1 #define PETSCMAT_DLL 2 3 /* 4 This is where the abstract matrix operations are defined 5 */ 6 7 #include "include/private/matimpl.h" /*I "petscmat.h" I*/ 8 #include "private/vecimpl.h" 9 10 /* Logging support */ 11 PetscCookie PETSCMAT_DLLEXPORT MAT_COOKIE = 0; 12 PetscEvent MAT_Mult = 0, MAT_Mults = 0, MAT_MultConstrained = 0, MAT_MultAdd = 0, MAT_MultTranspose = 0; 13 PetscEvent MAT_MultTransposeConstrained = 0, MAT_MultTransposeAdd = 0, MAT_Solve = 0, MAT_Solves = 0, MAT_SolveAdd = 0, MAT_SolveTranspose = 0, MAT_MatSolve = 0; 14 PetscEvent MAT_SolveTransposeAdd = 0, MAT_Relax = 0, MAT_ForwardSolve = 0, MAT_BackwardSolve = 0, MAT_LUFactor = 0, MAT_LUFactorSymbolic = 0; 15 PetscEvent MAT_LUFactorNumeric = 0, MAT_CholeskyFactor = 0, MAT_CholeskyFactorSymbolic = 0, MAT_CholeskyFactorNumeric = 0, MAT_ILUFactor = 0; 16 PetscEvent MAT_ILUFactorSymbolic = 0, MAT_ICCFactorSymbolic = 0, MAT_Copy = 0, MAT_Convert = 0, MAT_Scale = 0, MAT_AssemblyBegin = 0; 17 PetscEvent MAT_AssemblyEnd = 0, MAT_SetValues = 0, MAT_GetValues = 0, MAT_GetRow = 0, MAT_GetRowIJ = 0, MAT_GetSubMatrices = 0, MAT_GetColoring = 0, MAT_GetOrdering = 0, MAT_GetRedundantMatrix = 0; 18 PetscEvent MAT_IncreaseOverlap = 0, MAT_Partitioning = 0, MAT_ZeroEntries = 0, MAT_Load = 0, MAT_View = 0, MAT_AXPY = 0, MAT_FDColoringCreate = 0; 19 PetscEvent MAT_FDColoringApply = 0,MAT_Transpose = 0,MAT_FDColoringFunction = 0; 20 PetscEvent MAT_MatMult = 0, MAT_MatMultSymbolic = 0, MAT_MatMultNumeric = 0; 21 PetscEvent MAT_PtAP = 0, MAT_PtAPSymbolic = 0, MAT_PtAPNumeric = 0; 22 PetscEvent MAT_MatMultTranspose = 0, MAT_MatMultTransposeSymbolic = 0, MAT_MatMultTransposeNumeric = 0; 23 24 /* nasty global values for MatSetValue() */ 25 PetscInt PETSCMAT_DLLEXPORT MatSetValue_Row = 0; 26 PetscInt PETSCMAT_DLLEXPORT MatSetValue_Column = 0; 27 PetscScalar PETSCMAT_DLLEXPORT MatSetValue_Value = 0.0; 28 29 #undef __FUNCT__ 30 #define __FUNCT__ "MatRealPart" 31 /*@ 32 MatRealPart - Zeros out the imaginary part of the matrix 33 34 Collective on Mat 35 36 Input Parameters: 37 . mat - the matrix 38 39 Level: advanced 40 41 42 .seealso: MatImaginaryPart() 43 @*/ 44 45 PetscErrorCode PETSCMAT_DLLEXPORT MatRealPart(Mat mat) 46 { 47 PetscErrorCode ierr; 48 49 PetscFunctionBegin; 50 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 51 PetscValidType(mat,1); 52 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 53 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 54 if (!mat->ops->realpart) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 55 ierr = MatPreallocated(mat);CHKERRQ(ierr); 56 ierr = (*mat->ops->realpart)(mat);CHKERRQ(ierr); 57 PetscFunctionReturn(0); 58 } 59 60 #undef __FUNCT__ 61 #define __FUNCT__ "MatImaginaryPart" 62 /*@ 63 MatImaginaryPart - Moves the imaginary part of the matrix to the real part and zeros the imaginary part 64 65 Collective on Mat 66 67 Input Parameters: 68 . mat - the matrix 69 70 Level: advanced 71 72 73 .seealso: MatRealPart() 74 @*/ 75 76 PetscErrorCode PETSCMAT_DLLEXPORT MatImaginaryPart(Mat mat) 77 { 78 PetscErrorCode ierr; 79 80 PetscFunctionBegin; 81 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 82 PetscValidType(mat,1); 83 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 84 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 85 if (!mat->ops->imaginarypart) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 86 ierr = MatPreallocated(mat);CHKERRQ(ierr); 87 ierr = (*mat->ops->imaginarypart)(mat);CHKERRQ(ierr); 88 PetscFunctionReturn(0); 89 } 90 91 #undef __FUNCT__ 92 #define __FUNCT__ "MatGetRow" 93 /*@C 94 MatGetRow - Gets a row of a matrix. You MUST call MatRestoreRow() 95 for each row that you get to ensure that your application does 96 not bleed memory. 97 98 Not Collective 99 100 Input Parameters: 101 + mat - the matrix 102 - row - the row to get 103 104 Output Parameters: 105 + ncols - if not NULL, the number of nonzeros in the row 106 . cols - if not NULL, the column numbers 107 - vals - if not NULL, the values 108 109 Notes: 110 This routine is provided for people who need to have direct access 111 to the structure of a matrix. We hope that we provide enough 112 high-level matrix routines that few users will need it. 113 114 MatGetRow() always returns 0-based column indices, regardless of 115 whether the internal representation is 0-based (default) or 1-based. 116 117 For better efficiency, set cols and/or vals to PETSC_NULL if you do 118 not wish to extract these quantities. 119 120 The user can only examine the values extracted with MatGetRow(); 121 the values cannot be altered. To change the matrix entries, one 122 must use MatSetValues(). 123 124 You can only have one call to MatGetRow() outstanding for a particular 125 matrix at a time, per processor. MatGetRow() can only obtain rows 126 associated with the given processor, it cannot get rows from the 127 other processors; for that we suggest using MatGetSubMatrices(), then 128 MatGetRow() on the submatrix. The row indix passed to MatGetRows() 129 is in the global number of rows. 130 131 Fortran Notes: 132 The calling sequence from Fortran is 133 .vb 134 MatGetRow(matrix,row,ncols,cols,values,ierr) 135 Mat matrix (input) 136 integer row (input) 137 integer ncols (output) 138 integer cols(maxcols) (output) 139 double precision (or double complex) values(maxcols) output 140 .ve 141 where maxcols >= maximum nonzeros in any row of the matrix. 142 143 144 Caution: 145 Do not try to change the contents of the output arrays (cols and vals). 146 In some cases, this may corrupt the matrix. 147 148 Level: advanced 149 150 Concepts: matrices^row access 151 152 .seealso: MatRestoreRow(), MatSetValues(), MatGetValues(), MatGetSubMatrices(), MatGetDiagonal() 153 @*/ 154 155 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[]) 156 { 157 PetscErrorCode ierr; 158 PetscInt incols; 159 160 PetscFunctionBegin; 161 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 162 PetscValidType(mat,1); 163 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 164 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 165 if (!mat->ops->getrow) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 166 ierr = MatPreallocated(mat);CHKERRQ(ierr); 167 ierr = PetscLogEventBegin(MAT_GetRow,mat,0,0,0);CHKERRQ(ierr); 168 ierr = (*mat->ops->getrow)(mat,row,&incols,(PetscInt **)cols,(PetscScalar **)vals);CHKERRQ(ierr); 169 if (ncols) *ncols = incols; 170 ierr = PetscLogEventEnd(MAT_GetRow,mat,0,0,0);CHKERRQ(ierr); 171 PetscFunctionReturn(0); 172 } 173 174 #undef __FUNCT__ 175 #define __FUNCT__ "MatConjugate" 176 /*@ 177 MatConjugate - replaces the matrix values with their complex conjugates 178 179 Collective on Mat 180 181 Input Parameters: 182 . mat - the matrix 183 184 Level: advanced 185 186 .seealso: VecConjugate() 187 @*/ 188 PetscErrorCode PETSCMAT_DLLEXPORT MatConjugate(Mat mat) 189 { 190 PetscErrorCode ierr; 191 192 PetscFunctionBegin; 193 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 194 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 195 if (!mat->ops->conjugate) SETERRQ(PETSC_ERR_SUP,"Not provided for this matrix format, send email to petsc-maint@mcs.anl.gov"); 196 ierr = (*mat->ops->conjugate)(mat);CHKERRQ(ierr); 197 PetscFunctionReturn(0); 198 } 199 200 #undef __FUNCT__ 201 #define __FUNCT__ "MatRestoreRow" 202 /*@C 203 MatRestoreRow - Frees any temporary space allocated by MatGetRow(). 204 205 Not Collective 206 207 Input Parameters: 208 + mat - the matrix 209 . row - the row to get 210 . ncols, cols - the number of nonzeros and their columns 211 - vals - if nonzero the column values 212 213 Notes: 214 This routine should be called after you have finished examining the entries. 215 216 Fortran Notes: 217 The calling sequence from Fortran is 218 .vb 219 MatRestoreRow(matrix,row,ncols,cols,values,ierr) 220 Mat matrix (input) 221 integer row (input) 222 integer ncols (output) 223 integer cols(maxcols) (output) 224 double precision (or double complex) values(maxcols) output 225 .ve 226 Where maxcols >= maximum nonzeros in any row of the matrix. 227 228 In Fortran MatRestoreRow() MUST be called after MatGetRow() 229 before another call to MatGetRow() can be made. 230 231 Level: advanced 232 233 .seealso: MatGetRow() 234 @*/ 235 PetscErrorCode PETSCMAT_DLLEXPORT MatRestoreRow(Mat mat,PetscInt row,PetscInt *ncols,const PetscInt *cols[],const PetscScalar *vals[]) 236 { 237 PetscErrorCode ierr; 238 239 PetscFunctionBegin; 240 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 241 PetscValidIntPointer(ncols,3); 242 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 243 if (!mat->ops->restorerow) PetscFunctionReturn(0); 244 ierr = (*mat->ops->restorerow)(mat,row,ncols,(PetscInt **)cols,(PetscScalar **)vals);CHKERRQ(ierr); 245 PetscFunctionReturn(0); 246 } 247 248 #undef __FUNCT__ 249 #define __FUNCT__ "MatGetRowUpperTriangular" 250 /*@C 251 MatGetRowUpperTriangular - Sets a flag to enable calls to MatGetRow() for matrix in MATSBAIJ format. 252 You should call MatRestoreRowUpperTriangular() after calling MatGetRow/MatRestoreRow() to disable the flag. 253 254 Not Collective 255 256 Input Parameters: 257 + mat - the matrix 258 259 Notes: 260 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. 261 262 Level: advanced 263 264 Concepts: matrices^row access 265 266 .seealso: MatRestoreRowRowUpperTriangular() 267 @*/ 268 269 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowUpperTriangular(Mat mat) 270 { 271 PetscErrorCode ierr; 272 273 PetscFunctionBegin; 274 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 275 PetscValidType(mat,1); 276 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 277 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 278 if (!mat->ops->getrowuppertriangular) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 279 ierr = MatPreallocated(mat);CHKERRQ(ierr); 280 ierr = (*mat->ops->getrowuppertriangular)(mat);CHKERRQ(ierr); 281 PetscFunctionReturn(0); 282 } 283 284 #undef __FUNCT__ 285 #define __FUNCT__ "MatRestoreRowUpperTriangular" 286 /*@C 287 MatRestoreRowUpperTriangular - Disable calls to MatGetRow() for matrix in MATSBAIJ format. 288 289 Not Collective 290 291 Input Parameters: 292 + mat - the matrix 293 294 Notes: 295 This routine should be called after you have finished MatGetRow/MatRestoreRow(). 296 297 298 Level: advanced 299 300 .seealso: MatGetRowUpperTriangular() 301 @*/ 302 PetscErrorCode PETSCMAT_DLLEXPORT MatRestoreRowUpperTriangular(Mat mat) 303 { 304 PetscErrorCode ierr; 305 306 PetscFunctionBegin; 307 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 308 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 309 if (!mat->ops->restorerowuppertriangular) PetscFunctionReturn(0); 310 ierr = (*mat->ops->restorerowuppertriangular)(mat);CHKERRQ(ierr); 311 PetscFunctionReturn(0); 312 } 313 314 #undef __FUNCT__ 315 #define __FUNCT__ "MatSetOptionsPrefix" 316 /*@C 317 MatSetOptionsPrefix - Sets the prefix used for searching for all 318 Mat options in the database. 319 320 Collective on Mat 321 322 Input Parameter: 323 + A - the Mat context 324 - prefix - the prefix to prepend to all option names 325 326 Notes: 327 A hyphen (-) must NOT be given at the beginning of the prefix name. 328 The first character of all runtime options is AUTOMATICALLY the hyphen. 329 330 Level: advanced 331 332 .keywords: Mat, set, options, prefix, database 333 334 .seealso: MatSetFromOptions() 335 @*/ 336 PetscErrorCode PETSCMAT_DLLEXPORT MatSetOptionsPrefix(Mat A,const char prefix[]) 337 { 338 PetscErrorCode ierr; 339 340 PetscFunctionBegin; 341 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 342 ierr = PetscObjectSetOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 343 PetscFunctionReturn(0); 344 } 345 346 #undef __FUNCT__ 347 #define __FUNCT__ "MatAppendOptionsPrefix" 348 /*@C 349 MatAppendOptionsPrefix - Appends to the prefix used for searching for all 350 Mat options in the database. 351 352 Collective on Mat 353 354 Input Parameters: 355 + A - the Mat context 356 - prefix - the prefix to prepend to all option names 357 358 Notes: 359 A hyphen (-) must NOT be given at the beginning of the prefix name. 360 The first character of all runtime options is AUTOMATICALLY the hyphen. 361 362 Level: advanced 363 364 .keywords: Mat, append, options, prefix, database 365 366 .seealso: MatGetOptionsPrefix() 367 @*/ 368 PetscErrorCode PETSCMAT_DLLEXPORT MatAppendOptionsPrefix(Mat A,const char prefix[]) 369 { 370 PetscErrorCode ierr; 371 372 PetscFunctionBegin; 373 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 374 ierr = PetscObjectAppendOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 375 PetscFunctionReturn(0); 376 } 377 378 #undef __FUNCT__ 379 #define __FUNCT__ "MatGetOptionsPrefix" 380 /*@C 381 MatGetOptionsPrefix - Sets the prefix used for searching for all 382 Mat options in the database. 383 384 Not Collective 385 386 Input Parameter: 387 . A - the Mat context 388 389 Output Parameter: 390 . prefix - pointer to the prefix string used 391 392 Notes: On the fortran side, the user should pass in a string 'prefix' of 393 sufficient length to hold the prefix. 394 395 Level: advanced 396 397 .keywords: Mat, get, options, prefix, database 398 399 .seealso: MatAppendOptionsPrefix() 400 @*/ 401 PetscErrorCode PETSCMAT_DLLEXPORT MatGetOptionsPrefix(Mat A,const char *prefix[]) 402 { 403 PetscErrorCode ierr; 404 405 PetscFunctionBegin; 406 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 407 ierr = PetscObjectGetOptionsPrefix((PetscObject)A,prefix);CHKERRQ(ierr); 408 PetscFunctionReturn(0); 409 } 410 411 #undef __FUNCT__ 412 #define __FUNCT__ "MatSetUp" 413 /*@ 414 MatSetUp - Sets up the internal matrix data structures for the later use. 415 416 Collective on Mat 417 418 Input Parameters: 419 . A - the Mat context 420 421 Notes: 422 For basic use of the Mat classes the user need not explicitly call 423 MatSetUp(), since these actions will happen automatically. 424 425 Level: advanced 426 427 .keywords: Mat, setup 428 429 .seealso: MatCreate(), MatDestroy() 430 @*/ 431 PetscErrorCode PETSCMAT_DLLEXPORT MatSetUp(Mat A) 432 { 433 PetscErrorCode ierr; 434 435 PetscFunctionBegin; 436 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 437 ierr = MatSetUpPreallocation(A);CHKERRQ(ierr); 438 ierr = MatSetFromOptions(A);CHKERRQ(ierr); 439 PetscFunctionReturn(0); 440 } 441 442 #undef __FUNCT__ 443 #define __FUNCT__ "MatView" 444 /*@C 445 MatView - Visualizes a matrix object. 446 447 Collective on Mat 448 449 Input Parameters: 450 + mat - the matrix 451 - viewer - visualization context 452 453 Notes: 454 The available visualization contexts include 455 + PETSC_VIEWER_STDOUT_SELF - standard output (default) 456 . PETSC_VIEWER_STDOUT_WORLD - synchronized standard 457 output where only the first processor opens 458 the file. All other processors send their 459 data to the first processor to print. 460 - PETSC_VIEWER_DRAW_WORLD - graphical display of nonzero structure 461 462 The user can open alternative visualization contexts with 463 + PetscViewerASCIIOpen() - Outputs matrix to a specified file 464 . PetscViewerBinaryOpen() - Outputs matrix in binary to a 465 specified file; corresponding input uses MatLoad() 466 . PetscViewerDrawOpen() - Outputs nonzero matrix structure to 467 an X window display 468 - PetscViewerSocketOpen() - Outputs matrix to Socket viewer. 469 Currently only the sequential dense and AIJ 470 matrix types support the Socket viewer. 471 472 The user can call PetscViewerSetFormat() to specify the output 473 format of ASCII printed objects (when using PETSC_VIEWER_STDOUT_SELF, 474 PETSC_VIEWER_STDOUT_WORLD and PetscViewerASCIIOpen). Available formats include 475 + PETSC_VIEWER_ASCII_DEFAULT - default, prints matrix contents 476 . PETSC_VIEWER_ASCII_MATLAB - prints matrix contents in Matlab format 477 . PETSC_VIEWER_ASCII_DENSE - prints entire matrix including zeros 478 . PETSC_VIEWER_ASCII_COMMON - prints matrix contents, using a sparse 479 format common among all matrix types 480 . PETSC_VIEWER_ASCII_IMPL - prints matrix contents, using an implementation-specific 481 format (which is in many cases the same as the default) 482 . PETSC_VIEWER_ASCII_INFO - prints basic information about the matrix 483 size and structure (not the matrix entries) 484 . PETSC_VIEWER_ASCII_INFO_DETAIL - prints more detailed information about 485 the matrix structure 486 487 Options Database Keys: 488 + -mat_view_info - Prints info on matrix at conclusion of MatEndAssembly() 489 . -mat_view_info_detailed - Prints more detailed info 490 . -mat_view - Prints matrix in ASCII format 491 . -mat_view_matlab - Prints matrix in Matlab format 492 . -mat_view_draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX(). 493 . -display <name> - Sets display name (default is host) 494 . -draw_pause <sec> - Sets number of seconds to pause after display 495 . -mat_view_socket - Sends matrix to socket, can be accessed from Matlab (see users manual) 496 . -viewer_socket_machine <machine> 497 . -viewer_socket_port <port> 498 . -mat_view_binary - save matrix to file in binary format 499 - -viewer_binary_filename <name> 500 Level: beginner 501 502 Concepts: matrices^viewing 503 Concepts: matrices^plotting 504 Concepts: matrices^printing 505 506 .seealso: PetscViewerSetFormat(), PetscViewerASCIIOpen(), PetscViewerDrawOpen(), 507 PetscViewerSocketOpen(), PetscViewerBinaryOpen(), MatLoad() 508 @*/ 509 PetscErrorCode PETSCMAT_DLLEXPORT MatView(Mat mat,PetscViewer viewer) 510 { 511 PetscErrorCode ierr; 512 PetscInt rows,cols; 513 PetscTruth iascii; 514 const char *cstr; 515 PetscViewerFormat format; 516 517 PetscFunctionBegin; 518 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 519 PetscValidType(mat,1); 520 if (!viewer) { 521 ierr = PetscViewerASCIIGetStdout(mat->comm,&viewer);CHKERRQ(ierr); 522 } 523 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_COOKIE,2); 524 PetscCheckSameComm(mat,1,viewer,2); 525 if (!mat->assembled) SETERRQ(PETSC_ERR_ORDER,"Must call MatAssemblyBegin/End() before viewing matrix"); 526 ierr = MatPreallocated(mat);CHKERRQ(ierr); 527 528 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr); 529 if (iascii) { 530 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 531 if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 532 if (mat->prefix) { 533 ierr = PetscViewerASCIIPrintf(viewer,"Matrix Object:(%s)\n",mat->prefix);CHKERRQ(ierr); 534 } else { 535 ierr = PetscViewerASCIIPrintf(viewer,"Matrix Object:\n");CHKERRQ(ierr); 536 } 537 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 538 ierr = MatGetType(mat,&cstr);CHKERRQ(ierr); 539 ierr = MatGetSize(mat,&rows,&cols);CHKERRQ(ierr); 540 ierr = PetscViewerASCIIPrintf(viewer,"type=%s, rows=%D, cols=%D\n",cstr,rows,cols);CHKERRQ(ierr); 541 if (mat->ops->getinfo) { 542 MatInfo info; 543 ierr = MatGetInfo(mat,MAT_GLOBAL_SUM,&info);CHKERRQ(ierr); 544 ierr = PetscViewerASCIIPrintf(viewer,"total: nonzeros=%D, allocated nonzeros=%D\n", 545 (PetscInt)info.nz_used,(PetscInt)info.nz_allocated);CHKERRQ(ierr); 546 } 547 } 548 } 549 if (mat->ops->view) { 550 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 551 ierr = (*mat->ops->view)(mat,viewer);CHKERRQ(ierr); 552 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 553 } else if (!iascii) { 554 SETERRQ1(PETSC_ERR_SUP,"Viewer type %s not supported",((PetscObject)viewer)->type_name); 555 } 556 if (iascii) { 557 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 558 if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 559 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 560 } 561 } 562 PetscFunctionReturn(0); 563 } 564 565 #undef __FUNCT__ 566 #define __FUNCT__ "MatScaleSystem" 567 /*@C 568 MatScaleSystem - Scale a vector solution and right hand side to 569 match the scaling of a scaled matrix. 570 571 Collective on Mat 572 573 Input Parameter: 574 + mat - the matrix 575 . b - right hand side vector (or PETSC_NULL) 576 - x - solution vector (or PETSC_NULL) 577 578 579 Notes: 580 For AIJ, BAIJ, and BDiag matrix formats, the matrices are not 581 internally scaled, so this does nothing. For MPIROWBS it 582 permutes and diagonally scales. 583 584 The KSP methods automatically call this routine when required 585 (via PCPreSolve()) so it is rarely used directly. 586 587 Level: Developer 588 589 Concepts: matrices^scaling 590 591 .seealso: MatUseScaledForm(), MatUnScaleSystem() 592 @*/ 593 PetscErrorCode PETSCMAT_DLLEXPORT MatScaleSystem(Mat mat,Vec b,Vec x) 594 { 595 PetscErrorCode ierr; 596 597 PetscFunctionBegin; 598 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 599 PetscValidType(mat,1); 600 ierr = MatPreallocated(mat);CHKERRQ(ierr); 601 if (x) {PetscValidHeaderSpecific(x,VEC_COOKIE,2);PetscCheckSameComm(mat,1,x,2);} 602 if (b) {PetscValidHeaderSpecific(b,VEC_COOKIE,3);PetscCheckSameComm(mat,1,b,3);} 603 604 if (mat->ops->scalesystem) { 605 ierr = (*mat->ops->scalesystem)(mat,b,x);CHKERRQ(ierr); 606 } 607 PetscFunctionReturn(0); 608 } 609 610 #undef __FUNCT__ 611 #define __FUNCT__ "MatUnScaleSystem" 612 /*@C 613 MatUnScaleSystem - Unscales a vector solution and right hand side to 614 match the original scaling of a scaled matrix. 615 616 Collective on Mat 617 618 Input Parameter: 619 + mat - the matrix 620 . b - right hand side vector (or PETSC_NULL) 621 - x - solution vector (or PETSC_NULL) 622 623 624 Notes: 625 For AIJ, BAIJ, and BDiag matrix formats, the matrices are not 626 internally scaled, so this does nothing. For MPIROWBS it 627 permutes and diagonally scales. 628 629 The KSP methods automatically call this routine when required 630 (via PCPreSolve()) so it is rarely used directly. 631 632 Level: Developer 633 634 .seealso: MatUseScaledForm(), MatScaleSystem() 635 @*/ 636 PetscErrorCode PETSCMAT_DLLEXPORT MatUnScaleSystem(Mat mat,Vec b,Vec x) 637 { 638 PetscErrorCode ierr; 639 640 PetscFunctionBegin; 641 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 642 PetscValidType(mat,1); 643 ierr = MatPreallocated(mat);CHKERRQ(ierr); 644 if (x) {PetscValidHeaderSpecific(x,VEC_COOKIE,2);PetscCheckSameComm(mat,1,x,2);} 645 if (b) {PetscValidHeaderSpecific(b,VEC_COOKIE,3);PetscCheckSameComm(mat,1,b,3);} 646 if (mat->ops->unscalesystem) { 647 ierr = (*mat->ops->unscalesystem)(mat,b,x);CHKERRQ(ierr); 648 } 649 PetscFunctionReturn(0); 650 } 651 652 #undef __FUNCT__ 653 #define __FUNCT__ "MatUseScaledForm" 654 /*@C 655 MatUseScaledForm - For matrix storage formats that scale the 656 matrix (for example MPIRowBS matrices are diagonally scaled on 657 assembly) indicates matrix operations (MatMult() etc) are 658 applied using the scaled matrix. 659 660 Collective on Mat 661 662 Input Parameter: 663 + mat - the matrix 664 - scaled - PETSC_TRUE for applying the scaled, PETSC_FALSE for 665 applying the original matrix 666 667 Notes: 668 For scaled matrix formats, applying the original, unscaled matrix 669 will be slightly more expensive 670 671 Level: Developer 672 673 .seealso: MatScaleSystem(), MatUnScaleSystem() 674 @*/ 675 PetscErrorCode PETSCMAT_DLLEXPORT MatUseScaledForm(Mat mat,PetscTruth scaled) 676 { 677 PetscErrorCode ierr; 678 679 PetscFunctionBegin; 680 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 681 PetscValidType(mat,1); 682 ierr = MatPreallocated(mat);CHKERRQ(ierr); 683 if (mat->ops->usescaledform) { 684 ierr = (*mat->ops->usescaledform)(mat,scaled);CHKERRQ(ierr); 685 } 686 PetscFunctionReturn(0); 687 } 688 689 #undef __FUNCT__ 690 #define __FUNCT__ "MatDestroy" 691 /*@ 692 MatDestroy - Frees space taken by a matrix. 693 694 Collective on Mat 695 696 Input Parameter: 697 . A - the matrix 698 699 Level: beginner 700 701 @*/ 702 PetscErrorCode PETSCMAT_DLLEXPORT MatDestroy(Mat A) 703 { 704 PetscErrorCode ierr; 705 PetscFunctionBegin; 706 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 707 if (--A->refct > 0) PetscFunctionReturn(0); 708 ierr = MatPreallocated(A);CHKERRQ(ierr); 709 /* if memory was published with AMS then destroy it */ 710 ierr = PetscObjectDepublish(A);CHKERRQ(ierr); 711 if (A->ops->destroy) { 712 ierr = (*A->ops->destroy)(A);CHKERRQ(ierr); 713 } 714 if (A->mapping) { 715 ierr = ISLocalToGlobalMappingDestroy(A->mapping);CHKERRQ(ierr); 716 } 717 if (A->bmapping) { 718 ierr = ISLocalToGlobalMappingDestroy(A->bmapping);CHKERRQ(ierr); 719 } 720 ierr = PetscFree(A->rmap.range);CHKERRQ(ierr); 721 ierr = PetscFree(A->cmap.range);CHKERRQ(ierr); 722 if (A->spptr){ierr = PetscFree(A->spptr);CHKERRQ(ierr);} 723 ierr = PetscHeaderDestroy(A);CHKERRQ(ierr); 724 PetscFunctionReturn(0); 725 } 726 727 #undef __FUNCT__ 728 #define __FUNCT__ "MatValid" 729 /*@ 730 MatValid - Checks whether a matrix object is valid. 731 732 Collective on Mat 733 734 Input Parameter: 735 . m - the matrix to check 736 737 Output Parameter: 738 flg - flag indicating matrix status, either 739 PETSC_TRUE if matrix is valid, or PETSC_FALSE otherwise. 740 741 Level: developer 742 743 Concepts: matrices^validity 744 @*/ 745 PetscErrorCode PETSCMAT_DLLEXPORT MatValid(Mat m,PetscTruth *flg) 746 { 747 PetscFunctionBegin; 748 PetscValidIntPointer(flg,1); 749 if (!m) *flg = PETSC_FALSE; 750 else if (m->cookie != MAT_COOKIE) *flg = PETSC_FALSE; 751 else *flg = PETSC_TRUE; 752 PetscFunctionReturn(0); 753 } 754 755 #undef __FUNCT__ 756 #define __FUNCT__ "MatSetValues" 757 /*@ 758 MatSetValues - Inserts or adds a block of values into a matrix. 759 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 760 MUST be called after all calls to MatSetValues() have been completed. 761 762 Not Collective 763 764 Input Parameters: 765 + mat - the matrix 766 . v - a logically two-dimensional array of values 767 . m, idxm - the number of rows and their global indices 768 . n, idxn - the number of columns and their global indices 769 - addv - either ADD_VALUES or INSERT_VALUES, where 770 ADD_VALUES adds values to any existing entries, and 771 INSERT_VALUES replaces existing entries with new values 772 773 Notes: 774 By default the values, v, are row-oriented and unsorted. 775 See MatSetOption() for other options. 776 777 Calls to MatSetValues() with the INSERT_VALUES and ADD_VALUES 778 options cannot be mixed without intervening calls to the assembly 779 routines. 780 781 MatSetValues() uses 0-based row and column numbers in Fortran 782 as well as in C. 783 784 Negative indices may be passed in idxm and idxn, these rows and columns are 785 simply ignored. This allows easily inserting element stiffness matrices 786 with homogeneous Dirchlet boundary conditions that you don't want represented 787 in the matrix. 788 789 Efficiency Alert: 790 The routine MatSetValuesBlocked() may offer much better efficiency 791 for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ). 792 793 Level: beginner 794 795 Concepts: matrices^putting entries in 796 797 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 798 InsertMode, INSERT_VALUES, ADD_VALUES 799 @*/ 800 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv) 801 { 802 PetscErrorCode ierr; 803 804 PetscFunctionBegin; 805 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 806 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 807 PetscValidType(mat,1); 808 PetscValidIntPointer(idxm,3); 809 PetscValidIntPointer(idxn,5); 810 PetscValidScalarPointer(v,6); 811 ierr = MatPreallocated(mat);CHKERRQ(ierr); 812 if (mat->insertmode == NOT_SET_VALUES) { 813 mat->insertmode = addv; 814 } 815 #if defined(PETSC_USE_DEBUG) 816 else if (mat->insertmode != addv) { 817 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 818 } 819 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 820 #endif 821 822 if (mat->assembled) { 823 mat->was_assembled = PETSC_TRUE; 824 mat->assembled = PETSC_FALSE; 825 } 826 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 827 if (!mat->ops->setvalues) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 828 ierr = (*mat->ops->setvalues)(mat,m,idxm,n,idxn,v,addv);CHKERRQ(ierr); 829 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 830 PetscFunctionReturn(0); 831 } 832 833 834 #undef __FUNCT__ 835 #define __FUNCT__ "MatSetValuesRowLocal" 836 /*@ 837 MatSetValuesRowLocal - Inserts a row (block row for BAIJ matrices) of nonzero 838 values into a matrix 839 840 Not Collective 841 842 Input Parameters: 843 + mat - the matrix 844 . row - the (block) row to set 845 - v - a logically two-dimensional array of values 846 847 Notes: 848 By the values, v, are column-oriented (for the block version) and sorted 849 850 All the nonzeros in the row must be provided 851 852 The matrix must have previously had its column indices set 853 854 The row must belong to this process 855 856 Level: intermediate 857 858 Concepts: matrices^putting entries in 859 860 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 861 InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues(), MatSetValuesRow(), MatSetLocalToGlobalMapping() 862 @*/ 863 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesRowLocal(Mat mat,PetscInt row,const PetscScalar v[]) 864 { 865 PetscErrorCode ierr; 866 867 PetscFunctionBegin; 868 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 869 PetscValidType(mat,1); 870 PetscValidScalarPointer(v,2); 871 ierr = MatSetValuesRow(mat, mat->mapping->indices[row],v);CHKERRQ(ierr); 872 PetscFunctionReturn(0); 873 } 874 875 #undef __FUNCT__ 876 #define __FUNCT__ "MatSetValuesRow" 877 /*@ 878 MatSetValuesRow - Inserts a row (block row for BAIJ matrices) of nonzero 879 values into a matrix 880 881 Not Collective 882 883 Input Parameters: 884 + mat - the matrix 885 . row - the (block) row to set 886 - v - a logically two-dimensional array of values 887 888 Notes: 889 By the values, v, are column-oriented (for the block version) and sorted 890 891 All the nonzeros in the row must be provided 892 893 The matrix must have previously had its column indices set 894 895 The row must belong to this process 896 897 Level: intermediate 898 899 Concepts: matrices^putting entries in 900 901 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 902 InsertMode, INSERT_VALUES, ADD_VALUES, MatSetValues() 903 @*/ 904 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesRow(Mat mat,PetscInt row,const PetscScalar v[]) 905 { 906 PetscErrorCode ierr; 907 908 PetscFunctionBegin; 909 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 910 PetscValidType(mat,1); 911 PetscValidScalarPointer(v,2); 912 #if defined(PETSC_USE_DEBUG) 913 if (mat->insertmode == ADD_VALUES) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add and insert values"); 914 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 915 #endif 916 mat->insertmode = INSERT_VALUES; 917 918 if (mat->assembled) { 919 mat->was_assembled = PETSC_TRUE; 920 mat->assembled = PETSC_FALSE; 921 } 922 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 923 if (!mat->ops->setvaluesrow) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 924 ierr = (*mat->ops->setvaluesrow)(mat,row,v);CHKERRQ(ierr); 925 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 926 PetscFunctionReturn(0); 927 } 928 929 #undef __FUNCT__ 930 #define __FUNCT__ "MatSetValuesStencil" 931 /*@ 932 MatSetValuesStencil - Inserts or adds a block of values into a matrix. 933 Using structured grid indexing 934 935 Not Collective 936 937 Input Parameters: 938 + mat - the matrix 939 . v - a logically two-dimensional array of values 940 . m - number of rows being entered 941 . idxm - grid coordinates (and component number when dof > 1) for matrix rows being entered 942 . n - number of columns being entered 943 . idxn - grid coordinates (and component number when dof > 1) for matrix columns being entered 944 - addv - either ADD_VALUES or INSERT_VALUES, where 945 ADD_VALUES adds values to any existing entries, and 946 INSERT_VALUES replaces existing entries with new values 947 948 Notes: 949 By default the values, v, are row-oriented and unsorted. 950 See MatSetOption() for other options. 951 952 Calls to MatSetValuesStencil() with the INSERT_VALUES and ADD_VALUES 953 options cannot be mixed without intervening calls to the assembly 954 routines. 