Actual source code: itfunc.c
petsc-3.6.4 2016-04-12
2: /*
3: Interface KSP routines that the user calls.
4: */
6: #include <petsc/private/kspimpl.h> /*I "petscksp.h" I*/
7: #include <petscdm.h>
11: /*@
12: KSPComputeExtremeSingularValues - Computes the extreme singular values
13: for the preconditioned operator. Called after or during KSPSolve().
15: Not Collective
17: Input Parameter:
18: . ksp - iterative context obtained from KSPCreate()
20: Output Parameters:
21: . emin, emax - extreme singular values
23: Options Database Keys:
24: . -ksp_compute_singularvalues - compute extreme singular values and print when KSPSolve completes.
26: Notes:
27: One must call KSPSetComputeSingularValues() before calling KSPSetUp()
28: (or use the option -ksp_compute_eigenvalues) in order for this routine to work correctly.
30: Many users may just want to use the monitoring routine
31: KSPMonitorSingularValue() (which can be set with option -ksp_monitor_singular_value)
32: to print the extreme singular values at each iteration of the linear solve.
34: Estimates of the smallest singular value may be very inaccurate, especially if the Krylov method has not converged.
35: The largest singular value is usually accurate to within a few percent if the method has converged, but is still not
36: intended for eigenanalysis.
38: Disable restarts if using KSPGMRES, otherwise this estimate will only be using those iterations after the last
39: restart. See KSPGMRESSetRestart() for more details.
41: Level: advanced
43: .keywords: KSP, compute, extreme, singular, values
45: .seealso: KSPSetComputeSingularValues(), KSPMonitorSingularValue(), KSPComputeEigenvalues()
46: @*/
47: PetscErrorCode KSPComputeExtremeSingularValues(KSP ksp,PetscReal *emax,PetscReal *emin)
48: {
55: if (!ksp->calc_sings) SETERRQ(PetscObjectComm((PetscObject)ksp),4,"Singular values not requested before KSPSetUp()");
57: if (ksp->ops->computeextremesingularvalues) {
58: (*ksp->ops->computeextremesingularvalues)(ksp,emax,emin);
59: } else {
60: *emin = -1.0;
61: *emax = -1.0;
62: }
63: return(0);
64: }
68: /*@
69: KSPComputeEigenvalues - Computes the extreme eigenvalues for the
70: preconditioned operator. Called after or during KSPSolve().
72: Not Collective
74: Input Parameter:
75: + ksp - iterative context obtained from KSPCreate()
76: - n - size of arrays r and c. The number of eigenvalues computed (neig) will, in
77: general, be less than this.
79: Output Parameters:
80: + r - real part of computed eigenvalues, provided by user with a dimension of at least n
81: . c - complex part of computed eigenvalues, provided by user with a dimension of at least n
82: - neig - actual number of eigenvalues computed (will be less than or equal to n)
84: Options Database Keys:
85: + -ksp_compute_eigenvalues - Prints eigenvalues to stdout
86: - -ksp_plot_eigenvalues - Plots eigenvalues in an x-window display
88: Notes:
89: The number of eigenvalues estimated depends on the size of the Krylov space
90: generated during the KSPSolve() ; for example, with
91: CG it corresponds to the number of CG iterations, for GMRES it is the number
92: of GMRES iterations SINCE the last restart. Any extra space in r[] and c[]
93: will be ignored.
95: KSPComputeEigenvalues() does not usually provide accurate estimates; it is
96: intended only for assistance in understanding the convergence of iterative
97: methods, not for eigenanalysis. For accurate computation of eigenvalues we recommend using
98: the excellent package SLEPc.
100: One must call KSPSetComputeEigenvalues() before calling KSPSetUp()
101: in order for this routine to work correctly.
103: Many users may just want to use the monitoring routine
104: KSPMonitorSingularValue() (which can be set with option -ksp_monitor_singular_value)
105: to print the singular values at each iteration of the linear solve.
107: Level: advanced
109: .keywords: KSP, compute, extreme, singular, values
111: .seealso: KSPSetComputeSingularValues(), KSPMonitorSingularValue(), KSPComputeExtremeSingularValues()
112: @*/
113: PetscErrorCode KSPComputeEigenvalues(KSP ksp,PetscInt n,PetscReal r[],PetscReal c[],PetscInt *neig)
114: {
121: if (n<0) SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Requested < 0 Eigenvalues");
123: if (!ksp->calc_sings) SETERRQ(PetscObjectComm((PetscObject)ksp),4,"Eigenvalues not requested before KSPSetUp()");
125: if (ksp->ops->computeeigenvalues) {
126: (*ksp->ops->computeeigenvalues)(ksp,n,r,c,neig);
127: } else {
128: *neig = 0;
129: }
130: return(0);
131: }
135: /*@
136: KSPSetUpOnBlocks - Sets up the preconditioner for each block in
137: the block Jacobi, block Gauss-Seidel, and overlapping Schwarz
138: methods.
140: Collective on KSP
142: Input Parameter:
143: . ksp - the KSP context
145: Notes:
146: KSPSetUpOnBlocks() is a routine that the user can optinally call for
147: more precise profiling (via -log_summary) of the setup phase for these
148: block preconditioners. If the user does not call KSPSetUpOnBlocks(),
149: it will automatically be called from within KSPSolve().
151: Calling KSPSetUpOnBlocks() is the same as calling PCSetUpOnBlocks()
152: on the PC context within the KSP context.
154: Level: advanced
156: .keywords: KSP, setup, blocks
158: .seealso: PCSetUpOnBlocks(), KSPSetUp(), PCSetUp()
159: @*/
160: PetscErrorCode KSPSetUpOnBlocks(KSP ksp)
161: {
166: if (!ksp->pc) {KSPGetPC(ksp,&ksp->pc);}
167: PCSetUpOnBlocks(ksp->pc);
168: return(0);
169: }
173: /*@
174: KSPSetReusePreconditioner - reuse the current preconditioner, do not construct a new one even if the operator changes
176: Collective on KSP
178: Input Parameters:
179: + ksp - iterative context obtained from KSPCreate()
180: - flag - PETSC_TRUE to reuse the current preconditioner
182: Level: intermediate
184: .keywords: KSP, setup
186: .seealso: KSPCreate(), KSPSolve(), KSPDestroy(), PCSetReusePreconditioner()
187: @*/
188: PetscErrorCode KSPSetReusePreconditioner(KSP ksp,PetscBool flag)
189: {
194: PCSetReusePreconditioner(ksp->pc,flag);
195: return(0);
196: }
200: /*@
201: KSPSetSkipPCSetFromOptions - prevents KSPSetFromOptions() from call PCSetFromOptions(). This is used if the same PC is shared by more than one KSP so its options are not resetable for each KSP
203: Collective on KSP
205: Input Parameters:
206: + ksp - iterative context obtained from KSPCreate()
207: - flag - PETSC_TRUE to skip calling the PCSetFromOptions()
209: Level: intermediate
211: .keywords: KSP, setup
213: .seealso: KSPCreate(), KSPSolve(), KSPDestroy(), PCSetReusePreconditioner()
214: @*/
215: PetscErrorCode KSPSetSkipPCSetFromOptions(KSP ksp,PetscBool flag)
216: {
219: ksp->skippcsetfromoptions = flag;
220: return(0);
221: }
225: /*@
226: KSPSetUp - Sets up the internal data structures for the
227: later use of an iterative solver.
