERRQ(ierr); 7903 if (size > 1) { 7904 /* New vectors uses Mat rmap and does not create a new one */ 7905 ierr = PetscLayoutDestroy((*left)->map);CHKERRQ(ierr); 7906 (*left)->map = mat->rmap; 7907 mat->rmap->refcnt++; 7908 7909 ierr = VecSetType(*left,VECMPI);CHKERRQ(ierr); 7910 } else {ierr = VecSetType(*left,VECSEQ);CHKERRQ(ierr);} 7911 } 7912 } 7913 if (mat->rmapping) { 7914 if (right) {ierr = VecSetLocalToGlobalMapping(*right,mat->cmapping);CHKERRQ(ierr);} 7915 if (left) {ierr = VecSetLocalToGlobalMapping(*left,mat->rmapping);CHKERRQ(ierr);} 7916 } 7917 if (mat->rbmapping) { 7918 if (right) {ierr = VecSetLocalToGlobalMappingBlock(*right,mat->cbmapping);CHKERRQ(ierr);} 7919 if (left) {ierr = VecSetLocalToGlobalMappingBlock(*left,mat->rbmapping);CHKERRQ(ierr);} 7920 } 7921 PetscFunctionReturn(0); 7922 } 7923 7924 #undef __FUNCT__ 7925 #define __FUNCT__ "MatFactorInfoInitialize" 7926 /*@C 7927 MatFactorInfoInitialize - Initializes a MatFactorInfo data structure 7928 with default values. 7929 7930 Not Collective 7931 7932 Input Parameters: 7933 . info - the MatFactorInfo data structure 7934 7935 7936 Notes: The solvers are generally used through the KSP and PC objects, for example 7937 PCLU, PCILU, PCCHOLESKY, PCICC 7938 7939 Level: developer 7940 7941 .seealso: MatFactorInfo 7942 7943 Developer Note: fortran interface is not autogenerated as the f90 7944 interface defintion cannot be generated correctly [due to MatFactorInfo] 7945 7946 @*/ 7947 7948 PetscErrorCode MatFactorInfoInitialize(MatFactorInfo *info) 7949 { 7950 PetscErrorCode ierr; 7951 7952 PetscFunctionBegin; 7953 ierr = PetscMemzero(info,sizeof(MatFactorInfo));CHKERRQ(ierr); 7954 PetscFunctionReturn(0); 7955 } 7956 7957 #undef __FUNCT__ 7958 #define __FUNCT__ "MatPtAP" 7959 /*@ 7960 MatPtAP - Creates the matrix product C = P^T * A * P 7961 7962 Neighbor-wise Collective on Mat 7963 7964 Input Parameters: 7965 + A - the matrix 7966 . P - the projection matrix 7967 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 7968 - fill - expected fill as ratio of nnz(C)/nnz(A) 7969 7970 Output Parameters: 7971 . C - the product matrix 7972 7973 Notes: 7974 C will be created and must be destroyed by the user with MatDestroy(). 7975 7976 This routine is currently only implemented for pairs of AIJ matrices and classes 7977 which inherit from AIJ. 7978 7979 Level: intermediate 7980 7981 .seealso: MatPtAPSymbolic(), MatPtAPNumeric(), MatMatMult() 7982 @*/ 7983 PetscErrorCode MatPtAP(Mat A,Mat P,MatReuse scall,PetscReal fill,Mat *C) 7984 { 7985 PetscErrorCode ierr; 7986 7987 PetscFunctionBegin; 7988 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 7989 PetscValidType(A,1); 7990 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7991 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7992 PetscValidHeaderSpecific(P,MAT_CLASSID,2); 7993 PetscValidType(P,2); 7994 ierr = MatPreallocated(P);CHKERRQ(ierr); 7995 if (!P->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7996 if (P->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7997 PetscValidPointer(C,3); 7998 if (P->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N); 7999 if (fill < 1.0) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 8000 ierr = MatPreallocated(A);CHKERRQ(ierr); 8001 8002 if (!A->ops->ptap) { 8003 const MatType mattype; 8004 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 8005 SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"Matrix of type <%s> does not support PtAP",mattype); 8006 } 8007 ierr = PetscLogEventBegin(MAT_PtAP,A,P,0,0);CHKERRQ(ierr); 8008 ierr = (*A->ops->ptap)(A,P,scall,fill,C);CHKERRQ(ierr); 8009 ierr = PetscLogEventEnd(MAT_PtAP,A,P,0,0);CHKERRQ(ierr); 8010 8011 PetscFunctionReturn(0); 8012 } 8013 8014 #undef __FUNCT__ 8015 #define __FUNCT__ "MatPtAPNumeric" 8016 /*@ 8017 MatPtAPNumeric - Computes the matrix product C = P^T * A * P 8018 8019 Neighbor-wise Collective on Mat 8020 8021 Input Parameters: 8022 + A - the matrix 8023 - P - the projection matrix 8024 8025 Output Parameters: 8026 . C - the product matrix 8027 8028 Notes: 8029 C must have been created by calling MatPtAPSymbolic and must be destroyed by 8030 the user using MatDeatroy(). 