Actual source code: ts.c
1: #include <petsc/private/tsimpl.h>
2: #include <petscdmshell.h>
3: #include <petscdmda.h>
4: #include <petscviewer.h>
5: #include <petscdraw.h>
6: #include <petscconvest.h>
8: #define SkipSmallValue(a,b,tol) if (PetscAbsScalar(a)< tol || PetscAbsScalar(b)< tol) continue;
10: /* Logging support */
11: PetscClassId TS_CLASSID, DMTS_CLASSID;
12: PetscLogEvent TS_Step, TS_PseudoComputeTimeStep, TS_FunctionEval, TS_JacobianEval;
14: const char *const TSExactFinalTimeOptions[] = {"UNSPECIFIED","STEPOVER","INTERPOLATE","MATCHSTEP","TSExactFinalTimeOption","TS_EXACTFINALTIME_",NULL};
16: static PetscErrorCode TSAdaptSetDefaultType(TSAdapt adapt,TSAdaptType default_type)
17: {
20: if (!((PetscObject)adapt)->type_name) {
21: TSAdaptSetType(adapt,default_type);
22: }
23: return 0;
24: }
26: /*@
27: TSSetFromOptions - Sets various TS parameters from user options.
29: Collective on TS
31: Input Parameter:
32: . ts - the TS context obtained from TSCreate()
34: Options Database Keys:
35: + -ts_type <type> - TSEULER, TSBEULER, TSSUNDIALS, TSPSEUDO, TSCN, TSRK, TSTHETA, TSALPHA, TSGLLE, TSSSP, TSGLEE, TSBSYMP, TSIRK
36: . -ts_save_trajectory - checkpoint the solution at each time-step
37: . -ts_max_time <time> - maximum time to compute to
38: . -ts_max_steps <steps> - maximum number of time-steps to take
39: . -ts_init_time <time> - initial time to start computation
40: . -ts_final_time <time> - final time to compute to (deprecated: use -ts_max_time)
41: . -ts_dt <dt> - initial time step
42: . -ts_exact_final_time <stepover,interpolate,matchstep> - whether to stop at the exact given final time and how to compute the solution at that time
43: . -ts_max_snes_failures <maxfailures> - Maximum number of nonlinear solve failures allowed
44: . -ts_max_reject <maxrejects> - Maximum number of step rejections before step fails
45: . -ts_error_if_step_fails <true,false> - Error if no step succeeds
46: . -ts_rtol <rtol> - relative tolerance for local truncation error
47: . -ts_atol <atol> - Absolute tolerance for local truncation error
48: . -ts_rhs_jacobian_test_mult -mat_shell_test_mult_view - test the Jacobian at each iteration against finite difference with RHS function
49: . -ts_rhs_jacobian_test_mult_transpose -mat_shell_test_mult_transpose_view - test the Jacobian at each iteration against finite difference with RHS function
50: . -ts_adjoint_solve <yes,no> - After solving the ODE/DAE solve the adjoint problem (requires -ts_save_trajectory)
51: . -ts_fd_color - Use finite differences with coloring to compute IJacobian
52: . -ts_monitor - print information at each timestep
53: . -ts_monitor_cancel - Cancel all monitors
54: . -ts_monitor_lg_solution - Monitor solution graphically
55: . -ts_monitor_lg_error - Monitor error graphically
56: . -ts_monitor_error - Monitors norm of error
57: . -ts_monitor_lg_timestep - Monitor timestep size graphically
58: . -ts_monitor_lg_timestep_log - Monitor log timestep size graphically
59: . -ts_monitor_lg_snes_iterations - Monitor number nonlinear iterations for each timestep graphically
60: . -ts_monitor_lg_ksp_iterations - Monitor number nonlinear iterations for each timestep graphically
61: . -ts_monitor_sp_eig - Monitor eigenvalues of linearized operator graphically
62: . -ts_monitor_draw_solution - Monitor solution graphically
63: . -ts_monitor_draw_solution_phase <xleft,yleft,xright,yright> - Monitor solution graphically with phase diagram, requires problem with exactly 2 degrees of freedom
64: . -ts_monitor_draw_error - Monitor error graphically, requires use to have provided TSSetSolutionFunction()
65: . -ts_monitor_solution [ascii binary draw][:filename][:viewerformat] - monitors the solution at each timestep
66: . -ts_monitor_solution_vtk <filename.vts,filename.vtu> - Save each time step to a binary file, use filename-%%03D.vts (filename-%%03D.vtu)
67: - -ts_monitor_envelope - determine maximum and minimum value of each component of the solution over the solution time
69: Notes:
70: See SNESSetFromOptions() and KSPSetFromOptions() for how to control the nonlinear and linear solves used by the time-stepper.
72: Certain SNES options get reset for each new nonlinear solver, for example -snes_lag_jacobian <its> and -snes_lag_preconditioner <its>, in order
73: to retain them over the multiple nonlinear solves that TS uses you mush also provide -snes_lag_jacobian_persists true and
74: -snes_lag_preconditioner_persists true
76: Developer Note:
77: We should unify all the -ts_monitor options in the way that -xxx_view has been unified
79: Level: beginner
81: .seealso: TSGetType()
82: @*/
83: PetscErrorCode TSSetFromOptions(TS ts)
84: {
85: PetscBool opt,flg,tflg;
86: PetscErrorCode ierr;
87: char monfilename[PETSC_MAX_PATH_LEN];
88: PetscReal time_step;
89: TSExactFinalTimeOption eftopt;
90: char dir[16];
91: TSIFunction ifun;
92: const char *defaultType;
93: char typeName[256];
97: TSRegisterAll();
98: TSGetIFunction(ts,NULL,&ifun,NULL);
100: PetscObjectOptionsBegin((PetscObject)ts);
101: if (((PetscObject)ts)->type_name) defaultType = ((PetscObject)ts)->type_name;
102: else defaultType = ifun ? TSBEULER : TSEULER;
103: PetscOptionsFList("-ts_type","TS method","TSSetType",TSList,defaultType,typeName,256,&opt);
104: if (opt) {
105: TSSetType(ts,typeName);
106: } else {
107: TSSetType(ts,defaultType);
108: }
110: /* Handle generic TS options */
111: PetscOptionsDeprecated("-ts_final_time","-ts_max_time","3.10",NULL);
112: PetscOptionsReal("-ts_max_time","Maximum time to run to","TSSetMaxTime",ts->max_time,&ts->max_time,NULL);
113: PetscOptionsInt("-ts_max_steps","Maximum number of time steps","TSSetMaxSteps",ts->max_steps,&ts->max_steps,NULL);
114: PetscOptionsReal("-ts_init_time","Initial time","TSSetTime",ts->ptime,&ts->ptime,NULL);
115: PetscOptionsReal("-ts_dt","Initial time step","TSSetTimeStep",ts->time_step,&time_step,&flg);
116: if (flg) TSSetTimeStep(ts,time_step);
117: PetscOptionsEnum("-ts_exact_final_time","Option for handling of final time step","TSSetExactFinalTime",TSExactFinalTimeOptions,(PetscEnum)ts->exact_final_time,(PetscEnum*)&eftopt,&flg);
118: if (flg) TSSetExactFinalTime(ts,eftopt);
119: PetscOptionsInt("-ts_max_snes_failures","Maximum number of nonlinear solve failures","TSSetMaxSNESFailures",ts->max_snes_failures,&ts->max_snes_failures,NULL);
120: PetscOptionsInt("-ts_max_reject","Maximum number of step rejections before step fails","TSSetMaxStepRejections",ts->max_reject,&ts->max_reject,NULL);
121: PetscOptionsBool("-ts_error_if_step_fails","Error if no step succeeds","TSSetErrorIfStepFails",ts->errorifstepfailed,&ts->errorifstepfailed,NULL);
122: PetscOptionsReal("-ts_rtol","Relative tolerance for local truncation error","TSSetTolerances",ts->rtol,&ts->rtol,NULL);
123: PetscOptionsReal("-ts_atol","Absolute tolerance for local truncation error","TSSetTolerances",ts->atol,&ts->atol,NULL);
125: PetscOptionsBool("-ts_rhs_jacobian_test_mult","Test the RHS Jacobian for consistency with RHS at each solve ","None",ts->testjacobian,&ts->testjacobian,NULL);
126: PetscOptionsBool("-ts_rhs_jacobian_test_mult_transpose","Test the RHS Jacobian transpose for consistency with RHS at each solve ","None",ts->testjacobiantranspose,&ts->testjacobiantranspose,NULL);
127: PetscOptionsBool("-ts_use_splitrhsfunction","Use the split RHS function for multirate solvers ","TSSetUseSplitRHSFunction",ts->use_splitrhsfunction,&ts->use_splitrhsfunction,NULL);
128: #if defined(PETSC_HAVE_SAWS)
129: {
130: PetscBool set;
131: flg = PETSC_FALSE;
132: PetscOptionsBool("-ts_saws_block","Block for SAWs memory snooper at end of TSSolve","PetscObjectSAWsBlock",((PetscObject)ts)->amspublishblock,&flg,&set);
133: if (set) {
134: PetscObjectSAWsSetBlock((PetscObject)ts,flg);
135: }
136: }
137: #endif
139: /* Monitor options */
140: PetscOptionsInt("-ts_monitor_frequency", "Number of time steps between monitor output", "TSMonitorSetFrequency", ts->monitorFrequency, &ts->monitorFrequency, NULL);
141: TSMonitorSetFromOptions(ts,"-ts_monitor","Monitor time and timestep size","TSMonitorDefault",TSMonitorDefault,NULL);
142: TSMonitorSetFromOptions(ts,"-ts_monitor_extreme","Monitor extreme values of the solution","TSMonitorExtreme",TSMonitorExtreme,NULL);
143: TSMonitorSetFromOptions(ts,"-ts_monitor_solution","View the solution at each timestep","TSMonitorSolution",TSMonitorSolution,NULL);
144: TSMonitorSetFromOptions(ts,"-ts_dmswarm_monitor_moments","Monitor moments of particle distribution","TSDMSwarmMonitorMoments",TSDMSwarmMonitorMoments,NULL);
146: PetscOptionsString("-ts_monitor_python","Use Python function","TSMonitorSet",NULL,monfilename,sizeof(monfilename),&flg);
147: if (flg) PetscPythonMonitorSet((PetscObject)ts,monfilename);
149: PetscOptionsName("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",&opt);
150: if (opt) {
151: PetscInt howoften = 1;
152: DM dm;
153: PetscBool net;
155: PetscOptionsInt("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",howoften,&howoften,NULL);
156: TSGetDM(ts,&dm);
157: PetscObjectTypeCompare((PetscObject)dm,DMNETWORK,&net);
158: if (net) {
159: TSMonitorLGCtxNetwork ctx;
160: TSMonitorLGCtxNetworkCreate(ts,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);
161: TSMonitorSet(ts,TSMonitorLGCtxNetworkSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxNetworkDestroy);
162: PetscOptionsBool("-ts_monitor_lg_solution_semilogy","Plot the solution with a semi-log axis","",ctx->semilogy,&ctx->semilogy,NULL);
163: } else {
164: TSMonitorLGCtx ctx;
165: TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
166: TSMonitorSet(ts,TSMonitorLGSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
167: }
168: }
170: PetscOptionsName("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",&opt);
171: if (opt) {
172: TSMonitorLGCtx ctx;
173: PetscInt howoften = 1;
175: PetscOptionsInt("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",howoften,&howoften,NULL);
176: TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
177: TSMonitorSet(ts,TSMonitorLGError,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
178: }
179: TSMonitorSetFromOptions(ts,"-ts_monitor_error","View the error at each timestep","TSMonitorError",TSMonitorError,NULL);
181: PetscOptionsName("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",&opt);
182: if (opt) {
183: TSMonitorLGCtx ctx;
184: PetscInt howoften = 1;
186: PetscOptionsInt("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",howoften,&howoften,NULL);
187: TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
188: TSMonitorSet(ts,TSMonitorLGTimeStep,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
189: }
190: PetscOptionsName("-ts_monitor_lg_timestep_log","Monitor log timestep size graphically","TSMonitorLGTimeStep",&opt);
191: if (opt) {
192: TSMonitorLGCtx ctx;
193: PetscInt howoften = 1;
195: PetscOptionsInt("-ts_monitor_lg_timestep_log","Monitor log timestep size graphically","TSMonitorLGTimeStep",howoften,&howoften,NULL);
196: TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
197: TSMonitorSet(ts,TSMonitorLGTimeStep,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
198: ctx->semilogy = PETSC_TRUE;
199: }
201: PetscOptionsName("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",&opt);
202: if (opt) {
203: TSMonitorLGCtx ctx;
204: PetscInt howoften = 1;
206: PetscOptionsInt("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",howoften,&howoften,NULL);
207: TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
208: TSMonitorSet(ts,TSMonitorLGSNESIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
209: }
210: PetscOptionsName("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",&opt);
211: if (opt) {
212: TSMonitorLGCtx ctx;
213: PetscInt howoften = 1;
215: PetscOptionsInt("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",howoften,&howoften,NULL);
216: TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx);
217: TSMonitorSet(ts,TSMonitorLGKSPIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);
218: }
219: PetscOptionsName("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",&opt);
220: if (opt) {
221: TSMonitorSPEigCtx ctx;
222: PetscInt howoften = 1;
224: PetscOptionsInt("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",howoften,&howoften,NULL);
225: TSMonitorSPEigCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);
226: TSMonitorSet(ts,TSMonitorSPEig,ctx,(PetscErrorCode (*)(void**))TSMonitorSPEigCtxDestroy);
227: }
228: PetscOptionsName("-ts_monitor_sp_swarm","Display particle phase from the DMSwarm","TSMonitorSPSwarm",&opt);
229: if (opt) {
230: TSMonitorSPCtx ctx;
231: PetscInt howoften = 1, retain = 0;
232: PetscBool phase = PETSC_TRUE;
234: PetscOptionsInt("-ts_monitor_sp_swarm","Display particles phase from the DMSwarm", "TSMonitorSPSwarm", howoften, &howoften, NULL);
235: PetscOptionsInt("-ts_monitor_sp_swarm_retain", "Retain n points plotted to show trajectory, -1 for all points", "TSMonitorSPSwarm", retain, &retain, NULL);
236: PetscOptionsBool("-ts_monitor_sp_swarm_phase", "Plot in phase space rather than coordinate space", "TSMonitorSPSwarm", phase, &phase, NULL);
237: TSMonitorSPCtxCreate(PetscObjectComm((PetscObject) ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, retain, phase, &ctx);
238: TSMonitorSet(ts, TSMonitorSPSwarmSolution, ctx, (PetscErrorCode (*)(void**))TSMonitorSPCtxDestroy);
239: }
240: opt = PETSC_FALSE;
241: PetscOptionsName("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",&opt);
242: if (opt) {
243: TSMonitorDrawCtx ctx;
244: PetscInt howoften = 1;
246: PetscOptionsInt("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",howoften,&howoften,NULL);
247: TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Computed Solution",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);
248: TSMonitorSet(ts,TSMonitorDrawSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);
249: }
250: opt = PETSC_FALSE;
251: PetscOptionsName("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",&opt);
252: if (opt) {
253: TSMonitorDrawCtx ctx;
254: PetscReal bounds[4];
255: PetscInt n = 4;
256: PetscDraw draw;
257: PetscDrawAxis axis;
259: PetscOptionsRealArray("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",bounds,&n,NULL);
261: TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,300,300,1,&ctx);
262: PetscViewerDrawGetDraw(ctx->viewer,0,&draw);
263: PetscViewerDrawGetDrawAxis(ctx->viewer,0,&axis);
264: PetscDrawAxisSetLimits(axis,bounds[0],bounds[2],bounds[1],bounds[3]);
265: PetscDrawAxisSetLabels(axis,"Phase Diagram","Variable 1","Variable 2");
266: TSMonitorSet(ts,TSMonitorDrawSolutionPhase,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);
267: }
268: opt = PETSC_FALSE;
269: PetscOptionsName("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",&opt);
270: if (opt) {
271: TSMonitorDrawCtx ctx;
272: PetscInt howoften = 1;
274: PetscOptionsInt("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",howoften,&howoften,NULL);
275: TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Error",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);
276: TSMonitorSet(ts,TSMonitorDrawError,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);
277: }
278: opt = PETSC_FALSE;
279: PetscOptionsName("-ts_monitor_draw_solution_function","Monitor solution provided by TSMonitorSetSolutionFunction() graphically","TSMonitorDrawSolutionFunction",&opt);
280: if (opt) {
281: TSMonitorDrawCtx ctx;
282: PetscInt howoften = 1;
284: PetscOptionsInt("-ts_monitor_draw_solution_function","Monitor solution provided by TSMonitorSetSolutionFunction() graphically","TSMonitorDrawSolutionFunction",howoften,&howoften,NULL);
285: TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Solution provided by user function",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);
286: TSMonitorSet(ts,TSMonitorDrawSolutionFunction,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);
287: }
289: opt = PETSC_FALSE;
290: PetscOptionsString("-ts_monitor_solution_vtk","Save each time step to a binary file, use filename-%%03D.vts","TSMonitorSolutionVTK",NULL,monfilename,sizeof(monfilename),&flg);
291: if (flg) {
292: const char *ptr,*ptr2;
293: char *filetemplate;
295: /* Do some cursory validation of the input. */
296: PetscStrstr(monfilename,"%",(char**)&ptr);
298: for (ptr++; ptr && *ptr; ptr++) {
299: PetscStrchr("DdiouxX",*ptr,(char**)&ptr2);
301: if (ptr2) break;
302: }
303: PetscStrallocpy(monfilename,&filetemplate);
304: TSMonitorSet(ts,TSMonitorSolutionVTK,filetemplate,(PetscErrorCode (*)(void**))TSMonitorSolutionVTKDestroy);
305: }
307: PetscOptionsString("-ts_monitor_dmda_ray","Display a ray of the solution","None","y=0",dir,sizeof(dir),&flg);
308: if (flg) {
309: TSMonitorDMDARayCtx *rayctx;
310: int ray = 0;
311: DMDirection ddir;
312: DM da;
313: PetscMPIInt rank;
316: if (dir[0] == 'x') ddir = DM_X;
317: else if (dir[0] == 'y') ddir = DM_Y;
318: else SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Unknown ray %s",dir);
319: sscanf(dir+2,"%d",&ray);
321: PetscInfo(((PetscObject)ts),"Displaying DMDA ray %c = %d\n",dir[0],ray);
322: PetscNew(&rayctx);
323: TSGetDM(ts,&da);
324: DMDAGetRay(da,ddir,ray,&rayctx->ray,&rayctx->scatter);
325: MPI_Comm_rank(PetscObjectComm((PetscObject)ts),&rank);
326: if (rank == 0) {
327: PetscViewerDrawOpen(PETSC_COMM_SELF,NULL,NULL,0,0,600,300,&rayctx->viewer);
328: }
329: rayctx->lgctx = NULL;
330: TSMonitorSet(ts,TSMonitorDMDARay,rayctx,TSMonitorDMDARayDestroy);
331: }
332: PetscOptionsString("-ts_monitor_lg_dmda_ray","Display a ray of the solution","None","x=0",dir,sizeof(dir),&flg);
333: if (flg) {
334: TSMonitorDMDARayCtx *rayctx;
335: int ray = 0;
336: DMDirection ddir;
337: DM da;
338: PetscInt howoften = 1;
341: if (dir[0] == 'x') ddir = DM_X;
342: else if (dir[0] == 'y') ddir = DM_Y;
343: else SETERRQ(PetscObjectComm((PetscObject) ts), PETSC_ERR_ARG_WRONG, "Unknown ray direction %s", dir);
344: sscanf(dir+2, "%d", &ray);
346: PetscInfo(((PetscObject) ts),"Displaying LG DMDA ray %c = %d\n", dir[0], ray);
347: PetscNew(&rayctx);
348: TSGetDM(ts, &da);
349: DMDAGetRay(da, ddir, ray, &rayctx->ray, &rayctx->scatter);
350: TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&rayctx->lgctx);
351: TSMonitorSet(ts, TSMonitorLGDMDARay, rayctx, TSMonitorDMDARayDestroy);
352: }
354: PetscOptionsName("-ts_monitor_envelope","Monitor maximum and minimum value of each component of the solution","TSMonitorEnvelope",&opt);
355: if (opt) {
356: TSMonitorEnvelopeCtx ctx;
358: TSMonitorEnvelopeCtxCreate(ts,&ctx);
359: TSMonitorSet(ts,TSMonitorEnvelope,ctx,(PetscErrorCode (*)(void**))TSMonitorEnvelopeCtxDestroy);
360: }
361: flg = PETSC_FALSE;
362: PetscOptionsBool("-ts_monitor_cancel","Remove all monitors","TSMonitorCancel",flg,&flg,&opt);
363: if (opt && flg) TSMonitorCancel(ts);
365: flg = PETSC_FALSE;
366: PetscOptionsBool("-ts_fd_color", "Use finite differences with coloring to compute IJacobian", "TSComputeJacobianDefaultColor", flg, &flg, NULL);
367: if (flg) {
368: DM dm;
369: DMTS tdm;
371: TSGetDM(ts, &dm);
372: DMGetDMTS(dm, &tdm);
373: tdm->ijacobianctx = NULL;
374: TSSetIJacobian(ts, NULL, NULL, TSComputeIJacobianDefaultColor, NULL);
375: PetscInfo(ts, "Setting default finite difference coloring Jacobian matrix\n");
376: }
378: /* Handle specific TS options */
379: if (ts->ops->setfromoptions) {
380: (*ts->ops->setfromoptions)(PetscOptionsObject,ts);
381: }
383: /* Handle TSAdapt options */
384: TSGetAdapt(ts,&ts->adapt);
385: TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type);
386: TSAdaptSetFromOptions(PetscOptionsObject,ts->adapt);
388: /* TS trajectory must be set after TS, since it may use some TS options above */
389: tflg = ts->trajectory ? PETSC_TRUE : PETSC_FALSE;
390: PetscOptionsBool("-ts_save_trajectory","Save the solution at each timestep","TSSetSaveTrajectory",tflg,&tflg,NULL);
391: if (tflg) {
392: TSSetSaveTrajectory(ts);
393: }
395: TSAdjointSetFromOptions(PetscOptionsObject,ts);
397: /* process any options handlers added with PetscObjectAddOptionsHandler() */
398: PetscObjectProcessOptionsHandlers(PetscOptionsObject,(PetscObject)ts);
399: PetscOptionsEnd();
401: if (ts->trajectory) {
402: TSTrajectorySetFromOptions(ts->trajectory,ts);
403: }
405: /* why do we have to do this here and not during TSSetUp? */
406: TSGetSNES(ts,&ts->snes);
407: if (ts->problem_type == TS_LINEAR) {
408: PetscObjectTypeCompareAny((PetscObject)ts->snes,&flg,SNESKSPONLY,SNESKSPTRANSPOSEONLY,"");
409: if (!flg) SNESSetType(ts->snes,SNESKSPONLY);
410: }
411: SNESSetFromOptions(ts->snes);
412: return 0;
413: }
415: /*@
416: TSGetTrajectory - Gets the trajectory from a TS if it exists
418: Collective on TS
420: Input Parameters:
421: . ts - the TS context obtained from TSCreate()
423: Output Parameters:
424: . tr - the TSTrajectory object, if it exists
426: Note: This routine should be called after all TS options have been set
428: Level: advanced
430: .seealso: TSGetTrajectory(), TSAdjointSolve(), TSTrajectory, TSTrajectoryCreate()
432: @*/
433: PetscErrorCode TSGetTrajectory(TS ts,TSTrajectory *tr)
434: {
436: *tr = ts->trajectory;
437: return 0;
438: }
440: /*@
441: TSSetSaveTrajectory - Causes the TS to save its solutions as it iterates forward in time in a TSTrajectory object
443: Collective on TS
445: Input Parameter:
446: . ts - the TS context obtained from TSCreate()
448: Options Database:
449: + -ts_save_trajectory - saves the trajectory to a file
450: - -ts_trajectory_type type - set trajectory type
452: Note: This routine should be called after all TS options have been set
454: The TSTRAJECTORYVISUALIZATION files can be loaded into Python with $PETSC_DIR/lib/petsc/bin/PetscBinaryIOTrajectory.py and
455: MATLAB with $PETSC_DIR/share/petsc/matlab/PetscReadBinaryTrajectory.m
457: Level: intermediate
459: .seealso: TSGetTrajectory(), TSAdjointSolve()
461: @*/
462: PetscErrorCode TSSetSaveTrajectory(TS ts)
463: {
465: if (!ts->trajectory) {
466: TSTrajectoryCreate(PetscObjectComm((PetscObject)ts),&ts->trajectory);
467: }
468: return 0;
469: }
471: /*@
472: TSResetTrajectory - Destroys and recreates the internal TSTrajectory object
474: Collective on TS
476: Input Parameters:
477: . ts - the TS context obtained from TSCreate()
479: Level: intermediate
481: .seealso: TSGetTrajectory(), TSAdjointSolve(), TSRemoveTrajectory()
483: @*/
484: PetscErrorCode TSResetTrajectory(TS ts)
485: {
487: if (ts->trajectory) {
488: TSTrajectoryDestroy(&ts->trajectory);
489: TSTrajectoryCreate(PetscObjectComm((PetscObject)ts),&ts->trajectory);
490: }
491: return 0;
492: }
494: /*@
495: TSRemoveTrajectory - Destroys and removes the internal TSTrajectory object from TS
497: Collective on TS
499: Input Parameters:
500: . ts - the TS context obtained from TSCreate()
502: Level: intermediate
504: .seealso: TSResetTrajectory(), TSAdjointSolve()
506: @*/
507: PetscErrorCode TSRemoveTrajectory(TS ts)
508: {
510: if (ts->trajectory) {
511: TSTrajectoryDestroy(&ts->trajectory);
512: }
513: return 0;
514: }
516: /*@
517: TSComputeRHSJacobian - Computes the Jacobian matrix that has been
518: set with TSSetRHSJacobian().
520: Collective on TS
522: Input Parameters:
523: + ts - the TS context
524: . t - current timestep
525: - U - input vector
527: Output Parameters:
528: + A - Jacobian matrix
529: - B - optional preconditioning matrix
531: Notes:
532: Most users should not need to explicitly call this routine, as it
533: is used internally within the nonlinear solvers.