955 956 The grid coordinates are across the entire grid, not just the local portion 957 958 MatSetValuesStencil() uses 0-based row and column numbers in Fortran 959 as well as in C. 960 961 For setting/accessing vector values via array coordinates you can use the DAVecGetArray() routine 962 963 In order to use this routine you must either obtain the matrix with DAGetMatrix() 964 or call MatSetLocalToGlobalMapping() and MatSetStencil() first. 965 966 The columns and rows in the stencil passed in MUST be contained within the 967 ghost region of the given process as set with DACreateXXX() or MatSetStencil(). For example, 968 if you create a DA with an overlap of one grid level and on a particular process its first 969 local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the 970 first i index you can use in your column and row indices in MatSetStencil() is 5. 971 972 In Fortran idxm and idxn should be declared as 973 $ MatStencil idxm(4,m),idxn(4,n) 974 and the values inserted using 975 $ idxm(MatStencil_i,1) = i 976 $ idxm(MatStencil_j,1) = j 977 $ idxm(MatStencil_k,1) = k 978 $ idxm(MatStencil_c,1) = c 979 etc 980 981 For periodic boundary conditions use negative indices for values to the left (below 0; that are to be 982 obtained by wrapping values from right edge). For values to the right of the last entry using that index plus one 983 etc to obtain values that obtained by wrapping the values from the left edge. 984 985 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 986 a single value per point) you can skip filling those indices. 987 988 Inspired by the structured grid interface to the HYPRE package 989 (http://www.llnl.gov/CASC/hypre) 990 991 Efficiency Alert: 992 The routine MatSetValuesBlockedStencil() may offer much better efficiency 993 for users of block sparse formats (MATSEQBAIJ and MATMPIBAIJ). 994 995 Level: beginner 996 997 Concepts: matrices^putting entries in 998 999 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1000 MatSetValues(), MatSetValuesBlockedStencil(), MatSetStencil(), DAGetMatrix(), DAVecGetArray(), MatStencil 1001 @*/ 1002 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv) 1003 { 1004 PetscErrorCode ierr; 1005 PetscInt j,i,jdxm[128],jdxn[256],dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp; 1006 PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc); 1007 1008 PetscFunctionBegin; 1009 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1010 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1011 PetscValidType(mat,1); 1012 PetscValidIntPointer(idxm,3); 1013 PetscValidIntPointer(idxn,5); 1014 PetscValidScalarPointer(v,6); 1015 1016 if (m > 128) SETERRQ1(PETSC_ERR_SUP,"Can only set 128 rows at a time; trying to set %D",m); 1017 if (n > 256) SETERRQ1(PETSC_ERR_SUP,"Can only set 256 columns at a time; trying to set %D",n); 1018 1019 for (i=0; i<m; i++) { 1020 for (j=0; j<3-sdim; j++) dxm++; 1021 tmp = *dxm++ - starts[0]; 1022 for (j=0; j<dim-1; j++) { 1023 if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1024 else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1]; 1025 } 1026 if (mat->stencil.noc) dxm++; 1027 jdxm[i] = tmp; 1028 } 1029 for (i=0; i<n; i++) { 1030 for (j=0; j<3-sdim; j++) dxn++; 1031 tmp = *dxn++ - starts[0]; 1032 for (j=0; j<dim-1; j++) { 1033 if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1034 else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1]; 1035 } 1036 if (mat->stencil.noc) dxn++; 1037 jdxn[i] = tmp; 1038 } 1039 ierr = MatSetValuesLocal(mat,m,jdxm,n,jdxn,v,addv);CHKERRQ(ierr); 1040 PetscFunctionReturn(0); 1041 } 1042 1043 #undef __FUNCT__ 1044 #define __FUNCT__ "MatSetValuesBlockedStencil" 1045 /*@C 1046 MatSetValuesBlockedStencil - Inserts or adds a block of values into a matrix. 1047 Using structured grid indexing 1048 1049 Not Collective 1050 1051 Input Parameters: 1052 + mat - the matrix 1053 . v - a logically two-dimensional array of values 1054 . m - number of rows being entered 1055 . idxm - grid coordinates for matrix rows being entered 1056 . n - number of columns being entered 1057 . idxn - grid coordinates for matrix columns being entered 1058 - addv - either ADD_VALUES or INSERT_VALUES, where 1059 ADD_VALUES adds values to any existing entries, and 1060 INSERT_VALUES replaces existing entries with new values 1061 1062 Notes: 1063 By default the values, v, are row-oriented and unsorted. 1064 See MatSetOption() for other options. 1065 1066 Calls to MatSetValuesBlockedStencil() with the INSERT_VALUES and ADD_VALUES 1067 options cannot be mixed without intervening calls to the assembly 1068 routines. 1069 1070 The grid coordinates are across the entire grid, not just the local portion 1071 1072 MatSetValuesBlockedStencil() uses 0-based row and column numbers in Fortran 1073 as well as in C. 1074 1075 For setting/accessing vector values via array coordinates you can use the DAVecGetArray() routine 1076 1077 In order to use this routine you must either obtain the matrix with DAGetMatrix() 1078 or call MatSetLocalToGlobalMapping() and MatSetStencil() first. 1079 1080 The columns and rows in the stencil passed in MUST be contained within the 1081 ghost region of the given process as set with DACreateXXX() or MatSetStencil(). For example, 1082 if you create a DA with an overlap of one grid level and on a particular process its first 1083 local nonghost x logical coordinate is 6 (so its first ghost x logical coordinate is 5) the 1084 first i index you can use in your column and row indices in MatSetStencil() is 5. 1085 1086 In Fortran idxm and idxn should be declared as 1087 $ MatStencil idxm(4,m),idxn(4,n) 1088 and the values inserted using 1089 $ idxm(MatStencil_i,1) = i 1090 $ idxm(MatStencil_j,1) = j 1091 $ idxm(MatStencil_k,1) = k 1092 etc 1093 1094 Negative indices may be passed in idxm and idxn, these rows and columns are 1095 simply ignored. This allows easily inserting element stiffness matrices 1096 with homogeneous Dirchlet boundary conditions that you don't want represented 1097 in the matrix. 1098 1099 Inspired by the structured grid interface to the HYPRE package 1100 (http://www.llnl.gov/CASC/hypre) 1101 1102 Level: beginner 1103 1104 Concepts: matrices^putting entries in 1105 1106 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1107 MatSetValues(), MatSetValuesStencil(), MatSetStencil(), DAGetMatrix(), DAVecGetArray(), MatStencil 1108 @*/ 1109 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesBlockedStencil(Mat mat,PetscInt m,const MatStencil idxm[],PetscInt n,const MatStencil idxn[],const PetscScalar v[],InsertMode addv) 1110 { 1111 PetscErrorCode ierr; 1112 PetscInt j,i,jdxm[128],jdxn[256],dim = mat->stencil.dim,*dims = mat->stencil.dims+1,tmp; 1113 PetscInt *starts = mat->stencil.starts,*dxm = (PetscInt*)idxm,*dxn = (PetscInt*)idxn,sdim = dim - (1 - (PetscInt)mat->stencil.noc); 1114 1115 PetscFunctionBegin; 1116 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1117 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1118 PetscValidType(mat,1); 1119 PetscValidIntPointer(idxm,3); 1120 PetscValidIntPointer(idxn,5); 1121 PetscValidScalarPointer(v,6); 1122 1123 if (m > 128) SETERRQ1(PETSC_ERR_SUP,"Can only set 128 rows at a time; trying to set %D",m); 1124 if (n > 128) SETERRQ1(PETSC_ERR_SUP,"Can only set 256 columns at a time; trying to set %D",n); 1125 1126 for (i=0; i<m; i++) { 1127 for (j=0; j<3-sdim; j++) dxm++; 1128 tmp = *dxm++ - starts[0]; 1129 for (j=0; j<sdim-1; j++) { 1130 if ((*dxm++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1131 else tmp = tmp*dims[j] + *(dxm-1) - starts[j+1]; 1132 } 1133 dxm++; 1134 jdxm[i] = tmp; 1135 } 1136 for (i=0; i<n; i++) { 1137 for (j=0; j<3-sdim; j++) dxn++; 1138 tmp = *dxn++ - starts[0]; 1139 for (j=0; j<sdim-1; j++) { 1140 if ((*dxn++ - starts[j+1]) < 0 || tmp < 0) tmp = PETSC_MIN_INT; 1141 else tmp = tmp*dims[j] + *(dxn-1) - starts[j+1]; 1142 } 1143 dxn++; 1144 jdxn[i] = tmp; 1145 } 1146 ierr = MatSetValuesBlockedLocal(mat,m,jdxm,n,jdxn,v,addv);CHKERRQ(ierr); 1147 PetscFunctionReturn(0); 1148 } 1149 1150 #undef __FUNCT__ 1151 #define __FUNCT__ "MatSetStencil" 1152 /*@ 1153 MatSetStencil - Sets the grid information for setting values into a matrix via 1154 MatSetValuesStencil() 1155 1156 Not Collective 1157 1158 Input Parameters: 1159 + mat - the matrix 1160 . dim - dimension of the grid 1, 2, or 3 1161 . dims - number of grid points in x, y, and z direction, including ghost points on your processor 1162 . starts - starting point of ghost nodes on your processor in x, y, and z direction 1163 - dof - number of degrees of freedom per node 1164 1165 1166 Inspired by the structured grid interface to the HYPRE package 1167 (www.llnl.gov/CASC/hyper) 1168 1169 For matrices generated with DAGetMatrix() this routine is automatically called and so not needed by the 1170 user. 1171 1172 Level: beginner 1173 1174 Concepts: matrices^putting entries in 1175 1176 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal() 1177 MatSetValues(), MatSetValuesBlockedStencil(), MatSetValuesStencil() 1178 @*/ 1179 PetscErrorCode PETSCMAT_DLLEXPORT MatSetStencil(Mat mat,PetscInt dim,const PetscInt dims[],const PetscInt starts[],PetscInt dof) 1180 { 1181 PetscInt i; 1182 1183 PetscFunctionBegin; 1184 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1185 PetscValidIntPointer(dims,3); 1186 PetscValidIntPointer(starts,4); 1187 1188 mat->stencil.dim = dim + (dof > 1); 1189 for (i=0; i<dim; i++) { 1190 mat->stencil.dims[i] = dims[dim-i-1]; /* copy the values in backwards */ 1191 mat->stencil.starts[i] = starts[dim-i-1]; 1192 } 1193 mat->stencil.dims[dim] = dof; 1194 mat->stencil.starts[dim] = 0; 1195 mat->stencil.noc = (PetscTruth)(dof == 1); 1196 PetscFunctionReturn(0); 1197 } 1198 1199 #undef __FUNCT__ 1200 #define __FUNCT__ "MatSetValuesBlocked" 1201 /*@ 1202 MatSetValuesBlocked - Inserts or adds a block of values into a matrix. 1203 1204 Not Collective 1205 1206 Input Parameters: 1207 + mat - the matrix 1208 . v - a logically two-dimensional array of values 1209 . m, idxm - the number of block rows and their global block indices 1210 . n, idxn - the number of block columns and their global block indices 1211 - addv - either ADD_VALUES or INSERT_VALUES, where 1212 ADD_VALUES adds values to any existing entries, and 1213 INSERT_VALUES replaces existing entries with new values 1214 1215 Notes: 1216 The m and n count the NUMBER of blocks in the row direction and column direction, 1217 NOT the total number of rows/columns; for example, if the block size is 2 and 1218 you are passing in values for rows 2,3,4,5 then m would be 2 (not 4). 1219 1220 By default the values, v, are row-oriented and unsorted. So the layout of 1221 v is the same as for MatSetValues(). See MatSetOption() for other options. 1222 1223 Calls to MatSetValuesBlocked() with the INSERT_VALUES and ADD_VALUES 1224 options cannot be mixed without intervening calls to the assembly 1225 routines. 1226 1227 MatSetValuesBlocked() uses 0-based row and column numbers in Fortran 1228 as well as in C. 1229 1230 Negative indices may be passed in idxm and idxn, these rows and columns are 1231 simply ignored. This allows easily inserting element stiffness matrices 1232 with homogeneous Dirchlet boundary conditions that you don't want represented 1233 in the matrix. 1234 1235 Each time an entry is set within a sparse matrix via MatSetValues(), 1236 internal searching must be done to determine where to place the the 1237 data in the matrix storage space. By instead inserting blocks of 1238 entries via MatSetValuesBlocked(), the overhead of matrix assembly is 1239 reduced. 1240 1241 Example: 1242 $ Suppose m=n=2 and block size(bs) = 2 The matrix is 1243 $ 1244 $ 1 2 | 3 4 1245 $ 5 6 | 7 8 1246 $ - - - | - - - 1247 $ 9 10 | 11 12 1248 $ 13 14 | 15 16 1249 $ 1250 $ v[] should be passed in like 1251 $ v[] = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16] 1252 1253 Restrictions: 1254 MatSetValuesBlocked() is currently supported only for the BAIJ and SBAIJ formats 1255 1256 Level: intermediate 1257 1258 Concepts: matrices^putting entries in blocked 1259 1260 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal() 1261 @*/ 1262 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesBlocked(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv) 1263 { 1264 PetscErrorCode ierr; 1265 1266 PetscFunctionBegin; 1267 if (!m || !n) PetscFunctionReturn(0); /* no values to insert */ 1268 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1269 PetscValidType(mat,1); 1270 PetscValidIntPointer(idxm,3); 1271 PetscValidIntPointer(idxn,5); 1272 PetscValidScalarPointer(v,6); 1273 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1274 if (mat->insertmode == NOT_SET_VALUES) { 1275 mat->insertmode = addv; 1276 } 1277 #if defined(PETSC_USE_DEBUG) 1278 else if (mat->insertmode != addv) { 1279 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1280 } 1281 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1282 #endif 1283 1284 if (mat->assembled) { 1285 mat->was_assembled = PETSC_TRUE; 1286 mat->assembled = PETSC_FALSE; 1287 } 1288 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1289 if (!mat->ops->setvaluesblocked) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 1290 ierr = (*mat->ops->setvaluesblocked)(mat,m,idxm,n,idxn,v,addv);CHKERRQ(ierr); 1291 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1292 PetscFunctionReturn(0); 1293 } 1294 1295 #undef __FUNCT__ 1296 #define __FUNCT__ "MatGetValues" 1297 /*@ 1298 MatGetValues - Gets a block of values from a matrix. 1299 1300 Not Collective; currently only returns a local block 1301 1302 Input Parameters: 1303 + mat - the matrix 1304 . v - a logically two-dimensional array for storing the values 1305 . m, idxm - the number of rows and their global indices 1306 - n, idxn - the number of columns and their global indices 1307 1308 Notes: 1309 The user must allocate space (m*n PetscScalars) for the values, v. 1310 The values, v, are then returned in a row-oriented format, 1311 analogous to that used by default in MatSetValues(). 1312 1313 MatGetValues() uses 0-based row and column numbers in 1314 Fortran as well as in C. 1315 1316 MatGetValues() requires that the matrix has been assembled 1317 with MatAssemblyBegin()/MatAssemblyEnd(). Thus, calls to 1318 MatSetValues() and MatGetValues() CANNOT be made in succession 1319 without intermediate matrix assembly. 1320 1321 Negative row or column indices will be ignored and those locations in v[] will be 1322 left unchanged. 1323 1324 Level: advanced 1325 1326 Concepts: matrices^accessing values 1327 1328 .seealso: MatGetRow(), MatGetSubMatrices(), MatSetValues() 1329 @*/ 1330 PetscErrorCode PETSCMAT_DLLEXPORT MatGetValues(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[]) 1331 { 1332 PetscErrorCode ierr; 1333 1334 PetscFunctionBegin; 1335 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1336 PetscValidType(mat,1); 1337 PetscValidIntPointer(idxm,3); 1338 PetscValidIntPointer(idxn,5); 1339 PetscValidScalarPointer(v,6); 1340 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1341 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1342 if (!mat->ops->getvalues) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 1343 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1344 1345 ierr = PetscLogEventBegin(MAT_GetValues,mat,0,0,0);CHKERRQ(ierr); 1346 ierr = (*mat->ops->getvalues)(mat,m,idxm,n,idxn,v);CHKERRQ(ierr); 1347 ierr = PetscLogEventEnd(MAT_GetValues,mat,0,0,0);CHKERRQ(ierr); 1348 PetscFunctionReturn(0); 1349 } 1350 1351 #undef __FUNCT__ 1352 #define __FUNCT__ "MatSetLocalToGlobalMapping" 1353 /*@ 1354 MatSetLocalToGlobalMapping - Sets a local-to-global numbering for use by 1355 the routine MatSetValuesLocal() to allow users to insert matrix entries 1356 using a local (per-processor) numbering. 1357 1358 Not Collective 1359 1360 Input Parameters: 1361 + x - the matrix 1362 - mapping - mapping created with ISLocalToGlobalMappingCreate() 1363 or ISLocalToGlobalMappingCreateIS() 1364 1365 Level: intermediate 1366 1367 Concepts: matrices^local to global mapping 1368 Concepts: local to global mapping^for matrices 1369 1370 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesLocal() 1371 @*/ 1372 PetscErrorCode PETSCMAT_DLLEXPORT MatSetLocalToGlobalMapping(Mat x,ISLocalToGlobalMapping mapping) 1373 { 1374 PetscErrorCode ierr; 1375 PetscFunctionBegin; 1376 PetscValidHeaderSpecific(x,MAT_COOKIE,1); 1377 PetscValidType(x,1); 1378 PetscValidHeaderSpecific(mapping,IS_LTOGM_COOKIE,2); 1379 if (x->mapping) { 1380 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Mapping already set for matrix"); 1381 } 1382 ierr = MatPreallocated(x);CHKERRQ(ierr); 1383 1384 if (x->ops->setlocaltoglobalmapping) { 1385 ierr = (*x->ops->setlocaltoglobalmapping)(x,mapping);CHKERRQ(ierr); 1386 } else { 1387 ierr = PetscObjectReference((PetscObject)mapping);CHKERRQ(ierr); 1388 if (x->mapping) { ierr = ISLocalToGlobalMappingDestroy(x->mapping);CHKERRQ(ierr); } 1389 x->mapping = mapping; 1390 } 1391 PetscFunctionReturn(0); 1392 } 1393 1394 #undef __FUNCT__ 1395 #define __FUNCT__ "MatSetLocalToGlobalMappingBlock" 1396 /*@ 1397 MatSetLocalToGlobalMappingBlock - Sets a local-to-global numbering for use 1398 by the routine MatSetValuesBlockedLocal() to allow users to insert matrix 1399 entries using a local (per-processor) numbering. 1400 1401 Not Collective 1402 1403 Input Parameters: 1404 + x - the matrix 1405 - mapping - mapping created with ISLocalToGlobalMappingCreate() or 1406 ISLocalToGlobalMappingCreateIS() 1407 1408 Level: intermediate 1409 1410 Concepts: matrices^local to global mapping blocked 1411 Concepts: local to global mapping^for matrices, blocked 1412 1413 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetValuesBlockedLocal(), 1414 MatSetValuesBlocked(), MatSetValuesLocal() 1415 @*/ 1416 PetscErrorCode PETSCMAT_DLLEXPORT MatSetLocalToGlobalMappingBlock(Mat x,ISLocalToGlobalMapping mapping) 1417 { 1418 PetscErrorCode ierr; 1419 PetscFunctionBegin; 1420 PetscValidHeaderSpecific(x,MAT_COOKIE,1); 1421 PetscValidType(x,1); 1422 PetscValidHeaderSpecific(mapping,IS_LTOGM_COOKIE,2); 1423 if (x->bmapping) { 1424 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Mapping already set for matrix"); 1425 } 1426 ierr = PetscObjectReference((PetscObject)mapping);CHKERRQ(ierr); 1427 if (x->bmapping) { ierr = ISLocalToGlobalMappingDestroy(x->mapping);CHKERRQ(ierr); } 1428 x->bmapping = mapping; 1429 PetscFunctionReturn(0); 1430 } 1431 1432 #undef __FUNCT__ 1433 #define __FUNCT__ "MatSetValuesLocal" 1434 /*@ 1435 MatSetValuesLocal - Inserts or adds values into certain locations of a matrix, 1436 using a local ordering of the nodes. 1437 1438 Not Collective 1439 1440 Input Parameters: 1441 + x - the matrix 1442 . nrow, irow - number of rows and their local indices 1443 . ncol, icol - number of columns and their local indices 1444 . y - a logically two-dimensional array of values 1445 - addv - either INSERT_VALUES or ADD_VALUES, where 1446 ADD_VALUES adds values to any existing entries, and 1447 INSERT_VALUES replaces existing entries with new values 1448 1449 Notes: 1450 Before calling MatSetValuesLocal(), the user must first set the 1451 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 1452 1453 Calls to MatSetValuesLocal() with the INSERT_VALUES and ADD_VALUES 1454 options cannot be mixed without intervening calls to the assembly 1455 routines. 1456 1457 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 1458 MUST be called after all calls to MatSetValuesLocal() have been completed. 1459 1460 Level: intermediate 1461 1462 Concepts: matrices^putting entries in with local numbering 1463 1464 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValues(), MatSetLocalToGlobalMapping(), 1465 MatSetValueLocal() 1466 @*/ 1467 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv) 1468 { 1469 PetscErrorCode ierr; 1470 PetscInt irowm[2048],icolm[2048]; 1471 1472 PetscFunctionBegin; 1473 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1474 PetscValidType(mat,1); 1475 PetscValidIntPointer(irow,3); 1476 PetscValidIntPointer(icol,5); 1477 PetscValidScalarPointer(y,6); 1478 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1479 if (mat->insertmode == NOT_SET_VALUES) { 1480 mat->insertmode = addv; 1481 } 1482 #if defined(PETSC_USE_DEBUG) 1483 else if (mat->insertmode != addv) { 1484 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1485 } 1486 if (!mat->ops->setvalueslocal && (nrow > 2048 || ncol > 2048)) { 1487 SETERRQ2(PETSC_ERR_SUP,"Number column/row indices must be <= 2048: are %D %D",nrow,ncol); 1488 } 1489 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1490 #endif 1491 1492 if (mat->assembled) { 1493 mat->was_assembled = PETSC_TRUE; 1494 mat->assembled = PETSC_FALSE; 1495 } 1496 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1497 if (!mat->ops->setvalueslocal) { 1498 ierr = ISLocalToGlobalMappingApply(mat->mapping,nrow,irow,irowm);CHKERRQ(ierr); 1499 ierr = ISLocalToGlobalMappingApply(mat->mapping,ncol,icol,icolm);CHKERRQ(ierr); 1500 ierr = (*mat->ops->setvalues)(mat,nrow,irowm,ncol,icolm,y,addv);CHKERRQ(ierr); 1501 } else { 1502 ierr = (*mat->ops->setvalueslocal)(mat,nrow,irow,ncol,icol,y,addv);CHKERRQ(ierr); 1503 } 1504 mat->same_nonzero = PETSC_FALSE; 1505 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1506 PetscFunctionReturn(0); 1507 } 1508 1509 #undef __FUNCT__ 1510 #define __FUNCT__ "MatSetValuesBlockedLocal" 1511 /*@ 1512 MatSetValuesBlockedLocal - Inserts or adds values into certain locations of a matrix, 1513 using a local ordering of the nodes a block at a time. 1514 1515 Not Collective 1516 1517 Input Parameters: 1518 + x - the matrix 1519 . nrow, irow - number of rows and their local indices 1520 . ncol, icol - number of columns and their local indices 1521 . y - a logically two-dimensional array of values 1522 - addv - either INSERT_VALUES or ADD_VALUES, where 1523 ADD_VALUES adds values to any existing entries, and 1524 INSERT_VALUES replaces existing entries with new values 1525 1526 Notes: 1527 Before calling MatSetValuesBlockedLocal(), the user must first set the 1528 local-to-global mapping by calling MatSetLocalToGlobalMappingBlock(), 1529 where the mapping MUST be set for matrix blocks, not for matrix elements. 1530 1531 Calls to MatSetValuesBlockedLocal() with the INSERT_VALUES and ADD_VALUES 1532 options cannot be mixed without intervening calls to the assembly 1533 routines. 1534 1535 These values may be cached, so MatAssemblyBegin() and MatAssemblyEnd() 1536 MUST be called after all calls to MatSetValuesBlockedLocal() have been completed. 1537 1538 Level: intermediate 1539 1540 Concepts: matrices^putting blocked values in with local numbering 1541 1542 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesLocal(), MatSetLocalToGlobalMappingBlock(), MatSetValuesBlocked() 1543 @*/ 1544 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesBlockedLocal(Mat mat,PetscInt nrow,const PetscInt irow[],PetscInt ncol,const PetscInt icol[],const PetscScalar y[],InsertMode addv) 1545 { 1546 PetscErrorCode ierr; 1547 PetscInt irowm[2048],icolm[2048]; 1548 1549 PetscFunctionBegin; 1550 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1551 PetscValidType(mat,1); 1552 PetscValidIntPointer(irow,3); 1553 PetscValidIntPointer(icol,5); 1554 PetscValidScalarPointer(y,6); 1555 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1556 if (mat->insertmode == NOT_SET_VALUES) { 1557 mat->insertmode = addv; 1558 } 1559 #if defined(PETSC_USE_DEBUG) 1560 else if (mat->insertmode != addv) { 1561 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 1562 } 1563 if (!mat->bmapping) { 1564 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Local to global never set with MatSetLocalToGlobalMappingBlock()"); 1565 } 1566 if (nrow > 2048 || ncol > 2048) { 1567 SETERRQ2(PETSC_ERR_SUP,"Number column/row indices must be <= 2048: are %D %D",nrow,ncol); 1568 } 1569 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1570 #endif 1571 1572 if (mat->assembled) { 1573 mat->was_assembled = PETSC_TRUE; 1574 mat->assembled = PETSC_FALSE; 1575 } 1576 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1577 ierr = ISLocalToGlobalMappingApply(mat->bmapping,nrow,irow,irowm);CHKERRQ(ierr); 1578 ierr = ISLocalToGlobalMappingApply(mat->bmapping,ncol,icol,icolm);CHKERRQ(ierr); 1579 ierr = (*mat->ops->setvaluesblocked)(mat,nrow,irowm,ncol,icolm,y,addv);CHKERRQ(ierr); 1580 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 1581 PetscFunctionReturn(0); 1582 } 1583 1584 /* --------------------------------------------------------*/ 1585 #undef __FUNCT__ 1586 #define __FUNCT__ "MatMult" 1587 /*@ 1588 MatMult - Computes the matrix-vector product, y = Ax. 1589 1590 Collective on Mat and Vec 1591 1592 Input Parameters: 1593 + mat - the matrix 1594 - x - the vector to be multiplied 1595 1596 Output Parameters: 1597 . y - the result 1598 1599 Notes: 1600 The vectors x and y cannot be the same. I.e., one cannot 1601 call MatMult(A,y,y). 1602 1603 Level: beginner 1604 1605 Concepts: matrix-vector product 1606 1607 .seealso: MatMultTranspose(), MatMultAdd(), MatMultTransposeAdd() 1608 @*/ 1609 PetscErrorCode PETSCMAT_DLLEXPORT MatMult(Mat mat,Vec x,Vec y) 1610 { 1611 PetscErrorCode ierr; 1612 1613 PetscFunctionBegin; 1614 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1615 PetscValidType(mat,1); 1616 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 1617 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 1618 1619 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1620 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1621 if (x == y) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 1622 #ifndef PETSC_HAVE_CONSTRAINTS 1623 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 1624 if (mat->rmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap.N,y->map.N); 1625 if (mat->rmap.n != y->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: local dim %D %D",mat->rmap.n,y->map.n); 1626 #endif 1627 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1628 1629 if (mat->nullsp) { 1630 ierr = MatNullSpaceRemove(mat->nullsp,x,&x);CHKERRQ(ierr); 1631 } 1632 1633 if (!mat->ops->mult) SETERRQ(PETSC_ERR_SUP,"This matrix type does not have a multiply defined"); 1634 ierr = PetscLogEventBegin(MAT_Mult,mat,x,y,0);CHKERRQ(ierr); 1635 ierr = (*mat->ops->mult)(mat,x,y);CHKERRQ(ierr); 1636 ierr = PetscLogEventEnd(MAT_Mult,mat,x,y,0);CHKERRQ(ierr); 1637 1638 if (mat->nullsp) { 1639 ierr = MatNullSpaceRemove(mat->nullsp,y,PETSC_NULL);CHKERRQ(ierr); 1640 } 1641 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 1642 PetscFunctionReturn(0); 1643 } 1644 1645 #undef __FUNCT__ 1646 #define __FUNCT__ "MatMultTranspose" 1647 /*@ 1648 MatMultTranspose - Computes matrix transpose times a vector. 1649 1650 Collective on Mat and Vec 1651 1652 Input Parameters: 1653 + mat - the matrix 1654 - x - the vector to be multilplied 1655 1656 Output Parameters: 1657 . y - the result 1658 1659 Notes: 1660 The vectors x and y cannot be the same. I.e., one cannot 1661 call MatMultTranspose(A,y,y). 1662 1663 Level: beginner 1664 1665 Concepts: matrix vector product^transpose 1666 1667 .seealso: MatMult(), MatMultAdd(), MatMultTransposeAdd() 1668 @*/ 1669 PetscErrorCode PETSCMAT_DLLEXPORT MatMultTranspose(Mat mat,Vec x,Vec y) 1670 { 1671 PetscErrorCode ierr; 1672 1673 PetscFunctionBegin; 1674 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1675 PetscValidType(mat,1); 1676 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 1677 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 1678 1679 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1680 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1681 if (x == y) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 1682 #ifndef PETSC_HAVE_CONSTRAINTS 1683 if (mat->rmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap.N,x->map.N); 1684 if (mat->cmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap.N,y->map.N); 1685 #endif 1686 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1687 1688 if (!mat->ops->multtranspose) SETERRQ(PETSC_ERR_SUP,"This matrix type does not have a multiply tranpose defined"); 1689 ierr = PetscLogEventBegin(MAT_MultTranspose,mat,x,y,0);CHKERRQ(ierr); 1690 ierr = (*mat->ops->multtranspose)(mat,x,y);CHKERRQ(ierr); 1691 ierr = PetscLogEventEnd(MAT_MultTranspose,mat,x,y,0);CHKERRQ(ierr); 1692 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 1693 PetscFunctionReturn(0); 1694 } 1695 1696 #undef __FUNCT__ 1697 #define __FUNCT__ "MatMultAdd" 1698 /*@ 1699 MatMultAdd - Computes v3 = v2 + A * v1. 1700 1701 Collective on Mat and Vec 1702 1703 Input Parameters: 1704 + mat - the matrix 1705 - v1, v2 - the vectors 1706 1707 Output Parameters: 1708 . v3 - the result 1709 1710 Notes: 1711 The vectors v1 and v3 cannot be the same. I.e., one cannot 1712 call MatMultAdd(A,v1,v2,v1). 1713 1714 Level: beginner 1715 1716 Concepts: matrix vector product^addition 1717 1718 .seealso: MatMultTranspose(), MatMult(), MatMultTransposeAdd() 1719 @*/ 1720 PetscErrorCode PETSCMAT_DLLEXPORT MatMultAdd(Mat mat,Vec v1,Vec v2,Vec v3) 1721 { 1722 PetscErrorCode ierr; 1723 1724 PetscFunctionBegin; 1725 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1726 PetscValidType(mat,1); 1727 PetscValidHeaderSpecific(v1,VEC_COOKIE,2); 1728 PetscValidHeaderSpecific(v2,VEC_COOKIE,3); 1729 PetscValidHeaderSpecific(v3,VEC_COOKIE,4); 1730 1731 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1732 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1733 if (mat->cmap.N != v1->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->cmap.N,v1->map.N); 1734 if (mat->rmap.N != v2->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->rmap.N,v2->map.N); 1735 if (mat->rmap.N != v3->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->rmap.N,v3->map.N); 1736 if (mat->rmap.n != v3->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: local dim %D %D",mat->rmap.n,v3->map.n); 1737 if (mat->rmap.n != v2->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: local dim %D %D",mat->rmap.n,v2->map.n); 1738 if (v1 == v3) SETERRQ(PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors"); 1739 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1740 1741 ierr = PetscLogEventBegin(MAT_MultAdd,mat,v1,v2,v3);CHKERRQ(ierr); 1742 ierr = (*mat->ops->multadd)(mat,v1,v2,v3);CHKERRQ(ierr); 1743 ierr = PetscLogEventEnd(MAT_MultAdd,mat,v1,v2,v3);CHKERRQ(ierr); 1744 ierr = PetscObjectStateIncrease((PetscObject)v3);CHKERRQ(ierr); 1745 PetscFunctionReturn(0); 1746 } 1747 1748 #undef __FUNCT__ 1749 #define __FUNCT__ "MatMultTransposeAdd" 1750 /*@ 1751 MatMultTransposeAdd - Computes v3 = v2 + A' * v1. 1752 1753 Collective on Mat and Vec 1754 1755 Input Parameters: 1756 + mat - the matrix 1757 - v1, v2 - the vectors 1758 1759 Output Parameters: 1760 . v3 - the result 1761 1762 Notes: 1763 The vectors v1 and v3 cannot be the same. I.e., one cannot 1764 call MatMultTransposeAdd(A,v1,v2,v1). 1765 1766 Level: beginner 1767 1768 Concepts: matrix vector product^transpose and addition 1769 1770 .seealso: MatMultTranspose(), MatMultAdd(), MatMult() 1771 @*/ 1772 PetscErrorCode PETSCMAT_DLLEXPORT MatMultTransposeAdd(Mat mat,Vec v1,Vec v2,Vec v3) 1773 { 1774 PetscErrorCode ierr; 1775 1776 PetscFunctionBegin; 1777 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1778 PetscValidType(mat,1); 1779 PetscValidHeaderSpecific(v1,VEC_COOKIE,2); 1780 PetscValidHeaderSpecific(v2,VEC_COOKIE,3); 1781 PetscValidHeaderSpecific(v3,VEC_COOKIE,4); 1782 1783 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1784 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1785 if (!mat->ops->multtransposeadd) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 1786 if (v1 == v3) SETERRQ(PETSC_ERR_ARG_IDN,"v1 and v3 must be different vectors"); 1787 if (mat->rmap.N != v1->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v1: global dim %D %D",mat->rmap.N,v1->map.N); 1788 if (mat->cmap.N != v2->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v2: global dim %D %D",mat->cmap.N,v2->map.N); 1789 if (mat->cmap.N != v3->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec v3: global dim %D %D",mat->cmap.N,v3->map.N); 1790 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1791 1792 ierr = PetscLogEventBegin(MAT_MultTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 1793 ierr = (*mat->ops->multtransposeadd)(mat,v1,v2,v3);CHKERRQ(ierr); 1794 ierr = PetscLogEventEnd(MAT_MultTransposeAdd,mat,v1,v2,v3);CHKERRQ(ierr); 1795 ierr = PetscObjectStateIncrease((PetscObject)v3);CHKERRQ(ierr); 1796 PetscFunctionReturn(0); 1797 } 1798 1799 #undef __FUNCT__ 1800 #define __FUNCT__ "MatMultConstrained" 1801 /*@ 1802 MatMultConstrained - The inner multiplication routine for a 1803 constrained matrix P^T A P. 1804 1805 Collective on Mat and Vec 1806 1807 Input Parameters: 1808 + mat - the matrix 1809 - x - the vector to be multilplied 1810 1811 Output Parameters: 1812 . y - the result 1813 1814 Notes: 1815 The vectors x and y cannot be the same. I.e., one cannot 1816 call MatMult(A,y,y). 