229: Collective on KSP
231: Input Parameter:
232: . ksp - iterative context obtained from KSPCreate()
234: Level: developer
236: .keywords: KSP, setup
238: .seealso: KSPCreate(), KSPSolve(), KSPDestroy()
239: @*/
240: PetscErrorCode KSPSetUp(KSP ksp)
241: {
243: Mat A,B;
244: Mat mat,pmat;
245: MatNullSpace nullsp;
246:
250: /* reset the convergence flag from the previous solves */
251: ksp->reason = KSP_CONVERGED_ITERATING;
253: if (!((PetscObject)ksp)->type_name) {
254: KSPSetType(ksp,KSPGMRES);
255: }
256: KSPSetUpNorms_Private(ksp,&ksp->normtype,&ksp->pc_side);
258: if (ksp->dmActive && !ksp->setupstage) {
259: /* first time in so build matrix and vector data structures using DM */
260: if (!ksp->vec_rhs) {DMCreateGlobalVector(ksp->dm,&ksp->vec_rhs);}
261: if (!ksp->vec_sol) {DMCreateGlobalVector(ksp->dm,&ksp->vec_sol);}
262: DMCreateMatrix(ksp->dm,&A);
263: KSPSetOperators(ksp,A,A);
264: PetscObjectDereference((PetscObject)A);
265: }
267: if (ksp->dmActive) {
268: DMKSP kdm;
269: DMGetDMKSP(ksp->dm,&kdm);
271: if (kdm->ops->computeinitialguess && ksp->setupstage != KSP_SETUP_NEWRHS) {
272: /* only computes initial guess the first time through */
273: (*kdm->ops->computeinitialguess)(ksp,ksp->vec_sol,kdm->initialguessctx);
274: KSPSetInitialGuessNonzero(ksp,PETSC_TRUE);
275: }
276: if (kdm->ops->computerhs) {
277: (*kdm->ops->computerhs)(ksp,ksp->vec_rhs,kdm->rhsctx);
278: }
280: if (ksp->setupstage != KSP_SETUP_NEWRHS) {
281: if (kdm->ops->computeoperators) {
282: KSPGetOperators(ksp,&A,&B);
283: (*kdm->ops->computeoperators)(ksp,A,B,kdm->operatorsctx);
284: KSPSetOperators(ksp,A,B);
285: } else SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_WRONGSTATE,"You called KSPSetDM() but did not use DMKSPSetComputeOperators() or KSPSetDMActive(dm,PETSC_FALSE);");
286: }
287: }
289: if (ksp->setupstage == KSP_SETUP_NEWRHS) return(0);
290: PetscLogEventBegin(KSP_SetUp,ksp,ksp->vec_rhs,ksp->vec_sol,0);
292: switch (ksp->setupstage) {
293: case KSP_SETUP_NEW:
294: (*ksp->ops->setup)(ksp);
295: break;
296: case KSP_SETUP_NEWMATRIX: { /* This should be replaced with a more general mechanism */
297: } break;
298: default: break;
299: }
301: PCGetOperators(ksp->pc,&mat,&pmat);
302: /* scale the matrix if requested */
303: if (ksp->dscale) {
304: PetscScalar *xx;
305: PetscInt i,n;
306: PetscBool zeroflag = PETSC_FALSE;
307: if (!ksp->pc) {KSPGetPC(ksp,&ksp->pc);}
308: if (!ksp->diagonal) { /* allocate vector to hold diagonal */
309: MatCreateVecs(pmat,&ksp->diagonal,0);
310: }
311: MatGetDiagonal(pmat,ksp->diagonal);
312: VecGetLocalSize(ksp->diagonal,&n);
313: VecGetArray(ksp->diagonal,&xx);
314: for (i=0; i<n; i++) {
315: if (xx[i] != 0.0) xx[i] = 1.0/PetscSqrtReal(PetscAbsScalar(xx[i]));
316: else {
317: xx[i] = 1.0;
318: zeroflag = PETSC_TRUE;
319: }
320: }
321: VecRestoreArray(ksp->diagonal,&xx);
322: if (zeroflag) {
323: PetscInfo(ksp,"Zero detected in diagonal of matrix, using 1 at those locations\n");
324: }
325: MatDiagonalScale(pmat,ksp->diagonal,ksp->diagonal);
326: if (mat != pmat) {MatDiagonalScale(mat,ksp->diagonal,ksp->diagonal);}
327: ksp->dscalefix2 = PETSC_FALSE;
328: }
329: PetscLogEventEnd(KSP_SetUp,ksp,ksp->vec_rhs,ksp->vec_sol,0);
330: if (!ksp->pc) {KSPGetPC(ksp,&ksp->pc);}
331: PCSetErrorIfFailure(ksp->pc,ksp->errorifnotconverged);
332: PCSetUp(ksp->pc);
333: MatGetNullSpace(mat,&nullsp);
334: if (nullsp) {
335: PetscBool test = PETSC_FALSE;
336: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_test_null_space",&test,NULL);
337: if (test) {
338: MatNullSpaceTest(nullsp,mat,NULL);
339: }
340: }
341: ksp->setupstage = KSP_SETUP_NEWRHS;
342: return(0);
343: }
347: /*@
348: KSPReasonView - Displays the reason a KSP solve converged or diverged to a viewer
350: Collective on KSP
352: Parameter:
353: + ksp - iterative context obtained from KSPCreate()
354: - viewer - the viewer to display the reason
357: Options Database Keys:
358: . -ksp_converged_reason - print reason for converged or diverged, also prints number of iterations
360: Level: beginner
362: .keywords: KSP, solve, linear system
364: .seealso: KSPCreate(), KSPSetUp(), KSPDestroy(), KSPSetTolerances(), KSPConvergedDefault(),
365: KSPSolveTranspose(), KSPGetIterationNumber()
366: @*/
367: PetscErrorCode KSPReasonView(KSP ksp,PetscViewer viewer)
368: {
370: PetscBool isAscii;
373: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&isAscii);
374: if (isAscii) {
375: PetscViewerASCIIAddTab(viewer,((PetscObject)ksp)->tablevel);
376: if (ksp->reason > 0) {
377: if (((PetscObject) ksp)->prefix) {
378: PetscViewerASCIIPrintf(viewer,"Linear %s solve converged due to %s iterations %D\n",((PetscObject) ksp)->prefix,KSPConvergedReasons[ksp->reason],ksp->its);
379: } else {
380: PetscViewerASCIIPrintf(viewer,"Linear solve converged due to %s iterations %D\n",KSPConvergedReasons[ksp->reason],ksp->its);
381: }
382: } else {
383: if (((PetscObject) ksp)->prefix) {
384: PetscViewerASCIIPrintf(viewer,"Linear %s solve did not converge due to %s iterations %D\n",((PetscObject) ksp)->prefix,KSPConvergedReasons[ksp->reason],ksp->its);
385: } else {
386: PetscViewerASCIIPrintf(viewer,"Linear solve did not converge due to %s iterations %D\n",KSPConvergedReasons[ksp->reason],ksp->its);
387: }
388: }
389: PetscViewerASCIISubtractTab(viewer,((PetscObject)ksp)->tablevel);
390: }
391: return(0);
392: }
395: #if defined(PETSC_HAVE_THREADSAFETY)
396: #define KSPReasonViewFromOptions KSPReasonViewFromOptionsUnsafe
398: #else
400: #endif
401: /*@C
402: KSPReasonViewFromOptions - Processes command line options to determine if/how a KSPReason is to be viewed.
404: Collective on KSP
406: Input Parameters:
407: . ksp - the KSP object
409: Level: intermediate
411: @*/
412: PetscErrorCode KSPReasonViewFromOptions(KSP ksp)
413: {
414: PetscErrorCode ierr;
415: PetscViewer viewer;
416: PetscBool flg;
417: static PetscBool incall = PETSC_FALSE;
418: PetscViewerFormat format;
421: if (incall) return(0);
422: incall = PETSC_TRUE;
423: PetscOptionsGetViewer(PetscObjectComm((PetscObject)ksp),((PetscObject)ksp)->prefix,"-ksp_converged_reason",&viewer,&format,&flg);
424: if (flg) {
425: PetscViewerPushFormat(viewer,format);
426: KSPReasonView(ksp,viewer);
427: PetscViewerPopFormat(viewer);
428: PetscViewerDestroy(&viewer);
429: }
430: incall = PETSC_FALSE;
431: return(0);
432: }
434: #if defined(PETSC_HAVE_THREADSAFETY)
435: #undef KSPReasonViewFromOptions
436: PetscErrorCode KSPReasonViewFromOptions(KSP ksp)
437: {
439: #pragma omp critical
440: KSPReasonViewFromOptionsUnsafe(ksp);
441: return ierr;
442: }
443: #endif
445: #include <petscdraw.h>
448: /*@
449: KSPSolve - Solves linear system.
451: Collective on KSP
453: Parameter:
454: + ksp - iterative context obtained from KSPCreate()
455: . b - the right hand side vector
456: - x - the solution (this may be the same vector as b, then b will be overwritten with answer)
458: Options Database Keys:
459: + -ksp_compute_eigenvalues - compute preconditioned operators eigenvalues
460: . -ksp_plot_eigenvalues - plot the computed eigenvalues in an X-window
461: . -ksp_compute_eigenvalues_explicitly - compute the eigenvalues by forming the dense operator and useing LAPACK
462: . -ksp_plot_eigenvalues_explicitly - plot the explicitly computing eigenvalues
463: . -ksp_view_mat binary - save matrix to the default binary viewer
464: . -ksp_view_pmat binary - save matrix used to build preconditioner to the default binary viewer
465: . -ksp_view_rhs binary - save right hand side vector to the default binary viewer
466: . -ksp_view_solution binary - save computed solution vector to the default binary viewer
467: (can be read later with src/ksp/examples/tutorials/ex10.c for testing solvers)
468: . -ksp_view_mat_explicit - for matrix-free operators, computes the matrix entries and views them
469: . -ksp_view_preconditioned_operator_explicit - computes the product of the preconditioner and matrix as an explicit matrix and views it
470: . -ksp_converged_reason - print reason for converged or diverged, also prints number of iterations
471: . -ksp_final_residual - print 2-norm of true linear system residual at the end of the solution process
472: - -ksp_view - print the ksp data structure at the end of the system solution
474: Notes:
476: If one uses KSPSetDM() then x or b need not be passed. Use KSPGetSolution() to access the solution in this case.
478: The operator is specified with KSPSetOperators().
480: Call KSPGetConvergedReason() to determine if the solver converged or failed and
481: why. The number of iterations can be obtained from KSPGetIterationNumber().
483: If using a direct method (e.g., via the KSP solver
484: KSPPREONLY and a preconditioner such as PCLU/PCILU),
485: then its=1. See KSPSetTolerances() and KSPConvergedDefault()
486: for more details.
488: Understanding Convergence:
489: The routines KSPMonitorSet(), KSPComputeEigenvalues(), and
490: KSPComputeEigenvaluesExplicitly() provide information on additional
491: options to monitor convergence and print eigenvalue information.
