Actual source code: factor.c
petsc-3.3-p7 2013-05-11
2: #include <../src/ksp/pc/impls/factor/factor.h> /*I "petscpc.h" I*/
6: /*@
7: PCFactorSetUpMatSolverPackage - Can be called after KSPSetOperators() or PCSetOperators(), causes MatGetFactor() to be called so then one may
8: set the options for that particular factorization object.
10: Input Parameter:
11: . pc - the preconditioner context
13: Notes: After you have called this function (which has to be after the KSPSetOperators() or PCSetOperators()) you can call PCFactorGetMatrix() and then set factor options on that matrix.
15: .seealso: PCFactorSetMatSolverPackage(), PCFactorGetMatrix()
17: Level: intermediate
19: @*/
20: PetscErrorCode PCFactorSetUpMatSolverPackage(PC pc)
21: {
26: PetscTryMethod(pc,"PCFactorSetUpMatSolverPackage_C",(PC),(pc));
27: return(0);
28: }
32: /*@
33: PCFactorSetZeroPivot - Sets the size at which smaller pivots are declared to be zero
35: Logically Collective on PC
36:
37: Input Parameters:
38: + pc - the preconditioner context
39: - zero - all pivots smaller than this will be considered zero
41: Options Database Key:
42: . -pc_factor_zeropivot <zero> - Sets tolerance for what is considered a zero pivot
44: Level: intermediate
46: .keywords: PC, set, factorization, direct, fill
48: .seealso: PCFactorSetShiftType(), PCFactorSetShiftAmount()
49: @*/
50: PetscErrorCode PCFactorSetZeroPivot(PC pc,PetscReal zero)
51: {
57: PetscTryMethod(pc,"PCFactorSetZeroPivot_C",(PC,PetscReal),(pc,zero));
58: return(0);
59: }
63: /*@
64: PCFactorSetShiftType - adds a particular type of quantity to the diagonal of the matrix during
65: numerical factorization, thus the matrix has nonzero pivots
67: Logically Collective on PC
68:
69: Input Parameters:
70: + pc - the preconditioner context
71: - shifttype - type of shift; one of MAT_SHIFT_NONE, MAT_SHIFT_NONZERO, MAT_SHIFT_POSITIVE_DEFINITE, MAT_SHIFT_INBLOCKS
73: Options Database Key:
74: . -pc_factor_shift_type <shifttype> - Sets shift type or PETSC_DECIDE for the default; use '-help' for a list of available types
76: Level: intermediate
78: .keywords: PC, set, factorization,
80: .seealso: PCFactorSetZeroPivot(), PCFactorSetShiftAmount()
81: @*/
82: PetscErrorCode PCFactorSetShiftType(PC pc,MatFactorShiftType shifttype)
83: {
89: PetscTryMethod(pc,"PCFactorSetShiftType_C",(PC,MatFactorShiftType),(pc,shifttype));
90: return(0);
91: }
95: /*@
96: PCFactorSetShiftAmount - adds a quantity to the diagonal of the matrix during
97: numerical factorization, thus the matrix has nonzero pivots
99: Logically Collective on PC
100:
101: Input Parameters:
102: + pc - the preconditioner context
103: - shiftamount - amount of shift
105: Options Database Key:
106: . -pc_factor_shift_amount <shiftamount> - Sets shift amount or PETSC_DECIDE for the default
108: Level: intermediate
110: .keywords: PC, set, factorization,
112: .seealso: PCFactorSetZeroPivot(), PCFactorSetShiftType()
113: @*/
114: PetscErrorCode PCFactorSetShiftAmount(PC pc,PetscReal shiftamount)
115: {
121: PetscTryMethod(pc,"PCFactorSetShiftAmount_C",(PC,PetscReal),(pc,shiftamount));
122: return(0);
123: }
127: /*
128: PCFactorSetDropTolerance - The preconditioner will use an ILU
129: based on a drop tolerance. (Under development)
131: Logically Collective on PC
133: Input Parameters:
134: + pc - the preconditioner context
135: . dt - the drop tolerance, try from 1.e-10 to .1
136: . dtcol - tolerance for column pivot, good values [0.1 to 0.01]
137: - maxrowcount - the max number of nonzeros allowed in a row, best value
138: depends on the number of nonzeros in row of original matrix
140: Options Database Key:
141: . -pc_factor_drop_tolerance <dt,dtcol,maxrowcount> - Sets drop tolerance
143: Level: intermediate
145: There are NO default values for the 3 parameters, you must set them with reasonable values for your
146: matrix. We don't know how to compute reasonable values.