8031 8032 This routine is currently only implemented for pairs of AIJ matrices and classes 8033 which inherit from AIJ. C will be of type MATAIJ. 8034 8035 Level: intermediate 8036 8037 .seealso: MatPtAP(), MatPtAPSymbolic(), MatMatMultNumeric() 8038 @*/ 8039 PetscErrorCode MatPtAPNumeric(Mat A,Mat P,Mat C) 8040 { 8041 PetscErrorCode ierr; 8042 8043 PetscFunctionBegin; 8044 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8045 PetscValidType(A,1); 8046 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8047 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8048 PetscValidHeaderSpecific(P,MAT_CLASSID,2); 8049 PetscValidType(P,2); 8050 ierr = MatPreallocated(P);CHKERRQ(ierr); 8051 if (!P->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8052 if (P->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8053 PetscValidHeaderSpecific(C,MAT_CLASSID,3); 8054 PetscValidType(C,3); 8055 ierr = MatPreallocated(C);CHKERRQ(ierr); 8056 if (C->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8057 if (P->cmap->N!=C->rmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap->N,C->rmap->N); 8058 if (P->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N); 8059 if (A->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap->N,A->cmap->N); 8060 if (P->cmap->N!=C->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap->N,C->cmap->N); 8061 ierr = MatPreallocated(A);CHKERRQ(ierr); 8062 8063 ierr = PetscLogEventBegin(MAT_PtAPNumeric,A,P,0,0);CHKERRQ(ierr); 8064 ierr = (*A->ops->ptapnumeric)(A,P,C);CHKERRQ(ierr); 8065 ierr = PetscLogEventEnd(MAT_PtAPNumeric,A,P,0,0);CHKERRQ(ierr); 8066 PetscFunctionReturn(0); 8067 } 8068 8069 #undef __FUNCT__ 8070 #define __FUNCT__ "MatPtAPSymbolic" 8071 /*@ 8072 MatPtAPSymbolic - Creates the (i,j) structure of the matrix product C = P^T * A * P 8073 8074 Neighbor-wise Collective on Mat 8075 8076 Input Parameters: 8077 + A - the matrix 8078 - P - the projection matrix 8079 8080 Output Parameters: 8081 . C - the (i,j) structure of the product matrix 8082 8083 Notes: 8084 C will be created and must be destroyed by the user with MatDestroy(). 8085 8086 This routine is currently only implemented for pairs of SeqAIJ matrices and classes 8087 which inherit from SeqAIJ. C will be of type MATSEQAIJ. The product is computed using 8088 this (i,j) structure by calling MatPtAPNumeric(). 8089 8090 Level: intermediate 8091 8092 .seealso: MatPtAP(), MatPtAPNumeric(), MatMatMultSymbolic() 8093 @*/ 8094 PetscErrorCode MatPtAPSymbolic(Mat A,Mat P,PetscReal fill,Mat *C) 8095 { 8096 PetscErrorCode ierr; 8097 8098 PetscFunctionBegin; 8099 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8100 PetscValidType(A,1); 8101 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8102 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8103 if (fill <1.0) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 8104 PetscValidHeaderSpecific(P,MAT_CLASSID,2); 8105 PetscValidType(P,2); 8106 ierr = MatPreallocated(P);CHKERRQ(ierr); 8107 if (!P->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8108 if (P->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8109 PetscValidPointer(C,3); 8110 8111 if (P->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap->N,A->cmap->N); 8112 if (A->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap->N,A->cmap->N); 8113 ierr = MatPreallocated(A);CHKERRQ(ierr); 8114 ierr = PetscLogEventBegin(MAT_PtAPSymbolic,A,P,0,0);CHKERRQ(ierr); 8115 ierr = (*A->ops->ptapsymbolic)(A,P,fill,C);CHKERRQ(ierr); 8116 ierr = PetscLogEventEnd(MAT_PtAPSymbolic,A,P,0,0);CHKERRQ(ierr); 8117 8118 ierr = MatSetBlockSize(*C,A->rmap->bs);CHKERRQ(ierr); 8119 8120 PetscFunctionReturn(0); 8121 } 8122 8123 #undef __FUNCT__ 8124 #define __FUNCT__ "MatMatMult" 8125 /*@ 8126 MatMatMult - Performs Matrix-Matrix Multiplication C=A*B. 8127 8128 Neighbor-wise Collective on Mat 8129 8130 Input Parameters: 8131 + A - the left matrix 8132 . B - the right matrix 8133 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 8134 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)), use PETSC_DEFAULT if you do not have a good estimate 8135 if the result is a dense matrix this is irrelevent 8136 8137 Output Parameters: 8138 . C - the product matrix 8139 8140 Notes: 8141 Unless scall is MAT_REUSE_MATRIX C will be created. 8142 8143 MAT_REUSE_MATRIX can only be used if the matrices A and B have the same nonzero pattern as in the previous call 8144 8145 To determine the correct fill value, run with -info and search for the string "Fill ratio" to see the value 8146 actually needed. 8147 8148 If you have many matrices with the same non-zero structure to multiply, you 8149 should either 8150 $ 1) use MAT_REUSE_MATRIX in all calls but the first or 8151 $ 2) call MatMatMultSymbolic() once and then MatMatMultNumeric() for each product needed 8152 8153 Level: intermediate 8154 8155 .seealso: MatMatMultSymbolic(), MatMatMultNumeric(), MatPtAP() 8156 @*/ 8157 PetscErrorCode MatMatMult(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 8158 { 8159 PetscErrorCode ierr; 8160 PetscErrorCode (*fA)(Mat,Mat,MatReuse,PetscReal,Mat*); 8161 PetscErrorCode (*fB)(Mat,Mat,MatReuse,PetscReal,Mat*); 8162 PetscErrorCode (*mult)(Mat,Mat,MatReuse,PetscReal,Mat *)=PETSC_NULL; 8163 8164 PetscFunctionBegin; 8165 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8166 PetscValidType(A,1); 8167 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8168 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8169 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 8170 PetscValidType(B,2); 8171 ierr = MatPreallocated(B);CHKERRQ(ierr); 8172 if (!B->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8173 if (B->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8174 PetscValidPointer(C,3); 8175 if (B->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap->N,A->cmap->N); 8176 if (scall == MAT_REUSE_MATRIX){ 8177 PetscValidPointer(*C,5); 8178 PetscValidHeaderSpecific(*C,MAT_CLASSID,5); 8179 } 8180 if (fill == PETSC_DEFAULT || fill == PETSC_DECIDE) fill = 2.0; 8181 if (fill < 1.0) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 8182 ierr = MatPreallocated(A);CHKERRQ(ierr); 8183 8184 fA = A->ops->matmult; 8185 fB = B->ops->matmult; 8186 if (fB == fA) { 8187 if (!fB) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"MatMatMult not supported for B of type %s",((PetscObject)B)->type_name); 8188 mult = fB; 8189 } else { 8190 /* dispatch based on the type of A and B */ 8191 char multname[256]; 8192 ierr = PetscStrcpy(multname,"MatMatMult_");CHKERRQ(ierr); 8193 ierr = PetscStrcat(multname,((PetscObject)A)->type_name);CHKERRQ(ierr); 8194 ierr = PetscStrcat(multname,"_");CHKERRQ(ierr); 8195 ierr = PetscStrcat(multname,((PetscObject)B)->type_name);CHKERRQ(ierr); 8196 ierr = PetscStrcat(multname,"_C");CHKERRQ(ierr); /* e.g., multname = "MatMatMult_seqdense_seqaij_C" */ 8197 ierr = PetscObjectQueryFunction((PetscObject)B,multname,(void (**)(void))&mult);CHKERRQ(ierr); 8198 if (!mult) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_INCOMP,"MatMatMult requires A, %s, to be compatible with B, %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name); 8199 } 8200 ierr = PetscLogEventBegin(MAT_MatMult,A,B,0,0);CHKERRQ(ierr); 8201 ierr = (*mult)(A,B,scall,fill,C);CHKERRQ(ierr); 8202 ierr = PetscLogEventEnd(MAT_MatMult,A,B,0,0);CHKERRQ(ierr); 8203 PetscFunctionReturn(0); 8204 } 8205 8206 #undef __FUNCT__ 8207 #define __FUNCT__ "MatMatMultSymbolic" 8208 /*@ 8209 MatMatMultSymbolic - Performs construction, preallocation, and computes the ij structure 8210 of the matrix-matrix product C=A*B. Call this routine before calling MatMatMultNumeric(). 