xref: /petsc/src/ts/interface/ts.c (revision 9588d0235bf1a8b206fdb19328f30b65bdeca28e)
1 #include <petsc/private/tsimpl.h>        /*I "petscts.h"  I*/
2 #include <petscdmshell.h>
3 #include <petscdmda.h>
4 #include <petscviewer.h>
5 #include <petscdraw.h>
6 #include <petscconvest.h>
7 
8 #define SkipSmallValue(a,b,tol) if (PetscAbsScalar(a)< tol || PetscAbsScalar(b)< tol) continue;
9 
10 /* Logging support */
11 PetscClassId  TS_CLASSID, DMTS_CLASSID;
12 PetscLogEvent TS_Step, TS_PseudoComputeTimeStep, TS_FunctionEval, TS_JacobianEval;
13 
14 const char *const TSExactFinalTimeOptions[] = {"UNSPECIFIED","STEPOVER","INTERPOLATE","MATCHSTEP","TSExactFinalTimeOption","TS_EXACTFINALTIME_",NULL};
15 
16 static PetscErrorCode TSAdaptSetDefaultType(TSAdapt adapt,TSAdaptType default_type)
17 {
18   PetscFunctionBegin;
19   PetscValidHeaderSpecific(adapt,TSADAPT_CLASSID,1);
20   PetscValidCharPointer(default_type,2);
21   if (!((PetscObject)adapt)->type_name) PetscCall(TSAdaptSetType(adapt,default_type));
22   PetscFunctionReturn(0);
23 }
24 
25 /*@
26    TSSetFromOptions - Sets various TS parameters from user options.
27 
28    Collective on TS
29 
30    Input Parameter:
31 .  ts - the TS context obtained from TSCreate()
32 
33    Options Database Keys:
34 +  -ts_type <type> - TSEULER, TSBEULER, TSSUNDIALS, TSPSEUDO, TSCN, TSRK, TSTHETA, TSALPHA, TSGLLE, TSSSP, TSGLEE, TSBSYMP, TSIRK
35 .  -ts_save_trajectory - checkpoint the solution at each time-step
36 .  -ts_max_time <time> - maximum time to compute to
37 .  -ts_time_span <t0,...tf> - sets the time span, solutions are computed and stored for each indicated time
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-%%03" PetscInt_FMT ".vts (filename-%%03" PetscInt_FMT ".vtu)
67 -  -ts_monitor_envelope - determine maximum and minimum value of each component of the solution over the solution time
68 
69    Notes:
70      See SNESSetFromOptions() and KSPSetFromOptions() for how to control the nonlinear and linear solves used by the time-stepper.
71 
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
75 
76    Developer Note:
77      We should unify all the -ts_monitor options in the way that -xxx_view has been unified
78 
79    Level: beginner
80 
81 .seealso: `TSGetType()`
82 @*/
83 PetscErrorCode  TSSetFromOptions(TS ts)
84 {
85   PetscBool              opt,flg,tflg;
86   char                   monfilename[PETSC_MAX_PATH_LEN];
87   PetscReal              time_step,tspan[100];
88   PetscInt               nt = PETSC_STATIC_ARRAY_LENGTH(tspan);
89   TSExactFinalTimeOption eftopt;
90   char                   dir[16];
91   TSIFunction            ifun;
92   const char             *defaultType;
93   char                   typeName[256];
94 
95   PetscFunctionBegin;
96   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
97 
98   PetscCall(TSRegisterAll());
99   PetscCall(TSGetIFunction(ts,NULL,&ifun,NULL));
100 
101   PetscObjectOptionsBegin((PetscObject)ts);
102   if (((PetscObject)ts)->type_name) defaultType = ((PetscObject)ts)->type_name;
103   else defaultType = ifun ? TSBEULER : TSEULER;
104   PetscCall(PetscOptionsFList("-ts_type","TS method","TSSetType",TSList,defaultType,typeName,256,&opt));
105   if (opt) PetscCall(TSSetType(ts,typeName));
106   else     PetscCall(TSSetType(ts,defaultType));
107 
108   /* Handle generic TS options */
109   PetscCall(PetscOptionsDeprecated("-ts_final_time","-ts_max_time","3.10",NULL));
110   PetscCall(PetscOptionsReal("-ts_max_time","Maximum time to run to","TSSetMaxTime",ts->max_time,&ts->max_time,NULL));
111   PetscCall(PetscOptionsRealArray("-ts_time_span","Time span","TSSetTimeSpan",tspan,&nt,&flg));
112   if (flg) PetscCall(TSSetTimeSpan(ts,nt,tspan));
113   PetscCall(PetscOptionsInt("-ts_max_steps","Maximum number of time steps","TSSetMaxSteps",ts->max_steps,&ts->max_steps,NULL));
114   PetscCall(PetscOptionsReal("-ts_init_time","Initial time","TSSetTime",ts->ptime,&ts->ptime,NULL));
115   PetscCall(PetscOptionsReal("-ts_dt","Initial time step","TSSetTimeStep",ts->time_step,&time_step,&flg));
116   if (flg) PetscCall(TSSetTimeStep(ts,time_step));
117   PetscCall(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) PetscCall(TSSetExactFinalTime(ts,eftopt));
119   PetscCall(PetscOptionsInt("-ts_max_snes_failures","Maximum number of nonlinear solve failures","TSSetMaxSNESFailures",ts->max_snes_failures,&ts->max_snes_failures,NULL));
120   PetscCall(PetscOptionsInt("-ts_max_reject","Maximum number of step rejections before step fails","TSSetMaxStepRejections",ts->max_reject,&ts->max_reject,NULL));
121   PetscCall(PetscOptionsBool("-ts_error_if_step_fails","Error if no step succeeds","TSSetErrorIfStepFails",ts->errorifstepfailed,&ts->errorifstepfailed,NULL));
122   PetscCall(PetscOptionsReal("-ts_rtol","Relative tolerance for local truncation error","TSSetTolerances",ts->rtol,&ts->rtol,NULL));
123   PetscCall(PetscOptionsReal("-ts_atol","Absolute tolerance for local truncation error","TSSetTolerances",ts->atol,&ts->atol,NULL));
124 
125   PetscCall(PetscOptionsBool("-ts_rhs_jacobian_test_mult","Test the RHS Jacobian for consistency with RHS at each solve ","None",ts->testjacobian,&ts->testjacobian,NULL));
126   PetscCall(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   PetscCall(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     PetscCall(PetscOptionsBool("-ts_saws_block","Block for SAWs memory snooper at end of TSSolve","PetscObjectSAWsBlock",((PetscObject)ts)->amspublishblock,&flg,&set));
133     if (set) PetscCall(PetscObjectSAWsSetBlock((PetscObject)ts,flg));
134   }
135 #endif
136 
137   /* Monitor options */
138   PetscCall(PetscOptionsInt("-ts_monitor_frequency", "Number of time steps between monitor output", "TSMonitorSetFrequency", ts->monitorFrequency, &ts->monitorFrequency, NULL));
139   PetscCall(TSMonitorSetFromOptions(ts,"-ts_monitor","Monitor time and timestep size","TSMonitorDefault",TSMonitorDefault,NULL));
140   PetscCall(TSMonitorSetFromOptions(ts,"-ts_monitor_extreme","Monitor extreme values of the solution","TSMonitorExtreme",TSMonitorExtreme,NULL));
141   PetscCall(TSMonitorSetFromOptions(ts,"-ts_monitor_solution","View the solution at each timestep","TSMonitorSolution",TSMonitorSolution,NULL));
142   PetscCall(TSMonitorSetFromOptions(ts,"-ts_dmswarm_monitor_moments","Monitor moments of particle distribution","TSDMSwarmMonitorMoments",TSDMSwarmMonitorMoments,NULL));
143 
144   PetscCall(PetscOptionsString("-ts_monitor_python","Use Python function","TSMonitorSet",NULL,monfilename,sizeof(monfilename),&flg));
145   if (flg) PetscCall(PetscPythonMonitorSet((PetscObject)ts,monfilename));
146 
147   PetscCall(PetscOptionsName("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",&opt));
148   if (opt) {
149     PetscInt       howoften = 1;
150     DM             dm;
151     PetscBool      net;
152 
153     PetscCall(PetscOptionsInt("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",howoften,&howoften,NULL));
154     PetscCall(TSGetDM(ts,&dm));
155     PetscCall(PetscObjectTypeCompare((PetscObject)dm,DMNETWORK,&net));
156     if (net) {
157       TSMonitorLGCtxNetwork ctx;
158       PetscCall(TSMonitorLGCtxNetworkCreate(ts,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx));
159       PetscCall(TSMonitorSet(ts,TSMonitorLGCtxNetworkSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxNetworkDestroy));
160       PetscCall(PetscOptionsBool("-ts_monitor_lg_solution_semilogy","Plot the solution with a semi-log axis","",ctx->semilogy,&ctx->semilogy,NULL));
161     } else {
162       TSMonitorLGCtx ctx;
163       PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
164       PetscCall(TSMonitorSet(ts,TSMonitorLGSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
165     }
166   }
167 
168   PetscCall(PetscOptionsName("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",&opt));
169   if (opt) {
170     TSMonitorLGCtx ctx;
171     PetscInt       howoften = 1;
172 
173     PetscCall(PetscOptionsInt("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",howoften,&howoften,NULL));
174     PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
175     PetscCall(TSMonitorSet(ts,TSMonitorLGError,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
176   }
177   PetscCall(TSMonitorSetFromOptions(ts,"-ts_monitor_error","View the error at each timestep","TSMonitorError",TSMonitorError,NULL));
178 
179   PetscCall(PetscOptionsName("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",&opt));
180   if (opt) {
181     TSMonitorLGCtx ctx;
182     PetscInt       howoften = 1;
183 
184     PetscCall(PetscOptionsInt("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",howoften,&howoften,NULL));
185     PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
186     PetscCall(TSMonitorSet(ts,TSMonitorLGTimeStep,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
187   }
188   PetscCall(PetscOptionsName("-ts_monitor_lg_timestep_log","Monitor log timestep size graphically","TSMonitorLGTimeStep",&opt));
189   if (opt) {
190     TSMonitorLGCtx ctx;
191     PetscInt       howoften = 1;
192 
193     PetscCall(PetscOptionsInt("-ts_monitor_lg_timestep_log","Monitor log timestep size graphically","TSMonitorLGTimeStep",howoften,&howoften,NULL));
194     PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
195     PetscCall(TSMonitorSet(ts,TSMonitorLGTimeStep,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
196     ctx->semilogy = PETSC_TRUE;
197   }
198 
199   PetscCall(PetscOptionsName("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",&opt));
200   if (opt) {
201     TSMonitorLGCtx ctx;
202     PetscInt       howoften = 1;
203 
204     PetscCall(PetscOptionsInt("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",howoften,&howoften,NULL));
205     PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
206     PetscCall(TSMonitorSet(ts,TSMonitorLGSNESIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
207   }
208   PetscCall(PetscOptionsName("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",&opt));
209   if (opt) {
210     TSMonitorLGCtx ctx;
211     PetscInt       howoften = 1;
212 
213     PetscCall(PetscOptionsInt("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",howoften,&howoften,NULL));
214     PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,400,300,howoften,&ctx));
215     PetscCall(TSMonitorSet(ts,TSMonitorLGKSPIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy));
216   }
217   PetscCall(PetscOptionsName("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",&opt));
218   if (opt) {
219     TSMonitorSPEigCtx ctx;
220     PetscInt          howoften = 1;
221 
222     PetscCall(PetscOptionsInt("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",howoften,&howoften,NULL));
223     PetscCall(TSMonitorSPEigCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx));
224     PetscCall(TSMonitorSet(ts,TSMonitorSPEig,ctx,(PetscErrorCode (*)(void**))TSMonitorSPEigCtxDestroy));
225   }
226   PetscCall(PetscOptionsName("-ts_monitor_sp_swarm","Display particle phase from the DMSwarm","TSMonitorSPSwarm",&opt));
227   if (opt) {
228     TSMonitorSPCtx  ctx;
229     PetscInt        howoften = 1, retain = 0;
230     PetscBool       phase = PETSC_TRUE, create = PETSC_TRUE;
231 
232     for (PetscInt i = 0; i < ts->numbermonitors; ++i) if (ts->monitor[i] == TSMonitorSPSwarmSolution) {create = PETSC_FALSE;break;}
233     if (create) {
234       PetscCall(PetscOptionsInt("-ts_monitor_sp_swarm","Display particles phase from the DMSwarm", "TSMonitorSPSwarm", howoften, &howoften, NULL));
235       PetscCall(PetscOptionsInt("-ts_monitor_sp_swarm_retain", "Retain n points plotted to show trajectory, -1 for all points", "TSMonitorSPSwarm", retain, &retain, NULL));
236       PetscCall(PetscOptionsBool("-ts_monitor_sp_swarm_phase", "Plot in phase space rather than coordinate space", "TSMonitorSPSwarm", phase, &phase, NULL));
237       PetscCall(TSMonitorSPCtxCreate(PetscObjectComm((PetscObject) ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, retain, phase, &ctx));
238       PetscCall(TSMonitorSet(ts, TSMonitorSPSwarmSolution, ctx, (PetscErrorCode (*)(void**))TSMonitorSPCtxDestroy));
239     }
240   }
241   opt  = PETSC_FALSE;
242   PetscCall(PetscOptionsName("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",&opt));
243   if (opt) {
244     TSMonitorDrawCtx ctx;
245     PetscInt         howoften = 1;
246 
247     PetscCall(PetscOptionsInt("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",howoften,&howoften,NULL));
248     PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Computed Solution",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx));
249     PetscCall(TSMonitorSet(ts,TSMonitorDrawSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy));
250   }
251   opt  = PETSC_FALSE;
252   PetscCall(PetscOptionsName("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",&opt));
253   if (opt) {
254     TSMonitorDrawCtx ctx;
255     PetscReal        bounds[4];
256     PetscInt         n = 4;
257     PetscDraw        draw;
258     PetscDrawAxis    axis;
259 
260     PetscCall(PetscOptionsRealArray("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",bounds,&n,NULL));
261     PetscCheck(n == 4,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Must provide bounding box of phase field");
262     PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,300,300,1,&ctx));
263     PetscCall(PetscViewerDrawGetDraw(ctx->viewer,0,&draw));
264     PetscCall(PetscViewerDrawGetDrawAxis(ctx->viewer,0,&axis));
265     PetscCall(PetscDrawAxisSetLimits(axis,bounds[0],bounds[2],bounds[1],bounds[3]));
266     PetscCall(PetscDrawAxisSetLabels(axis,"Phase Diagram","Variable 1","Variable 2"));