535: Level: developer
537: .seealso: TSSetRHSJacobian(), KSPSetOperators()
538: @*/
539: PetscErrorCode TSComputeRHSJacobian(TS ts,PetscReal t,Vec U,Mat A,Mat B)
540: {
541: PetscObjectState Ustate;
542: PetscObjectId Uid;
543: DM dm;
544: DMTS tsdm;
545: TSRHSJacobian rhsjacobianfunc;
546: void *ctx;
547: TSRHSFunction rhsfunction;
552: TSGetDM(ts,&dm);
553: DMGetDMTS(dm,&tsdm);
554: DMTSGetRHSFunction(dm,&rhsfunction,NULL);
555: DMTSGetRHSJacobian(dm,&rhsjacobianfunc,&ctx);
556: PetscObjectStateGet((PetscObject)U,&Ustate);
557: PetscObjectGetId((PetscObject)U,&Uid);
559: if (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) && (rhsfunction != TSComputeRHSFunctionLinear)) return 0;
562: if (rhsjacobianfunc) {
563: PetscLogEventBegin(TS_JacobianEval,ts,U,A,B);
564: PetscStackPush("TS user Jacobian function");
565: (*rhsjacobianfunc)(ts,t,U,A,B,ctx);
566: PetscStackPop;
567: ts->rhsjacs++;
568: PetscLogEventEnd(TS_JacobianEval,ts,U,A,B);
569: } else {
570: MatZeroEntries(A);
571: if (B && A != B) MatZeroEntries(B);
572: }
573: ts->rhsjacobian.time = t;
574: ts->rhsjacobian.shift = 0;
575: ts->rhsjacobian.scale = 1.;
576: PetscObjectGetId((PetscObject)U,&ts->rhsjacobian.Xid);
577: PetscObjectStateGet((PetscObject)U,&ts->rhsjacobian.Xstate);
578: return 0;
579: }
581: /*@
582: TSComputeRHSFunction - Evaluates the right-hand-side function.
584: Collective on TS
586: Input Parameters:
587: + ts - the TS context
588: . t - current time
589: - U - state vector
591: Output Parameter:
592: . y - right hand side
594: Note:
595: Most users should not need to explicitly call this routine, as it
596: is used internally within the nonlinear solvers.
598: Level: developer
600: .seealso: TSSetRHSFunction(), TSComputeIFunction()
601: @*/
602: PetscErrorCode TSComputeRHSFunction(TS ts,PetscReal t,Vec U,Vec y)
603: {
604: TSRHSFunction rhsfunction;
605: TSIFunction ifunction;
606: void *ctx;
607: DM dm;
612: TSGetDM(ts,&dm);
613: DMTSGetRHSFunction(dm,&rhsfunction,&ctx);
614: DMTSGetIFunction(dm,&ifunction,NULL);
618: if (rhsfunction) {
619: PetscLogEventBegin(TS_FunctionEval,ts,U,y,0);
620: VecLockReadPush(U);
621: PetscStackPush("TS user right-hand-side function");
622: (*rhsfunction)(ts,t,U,y,ctx);
623: PetscStackPop;
624: VecLockReadPop(U);
625: ts->rhsfuncs++;
626: PetscLogEventEnd(TS_FunctionEval,ts,U,y,0);
627: } else {
628: VecZeroEntries(y);
629: }
630: return 0;
631: }
633: /*@
634: TSComputeSolutionFunction - Evaluates the solution function.
636: Collective on TS
638: Input Parameters:
639: + ts - the TS context
640: - t - current time
642: Output Parameter:
643: . U - the solution
645: Note:
646: Most users should not need to explicitly call this routine, as it
647: is used internally within the nonlinear solvers.
649: Level: developer
651: .seealso: TSSetSolutionFunction(), TSSetRHSFunction(), TSComputeIFunction()
652: @*/
653: PetscErrorCode TSComputeSolutionFunction(TS ts,PetscReal t,Vec U)
654: {
655: TSSolutionFunction solutionfunction;
656: void *ctx;
657: DM dm;
661: TSGetDM(ts,&dm);
662: DMTSGetSolutionFunction(dm,&solutionfunction,&ctx);
664: if (solutionfunction) {
665: PetscStackPush("TS user solution function");
666: (*solutionfunction)(ts,t,U,ctx);
667: PetscStackPop;
668: }
669: return 0;
670: }
671: /*@
672: TSComputeForcingFunction - Evaluates the forcing function.
674: Collective on TS
676: Input Parameters:
677: + ts - the TS context
678: - t - current time
680: Output Parameter:
681: . U - the function value
683: Note:
684: Most users should not need to explicitly call this routine, as it
685: is used internally within the nonlinear solvers.
687: Level: developer
689: .seealso: TSSetSolutionFunction(), TSSetRHSFunction(), TSComputeIFunction()
690: @*/
691: PetscErrorCode TSComputeForcingFunction(TS ts,PetscReal t,Vec U)
692: {
693: void *ctx;
694: DM dm;
695: TSForcingFunction forcing;
699: TSGetDM(ts,&dm);
700: DMTSGetForcingFunction(dm,&forcing,&ctx);
702: if (forcing) {
703: PetscStackPush("TS user forcing function");
704: (*forcing)(ts,t,U,ctx);
705: PetscStackPop;
706: }
707: return 0;
708: }
710: static PetscErrorCode TSGetRHSVec_Private(TS ts,Vec *Frhs)
711: {
712: Vec F;
714: *Frhs = NULL;
715: TSGetIFunction(ts,&F,NULL,NULL);
716: if (!ts->Frhs) {
717: VecDuplicate(F,&ts->Frhs);
718: }
719: *Frhs = ts->Frhs;
720: return 0;
721: }
723: PetscErrorCode TSGetRHSMats_Private(TS ts,Mat *Arhs,Mat *Brhs)
724: {
725: Mat A,B;
726: TSIJacobian ijacobian;
728: if (Arhs) *Arhs = NULL;
729: if (Brhs) *Brhs = NULL;
730: TSGetIJacobian(ts,&A,&B,&ijacobian,NULL);
731: if (Arhs) {
732: if (!ts->Arhs) {
733: if (ijacobian) {
734: MatDuplicate(A,MAT_DO_NOT_COPY_VALUES,&ts->Arhs);
735: TSSetMatStructure(ts,SAME_NONZERO_PATTERN);
736: } else {
737: ts->Arhs = A;
738: PetscObjectReference((PetscObject)A);
739: }
740: } else {
741: PetscBool flg;
742: SNESGetUseMatrixFree(ts->snes,NULL,&flg);
743: /* Handle case where user provided only RHSJacobian and used -snes_mf_operator */
744: if (flg && !ijacobian && ts->Arhs == ts->Brhs) {
745: PetscObjectDereference((PetscObject)ts->Arhs);
746: ts->Arhs = A;
747: PetscObjectReference((PetscObject)A);
748: }
749: }
750: *Arhs = ts->Arhs;
751: }
752: if (Brhs) {
753: if (!ts->Brhs) {
754: if (A != B) {
755: if (ijacobian) {
756: MatDuplicate(B,MAT_DO_NOT_COPY_VALUES,&ts->Brhs);
757: } else {
758: ts->Brhs = B;
759: PetscObjectReference((PetscObject)B);
760: }
761: } else {
762: PetscObjectReference((PetscObject)ts->Arhs);
763: ts->Brhs = ts->Arhs;
764: }
765: }
766: *Brhs = ts->Brhs;
767: }
768: return 0;
769: }
771: /*@
772: TSComputeIFunction - Evaluates the DAE residual written in implicit form F(t,U,Udot)=0
774: Collective on TS
776: Input Parameters:
777: + ts - the TS context
778: . t - current time
779: . U - state vector
780: . Udot - time derivative of state vector
781: - imex - flag indicates if the method is IMEX so that the RHSFunction should be kept separate
783: Output Parameter:
784: . Y - right hand side
786: Note:
787: Most users should not need to explicitly call this routine, as it
788: is used internally within the nonlinear solvers.
790: If the user did did not write their equations in implicit form, this
791: function recasts them in implicit form.
793: Level: developer
795: .seealso: TSSetIFunction(), TSComputeRHSFunction()
796: @*/
797: PetscErrorCode TSComputeIFunction(TS ts,PetscReal t,Vec U,Vec Udot,Vec Y,PetscBool imex)
798: {
799: TSIFunction ifunction;
800: TSRHSFunction rhsfunction;
801: void *ctx;
802: DM dm;
809: TSGetDM(ts,&dm);
810: DMTSGetIFunction(dm,&ifunction,&ctx);
811: DMTSGetRHSFunction(dm,&rhsfunction,NULL);
815: PetscLogEventBegin(TS_FunctionEval,ts,U,Udot,Y);
816: if (ifunction) {
817: PetscStackPush("TS user implicit function");
818: (*ifunction)(ts,t,U,Udot,Y,ctx);
819: PetscStackPop;
820: ts->ifuncs++;
821: }
822: if (imex) {
823: if (!ifunction) {
824: VecCopy(Udot,Y);
825: }
826: } else if (rhsfunction) {
827: if (ifunction) {
828: Vec Frhs;
829: TSGetRHSVec_Private(ts,&Frhs);
830: TSComputeRHSFunction(ts,t,U,Frhs);
831: VecAXPY(Y,-1,Frhs);
832: } else {
833: TSComputeRHSFunction(ts,t,U,Y);
834: VecAYPX(Y,-1,Udot);
835: }
836: }
837: PetscLogEventEnd(TS_FunctionEval,ts,U,Udot,Y);
838: return 0;
839: }
841: /*
842: TSRecoverRHSJacobian - Recover the Jacobian matrix so that one can call TSComputeRHSJacobian() on it.
844: Note:
845: This routine is needed when one switches from TSComputeIJacobian() to TSComputeRHSJacobian() because the Jacobian matrix may be shifted or scaled in TSComputeIJacobian().
847: */
848: static PetscErrorCode TSRecoverRHSJacobian(TS ts,Mat A,Mat B)
849: {
854: if (ts->rhsjacobian.shift) {
855: MatShift(A,-ts->rhsjacobian.shift);
856: }
857: if (ts->rhsjacobian.scale == -1.) {
858: MatScale(A,-1);
859: }
860: if (B && B == ts->Brhs && A != B) {
861: if (ts->rhsjacobian.shift) {
862: MatShift(B,-ts->rhsjacobian.shift);
863: }
864: if (ts->rhsjacobian.scale == -1.) {
865: MatScale(B,-1);
866: }
867: }
868: ts->rhsjacobian.shift = 0;
869: ts->rhsjacobian.scale = 1.;
870: return 0;
871: }
873: /*@
874: TSComputeIJacobian - Evaluates the Jacobian of the DAE
876: Collective on TS
878: Input
879: Input Parameters:
880: + ts - the TS context
881: . t - current timestep
882: . U - state vector
883: . Udot - time derivative of state vector
884: . shift - shift to apply, see note below
885: - imex - flag indicates if the method is IMEX so that the RHSJacobian should be kept separate
887: Output Parameters:
888: + A - Jacobian matrix
889: - B - matrix from which the preconditioner is constructed; often the same as A
891: Notes:
892: If F(t,U,Udot)=0 is the DAE, the required Jacobian is
894: dF/dU + shift*dF/dUdot
896: Most users should not need to explicitly call this routine, as it
897: is used internally within the nonlinear solvers.
899: Level: developer
901: .seealso: TSSetIJacobian()
902: @*/
903: PetscErrorCode TSComputeIJacobian(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,PetscBool imex)
904: {
905: TSIJacobian ijacobian;
906: TSRHSJacobian rhsjacobian;
907: DM dm;
908: void *ctx;
918: TSGetDM(ts,&dm);
919: DMTSGetIJacobian(dm,&ijacobian,&ctx);
920: DMTSGetRHSJacobian(dm,&rhsjacobian,NULL);
924: PetscLogEventBegin(TS_JacobianEval,ts,U,A,B);
925: if (ijacobian) {
926: PetscStackPush("TS user implicit Jacobian");
927: (*ijacobian)(ts,t,U,Udot,shift,A,B,ctx);
928: ts->ijacs++;
929: PetscStackPop;
930: }
931: if (imex) {
932: if (!ijacobian) { /* system was written as Udot = G(t,U) */
933: PetscBool assembled;
934: if (rhsjacobian) {
935: Mat Arhs = NULL;
936: TSGetRHSMats_Private(ts,&Arhs,NULL);
937: if (A == Arhs) {
939: ts->rhsjacobian.time = PETSC_MIN_REAL;
940: }
941: }
942: MatZeroEntries(A);
943: MatAssembled(A,&assembled);
944: if (!assembled) {
945: MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);
946: MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);
947: }
948: MatShift(A,shift);
949: if (A != B) {
950: MatZeroEntries(B);
951: MatAssembled(B,&assembled);
952: if (!assembled) {
953: MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
954: MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);
955: }
956: MatShift(B,shift);
957: }
958: }
959: } else {
960: Mat Arhs = NULL,Brhs = NULL;
961: if (rhsjacobian) { /* RHSJacobian needs to be converted to part of IJacobian if exists */
962: TSGetRHSMats_Private(ts,&Arhs,&Brhs);
963: }
964: if (Arhs == A) { /* No IJacobian matrix, so we only have the RHS matrix */
965: PetscObjectState Ustate;
966: PetscObjectId Uid;
967: TSRHSFunction rhsfunction;
969: DMTSGetRHSFunction(dm,&rhsfunction,NULL);
970: PetscObjectStateGet((PetscObject)U,&Ustate);
971: PetscObjectGetId((PetscObject)U,&Uid);
972: if ((rhsjacobian == TSComputeRHSJacobianConstant || (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) && rhsfunction != TSComputeRHSFunctionLinear)) && ts->rhsjacobian.scale == -1.) { /* No need to recompute RHSJacobian */
973: MatShift(A,shift-ts->rhsjacobian.shift); /* revert the old shift and add the new shift with a single call to MatShift */
974: if (A != B) {
975: MatShift(B,shift-ts->rhsjacobian.shift);
976: }
977: } else {
978: PetscBool flg;
980: if (ts->rhsjacobian.reuse) { /* Undo the damage */
981: /* MatScale has a short path for this case.
982: However, this code path is taken the first time TSComputeRHSJacobian is called
983: and the matrices have not been assembled yet */
984: TSRecoverRHSJacobian(ts,A,B);
985: }
986: TSComputeRHSJacobian(ts,t,U,A,B);
987: SNESGetUseMatrixFree(ts->snes,NULL,&flg);
988: /* since -snes_mf_operator uses the full SNES function it does not need to be shifted or scaled here */
989: if (!flg) {
990: MatScale(A,-1);
991: MatShift(A,shift);
992: }
993: if (A != B) {
994: MatScale(B,-1);
995: MatShift(B,shift);
996: }
997: }
998: ts->rhsjacobian.scale = -1;
999: ts->rhsjacobian.shift = shift;
1000: } else if (Arhs) { /* Both IJacobian and RHSJacobian */
1001: if (!ijacobian) { /* No IJacobian provided, but we have a separate RHS matrix */
1002: MatZeroEntries(A);
1003: MatShift(A,shift);
1004: if (A != B) {
1005: MatZeroEntries(B);
1006: MatShift(B,shift);
1007: }
1008: }
1009: TSComputeRHSJacobian(ts,t,U,Arhs,Brhs);
1010: MatAXPY(A,-1,Arhs,ts->axpy_pattern);
1011: if (A != B) {
1012: MatAXPY(B,-1,Brhs,ts->axpy_pattern);
1013: }
1014: }
1015: }
1016: PetscLogEventEnd(TS_JacobianEval,ts,U,A,B);
1017: return 0;
1018: }
1020: /*@C
1021: TSSetRHSFunction - Sets the routine for evaluating the function,
1022: where U_t = G(t,u).
1024: Logically Collective on TS
1026: Input Parameters:
1027: + ts - the TS context obtained from TSCreate()
1028: . r - vector to put the computed right hand side (or NULL to have it created)
1029: . f - routine for evaluating the right-hand-side function
1030: - ctx - [optional] user-defined context for private data for the
1031: function evaluation routine (may be NULL)
1033: Calling sequence of f:
1034: $ PetscErrorCode f(TS ts,PetscReal t,Vec u,Vec F,void *ctx);
1036: + ts - timestep context
1037: . t - current timestep
1038: . u - input vector
1039: . F - function vector
1040: - ctx - [optional] user-defined function context
1042: Level: beginner
1044: Notes:
1045: You must call this function or TSSetIFunction() to define your ODE. You cannot use this function when solving a DAE.
1047: .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSSetIFunction()
1048: @*/
1049: PetscErrorCode TSSetRHSFunction(TS ts,Vec r,PetscErrorCode (*f)(TS,PetscReal,Vec,Vec,void*),void *ctx)
1050: {
1051: SNES snes;
1052: Vec ralloc = NULL;
1053: DM dm;
1058: TSGetDM(ts,&dm);
1059: DMTSSetRHSFunction(dm,f,ctx);
1060: TSGetSNES(ts,&snes);
1061: if (!r && !ts->dm && ts->vec_sol) {
1062: VecDuplicate(ts->vec_sol,&ralloc);
1063: r = ralloc;
1064: }
1065: SNESSetFunction(snes,r,SNESTSFormFunction,ts);
1066: VecDestroy(&ralloc);
1067: return 0;
1068: }
1070: /*@C
1071: TSSetSolutionFunction - Provide a function that computes the solution of the ODE or DAE
1073: Logically Collective on TS
1075: Input Parameters:
1076: + ts - the TS context obtained from TSCreate()
1077: . f - routine for evaluating the solution
1078: - ctx - [optional] user-defined context for private data for the
1079: function evaluation routine (may be NULL)
1081: Calling sequence of f:
1082: $ PetscErrorCode f(TS ts,PetscReal t,Vec u,void *ctx);
1084: + t - current timestep
1085: . u - output vector
1086: - ctx - [optional] user-defined function context
1088: Options Database:
1089: + -ts_monitor_lg_error - create a graphical monitor of error history, requires user to have provided TSSetSolutionFunction()
1090: - -ts_monitor_draw_error - Monitor error graphically, requires user to have provided TSSetSolutionFunction()
1092: Notes:
1093: This routine is used for testing accuracy of time integration schemes when you already know the solution.
1094: If analytic solutions are not known for your system, consider using the Method of Manufactured Solutions to
1095: create closed-form solutions with non-physical forcing terms.
1097: For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history.
1099: Level: beginner
1101: .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSComputeSolutionFunction(), TSSetForcingFunction(), TSSetSolution(), TSGetSolution(), TSMonitorLGError(), TSMonitorDrawError()
1102: @*/
1103: PetscErrorCode TSSetSolutionFunction(TS ts,PetscErrorCode (*f)(TS,PetscReal,Vec,void*),void *ctx)
1104: {
1105: DM dm;
1108: TSGetDM(ts,&dm);
1109: DMTSSetSolutionFunction(dm,f,ctx);
1110: return 0;
1111: }
1113: /*@C
1114: TSSetForcingFunction - Provide a function that computes a forcing term for a ODE or PDE
1116: Logically Collective on TS
1118: Input Parameters:
1119: + ts - the TS context obtained from TSCreate()
1120: . func - routine for evaluating the forcing function
1121: - ctx - [optional] user-defined context for private data for the
1122: function evaluation routine (may be NULL)
1124: Calling sequence of func:
1125: $ PetscErrorCode func (TS ts,PetscReal t,Vec f,void *ctx);
1127: + t - current timestep
1128: . f - output vector
1129: - ctx - [optional] user-defined function context
1131: Notes:
1132: This routine is useful for testing accuracy of time integration schemes when using the Method of Manufactured Solutions to
1133: create closed-form solutions with a non-physical forcing term. It allows you to use the Method of Manufactored Solution without directly editing the
1134: definition of the problem you are solving and hence possibly introducing bugs.
1136: This replaces the ODE F(u,u_t,t) = 0 the TS is solving with F(u,u_t,t) - func(t) = 0
1138: This forcing function does not depend on the solution to the equations, it can only depend on spatial location, time, and possibly parameters, the
1139: parameters can be passed in the ctx variable.
1141: For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history.
1143: Level: beginner
1145: .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSComputeSolutionFunction(), TSSetSolutionFunction()
1146: @*/
1147: PetscErrorCode TSSetForcingFunction(TS ts,TSForcingFunction func,void *ctx)
1148: {
1149: DM dm;
1152: TSGetDM(ts,&dm);
1153: DMTSSetForcingFunction(dm,func,ctx);
1154: return 0;
1155: }
1157: /*@C
1158: TSSetRHSJacobian - Sets the function to compute the Jacobian of G,
1159: where U_t = G(U,t), as well as the location to store the matrix.
1161: Logically Collective on TS
1163: Input Parameters:
1164: + ts - the TS context obtained from TSCreate()
1165: . Amat - (approximate) Jacobian matrix
1166: . Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat)
1167: . f - the Jacobian evaluation routine
1168: - ctx - [optional] user-defined context for private data for the
1169: Jacobian evaluation routine (may be NULL)
1171: Calling sequence of f:
1172: $ PetscErrorCode f(TS ts,PetscReal t,Vec u,Mat A,Mat B,void *ctx);
1174: + t - current timestep
1175: . u - input vector
1176: . Amat - (approximate) Jacobian matrix
1177: . Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat)
1178: - ctx - [optional] user-defined context for matrix evaluation routine
1180: Notes:
1181: You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1183: The TS solver may modify the nonzero structure and the entries of the matrices Amat and Pmat between the calls to f()
1184: You should not assume the values are the same in the next call to f() as you set them in the previous call.
1186: Level: beginner
1188: .seealso: SNESComputeJacobianDefaultColor(), TSSetRHSFunction(), TSRHSJacobianSetReuse(), TSSetIJacobian()
1190: @*/
1191: PetscErrorCode TSSetRHSJacobian(TS ts,Mat Amat,Mat Pmat,TSRHSJacobian f,void *ctx)
1192: {
1193: SNES snes;
1194: DM dm;
1195: TSIJacobian ijacobian;
1203: TSGetDM(ts,&dm);
1204: DMTSSetRHSJacobian(dm,f,ctx);
1205: DMTSGetIJacobian(dm,&ijacobian,NULL);
1206: TSGetSNES(ts,&snes);
1207: if (!ijacobian) {
1208: SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts);
1209: }
1210: if (Amat) {
1211: PetscObjectReference((PetscObject)Amat);
1212: MatDestroy(&ts->Arhs);
1213: ts->Arhs = Amat;
1214: }
1215: if (Pmat) {
1216: PetscObjectReference((PetscObject)Pmat);
1217: MatDestroy(&ts->Brhs);
1218: ts->Brhs = Pmat;
1219: }
1220: return 0;
1221: }
1223: /*@C
1224: TSSetIFunction - Set the function to compute F(t,U,U_t) where F() = 0 is the DAE to be solved.
1226: Logically Collective on TS
1228: Input Parameters:
1229: + ts - the TS context obtained from TSCreate()
1230: . r - vector to hold the residual (or NULL to have it created internally)
1231: . f - the function evaluation routine
1232: - ctx - user-defined context for private data for the function evaluation routine (may be NULL)
1234: Calling sequence of f:
1235: $ PetscErrorCode f(TS ts,PetscReal t,Vec u,Vec u_t,Vec F,ctx);
1237: + t - time at step/stage being solved
1238: . u - state vector
1239: . u_t - time derivative of state vector
1240: . F - function vector
1241: - ctx - [optional] user-defined context for matrix evaluation routine
1243: Important:
1244: The user MUST call either this routine or TSSetRHSFunction() to define the ODE. When solving DAEs you must use this function.
1246: Level: beginner
1248: .seealso: TSSetRHSJacobian(), TSSetRHSFunction(), TSSetIJacobian()
1249: @*/
1250: PetscErrorCode TSSetIFunction(TS ts,Vec r,TSIFunction f,void *ctx)
1251: {
1252: SNES snes;
1253: Vec ralloc = NULL;
1254: DM dm;
1259: TSGetDM(ts,&dm);
1260: DMTSSetIFunction(dm,f,ctx);
1262: TSGetSNES(ts,&snes);
1263: if (!r && !ts->dm && ts->vec_sol) {
1264: VecDuplicate(ts->vec_sol,&ralloc);
1265: r = ralloc;
1266: }
1267: SNESSetFunction(snes,r,SNESTSFormFunction,ts);
1268: VecDestroy(&ralloc);
1269: return 0;
1270: }
1272: /*@C
1273: TSGetIFunction - Returns the vector where the implicit residual is stored and the function/context to compute it.
1275: Not Collective
1277: Input Parameter:
1278: . ts - the TS context
1280: Output Parameters:
1281: + r - vector to hold residual (or NULL)
1282: . func - the function to compute residual (or NULL)
1283: - ctx - the function context (or NULL)
1285: Level: advanced
1287: .seealso: TSSetIFunction(), SNESGetFunction()
1288: @*/
1289: PetscErrorCode TSGetIFunction(TS ts,Vec *r,TSIFunction *func,void **ctx)
1290: {
1291: SNES snes;
1292: DM dm;
1295: TSGetSNES(ts,&snes);
1296: SNESGetFunction(snes,r,NULL,NULL);
1297: TSGetDM(ts,&dm);
1298: DMTSGetIFunction(dm,func,ctx);
1299: return 0;
1300: }
1302: /*@C
1303: TSGetRHSFunction - Returns the vector where the right hand side is stored and the function/context to compute it.
1305: Not Collective
1307: Input Parameter:
1308: . ts - the TS context
1310: Output Parameters:
1311: + r - vector to hold computed right hand side (or NULL)
1312: . func - the function to compute right hand side (or NULL)
1313: - ctx - the function context (or NULL)
1315: Level: advanced
1317: .seealso: TSSetRHSFunction(), SNESGetFunction()
1318: @*/
1319: PetscErrorCode TSGetRHSFunction(TS ts,Vec *r,TSRHSFunction *func,void **ctx)
1320: {
1321: SNES snes;
1322: DM dm;
1325: TSGetSNES(ts,&snes);
1326: SNESGetFunction(snes,r,NULL,NULL);
1327: TSGetDM(ts,&dm);
1328: DMTSGetRHSFunction(dm,func,ctx);
1329: return 0;
1330: }
1332: /*@C
1333: TSSetIJacobian - Set the function to compute the matrix dF/dU + a*dF/dU_t where F(t,U,U_t) is the function
1334: provided with TSSetIFunction().