8211 8212 Neighbor-wise Collective on Mat 8213 8214 Input Parameters: 8215 + A - the left matrix 8216 . B - the right matrix 8217 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)), use PETSC_DEFAULT if you do not have a good estimate, 8218 if C is a dense matrix this is irrelevent 8219 8220 Output Parameters: 8221 . C - the product matrix 8222 8223 Notes: 8224 Unless scall is MAT_REUSE_MATRIX C will be created. 8225 8226 To determine the correct fill value, run with -info and search for the string "Fill ratio" to see the value 8227 actually needed. 8228 8229 This routine is currently implemented for 8230 - pairs of AIJ matrices and classes which inherit from AIJ, C will be of type AIJ 8231 - pairs of AIJ (A) and Dense (B) matrix, C will be of type Dense. 8232 - pairs of Dense (A) and AIJ (B) matrix, C will be of type Dense. 8233 8234 Level: intermediate 8235 8236 Developers Note: There are ways to estimate the number of nonzeros in the resulting product, see for example, http://arxiv.org/abs/1006.4173 8237 We should incorporate them into PETSc. 8238 8239 .seealso: MatMatMult(), MatMatMultNumeric() 8240 @*/ 8241 PetscErrorCode MatMatMultSymbolic(Mat A,Mat B,PetscReal fill,Mat *C) 8242 { 8243 PetscErrorCode ierr; 8244 PetscErrorCode (*Asymbolic)(Mat,Mat,PetscReal,Mat *); 8245 PetscErrorCode (*Bsymbolic)(Mat,Mat,PetscReal,Mat *); 8246 PetscErrorCode (*symbolic)(Mat,Mat,PetscReal,Mat *)=PETSC_NULL; 8247 8248 PetscFunctionBegin; 8249 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8250 PetscValidType(A,1); 8251 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8252 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8253 8254 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 8255 PetscValidType(B,2); 8256 ierr = MatPreallocated(B);CHKERRQ(ierr); 8257 if (!B->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8258 if (B->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8259 PetscValidPointer(C,3); 8260 8261 if (B->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap->N,A->cmap->N); 8262 if (fill == PETSC_DEFAULT) fill = 2.0; 8263 if (fill < 1.0) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Expected fill=%G must be > 1.0",fill); 8264 ierr = MatPreallocated(A);CHKERRQ(ierr); 8265 8266 Asymbolic = A->ops->matmultsymbolic; 8267 Bsymbolic = B->ops->matmultsymbolic; 8268 if (Asymbolic == Bsymbolic){ 8269 if (!Bsymbolic) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"C=A*B not implemented for B of type %s",((PetscObject)B)->type_name); 8270 symbolic = Bsymbolic; 8271 } else { /* dispatch based on the type of A and B */ 8272 char symbolicname[256]; 8273 ierr = PetscStrcpy(symbolicname,"MatMatMultSymbolic_");CHKERRQ(ierr); 8274 ierr = PetscStrcat(symbolicname,((PetscObject)A)->type_name);CHKERRQ(ierr); 8275 ierr = PetscStrcat(symbolicname,"_");CHKERRQ(ierr); 8276 ierr = PetscStrcat(symbolicname,((PetscObject)B)->type_name);CHKERRQ(ierr); 8277 ierr = PetscStrcat(symbolicname,"_C");CHKERRQ(ierr); 8278 ierr = PetscObjectQueryFunction((PetscObject)B,symbolicname,(void (**)(void))&symbolic);CHKERRQ(ierr); 8279 if (!symbolic) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_INCOMP,"MatMatMultSymbolic requires A, %s, to be compatible with B, %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name); 8280 } 8281 ierr = PetscLogEventBegin(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 8282 ierr = (*symbolic)(A,B,fill,C);CHKERRQ(ierr); 8283 ierr = PetscLogEventEnd(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 8284 PetscFunctionReturn(0); 8285 } 8286 8287 #undef __FUNCT__ 8288 #define __FUNCT__ "MatMatMultNumeric" 8289 /*@ 8290 MatMatMultNumeric - Performs the numeric matrix-matrix product. 