148: .keywords: PC, levels, reordering, factorization, incomplete, ILU
149: */
150: PetscErrorCode PCFactorSetDropTolerance(PC pc,PetscReal dt,PetscReal dtcol,PetscInt maxrowcount)
151: {
158: PetscTryMethod(pc,"PCFactorSetDropTolerance_C",(PC,PetscReal,PetscReal,PetscInt),(pc,dt,dtcol,maxrowcount));
159: return(0);
160: }
164: /*@
165: PCFactorSetLevels - Sets the number of levels of fill to use.
167: Logically Collective on PC
169: Input Parameters:
170: + pc - the preconditioner context
171: - levels - number of levels of fill
173: Options Database Key:
174: . -pc_factor_levels <levels> - Sets fill level
176: Level: intermediate
178: .keywords: PC, levels, fill, factorization, incomplete, ILU
179: @*/
180: PetscErrorCode PCFactorSetLevels(PC pc,PetscInt levels)
181: {
186: if (levels < 0) SETERRQ(((PetscObject)pc)->comm,PETSC_ERR_ARG_OUTOFRANGE,"negative levels");
188: PetscTryMethod(pc,"PCFactorSetLevels_C",(PC,PetscInt),(pc,levels));
189: return(0);
190: }
194: /*@
195: PCFactorSetAllowDiagonalFill - Causes all diagonal matrix entries to be
196: treated as level 0 fill even if there is no non-zero location.
198: Logically Collective on PC
200: Input Parameters:
201: + pc - the preconditioner context
203: Options Database Key:
204: . -pc_factor_diagonal_fill
206: Notes:
207: Does not apply with 0 fill.
209: Level: intermediate
211: .keywords: PC, levels, fill, factorization, incomplete, ILU
212: @*/
213: PetscErrorCode PCFactorSetAllowDiagonalFill(PC pc)
214: {
219: PetscTryMethod(pc,"PCFactorSetAllowDiagonalFill_C",(PC),(pc));
220: return(0);
221: }
225: /*@
226: PCFactorReorderForNonzeroDiagonal - reorders rows/columns of matrix to remove zeros from diagonal
228: Logically Collective on PC
229:
230: Input Parameters:
231: + pc - the preconditioner context
232: - tol - diagonal entries smaller than this in absolute value are considered zero
234: Options Database Key:
235: . -pc_factor_nonzeros_along_diagonal
237: Level: intermediate
239: .keywords: PC, set, factorization, direct, fill
241: .seealso: PCFactorSetFill(), PCFactorSetShiftNonzero(), PCFactorSetZeroPivot(), MatReorderForNonzeroDiagonal()
242: @*/
243: PetscErrorCode PCFactorReorderForNonzeroDiagonal(PC pc,PetscReal rtol)
244: {
250: PetscTryMethod(pc,"PCFactorReorderForNonzeroDiagonal_C",(PC,PetscReal),(pc,rtol));
251: return(0);
252: }
256: /*@C
257: PCFactorSetMatSolverPackage - sets the software that is used to perform the factorization
259: Logically Collective on PC
260:
261: Input Parameters:
262: + pc - the preconditioner context
263: - stype - for example, spooles, superlu, superlu_dist
265: Options Database Key:
266: . -pc_factor_mat_solver_package <stype> - spooles, petsc, superlu, superlu_dist, mumps
268: Level: intermediate
270: Note:
271: By default this will use the PETSc factorization if it exists
272:
274: .keywords: PC, set, factorization, direct, fill
276: .seealso: MatGetFactor(), MatSolverPackage, PCFactorGetMatSolverPackage()
278: @*/
279: PetscErrorCode PCFactorSetMatSolverPackage(PC pc,const MatSolverPackage stype)
280: {
285: PetscTryMethod(pc,"PCFactorSetMatSolverPackage_C",(PC,const MatSolverPackage),(pc,stype));