267     PetscCall(TSMonitorSet(ts,TSMonitorDrawSolutionPhase,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy));
268   }
269   opt  = PETSC_FALSE;
270   PetscCall(PetscOptionsName("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",&opt));
271   if (opt) {
272     TSMonitorDrawCtx ctx;
273     PetscInt         howoften = 1;
274 
275     PetscCall(PetscOptionsInt("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",howoften,&howoften,NULL));
276     PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Error",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx));
277     PetscCall(TSMonitorSet(ts,TSMonitorDrawError,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy));
278   }
279   opt  = PETSC_FALSE;
280   PetscCall(PetscOptionsName("-ts_monitor_draw_solution_function","Monitor solution provided by TSMonitorSetSolutionFunction() graphically","TSMonitorDrawSolutionFunction",&opt));
281   if (opt) {
282     TSMonitorDrawCtx ctx;
283     PetscInt         howoften = 1;
284 
285     PetscCall(PetscOptionsInt("-ts_monitor_draw_solution_function","Monitor solution provided by TSMonitorSetSolutionFunction() graphically","TSMonitorDrawSolutionFunction",howoften,&howoften,NULL));
286     PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),NULL,"Solution provided by user function",PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx));
287     PetscCall(TSMonitorSet(ts,TSMonitorDrawSolutionFunction,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy));
288   }
289 
290   opt  = PETSC_FALSE;
291   PetscCall(PetscOptionsString("-ts_monitor_solution_vtk","Save each time step to a binary file, use filename-%%03" PetscInt_FMT ".vts","TSMonitorSolutionVTK",NULL,monfilename,sizeof(monfilename),&flg));
292   if (flg) {
293     const char *ptr,*ptr2;
294     char       *filetemplate;
295     PetscCheck(monfilename[0],PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03" PetscInt_FMT ".vts");
296     /* Do some cursory validation of the input. */
297     PetscCall(PetscStrstr(monfilename,"%",(char**)&ptr));
298     PetscCheck(ptr,PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03" PetscInt_FMT ".vts");
299     for (ptr++; ptr && *ptr; ptr++) {
300       PetscCall(PetscStrchr("DdiouxX",*ptr,(char**)&ptr2));
301       PetscCheck(ptr2 || (*ptr >= '0' && *ptr <= '9'),PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Invalid file template argument to -ts_monitor_solution_vtk, should look like filename-%%03" PetscInt_FMT ".vts");
302       if (ptr2) break;
303     }
304     PetscCall(PetscStrallocpy(monfilename,&filetemplate));
305     PetscCall(TSMonitorSet(ts,TSMonitorSolutionVTK,filetemplate,(PetscErrorCode (*)(void**))TSMonitorSolutionVTKDestroy));
306   }
307 
308   PetscCall(PetscOptionsString("-ts_monitor_dmda_ray","Display a ray of the solution","None","y=0",dir,sizeof(dir),&flg));
309   if (flg) {
310     TSMonitorDMDARayCtx *rayctx;
311     int                  ray = 0;
312     DMDirection          ddir;
313     DM                   da;
314     PetscMPIInt          rank;
315 
316     PetscCheck(dir[1] == '=',PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Unknown ray %s",dir);
317     if (dir[0] == 'x') ddir = DM_X;
318     else if (dir[0] == 'y') ddir = DM_Y;
319     else SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Unknown ray %s",dir);
320     sscanf(dir+2,"%d",&ray);
321 
322     PetscCall(PetscInfo(((PetscObject)ts),"Displaying DMDA ray %c = %d\n",dir[0],ray));
323     PetscCall(PetscNew(&rayctx));
324     PetscCall(TSGetDM(ts,&da));
325     PetscCall(DMDAGetRay(da,ddir,ray,&rayctx->ray,&rayctx->scatter));
326     PetscCallMPI(MPI_Comm_rank(PetscObjectComm((PetscObject)ts),&rank));
327     if (rank == 0) PetscCall(PetscViewerDrawOpen(PETSC_COMM_SELF,NULL,NULL,0,0,600,300,&rayctx->viewer));
328     rayctx->lgctx = NULL;
329     PetscCall(TSMonitorSet(ts,TSMonitorDMDARay,rayctx,TSMonitorDMDARayDestroy));
330   }
331   PetscCall(PetscOptionsString("-ts_monitor_lg_dmda_ray","Display a ray of the solution","None","x=0",dir,sizeof(dir),&flg));
332   if (flg) {
333     TSMonitorDMDARayCtx *rayctx;
334     int                 ray = 0;
335     DMDirection         ddir;
336     DM                  da;
337     PetscInt            howoften = 1;
338 
339     PetscCheck(dir[1] == '=',PetscObjectComm((PetscObject) ts), PETSC_ERR_ARG_WRONG, "Malformed ray %s", dir);
340     if      (dir[0] == 'x') ddir = DM_X;
341     else if (dir[0] == 'y') ddir = DM_Y;
342     else SETERRQ(PetscObjectComm((PetscObject) ts), PETSC_ERR_ARG_WRONG, "Unknown ray direction %s", dir);
343     sscanf(dir+2, "%d", &ray);
344 
345     PetscCall(PetscInfo(((PetscObject) ts),"Displaying LG DMDA ray %c = %d\n", dir[0], ray));
346     PetscCall(PetscNew(&rayctx));
347     PetscCall(TSGetDM(ts, &da));
348     PetscCall(DMDAGetRay(da, ddir, ray, &rayctx->ray, &rayctx->scatter));
349     PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF,NULL,NULL,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&rayctx->lgctx));
350     PetscCall(TSMonitorSet(ts, TSMonitorLGDMDARay, rayctx, TSMonitorDMDARayDestroy));
351   }
352 
353   PetscCall(PetscOptionsName("-ts_monitor_envelope","Monitor maximum and minimum value of each component of the solution","TSMonitorEnvelope",&opt));
354   if (opt) {
355     TSMonitorEnvelopeCtx ctx;
356 
357     PetscCall(TSMonitorEnvelopeCtxCreate(ts,&ctx));
358     PetscCall(TSMonitorSet(ts,TSMonitorEnvelope,ctx,(PetscErrorCode (*)(void**))TSMonitorEnvelopeCtxDestroy));
359   }
360   flg  = PETSC_FALSE;
361   PetscCall(PetscOptionsBool("-ts_monitor_cancel","Remove all monitors","TSMonitorCancel",flg,&flg,&opt));
362   if (opt && flg) PetscCall(TSMonitorCancel(ts));
363 
364   flg  = PETSC_FALSE;
365   PetscCall(PetscOptionsBool("-ts_fd_color", "Use finite differences with coloring to compute IJacobian", "TSComputeJacobianDefaultColor", flg, &flg, NULL));
366   if (flg) {
367     DM   dm;
368 
369     PetscCall(TSGetDM(ts, &dm));    PetscCall(DMTSUnsetIJacobianContext_Internal(dm));
370     PetscCall(TSSetIJacobian(ts, NULL, NULL, TSComputeIJacobianDefaultColor, NULL));
371     PetscCall(PetscInfo(ts, "Setting default finite difference coloring Jacobian matrix\n"));
372   }
373 
374   /* Handle specific TS options */
375   PetscTryTypeMethod(ts,setfromoptions,PetscOptionsObject);
376 
377   /* Handle TSAdapt options */
378   PetscCall(TSGetAdapt(ts,&ts->adapt));
379   PetscCall(TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type));
380   PetscCall(TSAdaptSetFromOptions(ts->adapt,PetscOptionsObject));
381 
382   /* TS trajectory must be set after TS, since it may use some TS options above */
383   tflg = ts->trajectory ? PETSC_TRUE : PETSC_FALSE;
384   PetscCall(PetscOptionsBool("-ts_save_trajectory","Save the solution at each timestep","TSSetSaveTrajectory",tflg,&tflg,NULL));
385   if (tflg) PetscCall(TSSetSaveTrajectory(ts));
386 
387   PetscCall(TSAdjointSetFromOptions(ts,PetscOptionsObject));
388 
389   /* process any options handlers added with PetscObjectAddOptionsHandler() */
390   PetscCall(PetscObjectProcessOptionsHandlers((PetscObject)ts,PetscOptionsObject));
391   PetscOptionsEnd();
392 
393   if (ts->trajectory) PetscCall(TSTrajectorySetFromOptions(ts->trajectory,ts));
394 
395   /* why do we have to do this here and not during TSSetUp? */
396   PetscCall(TSGetSNES(ts,&ts->snes));
397   if (ts->problem_type == TS_LINEAR) {
398     PetscCall(PetscObjectTypeCompareAny((PetscObject)ts->snes,&flg,SNESKSPONLY,SNESKSPTRANSPOSEONLY,""));
399     if (!flg) PetscCall(SNESSetType(ts->snes,SNESKSPONLY));
400   }
401   PetscCall(SNESSetFromOptions(ts->snes));
402   PetscFunctionReturn(0);
403 }
404 
405 /*@
406    TSGetTrajectory - Gets the trajectory from a TS if it exists
407 
408    Collective on TS
409 
410    Input Parameters:
411 .  ts - the TS context obtained from TSCreate()
412 
413    Output Parameters:
414 .  tr - the TSTrajectory object, if it exists
415 
416    Note: This routine should be called after all TS options have been set
417 
418    Level: advanced
419 
420 .seealso: `TSGetTrajectory()`, `TSAdjointSolve()`, `TSTrajectory`, `TSTrajectoryCreate()`
421 
422 @*/
423 PetscErrorCode  TSGetTrajectory(TS ts,TSTrajectory *tr)
424 {
425   PetscFunctionBegin;
426   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
427   *tr = ts->trajectory;
428   PetscFunctionReturn(0);
429 }
430 
431 /*@
432    TSSetSaveTrajectory - Causes the TS to save its solutions as it iterates forward in time in a TSTrajectory object
433 
434    Collective on TS
435 
436    Input Parameter:
437 .  ts - the TS context obtained from TSCreate()
438 
439    Options Database:
440 +  -ts_save_trajectory - saves the trajectory to a file
441 -  -ts_trajectory_type type - set trajectory type
442 
443 Note: This routine should be called after all TS options have been set
444 
445     The TSTRAJECTORYVISUALIZATION files can be loaded into Python with $PETSC_DIR/lib/petsc/bin/PetscBinaryIOTrajectory.py and
446    MATLAB with $PETSC_DIR/share/petsc/matlab/PetscReadBinaryTrajectory.m
447 
448    Level: intermediate
449 
450 .seealso: `TSGetTrajectory()`, `TSAdjointSolve()`
451 
452 @*/
453 PetscErrorCode  TSSetSaveTrajectory(TS ts)
454 {
455   PetscFunctionBegin;
456   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
457   if (!ts->trajectory) PetscCall(TSTrajectoryCreate(PetscObjectComm((PetscObject)ts),&ts->trajectory));
458   PetscFunctionReturn(0);
459 }
460 
461 /*@
462    TSResetTrajectory - Destroys and recreates the internal TSTrajectory object
463 
464    Collective on TS
465 
466    Input Parameters:
467 .  ts - the TS context obtained from TSCreate()
468 
469    Level: intermediate
470 
471 .seealso: `TSGetTrajectory()`, `TSAdjointSolve()`, `TSRemoveTrajectory()`
472 
473 @*/
474 PetscErrorCode  TSResetTrajectory(TS ts)
475 {
476   PetscFunctionBegin;
477   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
478   if (ts->trajectory) {
479     PetscCall(TSTrajectoryDestroy(&ts->trajectory));
480     PetscCall(TSTrajectoryCreate(PetscObjectComm((PetscObject)ts),&ts->trajectory));
481   }
482   PetscFunctionReturn(0);
483 }
484 
485 /*@
486    TSRemoveTrajectory - Destroys and removes the internal TSTrajectory object from TS
487 
488    Collective on TS
489 
490    Input Parameters:
491 .  ts - the TS context obtained from TSCreate()
492 
493    Level: intermediate
494 
495 .seealso: `TSResetTrajectory()`, `TSAdjointSolve()`
496 
497 @*/
498 PetscErrorCode TSRemoveTrajectory(TS ts)
499 {
500   PetscFunctionBegin;
501   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
502   if (ts->trajectory) PetscCall(TSTrajectoryDestroy(&ts->trajectory));
503   PetscFunctionReturn(0);
504 }
505 
506 /*@
507    TSComputeRHSJacobian - Computes the Jacobian matrix that has been
508       set with TSSetRHSJacobian().
509 
510    Collective on TS
511 
512    Input Parameters:
513 +  ts - the TS context
514 .  t - current timestep
515 -  U - input vector
516 
517    Output Parameters:
518 +  A - Jacobian matrix
519 -  B - optional preconditioning matrix
520 
521    Notes:
522    Most users should not need to explicitly call this routine, as it
523    is used internally within the nonlinear solvers.
524 
525    Level: developer
526 
527 .seealso: `TSSetRHSJacobian()`, `KSPSetOperators()`
528 @*/
529 PetscErrorCode  TSComputeRHSJacobian(TS ts,PetscReal t,Vec U,Mat A,Mat B)
530 {
531   PetscObjectState Ustate;
532   PetscObjectId    Uid;
533   DM               dm;
534   DMTS             tsdm;
535   TSRHSJacobian    rhsjacobianfunc;
536   void             *ctx;
537   TSRHSFunction    rhsfunction;
538 
539   PetscFunctionBegin;
540   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
541   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
542   PetscCheckSameComm(ts,1,U,3);
543   PetscCall(TSGetDM(ts,&dm));
544   PetscCall(DMGetDMTS(dm,&tsdm));
545   PetscCall(DMTSGetRHSFunction(dm,&rhsfunction,NULL));
546   PetscCall(DMTSGetRHSJacobian(dm,&rhsjacobianfunc,&ctx));
547   PetscCall(PetscObjectStateGet((PetscObject)U,&Ustate));
548   PetscCall(PetscObjectGetId((PetscObject)U,&Uid));
549 
550   if (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) && (rhsfunction != TSComputeRHSFunctionLinear)) PetscFunctionReturn(0);
551 
552   PetscCheck(ts->rhsjacobian.shift == 0.0 || !ts->rhsjacobian.reuse,PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Should not call TSComputeRHSJacobian() on a shifted matrix (shift=%lf) when RHSJacobian is reusable.",(double)ts->rhsjacobian.shift);
553   if (rhsjacobianfunc) {
554     PetscCall(PetscLogEventBegin(TS_JacobianEval,ts,U,A,B));
555     PetscCallBack("TS callback Jacobian",(*rhsjacobianfunc)(ts,t,U,A,B,ctx));
556     ts->rhsjacs++;
557     PetscCall(PetscLogEventEnd(TS_JacobianEval,ts,U,A,B));
558   } else {
559     PetscCall(MatZeroEntries(A));
560     if (B && A != B) PetscCall(MatZeroEntries(B));
561   }
562   ts->rhsjacobian.time  = t;
563   ts->rhsjacobian.shift = 0;
564   ts->rhsjacobian.scale = 1.;
565   PetscCall(PetscObjectGetId((PetscObject)U,&ts->rhsjacobian.Xid));
566   PetscCall(PetscObjectStateGet((PetscObject)U,&ts->rhsjacobian.Xstate));
567   PetscFunctionReturn(0);
568 }
569 
570 /*@
571    TSComputeRHSFunction - Evaluates the right-hand-side function.
572 
573    Collective on TS
574 
575    Input Parameters:
576 +  ts - the TS context
577 .  t - current time
578 -  U - state vector
579 
580    Output Parameter:
581 .  y - right hand side
582 
583    Note:
584    Most users should not need to explicitly call this routine, as it
585    is used internally within the nonlinear solvers.