1336: Logically Collective on TS
1338: Input Parameters:
1339: + ts - the TS context obtained from TSCreate()
1340: . Amat - (approximate) Jacobian matrix
1341: . Pmat - matrix used to compute preconditioner (usually the same as Amat)
1342: . f - the Jacobian evaluation routine
1343: - ctx - user-defined context for private data for the Jacobian evaluation routine (may be NULL)
1345: Calling sequence of f:
1346: $ PetscErrorCode f(TS ts,PetscReal t,Vec U,Vec U_t,PetscReal a,Mat Amat,Mat Pmat,void *ctx);
1348: + t - time at step/stage being solved
1349: . U - state vector
1350: . U_t - time derivative of state vector
1351: . a - shift
1352: . Amat - (approximate) Jacobian of F(t,U,W+a*U), equivalent to dF/dU + a*dF/dU_t
1353: . Pmat - matrix used for constructing preconditioner, usually the same as Amat
1354: - ctx - [optional] user-defined context for matrix evaluation routine
1356: Notes:
1357: The matrices Amat and Pmat are exactly the matrices that are used by SNES for the nonlinear solve.
1359: If you know the operator Amat has a null space you can use MatSetNullSpace() and MatSetTransposeNullSpace() to supply the null
1360: space to Amat and the KSP solvers will automatically use that null space as needed during the solution process.
1362: The matrix dF/dU + a*dF/dU_t you provide turns out to be
1363: the Jacobian of F(t,U,W+a*U) where F(t,U,U_t) = 0 is the DAE to be solved.
1364: The time integrator internally approximates U_t by W+a*U where the positive "shift"
1365: a and vector W depend on the integration method, step size, and past states. For example with
1366: the backward Euler method a = 1/dt and W = -a*U(previous timestep) so
1367: W + a*U = a*(U - U(previous timestep)) = (U - U(previous timestep))/dt
1369: You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1371: The TS solver may modify the nonzero structure and the entries of the matrices Amat and Pmat between the calls to f()
1372: You should not assume the values are the same in the next call to f() as you set them in the previous call.
1374: Level: beginner
1376: .seealso: TSSetIFunction(), TSSetRHSJacobian(), SNESComputeJacobianDefaultColor(), SNESComputeJacobianDefault(), TSSetRHSFunction()
1378: @*/
1379: PetscErrorCode TSSetIJacobian(TS ts,Mat Amat,Mat Pmat,TSIJacobian f,void *ctx)
1380: {
1381: SNES snes;
1382: DM dm;
1390: TSGetDM(ts,&dm);
1391: DMTSSetIJacobian(dm,f,ctx);
1393: TSGetSNES(ts,&snes);
1394: SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts);
1395: return 0;
1396: }
1398: /*@
1399: TSRHSJacobianSetReuse - restore RHS Jacobian before re-evaluating. Without this flag, TS will change the sign and
1400: shift the RHS Jacobian for a finite-time-step implicit solve, in which case the user function will need to recompute
1401: the entire Jacobian. The reuse flag must be set if the evaluation function will assume that the matrix entries have
1402: not been changed by the TS.
1404: Logically Collective
1406: Input Parameters:
1407: + ts - TS context obtained from TSCreate()
1408: - reuse - PETSC_TRUE if the RHS Jacobian
1410: Level: intermediate
1412: .seealso: TSSetRHSJacobian(), TSComputeRHSJacobianConstant()
1413: @*/
1414: PetscErrorCode TSRHSJacobianSetReuse(TS ts,PetscBool reuse)
1415: {
1416: ts->rhsjacobian.reuse = reuse;
1417: return 0;
1418: }
1420: /*@C
1421: TSSetI2Function - Set the function to compute F(t,U,U_t,U_tt) where F = 0 is the DAE to be solved.
1423: Logically Collective on TS
1425: Input Parameters:
1426: + ts - the TS context obtained from TSCreate()
1427: . F - vector to hold the residual (or NULL to have it created internally)
1428: . fun - the function evaluation routine
1429: - ctx - user-defined context for private data for the function evaluation routine (may be NULL)
1431: Calling sequence of fun:
1432: $ PetscErrorCode fun(TS ts,PetscReal t,Vec U,Vec U_t,Vec U_tt,Vec F,ctx);
1434: + t - time at step/stage being solved
1435: . U - state vector
1436: . U_t - time derivative of state vector
1437: . U_tt - second time derivative of state vector
1438: . F - function vector
1439: - ctx - [optional] user-defined context for matrix evaluation routine (may be NULL)
1441: Level: beginner
1443: .seealso: TSSetI2Jacobian(), TSSetIFunction(), TSCreate(), TSSetRHSFunction()
1444: @*/
1445: PetscErrorCode TSSetI2Function(TS ts,Vec F,TSI2Function fun,void *ctx)
1446: {
1447: DM dm;
1451: TSSetIFunction(ts,F,NULL,NULL);
1452: TSGetDM(ts,&dm);
1453: DMTSSetI2Function(dm,fun,ctx);
1454: return 0;
1455: }
1457: /*@C
1458: TSGetI2Function - Returns the vector where the implicit residual is stored and the function/context to compute it.
1460: Not Collective
1462: Input Parameter:
1463: . ts - the TS context
1465: Output Parameters:
1466: + r - vector to hold residual (or NULL)
1467: . fun - the function to compute residual (or NULL)
1468: - ctx - the function context (or NULL)
1470: Level: advanced
1472: .seealso: TSSetIFunction(), SNESGetFunction(), TSCreate()
1473: @*/
1474: PetscErrorCode TSGetI2Function(TS ts,Vec *r,TSI2Function *fun,void **ctx)
1475: {
1476: SNES snes;
1477: DM dm;
1480: TSGetSNES(ts,&snes);
1481: SNESGetFunction(snes,r,NULL,NULL);
1482: TSGetDM(ts,&dm);
1483: DMTSGetI2Function(dm,fun,ctx);
1484: return 0;
1485: }
1487: /*@C
1488: TSSetI2Jacobian - Set the function to compute the matrix dF/dU + v*dF/dU_t + a*dF/dU_tt
1489: where F(t,U,U_t,U_tt) is the function you provided with TSSetI2Function().
1491: Logically Collective on TS
1493: Input Parameters:
1494: + ts - the TS context obtained from TSCreate()
1495: . J - Jacobian matrix
1496: . P - preconditioning matrix for J (may be same as J)
1497: . jac - the Jacobian evaluation routine
1498: - ctx - user-defined context for private data for the Jacobian evaluation routine (may be NULL)
1500: Calling sequence of jac:
1501: $ PetscErrorCode jac(TS ts,PetscReal t,Vec U,Vec U_t,Vec U_tt,PetscReal v,PetscReal a,Mat J,Mat P,void *ctx);
1503: + t - time at step/stage being solved
1504: . U - state vector
1505: . U_t - time derivative of state vector
1506: . U_tt - second time derivative of state vector
1507: . v - shift for U_t
1508: . a - shift for U_tt
1509: . J - Jacobian of G(U) = F(t,U,W+v*U,W'+a*U), equivalent to dF/dU + v*dF/dU_t + a*dF/dU_tt
1510: . P - preconditioning matrix for J, may be same as J
1511: - ctx - [optional] user-defined context for matrix evaluation routine
1513: Notes:
1514: The matrices J and P are exactly the matrices that are used by SNES for the nonlinear solve.
1516: The matrix dF/dU + v*dF/dU_t + a*dF/dU_tt you provide turns out to be
1517: the Jacobian of G(U) = F(t,U,W+v*U,W'+a*U) where F(t,U,U_t,U_tt) = 0 is the DAE to be solved.
1518: The time integrator internally approximates U_t by W+v*U and U_tt by W'+a*U where the positive "shift"
1519: parameters 'v' and 'a' and vectors W, W' depend on the integration method, step size, and past states.
1521: Level: beginner
1523: .seealso: TSSetI2Function(), TSGetI2Jacobian()
1524: @*/
1525: PetscErrorCode TSSetI2Jacobian(TS ts,Mat J,Mat P,TSI2Jacobian jac,void *ctx)
1526: {
1527: DM dm;
1532: TSSetIJacobian(ts,J,P,NULL,NULL);
1533: TSGetDM(ts,&dm);
1534: DMTSSetI2Jacobian(dm,jac,ctx);
1535: return 0;
1536: }
1538: /*@C
1539: TSGetI2Jacobian - Returns the implicit Jacobian at the present timestep.
1541: Not Collective, but parallel objects are returned if TS is parallel
1543: Input Parameter:
1544: . ts - The TS context obtained from TSCreate()
1546: Output Parameters:
1547: + J - The (approximate) Jacobian of F(t,U,U_t,U_tt)
1548: . P - The matrix from which the preconditioner is constructed, often the same as J
1549: . jac - The function to compute the Jacobian matrices
1550: - ctx - User-defined context for Jacobian evaluation routine
1552: Notes:
1553: You can pass in NULL for any return argument you do not need.
1555: Level: advanced
1557: .seealso: TSGetTimeStep(), TSGetMatrices(), TSGetTime(), TSGetStepNumber(), TSSetI2Jacobian(), TSGetI2Function(), TSCreate()
1559: @*/
1560: PetscErrorCode TSGetI2Jacobian(TS ts,Mat *J,Mat *P,TSI2Jacobian *jac,void **ctx)
1561: {
1562: SNES snes;
1563: DM dm;
1565: TSGetSNES(ts,&snes);
1566: SNESSetUpMatrices(snes);
1567: SNESGetJacobian(snes,J,P,NULL,NULL);
1568: TSGetDM(ts,&dm);
1569: DMTSGetI2Jacobian(dm,jac,ctx);
1570: return 0;
1571: }
1573: /*@
1574: TSComputeI2Function - Evaluates the DAE residual written in implicit form F(t,U,U_t,U_tt) = 0
1576: Collective on TS
1578: Input Parameters:
1579: + ts - the TS context
1580: . t - current time
1581: . U - state vector
1582: . V - time derivative of state vector (U_t)
1583: - A - second time derivative of state vector (U_tt)
1585: Output Parameter:
1586: . F - the residual vector
1588: Note:
1589: Most users should not need to explicitly call this routine, as it
1590: is used internally within the nonlinear solvers.
1592: Level: developer
1594: .seealso: TSSetI2Function(), TSGetI2Function()
1595: @*/
1596: PetscErrorCode TSComputeI2Function(TS ts,PetscReal t,Vec U,Vec V,Vec A,Vec F)
1597: {
1598: DM dm;
1599: TSI2Function I2Function;
1600: void *ctx;
1601: TSRHSFunction rhsfunction;
1609: TSGetDM(ts,&dm);
1610: DMTSGetI2Function(dm,&I2Function,&ctx);
1611: DMTSGetRHSFunction(dm,&rhsfunction,NULL);
1613: if (!I2Function) {
1614: TSComputeIFunction(ts,t,U,A,F,PETSC_FALSE);
1615: return 0;
1616: }
1618: PetscLogEventBegin(TS_FunctionEval,ts,U,V,F);
1620: PetscStackPush("TS user implicit function");
1621: I2Function(ts,t,U,V,A,F,ctx);
1622: PetscStackPop;
1624: if (rhsfunction) {
1625: Vec Frhs;
1626: TSGetRHSVec_Private(ts,&Frhs);
1627: TSComputeRHSFunction(ts,t,U,Frhs);
1628: VecAXPY(F,-1,Frhs);
1629: }
1631: PetscLogEventEnd(TS_FunctionEval,ts,U,V,F);
1632: return 0;
1633: }
1635: /*@
1636: TSComputeI2Jacobian - Evaluates the Jacobian of the DAE
1638: Collective on TS
1640: Input Parameters:
1641: + ts - the TS context
1642: . t - current timestep
1643: . U - state vector
1644: . V - time derivative of state vector
1645: . A - second time derivative of state vector
1646: . shiftV - shift to apply, see note below
1647: - shiftA - shift to apply, see note below
1649: Output Parameters:
1650: + J - Jacobian matrix
1651: - P - optional preconditioning matrix
1653: Notes:
1654: If F(t,U,V,A)=0 is the DAE, the required Jacobian is
1656: dF/dU + shiftV*dF/dV + shiftA*dF/dA
1658: Most users should not need to explicitly call this routine, as it
1659: is used internally within the nonlinear solvers.
1661: Level: developer
1663: .seealso: TSSetI2Jacobian()
1664: @*/
1665: PetscErrorCode TSComputeI2Jacobian(TS ts,PetscReal t,Vec U,Vec V,Vec A,PetscReal shiftV,PetscReal shiftA,Mat J,Mat P)
1666: {
1667: DM dm;
1668: TSI2Jacobian I2Jacobian;
1669: void *ctx;
1670: TSRHSJacobian rhsjacobian;
1679: TSGetDM(ts,&dm);
1680: DMTSGetI2Jacobian(dm,&I2Jacobian,&ctx);
1681: DMTSGetRHSJacobian(dm,&rhsjacobian,NULL);
1683: if (!I2Jacobian) {
1684: TSComputeIJacobian(ts,t,U,A,shiftA,J,P,PETSC_FALSE);
1685: return 0;
1686: }
1688: PetscLogEventBegin(TS_JacobianEval,ts,U,J,P);
1690: PetscStackPush("TS user implicit Jacobian");
1691: I2Jacobian(ts,t,U,V,A,shiftV,shiftA,J,P,ctx);
1692: PetscStackPop;
1694: if (rhsjacobian) {
1695: Mat Jrhs,Prhs;
1696: TSGetRHSMats_Private(ts,&Jrhs,&Prhs);
1697: TSComputeRHSJacobian(ts,t,U,Jrhs,Prhs);
1698: MatAXPY(J,-1,Jrhs,ts->axpy_pattern);
1699: if (P != J) MatAXPY(P,-1,Prhs,ts->axpy_pattern);
1700: }
1702: PetscLogEventEnd(TS_JacobianEval,ts,U,J,P);
1703: return 0;
1704: }
1706: /*@C
1707: TSSetTransientVariable - sets function to transform from state to transient variables
1709: Logically Collective
1711: Input Parameters:
1712: + ts - time stepping context on which to change the transient variable
1713: . tvar - a function that transforms to transient variables
1714: - ctx - a context for tvar
1716: Calling sequence of tvar:
1717: $ PetscErrorCode tvar(TS ts,Vec p,Vec c,void *ctx);
1719: + ts - timestep context
1720: . p - input vector (primitive form)
1721: . c - output vector, transient variables (conservative form)
1722: - ctx - [optional] user-defined function context
1724: Level: advanced
1726: Notes:
1727: This is typically used to transform from primitive to conservative variables so that a time integrator (e.g., TSBDF)
1728: can be conservative. In this context, primitive variables P are used to model the state (e.g., because they lead to
1729: well-conditioned formulations even in limiting cases such as low-Mach or zero porosity). The transient variable is
1730: C(P), specified by calling this function. An IFunction thus receives arguments (P, Cdot) and the IJacobian must be
1731: evaluated via the chain rule, as in
1733: dF/dP + shift * dF/dCdot dC/dP.
1735: .seealso: DMTSSetTransientVariable(), DMTSGetTransientVariable(), TSSetIFunction(), TSSetIJacobian()
1736: @*/
1737: PetscErrorCode TSSetTransientVariable(TS ts,TSTransientVariable tvar,void *ctx)
1738: {
1739: DM dm;
1742: TSGetDM(ts,&dm);
1743: DMTSSetTransientVariable(dm,tvar,ctx);
1744: return 0;
1745: }
1747: /*@
1748: TSComputeTransientVariable - transforms state (primitive) variables to transient (conservative) variables
1750: Logically Collective
1752: Input Parameters:
1753: + ts - TS on which to compute
1754: - U - state vector to be transformed to transient variables
1756: Output Parameters:
1757: . C - transient (conservative) variable
1759: Developer Notes:
1760: If DMTSSetTransientVariable() has not been called, then C is not modified in this routine and C=NULL is allowed.
1761: This makes it safe to call without a guard. One can use TSHasTransientVariable() to check if transient variables are
1762: being used.
1764: Level: developer
1766: .seealso: DMTSSetTransientVariable(), TSComputeIFunction(), TSComputeIJacobian()
1767: @*/
1768: PetscErrorCode TSComputeTransientVariable(TS ts,Vec U,Vec C)
1769: {
1770: DM dm;
1771: DMTS dmts;
1775: TSGetDM(ts,&dm);
1776: DMGetDMTS(dm,&dmts);
1777: if (dmts->ops->transientvar) {
1779: (*dmts->ops->transientvar)(ts,U,C,dmts->transientvarctx);
1780: }
1781: return 0;
1782: }
1784: /*@
1785: TSHasTransientVariable - determine whether transient variables have been set
1787: Logically Collective
1789: Input Parameters:
1790: . ts - TS on which to compute
1792: Output Parameters:
1793: . has - PETSC_TRUE if transient variables have been set
1795: Level: developer
1797: .seealso: DMTSSetTransientVariable(), TSComputeTransientVariable()
1798: @*/
1799: PetscErrorCode TSHasTransientVariable(TS ts,PetscBool *has)
1800: {
1801: DM dm;
1802: DMTS dmts;
1805: TSGetDM(ts,&dm);
1806: DMGetDMTS(dm,&dmts);
1807: *has = dmts->ops->transientvar ? PETSC_TRUE : PETSC_FALSE;
1808: return 0;
1809: }
1811: /*@
1812: TS2SetSolution - Sets the initial solution and time derivative vectors
1813: for use by the TS routines handling second order equations.
1815: Logically Collective on TS
1817: Input Parameters:
1818: + ts - the TS context obtained from TSCreate()
1819: . u - the solution vector
1820: - v - the time derivative vector
1822: Level: beginner
1824: @*/
1825: PetscErrorCode TS2SetSolution(TS ts,Vec u,Vec v)
1826: {
1830: TSSetSolution(ts,u);
1831: PetscObjectReference((PetscObject)v);
1832: VecDestroy(&ts->vec_dot);
1833: ts->vec_dot = v;
1834: return 0;
1835: }
1837: /*@
1838: TS2GetSolution - Returns the solution and time derivative at the present timestep
1839: for second order equations. It is valid to call this routine inside the function
1840: that you are evaluating in order to move to the new timestep. This vector not
1841: changed until the solution at the next timestep has been calculated.
1843: Not Collective, but Vec returned is parallel if TS is parallel
1845: Input Parameter:
1846: . ts - the TS context obtained from TSCreate()
1848: Output Parameters:
1849: + u - the vector containing the solution
1850: - v - the vector containing the time derivative
1852: Level: intermediate
1854: .seealso: TS2SetSolution(), TSGetTimeStep(), TSGetTime()
1856: @*/
1857: PetscErrorCode TS2GetSolution(TS ts,Vec *u,Vec *v)
1858: {
1862: if (u) *u = ts->vec_sol;
1863: if (v) *v = ts->vec_dot;
1864: return 0;
1865: }
1867: /*@C
1868: TSLoad - Loads a KSP that has been stored in binary with KSPView().
1870: Collective on PetscViewer
1872: Input Parameters:
1873: + newdm - the newly loaded TS, this needs to have been created with TSCreate() or
1874: some related function before a call to TSLoad().
1875: - viewer - binary file viewer, obtained from PetscViewerBinaryOpen()
1877: Level: intermediate
1879: Notes:
1880: The type is determined by the data in the file, any type set into the TS before this call is ignored.
1882: Notes for advanced users:
1883: Most users should not need to know the details of the binary storage
1884: format, since TSLoad() and TSView() completely hide these details.
1885: But for anyone who's interested, the standard binary matrix storage
1886: format is
1887: .vb
1888: has not yet been determined
1889: .ve
1891: .seealso: PetscViewerBinaryOpen(), TSView(), MatLoad(), VecLoad()
1892: @*/
1893: PetscErrorCode TSLoad(TS ts, PetscViewer viewer)
1894: {
1895: PetscBool isbinary;
1896: PetscInt classid;
1897: char type[256];
1898: DMTS sdm;
1899: DM dm;
1903: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);
1906: PetscViewerBinaryRead(viewer,&classid,1,NULL,PETSC_INT);
1908: PetscViewerBinaryRead(viewer,type,256,NULL,PETSC_CHAR);
1909: TSSetType(ts, type);
1910: if (ts->ops->load) {
1911: (*ts->ops->load)(ts,viewer);
1912: }
1913: DMCreate(PetscObjectComm((PetscObject)ts),&dm);
1914: DMLoad(dm,viewer);
1915: TSSetDM(ts,dm);
1916: DMCreateGlobalVector(ts->dm,&ts->vec_sol);
1917: VecLoad(ts->vec_sol,viewer);
1918: DMGetDMTS(ts->dm,&sdm);
1919: DMTSLoad(sdm,viewer);
1920: return 0;
1921: }
1923: #include <petscdraw.h>
1924: #if defined(PETSC_HAVE_SAWS)
1925: #include <petscviewersaws.h>
1926: #endif
1928: /*@C
1929: TSViewFromOptions - View from Options
1931: Collective on TS
1933: Input Parameters:
1934: + A - the application ordering context
1935: . obj - Optional object
1936: - name - command line option
1938: Level: intermediate
1939: .seealso: TS, TSView, PetscObjectViewFromOptions(), TSCreate()
1940: @*/
1941: PetscErrorCode TSViewFromOptions(TS A,PetscObject obj,const char name[])
1942: {
1944: PetscObjectViewFromOptions((PetscObject)A,obj,name);
1945: return 0;
1946: }
1948: /*@C
1949: TSView - Prints the TS data structure.
1951: Collective on TS
1953: Input Parameters:
1954: + ts - the TS context obtained from TSCreate()
1955: - viewer - visualization context
1957: Options Database Key:
1958: . -ts_view - calls TSView() at end of TSStep()
1960: Notes:
1961: The available visualization contexts include
1962: + PETSC_VIEWER_STDOUT_SELF - standard output (default)
1963: - PETSC_VIEWER_STDOUT_WORLD - synchronized standard
1964: output where only the first processor opens
1965: the file. All other processors send their
1966: data to the first processor to print.
1968: The user can open an alternative visualization context with
1969: PetscViewerASCIIOpen() - output to a specified file.
1971: In the debugger you can do "call TSView(ts,0)" to display the TS solver. (The same holds for any PETSc object viewer).
1973: Level: beginner
1975: .seealso: PetscViewerASCIIOpen()
1976: @*/
1977: PetscErrorCode TSView(TS ts,PetscViewer viewer)
1978: {
1979: TSType type;
1980: PetscBool iascii,isstring,isundials,isbinary,isdraw;
1981: DMTS sdm;
1982: #if defined(PETSC_HAVE_SAWS)
1983: PetscBool issaws;
1984: #endif
1987: if (!viewer) {
1988: PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)ts),&viewer);
1989: }
1993: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);
1994: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSTRING,&isstring);
1995: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);
1996: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);
1997: #if defined(PETSC_HAVE_SAWS)
1998: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSAWS,&issaws);
1999: #endif
2000: if (iascii) {
2001: PetscObjectPrintClassNamePrefixType((PetscObject)ts,viewer);
2002: if (ts->ops->view) {
2003: PetscViewerASCIIPushTab(viewer);
2004: (*ts->ops->view)(ts,viewer);
2005: PetscViewerASCIIPopTab(viewer);
2006: }
2007: if (ts->max_steps < PETSC_MAX_INT) {
2008: PetscViewerASCIIPrintf(viewer," maximum steps=%D\n",ts->max_steps);
2009: }
2010: if (ts->max_time < PETSC_MAX_REAL) {
2011: PetscViewerASCIIPrintf(viewer," maximum time=%g\n",(double)ts->max_time);
2012: }
2013: if (ts->ifuncs) {
2014: PetscViewerASCIIPrintf(viewer," total number of I function evaluations=%D\n",ts->ifuncs);
2015: }
2016: if (ts->ijacs) {
2017: PetscViewerASCIIPrintf(viewer," total number of I Jacobian evaluations=%D\n",ts->ijacs);
2018: }
2019: if (ts->rhsfuncs) {
2020: PetscViewerASCIIPrintf(viewer," total number of RHS function evaluations=%D\n",ts->rhsfuncs);
2021: }
2022: if (ts->rhsjacs) {
2023: PetscViewerASCIIPrintf(viewer," total number of RHS Jacobian evaluations=%D\n",ts->rhsjacs);
2024: }
2025: if (ts->usessnes) {
2026: PetscBool lin;
2027: if (ts->problem_type == TS_NONLINEAR) {
2028: PetscViewerASCIIPrintf(viewer," total number of nonlinear solver iterations=%D\n",ts->snes_its);
2029: }
2030: PetscViewerASCIIPrintf(viewer," total number of linear solver iterations=%D\n",ts->ksp_its);
2031: PetscObjectTypeCompareAny((PetscObject)ts->snes,&lin,SNESKSPONLY,SNESKSPTRANSPOSEONLY,"");
2032: PetscViewerASCIIPrintf(viewer," total number of %slinear solve failures=%D\n",lin ? "" : "non",ts->num_snes_failures);
2033: }
2034: PetscViewerASCIIPrintf(viewer," total number of rejected steps=%D\n",ts->reject);
2035: if (ts->vrtol) {
2036: PetscViewerASCIIPrintf(viewer," using vector of relative error tolerances, ");
2037: } else {
2038: PetscViewerASCIIPrintf(viewer," using relative error tolerance of %g, ",(double)ts->rtol);
2039: }
2040: if (ts->vatol) {
2041: PetscViewerASCIIPrintf(viewer," using vector of absolute error tolerances\n");
2042: } else {
2043: PetscViewerASCIIPrintf(viewer," using absolute error tolerance of %g\n",(double)ts->atol);
2044: }
2045: PetscViewerASCIIPushTab(viewer);
2046: TSAdaptView(ts->adapt,viewer);
2047: PetscViewerASCIIPopTab(viewer);
2048: } else if (isstring) {
2049: TSGetType(ts,&type);
2050: PetscViewerStringSPrintf(viewer," TSType: %-7.7s",type);
2051: if (ts->ops->view) (*ts->ops->view)(ts,viewer);
2052: } else if (isbinary) {
2053: PetscInt classid = TS_FILE_CLASSID;
2054: MPI_Comm comm;
2055: PetscMPIInt rank;
2056: char type[256];
2058: PetscObjectGetComm((PetscObject)ts,&comm);
2059: MPI_Comm_rank(comm,&rank);
2060: if (rank == 0) {
2061: PetscViewerBinaryWrite(viewer,&classid,1,PETSC_INT);
2062: PetscStrncpy(type,((PetscObject)ts)->type_name,256);
2063: PetscViewerBinaryWrite(viewer,type,256,PETSC_CHAR);
2064: }
2065: if (ts->ops->view) {
2066: (*ts->ops->view)(ts,viewer);
2067: }
2068: if (ts->adapt) TSAdaptView(ts->adapt,viewer);
2069: DMView(ts->dm,viewer);
2070: VecView(ts->vec_sol,viewer);
2071: DMGetDMTS(ts->dm,&sdm);
2072: DMTSView(sdm,viewer);
2073: } else if (isdraw) {
2074: PetscDraw draw;
2075: char str[36];
2076: PetscReal x,y,bottom,h;
2078: PetscViewerDrawGetDraw(viewer,0,&draw);
2079: PetscDrawGetCurrentPoint(draw,&x,&y);
2080: PetscStrcpy(str,"TS: ");
2081: PetscStrcat(str,((PetscObject)ts)->type_name);
2082: PetscDrawStringBoxed(draw,x,y,PETSC_DRAW_BLACK,PETSC_DRAW_BLACK,str,NULL,&h);
2083: bottom = y - h;
2084: PetscDrawPushCurrentPoint(draw,x,bottom);
2085: if (ts->ops->view) {
2086: (*ts->ops->view)(ts,viewer);
2087: }
2088: if (ts->adapt) TSAdaptView(ts->adapt,viewer);
2089: if (ts->snes) SNESView(ts->snes,viewer);
2090: PetscDrawPopCurrentPoint(draw);
2091: #if defined(PETSC_HAVE_SAWS)
2092: } else if (issaws) {
2093: PetscMPIInt rank;
2094: const char *name;
2096: PetscObjectGetName((PetscObject)ts,&name);
2097: MPI_Comm_rank(PETSC_COMM_WORLD,&rank);
2098: if (!((PetscObject)ts)->amsmem && rank == 0) {
2099: char dir[1024];
2101: PetscObjectViewSAWs((PetscObject)ts,viewer);
2102: PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time_step",name);
2103: PetscStackCallSAWs(SAWs_Register,(dir,&ts->steps,1,SAWs_READ,SAWs_INT));
2104: PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time",name);
2105: PetscStackCallSAWs(SAWs_Register,(dir,&ts->ptime,1,SAWs_READ,SAWs_DOUBLE));
2106: }
2107: if (ts->ops->view) {
2108: (*ts->ops->view)(ts,viewer);
2109: }
2110: #endif
2111: }
2112: if (ts->snes && ts->usessnes) {
2113: PetscViewerASCIIPushTab(viewer);
2114: SNESView(ts->snes,viewer);
2115: PetscViewerASCIIPopTab(viewer);
2116: }
2117: DMGetDMTS(ts->dm,&sdm);
2118: DMTSView(sdm,viewer);
2120: PetscViewerASCIIPushTab(viewer);
2121: PetscObjectTypeCompare((PetscObject)ts,TSSUNDIALS,&isundials);
2122: PetscViewerASCIIPopTab(viewer);
2123: return 0;
2124: }
2126: /*@
2127: TSSetApplicationContext - Sets an optional user-defined context for
2128: the timesteppers.