8291 Call this routine after first calling MatMatMultSymbolic(). 8292 8293 Neighbor-wise Collective on Mat 8294 8295 Input Parameters: 8296 + A - the left matrix 8297 - B - the right matrix 8298 8299 Output Parameters: 8300 . C - the product matrix, which was created by from MatMatMultSymbolic() or a call to MatMatMult(). 8301 8302 Notes: 8303 C must have been created with MatMatMultSymbolic(). 8304 8305 This routine is currently implemented for 8306 - pairs of AIJ matrices and classes which inherit from AIJ, C will be of type MATAIJ. 8307 - pairs of AIJ (A) and Dense (B) matrix, C will be of type Dense. 8308 - pairs of Dense (A) and AIJ (B) matrix, C will be of type Dense. 8309 8310 Level: intermediate 8311 8312 .seealso: MatMatMult(), MatMatMultSymbolic() 8313 @*/ 8314 PetscErrorCode MatMatMultNumeric(Mat A,Mat B,Mat C) 8315 { 8316 PetscErrorCode ierr; 8317 PetscErrorCode (*Anumeric)(Mat,Mat,Mat); 8318 PetscErrorCode (*Bnumeric)(Mat,Mat,Mat); 8319 PetscErrorCode (*numeric)(Mat,Mat,Mat)=PETSC_NULL; 8320 8321 PetscFunctionBegin; 8322 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8323 PetscValidType(A,1); 8324 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8325 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8326 8327 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 8328 PetscValidType(B,2); 8329 ierr = MatPreallocated(B);CHKERRQ(ierr); 8330 if (!B->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8331 if (B->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8332 8333 PetscValidHeaderSpecific(C,MAT_CLASSID,3); 8334 PetscValidType(C,3); 8335 ierr = MatPreallocated(C);CHKERRQ(ierr); 8336 if (!C->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8337 if (C->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8338 8339 if (B->cmap->N!=C->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->cmap->N,C->cmap->N); 8340 if (B->rmap->N!=A->cmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap->N,A->cmap->N); 8341 if (A->rmap->N!=C->rmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",A->rmap->N,C->rmap->N); 8342 ierr = MatPreallocated(A);CHKERRQ(ierr); 8343 8344 Anumeric = A->ops->matmultnumeric; 8345 Bnumeric = B->ops->matmultnumeric; 8346 if (Anumeric == Bnumeric){ 8347 if (!Bnumeric) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"MatMatMultNumeric not supported for B of type %s",((PetscObject)B)->type_name); 8348 numeric = Bnumeric; 8349 } else { 8350 char numericname[256]; 8351 ierr = PetscStrcpy(numericname,"MatMatMultNumeric_");CHKERRQ(ierr); 8352 ierr = PetscStrcat(numericname,((PetscObject)A)->type_name);CHKERRQ(ierr); 8353 ierr = PetscStrcat(numericname,"_");CHKERRQ(ierr); 8354 ierr = PetscStrcat(numericname,((PetscObject)B)->type_name);CHKERRQ(ierr); 8355 ierr = PetscStrcat(numericname,"_C");CHKERRQ(ierr); 8356 ierr = PetscObjectQueryFunction((PetscObject)B,numericname,(void (**)(void))&numeric);CHKERRQ(ierr); 8357 if (!numeric) 8358 SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_INCOMP,"MatMatMultNumeric requires A, %s, to be compatible with B, %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name); 8359 } 8360 ierr = PetscLogEventBegin(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 8361 ierr = (*numeric)(A,B,C);CHKERRQ(ierr); 8362 ierr = PetscLogEventEnd(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 8363 PetscFunctionReturn(0); 8364 } 8365 