586 
587    Level: developer
588 
589 .seealso: `TSSetRHSFunction()`, `TSComputeIFunction()`
590 @*/
591 PetscErrorCode TSComputeRHSFunction(TS ts,PetscReal t,Vec U,Vec y)
592 {
593   TSRHSFunction  rhsfunction;
594   TSIFunction    ifunction;
595   void           *ctx;
596   DM             dm;
597 
598   PetscFunctionBegin;
599   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
600   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
601   PetscValidHeaderSpecific(y,VEC_CLASSID,4);
602   PetscCall(TSGetDM(ts,&dm));
603   PetscCall(DMTSGetRHSFunction(dm,&rhsfunction,&ctx));
604   PetscCall(DMTSGetIFunction(dm,&ifunction,NULL));
605 
606   PetscCheck(rhsfunction || ifunction,PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSFunction() and / or TSSetIFunction()");
607 
608   if (rhsfunction) {
609     PetscCall(PetscLogEventBegin(TS_FunctionEval,ts,U,y,0));
610     PetscCall(VecLockReadPush(U));
611     PetscCallBack("TS callback right-hand-side",(*rhsfunction)(ts,t,U,y,ctx));
612     PetscCall(VecLockReadPop(U));
613     ts->rhsfuncs++;
614     PetscCall(PetscLogEventEnd(TS_FunctionEval,ts,U,y,0));
615   } else PetscCall(VecZeroEntries(y));
616   PetscFunctionReturn(0);
617 }
618 
619 /*@
620    TSComputeSolutionFunction - Evaluates the solution function.
621 
622    Collective on TS
623 
624    Input Parameters:
625 +  ts - the TS context
626 -  t - current time
627 
628    Output Parameter:
629 .  U - the solution
630 
631    Note:
632    Most users should not need to explicitly call this routine, as it
633    is used internally within the nonlinear solvers.
634 
635    Level: developer
636 
637 .seealso: `TSSetSolutionFunction()`, `TSSetRHSFunction()`, `TSComputeIFunction()`
638 @*/
639 PetscErrorCode TSComputeSolutionFunction(TS ts,PetscReal t,Vec U)
640 {
641   TSSolutionFunction solutionfunction;
642   void               *ctx;
643   DM                 dm;
644 
645   PetscFunctionBegin;
646   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
647   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
648   PetscCall(TSGetDM(ts,&dm));
649   PetscCall(DMTSGetSolutionFunction(dm,&solutionfunction,&ctx));
650 
651   if (solutionfunction) PetscCallBack("TS callback solution",(*solutionfunction)(ts,t,U,ctx));
652   PetscFunctionReturn(0);
653 }
654 /*@
655    TSComputeForcingFunction - Evaluates the forcing function.
656 
657    Collective on TS
658 
659    Input Parameters:
660 +  ts - the TS context
661 -  t - current time
662 
663    Output Parameter:
664 .  U - the function value
665 
666    Note:
667    Most users should not need to explicitly call this routine, as it
668    is used internally within the nonlinear solvers.
669 
670    Level: developer
671 
672 .seealso: `TSSetSolutionFunction()`, `TSSetRHSFunction()`, `TSComputeIFunction()`
673 @*/
674 PetscErrorCode TSComputeForcingFunction(TS ts,PetscReal t,Vec U)
675 {
676   void              *ctx;
677   DM                 dm;
678   TSForcingFunction  forcing;
679 
680   PetscFunctionBegin;
681   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
682   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
683   PetscCall(TSGetDM(ts,&dm));
684   PetscCall(DMTSGetForcingFunction(dm,&forcing,&ctx));
685 
686   if (forcing) PetscCallBack("TS callback forcing function",(*forcing)(ts,t,U,ctx));
687   PetscFunctionReturn(0);
688 }
689 
690 static PetscErrorCode TSGetRHSVec_Private(TS ts,Vec *Frhs)
691 {
692   Vec            F;
693 
694   PetscFunctionBegin;
695   *Frhs = NULL;
696   PetscCall(TSGetIFunction(ts,&F,NULL,NULL));
697   if (!ts->Frhs) PetscCall(VecDuplicate(F,&ts->Frhs));
698   *Frhs = ts->Frhs;
699   PetscFunctionReturn(0);
700 }
701 
702 PetscErrorCode TSGetRHSMats_Private(TS ts,Mat *Arhs,Mat *Brhs)
703 {
704   Mat            A,B;
705   TSIJacobian    ijacobian;
706 
707   PetscFunctionBegin;
708   if (Arhs) *Arhs = NULL;
709   if (Brhs) *Brhs = NULL;
710   PetscCall(TSGetIJacobian(ts,&A,&B,&ijacobian,NULL));
711   if (Arhs) {
712     if (!ts->Arhs) {
713       if (ijacobian) {
714         PetscCall(MatDuplicate(A,MAT_DO_NOT_COPY_VALUES,&ts->Arhs));
715         PetscCall(TSSetMatStructure(ts,SAME_NONZERO_PATTERN));
716       } else {
717         ts->Arhs = A;
718         PetscCall(PetscObjectReference((PetscObject)A));
719       }
720     } else {
721       PetscBool flg;
722       PetscCall(SNESGetUseMatrixFree(ts->snes,NULL,&flg));
723       /* Handle case where user provided only RHSJacobian and used -snes_mf_operator */
724       if (flg && !ijacobian && ts->Arhs == ts->Brhs) {
725         PetscCall(PetscObjectDereference((PetscObject)ts->Arhs));
726         ts->Arhs = A;
727         PetscCall(PetscObjectReference((PetscObject)A));
728       }
729     }
730     *Arhs = ts->Arhs;
731   }
732   if (Brhs) {
733     if (!ts->Brhs) {
734       if (A != B) {
735         if (ijacobian) {
736           PetscCall(MatDuplicate(B,MAT_DO_NOT_COPY_VALUES,&ts->Brhs));
737         } else {
738           ts->Brhs = B;
739           PetscCall(PetscObjectReference((PetscObject)B));
740         }
741       } else {
742         PetscCall(PetscObjectReference((PetscObject)ts->Arhs));
743         ts->Brhs = ts->Arhs;
744       }
745     }
746     *Brhs = ts->Brhs;
747   }
748   PetscFunctionReturn(0);
749 }
750 
751 /*@
752    TSComputeIFunction - Evaluates the DAE residual written in implicit form F(t,U,Udot)=0
753 
754    Collective on TS
755 
756    Input Parameters:
757 +  ts - the TS context
758 .  t - current time
759 .  U - state vector
760 .  Udot - time derivative of state vector
761 -  imex - flag indicates if the method is IMEX so that the RHSFunction should be kept separate
762 
763    Output Parameter:
764 .  Y - right hand side
765 
766    Note:
767    Most users should not need to explicitly call this routine, as it
768    is used internally within the nonlinear solvers.
769 
770    If the user did did not write their equations in implicit form, this
771    function recasts them in implicit form.
772 
773    Level: developer
774 
775 .seealso: `TSSetIFunction()`, `TSComputeRHSFunction()`
776 @*/
777 PetscErrorCode TSComputeIFunction(TS ts,PetscReal t,Vec U,Vec Udot,Vec Y,PetscBool imex)
778 {
779   TSIFunction    ifunction;
780   TSRHSFunction  rhsfunction;
781   void           *ctx;
782   DM             dm;
783 
784   PetscFunctionBegin;
785   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
786   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
787   PetscValidHeaderSpecific(Udot,VEC_CLASSID,4);
788   PetscValidHeaderSpecific(Y,VEC_CLASSID,5);
789 
790   PetscCall(TSGetDM(ts,&dm));
791   PetscCall(DMTSGetIFunction(dm,&ifunction,&ctx));
792   PetscCall(DMTSGetRHSFunction(dm,&rhsfunction,NULL));
793 
794   PetscCheck(rhsfunction || ifunction,PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSFunction() and / or TSSetIFunction()");
795 
796   PetscCall(PetscLogEventBegin(TS_FunctionEval,ts,U,Udot,Y));
797   if (ifunction) {
798     PetscCallBack("TS callback implicit function",(*ifunction)(ts,t,U,Udot,Y,ctx));
799     ts->ifuncs++;
800   }
801   if (imex) {
802     if (!ifunction) PetscCall(VecCopy(Udot,Y));
803   } else if (rhsfunction) {
804     if (ifunction) {
805       Vec Frhs;
806       PetscCall(TSGetRHSVec_Private(ts,&Frhs));
807       PetscCall(TSComputeRHSFunction(ts,t,U,Frhs));
808       PetscCall(VecAXPY(Y,-1,Frhs));
809     } else {
810       PetscCall(TSComputeRHSFunction(ts,t,U,Y));
811       PetscCall(VecAYPX(Y,-1,Udot));
812     }
813   }
814   PetscCall(PetscLogEventEnd(TS_FunctionEval,ts,U,Udot,Y));
815   PetscFunctionReturn(0);
816 }
817 
818 /*
819    TSRecoverRHSJacobian - Recover the Jacobian matrix so that one can call TSComputeRHSJacobian() on it.
820 
821    Note:
822    This routine is needed when one switches from TSComputeIJacobian() to TSComputeRHSJacobian() because the Jacobian matrix may be shifted or scaled in TSComputeIJacobian().
823 
824 */
825 static PetscErrorCode TSRecoverRHSJacobian(TS ts,Mat A,Mat B)
826 {
827   PetscFunctionBegin;
828   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
829   PetscCheck(A == ts->Arhs,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Invalid Amat");
830   PetscCheck(B == ts->Brhs,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Invalid Bmat");
831 
832   if (ts->rhsjacobian.shift) PetscCall(MatShift(A,-ts->rhsjacobian.shift));
833   if (ts->rhsjacobian.scale == -1.) {
834     PetscCall(MatScale(A,-1));
835   }
836   if (B && B == ts->Brhs && A != B) {
837     if (ts->rhsjacobian.shift) PetscCall(MatShift(B,-ts->rhsjacobian.shift));
838     if (ts->rhsjacobian.scale == -1.) PetscCall(MatScale(B,-1));
839   }
840   ts->rhsjacobian.shift = 0;
841   ts->rhsjacobian.scale = 1.;
842   PetscFunctionReturn(0);
843 }
844 
845 /*@
846    TSComputeIJacobian - Evaluates the Jacobian of the DAE
847 
848    Collective on TS
849 
850    Input
851       Input Parameters:
852 +  ts - the TS context
853 .  t - current timestep
854 .  U - state vector
855 .  Udot - time derivative of state vector
856 .  shift - shift to apply, see note below
857 -  imex - flag indicates if the method is IMEX so that the RHSJacobian should be kept separate
858 
859    Output Parameters:
860 +  A - Jacobian matrix
861 -  B - matrix from which the preconditioner is constructed; often the same as A
862 
863    Notes:
864    If F(t,U,Udot)=0 is the DAE, the required Jacobian is
865 
866    dF/dU + shift*dF/dUdot
867 
868    Most users should not need to explicitly call this routine, as it
869    is used internally within the nonlinear solvers.
870 
871    Level: developer
872 
873 .seealso: `TSSetIJacobian()`
874 @*/
875 PetscErrorCode TSComputeIJacobian(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,PetscBool imex)
876 {
877   TSIJacobian    ijacobian;
878   TSRHSJacobian  rhsjacobian;
879   DM             dm;
880   void           *ctx;
881 
882   PetscFunctionBegin;
883   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
884   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
885   PetscValidHeaderSpecific(Udot,VEC_CLASSID,4);
886   PetscValidPointer(A,6);
887   PetscValidHeaderSpecific(A,MAT_CLASSID,6);
888   PetscValidPointer(B,7);
889   PetscValidHeaderSpecific(B,MAT_CLASSID,7);
890 
891   PetscCall(TSGetDM(ts,&dm));
892   PetscCall(DMTSGetIJacobian(dm,&ijacobian,&ctx));
893   PetscCall(DMTSGetRHSJacobian(dm,&rhsjacobian,NULL));
894 
895   PetscCheck(rhsjacobian || ijacobian,PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSJacobian() and / or TSSetIJacobian()");
896 
897   PetscCall(PetscLogEventBegin(TS_JacobianEval,ts,U,A,B));
898   if (ijacobian) {
899     PetscCallBack("TS callback implicit Jacobian",(*ijacobian)(ts,t,U,Udot,shift,A,B,ctx));
900     ts->ijacs++;
901   }
902   if (imex) {
903     if (!ijacobian) {  /* system was written as Udot = G(t,U) */
904       PetscBool assembled;
905       if (rhsjacobian) {
906         Mat Arhs = NULL;
907         PetscCall(TSGetRHSMats_Private(ts,&Arhs,NULL));
908         if (A == Arhs) {
909           PetscCheck(rhsjacobian != TSComputeRHSJacobianConstant,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Unsupported operation! cannot use TSComputeRHSJacobianConstant"); /* there is no way to reconstruct shift*M-J since J cannot be reevaluated */
910           ts->rhsjacobian.time = PETSC_MIN_REAL;
911         }
912       }
913       PetscCall(MatZeroEntries(A));
914       PetscCall(MatAssembled(A,&assembled));
915       if (!assembled) {
916         PetscCall(MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY));
917         PetscCall(MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY));
918       }
919       PetscCall(MatShift(A,shift));
920       if (A != B) {
921         PetscCall(MatZeroEntries(B));
922         PetscCall(MatAssembled(B,&assembled));
923         if (!assembled) {
924           PetscCall(MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY));
925           PetscCall(MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY));
926         }
927         PetscCall(MatShift(B,shift));
928       }
929     }
930   } else {
931     Mat Arhs = NULL,Brhs = NULL;
932 
933     /* RHSJacobian needs to be converted to part of IJacobian if exists */
934     if (rhsjacobian) PetscCall(TSGetRHSMats_Private(ts,&Arhs,&Brhs));
935     if (Arhs == A) { /* No IJacobian matrix, so we only have the RHS matrix */
936       PetscObjectState Ustate;
937       PetscObjectId    Uid;
938       TSRHSFunction    rhsfunction;
939 
940       PetscCall(DMTSGetRHSFunction(dm,&rhsfunction,NULL));
941       PetscCall(PetscObjectStateGet((PetscObject)U,&Ustate));
942       PetscCall(PetscObjectGetId((PetscObject)U,&Uid));
943       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 */
944         PetscCall(MatShift(A,shift-ts->rhsjacobian.shift)); /* revert the old shift and add the new shift with a single call to MatShift */
945         if (A != B) PetscCall(MatShift(B,shift-ts->rhsjacobian.shift));
946       } else {
947         PetscBool flg;
948 
949         if (ts->rhsjacobian.reuse) { /* Undo the damage */
950           /* MatScale has a short path for this case.