1817 1818 Level: beginner 1819 1820 .keywords: matrix, multiply, matrix-vector product, constraint 1821 .seealso: MatMult(), MatMultTrans(), MatMultAdd(), MatMultTransAdd() 1822 @*/ 1823 PetscErrorCode PETSCMAT_DLLEXPORT MatMultConstrained(Mat mat,Vec x,Vec y) 1824 { 1825 PetscErrorCode ierr; 1826 1827 PetscFunctionBegin; 1828 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1829 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 1830 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 1831 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1832 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1833 if (x == y) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 1834 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 1835 if (mat->rmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap.N,y->map.N); 1836 if (mat->rmap.n != y->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: local dim %D %D",mat->rmap.n,y->map.n); 1837 1838 ierr = PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 1839 ierr = (*mat->ops->multconstrained)(mat,x,y);CHKERRQ(ierr); 1840 ierr = PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 1841 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 1842 1843 PetscFunctionReturn(0); 1844 } 1845 1846 #undef __FUNCT__ 1847 #define __FUNCT__ "MatMultTransposeConstrained" 1848 /*@ 1849 MatMultTransposeConstrained - The inner multiplication routine for a 1850 constrained matrix P^T A^T P. 1851 1852 Collective on Mat and Vec 1853 1854 Input Parameters: 1855 + mat - the matrix 1856 - x - the vector to be multilplied 1857 1858 Output Parameters: 1859 . y - the result 1860 1861 Notes: 1862 The vectors x and y cannot be the same. I.e., one cannot 1863 call MatMult(A,y,y). 1864 1865 Level: beginner 1866 1867 .keywords: matrix, multiply, matrix-vector product, constraint 1868 .seealso: MatMult(), MatMultTrans(), MatMultAdd(), MatMultTransAdd() 1869 @*/ 1870 PetscErrorCode PETSCMAT_DLLEXPORT MatMultTransposeConstrained(Mat mat,Vec x,Vec y) 1871 { 1872 PetscErrorCode ierr; 1873 1874 PetscFunctionBegin; 1875 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1876 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 1877 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 1878 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 1879 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 1880 if (x == y) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"x and y must be different vectors"); 1881 if (mat->rmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 1882 if (mat->cmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap.N,y->map.N); 1883 1884 ierr = PetscLogEventBegin(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 1885 ierr = (*mat->ops->multtransposeconstrained)(mat,x,y);CHKERRQ(ierr); 1886 ierr = PetscLogEventEnd(MAT_MultConstrained,mat,x,y,0);CHKERRQ(ierr); 1887 ierr = PetscObjectStateIncrease((PetscObject)y);CHKERRQ(ierr); 1888 1889 PetscFunctionReturn(0); 1890 } 1891 /* ------------------------------------------------------------*/ 1892 #undef __FUNCT__ 1893 #define __FUNCT__ "MatGetInfo" 1894 /*@ 1895 MatGetInfo - Returns information about matrix storage (number of 1896 nonzeros, memory, etc.). 1897 1898 Collective on Mat if MAT_GLOBAL_MAX or MAT_GLOBAL_SUM is used 1899 as the flag 1900 1901 Input Parameters: 1902 . mat - the matrix 1903 1904 Output Parameters: 1905 + flag - flag indicating the type of parameters to be returned 1906 (MAT_LOCAL - local matrix, MAT_GLOBAL_MAX - maximum over all processors, 1907 MAT_GLOBAL_SUM - sum over all processors) 1908 - info - matrix information context 1909 1910 Notes: 1911 The MatInfo context contains a variety of matrix data, including 1912 number of nonzeros allocated and used, number of mallocs during 1913 matrix assembly, etc. Additional information for factored matrices 1914 is provided (such as the fill ratio, number of mallocs during 1915 factorization, etc.). Much of this info is printed to PETSC_STDOUT 1916 when using the runtime options 1917 $ -info -mat_view_info 1918 1919 Example for C/C++ Users: 1920 See the file ${PETSC_DIR}/include/petscmat.h for a complete list of 1921 data within the MatInfo context. For example, 1922 .vb 1923 MatInfo info; 1924 Mat A; 1925 double mal, nz_a, nz_u; 1926 1927 MatGetInfo(A,MAT_LOCAL,&info); 1928 mal = info.mallocs; 1929 nz_a = info.nz_allocated; 1930 .ve 1931 1932 Example for Fortran Users: 1933 Fortran users should declare info as a double precision 1934 array of dimension MAT_INFO_SIZE, and then extract the parameters 1935 of interest. See the file ${PETSC_DIR}/include/finclude/petscmat.h 1936 a complete list of parameter names. 1937 .vb 1938 double precision info(MAT_INFO_SIZE) 1939 double precision mal, nz_a 1940 Mat A 1941 integer ierr 1942 1943 call MatGetInfo(A,MAT_LOCAL,info,ierr) 1944 mal = info(MAT_INFO_MALLOCS) 1945 nz_a = info(MAT_INFO_NZ_ALLOCATED) 1946 .ve 1947 1948 Level: intermediate 1949 1950 Concepts: matrices^getting information on 1951 1952 @*/ 1953 PetscErrorCode PETSCMAT_DLLEXPORT MatGetInfo(Mat mat,MatInfoType flag,MatInfo *info) 1954 { 1955 PetscErrorCode ierr; 1956 1957 PetscFunctionBegin; 1958 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 1959 PetscValidType(mat,1); 1960 PetscValidPointer(info,3); 1961 if (!mat->ops->getinfo) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 1962 ierr = MatPreallocated(mat);CHKERRQ(ierr); 1963 ierr = (*mat->ops->getinfo)(mat,flag,info);CHKERRQ(ierr); 1964 PetscFunctionReturn(0); 1965 } 1966 1967 /* ----------------------------------------------------------*/ 1968 #undef __FUNCT__ 1969 #define __FUNCT__ "MatILUDTFactor" 1970 /*@C 1971 MatILUDTFactor - Performs a drop tolerance ILU factorization. 1972 1973 Collective on Mat 1974 1975 Input Parameters: 1976 + mat - the matrix 1977 . row - row permutation 1978 . col - column permutation 1979 - info - information about the factorization to be done 1980 1981 Output Parameters: 1982 . fact - the factored matrix 1983 1984 Level: developer 1985 1986 Notes: 1987 Most users should employ the simplified KSP interface for linear solvers 1988 instead of working directly with matrix algebra routines such as this. 1989 See, e.g., KSPCreate(). 1990 1991 This is currently only supported for the SeqAIJ matrix format using code 1992 from Yousef Saad's SPARSEKIT2 package (translated to C with f2c) and/or 1993 Matlab. SPARSEKIT2 is copyrighted by Yousef Saad with the GNU copyright 1994 and thus can be distributed with PETSc. 1995 1996 Concepts: matrices^ILUDT factorization 1997 1998 .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 1999 @*/ 2000 PetscErrorCode PETSCMAT_DLLEXPORT MatILUDTFactor(Mat mat,IS row,IS col,MatFactorInfo *info,Mat *fact) 2001 { 2002 PetscErrorCode ierr; 2003 2004 PetscFunctionBegin; 2005 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2006 PetscValidType(mat,1); 2007 if (row) PetscValidHeaderSpecific(row,IS_COOKIE,2); 2008 if (col) PetscValidHeaderSpecific(col,IS_COOKIE,3); 2009 PetscValidPointer(info,4); 2010 PetscValidPointer(fact,5); 2011 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2012 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2013 if (!mat->ops->iludtfactor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2014 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2015 ierr = PetscLogEventBegin(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2016 ierr = (*mat->ops->iludtfactor)(mat,row,col,info,fact);CHKERRQ(ierr); 2017 ierr = PetscLogEventEnd(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2018 ierr = PetscObjectStateIncrease((PetscObject)*fact);CHKERRQ(ierr); 2019 2020 PetscFunctionReturn(0); 2021 } 2022 2023 #undef __FUNCT__ 2024 #define __FUNCT__ "MatLUFactor" 2025 /*@ 2026 MatLUFactor - Performs in-place LU factorization of matrix. 2027 2028 Collective on Mat 2029 2030 Input Parameters: 2031 + mat - the matrix 2032 . row - row permutation 2033 . col - column permutation 2034 - info - options for factorization, includes 2035 $ fill - expected fill as ratio of original fill. 2036 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2037 $ Run with the option -info to determine an optimal value to use 2038 2039 Notes: 2040 Most users should employ the simplified KSP interface for linear solvers 2041 instead of working directly with matrix algebra routines such as this. 2042 See, e.g., KSPCreate(). 2043 2044 This changes the state of the matrix to a factored matrix; it cannot be used 2045 for example with MatSetValues() unless one first calls MatSetUnfactored(). 2046 2047 Level: developer 2048 2049 Concepts: matrices^LU factorization 2050 2051 .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), 2052 MatGetOrdering(), MatSetUnfactored(), MatFactorInfo 2053 2054 @*/ 2055 PetscErrorCode PETSCMAT_DLLEXPORT MatLUFactor(Mat mat,IS row,IS col,MatFactorInfo *info) 2056 { 2057 PetscErrorCode ierr; 2058 2059 PetscFunctionBegin; 2060 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2061 if (row) PetscValidHeaderSpecific(row,IS_COOKIE,2); 2062 if (col) PetscValidHeaderSpecific(col,IS_COOKIE,3); 2063 PetscValidPointer(info,4); 2064 PetscValidType(mat,1); 2065 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2066 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2067 if (!mat->ops->lufactor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2068 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2069 2070 ierr = PetscLogEventBegin(MAT_LUFactor,mat,row,col,0);CHKERRQ(ierr); 2071 ierr = (*mat->ops->lufactor)(mat,row,col,info);CHKERRQ(ierr); 2072 ierr = PetscLogEventEnd(MAT_LUFactor,mat,row,col,0);CHKERRQ(ierr); 2073 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2074 PetscFunctionReturn(0); 2075 } 2076 2077 #undef __FUNCT__ 2078 #define __FUNCT__ "MatILUFactor" 2079 /*@ 2080 MatILUFactor - Performs in-place ILU factorization of matrix. 2081 2082 Collective on Mat 2083 2084 Input Parameters: 2085 + mat - the matrix 2086 . row - row permutation 2087 . col - column permutation 2088 - info - structure containing 2089 $ levels - number of levels of fill. 2090 $ expected fill - as ratio of original fill. 2091 $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices 2092 missing diagonal entries) 2093 2094 Notes: 2095 Probably really in-place only when level of fill is zero, otherwise allocates 2096 new space to store factored matrix and deletes previous memory. 2097 2098 Most users should employ the simplified KSP interface for linear solvers 2099 instead of working directly with matrix algebra routines such as this. 2100 See, e.g., KSPCreate(). 2101 2102 Level: developer 2103 2104 Concepts: matrices^ILU factorization 2105 2106 .seealso: MatILUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 2107 @*/ 2108 PetscErrorCode PETSCMAT_DLLEXPORT MatILUFactor(Mat mat,IS row,IS col,MatFactorInfo *info) 2109 { 2110 PetscErrorCode ierr; 2111 2112 PetscFunctionBegin; 2113 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2114 if (row) PetscValidHeaderSpecific(row,IS_COOKIE,2); 2115 if (col) PetscValidHeaderSpecific(col,IS_COOKIE,3); 2116 PetscValidPointer(info,4); 2117 PetscValidType(mat,1); 2118 if (mat->rmap.N != mat->cmap.N) SETERRQ(PETSC_ERR_ARG_WRONG,"matrix must be square"); 2119 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2120 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2121 if (!mat->ops->ilufactor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2122 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2123 2124 ierr = PetscLogEventBegin(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2125 ierr = (*mat->ops->ilufactor)(mat,row,col,info);CHKERRQ(ierr); 2126 ierr = PetscLogEventEnd(MAT_ILUFactor,mat,row,col,0);CHKERRQ(ierr); 2127 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2128 PetscFunctionReturn(0); 2129 } 2130 2131 #undef __FUNCT__ 2132 #define __FUNCT__ "MatLUFactorSymbolic" 2133 /*@ 2134 MatLUFactorSymbolic - Performs symbolic LU factorization of matrix. 2135 Call this routine before calling MatLUFactorNumeric(). 2136 2137 Collective on Mat 2138 2139 Input Parameters: 2140 + mat - the matrix 2141 . row, col - row and column permutations 2142 - info - options for factorization, includes 2143 $ fill - expected fill as ratio of original fill. 2144 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2145 $ Run with the option -info to determine an optimal value to use 2146 2147 Output Parameter: 2148 . fact - new matrix that has been symbolically factored 2149 2150 Notes: 2151 See the users manual for additional information about 2152 choosing the fill factor for better efficiency. 2153 2154 Most users should employ the simplified KSP interface for linear solvers 2155 instead of working directly with matrix algebra routines such as this. 2156 See, e.g., KSPCreate(). 2157 2158 Level: developer 2159 2160 Concepts: matrices^LU symbolic factorization 2161 2162 .seealso: MatLUFactor(), MatLUFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 2163 @*/ 2164 PetscErrorCode PETSCMAT_DLLEXPORT MatLUFactorSymbolic(Mat mat,IS row,IS col,MatFactorInfo *info,Mat *fact) 2165 { 2166 PetscErrorCode ierr; 2167 2168 PetscFunctionBegin; 2169 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2170 if (row) PetscValidHeaderSpecific(row,IS_COOKIE,2); 2171 if (col) PetscValidHeaderSpecific(col,IS_COOKIE,3); 2172 PetscValidPointer(info,4); 2173 PetscValidType(mat,1); 2174 PetscValidPointer(fact,5); 2175 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2176 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2177 if (!mat->ops->lufactorsymbolic) SETERRQ1(PETSC_ERR_SUP,"Matrix type %s symbolic LU",mat->type_name); 2178 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2179 2180 ierr = PetscLogEventBegin(MAT_LUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 2181 ierr = (*mat->ops->lufactorsymbolic)(mat,row,col,info,fact);CHKERRQ(ierr); 2182 ierr = PetscLogEventEnd(MAT_LUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 2183 ierr = PetscObjectStateIncrease((PetscObject)*fact);CHKERRQ(ierr); 2184 PetscFunctionReturn(0); 2185 } 2186 2187 #undef __FUNCT__ 2188 #define __FUNCT__ "MatLUFactorNumeric" 2189 /*@ 2190 MatLUFactorNumeric - Performs numeric LU factorization of a matrix. 2191 Call this routine after first calling MatLUFactorSymbolic(). 2192 2193 Collective on Mat 2194 2195 Input Parameters: 2196 + mat - the matrix 2197 . info - options for factorization 2198 - fact - the matrix generated for the factor, from MatLUFactorSymbolic() 2199 2200 Notes: 2201 See MatLUFactor() for in-place factorization. See 2202 MatCholeskyFactorNumeric() for the symmetric, positive definite case. 2203 2204 Most users should employ the simplified KSP interface for linear solvers 2205 instead of working directly with matrix algebra routines such as this. 2206 See, e.g., KSPCreate(). 2207 2208 Level: developer 2209 2210 Concepts: matrices^LU numeric factorization 2211 2212 .seealso: MatLUFactorSymbolic(), MatLUFactor(), MatCholeskyFactor() 2213 @*/ 2214 PetscErrorCode PETSCMAT_DLLEXPORT MatLUFactorNumeric(Mat mat,MatFactorInfo *info,Mat *fact) 2215 { 2216 PetscErrorCode ierr; 2217 2218 PetscFunctionBegin; 2219 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2220 PetscValidType(mat,1); 2221 PetscValidPointer(fact,2); 2222 PetscValidHeaderSpecific(*fact,MAT_COOKIE,2); 2223 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2224 if (mat->rmap.N != (*fact)->rmap.N || mat->cmap.N != (*fact)->cmap.N) { 2225 SETERRQ4(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); 2226 } 2227 if (!(*fact)->ops->lufactornumeric) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2228 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2229 ierr = PetscLogEventBegin(MAT_LUFactorNumeric,mat,*fact,0,0);CHKERRQ(ierr); 2230 ierr = (*(*fact)->ops->lufactornumeric)(mat,info,fact);CHKERRQ(ierr); 2231 ierr = PetscLogEventEnd(MAT_LUFactorNumeric,mat,*fact,0,0);CHKERRQ(ierr); 2232 2233 ierr = MatView_Private(*fact);CHKERRQ(ierr); 2234 ierr = PetscObjectStateIncrease((PetscObject)*fact);CHKERRQ(ierr); 2235 PetscFunctionReturn(0); 2236 } 2237 2238 #undef __FUNCT__ 2239 #define __FUNCT__ "MatCholeskyFactor" 2240 /*@ 2241 MatCholeskyFactor - Performs in-place Cholesky factorization of a 2242 symmetric matrix. 2243 2244 Collective on Mat 2245 2246 Input Parameters: 2247 + mat - the matrix 2248 . perm - row and column permutations 2249 - f - expected fill as ratio of original fill 2250 2251 Notes: 2252 See MatLUFactor() for the nonsymmetric case. See also 2253 MatCholeskyFactorSymbolic(), and MatCholeskyFactorNumeric(). 2254 2255 Most users should employ the simplified KSP interface for linear solvers 2256 instead of working directly with matrix algebra routines such as this. 2257 See, e.g., KSPCreate(). 2258 2259 Level: developer 2260 2261 Concepts: matrices^Cholesky factorization 2262 2263 .seealso: MatLUFactor(), MatCholeskyFactorSymbolic(), MatCholeskyFactorNumeric() 2264 MatGetOrdering() 2265 2266 @*/ 2267 PetscErrorCode PETSCMAT_DLLEXPORT MatCholeskyFactor(Mat mat,IS perm,MatFactorInfo *info) 2268 { 2269 PetscErrorCode ierr; 2270 2271 PetscFunctionBegin; 2272 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2273 PetscValidType(mat,1); 2274 PetscValidHeaderSpecific(perm,IS_COOKIE,2); 2275 PetscValidPointer(info,3); 2276 if (mat->rmap.N != mat->cmap.N) SETERRQ(PETSC_ERR_ARG_WRONG,"Matrix must be square"); 2277 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2278 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2279 if (!mat->ops->choleskyfactor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2280 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2281 2282 ierr = PetscLogEventBegin(MAT_CholeskyFactor,mat,perm,0,0);CHKERRQ(ierr); 2283 ierr = (*mat->ops->choleskyfactor)(mat,perm,info);CHKERRQ(ierr); 2284 ierr = PetscLogEventEnd(MAT_CholeskyFactor,mat,perm,0,0);CHKERRQ(ierr); 2285 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 2286 PetscFunctionReturn(0); 2287 } 2288 2289 #undef __FUNCT__ 2290 #define __FUNCT__ "MatCholeskyFactorSymbolic" 2291 /*@ 2292 MatCholeskyFactorSymbolic - Performs symbolic Cholesky factorization 2293 of a symmetric matrix. 2294 2295 Collective on Mat 2296 2297 Input Parameters: 2298 + mat - the matrix 2299 . perm - row and column permutations 2300 - info - options for factorization, includes 2301 $ fill - expected fill as ratio of original fill. 2302 $ dtcol - pivot tolerance (0 no pivot, 1 full column pivoting) 2303 $ Run with the option -info to determine an optimal value to use 2304 2305 Output Parameter: 2306 . fact - the factored matrix 2307 2308 Notes: 2309 See MatLUFactorSymbolic() for the nonsymmetric case. See also 2310 MatCholeskyFactor() and MatCholeskyFactorNumeric(). 2311 2312 Most users should employ the simplified KSP interface for linear solvers 2313 instead of working directly with matrix algebra routines such as this. 2314 See, e.g., KSPCreate(). 2315 2316 Level: developer 2317 2318 Concepts: matrices^Cholesky symbolic factorization 2319 2320 .seealso: MatLUFactorSymbolic(), MatCholeskyFactor(), MatCholeskyFactorNumeric() 2321 MatGetOrdering() 2322 2323 @*/ 2324 PetscErrorCode PETSCMAT_DLLEXPORT MatCholeskyFactorSymbolic(Mat mat,IS perm,MatFactorInfo *info,Mat *fact) 2325 { 2326 PetscErrorCode ierr; 2327 2328 PetscFunctionBegin; 2329 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2330 PetscValidType(mat,1); 2331 if (perm) PetscValidHeaderSpecific(perm,IS_COOKIE,2); 2332 PetscValidPointer(info,3); 2333 PetscValidPointer(fact,4); 2334 if (mat->rmap.N != mat->cmap.N) SETERRQ(PETSC_ERR_ARG_WRONG,"Matrix must be square"); 2335 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2336 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2337 if (!mat->ops->choleskyfactorsymbolic) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2338 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2339 2340 ierr = PetscLogEventBegin(MAT_CholeskyFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 2341 ierr = (*mat->ops->choleskyfactorsymbolic)(mat,perm,info,fact);CHKERRQ(ierr); 2342 ierr = PetscLogEventEnd(MAT_CholeskyFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 2343 ierr = PetscObjectStateIncrease((PetscObject)*fact);CHKERRQ(ierr); 2344 PetscFunctionReturn(0); 2345 } 2346 2347 #undef __FUNCT__ 2348 #define __FUNCT__ "MatCholeskyFactorNumeric" 2349 /*@ 2350 MatCholeskyFactorNumeric - Performs numeric Cholesky factorization 2351 of a symmetric matrix. Call this routine after first calling 2352 MatCholeskyFactorSymbolic(). 2353 2354 Collective on Mat 2355 2356 Input Parameter: 2357 . mat - the initial matrix 2358 . info - options for factorization 2359 - fact - the symbolic factor of mat 2360 2361 Output Parameter: 2362 . fact - the factored matrix 2363 2364 Notes: 2365 Most users should employ the simplified KSP interface for linear solvers 2366 instead of working directly with matrix algebra routines such as this. 2367 See, e.g., KSPCreate(). 2368 2369 Level: developer 2370 2371 Concepts: matrices^Cholesky numeric factorization 2372 2373 .seealso: MatCholeskyFactorSymbolic(), MatCholeskyFactor(), MatLUFactorNumeric() 2374 @*/ 2375 PetscErrorCode PETSCMAT_DLLEXPORT MatCholeskyFactorNumeric(Mat mat,MatFactorInfo *info,Mat *fact) 2376 { 2377 PetscErrorCode ierr; 2378 2379 PetscFunctionBegin; 2380 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2381 PetscValidType(mat,1); 2382 PetscValidPointer(fact,2); 2383 PetscValidHeaderSpecific(*fact,MAT_COOKIE,2); 2384 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2385 if (!(*fact)->ops->choleskyfactornumeric) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2386 if (mat->rmap.N != (*fact)->rmap.N || mat->cmap.N != (*fact)->cmap.N) { 2387 SETERRQ4(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); 2388 } 2389 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2390 2391 ierr = PetscLogEventBegin(MAT_CholeskyFactorNumeric,mat,*fact,0,0);CHKERRQ(ierr); 2392 ierr = (*(*fact)->ops->choleskyfactornumeric)(mat,info,fact);CHKERRQ(ierr); 2393 ierr = PetscLogEventEnd(MAT_CholeskyFactorNumeric,mat,*fact,0,0);CHKERRQ(ierr); 2394 2395 ierr = MatView_Private(*fact);CHKERRQ(ierr); 2396 ierr = PetscObjectStateIncrease((PetscObject)*fact);CHKERRQ(ierr); 2397 PetscFunctionReturn(0); 2398 } 2399 2400 /* ----------------------------------------------------------------*/ 2401 #undef __FUNCT__ 2402 #define __FUNCT__ "MatSolve" 2403 /*@ 2404 MatSolve - Solves A x = b, given a factored matrix. 2405 2406 Collective on Mat and Vec 2407 2408 Input Parameters: 2409 + mat - the factored matrix 2410 - b - the right-hand-side vector 2411 2412 Output Parameter: 2413 . x - the result vector 2414 2415 Notes: 2416 The vectors b and x cannot be the same. I.e., one cannot 2417 call MatSolve(A,x,x). 2418 2419 Notes: 2420 Most users should employ the simplified KSP interface for linear solvers 2421 instead of working directly with matrix algebra routines such as this. 2422 See, e.g., KSPCreate(). 2423 2424 Level: developer 2425 2426 Concepts: matrices^triangular solves 2427 2428 .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd() 2429 @*/ 2430 PetscErrorCode PETSCMAT_DLLEXPORT MatSolve(Mat mat,Vec b,Vec x) 2431 { 2432 PetscErrorCode ierr; 2433 2434 PetscFunctionBegin; 2435 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2436 PetscValidType(mat,1); 2437 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2438 PetscValidHeaderSpecific(x,VEC_COOKIE,3); 2439 PetscCheckSameComm(mat,1,b,2); 2440 PetscCheckSameComm(mat,1,x,3); 2441 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2442 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2443 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2444 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2445 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2446 if (!mat->rmap.N && !mat->cmap.N) PetscFunctionReturn(0); 2447 if (!mat->ops->solve) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2448 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2449 2450 ierr = PetscLogEventBegin(MAT_Solve,mat,b,x,0);CHKERRQ(ierr); 2451 ierr = (*mat->ops->solve)(mat,b,x);CHKERRQ(ierr); 2452 ierr = PetscLogEventEnd(MAT_Solve,mat,b,x,0);CHKERRQ(ierr); 2453 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2454 PetscFunctionReturn(0); 2455 } 2456 2457 #undef __FUNCT__ 2458 #define __FUNCT__ "MatMatSolve" 2459 /*@ 2460 MatMatSolve - Solves A X = B, given a factored matrix. 2461 2462 Collective on Mat 2463 2464 Input Parameters: 2465 + mat - the factored matrix 2466 - b - the right-hand-side vector 2467 2468 Output Parameter: 2469 . x - the result vector 2470 2471 Notes: 2472 The vectors b and x cannot be the same. I.e., one cannot 2473 call MatMatSolve(A,x,x). 2474 2475 Notes: 2476 Most users should employ the simplified KSP interface for linear solvers 2477 instead of working directly with matrix algebra routines such as this. 2478 See, e.g., KSPCreate(). 2479 2480 Level: developer 2481 2482 Concepts: matrices^triangular solves 2483 2484 .seealso: MatMatSolveAdd(), MatMatSolveTranspose(), MatMatSolveTransposeAdd() 2485 @*/ 2486 PetscErrorCode PETSCMAT_DLLEXPORT MatMatSolve(Mat A,Mat B,Mat X) 2487 { 2488 PetscErrorCode ierr; 2489 2490 PetscFunctionBegin; 2491 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 2492 PetscValidType(A,1); 2493 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 2494 PetscValidHeaderSpecific(X,MAT_COOKIE,3); 2495 PetscCheckSameComm(A,1,B,2); 2496 PetscCheckSameComm(A,1,X,3); 2497 if (X == B) SETERRQ(PETSC_ERR_ARG_IDN,"X and B must be different matrices"); 2498 if (!A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2499 if (A->cmap.N != X->rmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat A,Mat X: global dim %D %D",A->cmap.N,X->rmap.N); 2500 if (A->rmap.N != B->rmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat A,Mat B: global dim %D %D",A->rmap.N,B->rmap.N); 2501 if (A->rmap.n != B->rmap.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat A,Mat B: local dim %D %D",A->rmap.n,B->rmap.n); 2502 if (!A->rmap.N && !A->cmap.N) PetscFunctionReturn(0); 2503 if (!A->ops->matsolve) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",A->type_name); 2504 ierr = MatPreallocated(A);CHKERRQ(ierr); 2505 2506 ierr = PetscLogEventBegin(MAT_MatSolve,A,B,X,0);CHKERRQ(ierr); 2507 ierr = (*A->ops->matsolve)(A,B,X);CHKERRQ(ierr); 2508 ierr = PetscLogEventEnd(MAT_MatSolve,A,B,X,0);CHKERRQ(ierr); 2509 ierr = PetscObjectStateIncrease((PetscObject)X);CHKERRQ(ierr); 2510 PetscFunctionReturn(0); 2511 } 2512 2513 2514 #undef __FUNCT__ 2515 #define __FUNCT__ "MatForwardSolve" 2516 /* @ 2517 MatForwardSolve - Solves L x = b, given a factored matrix, A = LU, or 2518 U^T*D^(1/2) x = b, given a factored symmetric matrix, A = U^T*D*U, 2519 2520 Collective on Mat and Vec 2521 2522 Input Parameters: 2523 + mat - the factored matrix 2524 - b - the right-hand-side vector 2525 2526 Output Parameter: 2527 . x - the result vector 2528 2529 Notes: 2530 MatSolve() should be used for most applications, as it performs 2531 a forward solve followed by a backward solve. 2532 2533 The vectors b and x cannot be the same, i.e., one cannot 2534 call MatForwardSolve(A,x,x). 2535 2536 For matrix in seqsbaij format with block size larger than 1, 2537 the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet. 2538 MatForwardSolve() solves U^T*D y = b, and 2539 MatBackwardSolve() solves U x = y. 2540 Thus they do not provide a symmetric preconditioner. 2541 2542 Most users should employ the simplified KSP interface for linear solvers 2543 instead of working directly with matrix algebra routines such as this. 2544 See, e.g., KSPCreate(). 2545 2546 Level: developer 2547 2548 Concepts: matrices^forward solves 2549 2550 .seealso: MatSolve(), MatBackwardSolve() 2551 @ */ 2552 PetscErrorCode PETSCMAT_DLLEXPORT MatForwardSolve(Mat mat,Vec b,Vec x) 2553 { 2554 PetscErrorCode ierr; 2555 2556 PetscFunctionBegin; 2557 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2558 PetscValidType(mat,1); 2559 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2560 PetscValidHeaderSpecific(x,VEC_COOKIE,3); 2561 PetscCheckSameComm(mat,1,b,2); 2562 PetscCheckSameComm(mat,1,x,3); 2563 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2564 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2565 if (!mat->ops->forwardsolve) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2566 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2567 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2568 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2569 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2570 ierr = PetscLogEventBegin(MAT_ForwardSolve,mat,b,x,0);CHKERRQ(ierr); 2571 ierr = (*mat->ops->forwardsolve)(mat,b,x);CHKERRQ(ierr); 2572 ierr = PetscLogEventEnd(MAT_ForwardSolve,mat,b,x,0);CHKERRQ(ierr); 2573 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2574 PetscFunctionReturn(0); 2575 } 2576 2577 #undef __FUNCT__ 2578 #define __FUNCT__ "MatBackwardSolve" 2579 /* @ 2580 MatBackwardSolve - Solves U x = b, given a factored matrix, A = LU. 2581 D^(1/2) U x = b, given a factored symmetric matrix, A = U^T*D*U, 2582 2583 Collective on Mat and Vec 2584 2585 Input Parameters: 2586 + mat - the factored matrix 2587 - b - the right-hand-side vector 2588 2589 Output Parameter: 2590 . x - the result vector 2591 2592 Notes: 2593 MatSolve() should be used for most applications, as it performs 2594 a forward solve followed by a backward solve. 2595 2596 The vectors b and x cannot be the same. I.e., one cannot 2597 call MatBackwardSolve(A,x,x). 2598 2599 For matrix in seqsbaij format with block size larger than 1, 2600 the diagonal blocks are not implemented as D = D^(1/2) * D^(1/2) yet. 2601 MatForwardSolve() solves U^T*D y = b, and 2602 MatBackwardSolve() solves U x = y. 2603 Thus they do not provide a symmetric preconditioner. 2604 2605 Most users should employ the simplified KSP interface for linear solvers 2606 instead of working directly with matrix algebra routines such as this. 2607 See, e.g., KSPCreate(). 2608 2609 Level: developer 2610 2611 Concepts: matrices^backward solves 2612 2613 .seealso: MatSolve(), MatForwardSolve() 2614 @ */ 2615 PetscErrorCode PETSCMAT_DLLEXPORT MatBackwardSolve(Mat mat,Vec b,Vec x) 2616 { 2617 PetscErrorCode ierr; 2618 2619 PetscFunctionBegin; 2620 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2621 PetscValidType(mat,1); 2622 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2623 PetscValidHeaderSpecific(x,VEC_COOKIE,3); 2624 PetscCheckSameComm(mat,1,b,2); 2625 PetscCheckSameComm(mat,1,x,3); 2626 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2627 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2628 if (!mat->ops->backwardsolve) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2629 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2630 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2631 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2632 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2633 2634 ierr = PetscLogEventBegin(MAT_BackwardSolve,mat,b,x,0);CHKERRQ(ierr); 2635 ierr = (*mat->ops->backwardsolve)(mat,b,x);CHKERRQ(ierr); 2636 ierr = PetscLogEventEnd(MAT_BackwardSolve,mat,b,x,0);CHKERRQ(ierr); 2637 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2638 PetscFunctionReturn(0); 2639 } 2640 2641 #undef __FUNCT__ 2642 #define __FUNCT__ "MatSolveAdd" 2643 /*@ 2644 MatSolveAdd - Computes x = y + inv(A)*b, given a factored matrix. 