2130: Logically Collective on TS
2132: Input Parameters:
2133: + ts - the TS context obtained from TSCreate()
2134: - usrP - optional user context
2136: Fortran Notes:
2137: To use this from Fortran you must write a Fortran interface definition for this
2138: function that tells Fortran the Fortran derived data type that you are passing in as the ctx argument.
2140: Level: intermediate
2142: .seealso: TSGetApplicationContext()
2143: @*/
2144: PetscErrorCode TSSetApplicationContext(TS ts,void *usrP)
2145: {
2147: ts->user = usrP;
2148: return 0;
2149: }
2151: /*@
2152: TSGetApplicationContext - Gets the user-defined context for the
2153: timestepper.
2155: Not Collective
2157: Input Parameter:
2158: . ts - the TS context obtained from TSCreate()
2160: Output Parameter:
2161: . usrP - user context
2163: Fortran Notes:
2164: To use this from Fortran you must write a Fortran interface definition for this
2165: function that tells Fortran the Fortran derived data type that you are passing in as the ctx argument.
2167: Level: intermediate
2169: .seealso: TSSetApplicationContext()
2170: @*/
2171: PetscErrorCode TSGetApplicationContext(TS ts,void *usrP)
2172: {
2174: *(void**)usrP = ts->user;
2175: return 0;
2176: }
2178: /*@
2179: TSGetStepNumber - Gets the number of steps completed.
2181: Not Collective
2183: Input Parameter:
2184: . ts - the TS context obtained from TSCreate()
2186: Output Parameter:
2187: . steps - number of steps completed so far
2189: Level: intermediate
2191: .seealso: TSGetTime(), TSGetTimeStep(), TSSetPreStep(), TSSetPreStage(), TSSetPostStage(), TSSetPostStep()
2192: @*/
2193: PetscErrorCode TSGetStepNumber(TS ts,PetscInt *steps)
2194: {
2197: *steps = ts->steps;
2198: return 0;
2199: }
2201: /*@
2202: TSSetStepNumber - Sets the number of steps completed.
2204: Logically Collective on TS
2206: Input Parameters:
2207: + ts - the TS context
2208: - steps - number of steps completed so far
2210: Notes:
2211: For most uses of the TS solvers the user need not explicitly call
2212: TSSetStepNumber(), as the step counter is appropriately updated in
2213: TSSolve()/TSStep()/TSRollBack(). Power users may call this routine to
2214: reinitialize timestepping by setting the step counter to zero (and time
2215: to the initial time) to solve a similar problem with different initial
2216: conditions or parameters. Other possible use case is to continue
2217: timestepping from a previously interrupted run in such a way that TS
2218: monitors will be called with a initial nonzero step counter.
2220: Level: advanced
2222: .seealso: TSGetStepNumber(), TSSetTime(), TSSetTimeStep(), TSSetSolution()
2223: @*/
2224: PetscErrorCode TSSetStepNumber(TS ts,PetscInt steps)
2225: {
2229: ts->steps = steps;
2230: return 0;
2231: }
2233: /*@
2234: TSSetTimeStep - Allows one to reset the timestep at any time,
2235: useful for simple pseudo-timestepping codes.
2237: Logically Collective on TS
2239: Input Parameters:
2240: + ts - the TS context obtained from TSCreate()
2241: - time_step - the size of the timestep
2243: Level: intermediate
2245: .seealso: TSGetTimeStep(), TSSetTime()
2247: @*/
2248: PetscErrorCode TSSetTimeStep(TS ts,PetscReal time_step)
2249: {
2252: ts->time_step = time_step;
2253: return 0;
2254: }
2256: /*@
2257: TSSetExactFinalTime - Determines whether to adapt the final time step to
2258: match the exact final time, interpolate solution to the exact final time,
2259: or just return at the final time TS computed.
2261: Logically Collective on TS
2263: Input Parameters:
2264: + ts - the time-step context
2265: - eftopt - exact final time option
2267: $ TS_EXACTFINALTIME_STEPOVER - Don't do anything if final time is exceeded
2268: $ TS_EXACTFINALTIME_INTERPOLATE - Interpolate back to final time
2269: $ TS_EXACTFINALTIME_MATCHSTEP - Adapt final time step to match the final time
2271: Options Database:
2272: . -ts_exact_final_time <stepover,interpolate,matchstep> - select the final step at runtime
2274: Warning: If you use the option TS_EXACTFINALTIME_STEPOVER the solution may be at a very different time
2275: then the final time you selected.
2277: Level: beginner
2279: .seealso: TSExactFinalTimeOption, TSGetExactFinalTime()
2280: @*/
2281: PetscErrorCode TSSetExactFinalTime(TS ts,TSExactFinalTimeOption eftopt)
2282: {
2285: ts->exact_final_time = eftopt;
2286: return 0;
2287: }
2289: /*@
2290: TSGetExactFinalTime - Gets the exact final time option.
2292: Not Collective
2294: Input Parameter:
2295: . ts - the TS context
2297: Output Parameter:
2298: . eftopt - exact final time option
2300: Level: beginner
2302: .seealso: TSExactFinalTimeOption, TSSetExactFinalTime()
2303: @*/
2304: PetscErrorCode TSGetExactFinalTime(TS ts,TSExactFinalTimeOption *eftopt)
2305: {
2308: *eftopt = ts->exact_final_time;
2309: return 0;
2310: }
2312: /*@
2313: TSGetTimeStep - Gets the current timestep size.
2315: Not Collective
2317: Input Parameter:
2318: . ts - the TS context obtained from TSCreate()
2320: Output Parameter:
2321: . dt - the current timestep size
2323: Level: intermediate
2325: .seealso: TSSetTimeStep(), TSGetTime()
2327: @*/
2328: PetscErrorCode TSGetTimeStep(TS ts,PetscReal *dt)
2329: {
2332: *dt = ts->time_step;
2333: return 0;
2334: }
2336: /*@
2337: TSGetSolution - Returns the solution at the present timestep. It
2338: is valid to call this routine inside the function that you are evaluating
2339: in order to move to the new timestep. This vector not changed until
2340: the solution at the next timestep has been calculated.
2342: Not Collective, but Vec returned is parallel if TS is parallel
2344: Input Parameter:
2345: . ts - the TS context obtained from TSCreate()
2347: Output Parameter:
2348: . v - the vector containing the solution
2350: Note: If you used TSSetExactFinalTime(ts,TS_EXACTFINALTIME_MATCHSTEP); this does not return the solution at the requested
2351: final time. It returns the solution at the next timestep.
2353: Level: intermediate
2355: .seealso: TSGetTimeStep(), TSGetTime(), TSGetSolveTime(), TSGetSolutionComponents(), TSSetSolutionFunction()
2357: @*/
2358: PetscErrorCode TSGetSolution(TS ts,Vec *v)
2359: {
2362: *v = ts->vec_sol;
2363: return 0;
2364: }
2366: /*@
2367: TSGetSolutionComponents - Returns any solution components at the present
2368: timestep, if available for the time integration method being used.
2369: Solution components are quantities that share the same size and
2370: structure as the solution vector.
2372: Not Collective, but Vec returned is parallel if TS is parallel
2374: Parameters :
2375: + ts - the TS context obtained from TSCreate() (input parameter).
2376: . n - If v is PETSC_NULL, then the number of solution components is
2377: returned through n, else the n-th solution component is
2378: returned in v.
2379: - v - the vector containing the n-th solution component
2380: (may be PETSC_NULL to use this function to find out
2381: the number of solutions components).
2383: Level: advanced
2385: .seealso: TSGetSolution()
2387: @*/
2388: PetscErrorCode TSGetSolutionComponents(TS ts,PetscInt *n,Vec *v)
2389: {
2391: if (!ts->ops->getsolutioncomponents) *n = 0;
2392: else {
2393: (*ts->ops->getsolutioncomponents)(ts,n,v);
2394: }
2395: return 0;
2396: }
2398: /*@
2399: TSGetAuxSolution - Returns an auxiliary solution at the present
2400: timestep, if available for the time integration method being used.
2402: Not Collective, but Vec returned is parallel if TS is parallel
2404: Parameters :
2405: + ts - the TS context obtained from TSCreate() (input parameter).
2406: - v - the vector containing the auxiliary solution
2408: Level: intermediate
2410: .seealso: TSGetSolution()
2412: @*/
2413: PetscErrorCode TSGetAuxSolution(TS ts,Vec *v)
2414: {
2416: if (ts->ops->getauxsolution) {
2417: (*ts->ops->getauxsolution)(ts,v);
2418: } else {
2419: VecZeroEntries(*v);
2420: }
2421: return 0;
2422: }
2424: /*@
2425: TSGetTimeError - Returns the estimated error vector, if the chosen
2426: TSType has an error estimation functionality.
2428: Not Collective, but Vec returned is parallel if TS is parallel
2430: Note: MUST call after TSSetUp()
2432: Parameters :
2433: + ts - the TS context obtained from TSCreate() (input parameter).
2434: . n - current estimate (n=0) or previous one (n=-1)
2435: - v - the vector containing the error (same size as the solution).
2437: Level: intermediate
2439: .seealso: TSGetSolution(), TSSetTimeError()
2441: @*/
2442: PetscErrorCode TSGetTimeError(TS ts,PetscInt n,Vec *v)
2443: {
2445: if (ts->ops->gettimeerror) {
2446: (*ts->ops->gettimeerror)(ts,n,v);
2447: } else {
2448: VecZeroEntries(*v);
2449: }
2450: return 0;
2451: }
2453: /*@
2454: TSSetTimeError - Sets the estimated error vector, if the chosen
2455: TSType has an error estimation functionality. This can be used
2456: to restart such a time integrator with a given error vector.
2458: Not Collective, but Vec returned is parallel if TS is parallel
2460: Parameters :
2461: + ts - the TS context obtained from TSCreate() (input parameter).
2462: - v - the vector containing the error (same size as the solution).
2464: Level: intermediate
2466: .seealso: TSSetSolution(), TSGetTimeError)
2468: @*/
2469: PetscErrorCode TSSetTimeError(TS ts,Vec v)
2470: {
2473: if (ts->ops->settimeerror) {
2474: (*ts->ops->settimeerror)(ts,v);
2475: }
2476: return 0;
2477: }
2479: /* ----- Routines to initialize and destroy a timestepper ---- */
2480: /*@
2481: TSSetProblemType - Sets the type of problem to be solved.
2483: Not collective
2485: Input Parameters:
2486: + ts - The TS
2487: - type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms
2488: .vb
2489: U_t - A U = 0 (linear)
2490: U_t - A(t) U = 0 (linear)
2491: F(t,U,U_t) = 0 (nonlinear)
2492: .ve
2494: Level: beginner
2496: .seealso: TSSetUp(), TSProblemType, TS
2497: @*/
2498: PetscErrorCode TSSetProblemType(TS ts, TSProblemType type)
2499: {
2501: ts->problem_type = type;
2502: if (type == TS_LINEAR) {
2503: SNES snes;
2504: TSGetSNES(ts,&snes);
2505: SNESSetType(snes,SNESKSPONLY);
2506: }
2507: return 0;
2508: }
2510: /*@C
2511: TSGetProblemType - Gets the type of problem to be solved.
2513: Not collective
2515: Input Parameter:
2516: . ts - The TS
2518: Output Parameter:
2519: . type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms
2520: .vb
2521: M U_t = A U
2522: M(t) U_t = A(t) U
2523: F(t,U,U_t)
2524: .ve
2526: Level: beginner
2528: .seealso: TSSetUp(), TSProblemType, TS
2529: @*/
2530: PetscErrorCode TSGetProblemType(TS ts, TSProblemType *type)
2531: {
2534: *type = ts->problem_type;
2535: return 0;
2536: }
2538: /*
2539: Attempt to check/preset a default value for the exact final time option. This is needed at the beginning of TSSolve() and in TSSetUp()
2540: */
2541: static PetscErrorCode TSSetExactFinalTimeDefault(TS ts)
2542: {
2543: PetscBool isnone;
2545: TSGetAdapt(ts,&ts->adapt);
2546: TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type);
2548: PetscObjectTypeCompare((PetscObject)ts->adapt,TSADAPTNONE,&isnone);
2549: if (!isnone && ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) {
2550: ts->exact_final_time = TS_EXACTFINALTIME_MATCHSTEP;
2551: } else if (ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) {
2552: ts->exact_final_time = TS_EXACTFINALTIME_INTERPOLATE;
2553: }
2554: return 0;
2555: }
2557: /*@
2558: TSSetUp - Sets up the internal data structures for the later use of a timestepper.
2560: Collective on TS
2562: Input Parameter:
2563: . ts - the TS context obtained from TSCreate()
2565: Notes:
2566: For basic use of the TS solvers the user need not explicitly call
2567: TSSetUp(), since these actions will automatically occur during
2568: the call to TSStep() or TSSolve(). However, if one wishes to control this
2569: phase separately, TSSetUp() should be called after TSCreate()
2570: and optional routines of the form TSSetXXX(), but before TSStep() and TSSolve().
2572: Level: advanced
2574: .seealso: TSCreate(), TSStep(), TSDestroy(), TSSolve()
2575: @*/
2576: PetscErrorCode TSSetUp(TS ts)
2577: {
2578: DM dm;
2579: PetscErrorCode (*func)(SNES,Vec,Vec,void*);
2580: PetscErrorCode (*jac)(SNES,Vec,Mat,Mat,void*);
2581: TSIFunction ifun;
2582: TSIJacobian ijac;
2583: TSI2Jacobian i2jac;
2584: TSRHSJacobian rhsjac;
2587: if (ts->setupcalled) return 0;
2589: if (!((PetscObject)ts)->type_name) {
2590: TSGetIFunction(ts,NULL,&ifun,NULL);
2591: TSSetType(ts,ifun ? TSBEULER : TSEULER);
2592: }
2594: if (!ts->vec_sol) {
2595: if (ts->dm) {
2596: DMCreateGlobalVector(ts->dm,&ts->vec_sol);
2597: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call TSSetSolution() first");
2598: }
2600: if (!ts->Jacp && ts->Jacprhs) { /* IJacobianP shares the same matrix with RHSJacobianP if only RHSJacobianP is provided */
2601: PetscObjectReference((PetscObject)ts->Jacprhs);
2602: ts->Jacp = ts->Jacprhs;
2603: }
2605: if (ts->quadraturets) {
2606: TSSetUp(ts->quadraturets);
2607: VecDestroy(&ts->vec_costintegrand);
2608: VecDuplicate(ts->quadraturets->vec_sol,&ts->vec_costintegrand);
2609: }
2611: TSGetRHSJacobian(ts,NULL,NULL,&rhsjac,NULL);
2612: if (rhsjac == TSComputeRHSJacobianConstant) {
2613: Mat Amat,Pmat;
2614: SNES snes;
2615: TSGetSNES(ts,&snes);
2616: SNESGetJacobian(snes,&Amat,&Pmat,NULL,NULL);
2617: /* Matching matrices implies that an IJacobian is NOT set, because if it had been set, the IJacobian's matrix would
2618: * have displaced the RHS matrix */
2619: if (Amat && Amat == ts->Arhs) {
2620: /* we need to copy the values of the matrix because for the constant Jacobian case the user will never set the numerical values in this new location */
2621: MatDuplicate(ts->Arhs,MAT_COPY_VALUES,&Amat);
2622: SNESSetJacobian(snes,Amat,NULL,NULL,NULL);
2623: MatDestroy(&Amat);
2624: }
2625: if (Pmat && Pmat == ts->Brhs) {
2626: MatDuplicate(ts->Brhs,MAT_COPY_VALUES,&Pmat);
2627: SNESSetJacobian(snes,NULL,Pmat,NULL,NULL);
2628: MatDestroy(&Pmat);
2629: }
2630: }
2632: TSGetAdapt(ts,&ts->adapt);
2633: TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type);
2635: if (ts->ops->setup) {
2636: (*ts->ops->setup)(ts);
2637: }
2639: TSSetExactFinalTimeDefault(ts);
2641: /* In the case where we've set a DMTSFunction or what have you, we need the default SNESFunction
2642: to be set right but can't do it elsewhere due to the overreliance on ctx=ts.
2643: */
2644: TSGetDM(ts,&dm);
2645: DMSNESGetFunction(dm,&func,NULL);
2646: if (!func) {
2647: DMSNESSetFunction(dm,SNESTSFormFunction,ts);
2648: }
2649: /* If the SNES doesn't have a jacobian set and the TS has an ijacobian or rhsjacobian set, set the SNES to use it.
2650: Otherwise, the SNES will use coloring internally to form the Jacobian.
2651: */
2652: DMSNESGetJacobian(dm,&jac,NULL);
2653: DMTSGetIJacobian(dm,&ijac,NULL);
2654: DMTSGetI2Jacobian(dm,&i2jac,NULL);
2655: DMTSGetRHSJacobian(dm,&rhsjac,NULL);
2656: if (!jac && (ijac || i2jac || rhsjac)) {
2657: DMSNESSetJacobian(dm,SNESTSFormJacobian,ts);
2658: }
2660: /* if time integration scheme has a starting method, call it */
2661: if (ts->ops->startingmethod) {
2662: (*ts->ops->startingmethod)(ts);
2663: }
2665: ts->setupcalled = PETSC_TRUE;
2666: return 0;
2667: }
2669: /*@
2670: TSReset - Resets a TS context and removes any allocated Vecs and Mats.
2672: Collective on TS
2674: Input Parameter:
2675: . ts - the TS context obtained from TSCreate()
2677: Level: beginner
2679: .seealso: TSCreate(), TSSetup(), TSDestroy()
2680: @*/
2681: PetscErrorCode TSReset(TS ts)
2682: {
2683: TS_RHSSplitLink ilink = ts->tsrhssplit,next;
2687: if (ts->ops->reset) {
2688: (*ts->ops->reset)(ts);
2689: }
2690: if (ts->snes) SNESReset(ts->snes);
2691: if (ts->adapt) TSAdaptReset(ts->adapt);
2693: MatDestroy(&ts->Arhs);
2694: MatDestroy(&ts->Brhs);
2695: VecDestroy(&ts->Frhs);
2696: VecDestroy(&ts->vec_sol);
2697: VecDestroy(&ts->vec_dot);
2698: VecDestroy(&ts->vatol);
2699: VecDestroy(&ts->vrtol);
2700: VecDestroyVecs(ts->nwork,&ts->work);
2702: MatDestroy(&ts->Jacprhs);
2703: MatDestroy(&ts->Jacp);
2704: if (ts->forward_solve) {
2705: TSForwardReset(ts);
2706: }
2707: if (ts->quadraturets) {
2708: TSReset(ts->quadraturets);
2709: VecDestroy(&ts->vec_costintegrand);
2710: }
2711: while (ilink) {
2712: next = ilink->next;
2713: TSDestroy(&ilink->ts);
2714: PetscFree(ilink->splitname);
2715: ISDestroy(&ilink->is);
2716: PetscFree(ilink);
2717: ilink = next;
2718: }
2719: ts->tsrhssplit = NULL;
2720: ts->num_rhs_splits = 0;
2721: ts->setupcalled = PETSC_FALSE;
2722: return 0;
2723: }
2725: /*@C
2726: TSDestroy - Destroys the timestepper context that was created
2727: with TSCreate().
2729: Collective on TS
2731: Input Parameter:
2732: . ts - the TS context obtained from TSCreate()
2734: Level: beginner
2736: .seealso: TSCreate(), TSSetUp(), TSSolve()
2737: @*/
2738: PetscErrorCode TSDestroy(TS *ts)
2739: {
2740: if (!*ts) return 0;
2742: if (--((PetscObject)(*ts))->refct > 0) {*ts = NULL; return 0;}
2744: TSReset(*ts);
2745: TSAdjointReset(*ts);
2746: if ((*ts)->forward_solve) {
2747: TSForwardReset(*ts);
2748: }
2749: /* if memory was published with SAWs then destroy it */
2750: PetscObjectSAWsViewOff((PetscObject)*ts);
2751: if ((*ts)->ops->destroy) (*(*ts)->ops->destroy)((*ts));
2753: TSTrajectoryDestroy(&(*ts)->trajectory);
2755: TSAdaptDestroy(&(*ts)->adapt);
2756: TSEventDestroy(&(*ts)->event);
2758: SNESDestroy(&(*ts)->snes);
2759: DMDestroy(&(*ts)->dm);
2760: TSMonitorCancel((*ts));
2761: TSAdjointMonitorCancel((*ts));
2763: TSDestroy(&(*ts)->quadraturets);
2764: PetscHeaderDestroy(ts);
2765: return 0;
2766: }
2768: /*@
2769: TSGetSNES - Returns the SNES (nonlinear solver) associated with
2770: a TS (timestepper) context. Valid only for nonlinear problems.
2772: Not Collective, but SNES is parallel if TS is parallel
2774: Input Parameter:
2775: . ts - the TS context obtained from TSCreate()
2777: Output Parameter:
2778: . snes - the nonlinear solver context
2780: Notes:
2781: The user can then directly manipulate the SNES context to set various
2782: options, etc. Likewise, the user can then extract and manipulate the
2783: KSP, KSP, and PC contexts as well.
2785: TSGetSNES() does not work for integrators that do not use SNES; in
2786: this case TSGetSNES() returns NULL in snes.
2788: Level: beginner
2790: @*/
2791: PetscErrorCode TSGetSNES(TS ts,SNES *snes)
2792: {
2795: if (!ts->snes) {
2796: SNESCreate(PetscObjectComm((PetscObject)ts),&ts->snes);
2797: PetscObjectSetOptions((PetscObject)ts->snes,((PetscObject)ts)->options);
2798: SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts);
2799: PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->snes);
2800: PetscObjectIncrementTabLevel((PetscObject)ts->snes,(PetscObject)ts,1);
2801: if (ts->dm) SNESSetDM(ts->snes,ts->dm);
2802: if (ts->problem_type == TS_LINEAR) {
2803: SNESSetType(ts->snes,SNESKSPONLY);
2804: }
2805: }
2806: *snes = ts->snes;
2807: return 0;
2808: }
2810: /*@
2811: TSSetSNES - Set the SNES (nonlinear solver) to be used by the timestepping context
2813: Collective
2815: Input Parameters:
2816: + ts - the TS context obtained from TSCreate()
2817: - snes - the nonlinear solver context
2819: Notes:
2820: Most users should have the TS created by calling TSGetSNES()
2822: Level: developer
2824: @*/
2825: PetscErrorCode TSSetSNES(TS ts,SNES snes)
2826: {
2827: PetscErrorCode (*func)(SNES,Vec,Mat,Mat,void*);
2831: PetscObjectReference((PetscObject)snes);
2832: SNESDestroy(&ts->snes);
2834: ts->snes = snes;
2836: SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts);
2837: SNESGetJacobian(ts->snes,NULL,NULL,&func,NULL);
2838: if (func == SNESTSFormJacobian) {
2839: SNESSetJacobian(ts->snes,NULL,NULL,SNESTSFormJacobian,ts);
2840: }
2841: return 0;
2842: }
2844: /*@
2845: TSGetKSP - Returns the KSP (linear solver) associated with
2846: a TS (timestepper) context.