493: Level: beginner
495: .keywords: KSP, solve, linear system
497: .seealso: KSPCreate(), KSPSetUp(), KSPDestroy(), KSPSetTolerances(), KSPConvergedDefault(),
498: KSPSolveTranspose(), KSPGetIterationNumber()
499: @*/
500: PetscErrorCode KSPSolve(KSP ksp,Vec b,Vec x)
501: {
502: PetscErrorCode ierr;
503: PetscMPIInt rank;
504: PetscBool flag1,flag2,flag3,flg = PETSC_FALSE,inXisinB=PETSC_FALSE,guess_zero;
505: Mat mat,pmat;
506: MPI_Comm comm;
507: PetscInt pcreason;
508: MatNullSpace nullsp;
509: Vec btmp,vec_rhs=0;
515: comm = PetscObjectComm((PetscObject)ksp);
516: if (x && x == b) {
517: if (!ksp->guess_zero) SETERRQ(comm,PETSC_ERR_ARG_INCOMP,"Cannot use x == b with nonzero initial guess");
518: VecDuplicate(b,&x);
519: inXisinB = PETSC_TRUE;
520: }
521: if (b) {
522: PetscObjectReference((PetscObject)b);
523: VecDestroy(&ksp->vec_rhs);
524: ksp->vec_rhs = b;
525: }
526: if (x) {
527: PetscObjectReference((PetscObject)x);
528: VecDestroy(&ksp->vec_sol);
529: ksp->vec_sol = x;
530: }
531: KSPViewFromOptions(ksp,NULL,"-ksp_view_pre");
533: if (ksp->presolve) {
534: (*ksp->presolve)(ksp,ksp->vec_rhs,ksp->vec_sol,ksp->prectx);
535: }
536: PetscLogEventBegin(KSP_Solve,ksp,ksp->vec_rhs,ksp->vec_sol,0);
538: /* reset the residual history list if requested */
539: if (ksp->res_hist_reset) ksp->res_hist_len = 0;
540: ksp->transpose_solve = PETSC_FALSE;
542: if (ksp->guess) {
543: KSPFischerGuessFormGuess(ksp->guess,ksp->vec_rhs,ksp->vec_sol);
544: ksp->guess_zero = PETSC_FALSE;
545: }
546: /* KSPSetUp() scales the matrix if needed */
547: KSPSetUp(ksp);
548: PCGetSetUpFailedReason(ksp->pc,&pcreason);
549: if (pcreason) {
550: ksp->reason = KSP_DIVERGED_PCSETUP_FAILED;
551: goto skipsolve;
552: }
553: KSPSetUpOnBlocks(ksp);
554: VecLocked(ksp->vec_sol,3);
556: PCGetOperators(ksp->pc,&mat,&pmat);
557: /* diagonal scale RHS if called for */
558: if (ksp->dscale) {
559: VecPointwiseMult(ksp->vec_rhs,ksp->vec_rhs,ksp->diagonal);
560: /* second time in, but matrix was scaled back to original */
561: if (ksp->dscalefix && ksp->dscalefix2) {
562: Mat mat,pmat;
564: PCGetOperators(ksp->pc,&mat,&pmat);
565: MatDiagonalScale(pmat,ksp->diagonal,ksp->diagonal);
566: if (mat != pmat) {MatDiagonalScale(mat,ksp->diagonal,ksp->diagonal);}
567: }
569: /* scale initial guess */
570: if (!ksp->guess_zero) {
571: if (!ksp->truediagonal) {
572: VecDuplicate(ksp->diagonal,&ksp->truediagonal);
573: VecCopy(ksp->diagonal,ksp->truediagonal);
574: VecReciprocal(ksp->truediagonal);
575: }
576: VecPointwiseMult(ksp->vec_sol,ksp->vec_sol,ksp->truediagonal);
577: }
578: }
579: PCPreSolve(ksp->pc,ksp);
581: if (ksp->guess_zero) { VecSet(ksp->vec_sol,0.0);}
582: if (ksp->guess_knoll) {
583: PCApply(ksp->pc,ksp->vec_rhs,ksp->vec_sol);
584: KSP_RemoveNullSpace(ksp,ksp->vec_sol);
585: ksp->guess_zero = PETSC_FALSE;
586: }
588: /* can we mark the initial guess as zero for this solve? */
589: guess_zero = ksp->guess_zero;
590: if (!ksp->guess_zero) {
591: PetscReal norm;
593: VecNormAvailable(ksp->vec_sol,NORM_2,&flg,&norm);
594: if (flg && !norm) ksp->guess_zero = PETSC_TRUE;
595: }
596: MatGetTransposeNullSpace(pmat,&nullsp);
597: if (nullsp) {
598: VecDuplicate(ksp->vec_rhs,&btmp);
599: VecCopy(ksp->vec_rhs,btmp);
600: MatNullSpaceRemove(nullsp,btmp);
601: vec_rhs = ksp->vec_rhs;
602: ksp->vec_rhs = btmp;
603: }
604: VecLockPush(ksp->vec_rhs);
605: (*ksp->ops->solve)(ksp);
606: VecLockPop(ksp->vec_rhs);
607: if (nullsp) {
608: ksp->vec_rhs = vec_rhs;
609: VecDestroy(&btmp);
610: }
612: ksp->guess_zero = guess_zero;
615: if (!ksp->reason) SETERRQ(comm,PETSC_ERR_PLIB,"Internal error, solver returned without setting converged reason");
616: ksp->totalits += ksp->its;
618: skipsolve:
619: KSPReasonViewFromOptions(ksp);
620: PCPostSolve(ksp->pc,ksp);
622: /* diagonal scale solution if called for */
623: if (ksp->dscale) {
624: VecPointwiseMult(ksp->vec_sol,ksp->vec_sol,ksp->diagonal);
625: /* unscale right hand side and matrix */
626: if (ksp->dscalefix) {
627: Mat mat,pmat;
629: VecReciprocal(ksp->diagonal);
630: VecPointwiseMult(ksp->vec_rhs,ksp->vec_rhs,ksp->diagonal);
631: PCGetOperators(ksp->pc,&mat,&pmat);
632: MatDiagonalScale(pmat,ksp->diagonal,ksp->diagonal);
633: if (mat != pmat) {MatDiagonalScale(mat,ksp->diagonal,ksp->diagonal);}
634: VecReciprocal(ksp->diagonal);
635: ksp->dscalefix2 = PETSC_TRUE;
636: }
637: }
638: PetscLogEventEnd(KSP_Solve,ksp,ksp->vec_rhs,ksp->vec_sol,0);
639: if (ksp->postsolve) {
640: (*ksp->postsolve)(ksp,ksp->vec_rhs,ksp->vec_sol,ksp->postctx);
641: }
643: if (ksp->guess) {
644: KSPFischerGuessUpdate(ksp->guess,ksp->vec_sol);
645: }
647: MPI_Comm_rank(comm,&rank);
649: MatViewFromOptions(mat,(PetscObject)ksp,"-ksp_view_mat");
650: MatViewFromOptions(pmat,(PetscObject)ksp,"-ksp_view_pmat");
651: VecViewFromOptions(ksp->vec_rhs,(PetscObject)ksp,"-ksp_view_rhs");
653: flag1 = PETSC_FALSE;
654: flag2 = PETSC_FALSE;
655: flag3 = PETSC_FALSE;
656: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_compute_eigenvalues",&flag1,NULL);
657: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_plot_eigenvalues",&flag2,NULL);
658: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_plot_eigencontours",&flag3,NULL);
659: if (flag1 || flag2 || flag3) {
660: PetscInt nits,n,i,neig;
661: PetscReal *r,*c;
663: KSPGetIterationNumber(ksp,&nits);
664: n = nits+2;
666: if (!nits) {
667: PetscPrintf(comm,"Zero iterations in solver, cannot approximate any eigenvalues\n");
668: } else {
669: PetscMalloc2(n,&r,n,&c);
670: KSPComputeEigenvalues(ksp,n,r,c,&neig);
671: if (flag1) {
672: PetscPrintf(comm,"Iteratively computed eigenvalues\n");
673: for (i=0; i<neig; i++) {
674: if (c[i] >= 0.0) {
675: PetscPrintf(comm,"%g + %gi\n",(double)r[i],(double)c[i]);
676: } else {
677: PetscPrintf(comm,"%g - %gi\n",(double)r[i],-(double)c[i]);
678: }
679: }
680: }
681: if (flag2 && !rank) {
682: PetscDraw draw;
683: PetscDrawSP drawsp;
685: if (!ksp->eigviewer) {
686: PetscViewerDrawOpen(PETSC_COMM_SELF,0,"Iteratively Computed Eigenvalues",PETSC_DECIDE,PETSC_DECIDE,400,400,&ksp->eigviewer);
687: }
688: PetscViewerDrawGetDraw(ksp->eigviewer,0,&draw);
689: PetscDrawSPCreate(draw,1,&drawsp);
690: PetscDrawSPReset(drawsp);
691: for (i=0; i<neig; i++) {
692: PetscDrawSPAddPoint(drawsp,r+i,c+i);
693: }
694: PetscDrawSPDraw(drawsp,PETSC_TRUE);
695: PetscDrawSPDestroy(&drawsp);
696: }
697: if (flag3 && !rank) {
698: KSPPlotEigenContours_Private(ksp,neig,r,c);
699: }
700: PetscFree2(r,c);
701: }
702: }
704: flag1 = PETSC_FALSE;
705: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_compute_singularvalues",&flag1,NULL);
706: if (flag1) {
707: PetscInt nits;
709: KSPGetIterationNumber(ksp,&nits);
710: if (!nits) {
711: PetscPrintf(comm,"Zero iterations in solver, cannot approximate any singular values\n");
712: } else {
713: PetscReal emax,emin;
715: KSPComputeExtremeSingularValues(ksp,&emax,&emin);
716: PetscPrintf(comm,"Iteratively computed extreme singular values: max %g min %g max/min %g\n",(double)emax,(double)emin,(double)(emax/emin));
717: }
718: }
721: flag1 = PETSC_FALSE;
722: flag2 = PETSC_FALSE;
723: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_compute_eigenvalues_explicitly",&flag1,NULL);
724: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_plot_eigenvalues_explicitly",&flag2,NULL);
725: if (flag1 || flag2) {
726: PetscInt n,i;
727: PetscReal *r,*c;
728: VecGetSize(ksp->vec_sol,&n);
729: PetscMalloc2(n,&r,n,&c);
730: KSPComputeEigenvaluesExplicitly(ksp,n,r,c);
731: if (flag1) {
732: PetscPrintf(comm,"Explicitly computed eigenvalues\n");
733: for (i=0; i<n; i++) {
734: if (c[i] >= 0.0) {
735: PetscPrintf(comm,"%g + %gi\n",(double)r[i],(double)c[i]);
736: } else {
737: PetscPrintf(comm,"%g - %gi\n",(double)r[i],-(double)c[i]);
738: }
739: }
740: }
741: if (flag2 && !rank) {
742: PetscDraw draw;
743: PetscDrawSP drawsp;
745: if (!ksp->eigviewer) {
746: PetscViewerDrawOpen(PETSC_COMM_SELF,0,"Explicitly Computed Eigenvalues",0,320,400,400,&ksp->eigviewer);
747: }
748: PetscViewerDrawGetDraw(ksp->eigviewer,0,&draw);
749: PetscDrawSPCreate(draw,1,&drawsp);
750: PetscDrawSPReset(drawsp);
751: for (i=0; i<n; i++) {
752: PetscDrawSPAddPoint(drawsp,r+i,c+i);
753: }
754: PetscDrawSPDraw(drawsp,PETSC_TRUE);
755: PetscDrawSPDestroy(&drawsp);
756: }
757: PetscFree2(r,c);
758: }
760: PetscOptionsHasName(((PetscObject)ksp)->prefix,"-ksp_view_mat_explicit",&flag2);
761: if (flag2) {
762: Mat A,B;
763: PCGetOperators(ksp->pc,&A,NULL);
764: MatComputeExplicitOperator(A,&B);
765: MatViewFromOptions(B,(PetscObject)ksp,"-ksp_view_mat_explicit");
766: MatDestroy(&B);
767: }
768: PetscOptionsHasName(((PetscObject)ksp)->prefix,"-ksp_view_preconditioned_operator_explicit",&flag2);
769: if (flag2) {
770: Mat B;
771: KSPComputeExplicitOperator(ksp,&B);
772: MatViewFromOptions(B,(PetscObject)ksp,"-ksp_view_preconditioned_operator_explicit");
773: MatDestroy(&B);
774: }
775: KSPViewFromOptions(ksp,NULL,"-ksp_view");
777: flg = PETSC_FALSE;
778: PetscOptionsGetBool(((PetscObject)ksp)->prefix,"-ksp_final_residual",&flg,NULL);
779: if (flg) {
780: Mat A;
781: Vec t;
782: PetscReal norm;
783: if (ksp->dscale && !ksp->dscalefix) SETERRQ(comm,PETSC_ERR_ARG_WRONGSTATE,"Cannot compute final scale with -ksp_diagonal_scale except also with -ksp_diagonal_scale_fix");
784: PCGetOperators(ksp->pc,&A,NULL);
785: VecDuplicate(ksp->vec_rhs,&t);
786: KSP_MatMult(ksp,A,ksp->vec_sol,t);
787: VecAYPX(t, -1.0, ksp->vec_rhs);
788: VecNorm(t,NORM_2,&norm);
789: VecDestroy(&t);
790: PetscPrintf(comm,"KSP final norm of residual %g\n",(double)norm);
791: }
792: VecViewFromOptions(ksp->vec_sol,(PetscObject)ksp,"-ksp_view_solution");
794: if (inXisinB) {
795: VecCopy(x,b);
796: VecDestroy(&x);
797: }
798: PetscObjectSAWsBlock((PetscObject)ksp);
799: if (ksp->errorifnotconverged && ksp->reason < 0) SETERRQ(comm,PETSC_ERR_NOT_CONVERGED,"KSPSolve has not converged");
800: return(0);
801: }
805: /*@
806: KSPSolveTranspose - Solves the transpose of a linear system.