2645 2646 Collective on Mat and Vec 2647 2648 Input Parameters: 2649 + mat - the factored matrix 2650 . b - the right-hand-side vector 2651 - y - the vector to be added to 2652 2653 Output Parameter: 2654 . x - the result vector 2655 2656 Notes: 2657 The vectors b and x cannot be the same. I.e., one cannot 2658 call MatSolveAdd(A,x,y,x). 2659 2660 Most users should employ the simplified KSP interface for linear solvers 2661 instead of working directly with matrix algebra routines such as this. 2662 See, e.g., KSPCreate(). 2663 2664 Level: developer 2665 2666 Concepts: matrices^triangular solves 2667 2668 .seealso: MatSolve(), MatSolveTranspose(), MatSolveTransposeAdd() 2669 @*/ 2670 PetscErrorCode PETSCMAT_DLLEXPORT MatSolveAdd(Mat mat,Vec b,Vec y,Vec x) 2671 { 2672 PetscScalar one = 1.0; 2673 Vec tmp; 2674 PetscErrorCode ierr; 2675 2676 PetscFunctionBegin; 2677 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2678 PetscValidType(mat,1); 2679 PetscValidHeaderSpecific(y,VEC_COOKIE,2); 2680 PetscValidHeaderSpecific(b,VEC_COOKIE,3); 2681 PetscValidHeaderSpecific(x,VEC_COOKIE,4); 2682 PetscCheckSameComm(mat,1,b,2); 2683 PetscCheckSameComm(mat,1,y,2); 2684 PetscCheckSameComm(mat,1,x,3); 2685 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2686 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2687 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2688 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2689 if (mat->rmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->rmap.N,y->map.N); 2690 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2691 if (x->map.n != y->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Vec x,Vec y: local dim %D %D",x->map.n,y->map.n); 2692 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2693 2694 ierr = PetscLogEventBegin(MAT_SolveAdd,mat,b,x,y);CHKERRQ(ierr); 2695 if (mat->ops->solveadd) { 2696 ierr = (*mat->ops->solveadd)(mat,b,y,x);CHKERRQ(ierr); 2697 } else { 2698 /* do the solve then the add manually */ 2699 if (x != y) { 2700 ierr = MatSolve(mat,b,x);CHKERRQ(ierr); 2701 ierr = VecAXPY(x,one,y);CHKERRQ(ierr); 2702 } else { 2703 ierr = VecDuplicate(x,&tmp);CHKERRQ(ierr); 2704 ierr = PetscLogObjectParent(mat,tmp);CHKERRQ(ierr); 2705 ierr = VecCopy(x,tmp);CHKERRQ(ierr); 2706 ierr = MatSolve(mat,b,x);CHKERRQ(ierr); 2707 ierr = VecAXPY(x,one,tmp);CHKERRQ(ierr); 2708 ierr = VecDestroy(tmp);CHKERRQ(ierr); 2709 } 2710 } 2711 ierr = PetscLogEventEnd(MAT_SolveAdd,mat,b,x,y);CHKERRQ(ierr); 2712 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2713 PetscFunctionReturn(0); 2714 } 2715 2716 #undef __FUNCT__ 2717 #define __FUNCT__ "MatSolveTranspose" 2718 /*@ 2719 MatSolveTranspose - Solves A' x = b, given a factored matrix. 2720 2721 Collective on Mat and Vec 2722 2723 Input Parameters: 2724 + mat - the factored matrix 2725 - b - the right-hand-side vector 2726 2727 Output Parameter: 2728 . x - the result vector 2729 2730 Notes: 2731 The vectors b and x cannot be the same. I.e., one cannot 2732 call MatSolveTranspose(A,x,x). 2733 2734 Most users should employ the simplified KSP interface for linear solvers 2735 instead of working directly with matrix algebra routines such as this. 2736 See, e.g., KSPCreate(). 2737 2738 Level: developer 2739 2740 Concepts: matrices^triangular solves 2741 2742 .seealso: MatSolve(), MatSolveAdd(), MatSolveTransposeAdd() 2743 @*/ 2744 PetscErrorCode PETSCMAT_DLLEXPORT MatSolveTranspose(Mat mat,Vec b,Vec x) 2745 { 2746 PetscErrorCode ierr; 2747 2748 PetscFunctionBegin; 2749 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2750 PetscValidType(mat,1); 2751 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2752 PetscValidHeaderSpecific(x,VEC_COOKIE,3); 2753 PetscCheckSameComm(mat,1,b,2); 2754 PetscCheckSameComm(mat,1,x,3); 2755 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2756 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2757 if (!mat->ops->solvetranspose) SETERRQ1(PETSC_ERR_SUP,"Matrix type %s",mat->type_name); 2758 if (mat->rmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap.N,x->map.N); 2759 if (mat->cmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap.N,b->map.N); 2760 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2761 ierr = PetscLogEventBegin(MAT_SolveTranspose,mat,b,x,0);CHKERRQ(ierr); 2762 ierr = (*mat->ops->solvetranspose)(mat,b,x);CHKERRQ(ierr); 2763 ierr = PetscLogEventEnd(MAT_SolveTranspose,mat,b,x,0);CHKERRQ(ierr); 2764 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2765 PetscFunctionReturn(0); 2766 } 2767 2768 #undef __FUNCT__ 2769 #define __FUNCT__ "MatSolveTransposeAdd" 2770 /*@ 2771 MatSolveTransposeAdd - Computes x = y + inv(Transpose(A)) b, given a 2772 factored matrix. 2773 2774 Collective on Mat and Vec 2775 2776 Input Parameters: 2777 + mat - the factored matrix 2778 . b - the right-hand-side vector 2779 - y - the vector to be added to 2780 2781 Output Parameter: 2782 . x - the result vector 2783 2784 Notes: 2785 The vectors b and x cannot be the same. I.e., one cannot 2786 call MatSolveTransposeAdd(A,x,y,x). 2787 2788 Most users should employ the simplified KSP interface for linear solvers 2789 instead of working directly with matrix algebra routines such as this. 2790 See, e.g., KSPCreate(). 2791 2792 Level: developer 2793 2794 Concepts: matrices^triangular solves 2795 2796 .seealso: MatSolve(), MatSolveAdd(), MatSolveTranspose() 2797 @*/ 2798 PetscErrorCode PETSCMAT_DLLEXPORT MatSolveTransposeAdd(Mat mat,Vec b,Vec y,Vec x) 2799 { 2800 PetscScalar one = 1.0; 2801 PetscErrorCode ierr; 2802 Vec tmp; 2803 2804 PetscFunctionBegin; 2805 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2806 PetscValidType(mat,1); 2807 PetscValidHeaderSpecific(y,VEC_COOKIE,2); 2808 PetscValidHeaderSpecific(b,VEC_COOKIE,3); 2809 PetscValidHeaderSpecific(x,VEC_COOKIE,4); 2810 PetscCheckSameComm(mat,1,b,2); 2811 PetscCheckSameComm(mat,1,y,3); 2812 PetscCheckSameComm(mat,1,x,4); 2813 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 2814 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 2815 if (mat->rmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->rmap.N,x->map.N); 2816 if (mat->cmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->cmap.N,b->map.N); 2817 if (mat->cmap.N != y->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec y: global dim %D %D",mat->cmap.N,y->map.N); 2818 if (x->map.n != y->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Vec x,Vec y: local dim %D %D",x->map.n,y->map.n); 2819 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2820 2821 ierr = PetscLogEventBegin(MAT_SolveTransposeAdd,mat,b,x,y);CHKERRQ(ierr); 2822 if (mat->ops->solvetransposeadd) { 2823 ierr = (*mat->ops->solvetransposeadd)(mat,b,y,x);CHKERRQ(ierr); 2824 } else { 2825 /* do the solve then the add manually */ 2826 if (x != y) { 2827 ierr = MatSolveTranspose(mat,b,x);CHKERRQ(ierr); 2828 ierr = VecAXPY(x,one,y);CHKERRQ(ierr); 2829 } else { 2830 ierr = VecDuplicate(x,&tmp);CHKERRQ(ierr); 2831 ierr = PetscLogObjectParent(mat,tmp);CHKERRQ(ierr); 2832 ierr = VecCopy(x,tmp);CHKERRQ(ierr); 2833 ierr = MatSolveTranspose(mat,b,x);CHKERRQ(ierr); 2834 ierr = VecAXPY(x,one,tmp);CHKERRQ(ierr); 2835 ierr = VecDestroy(tmp);CHKERRQ(ierr); 2836 } 2837 } 2838 ierr = PetscLogEventEnd(MAT_SolveTransposeAdd,mat,b,x,y);CHKERRQ(ierr); 2839 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2840 PetscFunctionReturn(0); 2841 } 2842 /* ----------------------------------------------------------------*/ 2843 2844 #undef __FUNCT__ 2845 #define __FUNCT__ "MatRelax" 2846 /*@ 2847 MatRelax - Computes relaxation (SOR, Gauss-Seidel) sweeps. 2848 2849 Collective on Mat and Vec 2850 2851 Input Parameters: 2852 + mat - the matrix 2853 . b - the right hand side 2854 . omega - the relaxation factor 2855 . flag - flag indicating the type of SOR (see below) 2856 . shift - diagonal shift 2857 - its - the number of iterations 2858 - lits - the number of local iterations 2859 2860 Output Parameters: 2861 . x - the solution (can contain an initial guess) 2862 2863 SOR Flags: 2864 . SOR_FORWARD_SWEEP - forward SOR 2865 . SOR_BACKWARD_SWEEP - backward SOR 2866 . SOR_SYMMETRIC_SWEEP - SSOR (symmetric SOR) 2867 . SOR_LOCAL_FORWARD_SWEEP - local forward SOR 2868 . SOR_LOCAL_BACKWARD_SWEEP - local forward SOR 2869 . SOR_LOCAL_SYMMETRIC_SWEEP - local SSOR 2870 . SOR_APPLY_UPPER, SOR_APPLY_LOWER - applies 2871 upper/lower triangular part of matrix to 2872 vector (with omega) 2873 . SOR_ZERO_INITIAL_GUESS - zero initial guess 2874 2875 Notes: 2876 SOR_LOCAL_FORWARD_SWEEP, SOR_LOCAL_BACKWARD_SWEEP, and 2877 SOR_LOCAL_SYMMETRIC_SWEEP perform separate independent smoothings 2878 on each processor. 2879 2880 Application programmers will not generally use MatRelax() directly, 2881 but instead will employ the KSP/PC interface. 2882 2883 Notes for Advanced Users: 2884 The flags are implemented as bitwise inclusive or operations. 2885 For example, use (SOR_ZERO_INITIAL_GUESS | SOR_SYMMETRIC_SWEEP) 2886 to specify a zero initial guess for SSOR. 2887 2888 Most users should employ the simplified KSP interface for linear solvers 2889 instead of working directly with matrix algebra routines such as this. 2890 See, e.g., KSPCreate(). 2891 2892 See also, MatPBRelax(). This routine will automatically call the point block 2893 version if the point version is not available. 2894 2895 Level: developer 2896 2897 Concepts: matrices^relaxation 2898 Concepts: matrices^SOR 2899 Concepts: matrices^Gauss-Seidel 2900 2901 @*/ 2902 PetscErrorCode PETSCMAT_DLLEXPORT MatRelax(Mat mat,Vec b,PetscReal omega,MatSORType flag,PetscReal shift,PetscInt its,PetscInt lits,Vec x) 2903 { 2904 PetscErrorCode ierr; 2905 2906 PetscFunctionBegin; 2907 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2908 PetscValidType(mat,1); 2909 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2910 PetscValidHeaderSpecific(x,VEC_COOKIE,8); 2911 PetscCheckSameComm(mat,1,b,2); 2912 PetscCheckSameComm(mat,1,x,8); 2913 if (!mat->ops->relax && !mat->ops->pbrelax) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2914 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2915 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2916 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2917 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2918 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2919 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2920 ierr = PetscLogEventBegin(MAT_Relax,mat,b,x,0);CHKERRQ(ierr); 2921 if (mat->ops->relax) { 2922 ierr =(*mat->ops->relax)(mat,b,omega,flag,shift,its,lits,x);CHKERRQ(ierr); 2923 } else { 2924 ierr =(*mat->ops->pbrelax)(mat,b,omega,flag,shift,its,lits,x);CHKERRQ(ierr); 2925 } 2926 ierr = PetscLogEventEnd(MAT_Relax,mat,b,x,0);CHKERRQ(ierr); 2927 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2928 PetscFunctionReturn(0); 2929 } 2930 2931 #undef __FUNCT__ 2932 #define __FUNCT__ "MatPBRelax" 2933 /*@ 2934 MatPBRelax - Computes relaxation (SOR, Gauss-Seidel) sweeps. 2935 2936 Collective on Mat and Vec 2937 2938 See MatRelax() for usage 2939 2940 For multi-component PDEs where the Jacobian is stored in a point block format 2941 (with the PETSc BAIJ matrix formats) the relaxation is done one point block at 2942 a time. That is, the small (for example, 4 by 4) blocks along the diagonal are solved 2943 simultaneously (that is a 4 by 4 linear solve is done) to update all the values at a point. 2944 2945 Level: developer 2946 2947 @*/ 2948 PetscErrorCode PETSCMAT_DLLEXPORT MatPBRelax(Mat mat,Vec b,PetscReal omega,MatSORType flag,PetscReal shift,PetscInt its,PetscInt lits,Vec x) 2949 { 2950 PetscErrorCode ierr; 2951 2952 PetscFunctionBegin; 2953 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 2954 PetscValidType(mat,1); 2955 PetscValidHeaderSpecific(b,VEC_COOKIE,2); 2956 PetscValidHeaderSpecific(x,VEC_COOKIE,8); 2957 PetscCheckSameComm(mat,1,b,2); 2958 PetscCheckSameComm(mat,1,x,8); 2959 if (!mat->ops->pbrelax) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 2960 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 2961 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2962 if (mat->cmap.N != x->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec x: global dim %D %D",mat->cmap.N,x->map.N); 2963 if (mat->rmap.N != b->map.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: global dim %D %D",mat->rmap.N,b->map.N); 2964 if (mat->rmap.n != b->map.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"Mat mat,Vec b: local dim %D %D",mat->rmap.n,b->map.n); 2965 ierr = MatPreallocated(mat);CHKERRQ(ierr); 2966 2967 ierr = PetscLogEventBegin(MAT_Relax,mat,b,x,0);CHKERRQ(ierr); 2968 ierr =(*mat->ops->pbrelax)(mat,b,omega,flag,shift,its,lits,x);CHKERRQ(ierr); 2969 ierr = PetscLogEventEnd(MAT_Relax,mat,b,x,0);CHKERRQ(ierr); 2970 ierr = PetscObjectStateIncrease((PetscObject)x);CHKERRQ(ierr); 2971 PetscFunctionReturn(0); 2972 } 2973 2974 #undef __FUNCT__ 2975 #define __FUNCT__ "MatCopy_Basic" 2976 /* 2977 Default matrix copy routine. 2978 */ 2979 PetscErrorCode MatCopy_Basic(Mat A,Mat B,MatStructure str) 2980 { 2981 PetscErrorCode ierr; 2982 PetscInt i,rstart,rend,nz; 2983 const PetscInt *cwork; 2984 const PetscScalar *vwork; 2985 2986 PetscFunctionBegin; 2987 if (B->assembled) { 2988 ierr = MatZeroEntries(B);CHKERRQ(ierr); 2989 } 2990 ierr = MatGetOwnershipRange(A,&rstart,&rend);CHKERRQ(ierr); 2991 for (i=rstart; i<rend; i++) { 2992 ierr = MatGetRow(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 2993 ierr = MatSetValues(B,1,&i,nz,cwork,vwork,INSERT_VALUES);CHKERRQ(ierr); 2994 ierr = MatRestoreRow(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 2995 } 2996 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2997 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2998 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 2999 PetscFunctionReturn(0); 3000 } 3001 3002 #undef __FUNCT__ 3003 #define __FUNCT__ "MatCopy" 3004 /*@ 3005 MatCopy - Copys a matrix to another matrix. 3006 3007 Collective on Mat 3008 3009 Input Parameters: 3010 + A - the matrix 3011 - str - SAME_NONZERO_PATTERN or DIFFERENT_NONZERO_PATTERN 3012 3013 Output Parameter: 3014 . B - where the copy is put 3015 3016 Notes: 3017 If you use SAME_NONZERO_PATTERN then the two matrices had better have the 3018 same nonzero pattern or the routine will crash. 3019 3020 MatCopy() copies the matrix entries of a matrix to another existing 3021 matrix (after first zeroing the second matrix). A related routine is 3022 MatConvert(), which first creates a new matrix and then copies the data. 3023 3024 Level: intermediate 3025 3026 Concepts: matrices^copying 3027 3028 .seealso: MatConvert(), MatDuplicate() 3029 3030 @*/ 3031 PetscErrorCode PETSCMAT_DLLEXPORT MatCopy(Mat A,Mat B,MatStructure str) 3032 { 3033 PetscErrorCode ierr; 3034 3035 PetscFunctionBegin; 3036 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 3037 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 3038 PetscValidType(A,1); 3039 PetscValidType(B,2); 3040 MatPreallocated(B); 3041 PetscCheckSameComm(A,1,B,2); 3042 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3043 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3044 if (A->rmap.N != B->rmap.N || A->cmap.N != B->cmap.N) SETERRQ4(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); 3045 ierr = MatPreallocated(A);CHKERRQ(ierr); 3046 3047 ierr = PetscLogEventBegin(MAT_Copy,A,B,0,0);CHKERRQ(ierr); 3048 if (A->ops->copy) { 3049 ierr = (*A->ops->copy)(A,B,str);CHKERRQ(ierr); 3050 } else { /* generic conversion */ 3051 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 3052 } 3053 if (A->mapping) { 3054 if (B->mapping) {ierr = ISLocalToGlobalMappingDestroy(B->mapping);CHKERRQ(ierr);B->mapping = 0;} 3055 ierr = MatSetLocalToGlobalMapping(B,A->mapping);CHKERRQ(ierr); 3056 } 3057 if (A->bmapping) { 3058 if (B->bmapping) {ierr = ISLocalToGlobalMappingDestroy(B->bmapping);CHKERRQ(ierr);B->bmapping = 0;} 3059 ierr = MatSetLocalToGlobalMappingBlock(B,A->mapping);CHKERRQ(ierr); 3060 } 3061 ierr = PetscLogEventEnd(MAT_Copy,A,B,0,0);CHKERRQ(ierr); 3062 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 3063 PetscFunctionReturn(0); 3064 } 3065 3066 #undef __FUNCT__ 3067 #define __FUNCT__ "MatConvert" 3068 /*@C 3069 MatConvert - Converts a matrix to another matrix, either of the same 3070 or different type. 3071 3072 Collective on Mat 3073 3074 Input Parameters: 3075 + mat - the matrix 3076 . newtype - new matrix type. Use MATSAME to create a new matrix of the 3077 same type as the original matrix. 3078 - reuse - denotes if the destination matrix is to be created or reused. Currently 3079 MAT_REUSE_MATRIX is only supported for inplace conversion, otherwise use 3080 MAT_INITIAL_MATRIX. 3081 3082 Output Parameter: 3083 . M - pointer to place new matrix 3084 3085 Notes: 3086 MatConvert() first creates a new matrix and then copies the data from 3087 the first matrix. A related routine is MatCopy(), which copies the matrix 3088 entries of one matrix to another already existing matrix context. 3089 3090 Level: intermediate 3091 3092 Concepts: matrices^converting between storage formats 3093 3094 .seealso: MatCopy(), MatDuplicate() 3095 @*/ 3096 PetscErrorCode PETSCMAT_DLLEXPORT MatConvert(Mat mat, MatType newtype,MatReuse reuse,Mat *M) 3097 { 3098 PetscErrorCode ierr; 3099 PetscTruth sametype,issame,flg; 3100 char convname[256],mtype[256]; 3101 Mat B; 3102 3103 PetscFunctionBegin; 3104 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3105 PetscValidType(mat,1); 3106 PetscValidPointer(M,3); 3107 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3108 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3109 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3110 3111 ierr = PetscOptionsGetString(PETSC_NULL,"-matconvert_type",mtype,256,&flg);CHKERRQ(ierr); 3112 if (flg) { 3113 newtype = mtype; 3114 } 3115 ierr = PetscTypeCompare((PetscObject)mat,newtype,&sametype);CHKERRQ(ierr); 3116 ierr = PetscStrcmp(newtype,"same",&issame);CHKERRQ(ierr); 3117 if ((reuse==MAT_REUSE_MATRIX) && (mat != *M)) { 3118 SETERRQ(PETSC_ERR_SUP,"MAT_REUSE_MATRIX only supported for in-place conversion currently"); 3119 } 3120 if ((sametype || issame) && (reuse==MAT_INITIAL_MATRIX) && mat->ops->duplicate) { 3121 ierr = (*mat->ops->duplicate)(mat,MAT_COPY_VALUES,M);CHKERRQ(ierr); 3122 } else { 3123 PetscErrorCode (*conv)(Mat, MatType,MatReuse,Mat*)=PETSC_NULL; 3124 const char *prefix[3] = {"seq","mpi",""}; 3125 PetscInt i; 3126 /* 3127 Order of precedence: 3128 1) See if a specialized converter is known to the current matrix. 3129 2) See if a specialized converter is known to the desired matrix class. 3130 3) See if a good general converter is registered for the desired class 3131 (as of 6/27/03 only MATMPIADJ falls into this category). 3132 4) See if a good general converter is known for the current matrix. 3133 5) Use a really basic converter. 3134 */ 3135 3136 /* 1) See if a specialized converter is known to the current matrix and the desired class */ 3137 for (i=0; i<3; i++) { 3138 ierr = PetscStrcpy(convname,"MatConvert_");CHKERRQ(ierr); 3139 ierr = PetscStrcat(convname,mat->type_name);CHKERRQ(ierr); 3140 ierr = PetscStrcat(convname,"_");CHKERRQ(ierr); 3141 ierr = PetscStrcat(convname,prefix[i]);CHKERRQ(ierr); 3142 ierr = PetscStrcat(convname,newtype);CHKERRQ(ierr); 3143 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3144 ierr = PetscObjectQueryFunction((PetscObject)mat,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3145 if (conv) goto foundconv; 3146 } 3147 3148 /* 2) See if a specialized converter is known to the desired matrix class. */ 3149 ierr = MatCreate(mat->comm,&B);CHKERRQ(ierr); 3150 ierr = MatSetSizes(B,mat->rmap.n,mat->cmap.n,mat->rmap.N,mat->cmap.N);CHKERRQ(ierr); 3151 ierr = MatSetType(B,newtype);CHKERRQ(ierr); 3152 for (i=0; i<3; i++) { 3153 ierr = PetscStrcpy(convname,"MatConvert_");CHKERRQ(ierr); 3154 ierr = PetscStrcat(convname,mat->type_name);CHKERRQ(ierr); 3155 ierr = PetscStrcat(convname,"_");CHKERRQ(ierr); 3156 ierr = PetscStrcat(convname,prefix[i]);CHKERRQ(ierr); 3157 ierr = PetscStrcat(convname,newtype);CHKERRQ(ierr); 3158 ierr = PetscStrcat(convname,"_C");CHKERRQ(ierr); 3159 ierr = PetscObjectQueryFunction((PetscObject)B,convname,(void (**)(void))&conv);CHKERRQ(ierr); 3160 if (conv) { 3161 ierr = MatDestroy(B);CHKERRQ(ierr); 3162 goto foundconv; 3163 } 3164 } 3165 3166 /* 3) See if a good general converter is registered for the desired class */ 3167 conv = B->ops->convertfrom; 3168 ierr = MatDestroy(B);CHKERRQ(ierr); 3169 if (conv) goto foundconv; 3170 3171 /* 4) See if a good general converter is known for the current matrix */ 3172 if (mat->ops->convert) { 3173 conv = mat->ops->convert; 3174 } 3175 if (conv) goto foundconv; 3176 3177 /* 5) Use a really basic converter. */ 3178 conv = MatConvert_Basic; 3179 3180 foundconv: 3181 ierr = PetscLogEventBegin(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3182 ierr = (*conv)(mat,newtype,reuse,M);CHKERRQ(ierr); 3183 ierr = PetscLogEventEnd(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3184 } 3185 B = *M; 3186 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 3187 PetscFunctionReturn(0); 3188 } 3189 3190 3191 #undef __FUNCT__ 3192 #define __FUNCT__ "MatDuplicate" 3193 /*@ 3194 MatDuplicate - Duplicates a matrix including the non-zero structure. 3195 3196 Collective on Mat 3197 3198 Input Parameters: 3199 + mat - the matrix 3200 - op - either MAT_DO_NOT_COPY_VALUES or MAT_COPY_VALUES, cause it to copy nonzero 3201 values as well or not 3202 3203 Output Parameter: 3204 . M - pointer to place new matrix 3205 3206 Level: intermediate 3207 3208 Concepts: matrices^duplicating 3209 3210 .seealso: MatCopy(), MatConvert() 3211 @*/ 3212 PetscErrorCode PETSCMAT_DLLEXPORT MatDuplicate(Mat mat,MatDuplicateOption op,Mat *M) 3213 { 3214 PetscErrorCode ierr; 3215 Mat B; 3216 3217 PetscFunctionBegin; 3218 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3219 PetscValidType(mat,1); 3220 PetscValidPointer(M,3); 3221 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3222 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3223 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3224 3225 *M = 0; 3226 if (!mat->ops->duplicate) { 3227 SETERRQ(PETSC_ERR_SUP,"Not written for this matrix type"); 3228 } 3229 ierr = PetscLogEventBegin(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3230 ierr = (*mat->ops->duplicate)(mat,op,M);CHKERRQ(ierr); 3231 B = *M; 3232 if (mat->mapping) { 3233 ierr = MatSetLocalToGlobalMapping(B,mat->mapping);CHKERRQ(ierr); 3234 } 3235 if (mat->bmapping) { 3236 ierr = MatSetLocalToGlobalMappingBlock(B,mat->bmapping);CHKERRQ(ierr); 3237 } 3238 ierr = PetscMapCopy(mat->comm,&mat->rmap,&B->rmap);CHKERRQ(ierr); 3239 ierr = PetscMapCopy(mat->comm,&mat->cmap,&B->cmap);CHKERRQ(ierr); 3240 3241 ierr = PetscLogEventEnd(MAT_Convert,mat,0,0,0);CHKERRQ(ierr); 3242 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 3243 PetscFunctionReturn(0); 3244 } 3245 3246 #undef __FUNCT__ 3247 #define __FUNCT__ "MatGetDiagonal" 3248 /*@ 3249 MatGetDiagonal - Gets the diagonal of a matrix. 3250 3251 Collective on Mat and Vec 3252 3253 Input Parameters: 3254 + mat - the matrix 3255 - v - the vector for storing the diagonal 3256 3257 Output Parameter: 3258 . v - the diagonal of the matrix 3259 3260 Notes: The result of this call are the same as if one converted the matrix to dense format 3261 and found the minimum value in each row (i.e. the implicit zeros are counted as zeros). 3262 3263 Level: intermediate 3264 3265 Concepts: matrices^accessing diagonals 3266 3267 .seealso: MatGetRow(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs() 3268 @*/ 3269 PetscErrorCode PETSCMAT_DLLEXPORT MatGetDiagonal(Mat mat,Vec v) 3270 { 3271 PetscErrorCode ierr; 3272 3273 PetscFunctionBegin; 3274 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3275 PetscValidType(mat,1); 3276 PetscValidHeaderSpecific(v,VEC_COOKIE,2); 3277 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3278 if (!mat->ops->getdiagonal) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3279 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3280 3281 ierr = (*mat->ops->getdiagonal)(mat,v);CHKERRQ(ierr); 3282 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 3283 PetscFunctionReturn(0); 3284 } 3285 3286 #undef __FUNCT__ 3287 #define __FUNCT__ "MatGetRowMin" 3288 /*@ 3289 MatGetRowMin - Gets the minimum value (of the real part) of each 3290 row of the matrix 3291 3292 Collective on Mat and Vec 3293 3294 Input Parameters: 3295 . mat - the matrix 3296 3297 Output Parameter: 3298 + v - the vector for storing the maximums 3299 - idx - the indices of the column found for each row (optional) 3300 3301 Level: intermediate 3302 3303 Notes: The result of this call are the same as if one converted the matrix to dense format 3304 and found the minimum value in each row (i.e. the implicit zeros are counted as zeros). 3305 3306 This code is only implemented for a couple of matrix formats. 3307 3308 Concepts: matrices^getting row maximums 3309 3310 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs(), 3311 MatGetRowMax() 3312 @*/ 3313 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowMin(Mat mat,Vec v,PetscInt idx[]) 3314 { 3315 PetscErrorCode ierr; 3316 3317 PetscFunctionBegin; 3318 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3319 PetscValidType(mat,1); 3320 PetscValidHeaderSpecific(v,VEC_COOKIE,2); 3321 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3322 if (!mat->ops->getrowmax) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3323 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3324 3325 ierr = (*mat->ops->getrowmin)(mat,v,idx);CHKERRQ(ierr); 3326 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 3327 PetscFunctionReturn(0); 3328 } 3329 3330 #undef __FUNCT__ 3331 #define __FUNCT__ "MatGetRowMax" 3332 /*@ 3333 MatGetRowMax - Gets the maximum value (of the real part) of each 3334 row of the matrix 3335 3336 Collective on Mat and Vec 3337 3338 Input Parameters: 3339 . mat - the matrix 3340 3341 Output Parameter: 3342 + v - the vector for storing the maximums 3343 - idx - the indices of the column found for each row (optional) 3344 3345 Level: intermediate 3346 3347 Notes: The result of this call are the same as if one converted the matrix to dense format 3348 and found the minimum value in each row (i.e. the implicit zeros are counted as zeros). 3349 3350 This code is only implemented for a couple of matrix formats. 3351 3352 Concepts: matrices^getting row maximums 3353 3354 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMaxAbs(), MatGetRowMin() 3355 @*/ 3356 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowMax(Mat mat,Vec v,PetscInt idx[]) 3357 { 3358 PetscErrorCode ierr; 3359 3360 PetscFunctionBegin; 3361 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3362 PetscValidType(mat,1); 3363 PetscValidHeaderSpecific(v,VEC_COOKIE,2); 3364 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3365 if (!mat->ops->getrowmax) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3366 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3367 3368 ierr = (*mat->ops->getrowmax)(mat,v,idx);CHKERRQ(ierr); 3369 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 3370 PetscFunctionReturn(0); 3371 } 3372 3373 #undef __FUNCT__ 3374 #define __FUNCT__ "MatGetRowMaxAbs" 3375 /*@ 3376 MatGetRowMaxAbs - Gets the maximum value (in absolute value) of each 3377 row of the matrix 3378 3379 Collective on Mat and Vec 3380 3381 Input Parameters: 3382 . mat - the matrix 3383 3384 Output Parameter: 3385 + v - the vector for storing the maximums 3386 - idx - the indices of the column found for each row (optional) 3387 3388 Level: intermediate 3389 3390 Notes: if a row is completely empty or has only 0.0 values then the idx[] value for that 3391 row is 0 (the first column). 3392 3393 This code is only implemented for a couple of matrix formats. 3394 3395 Concepts: matrices^getting row maximums 3396 3397 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMax(), MatGetRowMin() 3398 @*/ 3399 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowMaxAbs(Mat mat,Vec v,PetscInt idx[]) 3400 { 3401 PetscErrorCode ierr; 3402 3403 PetscFunctionBegin; 3404 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3405 PetscValidType(mat,1); 3406 PetscValidHeaderSpecific(v,VEC_COOKIE,2); 3407 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3408 if (!mat->ops->getrowmaxabs) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3409 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3410 3411 ierr = (*mat->ops->getrowmaxabs)(mat,v,idx);CHKERRQ(ierr); 3412 ierr = PetscObjectStateIncrease((PetscObject)v);CHKERRQ(ierr); 3413 PetscFunctionReturn(0); 3414 } 3415 3416 #undef __FUNCT__ 3417 #define __FUNCT__ "MatGetRowSum" 3418 /*@ 3419 MatGetRowSum - Gets the sum of each row of the matrix 3420 3421 Collective on Mat and Vec 3422 3423 Input Parameters: 3424 . mat - the matrix 3425 3426 Output Parameter: 3427 . v - the vector for storing the maximums 3428 3429 Level: intermediate 3430 3431 Notes: This code is slow since it is not currently specialized for different formats 3432 3433 Concepts: matrices^getting row sums 3434 3435 .seealso: MatGetDiagonal(), MatGetSubMatrices(), MatGetSubmatrix(), MatGetRowMax(), MatGetRowMin() 3436 @*/ 3437 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowSum(Mat mat, Vec v) 3438 { 3439 PetscInt start, end, row; 3440 PetscScalar *array; 3441 PetscErrorCode ierr; 3442 3443 PetscFunctionBegin; 3444 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3445 PetscValidType(mat,1); 3446 PetscValidHeaderSpecific(v,VEC_COOKIE,2); 3447 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3448 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3449 ierr = MatGetOwnershipRange(mat, &start, &end);CHKERRQ(ierr); 3450 ierr = VecGetArray(v, &array);CHKERRQ(ierr); 3451 for(row = start; row < end; ++row) { 3452 PetscInt ncols, col; 3453 const PetscInt *cols; 3454 const PetscScalar *vals; 3455 3456 array[row - start] = 0.0; 3457 ierr = MatGetRow(mat, row, &ncols, &cols, &vals);CHKERRQ(ierr); 3458 for(col = 0; col < ncols; col++) { 3459 array[row - start] += vals[col]; 3460 } 3461 } 3462 ierr = VecRestoreArray(v, &array);CHKERRQ(ierr); 3463 ierr = PetscObjectStateIncrease((PetscObject) v);CHKERRQ(ierr); 3464 PetscFunctionReturn(0); 3465 } 3466 3467 #undef __FUNCT__ 3468 #define __FUNCT__ "MatTranspose" 3469 /*@C 3470 MatTranspose - Computes an in-place or out-of-place transpose of a matrix. 3471 3472 Collective on Mat 3473 3474 Input Parameter: 3475 . mat - the matrix to transpose 3476 3477 Output Parameters: 3478 . B - the transpose 3479 3480 Notes: 3481 If you pass in PETSC_NULL for B an in-place transpose in mat will be done 3482 3483 Level: intermediate 3484 3485 Concepts: matrices^transposing 3486 3487 .seealso: MatMultTranspose(), MatMultTransposeAdd(), MatIsTranspose() 3488 @*/ 3489 PetscErrorCode PETSCMAT_DLLEXPORT MatTranspose(Mat mat,Mat *B) 3490 { 3491 PetscErrorCode ierr; 3492 3493 PetscFunctionBegin; 3494 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3495 PetscValidType(mat,1); 3496 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3497 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3498 if (!mat->ops->transpose) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3499 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3500 3501 ierr = PetscLogEventBegin(MAT_Transpose,mat,0,0,0);CHKERRQ(ierr); 3502 ierr = (*mat->ops->transpose)(mat,B);CHKERRQ(ierr); 3503 ierr = PetscLogEventEnd(MAT_Transpose,mat,0,0,0);CHKERRQ(ierr); 3504 if (B) {ierr = PetscObjectStateIncrease((PetscObject)*B);CHKERRQ(ierr);} 3505 PetscFunctionReturn(0); 3506 } 3507 3508 #undef __FUNCT__ 3509 #define __FUNCT__ "MatIsTranspose" 3510 /*@C 3511 MatIsTranspose - Test whether a matrix is another one's transpose, 3512 or its own, in which case it tests symmetry. 