951              However, this code path is taken the first time TSComputeRHSJacobian is called
952              and the matrices have not been assembled yet */
953           PetscCall(TSRecoverRHSJacobian(ts,A,B));
954         }
955         PetscCall(TSComputeRHSJacobian(ts,t,U,A,B));
956         PetscCall(SNESGetUseMatrixFree(ts->snes,NULL,&flg));
957         /* since -snes_mf_operator uses the full SNES function it does not need to be shifted or scaled here */
958         if (!flg) {
959           PetscCall(MatScale(A,-1));
960           PetscCall(MatShift(A,shift));
961         }
962         if (A != B) {
963           PetscCall(MatScale(B,-1));
964           PetscCall(MatShift(B,shift));
965         }
966       }
967       ts->rhsjacobian.scale = -1;
968       ts->rhsjacobian.shift = shift;
969     } else if (Arhs) {          /* Both IJacobian and RHSJacobian */
970       if (!ijacobian) {         /* No IJacobian provided, but we have a separate RHS matrix */
971         PetscCall(MatZeroEntries(A));
972         PetscCall(MatShift(A,shift));
973         if (A != B) {
974           PetscCall(MatZeroEntries(B));
975           PetscCall(MatShift(B,shift));
976         }
977       }
978       PetscCall(TSComputeRHSJacobian(ts,t,U,Arhs,Brhs));
979       PetscCall(MatAXPY(A,-1,Arhs,ts->axpy_pattern));
980       if (A != B) PetscCall(MatAXPY(B,-1,Brhs,ts->axpy_pattern));
981     }
982   }
983   PetscCall(PetscLogEventEnd(TS_JacobianEval,ts,U,A,B));
984   PetscFunctionReturn(0);
985 }
986 
987 /*@C
988     TSSetRHSFunction - Sets the routine for evaluating the function,
989     where U_t = G(t,u).
990 
991     Logically Collective on TS
992 
993     Input Parameters:
994 +   ts - the TS context obtained from TSCreate()
995 .   r - vector to put the computed right hand side (or NULL to have it created)
996 .   f - routine for evaluating the right-hand-side function
997 -   ctx - [optional] user-defined context for private data for the
998           function evaluation routine (may be NULL)
999 
1000     Calling sequence of f:
1001 $     PetscErrorCode f(TS ts,PetscReal t,Vec u,Vec F,void *ctx);
1002 
1003 +   ts - timestep context
1004 .   t - current timestep
1005 .   u - input vector
1006 .   F - function vector
1007 -   ctx - [optional] user-defined function context
1008 
1009     Level: beginner
1010 
1011     Notes:
1012     You must call this function or TSSetIFunction() to define your ODE. You cannot use this function when solving a DAE.
1013 
1014 .seealso: `TSSetRHSJacobian()`, `TSSetIJacobian()`, `TSSetIFunction()`
1015 @*/
1016 PetscErrorCode  TSSetRHSFunction(TS ts,Vec r,PetscErrorCode (*f)(TS,PetscReal,Vec,Vec,void*),void *ctx)
1017 {
1018   SNES           snes;
1019   Vec            ralloc = NULL;
1020   DM             dm;
1021 
1022   PetscFunctionBegin;
1023   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1024   if (r) PetscValidHeaderSpecific(r,VEC_CLASSID,2);
1025 
1026   PetscCall(TSGetDM(ts,&dm));
1027   PetscCall(DMTSSetRHSFunction(dm,f,ctx));
1028   PetscCall(TSGetSNES(ts,&snes));
1029   if (!r && !ts->dm && ts->vec_sol) {
1030     PetscCall(VecDuplicate(ts->vec_sol,&ralloc));
1031     r = ralloc;
1032   }
1033   PetscCall(SNESSetFunction(snes,r,SNESTSFormFunction,ts));
1034   PetscCall(VecDestroy(&ralloc));
1035   PetscFunctionReturn(0);
1036 }
1037 
1038 /*@C
1039     TSSetSolutionFunction - Provide a function that computes the solution of the ODE or DAE
1040 
1041     Logically Collective on TS
1042 
1043     Input Parameters:
1044 +   ts - the TS context obtained from TSCreate()
1045 .   f - routine for evaluating the solution
1046 -   ctx - [optional] user-defined context for private data for the
1047           function evaluation routine (may be NULL)
1048 
1049     Calling sequence of f:
1050 $     PetscErrorCode f(TS ts,PetscReal t,Vec u,void *ctx);
1051 
1052 +   t - current timestep
1053 .   u - output vector
1054 -   ctx - [optional] user-defined function context
1055 
1056     Options Database:
1057 +  -ts_monitor_lg_error - create a graphical monitor of error history, requires user to have provided TSSetSolutionFunction()
1058 -  -ts_monitor_draw_error - Monitor error graphically, requires user to have provided TSSetSolutionFunction()
1059 
1060     Notes:
1061     This routine is used for testing accuracy of time integration schemes when you already know the solution.
1062     If analytic solutions are not known for your system, consider using the Method of Manufactured Solutions to
1063     create closed-form solutions with non-physical forcing terms.
1064 
1065     For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history.
1066 
1067     Level: beginner
1068 
1069 .seealso: `TSSetRHSJacobian()`, `TSSetIJacobian()`, `TSComputeSolutionFunction()`, `TSSetForcingFunction()`, `TSSetSolution()`, `TSGetSolution()`, `TSMonitorLGError()`, `TSMonitorDrawError()`
1070 @*/
1071 PetscErrorCode  TSSetSolutionFunction(TS ts,PetscErrorCode (*f)(TS,PetscReal,Vec,void*),void *ctx)
1072 {
1073   DM             dm;
1074 
1075   PetscFunctionBegin;
1076   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1077   PetscCall(TSGetDM(ts,&dm));
1078   PetscCall(DMTSSetSolutionFunction(dm,f,ctx));
1079   PetscFunctionReturn(0);
1080 }
1081 
1082 /*@C
1083     TSSetForcingFunction - Provide a function that computes a forcing term for a ODE or PDE
1084 
1085     Logically Collective on TS
1086 
1087     Input Parameters:
1088 +   ts - the TS context obtained from TSCreate()
1089 .   func - routine for evaluating the forcing function
1090 -   ctx - [optional] user-defined context for private data for the
1091           function evaluation routine (may be NULL)
1092 
1093     Calling sequence of func:
1094 $     PetscErrorCode func (TS ts,PetscReal t,Vec f,void *ctx);
1095 
1096 +   t - current timestep
1097 .   f - output vector
1098 -   ctx - [optional] user-defined function context
1099 
1100     Notes:
1101     This routine is useful for testing accuracy of time integration schemes when using the Method of Manufactured Solutions to
1102     create closed-form solutions with a non-physical forcing term. It allows you to use the Method of Manufactored Solution without directly editing the
1103     definition of the problem you are solving and hence possibly introducing bugs.
1104 
1105     This replaces the ODE F(u,u_t,t) = 0 the TS is solving with F(u,u_t,t) - func(t) = 0
1106 
1107     This forcing function does not depend on the solution to the equations, it can only depend on spatial location, time, and possibly parameters, the
1108     parameters can be passed in the ctx variable.
1109 
1110     For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history.
1111 
1112     Level: beginner
1113 
1114 .seealso: `TSSetRHSJacobian()`, `TSSetIJacobian()`, `TSComputeSolutionFunction()`, `TSSetSolutionFunction()`
1115 @*/
1116 PetscErrorCode  TSSetForcingFunction(TS ts,TSForcingFunction func,void *ctx)
1117 {
1118   DM             dm;
1119 
1120   PetscFunctionBegin;
1121   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1122   PetscCall(TSGetDM(ts,&dm));
1123   PetscCall(DMTSSetForcingFunction(dm,func,ctx));
1124   PetscFunctionReturn(0);
1125 }
1126 
1127 /*@C
1128    TSSetRHSJacobian - Sets the function to compute the Jacobian of G,
1129    where U_t = G(U,t), as well as the location to store the matrix.
1130 
1131    Logically Collective on TS
1132 
1133    Input Parameters:
1134 +  ts  - the TS context obtained from TSCreate()
1135 .  Amat - (approximate) Jacobian matrix
1136 .  Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat)
1137 .  f   - the Jacobian evaluation routine
1138 -  ctx - [optional] user-defined context for private data for the
1139          Jacobian evaluation routine (may be NULL)
1140 
1141    Calling sequence of f:
1142 $     PetscErrorCode f(TS ts,PetscReal t,Vec u,Mat A,Mat B,void *ctx);
1143 
1144 +  t - current timestep
1145 .  u - input vector
1146 .  Amat - (approximate) Jacobian matrix
1147 .  Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat)
1148 -  ctx - [optional] user-defined context for matrix evaluation routine
1149 
1150    Notes:
1151    You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1152 
1153    The TS solver may modify the nonzero structure and the entries of the matrices Amat and Pmat between the calls to f()
1154    You should not assume the values are the same in the next call to f() as you set them in the previous call.
1155 
1156    Level: beginner
1157 
1158 .seealso: `SNESComputeJacobianDefaultColor()`, `TSSetRHSFunction()`, `TSRHSJacobianSetReuse()`, `TSSetIJacobian()`
1159 
1160 @*/
1161 PetscErrorCode  TSSetRHSJacobian(TS ts,Mat Amat,Mat Pmat,TSRHSJacobian f,void *ctx)
1162 {
1163   SNES           snes;
1164   DM             dm;
1165   TSIJacobian    ijacobian;
1166 
1167   PetscFunctionBegin;
1168   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1169   if (Amat) PetscValidHeaderSpecific(Amat,MAT_CLASSID,2);
1170   if (Pmat) PetscValidHeaderSpecific(Pmat,MAT_CLASSID,3);
1171   if (Amat) PetscCheckSameComm(ts,1,Amat,2);
1172   if (Pmat) PetscCheckSameComm(ts,1,Pmat,3);
1173 
1174   PetscCall(TSGetDM(ts,&dm));
1175   PetscCall(DMTSSetRHSJacobian(dm,f,ctx));
1176   PetscCall(DMTSGetIJacobian(dm,&ijacobian,NULL));
1177   PetscCall(TSGetSNES(ts,&snes));
1178   if (!ijacobian) PetscCall(SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts));
1179   if (Amat) {
1180     PetscCall(PetscObjectReference((PetscObject)Amat));
1181     PetscCall(MatDestroy(&ts->Arhs));
1182     ts->Arhs = Amat;
1183   }
1184   if (Pmat) {
1185     PetscCall(PetscObjectReference((PetscObject)Pmat));
1186     PetscCall(MatDestroy(&ts->Brhs));
1187     ts->Brhs = Pmat;
1188   }
1189   PetscFunctionReturn(0);
1190 }
1191 
1192 /*@C
1193    TSSetIFunction - Set the function to compute F(t,U,U_t) where F() = 0 is the DAE to be solved.
1194 
1195    Logically Collective on TS
1196 
1197    Input Parameters:
1198 +  ts  - the TS context obtained from TSCreate()
1199 .  r   - vector to hold the residual (or NULL to have it created internally)
1200 .  f   - the function evaluation routine
1201 -  ctx - user-defined context for private data for the function evaluation routine (may be NULL)
1202 
1203    Calling sequence of f:
1204 $     PetscErrorCode f(TS ts,PetscReal t,Vec u,Vec u_t,Vec F,ctx);
1205 
1206 +  t   - time at step/stage being solved
1207 .  u   - state vector
1208 .  u_t - time derivative of state vector
1209 .  F   - function vector
1210 -  ctx - [optional] user-defined context for matrix evaluation routine
1211 
1212    Important:
1213    The user MUST call either this routine or TSSetRHSFunction() to define the ODE.  When solving DAEs you must use this function.
1214 
1215    Level: beginner
1216 
1217 .seealso: `TSSetRHSJacobian()`, `TSSetRHSFunction()`, `TSSetIJacobian()`
1218 @*/
1219 PetscErrorCode  TSSetIFunction(TS ts,Vec r,TSIFunction f,void *ctx)
1220 {
1221   SNES           snes;
1222   Vec            ralloc = NULL;
1223   DM             dm;
1224 
1225   PetscFunctionBegin;
1226   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1227   if (r) PetscValidHeaderSpecific(r,VEC_CLASSID,2);
1228 
1229   PetscCall(TSGetDM(ts,&dm));
1230   PetscCall(DMTSSetIFunction(dm,f,ctx));
1231 
1232   PetscCall(TSGetSNES(ts,&snes));
1233   if (!r && !ts->dm && ts->vec_sol) {
1234     PetscCall(VecDuplicate(ts->vec_sol,&ralloc));
1235     r  = ralloc;
1236   }
1237   PetscCall(SNESSetFunction(snes,r,SNESTSFormFunction,ts));
1238   PetscCall(VecDestroy(&ralloc));
1239   PetscFunctionReturn(0);
1240 }
1241 
1242 /*@C
1243    TSGetIFunction - Returns the vector where the implicit residual is stored and the function/context to compute it.