286: return(0);
287: }
291: /*@C
292: PCFactorGetMatSolverPackage - gets the software that is used to perform the factorization
294: Not Collective
295:
296: Input Parameter:
297: . pc - the preconditioner context
299: Output Parameter:
300: . stype - for example, spooles, superlu, superlu_dist (PETSC_NULL if the PC does not have a solver package)
302: Level: intermediate
305: .keywords: PC, set, factorization, direct, fill
307: .seealso: MatGetFactor(), MatSolverPackage, PCFactorGetMatSolverPackage()
309: @*/
310: PetscErrorCode PCFactorGetMatSolverPackage(PC pc,const MatSolverPackage *stype)
311: {
312: PetscErrorCode ierr,(*f)(PC,const MatSolverPackage*);
316: PetscObjectQueryFunction((PetscObject)pc,"PCFactorGetMatSolverPackage_C",(void(**)(void))&f);
317: if (f) {
318: (*f)(pc,stype);
319: } else {
320: *stype = PETSC_NULL;
321: }
322: return(0);
323: }
327: /*@
328: PCFactorSetFill - Indicate the amount of fill you expect in the factored matrix,
329: fill = number nonzeros in factor/number nonzeros in original matrix.
331: Not Collective, each process can expect a different amount of fill
332:
333: Input Parameters:
334: + pc - the preconditioner context
335: - fill - amount of expected fill
337: Options Database Key:
338: . -pc_factor_fill <fill> - Sets fill amount
340: Level: intermediate
342: Note:
343: For sparse matrix factorizations it is difficult to predict how much
344: fill to expect. By running with the option -info PETSc will print the
345: actual amount of fill used; allowing you to set the value accurately for
346: future runs. Default PETSc uses a value of 5.0
348: This parameter has NOTHING to do with the levels-of-fill of ILU(). That is set with PCFactorSetLevels() or -pc_factor_levels.
349:
351: .keywords: PC, set, factorization, direct, fill
353: @*/
354: PetscErrorCode PCFactorSetFill(PC pc,PetscReal fill)
355: {
360: if (fill < 1.0) SETERRQ(((PetscObject)pc)->comm,PETSC_ERR_ARG_OUTOFRANGE,"Fill factor cannot be less then 1.0");
361: PetscTryMethod(pc,"PCFactorSetFill_C",(PC,PetscReal),(pc,fill));
362: return(0);
363: }
367: /*@
368: PCFactorSetUseInPlace - Tells the system to do an in-place factorization.
369: For dense matrices, this enables the solution of much larger problems.
370: For sparse matrices the factorization cannot be done truly in-place
371: so this does not save memory during the factorization, but after the matrix
372: is factored, the original unfactored matrix is freed, thus recovering that
373: space.
375: Logically Collective on PC
377: Input Parameters:
378: . pc - the preconditioner context
380: Options Database Key:
381: . -pc_factor_in_place - Activates in-place factorization
383: Notes:
384: PCFactorSetUseInplace() can only be used with the KSP method KSPPREONLY or when
385: a different matrix is provided for the multiply and the preconditioner in
386: a call to KSPSetOperators().
387: This is because the Krylov space methods require an application of the
388: matrix multiplication, which is not possible here because the matrix has
389: been factored in-place, replacing the original matrix.