3513 3514 Collective on Mat 3515 3516 Input Parameter: 3517 + A - the matrix to test 3518 - B - the matrix to test against, this can equal the first parameter 3519 3520 Output Parameters: 3521 . flg - the result 3522 3523 Notes: 3524 Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm 3525 has a running time of the order of the number of nonzeros; the parallel 3526 test involves parallel copies of the block-offdiagonal parts of the matrix. 3527 3528 Level: intermediate 3529 3530 Concepts: matrices^transposing, matrix^symmetry 3531 3532 .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian() 3533 @*/ 3534 PetscErrorCode PETSCMAT_DLLEXPORT MatIsTranspose(Mat A,Mat B,PetscReal tol,PetscTruth *flg) 3535 { 3536 PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscTruth*),(*g)(Mat,Mat,PetscReal,PetscTruth*); 3537 3538 PetscFunctionBegin; 3539 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 3540 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 3541 PetscValidPointer(flg,3); 3542 ierr = PetscObjectQueryFunction((PetscObject)A,"MatIsTranspose_C",(void (**)(void))&f);CHKERRQ(ierr); 3543 ierr = PetscObjectQueryFunction((PetscObject)B,"MatIsTranspose_C",(void (**)(void))&g);CHKERRQ(ierr); 3544 if (f && g) { 3545 if (f==g) { 3546 ierr = (*f)(A,B,tol,flg);CHKERRQ(ierr); 3547 } else { 3548 SETERRQ(PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for symmetry test"); 3549 } 3550 } 3551 PetscFunctionReturn(0); 3552 } 3553 3554 #undef __FUNCT__ 3555 #define __FUNCT__ "MatIsHermitianTranspose" 3556 /*@C 3557 MatIsHermitianTranspose - Test whether a matrix is another one's Hermitian transpose, 3558 3559 Collective on Mat 3560 3561 Input Parameter: 3562 + A - the matrix to test 3563 - B - the matrix to test against, this can equal the first parameter 3564 3565 Output Parameters: 3566 . flg - the result 3567 3568 Notes: 3569 Only available for SeqAIJ/MPIAIJ matrices. The sequential algorithm 3570 has a running time of the order of the number of nonzeros; the parallel 3571 test involves parallel copies of the block-offdiagonal parts of the matrix. 3572 3573 Level: intermediate 3574 3575 Concepts: matrices^transposing, matrix^symmetry 3576 3577 .seealso: MatTranspose(), MatIsSymmetric(), MatIsHermitian(), MatIsTranspose() 3578 @*/ 3579 PetscErrorCode PETSCMAT_DLLEXPORT MatIsHermitianTranspose(Mat A,Mat B,PetscReal tol,PetscTruth *flg) 3580 { 3581 PetscErrorCode ierr,(*f)(Mat,Mat,PetscReal,PetscTruth*),(*g)(Mat,Mat,PetscReal,PetscTruth*); 3582 3583 PetscFunctionBegin; 3584 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 3585 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 3586 PetscValidPointer(flg,3); 3587 ierr = PetscObjectQueryFunction((PetscObject)A,"MatIsHermitianTranspose_C",(void (**)(void))&f);CHKERRQ(ierr); 3588 ierr = PetscObjectQueryFunction((PetscObject)B,"MatIsHermitianTranspose_C",(void (**)(void))&g);CHKERRQ(ierr); 3589 if (f && g) { 3590 if (f==g) { 3591 ierr = (*f)(A,B,tol,flg);CHKERRQ(ierr); 3592 } else { 3593 SETERRQ(PETSC_ERR_ARG_NOTSAMETYPE,"Matrices do not have the same comparator for Hermitian test"); 3594 } 3595 } 3596 PetscFunctionReturn(0); 3597 } 3598 3599 #undef __FUNCT__ 3600 #define __FUNCT__ "MatPermute" 3601 /*@C 3602 MatPermute - Creates a new matrix with rows and columns permuted from the 3603 original. 3604 3605 Collective on Mat 3606 3607 Input Parameters: 3608 + mat - the matrix to permute 3609 . row - row permutation, each processor supplies only the permutation for its rows 3610 - col - column permutation, each processor needs the entire column permutation, that is 3611 this is the same size as the total number of columns in the matrix 3612 3613 Output Parameters: 3614 . B - the permuted matrix 3615 3616 Level: advanced 3617 3618 Concepts: matrices^permuting 3619 3620 .seealso: MatGetOrdering() 3621 @*/ 3622 PetscErrorCode PETSCMAT_DLLEXPORT MatPermute(Mat mat,IS row,IS col,Mat *B) 3623 { 3624 PetscErrorCode ierr; 3625 3626 PetscFunctionBegin; 3627 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3628 PetscValidType(mat,1); 3629 PetscValidHeaderSpecific(row,IS_COOKIE,2); 3630 PetscValidHeaderSpecific(col,IS_COOKIE,3); 3631 PetscValidPointer(B,4); 3632 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3633 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3634 if (!mat->ops->permute) SETERRQ1(PETSC_ERR_SUP,"MatPermute not available for Mat type %s",mat->type_name); 3635 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3636 3637 ierr = (*mat->ops->permute)(mat,row,col,B);CHKERRQ(ierr); 3638 ierr = PetscObjectStateIncrease((PetscObject)*B);CHKERRQ(ierr); 3639 PetscFunctionReturn(0); 3640 } 3641 3642 #undef __FUNCT__ 3643 #define __FUNCT__ "MatPermuteSparsify" 3644 /*@C 3645 MatPermuteSparsify - Creates a new matrix with rows and columns permuted from the 3646 original and sparsified to the prescribed tolerance. 3647 3648 Collective on Mat 3649 3650 Input Parameters: 3651 + A - The matrix to permute 3652 . band - The half-bandwidth of the sparsified matrix, or PETSC_DECIDE 3653 . frac - The half-bandwidth as a fraction of the total size, or 0.0 3654 . tol - The drop tolerance 3655 . rowp - The row permutation 3656 - colp - The column permutation 3657 3658 Output Parameter: 3659 . B - The permuted, sparsified matrix 3660 3661 Level: advanced 3662 3663 Note: 3664 The default behavior (band = PETSC_DECIDE and frac = 0.0) is to 3665 restrict the half-bandwidth of the resulting matrix to 5% of the 3666 total matrix size. 3667 3668 .keywords: matrix, permute, sparsify 3669 3670 .seealso: MatGetOrdering(), MatPermute() 3671 @*/ 3672 PetscErrorCode PETSCMAT_DLLEXPORT MatPermuteSparsify(Mat A, PetscInt band, PetscReal frac, PetscReal tol, IS rowp, IS colp, Mat *B) 3673 { 3674 IS irowp, icolp; 3675 PetscInt *rows, *cols; 3676 PetscInt M, N, locRowStart, locRowEnd; 3677 PetscInt nz, newNz; 3678 const PetscInt *cwork; 3679 PetscInt *cnew; 3680 const PetscScalar *vwork; 3681 PetscScalar *vnew; 3682 PetscInt bw, issize; 3683 PetscInt row, locRow, newRow, col, newCol; 3684 PetscErrorCode ierr; 3685 3686 PetscFunctionBegin; 3687 PetscValidHeaderSpecific(A, MAT_COOKIE,1); 3688 PetscValidHeaderSpecific(rowp, IS_COOKIE,5); 3689 PetscValidHeaderSpecific(colp, IS_COOKIE,6); 3690 PetscValidPointer(B,7); 3691 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE, "Not for unassembled matrix"); 3692 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix"); 3693 if (!A->ops->permutesparsify) { 3694 ierr = MatGetSize(A, &M, &N);CHKERRQ(ierr); 3695 ierr = MatGetOwnershipRange(A, &locRowStart, &locRowEnd);CHKERRQ(ierr); 3696 ierr = ISGetSize(rowp, &issize);CHKERRQ(ierr); 3697 if (issize != M) SETERRQ2(PETSC_ERR_ARG_WRONG, "Wrong size %D for row permutation, should be %D", issize, M); 3698 ierr = ISGetSize(colp, &issize);CHKERRQ(ierr); 3699 if (issize != N) SETERRQ2(PETSC_ERR_ARG_WRONG, "Wrong size %D for column permutation, should be %D", issize, N); 3700 ierr = ISInvertPermutation(rowp, 0, &irowp);CHKERRQ(ierr); 3701 ierr = ISGetIndices(irowp, &rows);CHKERRQ(ierr); 3702 ierr = ISInvertPermutation(colp, 0, &icolp);CHKERRQ(ierr); 3703 ierr = ISGetIndices(icolp, &cols);CHKERRQ(ierr); 3704 ierr = PetscMalloc(N * sizeof(PetscInt), &cnew);CHKERRQ(ierr); 3705 ierr = PetscMalloc(N * sizeof(PetscScalar), &vnew);CHKERRQ(ierr); 3706 3707 /* Setup bandwidth to include */ 3708 if (band == PETSC_DECIDE) { 3709 if (frac <= 0.0) 3710 bw = (PetscInt) (M * 0.05); 3711 else 3712 bw = (PetscInt) (M * frac); 3713 } else { 3714 if (band <= 0) SETERRQ(PETSC_ERR_ARG_WRONG, "Bandwidth must be a positive integer"); 3715 bw = band; 3716 } 3717 3718 /* Put values into new matrix */ 3719 ierr = MatDuplicate(A, MAT_DO_NOT_COPY_VALUES, B);CHKERRQ(ierr); 3720 for(row = locRowStart, locRow = 0; row < locRowEnd; row++, locRow++) { 3721 ierr = MatGetRow(A, row, &nz, &cwork, &vwork);CHKERRQ(ierr); 3722 newRow = rows[locRow]+locRowStart; 3723 for(col = 0, newNz = 0; col < nz; col++) { 3724 newCol = cols[cwork[col]]; 3725 if ((newCol >= newRow - bw) && (newCol < newRow + bw) && (PetscAbsScalar(vwork[col]) >= tol)) { 3726 cnew[newNz] = newCol; 3727 vnew[newNz] = vwork[col]; 3728 newNz++; 3729 } 3730 } 3731 ierr = MatSetValues(*B, 1, &newRow, newNz, cnew, vnew, INSERT_VALUES);CHKERRQ(ierr); 3732 ierr = MatRestoreRow(A, row, &nz, &cwork, &vwork);CHKERRQ(ierr); 3733 } 3734 ierr = PetscFree(cnew);CHKERRQ(ierr); 3735 ierr = PetscFree(vnew);CHKERRQ(ierr); 3736 ierr = MatAssemblyBegin(*B, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3737 ierr = MatAssemblyEnd(*B, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3738 ierr = ISRestoreIndices(irowp, &rows);CHKERRQ(ierr); 3739 ierr = ISRestoreIndices(icolp, &cols);CHKERRQ(ierr); 3740 ierr = ISDestroy(irowp);CHKERRQ(ierr); 3741 ierr = ISDestroy(icolp);CHKERRQ(ierr); 3742 } else { 3743 ierr = (*A->ops->permutesparsify)(A, band, frac, tol, rowp, colp, B);CHKERRQ(ierr); 3744 } 3745 ierr = PetscObjectStateIncrease((PetscObject)*B);CHKERRQ(ierr); 3746 PetscFunctionReturn(0); 3747 } 3748 3749 #undef __FUNCT__ 3750 #define __FUNCT__ "MatEqual" 3751 /*@ 3752 MatEqual - Compares two matrices. 3753 3754 Collective on Mat 3755 3756 Input Parameters: 3757 + A - the first matrix 3758 - B - the second matrix 3759 3760 Output Parameter: 3761 . flg - PETSC_TRUE if the matrices are equal; PETSC_FALSE otherwise. 3762 3763 Level: intermediate 3764 3765 Concepts: matrices^equality between 3766 @*/ 3767 PetscErrorCode PETSCMAT_DLLEXPORT MatEqual(Mat A,Mat B,PetscTruth *flg) 3768 { 3769 PetscErrorCode ierr; 3770 3771 PetscFunctionBegin; 3772 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 3773 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 3774 PetscValidType(A,1); 3775 PetscValidType(B,2); 3776 MatPreallocated(B); 3777 PetscValidIntPointer(flg,3); 3778 PetscCheckSameComm(A,1,B,2); 3779 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3780 if (!B->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3781 if (A->rmap.N != B->rmap.N || A->cmap.N != B->cmap.N) SETERRQ4(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); 3782 if (!A->ops->equal) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",A->type_name); 3783 if (!B->ops->equal) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",B->type_name); 3784 if (A->ops->equal != B->ops->equal) SETERRQ2(PETSC_ERR_ARG_INCOMP,"A is type: %s\nB is type: %s",A->type_name,B->type_name); 3785 ierr = MatPreallocated(A);CHKERRQ(ierr); 3786 3787 ierr = (*A->ops->equal)(A,B,flg);CHKERRQ(ierr); 3788 PetscFunctionReturn(0); 3789 } 3790 3791 #undef __FUNCT__ 3792 #define __FUNCT__ "MatDiagonalScale" 3793 /*@ 3794 MatDiagonalScale - Scales a matrix on the left and right by diagonal 3795 matrices that are stored as vectors. Either of the two scaling 3796 matrices can be PETSC_NULL. 3797 3798 Collective on Mat 3799 3800 Input Parameters: 3801 + mat - the matrix to be scaled 3802 . l - the left scaling vector (or PETSC_NULL) 3803 - r - the right scaling vector (or PETSC_NULL) 3804 3805 Notes: 3806 MatDiagonalScale() computes A = LAR, where 3807 L = a diagonal matrix (stored as a vector), R = a diagonal matrix (stored as a vector) 3808 3809 Level: intermediate 3810 3811 Concepts: matrices^diagonal scaling 3812 Concepts: diagonal scaling of matrices 3813 3814 .seealso: MatScale() 3815 @*/ 3816 PetscErrorCode PETSCMAT_DLLEXPORT MatDiagonalScale(Mat mat,Vec l,Vec r) 3817 { 3818 PetscErrorCode ierr; 3819 3820 PetscFunctionBegin; 3821 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3822 PetscValidType(mat,1); 3823 if (!mat->ops->diagonalscale) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3824 if (l) {PetscValidHeaderSpecific(l,VEC_COOKIE,2);PetscCheckSameComm(mat,1,l,2);} 3825 if (r) {PetscValidHeaderSpecific(r,VEC_COOKIE,3);PetscCheckSameComm(mat,1,r,3);} 3826 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3827 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3828 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3829 3830 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 3831 ierr = (*mat->ops->diagonalscale)(mat,l,r);CHKERRQ(ierr); 3832 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 3833 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 3834 PetscFunctionReturn(0); 3835 } 3836 3837 #undef __FUNCT__ 3838 #define __FUNCT__ "MatScale" 3839 /*@ 3840 MatScale - Scales all elements of a matrix by a given number. 3841 3842 Collective on Mat 3843 3844 Input Parameters: 3845 + mat - the matrix to be scaled 3846 - a - the scaling value 3847 3848 Output Parameter: 3849 . mat - the scaled matrix 3850 3851 Level: intermediate 3852 3853 Concepts: matrices^scaling all entries 3854 3855 .seealso: MatDiagonalScale() 3856 @*/ 3857 PetscErrorCode PETSCMAT_DLLEXPORT MatScale(Mat mat,PetscScalar a) 3858 { 3859 PetscErrorCode ierr; 3860 3861 PetscFunctionBegin; 3862 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3863 PetscValidType(mat,1); 3864 if (!mat->ops->scale) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3865 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3866 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3867 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3868 3869 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 3870 ierr = (*mat->ops->scale)(mat,a);CHKERRQ(ierr); 3871 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 3872 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 3873 PetscFunctionReturn(0); 3874 } 3875 3876 #undef __FUNCT__ 3877 #define __FUNCT__ "MatNorm" 3878 /*@ 3879 MatNorm - Calculates various norms of a matrix. 3880 3881 Collective on Mat 3882 3883 Input Parameters: 3884 + mat - the matrix 3885 - type - the type of norm, NORM_1, NORM_FROBENIUS, NORM_INFINITY 3886 3887 Output Parameters: 3888 . nrm - the resulting norm 3889 3890 Level: intermediate 3891 3892 Concepts: matrices^norm 3893 Concepts: norm^of matrix 3894 @*/ 3895 PetscErrorCode PETSCMAT_DLLEXPORT MatNorm(Mat mat,NormType type,PetscReal *nrm) 3896 { 3897 PetscErrorCode ierr; 3898 3899 PetscFunctionBegin; 3900 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3901 PetscValidType(mat,1); 3902 PetscValidScalarPointer(nrm,3); 3903 3904 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3905 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3906 if (!mat->ops->norm) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 3907 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3908 3909 ierr = (*mat->ops->norm)(mat,type,nrm);CHKERRQ(ierr); 3910 PetscFunctionReturn(0); 3911 } 3912 3913 /* 3914 This variable is used to prevent counting of MatAssemblyBegin() that 3915 are called from within a MatAssemblyEnd(). 3916 */ 3917 static PetscInt MatAssemblyEnd_InUse = 0; 3918 #undef __FUNCT__ 3919 #define __FUNCT__ "MatAssemblyBegin" 3920 /*@ 3921 MatAssemblyBegin - Begins assembling the matrix. This routine should 3922 be called after completing all calls to MatSetValues(). 3923 3924 Collective on Mat 3925 3926 Input Parameters: 3927 + mat - the matrix 3928 - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY 3929 3930 Notes: 3931 MatSetValues() generally caches the values. The matrix is ready to 3932 use only after MatAssemblyBegin() and MatAssemblyEnd() have been called. 3933 Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES 3934 in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before 3935 using the matrix. 3936 3937 Level: beginner 3938 3939 Concepts: matrices^assembling 3940 3941 .seealso: MatAssemblyEnd(), MatSetValues(), MatAssembled() 3942 @*/ 3943 PetscErrorCode PETSCMAT_DLLEXPORT MatAssemblyBegin(Mat mat,MatAssemblyType type) 3944 { 3945 PetscErrorCode ierr; 3946 3947 PetscFunctionBegin; 3948 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3949 PetscValidType(mat,1); 3950 ierr = MatPreallocated(mat);CHKERRQ(ierr); 3951 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix.\nDid you forget to call MatSetUnfactored()?"); 3952 if (mat->assembled) { 3953 mat->was_assembled = PETSC_TRUE; 3954 mat->assembled = PETSC_FALSE; 3955 } 3956 if (!MatAssemblyEnd_InUse) { 3957 ierr = PetscLogEventBegin(MAT_AssemblyBegin,mat,0,0,0);CHKERRQ(ierr); 3958 if (mat->ops->assemblybegin){ierr = (*mat->ops->assemblybegin)(mat,type);CHKERRQ(ierr);} 3959 ierr = PetscLogEventEnd(MAT_AssemblyBegin,mat,0,0,0);CHKERRQ(ierr); 3960 } else { 3961 if (mat->ops->assemblybegin){ierr = (*mat->ops->assemblybegin)(mat,type);CHKERRQ(ierr);} 3962 } 3963 PetscFunctionReturn(0); 3964 } 3965 3966 #undef __FUNCT__ 3967 #define __FUNCT__ "MatAssembed" 3968 /*@ 3969 MatAssembled - Indicates if a matrix has been assembled and is ready for 3970 use; for example, in matrix-vector product. 3971 3972 Collective on Mat 3973 3974 Input Parameter: 3975 . mat - the matrix 3976 3977 Output Parameter: 3978 . assembled - PETSC_TRUE or PETSC_FALSE 3979 3980 Level: advanced 3981 3982 Concepts: matrices^assembled? 3983 3984 .seealso: MatAssemblyEnd(), MatSetValues(), MatAssemblyBegin() 3985 @*/ 3986 PetscErrorCode PETSCMAT_DLLEXPORT MatAssembled(Mat mat,PetscTruth *assembled) 3987 { 3988 PetscFunctionBegin; 3989 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 3990 PetscValidType(mat,1); 3991 PetscValidPointer(assembled,2); 3992 *assembled = mat->assembled; 3993 PetscFunctionReturn(0); 3994 } 3995 3996 #undef __FUNCT__ 3997 #define __FUNCT__ "MatView_Private" 3998 /* 3999 Processes command line options to determine if/how a matrix 4000 is to be viewed. Called by MatAssemblyEnd() and MatLoad(). 4001 */ 4002 PetscErrorCode MatView_Private(Mat mat) 4003 { 4004 PetscErrorCode ierr; 4005 PetscTruth flg1,flg2,flg3,flg4,flg6,flg7,flg8; 4006 static PetscTruth incall = PETSC_FALSE; 4007 #if defined(PETSC_USE_SOCKET_VIEWER) 4008 PetscTruth flg5; 4009 #endif 4010 4011 PetscFunctionBegin; 4012 if (incall) PetscFunctionReturn(0); 4013 incall = PETSC_TRUE; 4014 ierr = PetscOptionsBegin(mat->comm,mat->prefix,"Matrix Options","Mat");CHKERRQ(ierr); 4015 ierr = PetscOptionsName("-mat_view_info","Information on matrix size","MatView",&flg1);CHKERRQ(ierr); 4016 ierr = PetscOptionsName("-mat_view_info_detailed","Nonzeros in the matrix","MatView",&flg2);CHKERRQ(ierr); 4017 ierr = PetscOptionsName("-mat_view","Print matrix to stdout","MatView",&flg3);CHKERRQ(ierr); 4018 ierr = PetscOptionsName("-mat_view_matlab","Print matrix to stdout in a format Matlab can read","MatView",&flg4);CHKERRQ(ierr); 4019 #if defined(PETSC_USE_SOCKET_VIEWER) 4020 ierr = PetscOptionsName("-mat_view_socket","Send matrix to socket (can be read from matlab)","MatView",&flg5);CHKERRQ(ierr); 4021 #endif 4022 ierr = PetscOptionsName("-mat_view_binary","Save matrix to file in binary format","MatView",&flg6);CHKERRQ(ierr); 4023 ierr = PetscOptionsName("-mat_view_draw","Draw the matrix nonzero structure","MatView",&flg7);CHKERRQ(ierr); 4024 ierr = PetscOptionsEnd();CHKERRQ(ierr); 4025 4026 if (flg1) { 4027 PetscViewer viewer; 4028 4029 ierr = PetscViewerASCIIGetStdout(mat->comm,&viewer);CHKERRQ(ierr); 4030 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_INFO);CHKERRQ(ierr); 4031 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4032 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4033 } 4034 if (flg2) { 4035 PetscViewer viewer; 4036 4037 ierr = PetscViewerASCIIGetStdout(mat->comm,&viewer);CHKERRQ(ierr); 4038 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_INFO_DETAIL);CHKERRQ(ierr); 4039 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4040 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4041 } 4042 if (flg3) { 4043 PetscViewer viewer; 4044 4045 ierr = PetscViewerASCIIGetStdout(mat->comm,&viewer);CHKERRQ(ierr); 4046 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4047 } 4048 if (flg4) { 4049 PetscViewer viewer; 4050 4051 ierr = PetscViewerASCIIGetStdout(mat->comm,&viewer);CHKERRQ(ierr); 4052 ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_MATLAB);CHKERRQ(ierr); 4053 ierr = MatView(mat,viewer);CHKERRQ(ierr); 4054 ierr = PetscViewerPopFormat(viewer);CHKERRQ(ierr); 4055 } 4056 #if defined(PETSC_USE_SOCKET_VIEWER) 4057 if (flg5) { 4058 ierr = MatView(mat,PETSC_VIEWER_SOCKET_(mat->comm));CHKERRQ(ierr); 4059 ierr = PetscViewerFlush(PETSC_VIEWER_SOCKET_(mat->comm));CHKERRQ(ierr); 4060 } 4061 #endif 4062 if (flg6) { 4063 ierr = MatView(mat,PETSC_VIEWER_BINARY_(mat->comm));CHKERRQ(ierr); 4064 ierr = PetscViewerFlush(PETSC_VIEWER_BINARY_(mat->comm));CHKERRQ(ierr); 4065 } 4066 if (flg7) { 4067 ierr = PetscOptionsHasName(mat->prefix,"-mat_view_contour",&flg8);CHKERRQ(ierr); 4068 if (flg8) { 4069 PetscViewerPushFormat(PETSC_VIEWER_DRAW_(mat->comm),PETSC_VIEWER_DRAW_CONTOUR);CHKERRQ(ierr); 4070 } 4071 ierr = MatView(mat,PETSC_VIEWER_DRAW_(mat->comm));CHKERRQ(ierr); 4072 ierr = PetscViewerFlush(PETSC_VIEWER_DRAW_(mat->comm));CHKERRQ(ierr); 4073 if (flg8) { 4074 PetscViewerPopFormat(PETSC_VIEWER_DRAW_(mat->comm));CHKERRQ(ierr); 4075 } 4076 } 4077 incall = PETSC_FALSE; 4078 PetscFunctionReturn(0); 4079 } 4080 4081 #undef __FUNCT__ 4082 #define __FUNCT__ "MatAssemblyEnd" 4083 /*@ 4084 MatAssemblyEnd - Completes assembling the matrix. This routine should 4085 be called after MatAssemblyBegin(). 4086 4087 Collective on Mat 4088 4089 Input Parameters: 4090 + mat - the matrix 4091 - type - type of assembly, either MAT_FLUSH_ASSEMBLY or MAT_FINAL_ASSEMBLY 4092 4093 Options Database Keys: 4094 + -mat_view_info - Prints info on matrix at conclusion of MatEndAssembly() 4095 . -mat_view_info_detailed - Prints more detailed info 4096 . -mat_view - Prints matrix in ASCII format 4097 . -mat_view_matlab - Prints matrix in Matlab format 4098 . -mat_view_draw - PetscDraws nonzero structure of matrix, using MatView() and PetscDrawOpenX(). 4099 . -display <name> - Sets display name (default is host) 4100 . -draw_pause <sec> - Sets number of seconds to pause after display 4101 . -mat_view_socket - Sends matrix to socket, can be accessed from Matlab (see users manual) 4102 . -viewer_socket_machine <machine> 4103 . -viewer_socket_port <port> 4104 . -mat_view_binary - save matrix to file in binary format 4105 - -viewer_binary_filename <name> 4106 4107 Notes: 4108 MatSetValues() generally caches the values. The matrix is ready to 4109 use only after MatAssemblyBegin() and MatAssemblyEnd() have been called. 4110 Use MAT_FLUSH_ASSEMBLY when switching between ADD_VALUES and INSERT_VALUES 4111 in MatSetValues(); use MAT_FINAL_ASSEMBLY for the final assembly before 4112 using the matrix. 4113 4114 Level: beginner 4115 4116 .seealso: MatAssemblyBegin(), MatSetValues(), PetscDrawOpenX(), MatView(), MatAssembled(), PetscViewerSocketOpen() 4117 @*/ 4118 PetscErrorCode PETSCMAT_DLLEXPORT MatAssemblyEnd(Mat mat,MatAssemblyType type) 4119 { 4120 PetscErrorCode ierr; 4121 static PetscInt inassm = 0; 4122 PetscTruth flg; 4123 4124 PetscFunctionBegin; 4125 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4126 PetscValidType(mat,1); 4127 4128 inassm++; 4129 MatAssemblyEnd_InUse++; 4130 if (MatAssemblyEnd_InUse == 1) { /* Do the logging only the first time through */ 4131 ierr = PetscLogEventBegin(MAT_AssemblyEnd,mat,0,0,0);CHKERRQ(ierr); 4132 if (mat->ops->assemblyend) { 4133 ierr = (*mat->ops->assemblyend)(mat,type);CHKERRQ(ierr); 4134 } 4135 ierr = PetscLogEventEnd(MAT_AssemblyEnd,mat,0,0,0);CHKERRQ(ierr); 4136 } else { 4137 if (mat->ops->assemblyend) { 4138 ierr = (*mat->ops->assemblyend)(mat,type);CHKERRQ(ierr); 4139 } 4140 } 4141 4142 /* Flush assembly is not a true assembly */ 4143 if (type != MAT_FLUSH_ASSEMBLY) { 4144 mat->assembled = PETSC_TRUE; mat->num_ass++; 4145 } 4146 mat->insertmode = NOT_SET_VALUES; 4147 MatAssemblyEnd_InUse--; 4148 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4149 if (!mat->symmetric_eternal) { 4150 mat->symmetric_set = PETSC_FALSE; 4151 mat->hermitian_set = PETSC_FALSE; 4152 mat->structurally_symmetric_set = PETSC_FALSE; 4153 } 4154 if (inassm == 1 && type != MAT_FLUSH_ASSEMBLY) { 4155 ierr = MatView_Private(mat);CHKERRQ(ierr); 4156 ierr = PetscOptionsHasName(mat->prefix,"-mat_is_symmetric",&flg);CHKERRQ(ierr); 4157 if (flg) { 4158 PetscReal tol = 0.0; 4159 ierr = PetscOptionsGetReal(mat->prefix,"-mat_is_symmetric",&tol,PETSC_NULL);CHKERRQ(ierr); 4160 ierr = MatIsSymmetric(mat,tol,&flg);CHKERRQ(ierr); 4161 if (flg) { 4162 ierr = PetscPrintf(mat->comm,"Matrix is symmetric (tolerance %G)\n",tol);CHKERRQ(ierr); 4163 } else { 4164 ierr = PetscPrintf(mat->comm,"Matrix is not symmetric (tolerance %G)\n",tol);CHKERRQ(ierr); 4165 } 4166 } 4167 } 4168 inassm--; 4169 PetscFunctionReturn(0); 4170 } 4171 4172 4173 #undef __FUNCT__ 4174 #define __FUNCT__ "MatCompress" 4175 /*@ 4176 MatCompress - Tries to store the matrix in as little space as 4177 possible. May fail if memory is already fully used, since it 4178 tries to allocate new space. 4179 4180 Collective on Mat 4181 4182 Input Parameters: 4183 . mat - the matrix 4184 4185 Level: advanced 4186 4187 @*/ 4188 PetscErrorCode PETSCMAT_DLLEXPORT MatCompress(Mat mat) 4189 { 4190 PetscErrorCode ierr; 4191 4192 PetscFunctionBegin; 4193 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4194 PetscValidType(mat,1); 4195 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4196 if (mat->ops->compress) {ierr = (*mat->ops->compress)(mat);CHKERRQ(ierr);} 4197 PetscFunctionReturn(0); 4198 } 4199 4200 #undef __FUNCT__ 4201 #define __FUNCT__ "MatSetOption" 4202 /*@ 4203 MatSetOption - Sets a parameter option for a matrix. Some options 4204 may be specific to certain storage formats. Some options 4205 determine how values will be inserted (or added). Sorted, 4206 row-oriented input will generally assemble the fastest. The default 4207 is row-oriented, nonsorted input. 4208 4209 Collective on Mat 4210 4211 Input Parameters: 4212 + mat - the matrix 4213 - option - the option, one of those listed below (and possibly others), 4214 4215 Options Describing Matrix Structure: 4216 + MAT_SYMMETRIC - symmetric in terms of both structure and value 4217 . MAT_HERMITIAN - transpose is the complex conjugation 4218 . MAT_STRUCTURALLY_SYMMETRIC - symmetric nonzero structure 4219 - MAT_SYMMETRY_ETERNAL - if you would like the symmetry/Hermitian flag 4220 you set to be kept with all future use of the matrix 4221 including after MatAssemblyBegin/End() which could 4222 potentially change the symmetry structure, i.e. you 4223 KNOW the matrix will ALWAYS have the property you set. 4224 4225 4226 Options For Use with MatSetValues(): 4227 Insert a logically dense subblock, which can be 4228 . MAT_ROW_ORIENTED - row-oriented (default) 4229 4230 Not these options reflect the data you pass in with MatSetValues(); it has 4231 nothing to do with how the data is stored internally in the matrix 4232 data structure. 4233 4234 When (re)assembling a matrix, we can restrict the input for 4235 efficiency/debugging purposes. These options include 4236 + MAT_NEW_NONZERO_LOCATIONS - additional insertions will be 4237 allowed if they generate a new nonzero 4238 . MAT_NEW_DIAGONALS - new diagonals will be allowed (for block diagonal format only) 4239 . MAT_IGNORE_OFF_PROC_ENTRIES - drops off-processor entries 4240 . MAT_NEW_NONZERO_LOCATION_ERR - generates an error for new matrix entry 4241 - MAT_USE_HASH_TABLE - uses a hash table to speed up matrix assembly 4242 4243 Notes: 4244 Some options are relevant only for particular matrix types and 4245 are thus ignored by others. Other options are not supported by 4246 certain matrix types and will generate an error message if set. 4247 4248 If using a Fortran 77 module to compute a matrix, one may need to 4249 use the column-oriented option (or convert to the row-oriented 4250 format). 4251 4252 MAT_NEW_NONZERO_LOCATIONS set to PETSC_FALSE indicates that any add or insertion 4253 that would generate a new entry in the nonzero structure is instead 4254 ignored. Thus, if memory has not alredy been allocated for this particular 4255 data, then the insertion is ignored. For dense matrices, in which 4256 the entire array is allocated, no entries are ever ignored. 4257 Set after the first MatAssemblyEnd() 4258 4259 MAT_NEW_NONZERO_LOCATION_ERR indicates that any add or insertion 4260 that would generate a new entry in the nonzero structure instead produces 4261 an error. (Currently supported for AIJ and BAIJ formats only.) 4262 This is a useful flag when using SAME_NONZERO_PATTERN in calling 4263 KSPSetOperators() to ensure that the nonzero pattern truely does 4264 remain unchanged. Set after the first MatAssemblyEnd() 4265 4266 MAT_NEW_NONZERO_ALLOCATION_ERR indicates that any add or insertion 4267 that would generate a new entry that has not been preallocated will 4268 instead produce an error. (Currently supported for AIJ and BAIJ formats 4269 only.) This is a useful flag when debugging matrix memory preallocation. 4270 4271 MAT_IGNORE_OFF_PROC_ENTRIES indicates entries destined for 4272 other processors should be dropped, rather than stashed. 4273 This is useful if you know that the "owning" processor is also 4274 always generating the correct matrix entries, so that PETSc need 4275 not transfer duplicate entries generated on another processor. 4276 4277 MAT_USE_HASH_TABLE indicates that a hash table be used to improve the 4278 searches during matrix assembly. When this flag is set, the hash table 4279 is created during the first Matrix Assembly. This hash table is 4280 used the next time through, during MatSetVaules()/MatSetVaulesBlocked() 4281 to improve the searching of indices. MAT_NEW_NONZERO_LOCATIONS flag 4282 should be used with MAT_USE_HASH_TABLE flag. This option is currently 4283 supported by MATMPIBAIJ format only. 4284 4285 MAT_KEEP_ZEROED_ROWS indicates when MatZeroRows() is called the zeroed entries 4286 are kept in the nonzero structure 4287 4288 MAT_IGNORE_ZERO_ENTRIES - for AIJ/IS matrices this will stop zero values from creating 4289 a zero location in the matrix 4290 4291 MAT_USE_INODES - indicates using inode version of the code - works with AIJ and 4292 ROWBS matrix types 4293 4294 Level: intermediate 4295 4296 Concepts: matrices^setting options 4297 4298 @*/ 4299 PetscErrorCode PETSCMAT_DLLEXPORT MatSetOption(Mat mat,MatOption op,PetscTruth flg) 4300 { 4301 PetscErrorCode ierr; 4302 4303 PetscFunctionBegin; 4304 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4305 PetscValidType(mat,1); 4306 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4307 switch (op) { 4308 case MAT_SYMMETRIC: 4309 mat->symmetric = flg; 4310 if (flg) mat->structurally_symmetric = PETSC_TRUE; 4311 mat->symmetric_set = PETSC_TRUE; 4312 mat->structurally_symmetric_set = flg; 4313 break; 4314 case MAT_HERMITIAN: 4315 mat->hermitian = flg; 4316 if (flg) mat->structurally_symmetric = PETSC_TRUE; 4317 mat->hermitian_set = PETSC_TRUE; 4318 mat->structurally_symmetric_set = flg; 4319 break; 4320 case MAT_STRUCTURALLY_SYMMETRIC: 4321 mat->structurally_symmetric = flg; 4322 mat->structurally_symmetric_set = PETSC_TRUE; 4323 break; 4324 case MAT_SYMMETRY_ETERNAL: 4325 mat->symmetric_eternal = flg; 4326 break; 4327 default: 4328 break; 4329 } 4330 if (mat->ops->setoption) { 4331 ierr = (*mat->ops->setoption)(mat,op,flg);CHKERRQ(ierr); 4332 } 4333 PetscFunctionReturn(0); 4334 } 4335 4336 #undef __FUNCT__ 4337 #define __FUNCT__ "MatZeroEntries" 4338 /*@ 4339 MatZeroEntries - Zeros all entries of a matrix. For sparse matrices 4340 this routine retains the old nonzero structure. 