8366 #undef __FUNCT__ 8367 #define __FUNCT__ "MatMatMultTranspose" 8368 /*@ 8369 MatMatMultTranspose - Performs Matrix-Matrix Multiplication C=A^T*B. 8370 8371 Neighbor-wise Collective on Mat 8372 8373 Input Parameters: 8374 + A - the left matrix 8375 . B - the right matrix 8376 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 8377 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)), use PETSC_DEFAULT if not known 8378 8379 Output Parameters: 8380 . C - the product matrix 8381 8382 Notes: 8383 C will be created if MAT_INITIAL_MATRIX and must be destroyed by the user with MatDestroy(). 8384 8385 MAT_REUSE_MATRIX can only be used if the matrices A and B have the same nonzero pattern as in the previous call 8386 8387 To determine the correct fill value, run with -info and search for the string "Fill ratio" to see the value 8388 actually needed. 8389 8390 This routine is currently only implemented for pairs of SeqAIJ matrices and pairs of SeqDense matrices and classes 8391 which inherit from SeqAIJ. C will be of type MATSEQAIJ. 8392 8393 Level: intermediate 8394 8395 .seealso: MatMatMultTransposeSymbolic(), MatMatMultTransposeNumeric(), MatPtAP() 8396 @*/ 8397 PetscErrorCode MatMatMultTranspose(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 8398 { 8399 PetscErrorCode ierr; 8400 PetscErrorCode (*fA)(Mat,Mat,MatReuse,PetscReal,Mat*); 8401 PetscErrorCode (*fB)(Mat,Mat,MatReuse,PetscReal,Mat*); 8402 8403 PetscFunctionBegin; 8404 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 8405 PetscValidType(A,1); 8406 if (!A->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8407 if (A->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8408 PetscValidHeaderSpecific(B,MAT_CLASSID,2); 8409 PetscValidType(B,2); 8410 ierr = MatPreallocated(B);CHKERRQ(ierr); 8411 if (!B->assembled) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8412 if (B->factortype) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8413 PetscValidPointer(C,3); 8414 if (B->rmap->N!=A->rmap->N) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap->N,A->rmap->N); 8415 if (fill == PETSC_DEFAULT || fill == PETSC_DECIDE) fill = 2.0; 8416 if (fill < 1.0) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Expected fill=%G must be > 1.0",fill); 8417 ierr = MatPreallocated(A);CHKERRQ(ierr); 8418 8419 fA = A->ops->matmulttranspose; 8420 if (!fA) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"MatMatMultTranspose not supported for A of type %s",((PetscObject)A)->type_name); 8421 fB = B->ops->matmulttranspose; 8422 if (!fB) SETERRQ1(((PetscObject)A)->comm,PETSC_ERR_SUP,"MatMatMultTranspose not supported for B of type %s",((PetscObject)B)->type_name); 8423 if (fB!=fA) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_ARG_INCOMP,"MatMatMultTranspose requires A, %s, to be compatible with B, %s",((PetscObject)A)->type_name,((PetscObject)B)->type_name); 8424 8425 ierr = PetscLogEventBegin(MAT_MatMultTranspose,A,B,0,0);CHKERRQ(ierr); 8426 ierr = (*A->ops->matmulttranspose)(A,B,scall,fill,C);CHKERRQ(ierr); 8427 ierr = PetscLogEventEnd(MAT_MatMultTranspose,A,B,0,0);CHKERRQ(ierr); 8428 8429 PetscFunctionReturn(0); 8430 } 8431 8432 #undef __FUNCT__ 8433 #define __FUNCT__ "MatGetRedundantMatrix" 8434 /*@C 8435 MatGetRedundantMatrix - Create redundant matrices and put them into processors of subcommunicators. 8436 8437 Collective on Mat 8438 8439 Input Parameters: 8440 + mat - the matrix 8441 . nsubcomm - the number of subcommunicators (= number of redundant parallel or sequential matrices) 8442 . subcomm - MPI communicator split from the communicator where mat resides in 8443 . mlocal_red - number of local rows of the redundant matrix 8444 - reuse - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 8445 8446 Output Parameter: 8447 . matredundant - redundant matrix 8448 8449 Notes: 8450 MAT_REUSE_MATRIX can only be used when the nonzero structure of the 8451 original matrix has not changed from that last call to MatGetRedundantMatrix(). 