808: Collective on KSP
810: Input Parameter:
811: + ksp - iterative context obtained from KSPCreate()
812: . b - right hand side vector
813: - x - solution vector
815: Notes: For complex numbers this solve the non-Hermitian transpose system.
817: Developer Notes: We need to implement a KSPSolveHermitianTranspose()
819: Level: developer
821: .keywords: KSP, solve, linear system
823: .seealso: KSPCreate(), KSPSetUp(), KSPDestroy(), KSPSetTolerances(), KSPConvergedDefault(),
824: KSPSolve()
825: @*/
827: PetscErrorCode KSPSolveTranspose(KSP ksp,Vec b,Vec x)
828: {
830: PetscBool inXisinB=PETSC_FALSE;
836: if (x == b) {
837: VecDuplicate(b,&x);
838: inXisinB = PETSC_TRUE;
839: }
840: PetscObjectReference((PetscObject)b);
841: PetscObjectReference((PetscObject)x);
842: VecDestroy(&ksp->vec_rhs);
843: VecDestroy(&ksp->vec_sol);
845: ksp->vec_rhs = b;
846: ksp->vec_sol = x;
847: ksp->transpose_solve = PETSC_TRUE;
849: KSPSetUp(ksp);
850: if (ksp->guess_zero) { VecSet(ksp->vec_sol,0.0);}
851: (*ksp->ops->solve)(ksp);
852: if (!ksp->reason) SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_PLIB,"Internal error, solver returned without setting converged reason");
853: KSPReasonViewFromOptions(ksp);
854: if (inXisinB) {
855: VecCopy(x,b);
856: VecDestroy(&x);
857: }
858: if (ksp->errorifnotconverged && ksp->reason < 0) SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_NOT_CONVERGED,"KSPSolve has not converged");
859: return(0);
860: }
864: /*@
865: KSPReset - Resets a KSP context to the kspsetupcalled = 0 state and removes any allocated Vecs and Mats
867: Collective on KSP
869: Input Parameter:
870: . ksp - iterative context obtained from KSPCreate()
872: Level: beginner
874: .keywords: KSP, destroy
876: .seealso: KSPCreate(), KSPSetUp(), KSPSolve()
877: @*/
878: PetscErrorCode KSPReset(KSP ksp)
879: {
884: if (!ksp) return(0);
885: if (ksp->ops->reset) {
886: (*ksp->ops->reset)(ksp);
887: }
888: if (ksp->pc) {PCReset(ksp->pc);}
889: KSPFischerGuessDestroy(&ksp->guess);
890: VecDestroyVecs(ksp->nwork,&ksp->work);
891: VecDestroy(&ksp->vec_rhs);
892: VecDestroy(&ksp->vec_sol);
893: VecDestroy(&ksp->diagonal);
894: VecDestroy(&ksp->truediagonal);
896: ksp->setupstage = KSP_SETUP_NEW;
897: return(0);
898: }
902: /*@
903: KSPDestroy - Destroys KSP context.
905: Collective on KSP
907: Input Parameter:
908: . ksp - iterative context obtained from KSPCreate()
910: Level: beginner
912: .keywords: KSP, destroy
914: .seealso: KSPCreate(), KSPSetUp(), KSPSolve()
915: @*/
916: PetscErrorCode KSPDestroy(KSP *ksp)
917: {
919: PC pc;
922: if (!*ksp) return(0);
924: if (--((PetscObject)(*ksp))->refct > 0) {*ksp = 0; return(0);}
926: PetscObjectSAWsViewOff((PetscObject)*ksp);
927: /*
928: Avoid a cascading call to PCReset(ksp->pc) from the following call:
929: PCReset() shouldn't be called from KSPDestroy() as it is unprotected by pc's
930: refcount (and may be shared, e.g., by other ksps).
931: */
932: pc = (*ksp)->pc;
933: (*ksp)->pc = NULL;
934: KSPReset((*ksp));
935: (*ksp)->pc = pc;
936: if ((*ksp)->ops->destroy) {(*(*ksp)->ops->destroy)(*ksp);}
938: DMDestroy(&(*ksp)->dm);
939: PCDestroy(&(*ksp)->pc);
940: PetscFree((*ksp)->res_hist_alloc);
941: if ((*ksp)->convergeddestroy) {
942: (*(*ksp)->convergeddestroy)((*ksp)->cnvP);
943: }
944: KSPMonitorCancel((*ksp));
945: PetscViewerDestroy(&(*ksp)->eigviewer);
946: PetscHeaderDestroy(ksp);
947: return(0);
948: }
952: /*@
953: KSPSetPCSide - Sets the preconditioning side.
955: Logically Collective on KSP
957: Input Parameter:
958: . ksp - iterative context obtained from KSPCreate()
960: Output Parameter:
961: . side - the preconditioning side, where side is one of
962: .vb
963: PC_LEFT - left preconditioning (default)
964: PC_RIGHT - right preconditioning
965: PC_SYMMETRIC - symmetric preconditioning
966: .ve
968: Options Database Keys:
969: . -ksp_pc_side <right,left,symmetric>
971: Notes:
972: Left preconditioning is used by default for most Krylov methods except KSPFGMRES which only supports right preconditioning.
974: For methods changing the side of the preconditioner changes the norm type that is used, see KSPSetNormType().
976: Symmetric preconditioning is currently available only for the KSPQCG method. Note, however, that
977: symmetric preconditioning can be emulated by using either right or left
978: preconditioning and a pre or post processing step.
980: Setting the PC side often affects the default norm type. See KSPSetNormType() for details.
982: Level: intermediate
984: .keywords: KSP, set, right, left, symmetric, side, preconditioner, flag
986: .seealso: KSPGetPCSide(), KSPSetNormType(), KSPGetNormType()
987: @*/
988: PetscErrorCode KSPSetPCSide(KSP ksp,PCSide side)
989: {
993: ksp->pc_side = ksp->pc_side_set = side;
994: return(0);
995: }
999: /*@
1000: KSPGetPCSide - Gets the preconditioning side.
1002: Not Collective
1004: Input Parameter:
1005: . ksp - iterative context obtained from KSPCreate()
1007: Output Parameter:
1008: . side - the preconditioning side, where side is one of
1009: .vb
1010: PC_LEFT - left preconditioning (default)
1011: PC_RIGHT - right preconditioning
1012: PC_SYMMETRIC - symmetric preconditioning
1013: .ve
1015: Level: intermediate
1017: .keywords: KSP, get, right, left, symmetric, side, preconditioner, flag
1019: .seealso: KSPSetPCSide()
1020: @*/
1021: PetscErrorCode KSPGetPCSide(KSP ksp,PCSide *side)
1022: {
1028: KSPSetUpNorms_Private(ksp,&ksp->normtype,&ksp->pc_side);
1029: *side = ksp->pc_side;
1030: return(0);
1031: }
1035: /*@
1036: KSPGetTolerances - Gets the relative, absolute, divergence, and maximum
1037: iteration tolerances used by the default KSP convergence tests.