4341 4342 Collective on Mat 4343 4344 Input Parameters: 4345 . mat - the matrix 4346 4347 Level: intermediate 4348 4349 Concepts: matrices^zeroing 4350 4351 .seealso: MatZeroRows() 4352 @*/ 4353 PetscErrorCode PETSCMAT_DLLEXPORT MatZeroEntries(Mat mat) 4354 { 4355 PetscErrorCode ierr; 4356 4357 PetscFunctionBegin; 4358 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4359 PetscValidType(mat,1); 4360 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4361 if (mat->insertmode != NOT_SET_VALUES) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for matrices where you have set values but not yet assembled"); 4362 if (!mat->ops->zeroentries) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 4363 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4364 4365 ierr = PetscLogEventBegin(MAT_ZeroEntries,mat,0,0,0);CHKERRQ(ierr); 4366 ierr = (*mat->ops->zeroentries)(mat);CHKERRQ(ierr); 4367 ierr = PetscLogEventEnd(MAT_ZeroEntries,mat,0,0,0);CHKERRQ(ierr); 4368 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4369 PetscFunctionReturn(0); 4370 } 4371 4372 #undef __FUNCT__ 4373 #define __FUNCT__ "MatZeroRows" 4374 /*@C 4375 MatZeroRows - Zeros all entries (except possibly the main diagonal) 4376 of a set of rows of a matrix. 4377 4378 Collective on Mat 4379 4380 Input Parameters: 4381 + mat - the matrix 4382 . numRows - the number of rows to remove 4383 . rows - the global row indices 4384 - diag - value put in all diagonals of eliminated rows 4385 4386 Notes: 4387 For the AIJ and BAIJ matrix formats this removes the old nonzero structure, 4388 but does not release memory. For the dense and block diagonal 4389 formats this does not alter the nonzero structure. 4390 4391 If the option MatSetOption(mat,MAT_KEEP_ZEROED_ROWS,PETSC_TRUE) the nonzero structure 4392 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 4393 merely zeroed. 4394 4395 The user can set a value in the diagonal entry (or for the AIJ and 4396 row formats can optionally remove the main diagonal entry from the 4397 nonzero structure as well, by passing 0.0 as the final argument). 4398 4399 For the parallel case, all processes that share the matrix (i.e., 4400 those in the communicator used for matrix creation) MUST call this 4401 routine, regardless of whether any rows being zeroed are owned by 4402 them. 4403 4404 Each processor should list the rows that IT wants zeroed 4405 4406 Level: intermediate 4407 4408 Concepts: matrices^zeroing rows 4409 4410 .seealso: MatZeroRowsIS(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption() 4411 @*/ 4412 PetscErrorCode PETSCMAT_DLLEXPORT MatZeroRows(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag) 4413 { 4414 PetscErrorCode ierr; 4415 4416 PetscFunctionBegin; 4417 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4418 PetscValidType(mat,1); 4419 if (numRows) PetscValidIntPointer(rows,3); 4420 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4421 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4422 if (!mat->ops->zerorows) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 4423 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4424 4425 ierr = (*mat->ops->zerorows)(mat,numRows,rows,diag);CHKERRQ(ierr); 4426 ierr = MatView_Private(mat);CHKERRQ(ierr); 4427 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4428 PetscFunctionReturn(0); 4429 } 4430 4431 #undef __FUNCT__ 4432 #define __FUNCT__ "MatZeroRowsIS" 4433 /*@C 4434 MatZeroRowsIS - Zeros all entries (except possibly the main diagonal) 4435 of a set of rows of a matrix. 4436 4437 Collective on Mat 4438 4439 Input Parameters: 4440 + mat - the matrix 4441 . is - index set of rows to remove 4442 - diag - value put in all diagonals of eliminated rows 4443 4444 Notes: 4445 For the AIJ and BAIJ matrix formats this removes the old nonzero structure, 4446 but does not release memory. For the dense and block diagonal 4447 formats this does not alter the nonzero structure. 4448 4449 If the option MatSetOption(mat,MAT_KEEP_ZEROED_ROWS,PETSC_TRUE) the nonzero structure 4450 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 4451 merely zeroed. 4452 4453 The user can set a value in the diagonal entry (or for the AIJ and 4454 row formats can optionally remove the main diagonal entry from the 4455 nonzero structure as well, by passing 0.0 as the final argument). 4456 4457 For the parallel case, all processes that share the matrix (i.e., 4458 those in the communicator used for matrix creation) MUST call this 4459 routine, regardless of whether any rows being zeroed are owned by 4460 them. 4461 4462 Each processor should list the rows that IT wants zeroed 4463 4464 Level: intermediate 4465 4466 Concepts: matrices^zeroing rows 4467 4468 .seealso: MatZeroRows(), MatZeroEntries(), MatZeroRowsLocal(), MatSetOption() 4469 @*/ 4470 PetscErrorCode PETSCMAT_DLLEXPORT MatZeroRowsIS(Mat mat,IS is,PetscScalar diag) 4471 { 4472 PetscInt numRows; 4473 PetscInt *rows; 4474 PetscErrorCode ierr; 4475 4476 PetscFunctionBegin; 4477 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4478 PetscValidType(mat,1); 4479 PetscValidHeaderSpecific(is,IS_COOKIE,2); 4480 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 4481 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 4482 ierr = MatZeroRows(mat,numRows,rows,diag);CHKERRQ(ierr); 4483 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 4484 PetscFunctionReturn(0); 4485 } 4486 4487 #undef __FUNCT__ 4488 #define __FUNCT__ "MatZeroRowsLocal" 4489 /*@C 4490 MatZeroRowsLocal - Zeros all entries (except possibly the main diagonal) 4491 of a set of rows of a matrix; using local numbering of rows. 4492 4493 Collective on Mat 4494 4495 Input Parameters: 4496 + mat - the matrix 4497 . numRows - the number of rows to remove 4498 . rows - the global row indices 4499 - diag - value put in all diagonals of eliminated rows 4500 4501 Notes: 4502 Before calling MatZeroRowsLocal(), the user must first set the 4503 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 4504 4505 For the AIJ matrix formats this removes the old nonzero structure, 4506 but does not release memory. For the dense and block diagonal 4507 formats this does not alter the nonzero structure. 4508 4509 If the option MatSetOption(mat,MAT_KEEP_ZEROED_ROWS,PETSC_TRUE) the nonzero structure 4510 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 4511 merely zeroed. 4512 4513 The user can set a value in the diagonal entry (or for the AIJ and 4514 row formats can optionally remove the main diagonal entry from the 4515 nonzero structure as well, by passing 0.0 as the final argument). 4516 4517 Level: intermediate 4518 4519 Concepts: matrices^zeroing 4520 4521 .seealso: MatZeroRows(), MatZeroRowsLocalIS(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 4522 @*/ 4523 PetscErrorCode PETSCMAT_DLLEXPORT MatZeroRowsLocal(Mat mat,PetscInt numRows,const PetscInt rows[],PetscScalar diag) 4524 { 4525 PetscErrorCode ierr; 4526 4527 PetscFunctionBegin; 4528 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4529 PetscValidType(mat,1); 4530 if (numRows) PetscValidIntPointer(rows,3); 4531 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4532 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4533 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4534 4535 if (mat->ops->zerorowslocal) { 4536 ierr = (*mat->ops->zerorowslocal)(mat,numRows,rows,diag);CHKERRQ(ierr); 4537 } else { 4538 IS is, newis; 4539 PetscInt *newRows; 4540 4541 if (!mat->mapping) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Need to provide local to global mapping to matrix first"); 4542 ierr = ISCreateGeneral(PETSC_COMM_SELF,numRows,rows,&is);CHKERRQ(ierr); 4543 ierr = ISLocalToGlobalMappingApplyIS(mat->mapping,is,&newis);CHKERRQ(ierr); 4544 ierr = ISGetIndices(newis,&newRows);CHKERRQ(ierr); 4545 ierr = (*mat->ops->zerorows)(mat,numRows,newRows,diag);CHKERRQ(ierr); 4546 ierr = ISRestoreIndices(newis,&newRows);CHKERRQ(ierr); 4547 ierr = ISDestroy(newis);CHKERRQ(ierr); 4548 ierr = ISDestroy(is);CHKERRQ(ierr); 4549 } 4550 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 4551 PetscFunctionReturn(0); 4552 } 4553 4554 #undef __FUNCT__ 4555 #define __FUNCT__ "MatZeroRowsLocalIS" 4556 /*@C 4557 MatZeroRowsLocalIS - Zeros all entries (except possibly the main diagonal) 4558 of a set of rows of a matrix; using local numbering of rows. 4559 4560 Collective on Mat 4561 4562 Input Parameters: 4563 + mat - the matrix 4564 . is - index set of rows to remove 4565 - diag - value put in all diagonals of eliminated rows 4566 4567 Notes: 4568 Before calling MatZeroRowsLocalIS(), the user must first set the 4569 local-to-global mapping by calling MatSetLocalToGlobalMapping(). 4570 4571 For the AIJ matrix formats this removes the old nonzero structure, 4572 but does not release memory. For the dense and block diagonal 4573 formats this does not alter the nonzero structure. 4574 4575 If the option MatSetOption(mat,MAT_KEEP_ZEROED_ROWS,PETSC_TRUE) the nonzero structure 4576 of the matrix is not changed (even for AIJ and BAIJ matrices) the values are 4577 merely zeroed. 4578 4579 The user can set a value in the diagonal entry (or for the AIJ and 4580 row formats can optionally remove the main diagonal entry from the 4581 nonzero structure as well, by passing 0.0 as the final argument). 4582 4583 Level: intermediate 4584 4585 Concepts: matrices^zeroing 4586 4587 .seealso: MatZeroRows(), MatZeroRowsLocal(), MatZeroEntries(), MatZeroRows(), MatSetLocalToGlobalMapping 4588 @*/ 4589 PetscErrorCode PETSCMAT_DLLEXPORT MatZeroRowsLocalIS(Mat mat,IS is,PetscScalar diag) 4590 { 4591 PetscErrorCode ierr; 4592 PetscInt numRows; 4593 PetscInt *rows; 4594 4595 PetscFunctionBegin; 4596 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4597 PetscValidType(mat,1); 4598 PetscValidHeaderSpecific(is,IS_COOKIE,2); 4599 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4600 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4601 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4602 4603 ierr = ISGetLocalSize(is,&numRows);CHKERRQ(ierr); 4604 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 4605 ierr = MatZeroRowsLocal(mat,numRows,rows,diag);CHKERRQ(ierr); 4606 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 4607 PetscFunctionReturn(0); 4608 } 4609 4610 #undef __FUNCT__ 4611 #define __FUNCT__ "MatGetSize" 4612 /*@ 4613 MatGetSize - Returns the numbers of rows and columns in a matrix. 4614 4615 Not Collective 4616 4617 Input Parameter: 4618 . mat - the matrix 4619 4620 Output Parameters: 4621 + m - the number of global rows 4622 - n - the number of global columns 4623 4624 Note: both output parameters can be PETSC_NULL on input. 4625 4626 Level: beginner 4627 4628 Concepts: matrices^size 4629 4630 .seealso: MatGetLocalSize() 4631 @*/ 4632 PetscErrorCode PETSCMAT_DLLEXPORT MatGetSize(Mat mat,PetscInt *m,PetscInt* n) 4633 { 4634 PetscFunctionBegin; 4635 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4636 if (m) *m = mat->rmap.N; 4637 if (n) *n = mat->cmap.N; 4638 PetscFunctionReturn(0); 4639 } 4640 4641 #undef __FUNCT__ 4642 #define __FUNCT__ "MatGetLocalSize" 4643 /*@ 4644 MatGetLocalSize - Returns the number of rows and columns in a matrix 4645 stored locally. This information may be implementation dependent, so 4646 use with care. 4647 4648 Not Collective 4649 4650 Input Parameters: 4651 . mat - the matrix 4652 4653 Output Parameters: 4654 + m - the number of local rows 4655 - n - the number of local columns 4656 4657 Note: both output parameters can be PETSC_NULL on input. 4658 4659 Level: beginner 4660 4661 Concepts: matrices^local size 4662 4663 .seealso: MatGetSize() 4664 @*/ 4665 PetscErrorCode PETSCMAT_DLLEXPORT MatGetLocalSize(Mat mat,PetscInt *m,PetscInt* n) 4666 { 4667 PetscFunctionBegin; 4668 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4669 if (m) PetscValidIntPointer(m,2); 4670 if (n) PetscValidIntPointer(n,3); 4671 if (m) *m = mat->rmap.n; 4672 if (n) *n = mat->cmap.n; 4673 PetscFunctionReturn(0); 4674 } 4675 4676 #undef __FUNCT__ 4677 #define __FUNCT__ "MatGetOwnershipRangeColumn" 4678 /*@ 4679 MatGetOwnershipRangeColumn - Returns the range of matrix columns owned by 4680 this processor. 4681 4682 Not Collective 4683 4684 Input Parameters: 4685 . mat - the matrix 4686 4687 Output Parameters: 4688 + m - the global index of the first local column 4689 - n - one more than the global index of the last local column 4690 4691 Notes: both output parameters can be PETSC_NULL on input. 4692 4693 Level: developer 4694 4695 Concepts: matrices^column ownership 4696 @*/ 4697 PetscErrorCode PETSCMAT_DLLEXPORT MatGetOwnershipRangeColumn(Mat mat,PetscInt *m,PetscInt* n) 4698 { 4699 PetscErrorCode ierr; 4700 4701 PetscFunctionBegin; 4702 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4703 PetscValidType(mat,1); 4704 if (m) PetscValidIntPointer(m,2); 4705 if (n) PetscValidIntPointer(n,3); 4706 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4707 if (m) *m = mat->cmap.rstart; 4708 if (n) *n = mat->cmap.rend; 4709 PetscFunctionReturn(0); 4710 } 4711 4712 #undef __FUNCT__ 4713 #define __FUNCT__ "MatGetOwnershipRange" 4714 /*@ 4715 MatGetOwnershipRange - Returns the range of matrix rows owned by 4716 this processor, assuming that the matrix is laid out with the first 4717 n1 rows on the first processor, the next n2 rows on the second, etc. 4718 For certain parallel layouts this range may not be well defined. 4719 4720 Not Collective 4721 4722 Input Parameters: 4723 . mat - the matrix 4724 4725 Output Parameters: 4726 + m - the global index of the first local row 4727 - n - one more than the global index of the last local row 4728 4729 Note: both output parameters can be PETSC_NULL on input. 4730 4731 Level: beginner 4732 4733 Concepts: matrices^row ownership 4734 4735 .seealso: MatGetOwnershipRanges() 4736 4737 @*/ 4738 PetscErrorCode PETSCMAT_DLLEXPORT MatGetOwnershipRange(Mat mat,PetscInt *m,PetscInt* n) 4739 { 4740 PetscErrorCode ierr; 4741 4742 PetscFunctionBegin; 4743 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4744 PetscValidType(mat,1); 4745 if (m) PetscValidIntPointer(m,2); 4746 if (n) PetscValidIntPointer(n,3); 4747 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4748 if (m) *m = mat->rmap.rstart; 4749 if (n) *n = mat->rmap.rend; 4750 PetscFunctionReturn(0); 4751 } 4752 4753 #undef __FUNCT__ 4754 #define __FUNCT__ "MatGetOwnershipRanges" 4755 /*@C 4756 MatGetOwnershipRanges - Returns the range of matrix rows owned by 4757 each process 4758 4759 Not Collective 4760 4761 Input Parameters: 4762 . mat - the matrix 4763 4764 Output Parameters: 4765 . ranges - start of each processors portion plus one more then the total length at the end 4766 4767 Level: beginner 4768 4769 Concepts: matrices^row ownership 4770 4771 .seealso: MatGetOwnershipRange() 4772 4773 @*/ 4774 PetscErrorCode PETSCMAT_DLLEXPORT MatGetOwnershipRanges(Mat mat,const PetscInt **ranges) 4775 { 4776 PetscErrorCode ierr; 4777 4778 PetscFunctionBegin; 4779 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4780 PetscValidType(mat,1); 4781 ierr = PetscMapGetGlobalRange(&mat->rmap,ranges);CHKERRQ(ierr); 4782 PetscFunctionReturn(0); 4783 } 4784 4785 #undef __FUNCT__ 4786 #define __FUNCT__ "MatILUFactorSymbolic" 4787 /*@ 4788 MatILUFactorSymbolic - Performs symbolic ILU factorization of a matrix. 4789 Uses levels of fill only, not drop tolerance. Use MatLUFactorNumeric() 4790 to complete the factorization. 4791 4792 Collective on Mat 4793 4794 Input Parameters: 4795 + mat - the matrix 4796 . row - row permutation 4797 . column - column permutation 4798 - info - structure containing 4799 $ levels - number of levels of fill. 4800 $ expected fill - as ratio of original fill. 4801 $ 1 or 0 - indicating force fill on diagonal (improves robustness for matrices 4802 missing diagonal entries) 4803 4804 Output Parameters: 4805 . fact - new matrix that has been symbolically factored 4806 4807 Notes: 4808 See the users manual for additional information about 4809 choosing the fill factor for better efficiency. 4810 4811 Most users should employ the simplified KSP interface for linear solvers 4812 instead of working directly with matrix algebra routines such as this. 4813 See, e.g., KSPCreate(). 4814 4815 Level: developer 4816 4817 Concepts: matrices^symbolic LU factorization 4818 Concepts: matrices^factorization 4819 Concepts: LU^symbolic factorization 4820 4821 .seealso: MatLUFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor() 4822 MatGetOrdering(), MatFactorInfo 4823 4824 @*/ 4825 PetscErrorCode PETSCMAT_DLLEXPORT MatILUFactorSymbolic(Mat mat,IS row,IS col,MatFactorInfo *info,Mat *fact) 4826 { 4827 PetscErrorCode ierr; 4828 4829 PetscFunctionBegin; 4830 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4831 PetscValidType(mat,1); 4832 PetscValidHeaderSpecific(row,IS_COOKIE,2); 4833 PetscValidHeaderSpecific(col,IS_COOKIE,3); 4834 PetscValidPointer(info,4); 4835 PetscValidPointer(fact,5); 4836 if (info->levels < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Levels of fill negative %D",(PetscInt)info->levels); 4837 if (info->fill < 1.0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %G",info->fill); 4838 if (!mat->ops->ilufactorsymbolic) SETERRQ1(PETSC_ERR_SUP,"Matrix type %s symbolic ILU",mat->type_name); 4839 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4840 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4841 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4842 4843 ierr = PetscLogEventBegin(MAT_ILUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 4844 ierr = (*mat->ops->ilufactorsymbolic)(mat,row,col,info,fact);CHKERRQ(ierr); 4845 ierr = PetscLogEventEnd(MAT_ILUFactorSymbolic,mat,row,col,0);CHKERRQ(ierr); 4846 PetscFunctionReturn(0); 4847 } 4848 4849 #undef __FUNCT__ 4850 #define __FUNCT__ "MatICCFactorSymbolic" 4851 /*@ 4852 MatICCFactorSymbolic - Performs symbolic incomplete 4853 Cholesky factorization for a symmetric matrix. Use 4854 MatCholeskyFactorNumeric() to complete the factorization. 4855 4856 Collective on Mat 4857 4858 Input Parameters: 4859 + mat - the matrix 4860 . perm - row and column permutation 4861 - info - structure containing 4862 $ levels - number of levels of fill. 4863 $ expected fill - as ratio of original fill. 4864 4865 Output Parameter: 4866 . fact - the factored matrix 4867 4868 Notes: 4869 Most users should employ the KSP interface for linear solvers 4870 instead of working directly with matrix algebra routines such as this. 4871 See, e.g., KSPCreate(). 4872 4873 Level: developer 4874 4875 Concepts: matrices^symbolic incomplete Cholesky factorization 4876 Concepts: matrices^factorization 4877 Concepts: Cholsky^symbolic factorization 4878 4879 .seealso: MatCholeskyFactorNumeric(), MatCholeskyFactor(), MatFactorInfo 4880 @*/ 4881 PetscErrorCode PETSCMAT_DLLEXPORT MatICCFactorSymbolic(Mat mat,IS perm,MatFactorInfo *info,Mat *fact) 4882 { 4883 PetscErrorCode ierr; 4884 4885 PetscFunctionBegin; 4886 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4887 PetscValidType(mat,1); 4888 PetscValidHeaderSpecific(perm,IS_COOKIE,2); 4889 PetscValidPointer(info,3); 4890 PetscValidPointer(fact,4); 4891 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 4892 if (info->levels < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Levels negative %D",(PetscInt) info->levels); 4893 if (info->fill < 1.0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Expected fill less than 1.0 %G",info->fill); 4894 if (!mat->ops->iccfactorsymbolic) SETERRQ1(PETSC_ERR_SUP,"Matrix type %s symbolic ICC",mat->type_name); 4895 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 4896 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4897 4898 ierr = PetscLogEventBegin(MAT_ICCFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 4899 ierr = (*mat->ops->iccfactorsymbolic)(mat,perm,info,fact);CHKERRQ(ierr); 4900 ierr = PetscLogEventEnd(MAT_ICCFactorSymbolic,mat,perm,0,0);CHKERRQ(ierr); 4901 PetscFunctionReturn(0); 4902 } 4903 4904 #undef __FUNCT__ 4905 #define __FUNCT__ "MatGetArray" 4906 /*@C 4907 MatGetArray - Returns a pointer to the element values in the matrix. 4908 The result of this routine is dependent on the underlying matrix data 4909 structure, and may not even work for certain matrix types. You MUST 4910 call MatRestoreArray() when you no longer need to access the array. 4911 4912 Not Collective 4913 4914 Input Parameter: 4915 . mat - the matrix 4916 4917 Output Parameter: 4918 . v - the location of the values 4919 4920 4921 Fortran Note: 4922 This routine is used differently from Fortran, e.g., 4923 .vb 4924 Mat mat 4925 PetscScalar mat_array(1) 4926 PetscOffset i_mat 4927 PetscErrorCode ierr 4928 call MatGetArray(mat,mat_array,i_mat,ierr) 4929 4930 C Access first local entry in matrix; note that array is 4931 C treated as one dimensional 4932 value = mat_array(i_mat + 1) 4933 4934 [... other code ...] 4935 call MatRestoreArray(mat,mat_array,i_mat,ierr) 4936 .ve 4937 4938 See the Fortran chapter of the users manual and 4939 petsc/src/mat/examples/tests for details. 4940 4941 Level: advanced 4942 4943 Concepts: matrices^access array 4944 4945 .seealso: MatRestoreArray(), MatGetArrayF90(), MatGetRowIJ() 4946 @*/ 4947 PetscErrorCode PETSCMAT_DLLEXPORT MatGetArray(Mat mat,PetscScalar *v[]) 4948 { 4949 PetscErrorCode ierr; 4950 4951 PetscFunctionBegin; 4952 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 4953 PetscValidType(mat,1); 4954 PetscValidPointer(v,2); 4955 if (!mat->ops->getarray) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 4956 ierr = MatPreallocated(mat);CHKERRQ(ierr); 4957 ierr = (*mat->ops->getarray)(mat,v);CHKERRQ(ierr); 4958 CHKMEMQ; 4959 PetscFunctionReturn(0); 4960 } 4961 4962 #undef __FUNCT__ 4963 #define __FUNCT__ "MatRestoreArray" 4964 /*@C 4965 MatRestoreArray - Restores the matrix after MatGetArray() has been called. 4966 4967 Not Collective 4968 4969 Input Parameter: 4970 + mat - the matrix 4971 - v - the location of the values 4972 4973 Fortran Note: 4974 This routine is used differently from Fortran, e.g., 4975 .vb 4976 Mat mat 4977 PetscScalar mat_array(1) 4978 PetscOffset i_mat 4979 PetscErrorCode ierr 4980 call MatGetArray(mat,mat_array,i_mat,ierr) 4981 4982 C Access first local entry in matrix; note that array is 4983 C treated as one dimensional 4984 value = mat_array(i_mat + 1) 4985 4986 [... other code ...] 4987 call MatRestoreArray(mat,mat_array,i_mat,ierr) 4988 .ve 4989 4990 See the Fortran chapter of the users manual and 4991 petsc/src/mat/examples/tests for details 4992 4993 Level: advanced 4994 4995 .seealso: MatGetArray(), MatRestoreArrayF90() 4996 @*/ 4997 PetscErrorCode PETSCMAT_DLLEXPORT MatRestoreArray(Mat mat,PetscScalar *v[]) 4998 { 4999 PetscErrorCode ierr; 5000 5001 PetscFunctionBegin; 5002 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5003 PetscValidType(mat,1); 5004 PetscValidPointer(v,2); 5005 #if defined(PETSC_USE_DEBUG) 5006 CHKMEMQ; 5007 #endif 5008 if (!mat->ops->restorearray) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 5009 ierr = (*mat->ops->restorearray)(mat,v);CHKERRQ(ierr); 5010 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 5011 PetscFunctionReturn(0); 5012 } 5013 5014 #undef __FUNCT__ 5015 #define __FUNCT__ "MatGetSubMatrices" 5016 /*@C 5017 MatGetSubMatrices - Extracts several submatrices from a matrix. If submat 5018 points to an array of valid matrices, they may be reused to store the new 5019 submatrices. 5020 5021 Collective on Mat 5022 5023 Input Parameters: 5024 + mat - the matrix 5025 . n - the number of submatrixes to be extracted (on this processor, may be zero) 5026 . irow, icol - index sets of rows and columns to extract 5027 - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 5028 5029 Output Parameter: 5030 . submat - the array of submatrices 5031 5032 Notes: 5033 MatGetSubMatrices() can extract only sequential submatrices 5034 (from both sequential and parallel matrices). Use MatGetSubMatrix() 5035 to extract a parallel submatrix. 5036 5037 When extracting submatrices from a parallel matrix, each processor can 5038 form a different submatrix by setting the rows and columns of its 5039 individual index sets according to the local submatrix desired. 5040 5041 When finished using the submatrices, the user should destroy 5042 them with MatDestroyMatrices(). 5043 5044 MAT_REUSE_MATRIX can only be used when the nonzero structure of the 5045 original matrix has not changed from that last call to MatGetSubMatrices(). 5046 5047 This routine creates the matrices in submat; you should NOT create them before 5048 calling it. It also allocates the array of matrix pointers submat. 5049 5050 For BAIJ matrices the index sets must respect the block structure, that is if they 5051 request one row/column in a block, they must request all rows/columns that are in 5052 that block. For example, if the block size is 2 you cannot request just row 0 and 5053 column 0. 5054 5055 Fortran Note: 5056 The Fortran interface is slightly different from that given below; it 5057 requires one to pass in as submat a Mat (integer) array of size at least m. 5058 5059 Level: advanced 5060 5061 Concepts: matrices^accessing submatrices 5062 Concepts: submatrices 5063 5064 .seealso: MatDestroyMatrices(), MatGetSubMatrix(), MatGetRow(), MatGetDiagonal() 5065 @*/ 5066 PetscErrorCode PETSCMAT_DLLEXPORT MatGetSubMatrices(Mat mat,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *submat[]) 5067 { 5068 PetscErrorCode ierr; 5069 PetscInt i; 5070 PetscTruth eq; 5071 5072 PetscFunctionBegin; 5073 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5074 PetscValidType(mat,1); 5075 if (n) { 5076 PetscValidPointer(irow,3); 5077 PetscValidHeaderSpecific(*irow,IS_COOKIE,3); 5078 PetscValidPointer(icol,4); 5079 PetscValidHeaderSpecific(*icol,IS_COOKIE,4); 5080 } 5081 PetscValidPointer(submat,6); 5082 if (n && scall == MAT_REUSE_MATRIX) { 5083 PetscValidPointer(*submat,6); 5084 PetscValidHeaderSpecific(**submat,MAT_COOKIE,6); 5085 } 5086 if (!mat->ops->getsubmatrices) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 5087 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5088 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5089 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5090 5091 ierr = PetscLogEventBegin(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 5092 ierr = (*mat->ops->getsubmatrices)(mat,n,irow,icol,scall,submat);CHKERRQ(ierr); 5093 ierr = PetscLogEventEnd(MAT_GetSubMatrices,mat,0,0,0);CHKERRQ(ierr); 5094 for (i=0; i<n; i++) { 5095 if (mat->symmetric || mat->structurally_symmetric || mat->hermitian) { 5096 ierr = ISEqual(irow[i],icol[i],&eq);CHKERRQ(ierr); 5097 if (eq) { 5098 if (mat->symmetric){ 5099 ierr = MatSetOption((*submat)[i],MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 5100 } else if (mat->hermitian) { 5101 ierr = MatSetOption((*submat)[i],MAT_HERMITIAN,PETSC_TRUE);CHKERRQ(ierr); 5102 } else if (mat->structurally_symmetric) { 5103 ierr = MatSetOption((*submat)[i],MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 5104 } 5105 } 5106 } 5107 } 5108 PetscFunctionReturn(0); 5109 } 5110 5111 #undef __FUNCT__ 5112 #define __FUNCT__ "MatDestroyMatrices" 5113 /*@C 5114 MatDestroyMatrices - Destroys a set of matrices obtained with MatGetSubMatrices(). 5115 5116 Collective on Mat 5117 5118 Input Parameters: 5119 + n - the number of local matrices 5120 - mat - the matrices (note that this is a pointer to the array of matrices, just to match the calling 5121 sequence of MatGetSubMatrices()) 5122 5123 Level: advanced 5124 5125 Notes: Frees not only the matrices, but also the array that contains the matrices 5126 5127 .seealso: MatGetSubMatrices() 5128 @*/ 5129 PetscErrorCode PETSCMAT_DLLEXPORT MatDestroyMatrices(PetscInt n,Mat *mat[]) 5130 { 5131 PetscErrorCode ierr; 5132 PetscInt i; 5133 5134 PetscFunctionBegin; 5135 if (n < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Trying to destroy negative number of matrices %D",n); 5136 PetscValidPointer(mat,2); 5137 for (i=0; i<n; i++) { 5138 ierr = MatDestroy((*mat)[i]);CHKERRQ(ierr); 5139 } 5140 /* memory is allocated even if n = 0 */ 5141 ierr = PetscFree(*mat);CHKERRQ(ierr); 5142 PetscFunctionReturn(0); 5143 } 5144 5145 #undef __FUNCT__ 5146 #define __FUNCT__ "MatIncreaseOverlap" 5147 /*@ 5148 MatIncreaseOverlap - Given a set of submatrices indicated by index sets, 5149 replaces the index sets by larger ones that represent submatrices with 5150 additional overlap. 5151 5152 Collective on Mat 5153 5154 Input Parameters: 5155 + mat - the matrix 5156 . n - the number of index sets 5157 . is - the array of index sets (these index sets will changed during the call) 5158 - ov - the additional overlap requested 5159 5160 Level: developer 5161 5162 Concepts: overlap 5163 Concepts: ASM^computing overlap 5164 5165 .seealso: MatGetSubMatrices() 5166 @*/ 5167 PetscErrorCode PETSCMAT_DLLEXPORT MatIncreaseOverlap(Mat mat,PetscInt n,IS is[],PetscInt ov) 5168 { 5169 PetscErrorCode ierr; 5170 5171 PetscFunctionBegin; 5172 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5173 PetscValidType(mat,1); 5174 if (n < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Must have one or more domains, you have %D",n); 5175 if (n) { 5176 PetscValidPointer(is,3); 5177 PetscValidHeaderSpecific(*is,IS_COOKIE,3); 5178 } 5179 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 5180 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5181 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5182 5183 if (!ov) PetscFunctionReturn(0); 5184 if (!mat->ops->increaseoverlap) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 5185 ierr = PetscLogEventBegin(MAT_IncreaseOverlap,mat,0,0,0);CHKERRQ(ierr); 5186 ierr = (*mat->ops->increaseoverlap)(mat,n,is,ov);CHKERRQ(ierr); 5187 ierr = PetscLogEventEnd(MAT_IncreaseOverlap,mat,0,0,0);CHKERRQ(ierr); 5188 PetscFunctionReturn(0); 5189 } 5190 5191 #undef __FUNCT__ 5192 #define __FUNCT__ "MatGetBlockSize" 5193 /*@ 5194 MatGetBlockSize - Returns the matrix block size; useful especially for the 5195 block row and block diagonal formats. 5196 5197 Not Collective 5198 5199 Input Parameter: 5200 . mat - the matrix 5201 5202 Output Parameter: 5203 . bs - block size 5204 5205 Notes: 5206 Block diagonal formats are MATSEQBDIAG, MATMPIBDIAG. 5207 Block row formats are MATSEQBAIJ, MATMPIBAIJ, MATSEQSBAIJ, MATMPISBAIJ 5208 5209 Level: intermediate 5210 5211 Concepts: matrices^block size 5212 5213 .seealso: MatCreateSeqBAIJ(), MatCreateMPIBAIJ(), MatCreateSeqBDiag(), MatCreateMPIBDiag() 5214 @*/ 5215 PetscErrorCode PETSCMAT_DLLEXPORT MatGetBlockSize(Mat mat,PetscInt *bs) 5216 { 5217 PetscErrorCode ierr; 5218 5219 PetscFunctionBegin; 5220 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5221 PetscValidType(mat,1); 5222 PetscValidIntPointer(bs,2); 5223 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5224 *bs = mat->rmap.bs; 5225 PetscFunctionReturn(0); 5226 } 5227 5228 #undef __FUNCT__ 5229 #define __FUNCT__ "MatSetBlockSize" 5230 /*@ 5231 MatSetBlockSize - Sets the matrix block size; for many matrix types you 5232 cannot use this and MUST set the blocksize when you preallocate the matrix 5233 5234 Not Collective 5235 5236 Input Parameters: 5237 + mat - the matrix 5238 - bs - block size 5239 5240 Notes: 5241 Only works for shell and AIJ matrices 5242 5243 Level: intermediate 5244 5245 Concepts: matrices^block size 5246 5247 .seealso: MatCreateSeqBAIJ(), MatCreateMPIBAIJ(), MatCreateSeqBDiag(), MatCreateMPIBDiag(), MatGetBlockSize() 5248 @*/ 5249 PetscErrorCode PETSCMAT_DLLEXPORT MatSetBlockSize(Mat mat,PetscInt bs) 5250 { 5251 PetscErrorCode ierr; 5252 5253 PetscFunctionBegin; 5254 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5255 PetscValidType(mat,1); 5256 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5257 if (mat->ops->setblocksize) { 5258 mat->rmap.bs = bs; 5259 ierr = (*mat->ops->setblocksize)(mat,bs);CHKERRQ(ierr); 5260 } else { 5261 SETERRQ1(PETSC_ERR_ARG_INCOMP,"Cannot set the blocksize for matrix type %s",mat->type_name); 5262 } 5263 PetscFunctionReturn(0); 5264 } 5265 5266 #undef __FUNCT__ 5267 #define __FUNCT__ "MatGetRowIJ" 5268 /*@C 5269 MatGetRowIJ - Returns the compressed row storage i and j indices for sequential matrices. 