391: Level: intermediate
393: .keywords: PC, set, factorization, direct, inplace, in-place, LU
395: .seealso: PCILUSetUseInPlace()
396: @*/
397: PetscErrorCode PCFactorSetUseInPlace(PC pc)
398: {
403: PetscTryMethod(pc,"PCFactorSetUseInPlace_C",(PC),(pc));
404: return(0);
405: }
409: /*@C
410: PCFactorSetMatOrderingType - Sets the ordering routine (to reduce fill) to
411: be used in the LU factorization.
413: Logically Collective on PC
415: Input Parameters:
416: + pc - the preconditioner context
417: - ordering - the matrix ordering name, for example, MATORDERINGND or MATORDERINGRCM
419: Options Database Key:
420: . -pc_factor_mat_ordering_type <nd,rcm,...> - Sets ordering routine
422: Level: intermediate
424: Notes: nested dissection is used by default
426: For Cholesky and ICC and the SBAIJ format reorderings are not available,
427: since only the upper triangular part of the matrix is stored. You can use the
428: SeqAIJ format in this case to get reorderings.
430: @*/
431: PetscErrorCode PCFactorSetMatOrderingType(PC pc,const MatOrderingType ordering)
432: {
437: PetscTryMethod(pc,"PCFactorSetMatOrderingType_C",(PC,const MatOrderingType),(pc,ordering));
438: return(0);
439: }
443: /*@
444: PCFactorSetColumnPivot - Determines when column pivoting is done during matrix factorization.
445: For PETSc dense matrices column pivoting is always done, for PETSc sparse matrices
446: it is never done. For the MATLAB and SuperLU factorization this is used.
448: Logically Collective on PC
450: Input Parameters:
451: + pc - the preconditioner context
452: - dtcol - 0.0 implies no pivoting, 1.0 complete pivoting (slower, requires more memory but more stable)
454: Options Database Key:
455: . -pc_factor_pivoting <dtcol>
457: Level: intermediate
459: .seealso: PCILUSetMatOrdering(), PCFactorSetPivotInBlocks()
460: @*/
461: PetscErrorCode PCFactorSetColumnPivot(PC pc,PetscReal dtcol)
462: {
468: PetscTryMethod(pc,"PCFactorSetColumnPivot_C",(PC,PetscReal),(pc,dtcol));
469: return(0);
470: }
474: /*@
475: PCFactorSetPivotInBlocks - Determines if pivoting is done while factoring each block
476: with BAIJ or SBAIJ matrices
478: Logically Collective on PC
480: Input Parameters:
481: + pc - the preconditioner context
482: - pivot - PETSC_TRUE or PETSC_FALSE
484: Options Database Key:
485: . -pc_factor_pivot_in_blocks <true,false>
487: Level: intermediate
489: .seealso: PCILUSetMatOrdering(), PCFactorSetColumnPivot()
490: @*/
491: PetscErrorCode PCFactorSetPivotInBlocks(PC pc,PetscBool pivot)
492: {
498: PetscTryMethod(pc,"PCFactorSetPivotInBlocks_C",(PC,PetscBool),(pc,pivot));
499: return(0);
500: }
504: /*@
505: PCFactorSetReuseFill - When matrices with same different nonzero structure are factored,
506: this causes later ones to use the fill ratio computed in the initial factorization.
508: Logically Collective on PC
510: Input Parameters:
511: + pc - the preconditioner context
512: - flag - PETSC_TRUE to reuse else PETSC_FALSE
514: Options Database Key:
515: . -pc_factor_reuse_fill - Activates PCFactorSetReuseFill()
517: Level: intermediate
519: .keywords: PC, levels, reordering, factorization, incomplete, Cholesky
521: .seealso: PCFactorSetReuseOrdering()
522: @*/
523: PetscErrorCode PCFactorSetReuseFill(PC pc,PetscBool flag)
524: {
530: PetscTryMethod(pc,"PCFactorSetReuseFill_C",(PC,PetscBool),(pc,flag));
531: return(0);
532: }