1039: Not Collective
1041: Input Parameter:
1042: . ksp - the Krylov subspace context
1044: Output Parameters:
1045: + rtol - the relative convergence tolerance
1046: . abstol - the absolute convergence tolerance
1047: . dtol - the divergence tolerance
1048: - maxits - maximum number of iterations
1050: Notes:
1051: The user can specify NULL for any parameter that is not needed.
1053: Level: intermediate
1055: .keywords: KSP, get, tolerance, absolute, relative, divergence, convergence,
1056: maximum, iterations
1058: .seealso: KSPSetTolerances()
1059: @*/
1060: PetscErrorCode KSPGetTolerances(KSP ksp,PetscReal *rtol,PetscReal *abstol,PetscReal *dtol,PetscInt *maxits)
1061: {
1064: if (abstol) *abstol = ksp->abstol;
1065: if (rtol) *rtol = ksp->rtol;
1066: if (dtol) *dtol = ksp->divtol;
1067: if (maxits) *maxits = ksp->max_it;
1068: return(0);
1069: }
1073: /*@
1074: KSPSetTolerances - Sets the relative, absolute, divergence, and maximum
1075: iteration tolerances used by the default KSP convergence testers.
1077: Logically Collective on KSP
1079: Input Parameters:
1080: + ksp - the Krylov subspace context
1081: . rtol - the relative convergence tolerance, relative decrease in the (possibly preconditioned) residual norm
1082: . abstol - the absolute convergence tolerance absolute size of the (possibly preconditioned) residual norm
1083: . dtol - the divergence tolerance, amount (possibly preconditioned) residual norm can increase before KSPConvergedDefault() concludes that the method is diverging
1084: - maxits - maximum number of iterations to use
1086: Options Database Keys:
1087: + -ksp_atol <abstol> - Sets abstol
1088: . -ksp_rtol <rtol> - Sets rtol
1089: . -ksp_divtol <dtol> - Sets dtol
1090: - -ksp_max_it <maxits> - Sets maxits
1092: Notes:
1093: Use PETSC_DEFAULT to retain the default value of any of the tolerances.
1095: See KSPConvergedDefault() for details how these parameters are used in the default convergence test. See also KSPSetConvergenceTest()
1096: for setting user-defined stopping criteria.
1098: Level: intermediate
1100: .keywords: KSP, set, tolerance, absolute, relative, divergence,
1101: convergence, maximum, iterations
1103: .seealso: KSPGetTolerances(), KSPConvergedDefault(), KSPSetConvergenceTest()
1104: @*/
1105: PetscErrorCode KSPSetTolerances(KSP ksp,PetscReal rtol,PetscReal abstol,PetscReal dtol,PetscInt maxits)
1106: {
1114: if (rtol != PETSC_DEFAULT) {
1115: if (rtol < 0.0 || 1.0 <= rtol) SETERRQ1(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Relative tolerance %g must be non-negative and less than 1.0",(double)rtol);
1116: ksp->rtol = rtol;
1117: }
1118: if (abstol != PETSC_DEFAULT) {
1119: if (abstol < 0.0) SETERRQ1(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Absolute tolerance %g must be non-negative",(double)abstol);
1120: ksp->abstol = abstol;
1121: }
1122: if (dtol != PETSC_DEFAULT) {
1123: if (dtol < 0.0) SETERRQ1(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Divergence tolerance %g must be larger than 1.0",(double)dtol);
1124: ksp->divtol = dtol;
1125: }
1126: if (maxits != PETSC_DEFAULT) {
1127: if (maxits < 0) SETERRQ1(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Maximum number of iterations %D must be non-negative",maxits);
1128: ksp->max_it = maxits;
1129: }
1130: return(0);
1131: }
1135: /*@
1136: KSPSetInitialGuessNonzero - Tells the iterative solver that the
1137: initial guess is nonzero; otherwise KSP assumes the initial guess
1138: is to be zero (and thus zeros it out before solving).
1140: Logically Collective on KSP
1142: Input Parameters:
1143: + ksp - iterative context obtained from KSPCreate()
1144: - flg - PETSC_TRUE indicates the guess is non-zero, PETSC_FALSE indicates the guess is zero
1146: Options database keys:
1147: . -ksp_initial_guess_nonzero : use nonzero initial guess; this takes an optional truth value (0/1/no/yes/true/false)
1149: Level: beginner
1151: Notes:
1152: If this is not called the X vector is zeroed in the call to KSPSolve().
1154: .keywords: KSP, set, initial guess, nonzero
1156: .seealso: KSPGetInitialGuessNonzero(), KSPSetInitialGuessKnoll(), KSPGetInitialGuessKnoll()
1157: @*/
1158: PetscErrorCode KSPSetInitialGuessNonzero(KSP ksp,PetscBool flg)
1159: {
1163: ksp->guess_zero = (PetscBool) !(int)flg;
1164: return(0);
1165: }
1169: /*@
1170: KSPGetInitialGuessNonzero - Determines whether the KSP solver is using
1171: a zero initial guess.
1173: Not Collective
1175: Input Parameter:
1176: . ksp - iterative context obtained from KSPCreate()
1178: Output Parameter:
1179: . flag - PETSC_TRUE if guess is nonzero, else PETSC_FALSE
1181: Level: intermediate
1183: .keywords: KSP, set, initial guess, nonzero
1185: .seealso: KSPSetInitialGuessNonzero(), KSPSetInitialGuessKnoll(), KSPGetInitialGuessKnoll()
1186: @*/
1187: PetscErrorCode KSPGetInitialGuessNonzero(KSP ksp,PetscBool *flag)
1188: {
1192: if (ksp->guess_zero) *flag = PETSC_FALSE;
1193: else *flag = PETSC_TRUE;
1194: return(0);
1195: }
1199: /*@
1200: KSPSetErrorIfNotConverged - Causes KSPSolve() to generate an error if the solver has not converged.
1202: Logically Collective on KSP
1204: Input Parameters:
1205: + ksp - iterative context obtained from KSPCreate()
1206: - flg - PETSC_TRUE indicates you want the error generated
1208: Options database keys:
1209: . -ksp_error_if_not_converged : this takes an optional truth value (0/1/no/yes/true/false)
1211: Level: intermediate
1213: Notes:
1214: Normally PETSc continues if a linear solver fails to converge, you can call KSPGetConvergedReason() after a KSPSolve()
1215: to determine if it has converged.
1217: .keywords: KSP, set, initial guess, nonzero
1219: .seealso: KSPGetInitialGuessNonzero(), KSPSetInitialGuessKnoll(), KSPGetInitialGuessKnoll(), KSPGetErrorIfNotConverged()
1220: @*/
1221: PetscErrorCode KSPSetErrorIfNotConverged(KSP ksp,PetscBool flg)
1222: {
1226: ksp->errorifnotconverged = flg;
1227: return(0);
1228: }
1232: /*@
1233: KSPGetErrorIfNotConverged - Will KSPSolve() generate an error if the solver does not converge?
1235: Not Collective
1237: Input Parameter:
1238: . ksp - iterative context obtained from KSPCreate()
1240: Output Parameter:
1241: . flag - PETSC_TRUE if it will generate an error, else PETSC_FALSE
1243: Level: intermediate
1245: .keywords: KSP, set, initial guess, nonzero
1247: .seealso: KSPSetInitialGuessNonzero(), KSPSetInitialGuessKnoll(), KSPGetInitialGuessKnoll(), KSPSetErrorIfNotConverged()
1248: @*/
1249: PetscErrorCode KSPGetErrorIfNotConverged(KSP ksp,PetscBool *flag)
1250: {
1254: *flag = ksp->errorifnotconverged;
1255: return(0);
1256: }
1260: /*@
1261: KSPSetInitialGuessKnoll - Tells the iterative solver to use PCApply(pc,b,..) to compute the initial guess (The Knoll trick)
1263: Logically Collective on KSP
1265: Input Parameters:
1266: + ksp - iterative context obtained from KSPCreate()
1267: - flg - PETSC_TRUE or PETSC_FALSE
1269: Level: advanced
1272: .keywords: KSP, set, initial guess, nonzero
1274: .seealso: KSPGetInitialGuessKnoll(), KSPSetInitialGuessNonzero(), KSPGetInitialGuessNonzero()
1275: @*/
1276: PetscErrorCode KSPSetInitialGuessKnoll(KSP ksp,PetscBool flg)
1277: {
1281: ksp->guess_knoll = flg;
1282: return(0);
1283: }
1287: /*@
1288: KSPGetInitialGuessKnoll - Determines whether the KSP solver is using the Knoll trick (using PCApply(pc,b,...) to compute
1289: the initial guess
1291: Not Collective
1293: Input Parameter:
1294: . ksp - iterative context obtained from KSPCreate()
1296: Output Parameter:
1297: . flag - PETSC_TRUE if using Knoll trick, else PETSC_FALSE
1299: Level: advanced
1301: .keywords: KSP, set, initial guess, nonzero
1303: .seealso: KSPSetInitialGuessKnoll(), KSPSetInitialGuessNonzero(), KSPGetInitialGuessNonzero()
1304: @*/
1305: PetscErrorCode KSPGetInitialGuessKnoll(KSP ksp,PetscBool *flag)
1306: {
1310: *flag = ksp->guess_knoll;
1311: return(0);
1312: }
1316: /*@
1317: KSPGetComputeSingularValues - Gets the flag indicating whether the extreme singular
1318: values will be calculated via a Lanczos or Arnoldi process as the linear
1319: system is solved.
1321: Not Collective
1323: Input Parameter:
1324: . ksp - iterative context obtained from KSPCreate()
1326: Output Parameter:
1327: . flg - PETSC_TRUE or PETSC_FALSE
1329: Options Database Key:
1330: . -ksp_monitor_singular_value - Activates KSPSetComputeSingularValues()
1332: Notes:
1333: Currently this option is not valid for all iterative methods.
1335: Many users may just want to use the monitoring routine
1336: KSPMonitorSingularValue() (which can be set with option -ksp_monitor_singular_value)
1337: to print the singular values at each iteration of the linear solve.