5270 5271 Collective on Mat 5272 5273 Input Parameters: 5274 + mat - the matrix 5275 . shift - 0 or 1 indicating we want the indices starting at 0 or 1 5276 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 5277 symmetrized 5278 - blockcompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 5279 blockcompressed matrix is desired or not [inode, baij have blockcompressed 5280 nonzero structure which is different than the full nonzero structure] 5281 5282 Output Parameters: 5283 + n - number of rows in the (possibly compressed) matrix 5284 . ia - the row pointers [of length n+1] 5285 . ja - the column indices 5286 - done - indicates if the routine actually worked and returned appropriate ia[] and ja[] arrays; callers 5287 are responsible for handling the case when done == PETSC_FALSE and ia and ja are not set 5288 5289 Level: developer 5290 5291 Notes: You CANNOT change any of the ia[] or ja[] values. 5292 5293 Use MatRestoreRowIJ() when you are finished accessing the ia[] and ja[] values 5294 5295 .seealso: MatGetColumnIJ(), MatRestoreRowIJ(), MatGetArray() 5296 @*/ 5297 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRowIJ(Mat mat,PetscInt shift,PetscTruth symmetric,PetscTruth blockcompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscTruth *done) 5298 { 5299 PetscErrorCode ierr; 5300 5301 PetscFunctionBegin; 5302 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5303 PetscValidType(mat,1); 5304 PetscValidIntPointer(n,4); 5305 if (ia) PetscValidIntPointer(ia,5); 5306 if (ja) PetscValidIntPointer(ja,6); 5307 PetscValidIntPointer(done,7); 5308 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5309 if (!mat->ops->getrowij) *done = PETSC_FALSE; 5310 else { 5311 *done = PETSC_TRUE; 5312 ierr = PetscLogEventBegin(MAT_GetRowIJ,mat,0,0,0);CHKERRQ(ierr); 5313 ierr = (*mat->ops->getrowij)(mat,shift,symmetric,blockcompressed,n,ia,ja,done);CHKERRQ(ierr); 5314 ierr = PetscLogEventEnd(MAT_GetRowIJ,mat,0,0,0);CHKERRQ(ierr); 5315 } 5316 PetscFunctionReturn(0); 5317 } 5318 5319 #undef __FUNCT__ 5320 #define __FUNCT__ "MatGetColumnIJ" 5321 /*@C 5322 MatGetColumnIJ - Returns the compressed column storage i and j indices for sequential matrices. 5323 5324 Collective on Mat 5325 5326 Input Parameters: 5327 + mat - the matrix 5328 . shift - 1 or zero indicating we want the indices starting at 0 or 1 5329 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 5330 symmetrized 5331 - blockcompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 5332 blockcompressed matrix is desired or not [inode, baij have blockcompressed 5333 nonzero structure which is different than the full nonzero structure] 5334 5335 Output Parameters: 5336 + n - number of columns in the (possibly compressed) matrix 5337 . ia - the column pointers 5338 . ja - the row indices 5339 - done - PETSC_TRUE or PETSC_FALSE, indicating whether the values have been returned 5340 5341 Level: developer 5342 5343 .seealso: MatGetRowIJ(), MatRestoreColumnIJ() 5344 @*/ 5345 PetscErrorCode PETSCMAT_DLLEXPORT MatGetColumnIJ(Mat mat,PetscInt shift,PetscTruth symmetric,PetscTruth blockcompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscTruth *done) 5346 { 5347 PetscErrorCode ierr; 5348 5349 PetscFunctionBegin; 5350 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5351 PetscValidType(mat,1); 5352 PetscValidIntPointer(n,4); 5353 if (ia) PetscValidIntPointer(ia,5); 5354 if (ja) PetscValidIntPointer(ja,6); 5355 PetscValidIntPointer(done,7); 5356 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5357 if (!mat->ops->getcolumnij) *done = PETSC_FALSE; 5358 else { 5359 *done = PETSC_TRUE; 5360 ierr = (*mat->ops->getcolumnij)(mat,shift,symmetric,blockcompressed,n,ia,ja,done);CHKERRQ(ierr); 5361 } 5362 PetscFunctionReturn(0); 5363 } 5364 5365 #undef __FUNCT__ 5366 #define __FUNCT__ "MatRestoreRowIJ" 5367 /*@C 5368 MatRestoreRowIJ - Call after you are completed with the ia,ja indices obtained with 5369 MatGetRowIJ(). 5370 5371 Collective on Mat 5372 5373 Input Parameters: 5374 + mat - the matrix 5375 . shift - 1 or zero indicating we want the indices starting at 0 or 1 5376 . symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 5377 symmetrized 5378 - blockcompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 5379 blockcompressed matrix is desired or not [inode, baij have blockcompressed 5380 nonzero structure which is different than the full nonzero structure] 5381 5382 Output Parameters: 5383 + n - size of (possibly compressed) matrix 5384 . ia - the row pointers 5385 . ja - the column indices 5386 - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned 5387 5388 Level: developer 5389 5390 .seealso: MatGetRowIJ(), MatRestoreColumnIJ() 5391 @*/ 5392 PetscErrorCode PETSCMAT_DLLEXPORT MatRestoreRowIJ(Mat mat,PetscInt shift,PetscTruth symmetric,PetscTruth blockcompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscTruth *done) 5393 { 5394 PetscErrorCode ierr; 5395 5396 PetscFunctionBegin; 5397 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5398 PetscValidType(mat,1); 5399 if (ia) PetscValidIntPointer(ia,5); 5400 if (ja) PetscValidIntPointer(ja,6); 5401 PetscValidIntPointer(done,7); 5402 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5403 5404 if (!mat->ops->restorerowij) *done = PETSC_FALSE; 5405 else { 5406 *done = PETSC_TRUE; 5407 ierr = (*mat->ops->restorerowij)(mat,shift,symmetric,blockcompressed,n,ia,ja,done);CHKERRQ(ierr); 5408 } 5409 PetscFunctionReturn(0); 5410 } 5411 5412 #undef __FUNCT__ 5413 #define __FUNCT__ "MatRestoreColumnIJ" 5414 /*@C 5415 MatRestoreColumnIJ - Call after you are completed with the ia,ja indices obtained with 5416 MatGetColumnIJ(). 5417 5418 Collective on Mat 5419 5420 Input Parameters: 5421 + mat - the matrix 5422 . shift - 1 or zero indicating we want the indices starting at 0 or 1 5423 - symmetric - PETSC_TRUE or PETSC_FALSE indicating the matrix data structure should be 5424 symmetrized 5425 - blockcompressed - PETSC_TRUE or PETSC_FALSE indicating if the nonzero structure of the 5426 blockcompressed matrix is desired or not [inode, baij have blockcompressed 5427 nonzero structure which is different than the full nonzero structure] 5428 5429 Output Parameters: 5430 + n - size of (possibly compressed) matrix 5431 . ia - the column pointers 5432 . ja - the row indices 5433 - done - PETSC_TRUE or PETSC_FALSE indicated that the values have been returned 5434 5435 Level: developer 5436 5437 .seealso: MatGetColumnIJ(), MatRestoreRowIJ() 5438 @*/ 5439 PetscErrorCode PETSCMAT_DLLEXPORT MatRestoreColumnIJ(Mat mat,PetscInt shift,PetscTruth symmetric,PetscTruth blockcompressed,PetscInt *n,PetscInt *ia[],PetscInt* ja[],PetscTruth *done) 5440 { 5441 PetscErrorCode ierr; 5442 5443 PetscFunctionBegin; 5444 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5445 PetscValidType(mat,1); 5446 if (ia) PetscValidIntPointer(ia,5); 5447 if (ja) PetscValidIntPointer(ja,6); 5448 PetscValidIntPointer(done,7); 5449 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5450 5451 if (!mat->ops->restorecolumnij) *done = PETSC_FALSE; 5452 else { 5453 *done = PETSC_TRUE; 5454 ierr = (*mat->ops->restorecolumnij)(mat,shift,symmetric,blockcompressed,n,ia,ja,done);CHKERRQ(ierr); 5455 } 5456 PetscFunctionReturn(0); 5457 } 5458 5459 #undef __FUNCT__ 5460 #define __FUNCT__ "MatColoringPatch" 5461 /*@C 5462 MatColoringPatch -Used inside matrix coloring routines that 5463 use MatGetRowIJ() and/or MatGetColumnIJ(). 5464 5465 Collective on Mat 5466 5467 Input Parameters: 5468 + mat - the matrix 5469 . ncolors - max color value 5470 . n - number of entries in colorarray 5471 - colorarray - array indicating color for each column 5472 5473 Output Parameters: 5474 . iscoloring - coloring generated using colorarray information 5475 5476 Level: developer 5477 5478 .seealso: MatGetRowIJ(), MatGetColumnIJ() 5479 5480 @*/ 5481 PetscErrorCode PETSCMAT_DLLEXPORT MatColoringPatch(Mat mat,PetscInt ncolors,PetscInt n,ISColoringValue colorarray[],ISColoring *iscoloring) 5482 { 5483 PetscErrorCode ierr; 5484 5485 PetscFunctionBegin; 5486 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5487 PetscValidType(mat,1); 5488 PetscValidIntPointer(colorarray,4); 5489 PetscValidPointer(iscoloring,5); 5490 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5491 5492 if (!mat->ops->coloringpatch){ 5493 ierr = ISColoringCreate(mat->comm,ncolors,n,colorarray,iscoloring);CHKERRQ(ierr); 5494 } else { 5495 ierr = (*mat->ops->coloringpatch)(mat,ncolors,n,colorarray,iscoloring);CHKERRQ(ierr); 5496 } 5497 PetscFunctionReturn(0); 5498 } 5499 5500 5501 #undef __FUNCT__ 5502 #define __FUNCT__ "MatSetUnfactored" 5503 /*@ 5504 MatSetUnfactored - Resets a factored matrix to be treated as unfactored. 5505 5506 Collective on Mat 5507 5508 Input Parameter: 5509 . mat - the factored matrix to be reset 5510 5511 Notes: 5512 This routine should be used only with factored matrices formed by in-place 5513 factorization via ILU(0) (or by in-place LU factorization for the MATSEQDENSE 5514 format). This option can save memory, for example, when solving nonlinear 5515 systems with a matrix-free Newton-Krylov method and a matrix-based, in-place 5516 ILU(0) preconditioner. 5517 5518 Note that one can specify in-place ILU(0) factorization by calling 5519 .vb 5520 PCType(pc,PCILU); 5521 PCFactorSeUseInPlace(pc); 5522 .ve 5523 or by using the options -pc_type ilu -pc_factor_in_place 5524 5525 In-place factorization ILU(0) can also be used as a local 5526 solver for the blocks within the block Jacobi or additive Schwarz 5527 methods (runtime option: -sub_pc_factor_in_place). See the discussion 5528 of these preconditioners in the users manual for details on setting 5529 local solver options. 5530 5531 Most users should employ the simplified KSP interface for linear solvers 5532 instead of working directly with matrix algebra routines such as this. 5533 See, e.g., KSPCreate(). 5534 5535 Level: developer 5536 5537 .seealso: PCFactorSetUseInPlace() 5538 5539 Concepts: matrices^unfactored 5540 5541 @*/ 5542 PetscErrorCode PETSCMAT_DLLEXPORT MatSetUnfactored(Mat mat) 5543 { 5544 PetscErrorCode ierr; 5545 5546 PetscFunctionBegin; 5547 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5548 PetscValidType(mat,1); 5549 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5550 mat->factor = 0; 5551 if (!mat->ops->setunfactored) PetscFunctionReturn(0); 5552 ierr = (*mat->ops->setunfactored)(mat);CHKERRQ(ierr); 5553 PetscFunctionReturn(0); 5554 } 5555 5556 /*MC 5557 MatGetArrayF90 - Accesses a matrix array from Fortran90. 5558 5559 Synopsis: 5560 MatGetArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr) 5561 5562 Not collective 5563 5564 Input Parameter: 5565 . x - matrix 5566 5567 Output Parameters: 5568 + xx_v - the Fortran90 pointer to the array 5569 - ierr - error code 5570 5571 Example of Usage: 5572 .vb 5573 PetscScalar, pointer xx_v(:) 5574 .... 5575 call MatGetArrayF90(x,xx_v,ierr) 5576 a = xx_v(3) 5577 call MatRestoreArrayF90(x,xx_v,ierr) 5578 .ve 5579 5580 Notes: 5581 Not yet supported for all F90 compilers 5582 5583 Level: advanced 5584 5585 .seealso: MatRestoreArrayF90(), MatGetArray(), MatRestoreArray() 5586 5587 Concepts: matrices^accessing array 5588 5589 M*/ 5590 5591 /*MC 5592 MatRestoreArrayF90 - Restores a matrix array that has been 5593 accessed with MatGetArrayF90(). 5594 5595 Synopsis: 5596 MatRestoreArrayF90(Mat x,{Scalar, pointer :: xx_v(:)},integer ierr) 5597 5598 Not collective 5599 5600 Input Parameters: 5601 + x - matrix 5602 - xx_v - the Fortran90 pointer to the array 5603 5604 Output Parameter: 5605 . ierr - error code 5606 5607 Example of Usage: 5608 .vb 5609 PetscScalar, pointer xx_v(:) 5610 .... 5611 call MatGetArrayF90(x,xx_v,ierr) 5612 a = xx_v(3) 5613 call MatRestoreArrayF90(x,xx_v,ierr) 5614 .ve 5615 5616 Notes: 5617 Not yet supported for all F90 compilers 5618 5619 Level: advanced 5620 5621 .seealso: MatGetArrayF90(), MatGetArray(), MatRestoreArray() 5622 5623 M*/ 5624 5625 5626 #undef __FUNCT__ 5627 #define __FUNCT__ "MatGetSubMatrix" 5628 /*@ 5629 MatGetSubMatrix - Gets a single submatrix on the same number of processors 5630 as the original matrix. 5631 5632 Collective on Mat 5633 5634 Input Parameters: 5635 + mat - the original matrix 5636 . isrow - rows this processor should obtain 5637 . iscol - columns for all processors you wish to keep 5638 . csize - number of columns "local" to this processor (does nothing for sequential 5639 matrices). This should match the result from VecGetLocalSize(x,...) if you 5640 plan to use the matrix in a A*x; alternatively, you can use PETSC_DECIDE 5641 - cll - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 5642 5643 Output Parameter: 5644 . newmat - the new submatrix, of the same type as the old 5645 5646 Level: advanced 5647 5648 Notes: the iscol argument MUST be the same on each processor. You might be 5649 able to create the iscol argument with ISAllGather(). The rows is isrow will be 5650 sorted into the same order as the original matrix. 5651 5652 The first time this is called you should use a cll of MAT_INITIAL_MATRIX, 5653 the MatGetSubMatrix() routine will create the newmat for you. Any additional calls 5654 to this routine with a mat of the same nonzero structure and with a call of MAT_REUSE_MATRIX 5655 will reuse the matrix generated the first time. You should call MatDestroy() on newmat when 5656 you are finished using it. 5657 5658 Concepts: matrices^submatrices 5659 5660 .seealso: MatGetSubMatrices(), ISAllGather() 5661 @*/ 5662 PetscErrorCode PETSCMAT_DLLEXPORT MatGetSubMatrix(Mat mat,IS isrow,IS iscol,PetscInt csize,MatReuse cll,Mat *newmat) 5663 { 5664 PetscErrorCode ierr; 5665 PetscMPIInt size; 5666 Mat *local; 5667 5668 PetscFunctionBegin; 5669 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5670 PetscValidHeaderSpecific(isrow,IS_COOKIE,2); 5671 PetscValidHeaderSpecific(iscol,IS_COOKIE,3); 5672 PetscValidPointer(newmat,6); 5673 if (cll == MAT_REUSE_MATRIX) PetscValidHeaderSpecific(*newmat,MAT_COOKIE,6); 5674 PetscValidType(mat,1); 5675 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5676 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5677 ierr = MPI_Comm_size(mat->comm,&size);CHKERRQ(ierr); 5678 5679 /* if original matrix is on just one processor then use submatrix generated */ 5680 if (!mat->ops->getsubmatrix && size == 1 && cll == MAT_REUSE_MATRIX) { 5681 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_REUSE_MATRIX,&newmat);CHKERRQ(ierr); 5682 PetscFunctionReturn(0); 5683 } else if (!mat->ops->getsubmatrix && size == 1) { 5684 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_INITIAL_MATRIX,&local);CHKERRQ(ierr); 5685 *newmat = *local; 5686 ierr = PetscFree(local);CHKERRQ(ierr); 5687 PetscFunctionReturn(0); 5688 } 5689 5690 if (!mat->ops->getsubmatrix) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 5691 ierr = (*mat->ops->getsubmatrix)(mat,isrow,iscol,csize,cll,newmat);CHKERRQ(ierr); 5692 ierr = PetscObjectStateIncrease((PetscObject)*newmat);CHKERRQ(ierr); 5693 PetscFunctionReturn(0); 5694 } 5695 5696 #undef __FUNCT__ 5697 #define __FUNCT__ "MatGetSubMatrixRaw" 5698 /*@ 5699 MatGetSubMatrixRaw - Gets a single submatrix on the same number of processors 5700 as the original matrix. 5701 5702 Collective on Mat 5703 5704 Input Parameters: 5705 + mat - the original matrix 5706 . nrows - the number of rows this processor should obtain 5707 . rows - rows this processor should obtain 5708 . ncols - the number of columns for all processors you wish to keep 5709 . cols - columns for all processors you wish to keep 5710 . csize - number of columns "local" to this processor (does nothing for sequential 5711 matrices). This should match the result from VecGetLocalSize(x,...) if you 5712 plan to use the matrix in a A*x; alternatively, you can use PETSC_DECIDE 5713 - cll - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 5714 5715 Output Parameter: 5716 . newmat - the new submatrix, of the same type as the old 5717 5718 Level: advanced 5719 5720 Notes: the iscol argument MUST be the same on each processor. You might be 5721 able to create the iscol argument with ISAllGather(). 5722 5723 The first time this is called you should use a cll of MAT_INITIAL_MATRIX, 5724 the MatGetSubMatrix() routine will create the newmat for you. Any additional calls 5725 to this routine with a mat of the same nonzero structure and with a cll of MAT_REUSE_MATRIX 5726 will reuse the matrix generated the first time. 5727 5728 Concepts: matrices^submatrices 5729 5730 .seealso: MatGetSubMatrices(), ISAllGather() 5731 @*/ 5732 PetscErrorCode PETSCMAT_DLLEXPORT MatGetSubMatrixRaw(Mat mat,PetscInt nrows,const PetscInt rows[],PetscInt ncols,const PetscInt cols[],PetscInt csize,MatReuse cll,Mat *newmat) 5733 { 5734 IS isrow, iscol; 5735 PetscErrorCode ierr; 5736 5737 PetscFunctionBegin; 5738 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5739 PetscValidIntPointer(rows,2); 5740 PetscValidIntPointer(cols,3); 5741 PetscValidPointer(newmat,6); 5742 if (cll == MAT_REUSE_MATRIX) PetscValidHeaderSpecific(*newmat,MAT_COOKIE,6); 5743 PetscValidType(mat,1); 5744 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 5745 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5746 ierr = ISCreateGeneralWithArray(PETSC_COMM_SELF, nrows, (PetscInt *) rows, &isrow);CHKERRQ(ierr); 5747 ierr = ISCreateGeneralWithArray(PETSC_COMM_SELF, ncols, (PetscInt *) cols, &iscol);CHKERRQ(ierr); 5748 ierr = MatGetSubMatrix(mat, isrow, iscol, csize, cll, newmat);CHKERRQ(ierr); 5749 ierr = ISDestroy(isrow);CHKERRQ(ierr); 5750 ierr = ISDestroy(iscol);CHKERRQ(ierr); 5751 PetscFunctionReturn(0); 5752 } 5753 5754 #undef __FUNCT__ 5755 #define __FUNCT__ "MatStashSetInitialSize" 5756 /*@ 5757 MatStashSetInitialSize - sets the sizes of the matrix stash, that is 5758 used during the assembly process to store values that belong to 5759 other processors. 5760 5761 Not Collective 5762 5763 Input Parameters: 5764 + mat - the matrix 5765 . size - the initial size of the stash. 5766 - bsize - the initial size of the block-stash(if used). 5767 5768 Options Database Keys: 5769 + -matstash_initial_size <size> or <size0,size1,...sizep-1> 5770 - -matstash_block_initial_size <bsize> or <bsize0,bsize1,...bsizep-1> 5771 5772 Level: intermediate 5773 5774 Notes: 5775 The block-stash is used for values set with MatSetValuesBlocked() while 5776 the stash is used for values set with MatSetValues() 5777 5778 Run with the option -info and look for output of the form 5779 MatAssemblyBegin_MPIXXX:Stash has MM entries, uses nn mallocs. 5780 to determine the appropriate value, MM, to use for size and 5781 MatAssemblyBegin_MPIXXX:Block-Stash has BMM entries, uses nn mallocs. 5782 to determine the value, BMM to use for bsize 5783 5784 Concepts: stash^setting matrix size 5785 Concepts: matrices^stash 5786 5787 @*/ 5788 PetscErrorCode PETSCMAT_DLLEXPORT MatStashSetInitialSize(Mat mat,PetscInt size, PetscInt bsize) 5789 { 5790 PetscErrorCode ierr; 5791 5792 PetscFunctionBegin; 5793 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5794 PetscValidType(mat,1); 5795 ierr = MatStashSetInitialSize_Private(&mat->stash,size);CHKERRQ(ierr); 5796 ierr = MatStashSetInitialSize_Private(&mat->bstash,bsize);CHKERRQ(ierr); 5797 PetscFunctionReturn(0); 5798 } 5799 5800 #undef __FUNCT__ 5801 #define __FUNCT__ "MatInterpolateAdd" 5802 /*@ 5803 MatInterpolateAdd - w = y + A*x or A'*x depending on the shape of 5804 the matrix 5805 5806 Collective on Mat 5807 5808 Input Parameters: 5809 + mat - the matrix 5810 . x,y - the vectors 5811 - w - where the result is stored 5812 5813 Level: intermediate 5814 5815 Notes: 5816 w may be the same vector as y. 5817 5818 This allows one to use either the restriction or interpolation (its transpose) 5819 matrix to do the interpolation 5820 5821 Concepts: interpolation 5822 5823 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict() 5824 5825 @*/ 5826 PetscErrorCode PETSCMAT_DLLEXPORT MatInterpolateAdd(Mat A,Vec x,Vec y,Vec w) 5827 { 5828 PetscErrorCode ierr; 5829 PetscInt M,N; 5830 5831 PetscFunctionBegin; 5832 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 5833 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 5834 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 5835 PetscValidHeaderSpecific(w,VEC_COOKIE,4); 5836 PetscValidType(A,1); 5837 ierr = MatPreallocated(A);CHKERRQ(ierr); 5838 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 5839 if (N > M) { 5840 ierr = MatMultTransposeAdd(A,x,y,w);CHKERRQ(ierr); 5841 } else { 5842 ierr = MatMultAdd(A,x,y,w);CHKERRQ(ierr); 5843 } 5844 PetscFunctionReturn(0); 5845 } 5846 5847 #undef __FUNCT__ 5848 #define __FUNCT__ "MatInterpolate" 5849 /*@ 5850 MatInterpolate - y = A*x or A'*x depending on the shape of 5851 the matrix 5852 5853 Collective on Mat 5854 5855 Input Parameters: 5856 + mat - the matrix 5857 - x,y - the vectors 5858 5859 Level: intermediate 5860 5861 Notes: 5862 This allows one to use either the restriction or interpolation (its transpose) 5863 matrix to do the interpolation 5864 5865 Concepts: matrices^interpolation 5866 5867 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatRestrict() 5868 5869 @*/ 5870 PetscErrorCode PETSCMAT_DLLEXPORT MatInterpolate(Mat A,Vec x,Vec y) 5871 { 5872 PetscErrorCode ierr; 5873 PetscInt M,N; 5874 5875 PetscFunctionBegin; 5876 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 5877 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 5878 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 5879 PetscValidType(A,1); 5880 ierr = MatPreallocated(A);CHKERRQ(ierr); 5881 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 5882 if (N > M) { 5883 ierr = MatMultTranspose(A,x,y);CHKERRQ(ierr); 5884 } else { 5885 ierr = MatMult(A,x,y);CHKERRQ(ierr); 5886 } 5887 PetscFunctionReturn(0); 5888 } 5889 5890 #undef __FUNCT__ 5891 #define __FUNCT__ "MatRestrict" 5892 /*@ 5893 MatRestrict - y = A*x or A'*x 5894 5895 Collective on Mat 5896 5897 Input Parameters: 5898 + mat - the matrix 5899 - x,y - the vectors 5900 5901 Level: intermediate 5902 5903 Notes: 5904 This allows one to use either the restriction or interpolation (its transpose) 5905 matrix to do the restriction 5906 5907 Concepts: matrices^restriction 5908 5909 .seealso: MatMultAdd(), MatMultTransposeAdd(), MatInterpolate() 5910 5911 @*/ 5912 PetscErrorCode PETSCMAT_DLLEXPORT MatRestrict(Mat A,Vec x,Vec y) 5913 { 5914 PetscErrorCode ierr; 5915 PetscInt M,N; 5916 5917 PetscFunctionBegin; 5918 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 5919 PetscValidHeaderSpecific(x,VEC_COOKIE,2); 5920 PetscValidHeaderSpecific(y,VEC_COOKIE,3); 5921 PetscValidType(A,1); 5922 ierr = MatPreallocated(A);CHKERRQ(ierr); 5923 5924 ierr = MatGetSize(A,&M,&N);CHKERRQ(ierr); 5925 if (N > M) { 5926 ierr = MatMult(A,x,y);CHKERRQ(ierr); 5927 } else { 5928 ierr = MatMultTranspose(A,x,y);CHKERRQ(ierr); 5929 } 5930 PetscFunctionReturn(0); 5931 } 5932 5933 #undef __FUNCT__ 5934 #define __FUNCT__ "MatNullSpaceAttach" 5935 /*@C 5936 MatNullSpaceAttach - attaches a null space to a matrix. 5937 This null space will be removed from the resulting vector whenever 5938 MatMult() is called 5939 5940 Collective on Mat 5941 5942 Input Parameters: 5943 + mat - the matrix 5944 - nullsp - the null space object 5945 5946 Level: developer 5947 5948 Notes: 5949 Overwrites any previous null space that may have been attached 5950 5951 Concepts: null space^attaching to matrix 5952 5953 .seealso: MatCreate(), MatNullSpaceCreate() 5954 @*/ 5955 PetscErrorCode PETSCMAT_DLLEXPORT MatNullSpaceAttach(Mat mat,MatNullSpace nullsp) 5956 { 5957 PetscErrorCode ierr; 5958 5959 PetscFunctionBegin; 5960 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 5961 PetscValidType(mat,1); 5962 PetscValidHeaderSpecific(nullsp,MAT_NULLSPACE_COOKIE,2); 5963 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5964 ierr = PetscObjectReference((PetscObject)nullsp);CHKERRQ(ierr); 5965 if (mat->nullsp) { ierr = MatNullSpaceDestroy(mat->nullsp);CHKERRQ(ierr); } 5966 mat->nullsp = nullsp; 5967 PetscFunctionReturn(0); 5968 } 5969 5970 #undef __FUNCT__ 5971 #define __FUNCT__ "MatICCFactor" 5972 /*@ 5973 MatICCFactor - Performs in-place incomplete Cholesky factorization of matrix. 5974 5975 Collective on Mat 5976 5977 Input Parameters: 5978 + mat - the matrix 5979 . row - row/column permutation 5980 . fill - expected fill factor >= 1.0 5981 - level - level of fill, for ICC(k) 5982 5983 Notes: 5984 Probably really in-place only when level of fill is zero, otherwise allocates 5985 new space to store factored matrix and deletes previous memory. 5986 5987 Most users should employ the simplified KSP interface for linear solvers 5988 instead of working directly with matrix algebra routines such as this. 5989 See, e.g., KSPCreate(). 5990 5991 Level: developer 5992 5993 Concepts: matrices^incomplete Cholesky factorization 5994 Concepts: Cholesky factorization 5995 5996 .seealso: MatICCFactorSymbolic(), MatLUFactorNumeric(), MatCholeskyFactor() 5997 @*/ 5998 PetscErrorCode PETSCMAT_DLLEXPORT MatICCFactor(Mat mat,IS row,MatFactorInfo* info) 5999 { 6000 PetscErrorCode ierr; 6001 6002 PetscFunctionBegin; 6003 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6004 PetscValidType(mat,1); 6005 if (row) PetscValidHeaderSpecific(row,IS_COOKIE,2); 6006 PetscValidPointer(info,3); 6007 if (mat->rmap.N != mat->cmap.N) SETERRQ(PETSC_ERR_ARG_WRONG,"matrix must be square"); 6008 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6009 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6010 if (!mat->ops->iccfactor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6011 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6012 ierr = (*mat->ops->iccfactor)(mat,row,info);CHKERRQ(ierr); 6013 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 6014 PetscFunctionReturn(0); 6015 } 6016 6017 #undef __FUNCT__ 6018 #define __FUNCT__ "MatSetValuesAdic" 6019 /*@ 6020 MatSetValuesAdic - Sets values computed with ADIC automatic differentiation into a matrix. 6021 6022 Not Collective 6023 6024 Input Parameters: 6025 + mat - the matrix 6026 - v - the values compute with ADIC 6027 6028 Level: developer 6029 6030 Notes: 6031 Must call MatSetColoring() before using this routine. Also this matrix must already 6032 have its nonzero pattern determined. 6033 6034 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 6035 MatSetValues(), MatSetColoring(), MatSetValuesAdifor() 6036 @*/ 6037 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesAdic(Mat mat,void *v) 6038 { 6039 PetscErrorCode ierr; 6040 6041 PetscFunctionBegin; 6042 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6043 PetscValidType(mat,1); 6044 PetscValidPointer(mat,2); 6045 6046 if (!mat->assembled) { 6047 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 6048 } 6049 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 6050 if (!mat->ops->setvaluesadic) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6051 ierr = (*mat->ops->setvaluesadic)(mat,v);CHKERRQ(ierr); 6052 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 6053 ierr = MatView_Private(mat);CHKERRQ(ierr); 6054 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 6055 PetscFunctionReturn(0); 6056 } 6057 6058 6059 #undef __FUNCT__ 6060 #define __FUNCT__ "MatSetColoring" 6061 /*@ 6062 MatSetColoring - Sets a coloring used by calls to MatSetValuesAdic() 6063 6064 Not Collective 6065 6066 Input Parameters: 6067 + mat - the matrix 6068 - coloring - the coloring 6069 6070 Level: developer 6071 6072 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 6073 MatSetValues(), MatSetValuesAdic() 6074 @*/ 6075 PetscErrorCode PETSCMAT_DLLEXPORT MatSetColoring(Mat mat,ISColoring coloring) 6076 { 6077 PetscErrorCode ierr; 6078 6079 PetscFunctionBegin; 6080 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6081 PetscValidType(mat,1); 6082 PetscValidPointer(coloring,2); 6083 6084 if (!mat->assembled) { 6085 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 6086 } 6087 if (!mat->ops->setcoloring) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6088 ierr = (*mat->ops->setcoloring)(mat,coloring);CHKERRQ(ierr); 6089 PetscFunctionReturn(0); 6090 } 6091 6092 #undef __FUNCT__ 6093 #define __FUNCT__ "MatSetValuesAdifor" 6094 /*@ 6095 MatSetValuesAdifor - Sets values computed with automatic differentiation into a matrix. 6096 6097 Not Collective 6098 6099 Input Parameters: 6100 + mat - the matrix 6101 . nl - leading dimension of v 6102 - v - the values compute with ADIFOR 6103 6104 Level: developer 6105 6106 Notes: 6107 Must call MatSetColoring() before using this routine. Also this matrix must already 6108 have its nonzero pattern determined. 6109 6110 .seealso: MatSetOption(), MatAssemblyBegin(), MatAssemblyEnd(), MatSetValuesBlocked(), MatSetValuesLocal(), 6111 MatSetValues(), MatSetColoring() 6112 @*/ 6113 PetscErrorCode PETSCMAT_DLLEXPORT MatSetValuesAdifor(Mat mat,PetscInt nl,void *v) 6114 { 6115 PetscErrorCode ierr; 6116 6117 PetscFunctionBegin; 6118 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6119 PetscValidType(mat,1); 6120 PetscValidPointer(v,3); 6121 6122 if (!mat->assembled) { 6123 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 6124 } 6125 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 6126 if (!mat->ops->setvaluesadifor) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6127 ierr = (*mat->ops->setvaluesadifor)(mat,nl,v);CHKERRQ(ierr); 6128 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 6129 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 6130 PetscFunctionReturn(0); 6131 } 6132 6133 #undef __FUNCT__ 6134 #define __FUNCT__ "MatDiagonalScaleLocal" 6135 /*@ 6136 MatDiagonalScaleLocal - Scales columns of a matrix given the scaling values including the 6137 ghosted ones. 6138 6139 Not Collective 6140 6141 Input Parameters: 6142 + mat - the matrix 6143 - diag = the diagonal values, including ghost ones 6144 6145 Level: developer 6146 6147 Notes: Works only for MPIAIJ and MPIBAIJ matrices 6148 6149 .seealso: MatDiagonalScale() 6150 @*/ 6151 PetscErrorCode PETSCMAT_DLLEXPORT MatDiagonalScaleLocal(Mat mat,Vec diag) 6152 { 6153 PetscErrorCode ierr; 6154 PetscMPIInt size; 6155 6156 PetscFunctionBegin; 6157 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6158 PetscValidHeaderSpecific(diag,VEC_COOKIE,2); 6159 PetscValidType(mat,1); 6160 6161 if (!mat->assembled) { 6162 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Matrix must be already assembled"); 6163 } 6164 ierr = PetscLogEventBegin(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 6165 ierr = MPI_Comm_size(mat->comm,&size);CHKERRQ(ierr); 6166 if (size == 1) { 6167 PetscInt n,m; 6168 ierr = VecGetSize(diag,&n);CHKERRQ(ierr); 6169 ierr = MatGetSize(mat,0,&m);CHKERRQ(ierr); 6170 if (m == n) { 6171 ierr = MatDiagonalScale(mat,0,diag);CHKERRQ(ierr); 6172 } else { 6173 SETERRQ(PETSC_ERR_SUP,"Only supported for sequential matrices when no ghost points/periodic conditions"); 6174 } 6175 } else { 6176 PetscErrorCode (*f)(Mat,Vec); 6177 ierr = PetscObjectQueryFunction((PetscObject)mat,"MatDiagonalScaleLocal_C",(void (**)(void))&f);CHKERRQ(ierr); 6178 if (f) { 6179 ierr = (*f)(mat,diag);CHKERRQ(ierr); 6180 } else { 6181 SETERRQ(PETSC_ERR_SUP,"Only supported for MPIAIJ and MPIBAIJ parallel matrices"); 6182 } 6183 } 6184 ierr = PetscLogEventEnd(MAT_Scale,mat,0,0,0);CHKERRQ(ierr); 6185 ierr = PetscObjectStateIncrease((PetscObject)mat);CHKERRQ(ierr); 6186 PetscFunctionReturn(0); 6187 } 6188 6189 #undef __FUNCT__ 6190 #define __FUNCT__ "MatGetInertia" 6191 /*@ 6192 MatGetInertia - Gets the inertia from a factored matrix 6193 6194 Collective on Mat 6195 6196 Input Parameter: 6197 . mat - the matrix 6198 6199 Output Parameters: 6200 + nneg - number of negative eigenvalues 6201 . nzero - number of zero eigenvalues 6202 - npos - number of positive eigenvalues 6203 6204 Level: advanced 6205 6206 Notes: Matrix must have been factored by MatCholeskyFactor() 6207 6208 6209 @*/ 6210 PetscErrorCode PETSCMAT_DLLEXPORT MatGetInertia(Mat mat,PetscInt *nneg,PetscInt *nzero,PetscInt *npos) 6211 { 6212 PetscErrorCode ierr; 6213 6214 PetscFunctionBegin; 6215 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6216 PetscValidType(mat,1); 6217 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 6218 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Numeric factor mat is not assembled"); 6219 if (!mat->ops->getinertia) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6220 ierr = (*mat->ops->getinertia)(mat,nneg,nzero,npos);CHKERRQ(ierr); 6221 PetscFunctionReturn(0); 6222 } 6223 6224 /* ----------------------------------------------------------------*/ 6225 #undef __FUNCT__ 6226 #define __FUNCT__ "MatSolves" 6227 /*@ 6228 MatSolves - Solves A x = b, given a factored matrix, for a collection of vectors 6229 6230 Collective on Mat and Vecs 6231 6232 Input Parameters: 6233 + mat - the factored matrix 6234 - b - the right-hand-side vectors 6235 6236 Output Parameter: 6237 . x - the result vectors 6238 6239 Notes: 6240 The vectors b and x cannot be the same. I.e., one cannot 6241 call MatSolves(A,x,x). 