8452 8453 This routine creates the duplicated matrices in subcommunicators; you should NOT create them before 8454 calling it. 8455 8456 Only MPIAIJ matrix is supported. 8457 8458 Level: advanced 8459 8460 Concepts: subcommunicator 8461 Concepts: duplicate matrix 8462 8463 .seealso: MatDestroy() 8464 @*/ 8465 PetscErrorCode MatGetRedundantMatrix(Mat mat,PetscInt nsubcomm,MPI_Comm subcomm,PetscInt mlocal_red,MatReuse reuse,Mat *matredundant) 8466 { 8467 PetscErrorCode ierr; 8468 8469 PetscFunctionBegin; 8470 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 8471 if (nsubcomm && reuse == MAT_REUSE_MATRIX) { 8472 PetscValidPointer(*matredundant,6); 8473 PetscValidHeaderSpecific(*matredundant,MAT_CLASSID,6); 8474 } 8475 if (!mat->ops->getredundantmatrix) SETERRQ1(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Mat type %s",((PetscObject)mat)->type_name); 8476 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8477 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8478 ierr = MatPreallocated(mat);CHKERRQ(ierr); 8479 8480 ierr = PetscLogEventBegin(MAT_GetRedundantMatrix,mat,0,0,0);CHKERRQ(ierr); 8481 ierr = (*mat->ops->getredundantmatrix)(mat,nsubcomm,subcomm,mlocal_red,reuse,matredundant);CHKERRQ(ierr); 8482 ierr = PetscLogEventEnd(MAT_GetRedundantMatrix,mat,0,0,0);CHKERRQ(ierr); 8483 PetscFunctionReturn(0); 8484 } 8485 8486 #undef __FUNCT__ 8487 #define __FUNCT__ "MatGetMultiProcBlock" 8488 /*@C 8489 MatGetMultiProcBlock - Create multiple [bjacobi] 'parallel submatrices' from 8490 a given 'mat' object. Each submatrix can span multiple procs. 8491 8492 Collective on Mat 8493 8494 Input Parameters: 8495 + mat - the matrix 8496 - subcomm - the subcommunicator obtained by com_split(comm) 8497 8498 Output Parameter: 8499 . subMat - 'parallel submatrices each spans a given subcomm 8500 8501 Notes: 8502 The submatrix partition across processors is dicated by 'subComm' a 8503 communicator obtained by com_split(comm). The comm_split 8504 is not restriced to be grouped with consequitive original ranks. 8505 8506 Due the comm_split() usage, the parallel layout of the submatrices 8507 map directly to the layout of the original matrix [wrt the local 8508 row,col partitioning]. So the original 'DiagonalMat' naturally maps 8509 into the 'DiagonalMat' of the subMat, hence it is used directly from 8510 the subMat. However the offDiagMat looses some columns - and this is 8511 reconstructed with MatSetValues() 8512 8513 Level: advanced 8514 8515 Concepts: subcommunicator 8516 Concepts: submatrices 8517 8518 .seealso: MatGetSubMatrices() 8519 @*/ 8520 PetscErrorCode MatGetMultiProcBlock(Mat mat, MPI_Comm subComm, Mat* subMat) 8521 { 8522 PetscErrorCode ierr; 8523 PetscMPIInt commsize,subCommSize; 8524 8525 PetscFunctionBegin; 8526 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&commsize);CHKERRQ(ierr); 8527 ierr = MPI_Comm_size(subComm,&subCommSize);CHKERRQ(ierr); 8528 if (subCommSize > commsize) SETERRQ2(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"CommSize %D < SubCommZize %D",commsize,subCommSize); 8529 8530 ierr = PetscLogEventBegin(MAT_GetMultiProcBlock,mat,0,0,0);CHKERRQ(ierr); 8531 ierr = (*mat->ops->getmultiprocblock)(mat,subComm,subMat);CHKERRQ(ierr); 8532 ierr = PetscLogEventEnd(MAT_GetMultiProcBlock,mat,0,0,0);CHKERRQ(ierr); 8533 PetscFunctionReturn(0); 8534 } 8535 8536 #undef __FUNCT__ 8537 #define __FUNCT__ "MatGetLocalSubMatrix" 8538 /*@ 8539 MatGetLocalSubMatrix - Gets a reference to a submatrix specified in local numbering 8540 8541 Not Collective 8542 8543 Input Arguments: 8544 mat - matrix to extract local submatrix from 8545 isrow - local row indices for submatrix 8546 iscol - local column indices for submatrix 8547 8548 Output Arguments: 8549 submat - the submatrix 8550 8551 Level: intermediate 8552 8553 Notes: 8554 The submat should be returned with MatRestoreLocalSubMatrix(). 