1244 
1245    Not Collective
1246 
1247    Input Parameter:
1248 .  ts - the TS context
1249 
1250    Output Parameters:
1251 +  r - vector to hold residual (or NULL)
1252 .  func - the function to compute residual (or NULL)
1253 -  ctx - the function context (or NULL)
1254 
1255    Level: advanced
1256 
1257 .seealso: `TSSetIFunction()`, `SNESGetFunction()`
1258 @*/
1259 PetscErrorCode TSGetIFunction(TS ts,Vec *r,TSIFunction *func,void **ctx)
1260 {
1261   SNES           snes;
1262   DM             dm;
1263 
1264   PetscFunctionBegin;
1265   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1266   PetscCall(TSGetSNES(ts,&snes));
1267   PetscCall(SNESGetFunction(snes,r,NULL,NULL));
1268   PetscCall(TSGetDM(ts,&dm));
1269   PetscCall(DMTSGetIFunction(dm,func,ctx));
1270   PetscFunctionReturn(0);
1271 }
1272 
1273 /*@C
1274    TSGetRHSFunction - Returns the vector where the right hand side is stored and the function/context to compute it.
1275 
1276    Not Collective
1277 
1278    Input Parameter:
1279 .  ts - the TS context
1280 
1281    Output Parameters:
1282 +  r - vector to hold computed right hand side (or NULL)
1283 .  func - the function to compute right hand side (or NULL)
1284 -  ctx - the function context (or NULL)
1285 
1286    Level: advanced
1287 
1288 .seealso: `TSSetRHSFunction()`, `SNESGetFunction()`
1289 @*/
1290 PetscErrorCode TSGetRHSFunction(TS ts,Vec *r,TSRHSFunction *func,void **ctx)
1291 {
1292   SNES           snes;
1293   DM             dm;
1294 
1295   PetscFunctionBegin;
1296   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1297   PetscCall(TSGetSNES(ts,&snes));
1298   PetscCall(SNESGetFunction(snes,r,NULL,NULL));
1299   PetscCall(TSGetDM(ts,&dm));
1300   PetscCall(DMTSGetRHSFunction(dm,func,ctx));
1301   PetscFunctionReturn(0);
1302 }
1303 
1304 /*@C
1305    TSSetIJacobian - Set the function to compute the matrix dF/dU + a*dF/dU_t where F(t,U,U_t) is the function
1306         provided with TSSetIFunction().
1307 
1308    Logically Collective on TS
1309 
1310    Input Parameters:
1311 +  ts  - the TS context obtained from TSCreate()
1312 .  Amat - (approximate) Jacobian matrix
1313 .  Pmat - matrix used to compute preconditioner (usually the same as Amat)
1314 .  f   - the Jacobian evaluation routine
1315 -  ctx - user-defined context for private data for the Jacobian evaluation routine (may be NULL)
1316 
1317    Calling sequence of f:
1318 $    PetscErrorCode f(TS ts,PetscReal t,Vec U,Vec U_t,PetscReal a,Mat Amat,Mat Pmat,void *ctx);
1319 
1320 +  t    - time at step/stage being solved
1321 .  U    - state vector
1322 .  U_t  - time derivative of state vector
1323 .  a    - shift
1324 .  Amat - (approximate) Jacobian of F(t,U,W+a*U), equivalent to dF/dU + a*dF/dU_t
1325 .  Pmat - matrix used for constructing preconditioner, usually the same as Amat
1326 -  ctx  - [optional] user-defined context for matrix evaluation routine
1327 
1328    Notes:
1329    The matrices Amat and Pmat are exactly the matrices that are used by SNES for the nonlinear solve.
1330 
1331    If you know the operator Amat has a null space you can use MatSetNullSpace() and MatSetTransposeNullSpace() to supply the null
1332    space to Amat and the KSP solvers will automatically use that null space as needed during the solution process.
1333 
1334    The matrix dF/dU + a*dF/dU_t you provide turns out to be
1335    the Jacobian of F(t,U,W+a*U) where F(t,U,U_t) = 0 is the DAE to be solved.
1336    The time integrator internally approximates U_t by W+a*U where the positive "shift"
1337    a and vector W depend on the integration method, step size, and past states. For example with
1338    the backward Euler method a = 1/dt and W = -a*U(previous timestep) so
1339    W + a*U = a*(U - U(previous timestep)) = (U - U(previous timestep))/dt
1340 
1341    You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1342 
1343    The TS solver may modify the nonzero structure and the entries of the matrices Amat and Pmat between the calls to f()
1344    You should not assume the values are the same in the next call to f() as you set them in the previous call.
1345 
1346    Level: beginner
1347 
1348 .seealso: `TSSetIFunction()`, `TSSetRHSJacobian()`, `SNESComputeJacobianDefaultColor()`, `SNESComputeJacobianDefault()`, `TSSetRHSFunction()`
1349 
1350 @*/
1351 PetscErrorCode  TSSetIJacobian(TS ts,Mat Amat,Mat Pmat,TSIJacobian f,void *ctx)
1352 {
1353   SNES           snes;
1354   DM             dm;
1355 
1356   PetscFunctionBegin;
1357   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1358   if (Amat) PetscValidHeaderSpecific(Amat,MAT_CLASSID,2);
1359   if (Pmat) PetscValidHeaderSpecific(Pmat,MAT_CLASSID,3);
1360   if (Amat) PetscCheckSameComm(ts,1,Amat,2);
1361   if (Pmat) PetscCheckSameComm(ts,1,Pmat,3);
1362 
1363   PetscCall(TSGetDM(ts,&dm));
1364   PetscCall(DMTSSetIJacobian(dm,f,ctx));
1365 
1366   PetscCall(TSGetSNES(ts,&snes));
1367   PetscCall(SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts));
1368   PetscFunctionReturn(0);
1369 }
1370 
1371 /*@
1372    TSRHSJacobianSetReuse - restore RHS Jacobian before re-evaluating.  Without this flag, TS will change the sign and
1373    shift the RHS Jacobian for a finite-time-step implicit solve, in which case the user function will need to recompute
1374    the entire Jacobian.  The reuse flag must be set if the evaluation function will assume that the matrix entries have
1375    not been changed by the TS.
1376 
1377    Logically Collective
1378 
1379    Input Parameters:
1380 +  ts - TS context obtained from TSCreate()
1381 -  reuse - PETSC_TRUE if the RHS Jacobian
1382 
1383    Level: intermediate
1384 
1385 .seealso: `TSSetRHSJacobian()`, `TSComputeRHSJacobianConstant()`
1386 @*/
1387 PetscErrorCode TSRHSJacobianSetReuse(TS ts,PetscBool reuse)
1388 {
1389   PetscFunctionBegin;
1390   ts->rhsjacobian.reuse = reuse;
1391   PetscFunctionReturn(0);
1392 }
1393 
1394 /*@C
1395    TSSetI2Function - Set the function to compute F(t,U,U_t,U_tt) where F = 0 is the DAE to be solved.
1396 
1397    Logically Collective on TS
1398 
1399    Input Parameters:
1400 +  ts  - the TS context obtained from TSCreate()
1401 .  F   - vector to hold the residual (or NULL to have it created internally)
1402 .  fun - the function evaluation routine
1403 -  ctx - user-defined context for private data for the function evaluation routine (may be NULL)
1404 
1405    Calling sequence of fun:
1406 $     PetscErrorCode fun(TS ts,PetscReal t,Vec U,Vec U_t,Vec U_tt,Vec F,ctx);
1407 
1408 +  t    - time at step/stage being solved
1409 .  U    - state vector
1410 .  U_t  - time derivative of state vector
1411 .  U_tt - second time derivative of state vector
1412 .  F    - function vector
1413 -  ctx  - [optional] user-defined context for matrix evaluation routine (may be NULL)
1414 
1415    Level: beginner
1416 
1417 .seealso: `TSSetI2Jacobian()`, `TSSetIFunction()`, `TSCreate()`, `TSSetRHSFunction()`
1418 @*/
1419 PetscErrorCode TSSetI2Function(TS ts,Vec F,TSI2Function fun,void *ctx)
1420 {
1421   DM             dm;
1422 
1423   PetscFunctionBegin;
1424   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1425   if (F) PetscValidHeaderSpecific(F,VEC_CLASSID,2);
1426   PetscCall(TSSetIFunction(ts,F,NULL,NULL));
1427   PetscCall(TSGetDM(ts,&dm));
1428   PetscCall(DMTSSetI2Function(dm,fun,ctx));
1429   PetscFunctionReturn(0);
1430 }
1431 
1432 /*@C
1433   TSGetI2Function - Returns the vector where the implicit residual is stored and the function/context to compute it.
1434 
1435   Not Collective
1436 
1437   Input Parameter:
1438 . ts - the TS context
1439 
1440   Output Parameters:
1441 + r - vector to hold residual (or NULL)
1442 . fun - the function to compute residual (or NULL)
1443 - ctx - the function context (or NULL)
1444 
1445   Level: advanced
1446 
1447 .seealso: `TSSetIFunction()`, `SNESGetFunction()`, `TSCreate()`
1448 @*/
1449 PetscErrorCode TSGetI2Function(TS ts,Vec *r,TSI2Function *fun,void **ctx)
1450 {
1451   SNES           snes;
1452   DM             dm;
1453 
1454   PetscFunctionBegin;
1455   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1456   PetscCall(TSGetSNES(ts,&snes));
1457   PetscCall(SNESGetFunction(snes,r,NULL,NULL));
1458   PetscCall(TSGetDM(ts,&dm));
1459   PetscCall(DMTSGetI2Function(dm,fun,ctx));
1460   PetscFunctionReturn(0);
1461 }
1462 
1463 /*@C
1464    TSSetI2Jacobian - Set the function to compute the matrix dF/dU + v*dF/dU_t  + a*dF/dU_tt
1465         where F(t,U,U_t,U_tt) is the function you provided with TSSetI2Function().
1466 
1467    Logically Collective on TS
1468 
1469    Input Parameters:
1470 +  ts  - the TS context obtained from TSCreate()
1471 .  J   - Jacobian matrix
1472 .  P   - preconditioning matrix for J (may be same as J)
1473 .  jac - the Jacobian evaluation routine
1474 -  ctx - user-defined context for private data for the Jacobian evaluation routine (may be NULL)
1475 
1476    Calling sequence of jac:
1477 $    PetscErrorCode jac(TS ts,PetscReal t,Vec U,Vec U_t,Vec U_tt,PetscReal v,PetscReal a,Mat J,Mat P,void *ctx);
1478 
1479 +  t    - time at step/stage being solved
1480 .  U    - state vector
1481 .  U_t  - time derivative of state vector
1482 .  U_tt - second time derivative of state vector
1483 .  v    - shift for U_t
1484 .  a    - shift for U_tt
1485 .  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
1486 .  P    - preconditioning matrix for J, may be same as J
1487 -  ctx  - [optional] user-defined context for matrix evaluation routine
1488 
1489    Notes:
1490    The matrices J and P are exactly the matrices that are used by SNES for the nonlinear solve.
1491 
1492    The matrix dF/dU + v*dF/dU_t + a*dF/dU_tt you provide turns out to be
1493    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.
1494    The time integrator internally approximates U_t by W+v*U and U_tt by W'+a*U  where the positive "shift"
1495    parameters 'v' and 'a' and vectors W, W' depend on the integration method, step size, and past states.
1496 
1497    Level: beginner
1498 
1499 .seealso: `TSSetI2Function()`, `TSGetI2Jacobian()`
1500 @*/
1501 PetscErrorCode TSSetI2Jacobian(TS ts,Mat J,Mat P,TSI2Jacobian jac,void *ctx)
1502 {
1503   DM             dm;
1504 
1505   PetscFunctionBegin;
1506   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1507   if (J) PetscValidHeaderSpecific(J,MAT_CLASSID,2);
1508   if (P) PetscValidHeaderSpecific(P,MAT_CLASSID,3);
1509   PetscCall(TSSetIJacobian(ts,J,P,NULL,NULL));
1510   PetscCall(TSGetDM(ts,&dm));
1511   PetscCall(DMTSSetI2Jacobian(dm,jac,ctx));
1512   PetscFunctionReturn(0);
1513 }
1514 
1515 /*@C
1516   TSGetI2Jacobian - Returns the implicit Jacobian at the present timestep.
1517 
1518   Not Collective, but parallel objects are returned if TS is parallel
1519 
1520   Input Parameter:
1521 . ts  - The TS context obtained from TSCreate()
1522 
1523   Output Parameters:
1524 + J  - The (approximate) Jacobian of F(t,U,U_t,U_tt)
1525 . P - The matrix from which the preconditioner is constructed, often the same as J
1526 . jac - The function to compute the Jacobian matrices
1527 - ctx - User-defined context for Jacobian evaluation routine
1528 
1529   Notes:
1530     You can pass in NULL for any return argument you do not need.
1531 
1532   Level: advanced
1533 
1534 .seealso: `TSGetTimeStep()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`, `TSSetI2Jacobian()`, `TSGetI2Function()`, `TSCreate()`
1535 
1536 @*/
1537 PetscErrorCode  TSGetI2Jacobian(TS ts,Mat *J,Mat *P,TSI2Jacobian *jac,void **ctx)
1538 {
1539   SNES           snes;
1540   DM             dm;
1541 
1542   PetscFunctionBegin;
1543   PetscCall(TSGetSNES(ts,&snes));
1544   PetscCall(SNESSetUpMatrices(snes));
1545   PetscCall(SNESGetJacobian(snes,J,P,NULL,NULL));
1546   PetscCall(TSGetDM(ts,&dm));
1547   PetscCall(DMTSGetI2Jacobian(dm,jac,ctx));
1548   PetscFunctionReturn(0);
1549 }
1550 
1551 /*@
1552   TSComputeI2Function - Evaluates the DAE residual written in implicit form F(t,U,U_t,U_tt) = 0
1553 
1554   Collective on TS
1555 
1556   Input Parameters:
1557 + ts - the TS context
1558 . t - current time
1559 . U - state vector
1560 . V - time derivative of state vector (U_t)
1561 - A - second time derivative of state vector (U_tt)
1562 
1563   Output Parameter:
1564 . F - the residual vector
1565 
1566   Note:
1567   Most users should not need to explicitly call this routine, as it
1568   is used internally within the nonlinear solvers.