2848: Not Collective, but KSP is parallel if TS is parallel
2850: Input Parameter:
2851: . ts - the TS context obtained from TSCreate()
2853: Output Parameter:
2854: . ksp - the nonlinear solver context
2856: Notes:
2857: The user can then directly manipulate the KSP context to set various
2858: options, etc. Likewise, the user can then extract and manipulate the
2859: KSP and PC contexts as well.
2861: TSGetKSP() does not work for integrators that do not use KSP;
2862: in this case TSGetKSP() returns NULL in ksp.
2864: Level: beginner
2866: @*/
2867: PetscErrorCode TSGetKSP(TS ts,KSP *ksp)
2868: {
2869: SNES snes;
2875: TSGetSNES(ts,&snes);
2876: SNESGetKSP(snes,ksp);
2877: return 0;
2878: }
2880: /* ----------- Routines to set solver parameters ---------- */
2882: /*@
2883: TSSetMaxSteps - Sets the maximum number of steps to use.
2885: Logically Collective on TS
2887: Input Parameters:
2888: + ts - the TS context obtained from TSCreate()
2889: - maxsteps - maximum number of steps to use
2891: Options Database Keys:
2892: . -ts_max_steps <maxsteps> - Sets maxsteps
2894: Notes:
2895: The default maximum number of steps is 5000
2897: Level: intermediate
2899: .seealso: TSGetMaxSteps(), TSSetMaxTime(), TSSetExactFinalTime()
2900: @*/
2901: PetscErrorCode TSSetMaxSteps(TS ts,PetscInt maxsteps)
2902: {
2906: ts->max_steps = maxsteps;
2907: return 0;
2908: }
2910: /*@
2911: TSGetMaxSteps - Gets the maximum number of steps to use.
2913: Not Collective
2915: Input Parameters:
2916: . ts - the TS context obtained from TSCreate()
2918: Output Parameter:
2919: . maxsteps - maximum number of steps to use
2921: Level: advanced
2923: .seealso: TSSetMaxSteps(), TSGetMaxTime(), TSSetMaxTime()
2924: @*/
2925: PetscErrorCode TSGetMaxSteps(TS ts,PetscInt *maxsteps)
2926: {
2929: *maxsteps = ts->max_steps;
2930: return 0;
2931: }
2933: /*@
2934: TSSetMaxTime - Sets the maximum (or final) time for timestepping.
2936: Logically Collective on TS
2938: Input Parameters:
2939: + ts - the TS context obtained from TSCreate()
2940: - maxtime - final time to step to
2942: Options Database Keys:
2943: . -ts_max_time <maxtime> - Sets maxtime
2945: Notes:
2946: The default maximum time is 5.0
2948: Level: intermediate
2950: .seealso: TSGetMaxTime(), TSSetMaxSteps(), TSSetExactFinalTime()
2951: @*/
2952: PetscErrorCode TSSetMaxTime(TS ts,PetscReal maxtime)
2953: {
2956: ts->max_time = maxtime;
2957: return 0;
2958: }
2960: /*@
2961: TSGetMaxTime - Gets the maximum (or final) time for timestepping.
2963: Not Collective
2965: Input Parameters:
2966: . ts - the TS context obtained from TSCreate()
2968: Output Parameter:
2969: . maxtime - final time to step to
2971: Level: advanced
2973: .seealso: TSSetMaxTime(), TSGetMaxSteps(), TSSetMaxSteps()
2974: @*/
2975: PetscErrorCode TSGetMaxTime(TS ts,PetscReal *maxtime)
2976: {
2979: *maxtime = ts->max_time;
2980: return 0;
2981: }
2983: /*@
2984: TSSetInitialTimeStep - Deprecated, use TSSetTime() and TSSetTimeStep().
2986: Level: deprecated
2988: @*/
2989: PetscErrorCode TSSetInitialTimeStep(TS ts,PetscReal initial_time,PetscReal time_step)
2990: {
2992: TSSetTime(ts,initial_time);
2993: TSSetTimeStep(ts,time_step);
2994: return 0;
2995: }
2997: /*@
2998: TSGetDuration - Deprecated, use TSGetMaxSteps() and TSGetMaxTime().
3000: Level: deprecated
3002: @*/
3003: PetscErrorCode TSGetDuration(TS ts, PetscInt *maxsteps, PetscReal *maxtime)
3004: {
3006: if (maxsteps) {
3008: *maxsteps = ts->max_steps;
3009: }
3010: if (maxtime) {
3012: *maxtime = ts->max_time;
3013: }
3014: return 0;
3015: }
3017: /*@
3018: TSSetDuration - Deprecated, use TSSetMaxSteps() and TSSetMaxTime().
3020: Level: deprecated
3022: @*/
3023: PetscErrorCode TSSetDuration(TS ts,PetscInt maxsteps,PetscReal maxtime)
3024: {
3028: if (maxsteps >= 0) ts->max_steps = maxsteps;
3029: if (maxtime != PETSC_DEFAULT) ts->max_time = maxtime;
3030: return 0;
3031: }
3033: /*@
3034: TSGetTimeStepNumber - Deprecated, use TSGetStepNumber().
3036: Level: deprecated
3038: @*/
3039: PetscErrorCode TSGetTimeStepNumber(TS ts,PetscInt *steps) { return TSGetStepNumber(ts,steps); }
3041: /*@
3042: TSGetTotalSteps - Deprecated, use TSGetStepNumber().
3044: Level: deprecated
3046: @*/
3047: PetscErrorCode TSGetTotalSteps(TS ts,PetscInt *steps) { return TSGetStepNumber(ts,steps); }
3049: /*@
3050: TSSetSolution - Sets the initial solution vector
3051: for use by the TS routines.
3053: Logically Collective on TS
3055: Input Parameters:
3056: + ts - the TS context obtained from TSCreate()
3057: - u - the solution vector
3059: Level: beginner
3061: .seealso: TSSetSolutionFunction(), TSGetSolution(), TSCreate()
3062: @*/
3063: PetscErrorCode TSSetSolution(TS ts,Vec u)
3064: {
3065: DM dm;
3069: PetscObjectReference((PetscObject)u);
3070: VecDestroy(&ts->vec_sol);
3071: ts->vec_sol = u;
3073: TSGetDM(ts,&dm);
3074: DMShellSetGlobalVector(dm,u);
3075: return 0;
3076: }
3078: /*@C
3079: TSSetPreStep - Sets the general-purpose function
3080: called once at the beginning of each time step.
3082: Logically Collective on TS
3084: Input Parameters:
3085: + ts - The TS context obtained from TSCreate()
3086: - func - The function
3088: Calling sequence of func:
3089: .vb
3090: PetscErrorCode func (TS ts);
3091: .ve
3093: Level: intermediate
3095: .seealso: TSSetPreStage(), TSSetPostStage(), TSSetPostStep(), TSStep(), TSRestartStep()
3096: @*/
3097: PetscErrorCode TSSetPreStep(TS ts, PetscErrorCode (*func)(TS))
3098: {
3100: ts->prestep = func;
3101: return 0;
3102: }
3104: /*@
3105: TSPreStep - Runs the user-defined pre-step function.
3107: Collective on TS
3109: Input Parameters:
3110: . ts - The TS context obtained from TSCreate()
3112: Notes:
3113: TSPreStep() is typically used within time stepping implementations,
3114: so most users would not generally call this routine themselves.
3116: Level: developer
3118: .seealso: TSSetPreStep(), TSPreStage(), TSPostStage(), TSPostStep()
3119: @*/
3120: PetscErrorCode TSPreStep(TS ts)
3121: {
3123: if (ts->prestep) {
3124: Vec U;
3125: PetscObjectId idprev;
3126: PetscBool sameObject;
3127: PetscObjectState sprev,spost;
3129: TSGetSolution(ts,&U);
3130: PetscObjectGetId((PetscObject)U,&idprev);
3131: PetscObjectStateGet((PetscObject)U,&sprev);
3132: PetscStackCallStandard((*ts->prestep),ts);
3133: TSGetSolution(ts,&U);
3134: PetscObjectCompareId((PetscObject)U,idprev,&sameObject);
3135: PetscObjectStateGet((PetscObject)U,&spost);
3136: if (!sameObject || sprev != spost) TSRestartStep(ts);
3137: }
3138: return 0;
3139: }
3141: /*@C
3142: TSSetPreStage - Sets the general-purpose function
3143: called once at the beginning of each stage.
3145: Logically Collective on TS
3147: Input Parameters:
3148: + ts - The TS context obtained from TSCreate()
3149: - func - The function
3151: Calling sequence of func:
3152: .vb
3153: PetscErrorCode func(TS ts, PetscReal stagetime);
3154: .ve
3156: Level: intermediate
3158: Note:
3159: There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3160: The time step number being computed can be queried using TSGetStepNumber() and the total size of the step being
3161: attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime().
3163: .seealso: TSSetPostStage(), TSSetPreStep(), TSSetPostStep(), TSGetApplicationContext()
3164: @*/
3165: PetscErrorCode TSSetPreStage(TS ts, PetscErrorCode (*func)(TS,PetscReal))
3166: {
3168: ts->prestage = func;
3169: return 0;
3170: }
3172: /*@C
3173: TSSetPostStage - Sets the general-purpose function
3174: called once at the end of each stage.
3176: Logically Collective on TS
3178: Input Parameters:
3179: + ts - The TS context obtained from TSCreate()
3180: - func - The function
3182: Calling sequence of func:
3183: .vb
3184: PetscErrorCode func(TS ts, PetscReal stagetime, PetscInt stageindex, Vec* Y);
3185: .ve
3187: Level: intermediate
3189: Note:
3190: There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3191: The time step number being computed can be queried using TSGetStepNumber() and the total size of the step being
3192: attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime().
3194: .seealso: TSSetPreStage(), TSSetPreStep(), TSSetPostStep(), TSGetApplicationContext()
3195: @*/
3196: PetscErrorCode TSSetPostStage(TS ts, PetscErrorCode (*func)(TS,PetscReal,PetscInt,Vec*))
3197: {
3199: ts->poststage = func;
3200: return 0;
3201: }
3203: /*@C
3204: TSSetPostEvaluate - Sets the general-purpose function
3205: called once at the end of each step evaluation.
3207: Logically Collective on TS
3209: Input Parameters:
3210: + ts - The TS context obtained from TSCreate()
3211: - func - The function
3213: Calling sequence of func:
3214: .vb
3215: PetscErrorCode func(TS ts);
3216: .ve
3218: Level: intermediate
3220: Note:
3221: Semantically, TSSetPostEvaluate() differs from TSSetPostStep() since the function it sets is called before event-handling
3222: thus guaranteeing the same solution (computed by the time-stepper) will be passed to it. On the other hand, TSPostStep()
3223: may be passed a different solution, possibly changed by the event handler. TSPostEvaluate() is called after the next step
3224: solution is evaluated allowing to modify it, if need be. The solution can be obtained with TSGetSolution(), the time step
3225: with TSGetTimeStep(), and the time at the start of the step is available via TSGetTime()
3227: .seealso: TSSetPreStage(), TSSetPreStep(), TSSetPostStep(), TSGetApplicationContext()
3228: @*/
3229: PetscErrorCode TSSetPostEvaluate(TS ts, PetscErrorCode (*func)(TS))
3230: {
3232: ts->postevaluate = func;
3233: return 0;
3234: }
3236: /*@
3237: TSPreStage - Runs the user-defined pre-stage function set using TSSetPreStage()
3239: Collective on TS
3241: Input Parameters:
3242: . ts - The TS context obtained from TSCreate()
3243: stagetime - The absolute time of the current stage
3245: Notes:
3246: TSPreStage() is typically used within time stepping implementations,
3247: most users would not generally call this routine themselves.
3249: Level: developer
3251: .seealso: TSPostStage(), TSSetPreStep(), TSPreStep(), TSPostStep()
3252: @*/
3253: PetscErrorCode TSPreStage(TS ts, PetscReal stagetime)
3254: {
3256: if (ts->prestage) {
3257: PetscStackCallStandard((*ts->prestage),ts,stagetime);
3258: }
3259: return 0;
3260: }
3262: /*@
3263: TSPostStage - Runs the user-defined post-stage function set using TSSetPostStage()
3265: Collective on TS
3267: Input Parameters:
3268: . ts - The TS context obtained from TSCreate()
3269: stagetime - The absolute time of the current stage
3270: stageindex - Stage number
3271: Y - Array of vectors (of size = total number
3272: of stages) with the stage solutions
3274: Notes:
3275: TSPostStage() is typically used within time stepping implementations,
3276: most users would not generally call this routine themselves.
3278: Level: developer
3280: .seealso: TSPreStage(), TSSetPreStep(), TSPreStep(), TSPostStep()
3281: @*/
3282: PetscErrorCode TSPostStage(TS ts, PetscReal stagetime, PetscInt stageindex, Vec *Y)
3283: {
3285: if (ts->poststage) {
3286: PetscStackCallStandard((*ts->poststage),ts,stagetime,stageindex,Y);
3287: }
3288: return 0;
3289: }
3291: /*@
3292: TSPostEvaluate - Runs the user-defined post-evaluate function set using TSSetPostEvaluate()
3294: Collective on TS
3296: Input Parameters:
3297: . ts - The TS context obtained from TSCreate()
3299: Notes:
3300: TSPostEvaluate() is typically used within time stepping implementations,
3301: most users would not generally call this routine themselves.
3303: Level: developer
3305: .seealso: TSSetPostEvaluate(), TSSetPreStep(), TSPreStep(), TSPostStep()
3306: @*/
3307: PetscErrorCode TSPostEvaluate(TS ts)
3308: {
3310: if (ts->postevaluate) {
3311: Vec U;
3312: PetscObjectState sprev,spost;
3314: TSGetSolution(ts,&U);
3315: PetscObjectStateGet((PetscObject)U,&sprev);
3316: PetscStackCallStandard((*ts->postevaluate),ts);
3317: PetscObjectStateGet((PetscObject)U,&spost);
3318: if (sprev != spost) TSRestartStep(ts);
3319: }
3320: return 0;
3321: }
3323: /*@C
3324: TSSetPostStep - Sets the general-purpose function
3325: called once at the end of each time step.
3327: Logically Collective on TS
3329: Input Parameters:
3330: + ts - The TS context obtained from TSCreate()
3331: - func - The function
3333: Calling sequence of func:
3334: $ func (TS ts);
3336: Notes:
3337: The function set by TSSetPostStep() is called after each successful step. The solution vector X
3338: obtained by TSGetSolution() may be different than that computed at the step end if the event handler
3339: locates an event and TSPostEvent() modifies it. Use TSSetPostEvaluate() if an unmodified solution is needed instead.
3341: Level: intermediate
3343: .seealso: TSSetPreStep(), TSSetPreStage(), TSSetPostEvaluate(), TSGetTimeStep(), TSGetStepNumber(), TSGetTime(), TSRestartStep()
3344: @*/
3345: PetscErrorCode TSSetPostStep(TS ts, PetscErrorCode (*func)(TS))
3346: {
3348: ts->poststep = func;
3349: return 0;
3350: }
3352: /*@
3353: TSPostStep - Runs the user-defined post-step function.
3355: Collective on TS
3357: Input Parameters:
3358: . ts - The TS context obtained from TSCreate()
3360: Notes:
3361: TSPostStep() is typically used within time stepping implementations,
3362: so most users would not generally call this routine themselves.
3364: Level: developer
3366: @*/
3367: PetscErrorCode TSPostStep(TS ts)
3368: {
3370: if (ts->poststep) {
3371: Vec U;
3372: PetscObjectId idprev;
3373: PetscBool sameObject;
3374: PetscObjectState sprev,spost;
3376: TSGetSolution(ts,&U);
3377: PetscObjectGetId((PetscObject)U,&idprev);
3378: PetscObjectStateGet((PetscObject)U,&sprev);
3379: PetscStackCallStandard((*ts->poststep),ts);
3380: TSGetSolution(ts,&U);
3381: PetscObjectCompareId((PetscObject)U,idprev,&sameObject);
3382: PetscObjectStateGet((PetscObject)U,&spost);
3383: if (!sameObject || sprev != spost) TSRestartStep(ts);
3384: }
3385: return 0;
3386: }
3388: /*@
3389: TSInterpolate - Interpolate the solution computed during the previous step to an arbitrary location in the interval
3391: Collective on TS
3393: Input Parameters:
3394: + ts - time stepping context
3395: - t - time to interpolate to
3397: Output Parameter:
3398: . U - state at given time
3400: Level: intermediate
3402: Developer Notes:
3403: TSInterpolate() and the storing of previous steps/stages should be generalized to support delay differential equations and continuous adjoints.
3405: .seealso: TSSetExactFinalTime(), TSSolve()
3406: @*/
3407: PetscErrorCode TSInterpolate(TS ts,PetscReal t,Vec U)
3408: {
3413: (*ts->ops->interpolate)(ts,t,U);
3414: return 0;
3415: }
3417: /*@
3418: TSStep - Steps one time step
3420: Collective on TS
3422: Input Parameter:
3423: . ts - the TS context obtained from TSCreate()
3425: Level: developer
3427: Notes:
3428: The public interface for the ODE/DAE solvers is TSSolve(), you should almost for sure be using that routine and not this routine.
3430: The hook set using TSSetPreStep() is called before each attempt to take the step. In general, the time step size may
3431: be changed due to adaptive error controller or solve failures. Note that steps may contain multiple stages.
3433: This may over-step the final time provided in TSSetMaxTime() depending on the time-step used. TSSolve() interpolates to exactly the
3434: time provided in TSSetMaxTime(). One can use TSInterpolate() to determine an interpolated solution within the final timestep.
3436: .seealso: TSCreate(), TSSetUp(), TSDestroy(), TSSolve(), TSSetPreStep(), TSSetPreStage(), TSSetPostStage(), TSInterpolate()
3437: @*/
3438: PetscErrorCode TSStep(TS ts)
3439: {
3440: PetscErrorCode ierr;
3441: static PetscBool cite = PETSC_FALSE;
3442: PetscReal ptime;
3445: PetscCitationsRegister("@article{tspaper,\n"
3446: " title = {{PETSc/TS}: A Modern Scalable {DAE/ODE} Solver Library},\n"
3447: " author = {Abhyankar, Shrirang and Brown, Jed and Constantinescu, Emil and Ghosh, Debojyoti and Smith, Barry F. and Zhang, Hong},\n"
3448: " journal = {arXiv e-preprints},\n"
3449: " eprint = {1806.01437},\n"
3450: " archivePrefix = {arXiv},\n"
3451: " year = {2018}\n}\n",&cite);
3453: TSSetUp(ts);
3454: TSTrajectorySetUp(ts->trajectory,ts);
3461: if (!ts->steps) ts->ptime_prev = ts->ptime;
3462: ptime = ts->ptime; ts->ptime_prev_rollback = ts->ptime_prev;
3463: ts->reason = TS_CONVERGED_ITERATING;
3465: PetscLogEventBegin(TS_Step,ts,0,0,0);
3466: (*ts->ops->step)(ts);
3467: PetscLogEventEnd(TS_Step,ts,0,0,0);
3469: if (ts->reason >= 0) {
3470: ts->ptime_prev = ptime;
3471: ts->steps++;
3472: ts->steprollback = PETSC_FALSE;
3473: ts->steprestart = PETSC_FALSE;
3474: }
3476: if (!ts->reason) {
3477: if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
3478: else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
3479: }
3483: return 0;
3484: }
3486: /*@
3487: TSEvaluateWLTE - Evaluate the weighted local truncation error norm
3488: at the end of a time step with a given order of accuracy.
3490: Collective on TS
3492: Input Parameters:
3493: + ts - time stepping context
3494: - wnormtype - norm type, either NORM_2 or NORM_INFINITY
3496: Input/Output Parameter:
3497: . order - optional, desired order for the error evaluation or PETSC_DECIDE;
3498: on output, the actual order of the error evaluation
3500: Output Parameter:
3501: . wlte - the weighted local truncation error norm
3503: Level: advanced
3505: Notes:
3506: If the timestepper cannot evaluate the error in a particular step
3507: (eg. in the first step or restart steps after event handling),
3508: this routine returns wlte=-1.0 .
3510: .seealso: TSStep(), TSAdapt, TSErrorWeightedNorm()
3511: @*/
3512: PetscErrorCode TSEvaluateWLTE(TS ts,NormType wnormtype,PetscInt *order,PetscReal *wlte)
3513: {
3522: (*ts->ops->evaluatewlte)(ts,wnormtype,order,wlte);
3523: return 0;
3524: }
3526: /*@
3527: TSEvaluateStep - Evaluate the solution at the end of a time step with a given order of accuracy.
3529: Collective on TS
3531: Input Parameters:
3532: + ts - time stepping context
3533: . order - desired order of accuracy
3534: - done - whether the step was evaluated at this order (pass NULL to generate an error if not available)
3536: Output Parameter:
3537: . U - state at the end of the current step
3539: Level: advanced
3541: Notes:
3542: This function cannot be called until all stages have been evaluated.
3543: It is normally called by adaptive controllers before a step has been accepted and may also be called by the user after TSStep() has returned.
3545: .seealso: TSStep(), TSAdapt
3546: @*/
3547: PetscErrorCode TSEvaluateStep(TS ts,PetscInt order,Vec U,PetscBool *done)
3548: {
3553: (*ts->ops->evaluatestep)(ts,order,U,done);
3554: return 0;
3555: }
3557: /*@C
3558: TSGetComputeInitialCondition - Get the function used to automatically compute an initial condition for the timestepping.
3560: Not collective
3562: Input Parameter:
3563: . ts - time stepping context
3565: Output Parameter:
3566: . initConditions - The function which computes an initial condition
3568: Level: advanced
3570: Notes:
3571: The calling sequence for the function is
3572: $ initCondition(TS ts, Vec u)
3573: $ ts - The timestepping context
3574: $ u - The input vector in which the initial condition is stored
3576: .seealso: TSSetComputeInitialCondition(), TSComputeInitialCondition()
3577: @*/
3578: PetscErrorCode TSGetComputeInitialCondition(TS ts, PetscErrorCode (**initCondition)(TS, Vec))
3579: {
3582: *initCondition = ts->ops->initcondition;
3583: return 0;
3584: }
3586: /*@C
3587: TSSetComputeInitialCondition - Set the function used to automatically compute an initial condition for the timestepping.
3589: Logically collective on ts
3591: Input Parameters:
3592: + ts - time stepping context
3593: - initCondition - The function which computes an initial condition
3595: Level: advanced
3597: Calling sequence for initCondition:
3598: $ PetscErrorCode initCondition(TS ts, Vec u)
3600: + ts - The timestepping context
3601: - u - The input vector in which the initial condition is to be stored
3603: .seealso: TSGetComputeInitialCondition(), TSComputeInitialCondition()
3604: @*/
3605: PetscErrorCode TSSetComputeInitialCondition(TS ts, PetscErrorCode (*initCondition)(TS, Vec))
3606: {
3609: ts->ops->initcondition = initCondition;
3610: return 0;
3611: }
3613: /*@
3614: TSComputeInitialCondition - Compute an initial condition for the timestepping using the function previously set.
3616: Collective on ts
3618: Input Parameters:
3619: + ts - time stepping context
3620: - u - The Vec to store the condition in which will be used in TSSolve()
3622: Level: advanced
3624: .seealso: TSGetComputeInitialCondition(), TSSetComputeInitialCondition(), TSSolve()
3625: @*/
3626: PetscErrorCode TSComputeInitialCondition(TS ts, Vec u)
3627: {
3630: if (ts->ops->initcondition) (*ts->ops->initcondition)(ts, u);
3631: return 0;
3632: }
3634: /*@C
3635: TSGetComputeExactError - Get the function used to automatically compute the exact error for the timestepping.
3637: Not collective
3639: Input Parameter:
3640: . ts - time stepping context
3642: Output Parameter:
3643: . exactError - The function which computes the solution error
3645: Level: advanced
3647: Calling sequence for exactError:
3648: $ PetscErrorCode exactError(TS ts, Vec u)
3650: + ts - The timestepping context
3651: . u - The approximate solution vector
3652: - e - The input vector in which the error is stored
3654: .seealso: TSGetComputeExactError(), TSComputeExactError()
3655: @*/
3656: PetscErrorCode TSGetComputeExactError(TS ts, PetscErrorCode (**exactError)(TS, Vec, Vec))
3657: {
3660: *exactError = ts->ops->exacterror;
3661: return 0;
3662: }
3664: /*@C
3665: TSSetComputeExactError - Set the function used to automatically compute the exact error for the timestepping.
3667: Logically collective on ts
3669: Input Parameters:
3670: + ts - time stepping context
3671: - exactError - The function which computes the solution error
3673: Level: advanced
3675: Calling sequence for exactError:
3676: $ PetscErrorCode exactError(TS ts, Vec u)
3678: + ts - The timestepping context
3679: . u - The approximate solution vector
3680: - e - The input vector in which the error is stored
3682: .seealso: TSGetComputeExactError(), TSComputeExactError()
3683: @*/
3684: PetscErrorCode TSSetComputeExactError(TS ts, PetscErrorCode (*exactError)(TS, Vec, Vec))
3685: {
3688: ts->ops->exacterror = exactError;
3689: return 0;
3690: }
3692: /*@
3693: TSComputeExactError - Compute the solution error for the timestepping using the function previously set.
3695: Collective on ts
3697: Input Parameters:
3698: + ts - time stepping context
3699: . u - The approximate solution
3700: - e - The Vec used to store the error
3702: Level: advanced
3704: .seealso: TSGetComputeInitialCondition(), TSSetComputeInitialCondition(), TSSolve()
3705: @*/
3706: PetscErrorCode TSComputeExactError(TS ts, Vec u, Vec e)
3707: {
3711: if (ts->ops->exacterror) (*ts->ops->exacterror)(ts, u, e);
3712: return 0;
3713: }
3715: /*@
3716: TSSolve - Steps the requested number of timesteps.