1339: Level: advanced
1341: .keywords: KSP, set, compute, singular values
1343: .seealso: KSPComputeExtremeSingularValues(), KSPMonitorSingularValue()
1344: @*/
1345: PetscErrorCode KSPGetComputeSingularValues(KSP ksp,PetscBool *flg)
1346: {
1350: *flg = ksp->calc_sings;
1351: return(0);
1352: }
1356: /*@
1357: KSPSetComputeSingularValues - Sets a flag so that the extreme singular
1358: values will be calculated via a Lanczos or Arnoldi process as the linear
1359: system is solved.
1361: Logically Collective on KSP
1363: Input Parameters:
1364: + ksp - iterative context obtained from KSPCreate()
1365: - flg - PETSC_TRUE or PETSC_FALSE
1367: Options Database Key:
1368: . -ksp_monitor_singular_value - Activates KSPSetComputeSingularValues()
1370: Notes:
1371: Currently this option is not valid for all iterative methods.
1373: Many users may just want to use the monitoring routine
1374: KSPMonitorSingularValue() (which can be set with option -ksp_monitor_singular_value)
1375: to print the singular values at each iteration of the linear solve.
1377: Level: advanced
1379: .keywords: KSP, set, compute, singular values
1381: .seealso: KSPComputeExtremeSingularValues(), KSPMonitorSingularValue()
1382: @*/
1383: PetscErrorCode KSPSetComputeSingularValues(KSP ksp,PetscBool flg)
1384: {
1388: ksp->calc_sings = flg;
1389: return(0);
1390: }
1394: /*@
1395: KSPGetComputeEigenvalues - Gets the flag indicating that the extreme eigenvalues
1396: values will be calculated via a Lanczos or Arnoldi process as the linear
1397: system is solved.
1399: Not Collective
1401: Input Parameter:
1402: . ksp - iterative context obtained from KSPCreate()
1404: Output Parameter:
1405: . flg - PETSC_TRUE or PETSC_FALSE
1407: Notes:
1408: Currently this option is not valid for all iterative methods.
1410: Level: advanced
1412: .keywords: KSP, set, compute, eigenvalues
1414: .seealso: KSPComputeEigenvalues(), KSPComputeEigenvaluesExplicitly()
1415: @*/
1416: PetscErrorCode KSPGetComputeEigenvalues(KSP ksp,PetscBool *flg)
1417: {
1421: *flg = ksp->calc_sings;
1422: return(0);
1423: }
1427: /*@
1428: KSPSetComputeEigenvalues - Sets a flag so that the extreme eigenvalues
1429: values will be calculated via a Lanczos or Arnoldi process as the linear
1430: system is solved.
1432: Logically Collective on KSP
1434: Input Parameters:
1435: + ksp - iterative context obtained from KSPCreate()
1436: - flg - PETSC_TRUE or PETSC_FALSE
1438: Notes:
1439: Currently this option is not valid for all iterative methods.
1441: Level: advanced
1443: .keywords: KSP, set, compute, eigenvalues
1445: .seealso: KSPComputeEigenvalues(), KSPComputeEigenvaluesExplicitly()
1446: @*/
1447: PetscErrorCode KSPSetComputeEigenvalues(KSP ksp,PetscBool flg)
1448: {
1452: ksp->calc_sings = flg;
1453: return(0);
1454: }
1458: /*@
1459: KSPGetRhs - Gets the right-hand-side vector for the linear system to
1460: be solved.
1462: Not Collective
1464: Input Parameter:
1465: . ksp - iterative context obtained from KSPCreate()
1467: Output Parameter:
1468: . r - right-hand-side vector
1470: Level: developer
1472: .keywords: KSP, get, right-hand-side, rhs
1474: .seealso: KSPGetSolution(), KSPSolve()
1475: @*/
1476: PetscErrorCode KSPGetRhs(KSP ksp,Vec *r)
1477: {
1481: *r = ksp->vec_rhs;
1482: return(0);
1483: }
1487: /*@
1488: KSPGetSolution - Gets the location of the solution for the
1489: linear system to be solved. Note that this may not be where the solution
1490: is stored during the iterative process; see KSPBuildSolution().
1492: Not Collective
1494: Input Parameters:
1495: . ksp - iterative context obtained from KSPCreate()
1497: Output Parameters:
1498: . v - solution vector
1500: Level: developer
1502: .keywords: KSP, get, solution
1504: .seealso: KSPGetRhs(), KSPBuildSolution(), KSPSolve()
1505: @*/
1506: PetscErrorCode KSPGetSolution(KSP ksp,Vec *v)
1507: {
1511: *v = ksp->vec_sol;
1512: return(0);
1513: }
1517: /*@
1518: KSPSetPC - Sets the preconditioner to be used to calculate the
1519: application of the preconditioner on a vector.
1521: Collective on KSP
1523: Input Parameters:
1524: + ksp - iterative context obtained from KSPCreate()
1525: - pc - the preconditioner object
1527: Notes:
1528: Use KSPGetPC() to retrieve the preconditioner context (for example,
1529: to free it at the end of the computations).
1531: Level: developer
1533: .keywords: KSP, set, precondition, Binv
1535: .seealso: KSPGetPC()
1536: @*/
1537: PetscErrorCode KSPSetPC(KSP ksp,PC pc)
1538: {
1545: PetscObjectReference((PetscObject)pc);
1546: PCDestroy(&ksp->pc);
1547: ksp->pc = pc;
1548: PetscLogObjectParent((PetscObject)ksp,(PetscObject)ksp->pc);
1549: return(0);
1550: }
1554: /*@
1555: KSPGetPC - Returns a pointer to the preconditioner context
1556: set with KSPSetPC().
1558: Not Collective
1560: Input Parameters:
1561: . ksp - iterative context obtained from KSPCreate()
1563: Output Parameter:
1564: . pc - preconditioner context
1566: Level: developer
1568: .keywords: KSP, get, preconditioner, Binv
1570: .seealso: KSPSetPC()
1571: @*/
1572: PetscErrorCode KSPGetPC(KSP ksp,PC *pc)
1573: {
1579: if (!ksp->pc) {
1580: PCCreate(PetscObjectComm((PetscObject)ksp),&ksp->pc);
1581: PetscObjectIncrementTabLevel((PetscObject)ksp->pc,(PetscObject)ksp,0);
1582: PetscLogObjectParent((PetscObject)ksp,(PetscObject)ksp->pc);
1583: }
1584: *pc = ksp->pc;
1585: return(0);
1586: }
1590: /*@
1591: KSPMonitor - runs the user provided monitor routines, if they exist
1593: Collective on KSP
1595: Input Parameters:
1596: + ksp - iterative context obtained from KSPCreate()
1597: . it - iteration number
1598: - rnorm - relative norm of the residual
1600: Notes:
1601: This routine is called by the KSP implementations.
1602: It does not typically need to be called by the user.
1604: Level: developer
1606: .seealso: KSPMonitorSet()
1607: @*/
1608: PetscErrorCode KSPMonitor(KSP ksp,PetscInt it,PetscReal rnorm)
1609: {
1610: PetscInt i, n = ksp->numbermonitors;
1614: for (i=0; i<n; i++) {
1615: (*ksp->monitor[i])(ksp,it,rnorm,ksp->monitorcontext[i]);
1616: }
1617: return(0);
1618: }
1622: /*@C
1623: KSPMonitorSet - Sets an ADDITIONAL function to be called at every iteration to monitor
1624: the residual/error etc.
1626: Logically Collective on KSP
1628: Input Parameters:
1629: + ksp - iterative context obtained from KSPCreate()
1630: . monitor - pointer to function (if this is NULL, it turns off monitoring
1631: . mctx - [optional] context for private data for the
1632: monitor routine (use NULL if no context is desired)
1633: - monitordestroy - [optional] routine that frees monitor context
1634: (may be NULL)
1636: Calling Sequence of monitor:
1637: $ monitor (KSP ksp, int it, PetscReal rnorm, void *mctx)
1639: + ksp - iterative context obtained from KSPCreate()
1640: . it - iteration number
1641: . rnorm - (estimated) 2-norm of (preconditioned) residual
1642: - mctx - optional monitoring context, as set by KSPMonitorSet()
1644: Options Database Keys:
1645: + -ksp_monitor - sets KSPMonitorDefault()
1646: . -ksp_monitor_true_residual - sets KSPMonitorTrueResidualNorm()
1647: . -ksp_monitor_max - sets KSPMonitorTrueResidualMaxNorm()
1648: . -ksp_monitor_lg_residualnorm - sets line graph monitor,
1649: uses KSPMonitorLGResidualNormCreate()
1650: . -ksp_monitor_lg_true_residualnorm - sets line graph monitor,
1651: uses KSPMonitorLGResidualNormCreate()
1652: . -ksp_monitor_singular_value - sets KSPMonitorSingularValue()
1653: - -ksp_monitor_cancel - cancels all monitors that have
1654: been hardwired into a code by
1655: calls to KSPMonitorSet(), but
1656: does not cancel those set via
1657: the options database.
1659: Notes:
1660: The default is to do nothing. To print the residual, or preconditioned
1661: residual if KSPSetNormType(ksp,KSP_NORM_PRECONDITIONED) was called, use
1662: KSPMonitorDefault() as the monitoring routine, with a null monitoring
1663: context.
1665: Several different monitoring routines may be set by calling
1666: KSPMonitorSet() multiple times; all will be called in the
1667: order in which they were set.