1569 
1570   Level: developer
1571 
1572 .seealso: `TSSetI2Function()`, `TSGetI2Function()`
1573 @*/
1574 PetscErrorCode TSComputeI2Function(TS ts,PetscReal t,Vec U,Vec V,Vec A,Vec F)
1575 {
1576   DM             dm;
1577   TSI2Function   I2Function;
1578   void           *ctx;
1579   TSRHSFunction  rhsfunction;
1580 
1581   PetscFunctionBegin;
1582   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1583   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
1584   PetscValidHeaderSpecific(V,VEC_CLASSID,4);
1585   PetscValidHeaderSpecific(A,VEC_CLASSID,5);
1586   PetscValidHeaderSpecific(F,VEC_CLASSID,6);
1587 
1588   PetscCall(TSGetDM(ts,&dm));
1589   PetscCall(DMTSGetI2Function(dm,&I2Function,&ctx));
1590   PetscCall(DMTSGetRHSFunction(dm,&rhsfunction,NULL));
1591 
1592   if (!I2Function) {
1593     PetscCall(TSComputeIFunction(ts,t,U,A,F,PETSC_FALSE));
1594     PetscFunctionReturn(0);
1595   }
1596 
1597   PetscCall(PetscLogEventBegin(TS_FunctionEval,ts,U,V,F));
1598 
1599   PetscCallBack("TS callback implicit function",I2Function(ts,t,U,V,A,F,ctx));
1600 
1601   if (rhsfunction) {
1602     Vec Frhs;
1603     PetscCall(TSGetRHSVec_Private(ts,&Frhs));
1604     PetscCall(TSComputeRHSFunction(ts,t,U,Frhs));
1605     PetscCall(VecAXPY(F,-1,Frhs));
1606   }
1607 
1608   PetscCall(PetscLogEventEnd(TS_FunctionEval,ts,U,V,F));
1609   PetscFunctionReturn(0);
1610 }
1611 
1612 /*@
1613   TSComputeI2Jacobian - Evaluates the Jacobian of the DAE
1614 
1615   Collective on TS
1616 
1617   Input Parameters:
1618 + ts - the TS context
1619 . t - current timestep
1620 . U - state vector
1621 . V - time derivative of state vector
1622 . A - second time derivative of state vector
1623 . shiftV - shift to apply, see note below
1624 - shiftA - shift to apply, see note below
1625 
1626   Output Parameters:
1627 + J - Jacobian matrix
1628 - P - optional preconditioning matrix
1629 
1630   Notes:
1631   If F(t,U,V,A)=0 is the DAE, the required Jacobian is
1632 
1633   dF/dU + shiftV*dF/dV + shiftA*dF/dA
1634 
1635   Most users should not need to explicitly call this routine, as it
1636   is used internally within the nonlinear solvers.
1637 
1638   Level: developer
1639 
1640 .seealso: `TSSetI2Jacobian()`
1641 @*/
1642 PetscErrorCode TSComputeI2Jacobian(TS ts,PetscReal t,Vec U,Vec V,Vec A,PetscReal shiftV,PetscReal shiftA,Mat J,Mat P)
1643 {
1644   DM             dm;
1645   TSI2Jacobian   I2Jacobian;
1646   void           *ctx;
1647   TSRHSJacobian  rhsjacobian;
1648 
1649   PetscFunctionBegin;
1650   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1651   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
1652   PetscValidHeaderSpecific(V,VEC_CLASSID,4);
1653   PetscValidHeaderSpecific(A,VEC_CLASSID,5);
1654   PetscValidHeaderSpecific(J,MAT_CLASSID,8);
1655   PetscValidHeaderSpecific(P,MAT_CLASSID,9);
1656 
1657   PetscCall(TSGetDM(ts,&dm));
1658   PetscCall(DMTSGetI2Jacobian(dm,&I2Jacobian,&ctx));
1659   PetscCall(DMTSGetRHSJacobian(dm,&rhsjacobian,NULL));
1660 
1661   if (!I2Jacobian) {
1662     PetscCall(TSComputeIJacobian(ts,t,U,A,shiftA,J,P,PETSC_FALSE));
1663     PetscFunctionReturn(0);
1664   }
1665 
1666   PetscCall(PetscLogEventBegin(TS_JacobianEval,ts,U,J,P));
1667   PetscCallBack("TS callback implicit Jacobian",I2Jacobian(ts,t,U,V,A,shiftV,shiftA,J,P,ctx));
1668   if (rhsjacobian) {
1669     Mat Jrhs,Prhs;
1670     PetscCall(TSGetRHSMats_Private(ts,&Jrhs,&Prhs));
1671     PetscCall(TSComputeRHSJacobian(ts,t,U,Jrhs,Prhs));
1672     PetscCall(MatAXPY(J,-1,Jrhs,ts->axpy_pattern));
1673     if (P != J) PetscCall(MatAXPY(P,-1,Prhs,ts->axpy_pattern));
1674   }
1675 
1676   PetscCall(PetscLogEventEnd(TS_JacobianEval,ts,U,J,P));
1677   PetscFunctionReturn(0);
1678 }
1679 
1680 /*@C
1681    TSSetTransientVariable - sets function to transform from state to transient variables
1682 
1683    Logically Collective
1684 
1685    Input Parameters:
1686 +  ts - time stepping context on which to change the transient variable
1687 .  tvar - a function that transforms to transient variables
1688 -  ctx - a context for tvar
1689 
1690     Calling sequence of tvar:
1691 $     PetscErrorCode tvar(TS ts,Vec p,Vec c,void *ctx);
1692 
1693 +   ts - timestep context
1694 .   p - input vector (primitive form)
1695 .   c - output vector, transient variables (conservative form)
1696 -   ctx - [optional] user-defined function context
1697 
1698    Level: advanced
1699 
1700    Notes:
1701    This is typically used to transform from primitive to conservative variables so that a time integrator (e.g., TSBDF)
1702    can be conservative.  In this context, primitive variables P are used to model the state (e.g., because they lead to
1703    well-conditioned formulations even in limiting cases such as low-Mach or zero porosity).  The transient variable is
1704    C(P), specified by calling this function.  An IFunction thus receives arguments (P, Cdot) and the IJacobian must be
1705    evaluated via the chain rule, as in
1706 
1707      dF/dP + shift * dF/dCdot dC/dP.
1708 
1709 .seealso: `DMTSSetTransientVariable()`, `DMTSGetTransientVariable()`, `TSSetIFunction()`, `TSSetIJacobian()`
1710 @*/
1711 PetscErrorCode TSSetTransientVariable(TS ts,TSTransientVariable tvar,void *ctx)
1712 {
1713   DM             dm;
1714 
1715   PetscFunctionBegin;
1716   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1717   PetscCall(TSGetDM(ts,&dm));
1718   PetscCall(DMTSSetTransientVariable(dm,tvar,ctx));
1719   PetscFunctionReturn(0);
1720 }
1721 
1722 /*@
1723    TSComputeTransientVariable - transforms state (primitive) variables to transient (conservative) variables
1724 
1725    Logically Collective
1726 
1727    Input Parameters:
1728 +  ts - TS on which to compute
1729 -  U - state vector to be transformed to transient variables
1730 
1731    Output Parameters:
1732 .  C - transient (conservative) variable
1733 
1734    Developer Notes:
1735    If DMTSSetTransientVariable() has not been called, then C is not modified in this routine and C=NULL is allowed.
1736    This makes it safe to call without a guard.  One can use TSHasTransientVariable() to check if transient variables are
1737    being used.
1738 
1739    Level: developer
1740 
1741 .seealso: `DMTSSetTransientVariable()`, `TSComputeIFunction()`, `TSComputeIJacobian()`
1742 @*/
1743 PetscErrorCode TSComputeTransientVariable(TS ts,Vec U,Vec C)
1744 {
1745   DM             dm;
1746   DMTS           dmts;
1747 
1748   PetscFunctionBegin;
1749   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1750   PetscValidHeaderSpecific(U,VEC_CLASSID,2);
1751   PetscCall(TSGetDM(ts,&dm));
1752   PetscCall(DMGetDMTS(dm,&dmts));
1753   if (dmts->ops->transientvar) {
1754     PetscValidHeaderSpecific(C,VEC_CLASSID,3);
1755     PetscCall((*dmts->ops->transientvar)(ts,U,C,dmts->transientvarctx));
1756   }
1757   PetscFunctionReturn(0);
1758 }
1759 
1760 /*@
1761    TSHasTransientVariable - determine whether transient variables have been set
1762 
1763    Logically Collective
1764 
1765    Input Parameters:
1766 .  ts - TS on which to compute
1767 
1768    Output Parameters:
1769 .  has - PETSC_TRUE if transient variables have been set
1770 
1771    Level: developer
1772 
1773 .seealso: `DMTSSetTransientVariable()`, `TSComputeTransientVariable()`
1774 @*/
1775 PetscErrorCode TSHasTransientVariable(TS ts,PetscBool *has)
1776 {
1777   DM             dm;
1778   DMTS           dmts;
1779 
1780   PetscFunctionBegin;
1781   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1782   PetscCall(TSGetDM(ts,&dm));
1783   PetscCall(DMGetDMTS(dm,&dmts));
1784   *has = dmts->ops->transientvar ? PETSC_TRUE : PETSC_FALSE;
1785   PetscFunctionReturn(0);
1786 }
1787 
1788 /*@
1789    TS2SetSolution - Sets the initial solution and time derivative vectors
1790    for use by the TS routines handling second order equations.
1791 
1792    Logically Collective on TS
1793 
1794    Input Parameters:
1795 +  ts - the TS context obtained from TSCreate()
1796 .  u - the solution vector
1797 -  v - the time derivative vector
1798 
1799    Level: beginner
1800 
1801 @*/
1802 PetscErrorCode  TS2SetSolution(TS ts,Vec u,Vec v)
1803 {
1804   PetscFunctionBegin;
1805   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1806   PetscValidHeaderSpecific(u,VEC_CLASSID,2);
1807   PetscValidHeaderSpecific(v,VEC_CLASSID,3);
1808   PetscCall(TSSetSolution(ts,u));
1809   PetscCall(PetscObjectReference((PetscObject)v));
1810   PetscCall(VecDestroy(&ts->vec_dot));
1811   ts->vec_dot = v;
1812   PetscFunctionReturn(0);
1813 }
1814 
1815 /*@
1816    TS2GetSolution - Returns the solution and time derivative at the present timestep
1817    for second order equations. It is valid to call this routine inside the function
1818    that you are evaluating in order to move to the new timestep. This vector not
1819    changed until the solution at the next timestep has been calculated.
1820 
1821    Not Collective, but Vec returned is parallel if TS is parallel
1822 
1823    Input Parameter:
1824 .  ts - the TS context obtained from TSCreate()
1825 
1826    Output Parameters:
1827 +  u - the vector containing the solution
1828 -  v - the vector containing the time derivative
1829 
1830    Level: intermediate
1831 
1832 .seealso: `TS2SetSolution()`, `TSGetTimeStep()`, `TSGetTime()`
1833 
1834 @*/
1835 PetscErrorCode  TS2GetSolution(TS ts,Vec *u,Vec *v)
1836 {
1837   PetscFunctionBegin;
1838   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1839   if (u) PetscValidPointer(u,2);
1840   if (v) PetscValidPointer(v,3);
1841   if (u) *u = ts->vec_sol;
1842   if (v) *v = ts->vec_dot;
1843   PetscFunctionReturn(0);
1844 }
1845 
1846 /*@C
1847   TSLoad - Loads a KSP that has been stored in binary  with KSPView().
1848 
1849   Collective on PetscViewer
1850 
1851   Input Parameters:
1852 + newdm - the newly loaded TS, this needs to have been created with TSCreate() or
1853            some related function before a call to TSLoad().
1854 - viewer - binary file viewer, obtained from PetscViewerBinaryOpen()
1855 
1856    Level: intermediate
1857 
1858   Notes:
1859    The type is determined by the data in the file, any type set into the TS before this call is ignored.
1860 
1861   Notes for advanced users:
1862   Most users should not need to know the details of the binary storage
1863   format, since TSLoad() and TSView() completely hide these details.
1864   But for anyone who's interested, the standard binary matrix storage
1865   format is
1866 .vb
1867      has not yet been determined
1868 .ve
1869 
1870 .seealso: `PetscViewerBinaryOpen()`, `TSView()`, `MatLoad()`, `VecLoad()`
1871 @*/
1872 PetscErrorCode  TSLoad(TS ts, PetscViewer viewer)
1873 {
1874   PetscBool      isbinary;
1875   PetscInt       classid;
1876   char           type[256];
1877   DMTS           sdm;
1878   DM             dm;
1879 
1880   PetscFunctionBegin;
1881   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1882   PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2);
1883   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary));
1884   PetscCheck(isbinary,PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Invalid viewer; open viewer with PetscViewerBinaryOpen()");
1885 
1886   PetscCall(PetscViewerBinaryRead(viewer,&classid,1,NULL,PETSC_INT));
1887   PetscCheck(classid == TS_FILE_CLASSID,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Not TS next in file");
1888   PetscCall(PetscViewerBinaryRead(viewer,type,256,NULL,PETSC_CHAR));
1889   PetscCall(TSSetType(ts, type));
1890   PetscTryTypeMethod(ts,load,viewer);
1891   PetscCall(DMCreate(PetscObjectComm((PetscObject)ts),&dm));
1892   PetscCall(DMLoad(dm,viewer));
1893   PetscCall(TSSetDM(ts,dm));
1894   PetscCall(DMCreateGlobalVector(ts->dm,&ts->vec_sol));
1895   PetscCall(VecLoad(ts->vec_sol,viewer));
1896   PetscCall(DMGetDMTS(ts->dm,&sdm));
1897   PetscCall(DMTSLoad(sdm,viewer));
1898   PetscFunctionReturn(0);
1899 }
1900 
1901 #include <petscdraw.h>
1902 #if defined(PETSC_HAVE_SAWS)
1903 #include <petscviewersaws.h>
1904 #endif
1905 
1906 /*@C
1907    TSViewFromOptions - View from Options
1908 
1909    Collective on TS
1910 
1911    Input Parameters:
1912 +  A - the application ordering context
1913 .  obj - Optional object
1914 -  name - command line option
1915 
1916    Level: intermediate
1917 .seealso: `TS`, `TSView`, `PetscObjectViewFromOptions()`, `TSCreate()`
1918 @*/
1919 PetscErrorCode  TSViewFromOptions(TS A,PetscObject obj,const char name[])
1920 {
1921   PetscFunctionBegin;
1922   PetscValidHeaderSpecific(A,TS_CLASSID,1);
1923   PetscCall(PetscObjectViewFromOptions((PetscObject)A,obj,name));
1924   PetscFunctionReturn(0);
1925 }
1926 
1927 /*@C
1928     TSView - Prints the TS data structure.
1929 
1930     Collective on TS
1931 
1932     Input Parameters:
1933 +   ts - the TS context obtained from TSCreate()
1934 -   viewer - visualization context
1935 
1936     Options Database Key:
1937 .   -ts_view - calls TSView() at end of TSStep()
1938 
1939     Notes:
1940     The available visualization contexts include
1941 +     PETSC_VIEWER_STDOUT_SELF - standard output (default)
1942 -     PETSC_VIEWER_STDOUT_WORLD - synchronized standard
1943          output where only the first processor opens
1944          the file.  All other processors send their
1945          data to the first processor to print.