3718: Collective on TS
3720: Input Parameters:
3721: + ts - the TS context obtained from TSCreate()
3722: - u - the solution vector (can be null if TSSetSolution() was used and TSSetExactFinalTime(ts,TS_EXACTFINALTIME_MATCHSTEP) was not used,
3723: otherwise must contain the initial conditions and will contain the solution at the final requested time
3725: Level: beginner
3727: Notes:
3728: The final time returned by this function may be different from the time of the internally
3729: held state accessible by TSGetSolution() and TSGetTime() because the method may have
3730: stepped over the final time.
3732: .seealso: TSCreate(), TSSetSolution(), TSStep(), TSGetTime(), TSGetSolveTime()
3733: @*/
3734: PetscErrorCode TSSolve(TS ts,Vec u)
3735: {
3736: Vec solution;
3741: TSSetExactFinalTimeDefault(ts);
3742: if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE && u) { /* Need ts->vec_sol to be distinct so it is not overwritten when we interpolate at the end */
3743: if (!ts->vec_sol || u == ts->vec_sol) {
3744: VecDuplicate(u,&solution);
3745: TSSetSolution(ts,solution);
3746: VecDestroy(&solution); /* grant ownership */
3747: }
3748: VecCopy(u,ts->vec_sol);
3750: } else if (u) {
3751: TSSetSolution(ts,u);
3752: }
3753: TSSetUp(ts);
3754: TSTrajectorySetUp(ts->trajectory,ts);
3760: if (ts->forward_solve) {
3761: TSForwardSetUp(ts);
3762: }
3764: /* reset number of steps only when the step is not restarted. ARKIMEX
3765: restarts the step after an event. Resetting these counters in such case causes
3766: TSTrajectory to incorrectly save the output files
3767: */
3768: /* reset time step and iteration counters */
3769: if (!ts->steps) {
3770: ts->ksp_its = 0;
3771: ts->snes_its = 0;
3772: ts->num_snes_failures = 0;
3773: ts->reject = 0;
3774: ts->steprestart = PETSC_TRUE;
3775: ts->steprollback = PETSC_FALSE;
3776: ts->rhsjacobian.time = PETSC_MIN_REAL;
3777: }
3779: /* make sure initial time step does not overshoot final time */
3780: if (ts->exact_final_time == TS_EXACTFINALTIME_MATCHSTEP) {
3781: PetscReal maxdt = ts->max_time-ts->ptime;
3782: PetscReal dt = ts->time_step;
3784: ts->time_step = dt >= maxdt ? maxdt : (PetscIsCloseAtTol(dt,maxdt,10*PETSC_MACHINE_EPSILON,0) ? maxdt : dt);
3785: }
3786: ts->reason = TS_CONVERGED_ITERATING;
3788: {
3789: PetscViewer viewer;
3790: PetscViewerFormat format;
3791: PetscBool flg;
3792: static PetscBool incall = PETSC_FALSE;
3794: if (!incall) {
3795: /* Estimate the convergence rate of the time discretization */
3796: PetscOptionsGetViewer(PetscObjectComm((PetscObject) ts),((PetscObject)ts)->options, ((PetscObject) ts)->prefix, "-ts_convergence_estimate", &viewer, &format, &flg);
3797: if (flg) {
3798: PetscConvEst conv;
3799: DM dm;
3800: PetscReal *alpha; /* Convergence rate of the solution error for each field in the L_2 norm */
3801: PetscInt Nf;
3802: PetscBool checkTemporal = PETSC_TRUE;
3804: incall = PETSC_TRUE;
3805: PetscOptionsGetBool(((PetscObject)ts)->options, ((PetscObject) ts)->prefix, "-ts_convergence_temporal", &checkTemporal, &flg);
3806: TSGetDM(ts, &dm);
3807: DMGetNumFields(dm, &Nf);
3808: PetscCalloc1(PetscMax(Nf, 1), &alpha);
3809: PetscConvEstCreate(PetscObjectComm((PetscObject) ts), &conv);
3810: PetscConvEstUseTS(conv, checkTemporal);
3811: PetscConvEstSetSolver(conv, (PetscObject) ts);
3812: PetscConvEstSetFromOptions(conv);
3813: PetscConvEstSetUp(conv);
3814: PetscConvEstGetConvRate(conv, alpha);
3815: PetscViewerPushFormat(viewer, format);
3816: PetscConvEstRateView(conv, alpha, viewer);
3817: PetscViewerPopFormat(viewer);
3818: PetscViewerDestroy(&viewer);
3819: PetscConvEstDestroy(&conv);
3820: PetscFree(alpha);
3821: incall = PETSC_FALSE;
3822: }
3823: }
3824: }
3826: TSViewFromOptions(ts,NULL,"-ts_view_pre");
3828: if (ts->ops->solve) { /* This private interface is transitional and should be removed when all implementations are updated. */
3829: (*ts->ops->solve)(ts);
3830: if (u) VecCopy(ts->vec_sol,u);
3831: ts->solvetime = ts->ptime;
3832: solution = ts->vec_sol;
3833: } else { /* Step the requested number of timesteps. */
3834: if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
3835: else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
3837: if (!ts->steps) {
3838: TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol);
3839: TSEventInitialize(ts->event,ts,ts->ptime,ts->vec_sol);
3840: }
3842: while (!ts->reason) {
3843: TSMonitor(ts,ts->steps,ts->ptime,ts->vec_sol);
3844: if (!ts->steprollback) {
3845: TSPreStep(ts);
3846: }
3847: TSStep(ts);
3848: if (ts->testjacobian) {
3849: TSRHSJacobianTest(ts,NULL);
3850: }
3851: if (ts->testjacobiantranspose) {
3852: TSRHSJacobianTestTranspose(ts,NULL);
3853: }
3854: if (ts->quadraturets && ts->costintegralfwd) { /* Must evaluate the cost integral before event is handled. The cost integral value can also be rolled back. */
3855: if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
3856: TSForwardCostIntegral(ts);
3857: if (ts->reason >= 0) ts->steps++;
3858: }
3859: if (ts->forward_solve) { /* compute forward sensitivities before event handling because postevent() may change RHS and jump conditions may have to be applied */
3860: if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
3861: TSForwardStep(ts);
3862: if (ts->reason >= 0) ts->steps++;
3863: }
3864: TSPostEvaluate(ts);
3865: TSEventHandler(ts); /* The right-hand side may be changed due to event. Be careful with Any computation using the RHS information after this point. */
3866: if (ts->steprollback) {
3867: TSPostEvaluate(ts);
3868: }
3869: if (!ts->steprollback) {
3870: TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol);
3871: TSPostStep(ts);
3872: }
3873: }
3874: TSMonitor(ts,ts->steps,ts->ptime,ts->vec_sol);
3876: if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE && ts->ptime > ts->max_time) {
3877: TSInterpolate(ts,ts->max_time,u);
3878: ts->solvetime = ts->max_time;
3879: solution = u;
3880: TSMonitor(ts,-1,ts->solvetime,solution);
3881: } else {
3882: if (u) VecCopy(ts->vec_sol,u);
3883: ts->solvetime = ts->ptime;
3884: solution = ts->vec_sol;
3885: }
3886: }
3888: TSViewFromOptions(ts,NULL,"-ts_view");
3889: VecViewFromOptions(solution,(PetscObject)ts,"-ts_view_solution");
3890: PetscObjectSAWsBlock((PetscObject)ts);
3891: if (ts->adjoint_solve) {
3892: TSAdjointSolve(ts);
3893: }
3894: return 0;
3895: }
3897: /*@
3898: TSGetTime - Gets the time of the most recently completed step.
3900: Not Collective
3902: Input Parameter:
3903: . ts - the TS context obtained from TSCreate()
3905: Output Parameter:
3906: . t - the current time. This time may not corresponds to the final time set with TSSetMaxTime(), use TSGetSolveTime().
3908: Level: beginner
3910: Note:
3911: When called during time step evaluation (e.g. during residual evaluation or via hooks set using TSSetPreStep(),
3912: TSSetPreStage(), TSSetPostStage(), or TSSetPostStep()), the time is the time at the start of the step being evaluated.
3914: .seealso: TSGetSolveTime(), TSSetTime(), TSGetTimeStep(), TSGetStepNumber()
3916: @*/
3917: PetscErrorCode TSGetTime(TS ts,PetscReal *t)
3918: {
3921: *t = ts->ptime;
3922: return 0;
3923: }
3925: /*@
3926: TSGetPrevTime - Gets the starting time of the previously completed step.
3928: Not Collective
3930: Input Parameter:
3931: . ts - the TS context obtained from TSCreate()
3933: Output Parameter:
3934: . t - the previous time
3936: Level: beginner
3938: .seealso: TSGetTime(), TSGetSolveTime(), TSGetTimeStep()
3940: @*/
3941: PetscErrorCode TSGetPrevTime(TS ts,PetscReal *t)
3942: {
3945: *t = ts->ptime_prev;
3946: return 0;
3947: }
3949: /*@
3950: TSSetTime - Allows one to reset the time.
3952: Logically Collective on TS
3954: Input Parameters:
3955: + ts - the TS context obtained from TSCreate()
3956: - time - the time
3958: Level: intermediate
3960: .seealso: TSGetTime(), TSSetMaxSteps()
3962: @*/
3963: PetscErrorCode TSSetTime(TS ts, PetscReal t)
3964: {
3967: ts->ptime = t;
3968: return 0;
3969: }
3971: /*@C
3972: TSSetOptionsPrefix - Sets the prefix used for searching for all
3973: TS options in the database.
3975: Logically Collective on TS
3977: Input Parameters:
3978: + ts - The TS context
3979: - prefix - The prefix to prepend to all option names
3981: Notes:
3982: A hyphen (-) must NOT be given at the beginning of the prefix name.
3983: The first character of all runtime options is AUTOMATICALLY the
3984: hyphen.
3986: Level: advanced
3988: .seealso: TSSetFromOptions()
3990: @*/
3991: PetscErrorCode TSSetOptionsPrefix(TS ts,const char prefix[])
3992: {
3993: SNES snes;
3996: PetscObjectSetOptionsPrefix((PetscObject)ts,prefix);
3997: TSGetSNES(ts,&snes);
3998: SNESSetOptionsPrefix(snes,prefix);
3999: return 0;
4000: }
4002: /*@C
4003: TSAppendOptionsPrefix - Appends to the prefix used for searching for all
4004: TS options in the database.
4006: Logically Collective on TS
4008: Input Parameters:
4009: + ts - The TS context
4010: - prefix - The prefix to prepend to all option names
4012: Notes:
4013: A hyphen (-) must NOT be given at the beginning of the prefix name.
4014: The first character of all runtime options is AUTOMATICALLY the
4015: hyphen.
4017: Level: advanced
4019: .seealso: TSGetOptionsPrefix()
4021: @*/
4022: PetscErrorCode TSAppendOptionsPrefix(TS ts,const char prefix[])
4023: {
4024: SNES snes;
4027: PetscObjectAppendOptionsPrefix((PetscObject)ts,prefix);
4028: TSGetSNES(ts,&snes);
4029: SNESAppendOptionsPrefix(snes,prefix);
4030: return 0;
4031: }
4033: /*@C
4034: TSGetOptionsPrefix - Sets the prefix used for searching for all
4035: TS options in the database.
4037: Not Collective
4039: Input Parameter:
4040: . ts - The TS context
4042: Output Parameter:
4043: . prefix - A pointer to the prefix string used
4045: Notes:
4046: On the fortran side, the user should pass in a string 'prifix' of
4047: sufficient length to hold the prefix.
4049: Level: intermediate
4051: .seealso: TSAppendOptionsPrefix()
4052: @*/
4053: PetscErrorCode TSGetOptionsPrefix(TS ts,const char *prefix[])
4054: {
4057: PetscObjectGetOptionsPrefix((PetscObject)ts,prefix);
4058: return 0;
4059: }
4061: /*@C
4062: TSGetRHSJacobian - Returns the Jacobian J at the present timestep.
4064: Not Collective, but parallel objects are returned if TS is parallel
4066: Input Parameter:
4067: . ts - The TS context obtained from TSCreate()
4069: Output Parameters:
4070: + Amat - The (approximate) Jacobian J of G, where U_t = G(U,t) (or NULL)
4071: . Pmat - The matrix from which the preconditioner is constructed, usually the same as Amat (or NULL)
4072: . func - Function to compute the Jacobian of the RHS (or NULL)
4073: - ctx - User-defined context for Jacobian evaluation routine (or NULL)
4075: Notes:
4076: You can pass in NULL for any return argument you do not need.
4078: Level: intermediate
4080: .seealso: TSGetTimeStep(), TSGetMatrices(), TSGetTime(), TSGetStepNumber()
4082: @*/
4083: PetscErrorCode TSGetRHSJacobian(TS ts,Mat *Amat,Mat *Pmat,TSRHSJacobian *func,void **ctx)
4084: {
4085: DM dm;
4087: if (Amat || Pmat) {
4088: SNES snes;
4089: TSGetSNES(ts,&snes);
4090: SNESSetUpMatrices(snes);
4091: SNESGetJacobian(snes,Amat,Pmat,NULL,NULL);
4092: }
4093: TSGetDM(ts,&dm);
4094: DMTSGetRHSJacobian(dm,func,ctx);
4095: return 0;
4096: }
4098: /*@C
4099: TSGetIJacobian - Returns the implicit Jacobian at the present timestep.
4101: Not Collective, but parallel objects are returned if TS is parallel
4103: Input Parameter:
4104: . ts - The TS context obtained from TSCreate()
4106: Output Parameters:
4107: + Amat - The (approximate) Jacobian of F(t,U,U_t)
4108: . Pmat - The matrix from which the preconditioner is constructed, often the same as Amat
4109: . f - The function to compute the matrices
4110: - ctx - User-defined context for Jacobian evaluation routine
4112: Notes:
4113: You can pass in NULL for any return argument you do not need.
4115: Level: advanced
4117: .seealso: TSGetTimeStep(), TSGetRHSJacobian(), TSGetMatrices(), TSGetTime(), TSGetStepNumber()
4119: @*/
4120: PetscErrorCode TSGetIJacobian(TS ts,Mat *Amat,Mat *Pmat,TSIJacobian *f,void **ctx)
4121: {
4122: DM dm;
4124: if (Amat || Pmat) {
4125: SNES snes;
4126: TSGetSNES(ts,&snes);
4127: SNESSetUpMatrices(snes);
4128: SNESGetJacobian(snes,Amat,Pmat,NULL,NULL);
4129: }
4130: TSGetDM(ts,&dm);
4131: DMTSGetIJacobian(dm,f,ctx);
4132: return 0;
4133: }
4135: #include <petsc/private/dmimpl.h>
4136: /*@
4137: TSSetDM - Sets the DM that may be used by some nonlinear solvers or preconditioners under the TS
4139: Logically Collective on ts
4141: Input Parameters:
4142: + ts - the ODE integrator object
4143: - dm - the dm, cannot be NULL
4145: Notes:
4146: A DM can only be used for solving one problem at a time because information about the problem is stored on the DM,
4147: even when not using interfaces like DMTSSetIFunction(). Use DMClone() to get a distinct DM when solving
4148: different problems using the same function space.
4150: Level: intermediate
4152: .seealso: TSGetDM(), SNESSetDM(), SNESGetDM()
4153: @*/
4154: PetscErrorCode TSSetDM(TS ts,DM dm)
4155: {
4156: SNES snes;
4157: DMTS tsdm;
4161: PetscObjectReference((PetscObject)dm);
4162: if (ts->dm) { /* Move the DMTS context over to the new DM unless the new DM already has one */
4163: if (ts->dm->dmts && !dm->dmts) {
4164: DMCopyDMTS(ts->dm,dm);
4165: DMGetDMTS(ts->dm,&tsdm);
4166: if (tsdm->originaldm == ts->dm) { /* Grant write privileges to the replacement DM */
4167: tsdm->originaldm = dm;
4168: }
4169: }
4170: DMDestroy(&ts->dm);
4171: }
4172: ts->dm = dm;
4174: TSGetSNES(ts,&snes);
4175: SNESSetDM(snes,dm);
4176: return 0;
4177: }
4179: /*@
4180: TSGetDM - Gets the DM that may be used by some preconditioners
4182: Not Collective
4184: Input Parameter:
4185: . ts - the preconditioner context
4187: Output Parameter:
4188: . dm - the dm
4190: Level: intermediate
4192: .seealso: TSSetDM(), SNESSetDM(), SNESGetDM()
4193: @*/
4194: PetscErrorCode TSGetDM(TS ts,DM *dm)
4195: {
4197: if (!ts->dm) {
4198: DMShellCreate(PetscObjectComm((PetscObject)ts),&ts->dm);
4199: if (ts->snes) SNESSetDM(ts->snes,ts->dm);
4200: }
4201: *dm = ts->dm;
4202: return 0;
4203: }
4205: /*@
4206: SNESTSFormFunction - Function to evaluate nonlinear residual
4208: Logically Collective on SNES
4210: Input Parameters:
4211: + snes - nonlinear solver
4212: . U - the current state at which to evaluate the residual
4213: - ctx - user context, must be a TS
4215: Output Parameter:
4216: . F - the nonlinear residual
4218: Notes:
4219: This function is not normally called by users and is automatically registered with the SNES used by TS.
4220: It is most frequently passed to MatFDColoringSetFunction().
4222: Level: advanced
4224: .seealso: SNESSetFunction(), MatFDColoringSetFunction()
4225: @*/
4226: PetscErrorCode SNESTSFormFunction(SNES snes,Vec U,Vec F,void *ctx)
4227: {
4228: TS ts = (TS)ctx;
4234: (ts->ops->snesfunction)(snes,U,F,ts);
4235: return 0;
4236: }
4238: /*@
4239: SNESTSFormJacobian - Function to evaluate the Jacobian
4241: Collective on SNES
4243: Input Parameters:
4244: + snes - nonlinear solver
4245: . U - the current state at which to evaluate the residual
4246: - ctx - user context, must be a TS
4248: Output Parameters:
4249: + A - the Jacobian
4250: - B - the preconditioning matrix (may be the same as A)
4252: Notes:
4253: This function is not normally called by users and is automatically registered with the SNES used by TS.
4255: Level: developer
4257: .seealso: SNESSetJacobian()
4258: @*/
4259: PetscErrorCode SNESTSFormJacobian(SNES snes,Vec U,Mat A,Mat B,void *ctx)
4260: {
4261: TS ts = (TS)ctx;
4270: (ts->ops->snesjacobian)(snes,U,A,B,ts);
4271: return 0;
4272: }
4274: /*@C
4275: TSComputeRHSFunctionLinear - Evaluate the right hand side via the user-provided Jacobian, for linear problems Udot = A U only
4277: Collective on TS
4279: Input Parameters:
4280: + ts - time stepping context
4281: . t - time at which to evaluate
4282: . U - state at which to evaluate
4283: - ctx - context
4285: Output Parameter:
4286: . F - right hand side
4288: Level: intermediate
4290: Notes:
4291: This function is intended to be passed to TSSetRHSFunction() to evaluate the right hand side for linear problems.
4292: The matrix (and optionally the evaluation context) should be passed to TSSetRHSJacobian().
4294: .seealso: TSSetRHSFunction(), TSSetRHSJacobian(), TSComputeRHSJacobianConstant()
4295: @*/
4296: PetscErrorCode TSComputeRHSFunctionLinear(TS ts,PetscReal t,Vec U,Vec F,void *ctx)
4297: {
4298: Mat Arhs,Brhs;
4300: TSGetRHSMats_Private(ts,&Arhs,&Brhs);
4301: /* undo the damage caused by shifting */
4302: TSRecoverRHSJacobian(ts,Arhs,Brhs);
4303: TSComputeRHSJacobian(ts,t,U,Arhs,Brhs);
4304: MatMult(Arhs,U,F);
4305: return 0;
4306: }
4308: /*@C
4309: TSComputeRHSJacobianConstant - Reuses a Jacobian that is time-independent.
4311: Collective on TS
4313: Input Parameters:
4314: + ts - time stepping context
4315: . t - time at which to evaluate
4316: . U - state at which to evaluate
4317: - ctx - context
4319: Output Parameters:
4320: + A - pointer to operator
4321: - B - pointer to preconditioning matrix
4323: Level: intermediate
4325: Notes:
4326: This function is intended to be passed to TSSetRHSJacobian() to evaluate the Jacobian for linear time-independent problems.
4328: .seealso: TSSetRHSFunction(), TSSetRHSJacobian(), TSComputeRHSFunctionLinear()
4329: @*/
4330: PetscErrorCode TSComputeRHSJacobianConstant(TS ts,PetscReal t,Vec U,Mat A,Mat B,void *ctx)
4331: {
4332: return 0;
4333: }
4335: /*@C
4336: TSComputeIFunctionLinear - Evaluate the left hand side via the user-provided Jacobian, for linear problems only
4338: Collective on TS
4340: Input Parameters:
4341: + ts - time stepping context
4342: . t - time at which to evaluate
4343: . U - state at which to evaluate
4344: . Udot - time derivative of state vector
4345: - ctx - context
4347: Output Parameter:
4348: . F - left hand side
4350: Level: intermediate
4352: Notes:
4353: The assumption here is that the left hand side is of the form A*Udot (and not A*Udot + B*U). For other cases, the
4354: user is required to write their own TSComputeIFunction.
4355: This function is intended to be passed to TSSetIFunction() to evaluate the left hand side for linear problems.
4356: The matrix (and optionally the evaluation context) should be passed to TSSetIJacobian().
4358: Note that using this function is NOT equivalent to using TSComputeRHSFunctionLinear() since that solves Udot = A U
4360: .seealso: TSSetIFunction(), TSSetIJacobian(), TSComputeIJacobianConstant(), TSComputeRHSFunctionLinear()
4361: @*/
4362: PetscErrorCode TSComputeIFunctionLinear(TS ts,PetscReal t,Vec U,Vec Udot,Vec F,void *ctx)
4363: {
4364: Mat A,B;
4366: TSGetIJacobian(ts,&A,&B,NULL,NULL);
4367: TSComputeIJacobian(ts,t,U,Udot,1.0,A,B,PETSC_TRUE);
4368: MatMult(A,Udot,F);
4369: return 0;
4370: }
4372: /*@C
4373: TSComputeIJacobianConstant - Reuses a time-independent for a semi-implicit DAE or ODE
4375: Collective on TS
4377: Input Parameters:
4378: + ts - time stepping context
4379: . t - time at which to evaluate
4380: . U - state at which to evaluate
4381: . Udot - time derivative of state vector
4382: . shift - shift to apply
4383: - ctx - context
4385: Output Parameters:
4386: + A - pointer to operator
4387: - B - pointer to preconditioning matrix
4389: Level: advanced
4391: Notes:
4392: This function is intended to be passed to TSSetIJacobian() to evaluate the Jacobian for linear time-independent problems.
4394: It is only appropriate for problems of the form
4396: $ M Udot = F(U,t)
4398: where M is constant and F is non-stiff. The user must pass M to TSSetIJacobian(). The current implementation only
4399: works with IMEX time integration methods such as TSROSW and TSARKIMEX, since there is no support for de-constructing
4400: an implicit operator of the form
4402: $ shift*M + J
4404: where J is the Jacobian of -F(U). Support may be added in a future version of PETSc, but for now, the user must store
4405: a copy of M or reassemble it when requested.
4407: .seealso: TSSetIFunction(), TSSetIJacobian(), TSComputeIFunctionLinear()
4408: @*/
4409: PetscErrorCode TSComputeIJacobianConstant(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,void *ctx)
4410: {
4411: MatScale(A, shift / ts->ijacobian.shift);
4412: ts->ijacobian.shift = shift;
4413: return 0;
4414: }
4416: /*@
4417: TSGetEquationType - Gets the type of the equation that TS is solving.
4419: Not Collective
4421: Input Parameter:
4422: . ts - the TS context
4424: Output Parameter:
4425: . equation_type - see TSEquationType
4427: Level: beginner
4429: .seealso: TSSetEquationType(), TSEquationType
4430: @*/
4431: PetscErrorCode TSGetEquationType(TS ts,TSEquationType *equation_type)
4432: {
4435: *equation_type = ts->equation_type;
4436: return 0;
4437: }
4439: /*@
4440: TSSetEquationType - Sets the type of the equation that TS is solving.
4442: Not Collective
4444: Input Parameters:
4445: + ts - the TS context
4446: - equation_type - see TSEquationType
4448: Level: advanced
4450: .seealso: TSGetEquationType(), TSEquationType
4451: @*/
4452: PetscErrorCode TSSetEquationType(TS ts,TSEquationType equation_type)
4453: {
4455: ts->equation_type = equation_type;
4456: return 0;
4457: }
4459: /*@
4460: TSGetConvergedReason - Gets the reason the TS iteration was stopped.
4462: Not Collective
4464: Input Parameter:
4465: . ts - the TS context
4467: Output Parameter:
4468: . reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the
4469: manual pages for the individual convergence tests for complete lists
4471: Level: beginner
4473: Notes:
4474: Can only be called after the call to TSSolve() is complete.
4476: .seealso: TSSetConvergenceTest(), TSConvergedReason
4477: @*/
4478: PetscErrorCode TSGetConvergedReason(TS ts,TSConvergedReason *reason)
4479: {
4482: *reason = ts->reason;
4483: return 0;
4484: }
4486: /*@
4487: TSSetConvergedReason - Sets the reason for handling the convergence of TSSolve.
4489: Logically Collective; reason must contain common value
4491: Input Parameters:
4492: + ts - the TS context
4493: - reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the
4494: manual pages for the individual convergence tests for complete lists
4496: Level: advanced
4498: Notes:
4499: Can only be called while TSSolve() is active.
4501: .seealso: TSConvergedReason
4502: @*/
4503: PetscErrorCode TSSetConvergedReason(TS ts,TSConvergedReason reason)
4504: {
4506: ts->reason = reason;
4507: return 0;
4508: }
4510: /*@
4511: TSGetSolveTime - Gets the time after a call to TSSolve()
4513: Not Collective
4515: Input Parameter:
4516: . ts - the TS context
4518: Output Parameter:
4519: . ftime - the final time. This time corresponds to the final time set with TSSetMaxTime()
4521: Level: beginner
4523: Notes:
4524: Can only be called after the call to TSSolve() is complete.