1669: Fortran notes: Only a single monitor function can be set for each KSP object
1671: Level: beginner
1673: .keywords: KSP, set, monitor
1675: .seealso: KSPMonitorDefault(), KSPMonitorLGResidualNormCreate(), KSPMonitorCancel()
1676: @*/
1677: PetscErrorCode KSPMonitorSet(KSP ksp,PetscErrorCode (*monitor)(KSP,PetscInt,PetscReal,void*),void *mctx,PetscErrorCode (*monitordestroy)(void**))
1678: {
1679: PetscInt i;
1684: if (ksp->numbermonitors >= MAXKSPMONITORS) SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_OUTOFRANGE,"Too many KSP monitors set");
1685: for (i=0; i<ksp->numbermonitors;i++) {
1686: if (monitor == ksp->monitor[i] && monitordestroy == ksp->monitordestroy[i] && mctx == ksp->monitorcontext[i]) {
1687: if (monitordestroy) {
1688: (*monitordestroy)(&mctx);
1689: }
1690: return(0);
1691: }
1692: }
1693: ksp->monitor[ksp->numbermonitors] = monitor;
1694: ksp->monitordestroy[ksp->numbermonitors] = monitordestroy;
1695: ksp->monitorcontext[ksp->numbermonitors++] = (void*)mctx;
1696: return(0);
1697: }
1701: /*@
1702: KSPMonitorCancel - Clears all monitors for a KSP object.
1704: Logically Collective on KSP
1706: Input Parameters:
1707: . ksp - iterative context obtained from KSPCreate()
1709: Options Database Key:
1710: . -ksp_monitor_cancel - Cancels all monitors that have
1711: been hardwired into a code by calls to KSPMonitorSet(),
1712: but does not cancel those set via the options database.
1714: Level: intermediate
1716: .keywords: KSP, set, monitor
1718: .seealso: KSPMonitorDefault(), KSPMonitorLGResidualNormCreate(), KSPMonitorSet()
1719: @*/
1720: PetscErrorCode KSPMonitorCancel(KSP ksp)
1721: {
1723: PetscInt i;
1727: for (i=0; i<ksp->numbermonitors; i++) {
1728: if (ksp->monitordestroy[i]) {
1729: (*ksp->monitordestroy[i])(&ksp->monitorcontext[i]);
1730: }
1731: }
1732: ksp->numbermonitors = 0;
1733: return(0);
1734: }
1738: /*@C
1739: KSPGetMonitorContext - Gets the monitoring context, as set by
1740: KSPMonitorSet() for the FIRST monitor only.
1742: Not Collective
1744: Input Parameter:
1745: . ksp - iterative context obtained from KSPCreate()
1747: Output Parameter:
1748: . ctx - monitoring context
1750: Level: intermediate
1752: .keywords: KSP, get, monitor, context
1754: .seealso: KSPMonitorDefault(), KSPMonitorLGResidualNormCreate()
1755: @*/
1756: PetscErrorCode KSPGetMonitorContext(KSP ksp,void **ctx)
1757: {
1760: *ctx = (ksp->monitorcontext[0]);
1761: return(0);
1762: }
1766: /*@
1767: KSPSetResidualHistory - Sets the array used to hold the residual history.
1768: If set, this array will contain the residual norms computed at each
1769: iteration of the solver.
1771: Not Collective
1773: Input Parameters:
1774: + ksp - iterative context obtained from KSPCreate()
1775: . a - array to hold history
1776: . na - size of a
1777: - reset - PETSC_TRUE indicates the history counter is reset to zero
1778: for each new linear solve
1780: Level: advanced
1782: Notes: The array is NOT freed by PETSc so the user needs to keep track of
1783: it and destroy once the KSP object is destroyed.
1785: If 'a' is NULL then space is allocated for the history. If 'na' PETSC_DECIDE or PETSC_DEFAULT then a
1786: default array of length 10000 is allocated.
1788: .keywords: KSP, set, residual, history, norm
1790: .seealso: KSPGetResidualHistory()
1792: @*/
1793: PetscErrorCode KSPSetResidualHistory(KSP ksp,PetscReal a[],PetscInt na,PetscBool reset)
1794: {
1800: PetscFree(ksp->res_hist_alloc);
1801: if (na != PETSC_DECIDE && na != PETSC_DEFAULT && a) {
1802: ksp->res_hist = a;
1803: ksp->res_hist_max = na;
1804: } else {
1805: if (na != PETSC_DECIDE && na != PETSC_DEFAULT) ksp->res_hist_max = na;
1806: else ksp->res_hist_max = 10000; /* like default ksp->max_it */
1807: PetscCalloc1(ksp->res_hist_max,&ksp->res_hist_alloc);
1809: ksp->res_hist = ksp->res_hist_alloc;
1810: }
1811: ksp->res_hist_len = 0;
1812: ksp->res_hist_reset = reset;
1813: return(0);
1814: }
1818: /*@C
1819: KSPGetResidualHistory - Gets the array used to hold the residual history
1820: and the number of residuals it contains.
1822: Not Collective
1824: Input Parameter:
1825: . ksp - iterative context obtained from KSPCreate()
1827: Output Parameters:
1828: + a - pointer to array to hold history (or NULL)
1829: - na - number of used entries in a (or NULL)
1831: Level: advanced
1833: Notes:
1834: Can only be called after a KSPSetResidualHistory() otherwise a and na are set to zero
1836: The Fortran version of this routine has a calling sequence
1837: $ call KSPGetResidualHistory(KSP ksp, integer na, integer ierr)
1838: note that you have passed a Fortran array into KSPSetResidualHistory() and you need
1839: to access the residual values from this Fortran array you provided. Only the na (number of
1840: residual norms currently held) is set.
1842: .keywords: KSP, get, residual, history, norm
1844: .seealso: KSPGetResidualHistory()
1846: @*/
1847: PetscErrorCode KSPGetResidualHistory(KSP ksp,PetscReal *a[],PetscInt *na)
1848: {
1851: if (a) *a = ksp->res_hist;
1852: if (na) *na = ksp->res_hist_len;
1853: return(0);
1854: }
1858: /*@C
1859: KSPSetConvergenceTest - Sets the function to be used to determine
1860: convergence.
1862: Logically Collective on KSP
1864: Input Parameters:
1865: + ksp - iterative context obtained from KSPCreate()
1866: . converge - pointer to int function
1867: . cctx - context for private data for the convergence routine (may be null)
1868: - destroy - a routine for destroying the context (may be null)
1870: Calling sequence of converge:
1871: $ converge (KSP ksp, int it, PetscReal rnorm, KSPConvergedReason *reason,void *mctx)
1873: + ksp - iterative context obtained from KSPCreate()
1874: . it - iteration number
1875: . rnorm - (estimated) 2-norm of (preconditioned) residual
1876: . reason - the reason why it has converged or diverged
1877: - cctx - optional convergence context, as set by KSPSetConvergenceTest()
1880: Notes:
1881: Must be called after the KSP type has been set so put this after
1882: a call to KSPSetType(), or KSPSetFromOptions().
1884: The default convergence test, KSPConvergedDefault(), aborts if the
1885: residual grows to more than 10000 times the initial residual.
1887: The default is a combination of relative and absolute tolerances.
1888: The residual value that is tested may be an approximation; routines
1889: that need exact values should compute them.
1891: In the default PETSc convergence test, the precise values of reason
1892: are macros such as KSP_CONVERGED_RTOL, which are defined in petscksp.h.
1894: Level: advanced
1896: .keywords: KSP, set, convergence, test, context
1898: .seealso: KSPConvergedDefault(), KSPGetConvergenceContext(), KSPSetTolerances()
1899: @*/
1900: PetscErrorCode KSPSetConvergenceTest(KSP ksp,PetscErrorCode (*converge)(KSP,PetscInt,PetscReal,KSPConvergedReason*,void*),void *cctx,PetscErrorCode (*destroy)(void*))
1901: {
1906: if (ksp->convergeddestroy) {
1907: (*ksp->convergeddestroy)(ksp->cnvP);
1908: }
1909: ksp->converged = converge;
1910: ksp->convergeddestroy = destroy;
1911: ksp->cnvP = (void*)cctx;
1912: return(0);
1913: }
1917: /*@C
1918: KSPGetConvergenceContext - Gets the convergence context set with
1919: KSPSetConvergenceTest().
1921: Not Collective
1923: Input Parameter:
1924: . ksp - iterative context obtained from KSPCreate()
1926: Output Parameter:
1927: . ctx - monitoring context
1929: Level: advanced
1931: .keywords: KSP, get, convergence, test, context
1933: .seealso: KSPConvergedDefault(), KSPSetConvergenceTest()
1934: @*/
1935: PetscErrorCode KSPGetConvergenceContext(KSP ksp,void **ctx)
1936: {
1939: *ctx = ksp->cnvP;
1940: return(0);
1941: }
1945: /*@C
1946: KSPBuildSolution - Builds the approximate solution in a vector provided.
1947: This routine is NOT commonly needed (see KSPSolve()).
1949: Collective on KSP
1951: Input Parameter:
1952: . ctx - iterative context obtained from KSPCreate()
1954: Output Parameter:
1955: Provide exactly one of
1956: + v - location to stash solution.
1957: - V - the solution is returned in this location. This vector is created
1958: internally. This vector should NOT be destroyed by the user with
1959: VecDestroy().
1961: Notes:
1962: This routine can be used in one of two ways
1963: .vb
1964: KSPBuildSolution(ksp,NULL,&V);
1965: or
1966: KSPBuildSolution(ksp,v,NULL); or KSPBuildSolution(ksp,v,&v);
1967: .ve
1968: In the first case an internal vector is allocated to store the solution
1969: (the user cannot destroy this vector). In the second case the solution
1970: is generated in the vector that the user provides. Note that for certain
1971: methods, such as KSPCG, the second case requires a copy of the solution,
1972: while in the first case the call is essentially free since it simply
1973: returns the vector where the solution already is stored. For some methods
1974: like GMRES this is a reasonably expensive operation and should only be
1975: used in truly needed.