8555 8556 Depending on the format of mat, the returned submat may not implement MatMult(). Its communicator may be 8557 the same as mat, it may be PETSC_COMM_SELF, or some other subcomm of mat's. 8558 8559 The submat always implements MatSetValuesLocal(). If isrow and iscol have the same block size, then 8560 MatSetValuesBlockedLocal() will also be implemented. 8561 8562 .seealso: MatRestoreLocalSubMatrix(), MatCreateLocalRef() 8563 @*/ 8564 PetscErrorCode MatGetLocalSubMatrix(Mat mat,IS isrow,IS iscol,Mat *submat) 8565 { 8566 PetscErrorCode ierr; 8567 8568 PetscFunctionBegin; 8569 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 8570 PetscValidHeaderSpecific(isrow,IS_CLASSID,2); 8571 PetscValidHeaderSpecific(iscol,IS_CLASSID,3); 8572 PetscCheckSameComm(isrow,2,iscol,3); 8573 PetscValidPointer(submat,4); 8574 8575 if (mat->ops->getlocalsubmatrix) { 8576 ierr = (*mat->ops->getlocalsubmatrix)(mat,isrow,iscol,submat);CHKERRQ(ierr); 8577 } else { 8578 ierr = MatCreateLocalRef(mat,isrow,iscol,submat);CHKERRQ(ierr); 8579 } 8580 PetscFunctionReturn(0); 8581 } 8582 8583 #undef __FUNCT__ 8584 #define __FUNCT__ "MatRestoreLocalSubMatrix" 8585 /*@ 8586 MatRestoreLocalSubMatrix - Restores a reference to a submatrix specified in local numbering 8587 8588 Not Collective 8589 8590 Input Arguments: 8591 mat - matrix to extract local submatrix from 8592 isrow - local row indices for submatrix 8593 iscol - local column indices for submatrix 8594 submat - the submatrix 8595 8596 Level: intermediate 8597 8598 .seealso: MatGetLocalSubMatrix() 8599 @*/ 8600 PetscErrorCode MatRestoreLocalSubMatrix(Mat mat,IS isrow,IS iscol,Mat *submat) 8601 { 8602 PetscErrorCode ierr; 8603 8604 PetscFunctionBegin; 8605 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 8606 PetscValidHeaderSpecific(isrow,IS_CLASSID,2); 8607 PetscValidHeaderSpecific(iscol,IS_CLASSID,3); 8608 PetscCheckSameComm(isrow,2,iscol,3); 8609 PetscValidPointer(submat,4); 8610 if (*submat) {PetscValidHeaderSpecific(*submat,MAT_CLASSID,4);} 8611 8612 if (mat->ops->restorelocalsubmatrix) { 8613 ierr = (*mat->ops->restorelocalsubmatrix)(mat,isrow,iscol,submat);CHKERRQ(ierr); 8614 } else { 8615 ierr = MatDestroy(*submat);CHKERRQ(ierr); 8616 } 8617 *submat = PETSC_NULL; 8618 PetscFunctionReturn(0); 8619 } 8620 8621 /* --------------------------------------------------------*/ 8622 #undef __FUNCT__ 8623 #define __FUNCT__ "MatFindZeroDiagonals" 8624 /*@ 8625 MatFindZeroDiagonals - Finds all the rows of a matrix that have zero or no entry in the matrix 8626 8627 Collective on Mat 8628 8629 Input Parameter: 8630 . mat - the matrix 8631 8632 Output Parameter: 8633 . is - if any rows have zero diagonals this contains the list of them 8634 8635 Level: developer 8636 8637 Concepts: matrix-vector product 8638 8639 .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 8640 @*/ 8641 PetscErrorCode MatFindZeroDiagonals(Mat mat,IS *is) 8642 { 8643 PetscErrorCode ierr; 8644 8645 PetscFunctionBegin; 8646 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 8647 PetscValidType(mat,1); 8648 if (!mat->assembled) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 8649 if (mat->factortype) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 8650 8651 if (!mat->ops->findzerodiagonals) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"This matrix type does not have a find zero diagonals defined"); 8652 ierr = (*mat->ops->findzerodiagonals)(mat,is);CHKERRQ(ierr); 8653 PetscFunctionReturn(0); 8654 } 8655