4526: .seealso: TSSetConvergenceTest(), TSConvergedReason
4527: @*/
4528: PetscErrorCode TSGetSolveTime(TS ts,PetscReal *ftime)
4529: {
4532: *ftime = ts->solvetime;
4533: return 0;
4534: }
4536: /*@
4537: TSGetSNESIterations - Gets the total number of nonlinear iterations
4538: used by the time integrator.
4540: Not Collective
4542: Input Parameter:
4543: . ts - TS context
4545: Output Parameter:
4546: . nits - number of nonlinear iterations
4548: Notes:
4549: This counter is reset to zero for each successive call to TSSolve().
4551: Level: intermediate
4553: .seealso: TSGetKSPIterations()
4554: @*/
4555: PetscErrorCode TSGetSNESIterations(TS ts,PetscInt *nits)
4556: {
4559: *nits = ts->snes_its;
4560: return 0;
4561: }
4563: /*@
4564: TSGetKSPIterations - Gets the total number of linear iterations
4565: used by the time integrator.
4567: Not Collective
4569: Input Parameter:
4570: . ts - TS context
4572: Output Parameter:
4573: . lits - number of linear iterations
4575: Notes:
4576: This counter is reset to zero for each successive call to TSSolve().
4578: Level: intermediate
4580: .seealso: TSGetSNESIterations(), SNESGetKSPIterations()
4581: @*/
4582: PetscErrorCode TSGetKSPIterations(TS ts,PetscInt *lits)
4583: {
4586: *lits = ts->ksp_its;
4587: return 0;
4588: }
4590: /*@
4591: TSGetStepRejections - Gets the total number of rejected steps.
4593: Not Collective
4595: Input Parameter:
4596: . ts - TS context
4598: Output Parameter:
4599: . rejects - number of steps rejected
4601: Notes:
4602: This counter is reset to zero for each successive call to TSSolve().
4604: Level: intermediate
4606: .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetSNESFailures(), TSSetMaxSNESFailures(), TSSetErrorIfStepFails()
4607: @*/
4608: PetscErrorCode TSGetStepRejections(TS ts,PetscInt *rejects)
4609: {
4612: *rejects = ts->reject;
4613: return 0;
4614: }
4616: /*@
4617: TSGetSNESFailures - Gets the total number of failed SNES solves
4619: Not Collective
4621: Input Parameter:
4622: . ts - TS context
4624: Output Parameter:
4625: . fails - number of failed nonlinear solves
4627: Notes:
4628: This counter is reset to zero for each successive call to TSSolve().
4630: Level: intermediate
4632: .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSSetMaxSNESFailures()
4633: @*/
4634: PetscErrorCode TSGetSNESFailures(TS ts,PetscInt *fails)
4635: {
4638: *fails = ts->num_snes_failures;
4639: return 0;
4640: }
4642: /*@
4643: TSSetMaxStepRejections - Sets the maximum number of step rejections before a step fails
4645: Not Collective
4647: Input Parameters:
4648: + ts - TS context
4649: - rejects - maximum number of rejected steps, pass -1 for unlimited
4651: Notes:
4652: The counter is reset to zero for each step
4654: Options Database Key:
4655: . -ts_max_reject - Maximum number of step rejections before a step fails
4657: Level: intermediate
4659: .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxSNESFailures(), TSGetStepRejections(), TSGetSNESFailures(), TSSetErrorIfStepFails(), TSGetConvergedReason()
4660: @*/
4661: PetscErrorCode TSSetMaxStepRejections(TS ts,PetscInt rejects)
4662: {
4664: ts->max_reject = rejects;
4665: return 0;
4666: }
4668: /*@
4669: TSSetMaxSNESFailures - Sets the maximum number of failed SNES solves
4671: Not Collective
4673: Input Parameters:
4674: + ts - TS context
4675: - fails - maximum number of failed nonlinear solves, pass -1 for unlimited
4677: Notes:
4678: The counter is reset to zero for each successive call to TSSolve().
4680: Options Database Key:
4681: . -ts_max_snes_failures - Maximum number of nonlinear solve failures
4683: Level: intermediate
4685: .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSGetSNESFailures(), SNESGetConvergedReason(), TSGetConvergedReason()
4686: @*/
4687: PetscErrorCode TSSetMaxSNESFailures(TS ts,PetscInt fails)
4688: {
4690: ts->max_snes_failures = fails;
4691: return 0;
4692: }
4694: /*@
4695: TSSetErrorIfStepFails - Error if no step succeeds
4697: Not Collective
4699: Input Parameters:
4700: + ts - TS context
4701: - err - PETSC_TRUE to error if no step succeeds, PETSC_FALSE to return without failure
4703: Options Database Key:
4704: . -ts_error_if_step_fails - Error if no step succeeds
4706: Level: intermediate
4708: .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSGetSNESFailures(), TSSetErrorIfStepFails(), TSGetConvergedReason()
4709: @*/
4710: PetscErrorCode TSSetErrorIfStepFails(TS ts,PetscBool err)
4711: {
4713: ts->errorifstepfailed = err;
4714: return 0;
4715: }
4717: /*@
4718: TSGetAdapt - Get the adaptive controller context for the current method
4720: Collective on TS if controller has not been created yet
4722: Input Parameter:
4723: . ts - time stepping context
4725: Output Parameter:
4726: . adapt - adaptive controller
4728: Level: intermediate
4730: .seealso: TSAdapt, TSAdaptSetType(), TSAdaptChoose()
4731: @*/
4732: PetscErrorCode TSGetAdapt(TS ts,TSAdapt *adapt)
4733: {
4736: if (!ts->adapt) {
4737: TSAdaptCreate(PetscObjectComm((PetscObject)ts),&ts->adapt);
4738: PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->adapt);
4739: PetscObjectIncrementTabLevel((PetscObject)ts->adapt,(PetscObject)ts,1);
4740: }
4741: *adapt = ts->adapt;
4742: return 0;
4743: }
4745: /*@
4746: TSSetTolerances - Set tolerances for local truncation error when using adaptive controller
4748: Logically Collective
4750: Input Parameters:
4751: + ts - time integration context
4752: . atol - scalar absolute tolerances, PETSC_DECIDE to leave current value
4753: . vatol - vector of absolute tolerances or NULL, used in preference to atol if present
4754: . rtol - scalar relative tolerances, PETSC_DECIDE to leave current value
4755: - vrtol - vector of relative tolerances or NULL, used in preference to atol if present
4757: Options Database keys:
4758: + -ts_rtol <rtol> - relative tolerance for local truncation error
4759: - -ts_atol <atol> - Absolute tolerance for local truncation error
4761: Notes:
4762: With PETSc's implicit schemes for DAE problems, the calculation of the local truncation error
4763: (LTE) includes both the differential and the algebraic variables. If one wants the LTE to be
4764: computed only for the differential or the algebraic part then this can be done using the vector of
4765: tolerances vatol. For example, by setting the tolerance vector with the desired tolerance for the
4766: differential part and infinity for the algebraic part, the LTE calculation will include only the
4767: differential variables.
4769: Level: beginner
4771: .seealso: TS, TSAdapt, TSErrorWeightedNorm(), TSGetTolerances()
4772: @*/
4773: PetscErrorCode TSSetTolerances(TS ts,PetscReal atol,Vec vatol,PetscReal rtol,Vec vrtol)
4774: {
4775: if (atol != PETSC_DECIDE && atol != PETSC_DEFAULT) ts->atol = atol;
4776: if (vatol) {
4777: PetscObjectReference((PetscObject)vatol);
4778: VecDestroy(&ts->vatol);
4779: ts->vatol = vatol;
4780: }
4781: if (rtol != PETSC_DECIDE && rtol != PETSC_DEFAULT) ts->rtol = rtol;
4782: if (vrtol) {
4783: PetscObjectReference((PetscObject)vrtol);
4784: VecDestroy(&ts->vrtol);
4785: ts->vrtol = vrtol;
4786: }
4787: return 0;
4788: }
4790: /*@
4791: TSGetTolerances - Get tolerances for local truncation error when using adaptive controller
4793: Logically Collective
4795: Input Parameter:
4796: . ts - time integration context
4798: Output Parameters:
4799: + atol - scalar absolute tolerances, NULL to ignore
4800: . vatol - vector of absolute tolerances, NULL to ignore
4801: . rtol - scalar relative tolerances, NULL to ignore
4802: - vrtol - vector of relative tolerances, NULL to ignore
4804: Level: beginner
4806: .seealso: TS, TSAdapt, TSErrorWeightedNorm(), TSSetTolerances()
4807: @*/
4808: PetscErrorCode TSGetTolerances(TS ts,PetscReal *atol,Vec *vatol,PetscReal *rtol,Vec *vrtol)
4809: {
4810: if (atol) *atol = ts->atol;
4811: if (vatol) *vatol = ts->vatol;
4812: if (rtol) *rtol = ts->rtol;
4813: if (vrtol) *vrtol = ts->vrtol;
4814: return 0;
4815: }
4817: /*@
4818: TSErrorWeightedNorm2 - compute a weighted 2-norm of the difference between two state vectors
4820: Collective on TS
4822: Input Parameters:
4823: + ts - time stepping context
4824: . U - state vector, usually ts->vec_sol
4825: - Y - state vector to be compared to U
4827: Output Parameters:
4828: + norm - weighted norm, a value of 1.0 means that the error matches the tolerances
4829: . norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
4830: - normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
4832: Level: developer
4834: .seealso: TSErrorWeightedNorm(), TSErrorWeightedNormInfinity()
4835: @*/
4836: PetscErrorCode TSErrorWeightedNorm2(TS ts,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
4837: {
4838: PetscInt i,n,N,rstart;
4839: PetscInt n_loc,na_loc,nr_loc;
4840: PetscReal n_glb,na_glb,nr_glb;
4841: const PetscScalar *u,*y;
4842: PetscReal sum,suma,sumr,gsum,gsuma,gsumr,diff;
4843: PetscReal tol,tola,tolr;
4844: PetscReal err_loc[6],err_glb[6];
4857: VecGetSize(U,&N);
4858: VecGetLocalSize(U,&n);
4859: VecGetOwnershipRange(U,&rstart,NULL);
4860: VecGetArrayRead(U,&u);
4861: VecGetArrayRead(Y,&y);
4862: sum = 0.; n_loc = 0;
4863: suma = 0.; na_loc = 0;
4864: sumr = 0.; nr_loc = 0;
4865: if (ts->vatol && ts->vrtol) {
4866: const PetscScalar *atol,*rtol;
4867: VecGetArrayRead(ts->vatol,&atol);
4868: VecGetArrayRead(ts->vrtol,&rtol);
4869: for (i=0; i<n; i++) {
4870: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4871: diff = PetscAbsScalar(y[i] - u[i]);
4872: tola = PetscRealPart(atol[i]);
4873: if (tola>0.) {
4874: suma += PetscSqr(diff/tola);
4875: na_loc++;
4876: }
4877: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4878: if (tolr>0.) {
4879: sumr += PetscSqr(diff/tolr);
4880: nr_loc++;
4881: }
4882: tol=tola+tolr;
4883: if (tol>0.) {
4884: sum += PetscSqr(diff/tol);
4885: n_loc++;
4886: }
4887: }
4888: VecRestoreArrayRead(ts->vatol,&atol);
4889: VecRestoreArrayRead(ts->vrtol,&rtol);
4890: } else if (ts->vatol) { /* vector atol, scalar rtol */
4891: const PetscScalar *atol;
4892: VecGetArrayRead(ts->vatol,&atol);
4893: for (i=0; i<n; i++) {
4894: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4895: diff = PetscAbsScalar(y[i] - u[i]);
4896: tola = PetscRealPart(atol[i]);
4897: if (tola>0.) {
4898: suma += PetscSqr(diff/tola);
4899: na_loc++;
4900: }
4901: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4902: if (tolr>0.) {
4903: sumr += PetscSqr(diff/tolr);
4904: nr_loc++;
4905: }
4906: tol=tola+tolr;
4907: if (tol>0.) {
4908: sum += PetscSqr(diff/tol);
4909: n_loc++;
4910: }
4911: }
4912: VecRestoreArrayRead(ts->vatol,&atol);
4913: } else if (ts->vrtol) { /* scalar atol, vector rtol */
4914: const PetscScalar *rtol;
4915: VecGetArrayRead(ts->vrtol,&rtol);
4916: for (i=0; i<n; i++) {
4917: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4918: diff = PetscAbsScalar(y[i] - u[i]);
4919: tola = ts->atol;
4920: if (tola>0.) {
4921: suma += PetscSqr(diff/tola);
4922: na_loc++;
4923: }
4924: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4925: if (tolr>0.) {
4926: sumr += PetscSqr(diff/tolr);
4927: nr_loc++;
4928: }
4929: tol=tola+tolr;
4930: if (tol>0.) {
4931: sum += PetscSqr(diff/tol);
4932: n_loc++;
4933: }
4934: }
4935: VecRestoreArrayRead(ts->vrtol,&rtol);
4936: } else { /* scalar atol, scalar rtol */
4937: for (i=0; i<n; i++) {
4938: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4939: diff = PetscAbsScalar(y[i] - u[i]);
4940: tola = ts->atol;
4941: if (tola>0.) {
4942: suma += PetscSqr(diff/tola);
4943: na_loc++;
4944: }
4945: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4946: if (tolr>0.) {
4947: sumr += PetscSqr(diff/tolr);
4948: nr_loc++;
4949: }
4950: tol=tola+tolr;
4951: if (tol>0.) {
4952: sum += PetscSqr(diff/tol);
4953: n_loc++;
4954: }
4955: }
4956: }
4957: VecRestoreArrayRead(U,&u);
4958: VecRestoreArrayRead(Y,&y);
4960: err_loc[0] = sum;
4961: err_loc[1] = suma;
4962: err_loc[2] = sumr;
4963: err_loc[3] = (PetscReal)n_loc;
4964: err_loc[4] = (PetscReal)na_loc;
4965: err_loc[5] = (PetscReal)nr_loc;
4967: MPIU_Allreduce(err_loc,err_glb,6,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)ts));
4969: gsum = err_glb[0];
4970: gsuma = err_glb[1];
4971: gsumr = err_glb[2];
4972: n_glb = err_glb[3];
4973: na_glb = err_glb[4];
4974: nr_glb = err_glb[5];
4976: *norm = 0.;
4977: if (n_glb>0.) {*norm = PetscSqrtReal(gsum / n_glb);}
4978: *norma = 0.;
4979: if (na_glb>0.) {*norma = PetscSqrtReal(gsuma / na_glb);}
4980: *normr = 0.;
4981: if (nr_glb>0.) {*normr = PetscSqrtReal(gsumr / nr_glb);}
4986: return 0;
4987: }
4989: /*@
4990: TSErrorWeightedNormInfinity - compute a weighted infinity-norm of the difference between two state vectors
4992: Collective on TS
4994: Input Parameters:
4995: + ts - time stepping context
4996: . U - state vector, usually ts->vec_sol
4997: - Y - state vector to be compared to U
4999: Output Parameters:
5000: + norm - weighted norm, a value of 1.0 means that the error matches the tolerances
5001: . norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
5002: - normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
5004: Level: developer
5006: .seealso: TSErrorWeightedNorm(), TSErrorWeightedNorm2()
5007: @*/
5008: PetscErrorCode TSErrorWeightedNormInfinity(TS ts,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5009: {
5010: PetscInt i,n,N,rstart;
5011: const PetscScalar *u,*y;
5012: PetscReal max,gmax,maxa,gmaxa,maxr,gmaxr;
5013: PetscReal tol,tola,tolr,diff;
5014: PetscReal err_loc[3],err_glb[3];
5027: VecGetSize(U,&N);
5028: VecGetLocalSize(U,&n);
5029: VecGetOwnershipRange(U,&rstart,NULL);
5030: VecGetArrayRead(U,&u);
5031: VecGetArrayRead(Y,&y);
5033: max=0.;
5034: maxa=0.;
5035: maxr=0.;
5037: if (ts->vatol && ts->vrtol) { /* vector atol, vector rtol */
5038: const PetscScalar *atol,*rtol;
5039: VecGetArrayRead(ts->vatol,&atol);
5040: VecGetArrayRead(ts->vrtol,&rtol);
5042: for (i=0; i<n; i++) {
5043: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5044: diff = PetscAbsScalar(y[i] - u[i]);
5045: tola = PetscRealPart(atol[i]);
5046: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5047: tol = tola+tolr;
5048: if (tola>0.) {
5049: maxa = PetscMax(maxa,diff / tola);
5050: }
5051: if (tolr>0.) {
5052: maxr = PetscMax(maxr,diff / tolr);
5053: }
5054: if (tol>0.) {
5055: max = PetscMax(max,diff / tol);
5056: }
5057: }
5058: VecRestoreArrayRead(ts->vatol,&atol);
5059: VecRestoreArrayRead(ts->vrtol,&rtol);
5060: } else if (ts->vatol) { /* vector atol, scalar rtol */
5061: const PetscScalar *atol;
5062: VecGetArrayRead(ts->vatol,&atol);
5063: for (i=0; i<n; i++) {
5064: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5065: diff = PetscAbsScalar(y[i] - u[i]);
5066: tola = PetscRealPart(atol[i]);
5067: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5068: tol = tola+tolr;
5069: if (tola>0.) {
5070: maxa = PetscMax(maxa,diff / tola);
5071: }
5072: if (tolr>0.) {
5073: maxr = PetscMax(maxr,diff / tolr);
5074: }
5075: if (tol>0.) {
5076: max = PetscMax(max,diff / tol);
5077: }
5078: }
5079: VecRestoreArrayRead(ts->vatol,&atol);
5080: } else if (ts->vrtol) { /* scalar atol, vector rtol */
5081: const PetscScalar *rtol;
5082: VecGetArrayRead(ts->vrtol,&rtol);
5084: for (i=0; i<n; i++) {
5085: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5086: diff = PetscAbsScalar(y[i] - u[i]);
5087: tola = ts->atol;
5088: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5089: tol = tola+tolr;
5090: if (tola>0.) {
5091: maxa = PetscMax(maxa,diff / tola);
5092: }
5093: if (tolr>0.) {
5094: maxr = PetscMax(maxr,diff / tolr);
5095: }
5096: if (tol>0.) {
5097: max = PetscMax(max,diff / tol);
5098: }
5099: }
5100: VecRestoreArrayRead(ts->vrtol,&rtol);
5101: } else { /* scalar atol, scalar rtol */
5103: for (i=0; i<n; i++) {
5104: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5105: diff = PetscAbsScalar(y[i] - u[i]);
5106: tola = ts->atol;
5107: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5108: tol = tola+tolr;
5109: if (tola>0.) {
5110: maxa = PetscMax(maxa,diff / tola);
5111: }
5112: if (tolr>0.) {
5113: maxr = PetscMax(maxr,diff / tolr);
5114: }
5115: if (tol>0.) {
5116: max = PetscMax(max,diff / tol);
5117: }
5118: }
5119: }
5120: VecRestoreArrayRead(U,&u);
5121: VecRestoreArrayRead(Y,&y);
5122: err_loc[0] = max;
5123: err_loc[1] = maxa;
5124: err_loc[2] = maxr;
5125: MPIU_Allreduce(err_loc,err_glb,3,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)ts));
5126: gmax = err_glb[0];
5127: gmaxa = err_glb[1];
5128: gmaxr = err_glb[2];
5130: *norm = gmax;
5131: *norma = gmaxa;
5132: *normr = gmaxr;
5136: return 0;
5137: }
5139: /*@
5140: TSErrorWeightedNorm - compute a weighted norm of the difference between two state vectors based on supplied absolute and relative tolerances
5142: Collective on TS
5144: Input Parameters:
5145: + ts - time stepping context
5146: . U - state vector, usually ts->vec_sol
5147: . Y - state vector to be compared to U
5148: - wnormtype - norm type, either NORM_2 or NORM_INFINITY
5150: Output Parameters:
5151: + norm - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5152: . norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5153: - normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5155: Options Database Keys:
5156: . -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5158: Level: developer
5160: .seealso: TSErrorWeightedNormInfinity(), TSErrorWeightedNorm2(), TSErrorWeightedENorm
5161: @*/
5162: PetscErrorCode TSErrorWeightedNorm(TS ts,Vec U,Vec Y,NormType wnormtype,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5163: {
5164: if (wnormtype == NORM_2) {
5165: TSErrorWeightedNorm2(ts,U,Y,norm,norma,normr);
5166: } else if (wnormtype == NORM_INFINITY) {
5167: TSErrorWeightedNormInfinity(ts,U,Y,norm,norma,normr);
5168: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]);
5169: return 0;
5170: }
5172: /*@
5173: TSErrorWeightedENorm2 - compute a weighted 2 error norm based on supplied absolute and relative tolerances
5175: Collective on TS
5177: Input Parameters:
5178: + ts - time stepping context
5179: . E - error vector
5180: . U - state vector, usually ts->vec_sol
5181: - Y - state vector, previous time step
5183: Output Parameters:
5184: + norm - weighted norm, a value of 1.0 means that the error matches the tolerances
5185: . norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
5186: - normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
5188: Level: developer
5190: .seealso: TSErrorWeightedENorm(), TSErrorWeightedENormInfinity()
5191: @*/
5192: PetscErrorCode TSErrorWeightedENorm2(TS ts,Vec E,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5193: {
5194: PetscInt i,n,N,rstart;
5195: PetscInt n_loc,na_loc,nr_loc;
5196: PetscReal n_glb,na_glb,nr_glb;
5197: const PetscScalar *e,*u,*y;
5198: PetscReal err,sum,suma,sumr,gsum,gsuma,gsumr;
5199: PetscReal tol,tola,tolr;
5200: PetscReal err_loc[6],err_glb[6];
5215: VecGetSize(E,&N);
5216: VecGetLocalSize(E,&n);
5217: VecGetOwnershipRange(E,&rstart,NULL);
5218: VecGetArrayRead(E,&e);
5219: VecGetArrayRead(U,&u);
5220: VecGetArrayRead(Y,&y);
5221: sum = 0.; n_loc = 0;
5222: suma = 0.; na_loc = 0;
5223: sumr = 0.; nr_loc = 0;
5224: if (ts->vatol && ts->vrtol) {
5225: const PetscScalar *atol,*rtol;
5226: VecGetArrayRead(ts->vatol,&atol);
5227: VecGetArrayRead(ts->vrtol,&rtol);
5228: for (i=0; i<n; i++) {
5229: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5230: err = PetscAbsScalar(e[i]);
5231: tola = PetscRealPart(atol[i]);
5232: if (tola>0.) {
5233: suma += PetscSqr(err/tola);
5234: na_loc++;
5235: }
5236: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5237: if (tolr>0.) {
5238: sumr += PetscSqr(err/tolr);
5239: nr_loc++;
5240: }
5241: tol=tola+tolr;
5242: if (tol>0.) {
5243: sum += PetscSqr(err/tol);
5244: n_loc++;
5245: }
5246: }
5247: VecRestoreArrayRead(ts->vatol,&atol);
5248: VecRestoreArrayRead(ts->vrtol,&rtol);
5249: } else if (ts->vatol) { /* vector atol, scalar rtol */
5250: const PetscScalar *atol;
5251: VecGetArrayRead(ts->vatol,&atol);
5252: for (i=0; i<n; i++) {
5253: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5254: err = PetscAbsScalar(e[i]);
5255: tola = PetscRealPart(atol[i]);
5256: if (tola>0.) {
5257: suma += PetscSqr(err/tola);
5258: na_loc++;
5259: }
5260: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5261: if (tolr>0.) {
5262: sumr += PetscSqr(err/tolr);
5263: nr_loc++;
5264: }
5265: tol=tola+tolr;
5266: if (tol>0.) {
5267: sum += PetscSqr(err/tol);
5268: n_loc++;
5269: }
5270: }
5271: VecRestoreArrayRead(ts->vatol,&atol);
5272: } else if (ts->vrtol) { /* scalar atol, vector rtol */
5273: const PetscScalar *rtol;
5274: VecGetArrayRead(ts->vrtol,&rtol);
5275: for (i=0; i<n; i++) {
5276: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5277: err = PetscAbsScalar(e[i]);
5278: tola = ts->atol;
5279: if (tola>0.) {
5280: suma += PetscSqr(err/tola);
5281: na_loc++;
5282: }
5283: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5284: if (tolr>0.) {
5285: sumr += PetscSqr(err/tolr);
5286: nr_loc++;
5287: }
5288: tol=tola+tolr;
5289: if (tol>0.) {
5290: sum += PetscSqr(err/tol);
5291: n_loc++;
5292: }
5293: }
5294: VecRestoreArrayRead(ts->vrtol,&rtol);
5295: } else { /* scalar atol, scalar rtol */
5296: for (i=0; i<n; i++) {
5297: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5298: err = PetscAbsScalar(e[i]);
5299: tola = ts->atol;
5300: if (tola>0.) {
5301: suma += PetscSqr(err/tola);
5302: na_loc++;
5303: }
5304: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5305: if (tolr>0.) {
5306: sumr += PetscSqr(err/tolr);
5307: nr_loc++;
5308: }
5309: tol=tola+tolr;
5310: if (tol>0.) {
5311: sum += PetscSqr(err/tol);
5312: n_loc++;
5313: }
5314: }
5315: }
5316: VecRestoreArrayRead(E,&e);
5317: VecRestoreArrayRead(U,&u);
5318: VecRestoreArrayRead(Y,&y);
5320: err_loc[0] = sum;
5321: err_loc[1] = suma;
5322: err_loc[2] = sumr;
5323: err_loc[3] = (PetscReal)n_loc;
5324: err_loc[4] = (PetscReal)na_loc;
5325: err_loc[5] = (PetscReal)nr_loc;
5327: MPIU_Allreduce(err_loc,err_glb,6,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)ts));
5329: gsum = err_glb[0];
5330: gsuma = err_glb[1];
5331: gsumr = err_glb[2];
5332: n_glb = err_glb[3];
5333: na_glb = err_glb[4];
5334: nr_glb = err_glb[5];
5336: *norm = 0.;
5337: if (n_glb>0.) {*norm = PetscSqrtReal(gsum / n_glb);}
5338: *norma = 0.;
5339: if (na_glb>0.) {*norma = PetscSqrtReal(gsuma / na_glb);}