1977: Level: advanced
1979: .keywords: KSP, build, solution
1981: .seealso: KSPGetSolution(), KSPBuildResidual()
1982: @*/
1983: PetscErrorCode KSPBuildSolution(KSP ksp,Vec v,Vec *V)
1984: {
1989: if (!V && !v) SETERRQ(PetscObjectComm((PetscObject)ksp),PETSC_ERR_ARG_WRONG,"Must provide either v or V");
1990: if (!V) V = &v;
1991: (*ksp->ops->buildsolution)(ksp,v,V);
1992: return(0);
1993: }
1997: /*@C
1998: KSPBuildResidual - Builds the residual in a vector provided.
2000: Collective on KSP
2002: Input Parameter:
2003: . ksp - iterative context obtained from KSPCreate()
2005: Output Parameters:
2006: + v - optional location to stash residual. If v is not provided,
2007: then a location is generated.
2008: . t - work vector. If not provided then one is generated.
2009: - V - the residual
2011: Notes:
2012: Regardless of whether or not v is provided, the residual is
2013: returned in V.
2015: Level: advanced
2017: .keywords: KSP, build, residual
2019: .seealso: KSPBuildSolution()
2020: @*/
2021: PetscErrorCode KSPBuildResidual(KSP ksp,Vec t,Vec v,Vec *V)
2022: {
2024: PetscBool flag = PETSC_FALSE;
2025: Vec w = v,tt = t;
2029: if (!w) {
2030: VecDuplicate(ksp->vec_rhs,&w);
2031: PetscLogObjectParent((PetscObject)ksp,(PetscObject)w);
2032: }
2033: if (!tt) {
2034: VecDuplicate(ksp->vec_sol,&tt); flag = PETSC_TRUE;
2035: PetscLogObjectParent((PetscObject)ksp,(PetscObject)tt);
2036: }
2037: (*ksp->ops->buildresidual)(ksp,tt,w,V);
2038: if (flag) {VecDestroy(&tt);}
2039: return(0);
2040: }
2044: /*@
2045: KSPSetDiagonalScale - Tells KSP to symmetrically diagonally scale the system
2046: before solving. This actually CHANGES the matrix (and right hand side).
2048: Logically Collective on KSP
2050: Input Parameter:
2051: + ksp - the KSP context
2052: - scale - PETSC_TRUE or PETSC_FALSE
2054: Options Database Key:
2055: + -ksp_diagonal_scale -
2056: - -ksp_diagonal_scale_fix - scale the matrix back AFTER the solve
2059: Notes: Scales the matrix by D^(-1/2) A D^(-1/2) [D^(1/2) x ] = D^(-1/2) b
2060: where D_{ii} is 1/abs(A_{ii}) unless A_{ii} is zero and then it is 1.
2062: BE CAREFUL with this routine: it actually scales the matrix and right
2063: hand side that define the system. After the system is solved the matrix
2064: and right hand side remain scaled unless you use KSPSetDiagonalScaleFix()
2066: This should NOT be used within the SNES solves if you are using a line
2067: search.
2069: If you use this with the PCType Eisenstat preconditioner than you can
2070: use the PCEisenstatSetNoDiagonalScaling() option, or -pc_eisenstat_no_diagonal_scaling
2071: to save some unneeded, redundant flops.
2073: Level: intermediate
2075: .keywords: KSP, set, options, prefix, database
2077: .seealso: KSPGetDiagonalScale(), KSPSetDiagonalScaleFix()
2078: @*/
2079: PetscErrorCode KSPSetDiagonalScale(KSP ksp,PetscBool scale)
2080: {
2084: ksp->dscale = scale;
2085: return(0);
2086: }
2090: /*@
2091: KSPGetDiagonalScale - Checks if KSP solver scales the matrix and
2092: right hand side
2094: Not Collective
2096: Input Parameter:
2097: . ksp - the KSP context
2099: Output Parameter:
2100: . scale - PETSC_TRUE or PETSC_FALSE
2102: Notes:
2103: BE CAREFUL with this routine: it actually scales the matrix and right
2104: hand side that define the system. After the system is solved the matrix
2105: and right hand side remain scaled unless you use KSPSetDiagonalScaleFix()
2107: Level: intermediate
2109: .keywords: KSP, set, options, prefix, database
2111: .seealso: KSPSetDiagonalScale(), KSPSetDiagonalScaleFix()
2112: @*/
2113: PetscErrorCode KSPGetDiagonalScale(KSP ksp,PetscBool *scale)
2114: {
2118: *scale = ksp->dscale;
2119: return(0);
2120: }
2124: /*@
2125: KSPSetDiagonalScaleFix - Tells KSP to diagonally scale the system
2126: back after solving.
2128: Logically Collective on KSP
2130: Input Parameter:
2131: + ksp - the KSP context
2132: - fix - PETSC_TRUE to scale back after the system solve, PETSC_FALSE to not
2133: rescale (default)
2135: Notes:
2136: Must be called after KSPSetDiagonalScale()
2138: Using this will slow things down, because it rescales the matrix before and
2139: after each linear solve. This is intended mainly for testing to allow one
2140: to easily get back the original system to make sure the solution computed is
2141: accurate enough.
2143: Level: intermediate
2145: .keywords: KSP, set, options, prefix, database
2147: .seealso: KSPGetDiagonalScale(), KSPSetDiagonalScale(), KSPGetDiagonalScaleFix()
2148: @*/
2149: PetscErrorCode KSPSetDiagonalScaleFix(KSP ksp,PetscBool fix)
2150: {
2154: ksp->dscalefix = fix;
2155: return(0);
2156: }
2160: /*@
2161: KSPGetDiagonalScaleFix - Determines if KSP diagonally scales the system
2162: back after solving.
2164: Not Collective
2166: Input Parameter:
2167: . ksp - the KSP context
2169: Output Parameter:
2170: . fix - PETSC_TRUE to scale back after the system solve, PETSC_FALSE to not
2171: rescale (default)
2173: Notes:
2174: Must be called after KSPSetDiagonalScale()
2176: If PETSC_TRUE will slow things down, because it rescales the matrix before and
2177: after each linear solve. This is intended mainly for testing to allow one
2178: to easily get back the original system to make sure the solution computed is
2179: accurate enough.
2181: Level: intermediate
2183: .keywords: KSP, set, options, prefix, database
2185: .seealso: KSPGetDiagonalScale(), KSPSetDiagonalScale(), KSPSetDiagonalScaleFix()
2186: @*/
2187: PetscErrorCode KSPGetDiagonalScaleFix(KSP ksp,PetscBool *fix)
2188: {
2192: *fix = ksp->dscalefix;
2193: return(0);
2194: }
2198: /*@C
2199: KSPSetComputeOperators - set routine to compute the linear operators
2201: Logically Collective
2203: Input Arguments:
2204: + ksp - the KSP context
2205: . func - function to compute the operators
2206: - ctx - optional context
2208: Calling sequence of func:
2209: $ func(KSP ksp,Mat A,Mat B,void *ctx)
2211: + ksp - the KSP context
2212: . A - the linear operator
2213: . B - preconditioning matrix
2214: - ctx - optional user-provided context
2216: Notes: The user provided func() will be called automatically at the very next call to KSPSolve(). It will not be called at future KSPSolve() calls
2217: unless either KSPSetComputeOperators() or KSPSetOperators() is called before that KSPSolve() is called.
2219: To reuse the same preconditioner for the next KSPSolve() and not compute a new one based on the most recently computed matrix call KSPSetReusePreconditioner()
2221: Level: beginner
2223: .seealso: KSPSetOperators(), KSPSetComputeRHS(), DMKSPSetComputeOperators(), KSPSetComputeInitialGuess()
2224: @*/
2225: PetscErrorCode KSPSetComputeOperators(KSP ksp,PetscErrorCode (*func)(KSP,Mat,Mat,void*),void *ctx)
2226: {
2228: DM dm;
2232: KSPGetDM(ksp,&dm);
2233: DMKSPSetComputeOperators(dm,func,ctx);
2234: if (ksp->setupstage == KSP_SETUP_NEWRHS) ksp->setupstage = KSP_SETUP_NEWMATRIX;
2235: return(0);
2236: }
2240: /*@C
2241: KSPSetComputeRHS - set routine to compute the right hand side of the linear system
2243: Logically Collective
2245: Input Arguments:
2246: + ksp - the KSP context
2247: . func - function to compute the right hand side
2248: - ctx - optional context
2250: Calling sequence of func:
2251: $ func(KSP ksp,Vec b,void *ctx)
2253: + ksp - the KSP context
2254: . b - right hand side of linear system
2255: - ctx - optional user-provided context
2257: Notes: The routine you provide will be called EACH you call KSPSolve() to prepare the new right hand side for that solve
2259: Level: beginner
2261: .seealso: KSPSolve(), DMKSPSetComputeRHS(), KSPSetComputeOperators()
2262: @*/
2263: PetscErrorCode KSPSetComputeRHS(KSP ksp,PetscErrorCode (*func)(KSP,Vec,void*),void *ctx)
2264: {
2266: DM dm;
2270: KSPGetDM(ksp,&dm);
2271: DMKSPSetComputeRHS(dm,func,ctx);
2272: return(0);
2273: }
2277: /*@C
2278: KSPSetComputeInitialGuess - set routine to compute the initial guess of the linear system
2280: Logically Collective
2282: Input Arguments:
2283: + ksp - the KSP context
2284: . func - function to compute the initial guess
2285: - ctx - optional context
2287: Calling sequence of func:
2288: $ func(KSP ksp,Vec x,void *ctx)
2290: + ksp - the KSP context
2291: . x - solution vector
2292: - ctx - optional user-provided context
2294: Level: beginner
2296: .seealso: KSPSolve(), KSPSetComputeRHS(), KSPSetComputeOperators(), DMKSPSetComputeInitialGuess()
2297: @*/
2298: PetscErrorCode KSPSetComputeInitialGuess(KSP ksp,PetscErrorCode (*func)(KSP,Vec,void*),void *ctx)
2299: {
2301: DM dm;
2305: KSPGetDM(ksp,&dm);
2306: DMKSPSetComputeInitialGuess(dm,func,ctx);
2307: return(0);
2308: }