1946 
1947     The user can open an alternative visualization context with
1948     PetscViewerASCIIOpen() - output to a specified file.
1949 
1950     In the debugger you can do "call TSView(ts,0)" to display the TS solver. (The same holds for any PETSc object viewer).
1951 
1952     Level: beginner
1953 
1954 .seealso: `PetscViewerASCIIOpen()`
1955 @*/
1956 PetscErrorCode  TSView(TS ts,PetscViewer viewer)
1957 {
1958   TSType         type;
1959   PetscBool      iascii,isstring,isundials,isbinary,isdraw;
1960   DMTS           sdm;
1961 #if defined(PETSC_HAVE_SAWS)
1962   PetscBool      issaws;
1963 #endif
1964 
1965   PetscFunctionBegin;
1966   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
1967   if (!viewer) PetscCall(PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)ts),&viewer));
1968   PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2);
1969   PetscCheckSameComm(ts,1,viewer,2);
1970 
1971   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii));
1972   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSTRING,&isstring));
1973   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary));
1974   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw));
1975 #if defined(PETSC_HAVE_SAWS)
1976   PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSAWS,&issaws));
1977 #endif
1978   if (iascii) {
1979     PetscCall(PetscObjectPrintClassNamePrefixType((PetscObject)ts,viewer));
1980     if (ts->ops->view) {
1981       PetscCall(PetscViewerASCIIPushTab(viewer));
1982       PetscUseTypeMethod(ts,view ,viewer);
1983       PetscCall(PetscViewerASCIIPopTab(viewer));
1984     }
1985     if (ts->max_steps < PETSC_MAX_INT) PetscCall(PetscViewerASCIIPrintf(viewer,"  maximum steps=%" PetscInt_FMT "\n",ts->max_steps));
1986     if (ts->max_time < PETSC_MAX_REAL) PetscCall(PetscViewerASCIIPrintf(viewer,"  maximum time=%g\n",(double)ts->max_time));
1987     if (ts->ifuncs) PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of I function evaluations=%" PetscInt_FMT "\n",ts->ifuncs));
1988     if (ts->ijacs) PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of I Jacobian evaluations=%" PetscInt_FMT "\n",ts->ijacs));
1989     if (ts->rhsfuncs) PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of RHS function evaluations=%" PetscInt_FMT "\n",ts->rhsfuncs));
1990     if (ts->rhsjacs) PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of RHS Jacobian evaluations=%" PetscInt_FMT "\n",ts->rhsjacs));
1991     if (ts->usessnes) {
1992       PetscBool lin;
1993       if (ts->problem_type == TS_NONLINEAR) PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of nonlinear solver iterations=%" PetscInt_FMT "\n",ts->snes_its));
1994       PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of linear solver iterations=%" PetscInt_FMT "\n",ts->ksp_its));
1995       PetscCall(PetscObjectTypeCompareAny((PetscObject)ts->snes,&lin,SNESKSPONLY,SNESKSPTRANSPOSEONLY,""));
1996       PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of %slinear solve failures=%" PetscInt_FMT "\n",lin ? "" : "non",ts->num_snes_failures));
1997     }
1998     PetscCall(PetscViewerASCIIPrintf(viewer,"  total number of rejected steps=%" PetscInt_FMT "\n",ts->reject));
1999     if (ts->vrtol) PetscCall(PetscViewerASCIIPrintf(viewer,"  using vector of relative error tolerances, "));
2000     else PetscCall(PetscViewerASCIIPrintf(viewer,"  using relative error tolerance of %g, ",(double)ts->rtol));
2001     if (ts->vatol) PetscCall(PetscViewerASCIIPrintf(viewer,"  using vector of absolute error tolerances\n"));
2002     else PetscCall(PetscViewerASCIIPrintf(viewer,"  using absolute error tolerance of %g\n",(double)ts->atol));
2003     PetscCall(PetscViewerASCIIPushTab(viewer));
2004     PetscCall(TSAdaptView(ts->adapt,viewer));
2005     PetscCall(PetscViewerASCIIPopTab(viewer));
2006   } else if (isstring) {
2007     PetscCall(TSGetType(ts,&type));
2008     PetscCall(PetscViewerStringSPrintf(viewer," TSType: %-7.7s",type));
2009     PetscTryTypeMethod(ts,view,viewer);
2010   } else if (isbinary) {
2011     PetscInt    classid = TS_FILE_CLASSID;
2012     MPI_Comm    comm;
2013     PetscMPIInt rank;
2014     char        type[256];
2015 
2016     PetscCall(PetscObjectGetComm((PetscObject)ts,&comm));
2017     PetscCallMPI(MPI_Comm_rank(comm,&rank));
2018     if (rank == 0) {
2019       PetscCall(PetscViewerBinaryWrite(viewer,&classid,1,PETSC_INT));
2020       PetscCall(PetscStrncpy(type,((PetscObject)ts)->type_name,256));
2021       PetscCall(PetscViewerBinaryWrite(viewer,type,256,PETSC_CHAR));
2022     }
2023     PetscTryTypeMethod(ts,view,viewer);
2024     if (ts->adapt) PetscCall(TSAdaptView(ts->adapt,viewer));
2025     PetscCall(DMView(ts->dm,viewer));
2026     PetscCall(VecView(ts->vec_sol,viewer));
2027     PetscCall(DMGetDMTS(ts->dm,&sdm));
2028     PetscCall(DMTSView(sdm,viewer));
2029   } else if (isdraw) {
2030     PetscDraw draw;
2031     char      str[36];
2032     PetscReal x,y,bottom,h;
2033 
2034     PetscCall(PetscViewerDrawGetDraw(viewer,0,&draw));
2035     PetscCall(PetscDrawGetCurrentPoint(draw,&x,&y));
2036     PetscCall(PetscStrcpy(str,"TS: "));
2037     PetscCall(PetscStrcat(str,((PetscObject)ts)->type_name));
2038     PetscCall(PetscDrawStringBoxed(draw,x,y,PETSC_DRAW_BLACK,PETSC_DRAW_BLACK,str,NULL,&h));
2039     bottom = y - h;
2040     PetscCall(PetscDrawPushCurrentPoint(draw,x,bottom));
2041     PetscTryTypeMethod(ts,view,viewer);
2042     if (ts->adapt) PetscCall(TSAdaptView(ts->adapt,viewer));
2043     if (ts->snes)  PetscCall(SNESView(ts->snes,viewer));
2044     PetscCall(PetscDrawPopCurrentPoint(draw));
2045 #if defined(PETSC_HAVE_SAWS)
2046   } else if (issaws) {
2047     PetscMPIInt rank;
2048     const char  *name;
2049 
2050     PetscCall(PetscObjectGetName((PetscObject)ts,&name));
2051     PetscCallMPI(MPI_Comm_rank(PETSC_COMM_WORLD,&rank));
2052     if (!((PetscObject)ts)->amsmem && rank == 0) {
2053       char       dir[1024];
2054 
2055       PetscCall(PetscObjectViewSAWs((PetscObject)ts,viewer));
2056       PetscCall(PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time_step",name));
2057       PetscCallSAWs(SAWs_Register,(dir,&ts->steps,1,SAWs_READ,SAWs_INT));
2058       PetscCall(PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time",name));
2059       PetscCallSAWs(SAWs_Register,(dir,&ts->ptime,1,SAWs_READ,SAWs_DOUBLE));
2060     }
2061     PetscTryTypeMethod(ts,view,viewer);
2062 #endif
2063   }
2064   if (ts->snes && ts->usessnes)  {
2065     PetscCall(PetscViewerASCIIPushTab(viewer));
2066     PetscCall(SNESView(ts->snes,viewer));
2067     PetscCall(PetscViewerASCIIPopTab(viewer));
2068   }
2069   PetscCall(DMGetDMTS(ts->dm,&sdm));
2070   PetscCall(DMTSView(sdm,viewer));
2071 
2072   PetscCall(PetscViewerASCIIPushTab(viewer));
2073   PetscCall(PetscObjectTypeCompare((PetscObject)ts,TSSUNDIALS,&isundials));
2074   PetscCall(PetscViewerASCIIPopTab(viewer));
2075   PetscFunctionReturn(0);
2076 }
2077 
2078 /*@
2079    TSSetApplicationContext - Sets an optional user-defined context for
2080    the timesteppers.
2081 
2082    Logically Collective on TS
2083 
2084    Input Parameters:
2085 +  ts - the TS context obtained from TSCreate()
2086 -  usrP - optional user context
2087 
2088    Fortran Notes:
2089     To use this from Fortran you must write a Fortran interface definition for this
2090     function that tells Fortran the Fortran derived data type that you are passing in as the ctx argument.
2091 
2092    Level: intermediate
2093 
2094 .seealso: `TSGetApplicationContext()`
2095 @*/
2096 PetscErrorCode  TSSetApplicationContext(TS ts,void *usrP)
2097 {
2098   PetscFunctionBegin;
2099   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2100   ts->user = usrP;
2101   PetscFunctionReturn(0);
2102 }
2103 
2104 /*@
2105     TSGetApplicationContext - Gets the user-defined context for the
2106     timestepper.
2107 
2108     Not Collective
2109 
2110     Input Parameter:
2111 .   ts - the TS context obtained from TSCreate()
2112 
2113     Output Parameter:
2114 .   usrP - user context
2115 
2116    Fortran Notes:
2117     To use this from Fortran you must write a Fortran interface definition for this
2118     function that tells Fortran the Fortran derived data type that you are passing in as the ctx argument.
2119 
2120     Level: intermediate
2121 
2122 .seealso: `TSSetApplicationContext()`
2123 @*/
2124 PetscErrorCode  TSGetApplicationContext(TS ts,void *usrP)
2125 {
2126   PetscFunctionBegin;
2127   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2128   *(void**)usrP = ts->user;
2129   PetscFunctionReturn(0);
2130 }
2131 
2132 /*@
2133    TSGetStepNumber - Gets the number of steps completed.
2134 
2135    Not Collective
2136 
2137    Input Parameter:
2138 .  ts - the TS context obtained from TSCreate()
2139 
2140    Output Parameter:
2141 .  steps - number of steps completed so far
2142 
2143    Level: intermediate
2144 
2145 .seealso: `TSGetTime()`, `TSGetTimeStep()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostStage()`, `TSSetPostStep()`
2146 @*/
2147 PetscErrorCode TSGetStepNumber(TS ts,PetscInt *steps)
2148 {
2149   PetscFunctionBegin;
2150   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2151   PetscValidIntPointer(steps,2);
2152   *steps = ts->steps;
2153   PetscFunctionReturn(0);
2154 }
2155 
2156 /*@
2157    TSSetStepNumber - Sets the number of steps completed.
2158 
2159    Logically Collective on TS
2160 
2161    Input Parameters:
2162 +  ts - the TS context
2163 -  steps - number of steps completed so far
2164 
2165    Notes:
2166    For most uses of the TS solvers the user need not explicitly call
2167    TSSetStepNumber(), as the step counter is appropriately updated in
2168    TSSolve()/TSStep()/TSRollBack(). Power users may call this routine to
2169    reinitialize timestepping by setting the step counter to zero (and time
2170    to the initial time) to solve a similar problem with different initial
2171    conditions or parameters. Other possible use case is to continue
2172    timestepping from a previously interrupted run in such a way that TS
2173    monitors will be called with a initial nonzero step counter.
2174 
2175    Level: advanced
2176 
2177 .seealso: `TSGetStepNumber()`, `TSSetTime()`, `TSSetTimeStep()`, `TSSetSolution()`
2178 @*/
2179 PetscErrorCode TSSetStepNumber(TS ts,PetscInt steps)
2180 {
2181   PetscFunctionBegin;
2182   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2183   PetscValidLogicalCollectiveInt(ts,steps,2);
2184   PetscCheck(steps >= 0,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Step number must be non-negative");
2185   ts->steps = steps;
2186   PetscFunctionReturn(0);
2187 }
2188 
2189 /*@
2190    TSSetTimeStep - Allows one to reset the timestep at any time,
2191    useful for simple pseudo-timestepping codes.
2192 
2193    Logically Collective on TS
2194 
2195    Input Parameters:
2196 +  ts - the TS context obtained from TSCreate()
2197 -  time_step - the size of the timestep
2198 
2199    Level: intermediate
2200 
2201 .seealso: `TSGetTimeStep()`, `TSSetTime()`
2202 
2203 @*/
2204 PetscErrorCode  TSSetTimeStep(TS ts,PetscReal time_step)
2205 {
2206   PetscFunctionBegin;
2207   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2208   PetscValidLogicalCollectiveReal(ts,time_step,2);
2209   ts->time_step = time_step;
2210   PetscFunctionReturn(0);
2211 }
2212 
2213 /*@
2214    TSSetExactFinalTime - Determines whether to adapt the final time step to
2215      match the exact final time, interpolate solution to the exact final time,
2216      or just return at the final time TS computed.
2217 
2218   Logically Collective on TS
2219 
2220    Input Parameters:
2221 +   ts - the time-step context
2222 -   eftopt - exact final time option
2223 
2224 $  TS_EXACTFINALTIME_STEPOVER    - Don't do anything if final time is exceeded
2225 $  TS_EXACTFINALTIME_INTERPOLATE - Interpolate back to final time
2226 $  TS_EXACTFINALTIME_MATCHSTEP - Adapt final time step to match the final time
2227 
2228    Options Database:
2229 .   -ts_exact_final_time <stepover,interpolate,matchstep> - select the final step at runtime
2230 
2231    Warning: If you use the option TS_EXACTFINALTIME_STEPOVER the solution may be at a very different time
2232     then the final time you selected.