5340: *normr = 0.;
5341: if (nr_glb>0.) {*normr = PetscSqrtReal(gsumr / nr_glb);}
5346: return 0;
5347: }
5349: /*@
5350: TSErrorWeightedENormInfinity - compute a weighted infinity error norm based on supplied absolute and relative tolerances
5351: Collective on TS
5353: Input Parameters:
5354: + ts - time stepping context
5355: . E - error vector
5356: . U - state vector, usually ts->vec_sol
5357: - Y - state vector, previous time step
5359: Output Parameters:
5360: + norm - weighted norm, a value of 1.0 means that the error matches the tolerances
5361: . norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
5362: - normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
5364: Level: developer
5366: .seealso: TSErrorWeightedENorm(), TSErrorWeightedENorm2()
5367: @*/
5368: PetscErrorCode TSErrorWeightedENormInfinity(TS ts,Vec E,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5369: {
5370: PetscInt i,n,N,rstart;
5371: const PetscScalar *e,*u,*y;
5372: PetscReal err,max,gmax,maxa,gmaxa,maxr,gmaxr;
5373: PetscReal tol,tola,tolr;
5374: PetscReal err_loc[3],err_glb[3];
5389: VecGetSize(E,&N);
5390: VecGetLocalSize(E,&n);
5391: VecGetOwnershipRange(E,&rstart,NULL);
5392: VecGetArrayRead(E,&e);
5393: VecGetArrayRead(U,&u);
5394: VecGetArrayRead(Y,&y);
5396: max=0.;
5397: maxa=0.;
5398: maxr=0.;
5400: if (ts->vatol && ts->vrtol) { /* vector atol, vector rtol */
5401: const PetscScalar *atol,*rtol;
5402: VecGetArrayRead(ts->vatol,&atol);
5403: VecGetArrayRead(ts->vrtol,&rtol);
5405: for (i=0; i<n; i++) {
5406: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5407: err = PetscAbsScalar(e[i]);
5408: tola = PetscRealPart(atol[i]);
5409: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5410: tol = tola+tolr;
5411: if (tola>0.) {
5412: maxa = PetscMax(maxa,err / tola);
5413: }
5414: if (tolr>0.) {
5415: maxr = PetscMax(maxr,err / tolr);
5416: }
5417: if (tol>0.) {
5418: max = PetscMax(max,err / tol);
5419: }
5420: }
5421: VecRestoreArrayRead(ts->vatol,&atol);
5422: VecRestoreArrayRead(ts->vrtol,&rtol);
5423: } else if (ts->vatol) { /* vector atol, scalar rtol */
5424: const PetscScalar *atol;
5425: VecGetArrayRead(ts->vatol,&atol);
5426: for (i=0; i<n; i++) {
5427: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5428: err = PetscAbsScalar(e[i]);
5429: tola = PetscRealPart(atol[i]);
5430: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5431: tol = tola+tolr;
5432: if (tola>0.) {
5433: maxa = PetscMax(maxa,err / tola);
5434: }
5435: if (tolr>0.) {
5436: maxr = PetscMax(maxr,err / tolr);
5437: }
5438: if (tol>0.) {
5439: max = PetscMax(max,err / tol);
5440: }
5441: }
5442: VecRestoreArrayRead(ts->vatol,&atol);
5443: } else if (ts->vrtol) { /* scalar atol, vector rtol */
5444: const PetscScalar *rtol;
5445: VecGetArrayRead(ts->vrtol,&rtol);
5447: for (i=0; i<n; i++) {
5448: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5449: err = PetscAbsScalar(e[i]);
5450: tola = ts->atol;
5451: tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5452: tol = tola+tolr;
5453: if (tola>0.) {
5454: maxa = PetscMax(maxa,err / tola);
5455: }
5456: if (tolr>0.) {
5457: maxr = PetscMax(maxr,err / tolr);
5458: }
5459: if (tol>0.) {
5460: max = PetscMax(max,err / tol);
5461: }
5462: }
5463: VecRestoreArrayRead(ts->vrtol,&rtol);
5464: } else { /* scalar atol, scalar rtol */
5466: for (i=0; i<n; i++) {
5467: SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5468: err = PetscAbsScalar(e[i]);
5469: tola = ts->atol;
5470: tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5471: tol = tola+tolr;
5472: if (tola>0.) {
5473: maxa = PetscMax(maxa,err / tola);
5474: }
5475: if (tolr>0.) {
5476: maxr = PetscMax(maxr,err / tolr);
5477: }
5478: if (tol>0.) {
5479: max = PetscMax(max,err / tol);
5480: }
5481: }
5482: }
5483: VecRestoreArrayRead(E,&e);
5484: VecRestoreArrayRead(U,&u);
5485: VecRestoreArrayRead(Y,&y);
5486: err_loc[0] = max;
5487: err_loc[1] = maxa;
5488: err_loc[2] = maxr;
5489: MPIU_Allreduce(err_loc,err_glb,3,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)ts));
5490: gmax = err_glb[0];
5491: gmaxa = err_glb[1];
5492: gmaxr = err_glb[2];
5494: *norm = gmax;
5495: *norma = gmaxa;
5496: *normr = gmaxr;
5500: return 0;
5501: }
5503: /*@
5504: TSErrorWeightedENorm - compute a weighted error norm based on supplied absolute and relative tolerances
5506: Collective on TS
5508: Input Parameters:
5509: + ts - time stepping context
5510: . E - error vector
5511: . U - state vector, usually ts->vec_sol
5512: . Y - state vector, previous time step
5513: - wnormtype - norm type, either NORM_2 or NORM_INFINITY
5515: Output Parameters:
5516: + norm - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5517: . norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5518: - normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5520: Options Database Keys:
5521: . -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5523: Level: developer
5525: .seealso: TSErrorWeightedENormInfinity(), TSErrorWeightedENorm2(), TSErrorWeightedNormInfinity(), TSErrorWeightedNorm2()
5526: @*/
5527: PetscErrorCode TSErrorWeightedENorm(TS ts,Vec E,Vec U,Vec Y,NormType wnormtype,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5528: {
5529: if (wnormtype == NORM_2) {
5530: TSErrorWeightedENorm2(ts,E,U,Y,norm,norma,normr);
5531: } else if (wnormtype == NORM_INFINITY) {
5532: TSErrorWeightedENormInfinity(ts,E,U,Y,norm,norma,normr);
5533: } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]);
5534: return 0;
5535: }
5537: /*@
5538: TSSetCFLTimeLocal - Set the local CFL constraint relative to forward Euler
5540: Logically Collective on TS
5542: Input Parameters:
5543: + ts - time stepping context
5544: - cfltime - maximum stable time step if using forward Euler (value can be different on each process)
5546: Note:
5547: After calling this function, the global CFL time can be obtained by calling TSGetCFLTime()
5549: Level: intermediate
5551: .seealso: TSGetCFLTime(), TSADAPTCFL
5552: @*/
5553: PetscErrorCode TSSetCFLTimeLocal(TS ts,PetscReal cfltime)
5554: {
5556: ts->cfltime_local = cfltime;
5557: ts->cfltime = -1.;
5558: return 0;
5559: }
5561: /*@
5562: TSGetCFLTime - Get the maximum stable time step according to CFL criteria applied to forward Euler
5564: Collective on TS
5566: Input Parameter:
5567: . ts - time stepping context
5569: Output Parameter:
5570: . cfltime - maximum stable time step for forward Euler
5572: Level: advanced
5574: .seealso: TSSetCFLTimeLocal()
5575: @*/
5576: PetscErrorCode TSGetCFLTime(TS ts,PetscReal *cfltime)
5577: {
5578: if (ts->cfltime < 0) {
5579: MPIU_Allreduce(&ts->cfltime_local,&ts->cfltime,1,MPIU_REAL,MPIU_MIN,PetscObjectComm((PetscObject)ts));
5580: }
5581: *cfltime = ts->cfltime;
5582: return 0;
5583: }
5585: /*@
5586: TSVISetVariableBounds - Sets the lower and upper bounds for the solution vector. xl <= x <= xu
5588: Input Parameters:
5589: + ts - the TS context.
5590: . xl - lower bound.
5591: - xu - upper bound.
5593: Notes:
5594: If this routine is not called then the lower and upper bounds are set to
5595: PETSC_NINFINITY and PETSC_INFINITY respectively during SNESSetUp().
5597: Level: advanced
5599: @*/
5600: PetscErrorCode TSVISetVariableBounds(TS ts, Vec xl, Vec xu)
5601: {
5602: SNES snes;
5604: TSGetSNES(ts,&snes);
5605: SNESVISetVariableBounds(snes,xl,xu);
5606: return 0;
5607: }
5609: /*@
5610: TSComputeLinearStability - computes the linear stability function at a point
5612: Collective on TS
5614: Input Parameters:
5615: + ts - the TS context
5616: - xr,xi - real and imaginary part of input arguments
5618: Output Parameters:
5619: . yr,yi - real and imaginary part of function value
5621: Level: developer
5623: .seealso: TSSetRHSFunction(), TSComputeIFunction()
5624: @*/
5625: PetscErrorCode TSComputeLinearStability(TS ts,PetscReal xr,PetscReal xi,PetscReal *yr,PetscReal *yi)
5626: {
5629: (*ts->ops->linearstability)(ts,xr,xi,yr,yi);
5630: return 0;
5631: }
5633: /*@
5634: TSRestartStep - Flags the solver to restart the next step
5636: Collective on TS
5638: Input Parameter:
5639: . ts - the TS context obtained from TSCreate()
5641: Level: advanced
5643: Notes:
5644: Multistep methods like BDF or Runge-Kutta methods with FSAL property require restarting the solver in the event of
5645: discontinuities. These discontinuities may be introduced as a consequence of explicitly modifications to the solution
5646: vector (which PETSc attempts to detect and handle) or problem coefficients (which PETSc is not able to detect). For
5647: the sake of correctness and maximum safety, users are expected to call TSRestart() whenever they introduce
5648: discontinuities in callback routines (e.g. prestep and poststep routines, or implicit/rhs function routines with
5649: discontinuous source terms).
5651: .seealso: TSSolve(), TSSetPreStep(), TSSetPostStep()
5652: @*/
5653: PetscErrorCode TSRestartStep(TS ts)
5654: {
5656: ts->steprestart = PETSC_TRUE;
5657: return 0;
5658: }
5660: /*@
5661: TSRollBack - Rolls back one time step
5663: Collective on TS
5665: Input Parameter:
5666: . ts - the TS context obtained from TSCreate()
5668: Level: advanced
5670: .seealso: TSCreate(), TSSetUp(), TSDestroy(), TSSolve(), TSSetPreStep(), TSSetPreStage(), TSInterpolate()
5671: @*/
5672: PetscErrorCode TSRollBack(TS ts)
5673: {
5677: (*ts->ops->rollback)(ts);
5678: ts->time_step = ts->ptime - ts->ptime_prev;
5679: ts->ptime = ts->ptime_prev;
5680: ts->ptime_prev = ts->ptime_prev_rollback;
5681: ts->steps--;
5682: ts->steprollback = PETSC_TRUE;
5683: return 0;
5684: }
5686: /*@
5687: TSGetStages - Get the number of stages and stage values
5689: Input Parameter:
5690: . ts - the TS context obtained from TSCreate()
5692: Output Parameters:
5693: + ns - the number of stages
5694: - Y - the current stage vectors
5696: Level: advanced
5698: Notes: Both ns and Y can be NULL.
5700: .seealso: TSCreate()
5701: @*/
5702: PetscErrorCode TSGetStages(TS ts,PetscInt *ns,Vec **Y)
5703: {
5707: if (!ts->ops->getstages) {
5708: if (ns) *ns = 0;
5709: if (Y) *Y = NULL;
5710: } else {
5711: (*ts->ops->getstages)(ts,ns,Y);
5712: }
5713: return 0;
5714: }
5716: /*@C
5717: TSComputeIJacobianDefaultColor - Computes the Jacobian using finite differences and coloring to exploit matrix sparsity.
5719: Collective on SNES
5721: Input Parameters:
5722: + ts - the TS context
5723: . t - current timestep
5724: . U - state vector
5725: . Udot - time derivative of state vector
5726: . shift - shift to apply, see note below
5727: - ctx - an optional user context
5729: Output Parameters:
5730: + J - Jacobian matrix (not altered in this routine)
5731: - B - newly computed Jacobian matrix to use with preconditioner (generally the same as J)
5733: Level: intermediate
5735: Notes:
5736: If F(t,U,Udot)=0 is the DAE, the required Jacobian is
5738: dF/dU + shift*dF/dUdot
5740: Most users should not need to explicitly call this routine, as it
5741: is used internally within the nonlinear solvers.
5743: This will first try to get the coloring from the DM. If the DM type has no coloring
5744: routine, then it will try to get the coloring from the matrix. This requires that the
5745: matrix have nonzero entries precomputed.
5747: .seealso: TSSetIJacobian(), MatFDColoringCreate(), MatFDColoringSetFunction()
5748: @*/
5749: PetscErrorCode TSComputeIJacobianDefaultColor(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat J,Mat B,void *ctx)
5750: {
5751: SNES snes;
5752: MatFDColoring color;
5753: PetscBool hascolor, matcolor = PETSC_FALSE;
5755: PetscOptionsGetBool(((PetscObject)ts)->options,((PetscObject) ts)->prefix, "-ts_fd_color_use_mat", &matcolor, NULL);
5756: PetscObjectQuery((PetscObject) B, "TSMatFDColoring", (PetscObject *) &color);
5757: if (!color) {
5758: DM dm;
5759: ISColoring iscoloring;
5761: TSGetDM(ts, &dm);
5762: DMHasColoring(dm, &hascolor);
5763: if (hascolor && !matcolor) {
5764: DMCreateColoring(dm, IS_COLORING_GLOBAL, &iscoloring);
5765: MatFDColoringCreate(B, iscoloring, &color);
5766: MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts);
5767: MatFDColoringSetFromOptions(color);
5768: MatFDColoringSetUp(B, iscoloring, color);
5769: ISColoringDestroy(&iscoloring);
5770: } else {
5771: MatColoring mc;
5773: MatColoringCreate(B, &mc);
5774: MatColoringSetDistance(mc, 2);
5775: MatColoringSetType(mc, MATCOLORINGSL);
5776: MatColoringSetFromOptions(mc);
5777: MatColoringApply(mc, &iscoloring);
5778: MatColoringDestroy(&mc);
5779: MatFDColoringCreate(B, iscoloring, &color);
5780: MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts);
5781: MatFDColoringSetFromOptions(color);
5782: MatFDColoringSetUp(B, iscoloring, color);
5783: ISColoringDestroy(&iscoloring);
5784: }
5785: PetscObjectCompose((PetscObject) B, "TSMatFDColoring", (PetscObject) color);
5786: PetscObjectDereference((PetscObject) color);
5787: }
5788: TSGetSNES(ts, &snes);
5789: MatFDColoringApply(B, color, U, snes);
5790: if (J != B) {
5791: MatAssemblyBegin(J, MAT_FINAL_ASSEMBLY);
5792: MatAssemblyEnd(J, MAT_FINAL_ASSEMBLY);
5793: }
5794: return 0;
5795: }
5797: /*@
5798: TSSetFunctionDomainError - Set a function that tests if the current state vector is valid
5800: Input Parameters:
5801: + ts - the TS context
5802: - func - function called within TSFunctionDomainError
5804: Calling sequence of func:
5805: $ PetscErrorCode func(TS ts,PetscReal time,Vec state,PetscBool reject)
5807: + ts - the TS context
5808: . time - the current time (of the stage)
5809: . state - the state to check if it is valid
5810: - reject - (output parameter) PETSC_FALSE if the state is acceptable, PETSC_TRUE if not acceptable
5812: Level: intermediate
5814: Notes:
5815: If an implicit ODE solver is being used then, in addition to providing this routine, the
5816: user's code should call SNESSetFunctionDomainError() when domain errors occur during
5817: function evaluations where the functions are provided by TSSetIFunction() or TSSetRHSFunction().
5818: Use TSGetSNES() to obtain the SNES object
5820: Developer Notes:
5821: The naming of this function is inconsistent with the SNESSetFunctionDomainError()
5822: since one takes a function pointer and the other does not.
5824: .seealso: TSAdaptCheckStage(), TSFunctionDomainError(), SNESSetFunctionDomainError(), TSGetSNES()
5825: @*/
5827: PetscErrorCode TSSetFunctionDomainError(TS ts, PetscErrorCode (*func)(TS,PetscReal,Vec,PetscBool*))
5828: {
5830: ts->functiondomainerror = func;
5831: return 0;
5832: }
5834: /*@
5835: TSFunctionDomainError - Checks if the current state is valid
5837: Input Parameters:
5838: + ts - the TS context
5839: . stagetime - time of the simulation
5840: - Y - state vector to check.
5842: Output Parameter:
5843: . accept - Set to PETSC_FALSE if the current state vector is valid.
5845: Note:
5846: This function is called by the TS integration routines and calls the user provided function (set with TSSetFunctionDomainError())
5847: to check if the current state is valid.
5849: Level: developer
5851: .seealso: TSSetFunctionDomainError()
5852: @*/
5853: PetscErrorCode TSFunctionDomainError(TS ts,PetscReal stagetime,Vec Y,PetscBool* accept)
5854: {
5856: *accept = PETSC_TRUE;
5857: if (ts->functiondomainerror) {
5858: PetscStackCallStandard((*ts->functiondomainerror),ts,stagetime,Y,accept);
5859: }
5860: return 0;
5861: }
5863: /*@C
5864: TSClone - This function clones a time step object.
5866: Collective
5868: Input Parameter:
5869: . tsin - The input TS
5871: Output Parameter:
5872: . tsout - The output TS (cloned)
5874: Notes:
5875: This function is used to create a clone of a TS object. It is used in ARKIMEX for initializing the slope for first stage explicit methods. It will likely be replaced in the future with a mechanism of switching methods on the fly.
5877: When using TSDestroy() on a clone the user has to first reset the correct TS reference in the embedded SNES object: e.g.: by running SNES snes_dup=NULL; TSGetSNES(ts,&snes_dup); TSSetSNES(ts,snes_dup);
5879: Level: developer
5881: .seealso: TSCreate(), TSSetType(), TSSetUp(), TSDestroy(), TSSetProblemType()
5882: @*/
5883: PetscErrorCode TSClone(TS tsin, TS *tsout)
5884: {
5885: TS t;
5886: SNES snes_start;
5887: DM dm;
5888: TSType type;
5891: *tsout = NULL;
5893: PetscHeaderCreate(t, TS_CLASSID, "TS", "Time stepping", "TS", PetscObjectComm((PetscObject)tsin), TSDestroy, TSView);
5895: /* General TS description */
5896: t->numbermonitors = 0;
5897: t->monitorFrequency = 1;
5898: t->setupcalled = 0;
5899: t->ksp_its = 0;
5900: t->snes_its = 0;
5901: t->nwork = 0;
5902: t->rhsjacobian.time = PETSC_MIN_REAL;
5903: t->rhsjacobian.scale = 1.;
5904: t->ijacobian.shift = 1.;
5906: TSGetSNES(tsin,&snes_start);
5907: TSSetSNES(t,snes_start);
5909: TSGetDM(tsin,&dm);
5910: TSSetDM(t,dm);
5912: t->adapt = tsin->adapt;
5913: PetscObjectReference((PetscObject)t->adapt);
5915: t->trajectory = tsin->trajectory;
5916: PetscObjectReference((PetscObject)t->trajectory);
5918: t->event = tsin->event;
5919: if (t->event) t->event->refct++;
5921: t->problem_type = tsin->problem_type;
5922: t->ptime = tsin->ptime;
5923: t->ptime_prev = tsin->ptime_prev;
5924: t->time_step = tsin->time_step;
5925: t->max_time = tsin->max_time;
5926: t->steps = tsin->steps;
5927: t->max_steps = tsin->max_steps;
5928: t->equation_type = tsin->equation_type;
5929: t->atol = tsin->atol;
5930: t->rtol = tsin->rtol;
5931: t->max_snes_failures = tsin->max_snes_failures;
5932: t->max_reject = tsin->max_reject;
5933: t->errorifstepfailed = tsin->errorifstepfailed;
5935: TSGetType(tsin,&type);
5936: TSSetType(t,type);
5938: t->vec_sol = NULL;
5940: t->cfltime = tsin->cfltime;
5941: t->cfltime_local = tsin->cfltime_local;
5942: t->exact_final_time = tsin->exact_final_time;
5944: PetscMemcpy(t->ops,tsin->ops,sizeof(struct _TSOps));
5946: if (((PetscObject)tsin)->fortran_func_pointers) {
5947: PetscInt i;
5948: PetscMalloc((10)*sizeof(void(*)(void)),&((PetscObject)t)->fortran_func_pointers);
5949: for (i=0; i<10; i++) {
5950: ((PetscObject)t)->fortran_func_pointers[i] = ((PetscObject)tsin)->fortran_func_pointers[i];
5951: }
5952: }
5953: *tsout = t;
5954: return 0;
5955: }
5957: static PetscErrorCode RHSWrapperFunction_TSRHSJacobianTest(void* ctx,Vec x,Vec y)
5958: {
5959: TS ts = (TS) ctx;
5961: TSComputeRHSFunction(ts,0,x,y);
5962: return 0;
5963: }
5965: /*@
5966: TSRHSJacobianTest - Compares the multiply routine provided to the MATSHELL with differencing on the TS given RHS function.
5968: Logically Collective on TS
5970: Input Parameters:
5971: TS - the time stepping routine
5973: Output Parameter:
5974: . flg - PETSC_TRUE if the multiply is likely correct
5976: Options Database:
5977: . -ts_rhs_jacobian_test_mult -mat_shell_test_mult_view - run the test at each timestep of the integrator
5979: Level: advanced
5981: Notes:
5982: This only works for problems defined only the RHS function and Jacobian NOT IFunction and IJacobian
5984: .seealso: MatCreateShell(), MatShellGetContext(), MatShellGetOperation(), MatShellTestMultTranspose(), TSRHSJacobianTestTranspose()
5985: @*/
5986: PetscErrorCode TSRHSJacobianTest(TS ts,PetscBool *flg)
5987: {
5988: Mat J,B;
5989: TSRHSJacobian func;
5990: void* ctx;
5992: TSGetRHSJacobian(ts,&J,&B,&func,&ctx);
5993: (*func)(ts,0.0,ts->vec_sol,J,B,ctx);
5994: MatShellTestMult(J,RHSWrapperFunction_TSRHSJacobianTest,ts->vec_sol,ts,flg);
5995: return 0;
5996: }
5998: /*@C
5999: TSRHSJacobianTestTranspose - Compares the multiply transpose routine provided to the MATSHELL with differencing on the TS given RHS function.
6001: Logically Collective on TS
6003: Input Parameters:
6004: TS - the time stepping routine
6006: Output Parameter:
6007: . flg - PETSC_TRUE if the multiply is likely correct
6009: Options Database:
6010: . -ts_rhs_jacobian_test_mult_transpose -mat_shell_test_mult_transpose_view - run the test at each timestep of the integrator
6012: Notes:
6013: This only works for problems defined only the RHS function and Jacobian NOT IFunction and IJacobian
6015: Level: advanced
6017: .seealso: MatCreateShell(), MatShellGetContext(), MatShellGetOperation(), MatShellTestMultTranspose(), TSRHSJacobianTest()
6018: @*/
6019: PetscErrorCode TSRHSJacobianTestTranspose(TS ts,PetscBool *flg)
6020: {
6021: Mat J,B;
6022: void *ctx;
6023: TSRHSJacobian func;
6025: TSGetRHSJacobian(ts,&J,&B,&func,&ctx);
6026: (*func)(ts,0.0,ts->vec_sol,J,B,ctx);
6027: MatShellTestMultTranspose(J,RHSWrapperFunction_TSRHSJacobianTest,ts->vec_sol,ts,flg);
6028: return 0;
6029: }
6031: /*@
6032: TSSetUseSplitRHSFunction - Use the split RHSFunction when a multirate method is used.
6034: Logically collective
6036: Input Parameters:
6037: + ts - timestepping context
6038: - use_splitrhsfunction - PETSC_TRUE indicates that the split RHSFunction will be used
6040: Options Database:
6041: . -ts_use_splitrhsfunction - <true,false>
6043: Notes:
6044: This is only useful for multirate methods
6046: Level: intermediate
6048: .seealso: TSGetUseSplitRHSFunction()
6049: @*/
6050: PetscErrorCode TSSetUseSplitRHSFunction(TS ts, PetscBool use_splitrhsfunction)
6051: {
6053: ts->use_splitrhsfunction = use_splitrhsfunction;
6054: return 0;
6055: }
6057: /*@
6058: TSGetUseSplitRHSFunction - Gets whether to use the split RHSFunction when a multirate method is used.
6060: Not collective
6062: Input Parameter:
6063: . ts - timestepping context
6065: Output Parameter:
6066: . use_splitrhsfunction - PETSC_TRUE indicates that the split RHSFunction will be used
6068: Level: intermediate
6070: .seealso: TSSetUseSplitRHSFunction()
6071: @*/
6072: PetscErrorCode TSGetUseSplitRHSFunction(TS ts, PetscBool *use_splitrhsfunction)
6073: {
6075: *use_splitrhsfunction = ts->use_splitrhsfunction;
6076: return 0;
6077: }
6079: /*@
6080: TSSetMatStructure - sets the relationship between the nonzero structure of the RHS Jacobian matrix to the IJacobian matrix.
6082: Logically Collective on ts
6084: Input Parameters:
6085: + ts - the time-stepper
6086: - str - the structure (the default is UNKNOWN_NONZERO_PATTERN)
6088: Level: intermediate
6090: Notes:
6091: When the relationship between the nonzero structures is known and supplied the solution process can be much faster
6093: .seealso: MatAXPY(), MatStructure
6094: @*/
6095: PetscErrorCode TSSetMatStructure(TS ts,MatStructure str)
6096: {
6098: ts->axpy_pattern = str;
6099: return 0;
6100: }