6242 6243 Notes: 6244 Most users should employ the simplified KSP interface for linear solvers 6245 instead of working directly with matrix algebra routines such as this. 6246 See, e.g., KSPCreate(). 6247 6248 Level: developer 6249 6250 Concepts: matrices^triangular solves 6251 6252 .seealso: MatSolveAdd(), MatSolveTranspose(), MatSolveTransposeAdd(), MatSolve() 6253 @*/ 6254 PetscErrorCode PETSCMAT_DLLEXPORT MatSolves(Mat mat,Vecs b,Vecs x) 6255 { 6256 PetscErrorCode ierr; 6257 6258 PetscFunctionBegin; 6259 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6260 PetscValidType(mat,1); 6261 if (x == b) SETERRQ(PETSC_ERR_ARG_IDN,"x and b must be different vectors"); 6262 if (!mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Unfactored matrix"); 6263 if (!mat->rmap.N && !mat->cmap.N) PetscFunctionReturn(0); 6264 6265 if (!mat->ops->solves) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 6266 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6267 ierr = PetscLogEventBegin(MAT_Solves,mat,0,0,0);CHKERRQ(ierr); 6268 ierr = (*mat->ops->solves)(mat,b,x);CHKERRQ(ierr); 6269 ierr = PetscLogEventEnd(MAT_Solves,mat,0,0,0);CHKERRQ(ierr); 6270 PetscFunctionReturn(0); 6271 } 6272 6273 #undef __FUNCT__ 6274 #define __FUNCT__ "MatIsSymmetric" 6275 /*@ 6276 MatIsSymmetric - Test whether a matrix is symmetric 6277 6278 Collective on Mat 6279 6280 Input Parameter: 6281 + A - the matrix to test 6282 - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact transpose) 6283 6284 Output Parameters: 6285 . flg - the result 6286 6287 Level: intermediate 6288 6289 Concepts: matrix^symmetry 6290 6291 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown() 6292 @*/ 6293 PetscErrorCode PETSCMAT_DLLEXPORT MatIsSymmetric(Mat A,PetscReal tol,PetscTruth *flg) 6294 { 6295 PetscErrorCode ierr; 6296 6297 PetscFunctionBegin; 6298 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6299 PetscValidPointer(flg,2); 6300 if (!A->symmetric_set) { 6301 if (!A->ops->issymmetric) { 6302 MatType mattype; 6303 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 6304 SETERRQ1(PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for symmetric",mattype); 6305 } 6306 ierr = (*A->ops->issymmetric)(A,tol,&A->symmetric);CHKERRQ(ierr); 6307 A->symmetric_set = PETSC_TRUE; 6308 if (A->symmetric) { 6309 A->structurally_symmetric_set = PETSC_TRUE; 6310 A->structurally_symmetric = PETSC_TRUE; 6311 } 6312 } 6313 *flg = A->symmetric; 6314 PetscFunctionReturn(0); 6315 } 6316 6317 #undef __FUNCT__ 6318 #define __FUNCT__ "MatIsHermitian" 6319 /*@ 6320 MatIsHermitian - Test whether a matrix is Hermitian 6321 6322 Collective on Mat 6323 6324 Input Parameter: 6325 + A - the matrix to test 6326 - tol - difference between value and its transpose less than this amount counts as equal (use 0.0 for exact Hermitian) 6327 6328 Output Parameters: 6329 . flg - the result 6330 6331 Level: intermediate 6332 6333 Concepts: matrix^symmetry 6334 6335 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetricKnown() 6336 @*/ 6337 PetscErrorCode PETSCMAT_DLLEXPORT MatIsHermitian(Mat A,PetscReal tol,PetscTruth *flg) 6338 { 6339 PetscErrorCode ierr; 6340 6341 PetscFunctionBegin; 6342 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6343 PetscValidPointer(flg,2); 6344 if (!A->hermitian_set) { 6345 if (!A->ops->ishermitian) { 6346 MatType mattype; 6347 ierr = MatGetType(A,&mattype);CHKERRQ(ierr); 6348 SETERRQ1(PETSC_ERR_SUP,"Matrix of type <%s> does not support checking for Hermitian",mattype); 6349 } 6350 ierr = (*A->ops->ishermitian)(A,tol,&A->hermitian);CHKERRQ(ierr); 6351 A->hermitian_set = PETSC_TRUE; 6352 if (A->hermitian) { 6353 A->structurally_symmetric_set = PETSC_TRUE; 6354 A->structurally_symmetric = PETSC_TRUE; 6355 } 6356 } 6357 *flg = A->hermitian; 6358 PetscFunctionReturn(0); 6359 } 6360 6361 #undef __FUNCT__ 6362 #define __FUNCT__ "MatIsSymmetricKnown" 6363 /*@ 6364 MatIsSymmetricKnown - Checks the flag on the matrix to see if it is symmetric. 6365 6366 Collective on Mat 6367 6368 Input Parameter: 6369 . A - the matrix to check 6370 6371 Output Parameters: 6372 + set - if the symmetric flag is set (this tells you if the next flag is valid) 6373 - flg - the result 6374 6375 Level: advanced 6376 6377 Concepts: matrix^symmetry 6378 6379 Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsSymmetric() 6380 if you want it explicitly checked 6381 6382 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric() 6383 @*/ 6384 PetscErrorCode PETSCMAT_DLLEXPORT MatIsSymmetricKnown(Mat A,PetscTruth *set,PetscTruth *flg) 6385 { 6386 PetscFunctionBegin; 6387 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6388 PetscValidPointer(set,2); 6389 PetscValidPointer(flg,3); 6390 if (A->symmetric_set) { 6391 *set = PETSC_TRUE; 6392 *flg = A->symmetric; 6393 } else { 6394 *set = PETSC_FALSE; 6395 } 6396 PetscFunctionReturn(0); 6397 } 6398 6399 #undef __FUNCT__ 6400 #define __FUNCT__ "MatIsHermitianKnown" 6401 /*@ 6402 MatIsHermitianKnown - Checks the flag on the matrix to see if it is hermitian. 6403 6404 Collective on Mat 6405 6406 Input Parameter: 6407 . A - the matrix to check 6408 6409 Output Parameters: 6410 + set - if the hermitian flag is set (this tells you if the next flag is valid) 6411 - flg - the result 6412 6413 Level: advanced 6414 6415 Concepts: matrix^symmetry 6416 6417 Note: Does not check the matrix values directly, so this may return unknown (set = PETSC_FALSE). Use MatIsHermitian() 6418 if you want it explicitly checked 6419 6420 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsStructurallySymmetric(), MatSetOption(), MatIsSymmetric() 6421 @*/ 6422 PetscErrorCode PETSCMAT_DLLEXPORT MatIsHermitianKnown(Mat A,PetscTruth *set,PetscTruth *flg) 6423 { 6424 PetscFunctionBegin; 6425 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6426 PetscValidPointer(set,2); 6427 PetscValidPointer(flg,3); 6428 if (A->hermitian_set) { 6429 *set = PETSC_TRUE; 6430 *flg = A->hermitian; 6431 } else { 6432 *set = PETSC_FALSE; 6433 } 6434 PetscFunctionReturn(0); 6435 } 6436 6437 #undef __FUNCT__ 6438 #define __FUNCT__ "MatIsStructurallySymmetric" 6439 /*@ 6440 MatIsStructurallySymmetric - Test whether a matrix is structurally symmetric 6441 6442 Collective on Mat 6443 6444 Input Parameter: 6445 . A - the matrix to test 6446 6447 Output Parameters: 6448 . flg - the result 6449 6450 Level: intermediate 6451 6452 Concepts: matrix^symmetry 6453 6454 .seealso: MatTranspose(), MatIsTranspose(), MatIsHermitian(), MatIsSymmetric(), MatSetOption() 6455 @*/ 6456 PetscErrorCode PETSCMAT_DLLEXPORT MatIsStructurallySymmetric(Mat A,PetscTruth *flg) 6457 { 6458 PetscErrorCode ierr; 6459 6460 PetscFunctionBegin; 6461 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6462 PetscValidPointer(flg,2); 6463 if (!A->structurally_symmetric_set) { 6464 if (!A->ops->isstructurallysymmetric) SETERRQ(PETSC_ERR_SUP,"Matrix does not support checking for structural symmetric"); 6465 ierr = (*A->ops->isstructurallysymmetric)(A,&A->structurally_symmetric);CHKERRQ(ierr); 6466 A->structurally_symmetric_set = PETSC_TRUE; 6467 } 6468 *flg = A->structurally_symmetric; 6469 PetscFunctionReturn(0); 6470 } 6471 6472 #undef __FUNCT__ 6473 #define __FUNCT__ "MatStashGetInfo" 6474 extern PetscErrorCode MatStashGetInfo_Private(MatStash*,PetscInt*,PetscInt*); 6475 /*@ 6476 MatStashGetInfo - Gets how many values are currently in the vector stash, i.e. need 6477 to be communicated to other processors during the MatAssemblyBegin/End() process 6478 6479 Not collective 6480 6481 Input Parameter: 6482 . vec - the vector 6483 6484 Output Parameters: 6485 + nstash - the size of the stash 6486 . reallocs - the number of additional mallocs incurred. 6487 . bnstash - the size of the block stash 6488 - breallocs - the number of additional mallocs incurred.in the block stash 6489 6490 Level: advanced 6491 6492 .seealso: MatAssemblyBegin(), MatAssemblyEnd(), Mat, MatStashSetInitialSize() 6493 6494 @*/ 6495 PetscErrorCode PETSCMAT_DLLEXPORT MatStashGetInfo(Mat mat,PetscInt *nstash,PetscInt *reallocs,PetscInt *bnstash,PetscInt *breallocs) 6496 { 6497 PetscErrorCode ierr; 6498 PetscFunctionBegin; 6499 ierr = MatStashGetInfo_Private(&mat->stash,nstash,reallocs);CHKERRQ(ierr); 6500 ierr = MatStashGetInfo_Private(&mat->bstash,bnstash,breallocs);CHKERRQ(ierr); 6501 PetscFunctionReturn(0); 6502 } 6503 6504 #undef __FUNCT__ 6505 #define __FUNCT__ "MatGetVecs" 6506 /*@ 6507 MatGetVecs - Get vector(s) compatible with the matrix, i.e. with the same 6508 parallel layout 6509 6510 Collective on Mat 6511 6512 Input Parameter: 6513 . mat - the matrix 6514 6515 Output Parameter: 6516 + right - (optional) vector that the matrix can be multiplied against 6517 - left - (optional) vector that the matrix vector product can be stored in 6518 6519 Level: advanced 6520 6521 .seealso: MatCreate() 6522 @*/ 6523 PetscErrorCode PETSCMAT_DLLEXPORT MatGetVecs(Mat mat,Vec *right,Vec *left) 6524 { 6525 PetscErrorCode ierr; 6526 6527 PetscFunctionBegin; 6528 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 6529 PetscValidType(mat,1); 6530 ierr = MatPreallocated(mat);CHKERRQ(ierr); 6531 if (mat->ops->getvecs) { 6532 ierr = (*mat->ops->getvecs)(mat,right,left);CHKERRQ(ierr); 6533 } else { 6534 PetscMPIInt size; 6535 ierr = MPI_Comm_size(mat->comm, &size);CHKERRQ(ierr); 6536 if (right) { 6537 ierr = VecCreate(mat->comm,right);CHKERRQ(ierr); 6538 ierr = VecSetSizes(*right,mat->cmap.n,PETSC_DETERMINE);CHKERRQ(ierr); 6539 if (size > 1) {ierr = VecSetType(*right,VECMPI);CHKERRQ(ierr);} 6540 else {ierr = VecSetType(*right,VECSEQ);CHKERRQ(ierr);} 6541 } 6542 if (left) { 6543 ierr = VecCreate(mat->comm,left);CHKERRQ(ierr); 6544 ierr = VecSetSizes(*left,mat->rmap.n,PETSC_DETERMINE);CHKERRQ(ierr); 6545 if (size > 1) {ierr = VecSetType(*left,VECMPI);CHKERRQ(ierr);} 6546 else {ierr = VecSetType(*left,VECSEQ);CHKERRQ(ierr);} 6547 } 6548 } 6549 if (right) {ierr = VecSetBlockSize(*right,mat->rmap.bs);CHKERRQ(ierr);} 6550 if (left) {ierr = VecSetBlockSize(*left,mat->rmap.bs);CHKERRQ(ierr);} 6551 if (mat->mapping) { 6552 if (right) {ierr = VecSetLocalToGlobalMapping(*right,mat->mapping);CHKERRQ(ierr);} 6553 if (left) {ierr = VecSetLocalToGlobalMapping(*left,mat->mapping);CHKERRQ(ierr);} 6554 } 6555 if (mat->bmapping) { 6556 if (right) {ierr = VecSetLocalToGlobalMappingBlock(*right,mat->bmapping);CHKERRQ(ierr);} 6557 if (left) {ierr = VecSetLocalToGlobalMappingBlock(*left,mat->bmapping);CHKERRQ(ierr);} 6558 } 6559 PetscFunctionReturn(0); 6560 } 6561 6562 #undef __FUNCT__ 6563 #define __FUNCT__ "MatFactorInfoInitialize" 6564 /*@ 6565 MatFactorInfoInitialize - Initializes a MatFactorInfo data structure 6566 with default values. 6567 6568 Not Collective 6569 6570 Input Parameters: 6571 . info - the MatFactorInfo data structure 6572 6573 6574 Notes: The solvers are generally used through the KSP and PC objects, for example 6575 PCLU, PCILU, PCCHOLESKY, PCICC 6576 6577 Level: developer 6578 6579 .seealso: MatFactorInfo 6580 @*/ 6581 6582 PetscErrorCode PETSCMAT_DLLEXPORT MatFactorInfoInitialize(MatFactorInfo *info) 6583 { 6584 PetscErrorCode ierr; 6585 6586 PetscFunctionBegin; 6587 ierr = PetscMemzero(info,sizeof(MatFactorInfo));CHKERRQ(ierr); 6588 PetscFunctionReturn(0); 6589 } 6590 6591 #undef __FUNCT__ 6592 #define __FUNCT__ "MatPtAP" 6593 /*@ 6594 MatPtAP - Creates the matrix projection C = P^T * A * P 6595 6596 Collective on Mat 6597 6598 Input Parameters: 6599 + A - the matrix 6600 . P - the projection matrix 6601 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 6602 - fill - expected fill as ratio of nnz(C)/nnz(A) 6603 6604 Output Parameters: 6605 . C - the product matrix 6606 6607 Notes: 6608 C will be created and must be destroyed by the user with MatDestroy(). 6609 6610 This routine is currently only implemented for pairs of AIJ matrices and classes 6611 which inherit from AIJ. 6612 6613 Level: intermediate 6614 6615 .seealso: MatPtAPSymbolic(), MatPtAPNumeric(), MatMatMult() 6616 @*/ 6617 PetscErrorCode PETSCMAT_DLLEXPORT MatPtAP(Mat A,Mat P,MatReuse scall,PetscReal fill,Mat *C) 6618 { 6619 PetscErrorCode ierr; 6620 6621 PetscFunctionBegin; 6622 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6623 PetscValidType(A,1); 6624 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6625 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6626 PetscValidHeaderSpecific(P,MAT_COOKIE,2); 6627 PetscValidType(P,2); 6628 MatPreallocated(P); 6629 if (!P->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6630 if (P->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6631 PetscValidPointer(C,3); 6632 if (P->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap.N,A->cmap.N); 6633 if (fill < 1.0) SETERRQ1(PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 6634 ierr = MatPreallocated(A);CHKERRQ(ierr); 6635 6636 ierr = PetscLogEventBegin(MAT_PtAP,A,P,0,0);CHKERRQ(ierr); 6637 ierr = (*A->ops->ptap)(A,P,scall,fill,C);CHKERRQ(ierr); 6638 ierr = PetscLogEventEnd(MAT_PtAP,A,P,0,0);CHKERRQ(ierr); 6639 6640 PetscFunctionReturn(0); 6641 } 6642 6643 #undef __FUNCT__ 6644 #define __FUNCT__ "MatPtAPNumeric" 6645 /*@ 6646 MatPtAPNumeric - Computes the matrix projection C = P^T * A * P 6647 6648 Collective on Mat 6649 6650 Input Parameters: 6651 + A - the matrix 6652 - P - the projection matrix 6653 6654 Output Parameters: 6655 . C - the product matrix 6656 6657 Notes: 6658 C must have been created by calling MatPtAPSymbolic and must be destroyed by 6659 the user using MatDeatroy(). 6660 6661 This routine is currently only implemented for pairs of AIJ matrices and classes 6662 which inherit from AIJ. C will be of type MATAIJ. 6663 6664 Level: intermediate 6665 6666 .seealso: MatPtAP(), MatPtAPSymbolic(), MatMatMultNumeric() 6667 @*/ 6668 PetscErrorCode PETSCMAT_DLLEXPORT MatPtAPNumeric(Mat A,Mat P,Mat C) 6669 { 6670 PetscErrorCode ierr; 6671 6672 PetscFunctionBegin; 6673 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6674 PetscValidType(A,1); 6675 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6676 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6677 PetscValidHeaderSpecific(P,MAT_COOKIE,2); 6678 PetscValidType(P,2); 6679 MatPreallocated(P); 6680 if (!P->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6681 if (P->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6682 PetscValidHeaderSpecific(C,MAT_COOKIE,3); 6683 PetscValidType(C,3); 6684 MatPreallocated(C); 6685 if (C->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6686 if (P->cmap.N!=C->rmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap.N,C->rmap.N); 6687 if (P->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap.N,A->cmap.N); 6688 if (A->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap.N,A->cmap.N); 6689 if (P->cmap.N!=C->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->cmap.N,C->cmap.N); 6690 ierr = MatPreallocated(A);CHKERRQ(ierr); 6691 6692 ierr = PetscLogEventBegin(MAT_PtAPNumeric,A,P,0,0);CHKERRQ(ierr); 6693 ierr = (*A->ops->ptapnumeric)(A,P,C);CHKERRQ(ierr); 6694 ierr = PetscLogEventEnd(MAT_PtAPNumeric,A,P,0,0);CHKERRQ(ierr); 6695 PetscFunctionReturn(0); 6696 } 6697 6698 #undef __FUNCT__ 6699 #define __FUNCT__ "MatPtAPSymbolic" 6700 /*@ 6701 MatPtAPSymbolic - Creates the (i,j) structure of the matrix projection C = P^T * A * P 6702 6703 Collective on Mat 6704 6705 Input Parameters: 6706 + A - the matrix 6707 - P - the projection matrix 6708 6709 Output Parameters: 6710 . C - the (i,j) structure of the product matrix 6711 6712 Notes: 6713 C will be created and must be destroyed by the user with MatDestroy(). 6714 6715 This routine is currently only implemented for pairs of SeqAIJ matrices and classes 6716 which inherit from SeqAIJ. C will be of type MATSEQAIJ. The product is computed using 6717 this (i,j) structure by calling MatPtAPNumeric(). 6718 6719 Level: intermediate 6720 6721 .seealso: MatPtAP(), MatPtAPNumeric(), MatMatMultSymbolic() 6722 @*/ 6723 PetscErrorCode PETSCMAT_DLLEXPORT MatPtAPSymbolic(Mat A,Mat P,PetscReal fill,Mat *C) 6724 { 6725 PetscErrorCode ierr; 6726 6727 PetscFunctionBegin; 6728 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6729 PetscValidType(A,1); 6730 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6731 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6732 if (fill <1.0) SETERRQ1(PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 6733 PetscValidHeaderSpecific(P,MAT_COOKIE,2); 6734 PetscValidType(P,2); 6735 MatPreallocated(P); 6736 if (!P->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6737 if (P->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6738 PetscValidPointer(C,3); 6739 6740 if (P->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",P->rmap.N,A->cmap.N); 6741 if (A->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix 'A' must be square, %D != %D",A->rmap.N,A->cmap.N); 6742 ierr = MatPreallocated(A);CHKERRQ(ierr); 6743 ierr = PetscLogEventBegin(MAT_PtAPSymbolic,A,P,0,0);CHKERRQ(ierr); 6744 ierr = (*A->ops->ptapsymbolic)(A,P,fill,C);CHKERRQ(ierr); 6745 ierr = PetscLogEventEnd(MAT_PtAPSymbolic,A,P,0,0);CHKERRQ(ierr); 6746 6747 ierr = MatSetBlockSize(*C,A->rmap.bs);CHKERRQ(ierr); 6748 6749 PetscFunctionReturn(0); 6750 } 6751 6752 #undef __FUNCT__ 6753 #define __FUNCT__ "MatMatMult" 6754 /*@ 6755 MatMatMult - Performs Matrix-Matrix Multiplication C=A*B. 6756 6757 Collective on Mat 6758 6759 Input Parameters: 6760 + A - the left matrix 6761 . B - the right matrix 6762 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 6763 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)) 6764 6765 Output Parameters: 6766 . C - the product matrix 6767 6768 Notes: 6769 C will be created and must be destroyed by the user with MatDestroy(). 6770 Unless scall is MAT_REUSE_MATRIX 6771 6772 If you have many matrices with the same non-zero structure to multiply, you 6773 should either 6774 $ 1) use MAT_REUSE_MATRIX in all calls but the first or 6775 $ 2) call MatMatMultSymbolic() once and then MatMatMultNumeric() for each product needed 6776 6777 Level: intermediate 6778 6779 .seealso: MatMatMultSymbolic(), MatMatMultNumeric(), MatPtAP() 6780 @*/ 6781 PetscErrorCode PETSCMAT_DLLEXPORT MatMatMult(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 6782 { 6783 PetscErrorCode ierr; 6784 PetscErrorCode (*fA)(Mat,Mat,MatReuse,PetscReal,Mat*); 6785 PetscErrorCode (*fB)(Mat,Mat,MatReuse,PetscReal,Mat*); 6786 PetscErrorCode (*mult)(Mat,Mat,MatReuse,PetscReal,Mat *)=PETSC_NULL; 6787 6788 PetscFunctionBegin; 6789 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6790 PetscValidType(A,1); 6791 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6792 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6793 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 6794 PetscValidType(B,2); 6795 MatPreallocated(B); 6796 if (!B->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6797 if (B->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6798 PetscValidPointer(C,3); 6799 if (B->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap.N,A->cmap.N); 6800 if (fill == PETSC_DEFAULT) fill = 2.0; 6801 if (fill < 1.0) SETERRQ1(PETSC_ERR_ARG_SIZ,"Expected fill=%G must be >= 1.0",fill); 6802 ierr = MatPreallocated(A);CHKERRQ(ierr); 6803 6804 fA = A->ops->matmult; 6805 fB = B->ops->matmult; 6806 if (fB == fA) { 6807 if (!fB) SETERRQ1(PETSC_ERR_SUP,"MatMatMult not supported for B of type %s",B->type_name); 6808 mult = fB; 6809 } else { 6810 /* dispatch based on the type of A and B */ 6811 char multname[256]; 6812 ierr = PetscStrcpy(multname,"MatMatMult_");CHKERRQ(ierr); 6813 ierr = PetscStrcat(multname,A->type_name);CHKERRQ(ierr); 6814 ierr = PetscStrcat(multname,"_");CHKERRQ(ierr); 6815 ierr = PetscStrcat(multname,B->type_name);CHKERRQ(ierr); 6816 ierr = PetscStrcat(multname,"_C");CHKERRQ(ierr); /* e.g., multname = "MatMatMult_aij_dense_C" */ 6817 ierr = PetscObjectQueryFunction((PetscObject)B,multname,(void (**)(void))&mult);CHKERRQ(ierr); 6818 if (!mult) SETERRQ2(PETSC_ERR_ARG_INCOMP,"MatMatMult requires A, %s, to be compatible with B, %s",A->type_name,B->type_name); 6819 } 6820 ierr = PetscLogEventBegin(MAT_MatMult,A,B,0,0);CHKERRQ(ierr); 6821 ierr = (*mult)(A,B,scall,fill,C);CHKERRQ(ierr); 6822 ierr = PetscLogEventEnd(MAT_MatMult,A,B,0,0);CHKERRQ(ierr); 6823 PetscFunctionReturn(0); 6824 } 6825 6826 #undef __FUNCT__ 6827 #define __FUNCT__ "MatMatMultSymbolic" 6828 /*@ 6829 MatMatMultSymbolic - Performs construction, preallocation, and computes the ij structure 6830 of the matrix-matrix product C=A*B. Call this routine before calling MatMatMultNumeric(). 6831 6832 Collective on Mat 6833 6834 Input Parameters: 6835 + A - the left matrix 6836 . B - the right matrix 6837 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)) 6838 6839 Output Parameters: 6840 . C - the matrix containing the ij structure of product matrix 6841 6842 Notes: 6843 C will be created and must be destroyed by the user with MatDestroy(). 6844 6845 This routine is currently implemented for 6846 - pairs of AIJ matrices and classes which inherit from AIJ, C will be of type MATAIJ. 6847 - pairs of AIJ (A) and Dense (B) matrix, C will be of type MATDENSE. 6848 6849 Level: intermediate 6850 6851 .seealso: MatMatMult(), MatMatMultNumeric() 6852 @*/ 6853 PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultSymbolic(Mat A,Mat B,PetscReal fill,Mat *C) 6854 { 6855 PetscErrorCode ierr; 6856 PetscErrorCode (*Asymbolic)(Mat,Mat,PetscReal,Mat *); 6857 PetscErrorCode (*Bsymbolic)(Mat,Mat,PetscReal,Mat *); 6858 PetscErrorCode (*symbolic)(Mat,Mat,PetscReal,Mat *)=PETSC_NULL; 6859 6860 PetscFunctionBegin; 6861 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6862 PetscValidType(A,1); 6863 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6864 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6865 6866 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 6867 PetscValidType(B,2); 6868 MatPreallocated(B); 6869 if (!B->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6870 if (B->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6871 PetscValidPointer(C,3); 6872 6873 if (B->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap.N,A->cmap.N); 6874 if (fill == PETSC_DEFAULT) fill = 2.0; 6875 if (fill < 1.0) SETERRQ1(PETSC_ERR_ARG_SIZ,"Expected fill=%G must be > 1.0",fill); 6876 ierr = MatPreallocated(A);CHKERRQ(ierr); 6877 6878 Asymbolic = A->ops->matmultsymbolic; 6879 Bsymbolic = B->ops->matmultsymbolic; 6880 if (Asymbolic == Bsymbolic){ 6881 if (!Bsymbolic) SETERRQ1(PETSC_ERR_SUP,"C=A*B not implemented for B of type %s",B->type_name); 6882 symbolic = Bsymbolic; 6883 } else { /* dispatch based on the type of A and B */ 6884 char symbolicname[256]; 6885 ierr = PetscStrcpy(symbolicname,"MatMatMultSymbolic_");CHKERRQ(ierr); 6886 ierr = PetscStrcat(symbolicname,A->type_name);CHKERRQ(ierr); 6887 ierr = PetscStrcat(symbolicname,"_");CHKERRQ(ierr); 6888 ierr = PetscStrcat(symbolicname,B->type_name);CHKERRQ(ierr); 6889 ierr = PetscStrcat(symbolicname,"_C");CHKERRQ(ierr); 6890 ierr = PetscObjectQueryFunction((PetscObject)B,symbolicname,(void (**)(void))&symbolic);CHKERRQ(ierr); 6891 if (!symbolic) SETERRQ2(PETSC_ERR_ARG_INCOMP,"MatMatMultSymbolic requires A, %s, to be compatible with B, %s",A->type_name,B->type_name); 6892 } 6893 ierr = PetscLogEventBegin(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 6894 ierr = (*symbolic)(A,B,fill,C);CHKERRQ(ierr); 6895 ierr = PetscLogEventEnd(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 6896 PetscFunctionReturn(0); 6897 } 6898 6899 #undef __FUNCT__ 6900 #define __FUNCT__ "MatMatMultNumeric" 6901 /*@ 6902 MatMatMultNumeric - Performs the numeric matrix-matrix product. 6903 Call this routine after first calling MatMatMultSymbolic(). 6904 6905 Collective on Mat 6906 6907 Input Parameters: 6908 + A - the left matrix 6909 - B - the right matrix 6910 6911 Output Parameters: 6912 . C - the product matrix, whose ij structure was defined from MatMatMultSymbolic(). 6913 6914 Notes: 6915 C must have been created with MatMatMultSymbolic. 6916 6917 This routine is currently implemented for 6918 - pairs of AIJ matrices and classes which inherit from AIJ, C will be of type MATAIJ. 6919 - pairs of AIJ (A) and Dense (B) matrix, C will be of type MATDENSE. 6920 6921 Level: intermediate 6922 6923 .seealso: MatMatMult(), MatMatMultSymbolic() 6924 @*/ 6925 PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultNumeric(Mat A,Mat B,Mat C) 6926 { 6927 PetscErrorCode ierr; 6928 PetscErrorCode (*Anumeric)(Mat,Mat,Mat); 6929 PetscErrorCode (*Bnumeric)(Mat,Mat,Mat); 6930 PetscErrorCode (*numeric)(Mat,Mat,Mat)=PETSC_NULL; 6931 6932 PetscFunctionBegin; 6933 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 6934 PetscValidType(A,1); 6935 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6936 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6937 6938 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 6939 PetscValidType(B,2); 6940 MatPreallocated(B); 6941 if (!B->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6942 if (B->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6943 6944 PetscValidHeaderSpecific(C,MAT_COOKIE,3); 6945 PetscValidType(C,3); 6946 MatPreallocated(C); 6947 if (!C->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 6948 if (C->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 6949 6950 if (B->cmap.N!=C->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->cmap.N,C->cmap.N); 6951 if (B->rmap.N!=A->cmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap.N,A->cmap.N); 6952 if (A->rmap.N!=C->rmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",A->rmap.N,C->rmap.N); 6953 ierr = MatPreallocated(A);CHKERRQ(ierr); 6954 6955 Anumeric = A->ops->matmultnumeric; 6956 Bnumeric = B->ops->matmultnumeric; 6957 if (Anumeric == Bnumeric){ 6958 if (!Bnumeric) SETERRQ1(PETSC_ERR_SUP,"MatMatMultNumeric not supported for B of type %s",B->type_name); 6959 numeric = Bnumeric; 6960 } else { 6961 char numericname[256]; 6962 ierr = PetscStrcpy(numericname,"MatMatMultNumeric_");CHKERRQ(ierr); 6963 ierr = PetscStrcat(numericname,A->type_name);CHKERRQ(ierr); 6964 ierr = PetscStrcat(numericname,"_");CHKERRQ(ierr); 6965 ierr = PetscStrcat(numericname,B->type_name);CHKERRQ(ierr); 6966 ierr = PetscStrcat(numericname,"_C");CHKERRQ(ierr); 6967 ierr = PetscObjectQueryFunction((PetscObject)B,numericname,(void (**)(void))&numeric);CHKERRQ(ierr); 6968 if (!numeric) 6969 SETERRQ2(PETSC_ERR_ARG_INCOMP,"MatMatMultNumeric requires A, %s, to be compatible with B, %s",A->type_name,B->type_name); 6970 } 6971 ierr = PetscLogEventBegin(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 6972 ierr = (*numeric)(A,B,C);CHKERRQ(ierr); 6973 ierr = PetscLogEventEnd(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 6974 PetscFunctionReturn(0); 6975 } 6976 6977 #undef __FUNCT__ 6978 #define __FUNCT__ "MatMatMultTranspose" 6979 /*@ 6980 MatMatMultTranspose - Performs Matrix-Matrix Multiplication C=A^T*B. 6981 6982 Collective on Mat 6983 6984 Input Parameters: 6985 + A - the left matrix 6986 . B - the right matrix 6987 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 6988 - fill - expected fill as ratio of nnz(C)/(nnz(A) + nnz(B)) 6989 6990 Output Parameters: 6991 . C - the product matrix 6992 6993 Notes: 6994 C will be created and must be destroyed by the user with MatDestroy(). 6995 6996 This routine is currently only implemented for pairs of SeqAIJ matrices and classes 6997 which inherit from SeqAIJ. C will be of type MATSEQAIJ. 6998 6999 Level: intermediate 7000 7001 .seealso: MatMatMultTransposeSymbolic(), MatMatMultTransposeNumeric(), MatPtAP() 7002 @*/ 7003 PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultTranspose(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 7004 { 7005 PetscErrorCode ierr; 7006 PetscErrorCode (*fA)(Mat,Mat,MatReuse,PetscReal,Mat*); 7007 PetscErrorCode (*fB)(Mat,Mat,MatReuse,PetscReal,Mat*); 7008 7009 PetscFunctionBegin; 7010 PetscValidHeaderSpecific(A,MAT_COOKIE,1); 7011 PetscValidType(A,1); 7012 if (!A->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7013 if (A->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7014 PetscValidHeaderSpecific(B,MAT_COOKIE,2); 7015 PetscValidType(B,2); 7016 MatPreallocated(B); 7017 if (!B->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7018 if (B->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7019 PetscValidPointer(C,3); 7020 if (B->rmap.N!=A->rmap.N) SETERRQ2(PETSC_ERR_ARG_SIZ,"Matrix dimensions are incompatible, %D != %D",B->rmap.N,A->rmap.N); 7021 if (fill < 1.0) SETERRQ1(PETSC_ERR_ARG_SIZ,"Expected fill=%G must be > 1.0",fill); 7022 ierr = MatPreallocated(A);CHKERRQ(ierr); 7023 7024 fA = A->ops->matmulttranspose; 7025 if (!fA) SETERRQ1(PETSC_ERR_SUP,"MatMatMultTranspose not supported for A of type %s",A->type_name); 7026 fB = B->ops->matmulttranspose; 7027 if (!fB) SETERRQ1(PETSC_ERR_SUP,"MatMatMultTranspose not supported for B of type %s",B->type_name); 7028 if (fB!=fA) SETERRQ2(PETSC_ERR_ARG_INCOMP,"MatMatMultTranspose requires A, %s, to be compatible with B, %s",A->type_name,B->type_name); 7029 7030 ierr = PetscLogEventBegin(MAT_MatMultTranspose,A,B,0,0);CHKERRQ(ierr); 7031 ierr = (*A->ops->matmulttranspose)(A,B,scall,fill,C);CHKERRQ(ierr); 7032 ierr = PetscLogEventEnd(MAT_MatMultTranspose,A,B,0,0);CHKERRQ(ierr); 7033 7034 PetscFunctionReturn(0); 7035 } 7036 7037 #undef __FUNCT__ 7038 #define __FUNCT__ "MatGetRedundantMatrix" 7039 /*@C 7040 MatGetRedundantMatrix - Create redundant matrices and put them into processors of subcommunicators. 7041 7042 Collective on Mat 7043 7044 Input Parameters: 7045 + mat - the matrix 7046 . nsubcomm - the number of subcommunicators (= number of redundant pareallel or sequential matrices) 7047 . subcomm - MPI communicator split from the communicator where mat resides in 7048 . mlocal_red - number of local rows of the redundant matrix 7049 - reuse - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 7050 7051 Output Parameter: 7052 . matredundant - redundant matrix 7053 7054 Notes: 7055 MAT_REUSE_MATRIX can only be used when the nonzero structure of the 7056 original matrix has not changed from that last call to MatGetRedundantMatrix(). 7057 7058 This routine creates the duplicated matrices in subcommunicators; you should NOT create them before 7059 calling it. 7060 7061 Only MPIAIJ matrix is supported. 7062 7063 Level: advanced 7064 7065 Concepts: subcommunicator 7066 Concepts: duplicate matrix 7067 7068 .seealso: MatDestroy() 7069 @*/ 7070 PetscErrorCode PETSCMAT_DLLEXPORT MatGetRedundantMatrix(Mat mat,PetscInt nsubcomm,MPI_Comm subcomm,PetscInt mlocal_red,MatReuse reuse,Mat *matredundant) 7071 { 7072 PetscErrorCode ierr; 7073 7074 PetscFunctionBegin; 7075 PetscValidHeaderSpecific(mat,MAT_COOKIE,1); 7076 if (nsubcomm && reuse == MAT_REUSE_MATRIX) { 7077 PetscValidPointer(*matredundant,6); 7078 PetscValidHeaderSpecific(*matredundant,MAT_COOKIE,6); 7079 } 7080 if (!mat->ops->getredundantmatrix) SETERRQ1(PETSC_ERR_SUP,"Mat type %s",mat->type_name); 7081 if (!mat->assembled) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 7082 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 7083 ierr = MatPreallocated(mat);CHKERRQ(ierr); 7084 7085 ierr = PetscLogEventBegin(MAT_GetRedundantMatrix,mat,0,0,0);CHKERRQ(ierr); 7086 ierr = (*mat->ops->getredundantmatrix)(mat,nsubcomm,subcomm,mlocal_red,reuse,matredundant);CHKERRQ(ierr); 7087 ierr = PetscLogEventEnd(MAT_GetRedundantMatrix,mat,0,0,0);CHKERRQ(ierr); 7088 PetscFunctionReturn(0); 7089 } 7090