2233 
2234    Level: beginner
2235 
2236 .seealso: `TSExactFinalTimeOption`, `TSGetExactFinalTime()`
2237 @*/
2238 PetscErrorCode TSSetExactFinalTime(TS ts,TSExactFinalTimeOption eftopt)
2239 {
2240   PetscFunctionBegin;
2241   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2242   PetscValidLogicalCollectiveEnum(ts,eftopt,2);
2243   ts->exact_final_time = eftopt;
2244   PetscFunctionReturn(0);
2245 }
2246 
2247 /*@
2248    TSGetExactFinalTime - Gets the exact final time option.
2249 
2250    Not Collective
2251 
2252    Input Parameter:
2253 .  ts - the TS context
2254 
2255    Output Parameter:
2256 .  eftopt - exact final time option
2257 
2258    Level: beginner
2259 
2260 .seealso: `TSExactFinalTimeOption`, `TSSetExactFinalTime()`
2261 @*/
2262 PetscErrorCode TSGetExactFinalTime(TS ts,TSExactFinalTimeOption *eftopt)
2263 {
2264   PetscFunctionBegin;
2265   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2266   PetscValidPointer(eftopt,2);
2267   *eftopt = ts->exact_final_time;
2268   PetscFunctionReturn(0);
2269 }
2270 
2271 /*@
2272    TSGetTimeStep - Gets the current timestep size.
2273 
2274    Not Collective
2275 
2276    Input Parameter:
2277 .  ts - the TS context obtained from TSCreate()
2278 
2279    Output Parameter:
2280 .  dt - the current timestep size
2281 
2282    Level: intermediate
2283 
2284 .seealso: `TSSetTimeStep()`, `TSGetTime()`
2285 
2286 @*/
2287 PetscErrorCode  TSGetTimeStep(TS ts,PetscReal *dt)
2288 {
2289   PetscFunctionBegin;
2290   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2291   PetscValidRealPointer(dt,2);
2292   *dt = ts->time_step;
2293   PetscFunctionReturn(0);
2294 }
2295 
2296 /*@
2297    TSGetSolution - Returns the solution at the present timestep. It
2298    is valid to call this routine inside the function that you are evaluating
2299    in order to move to the new timestep. This vector not changed until
2300    the solution at the next timestep has been calculated.
2301 
2302    Not Collective, but Vec returned is parallel if TS is parallel
2303 
2304    Input Parameter:
2305 .  ts - the TS context obtained from TSCreate()
2306 
2307    Output Parameter:
2308 .  v - the vector containing the solution
2309 
2310    Note: If you used TSSetExactFinalTime(ts,TS_EXACTFINALTIME_MATCHSTEP); this does not return the solution at the requested
2311    final time. It returns the solution at the next timestep.
2312 
2313    Level: intermediate
2314 
2315 .seealso: `TSGetTimeStep()`, `TSGetTime()`, `TSGetSolveTime()`, `TSGetSolutionComponents()`, `TSSetSolutionFunction()`
2316 
2317 @*/
2318 PetscErrorCode  TSGetSolution(TS ts,Vec *v)
2319 {
2320   PetscFunctionBegin;
2321   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2322   PetscValidPointer(v,2);
2323   *v = ts->vec_sol;
2324   PetscFunctionReturn(0);
2325 }
2326 
2327 /*@
2328    TSGetSolutionComponents - Returns any solution components at the present
2329    timestep, if available for the time integration method being used.
2330    Solution components are quantities that share the same size and
2331    structure as the solution vector.
2332 
2333    Not Collective, but Vec returned is parallel if TS is parallel
2334 
2335    Parameters :
2336 +  ts - the TS context obtained from TSCreate() (input parameter).
2337 .  n - If v is PETSC_NULL, then the number of solution components is
2338        returned through n, else the n-th solution component is
2339        returned in v.
2340 -  v - the vector containing the n-th solution component
2341        (may be PETSC_NULL to use this function to find out
2342         the number of solutions components).
2343 
2344    Level: advanced
2345 
2346 .seealso: `TSGetSolution()`
2347 
2348 @*/
2349 PetscErrorCode  TSGetSolutionComponents(TS ts,PetscInt *n,Vec *v)
2350 {
2351   PetscFunctionBegin;
2352   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2353   if (!ts->ops->getsolutioncomponents) *n = 0;
2354   else PetscUseTypeMethod(ts,getsolutioncomponents ,n,v);
2355   PetscFunctionReturn(0);
2356 }
2357 
2358 /*@
2359    TSGetAuxSolution - Returns an auxiliary solution at the present
2360    timestep, if available for the time integration method being used.
2361 
2362    Not Collective, but Vec returned is parallel if TS is parallel
2363 
2364    Parameters :
2365 +  ts - the TS context obtained from TSCreate() (input parameter).
2366 -  v - the vector containing the auxiliary solution
2367 
2368    Level: intermediate
2369 
2370 .seealso: `TSGetSolution()`
2371 
2372 @*/
2373 PetscErrorCode  TSGetAuxSolution(TS ts,Vec *v)
2374 {
2375   PetscFunctionBegin;
2376   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2377   if (ts->ops->getauxsolution) PetscUseTypeMethod(ts,getauxsolution,v);
2378   else PetscCall(VecZeroEntries(*v));
2379   PetscFunctionReturn(0);
2380 }
2381 
2382 /*@
2383    TSGetTimeError - Returns the estimated error vector, if the chosen
2384    TSType has an error estimation functionality.
2385 
2386    Not Collective, but Vec returned is parallel if TS is parallel
2387 
2388    Note: MUST call after TSSetUp()
2389 
2390    Parameters :
2391 +  ts - the TS context obtained from TSCreate() (input parameter).
2392 .  n - current estimate (n=0) or previous one (n=-1)
2393 -  v - the vector containing the error (same size as the solution).
2394 
2395    Level: intermediate
2396 
2397 .seealso: `TSGetSolution()`, `TSSetTimeError()`
2398 
2399 @*/
2400 PetscErrorCode  TSGetTimeError(TS ts,PetscInt n,Vec *v)
2401 {
2402   PetscFunctionBegin;
2403   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2404   if (ts->ops->gettimeerror) PetscUseTypeMethod(ts,gettimeerror,n,v);
2405   else PetscCall(VecZeroEntries(*v));
2406   PetscFunctionReturn(0);
2407 }
2408 
2409 /*@
2410    TSSetTimeError - Sets the estimated error vector, if the chosen
2411    TSType has an error estimation functionality. This can be used
2412    to restart such a time integrator with a given error vector.
2413 
2414    Not Collective, but Vec returned is parallel if TS is parallel
2415 
2416    Parameters :
2417 +  ts - the TS context obtained from TSCreate() (input parameter).
2418 -  v - the vector containing the error (same size as the solution).
2419 
2420    Level: intermediate
2421 
2422 .seealso: `TSSetSolution()`, `TSGetTimeError)`
2423 
2424 @*/
2425 PetscErrorCode  TSSetTimeError(TS ts,Vec v)
2426 {
2427   PetscFunctionBegin;
2428   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2429   PetscCheck(ts->setupcalled,PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call TSSetUp() first");
2430   PetscTryTypeMethod(ts,settimeerror,v);
2431   PetscFunctionReturn(0);
2432 }
2433 
2434 /* ----- Routines to initialize and destroy a timestepper ---- */
2435 /*@
2436   TSSetProblemType - Sets the type of problem to be solved.
2437 
2438   Not collective
2439 
2440   Input Parameters:
2441 + ts   - The TS
2442 - type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms
2443 .vb
2444          U_t - A U = 0      (linear)
2445          U_t - A(t) U = 0   (linear)
2446          F(t,U,U_t) = 0     (nonlinear)
2447 .ve
2448 
2449    Level: beginner
2450 
2451 .seealso: `TSSetUp()`, `TSProblemType`, `TS`
2452 @*/
2453 PetscErrorCode  TSSetProblemType(TS ts, TSProblemType type)
2454 {
2455   PetscFunctionBegin;
2456   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
2457   ts->problem_type = type;
2458   if (type == TS_LINEAR) {
2459     SNES snes;
2460     PetscCall(TSGetSNES(ts,&snes));
2461     PetscCall(SNESSetType(snes,SNESKSPONLY));
2462   }
2463   PetscFunctionReturn(0);
2464 }
2465 
2466 /*@C
2467   TSGetProblemType - Gets the type of problem to be solved.
2468 
2469   Not collective
2470 
2471   Input Parameter:
2472 . ts   - The TS
2473 
2474   Output Parameter:
2475 . type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms
2476 .vb
2477          M U_t = A U
2478          M(t) U_t = A(t) U
2479          F(t,U,U_t)
2480 .ve
2481 
2482    Level: beginner
2483 
2484 .seealso: `TSSetUp()`, `TSProblemType`, `TS`
2485 @*/
2486 PetscErrorCode  TSGetProblemType(TS ts, TSProblemType *type)
2487 {
2488   PetscFunctionBegin;
2489   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
2490   PetscValidIntPointer(type,2);
2491   *type = ts->problem_type;
2492   PetscFunctionReturn(0);
2493 }
2494 
2495 /*
2496     Attempt to check/preset a default value for the exact final time option. This is needed at the beginning of TSSolve() and in TSSetUp()
2497 */
2498 static PetscErrorCode TSSetExactFinalTimeDefault(TS ts)
2499 {
2500   PetscBool      isnone;
2501 
2502   PetscFunctionBegin;
2503   PetscCall(TSGetAdapt(ts,&ts->adapt));
2504   PetscCall(TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type));
2505 
2506   PetscCall(PetscObjectTypeCompare((PetscObject)ts->adapt,TSADAPTNONE,&isnone));
2507   if (!isnone && ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) ts->exact_final_time = TS_EXACTFINALTIME_MATCHSTEP;
2508   else if (ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) ts->exact_final_time = TS_EXACTFINALTIME_INTERPOLATE;
2509   PetscFunctionReturn(0);
2510 }
2511 
2512 /*@
2513    TSSetUp - Sets up the internal data structures for the later use of a timestepper.
2514 
2515    Collective on TS
2516 
2517    Input Parameter:
2518 .  ts - the TS context obtained from TSCreate()
2519 
2520    Notes:
2521    For basic use of the TS solvers the user need not explicitly call
2522    TSSetUp(), since these actions will automatically occur during
2523    the call to TSStep() or TSSolve().  However, if one wishes to control this
2524    phase separately, TSSetUp() should be called after TSCreate()
2525    and optional routines of the form TSSetXXX(), but before TSStep() and TSSolve().
2526 
2527    Level: advanced
2528 
2529 .seealso: `TSCreate()`, `TSStep()`, `TSDestroy()`, `TSSolve()`
2530 @*/
2531 PetscErrorCode  TSSetUp(TS ts)
2532 {
2533   DM             dm;
2534   PetscErrorCode (*func)(SNES,Vec,Vec,void*);
2535   PetscErrorCode (*jac)(SNES,Vec,Mat,Mat,void*);
2536   TSIFunction    ifun;
2537   TSIJacobian    ijac;
2538   TSI2Jacobian   i2jac;
2539   TSRHSJacobian  rhsjac;
2540 
2541   PetscFunctionBegin;
2542   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2543   if (ts->setupcalled) PetscFunctionReturn(0);
2544 
2545   if (!((PetscObject)ts)->type_name) {
2546     PetscCall(TSGetIFunction(ts,NULL,&ifun,NULL));
2547     PetscCall(TSSetType(ts,ifun ? TSBEULER : TSEULER));
2548   }
2549 
2550   if (!ts->vec_sol) {
2551     PetscCheck(ts->dm,PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call TSSetSolution() first");
2552     PetscCall(DMCreateGlobalVector(ts->dm,&ts->vec_sol));
2553   }
2554 
2555   if (ts->tspan) {
2556     if (!ts->tspan->vecs_sol) PetscCall(VecDuplicateVecs(ts->vec_sol,ts->tspan->num_span_times,&ts->tspan->vecs_sol));
2557   }
2558   if (!ts->Jacp && ts->Jacprhs) { /* IJacobianP shares the same matrix with RHSJacobianP if only RHSJacobianP is provided */
2559     PetscCall(PetscObjectReference((PetscObject)ts->Jacprhs));
2560     ts->Jacp = ts->Jacprhs;
2561   }
2562 
2563   if (ts->quadraturets) {
2564     PetscCall(TSSetUp(ts->quadraturets));
2565     PetscCall(VecDestroy(&ts->vec_costintegrand));
2566     PetscCall(VecDuplicate(ts->quadraturets->vec_sol,&ts->vec_costintegrand));
2567   }
2568 
2569   PetscCall(TSGetRHSJacobian(ts,NULL,NULL,&rhsjac,NULL));
2570   if (rhsjac == TSComputeRHSJacobianConstant) {
2571     Mat Amat,Pmat;
2572     SNES snes;
2573     PetscCall(TSGetSNES(ts,&snes));
2574     PetscCall(SNESGetJacobian(snes,&Amat,&Pmat,NULL,NULL));
2575     /* Matching matrices implies that an IJacobian is NOT set, because if it had been set, the IJacobian's matrix would
2576      * have displaced the RHS matrix */
2577     if (Amat && Amat == ts->Arhs) {
2578       /* 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 */
2579       PetscCall(MatDuplicate(ts->Arhs,MAT_COPY_VALUES,&Amat));
2580       PetscCall(SNESSetJacobian(snes,Amat,NULL,NULL,NULL));
2581       PetscCall(MatDestroy(&Amat));
2582     }
2583     if (Pmat && Pmat == ts->Brhs) {
2584       PetscCall(MatDuplicate(ts->Brhs,MAT_COPY_VALUES,&Pmat));
2585       PetscCall(SNESSetJacobian(snes,NULL,Pmat,NULL,NULL));
2586       PetscCall(MatDestroy(&Pmat));
2587     }
2588   }
2589 
2590   PetscCall(TSGetAdapt(ts,&ts->adapt));
2591   PetscCall(TSAdaptSetDefaultType(ts->adapt,ts->default_adapt_type));
2592 
2593   PetscTryTypeMethod(ts,setup);
2594 
2595   PetscCall(TSSetExactFinalTimeDefault(ts));
2596 
2597   /* In the case where we've set a DMTSFunction or what have you, we need the default SNESFunction
2598      to be set right but can't do it elsewhere due to the overreliance on ctx=ts.
2599    */
2600   PetscCall(TSGetDM(ts,&dm));
2601   PetscCall(DMSNESGetFunction(dm,&func,NULL));
2602   if (!func) PetscCall(DMSNESSetFunction(dm,SNESTSFormFunction,ts));
2603 
2604   /* If the SNES doesn't have a jacobian set and the TS has an ijacobian or rhsjacobian set, set the SNES to use it.
2605      Otherwise, the SNES will use coloring internally to form the Jacobian.
2606    */
2607   PetscCall(DMSNESGetJacobian(dm,&jac,NULL));
2608   PetscCall(DMTSGetIJacobian(dm,&ijac,NULL));
2609   PetscCall(DMTSGetI2Jacobian(dm,&i2jac,NULL));
2610   PetscCall(DMTSGetRHSJacobian(dm,&rhsjac,NULL));
2611   if (!jac && (ijac || i2jac || rhsjac)) PetscCall(DMSNESSetJacobian(dm,SNESTSFormJacobian,ts));
2612 
2613   /* if time integration scheme has a starting method, call it */
2614   PetscTryTypeMethod(ts,startingmethod);
2615 
2616   ts->setupcalled = PETSC_TRUE;
2617   PetscFunctionReturn(0);
2618 }
2619 
2620 /*@
2621    TSReset - Resets a TS context and removes any allocated Vecs and Mats.
2622 
2623    Collective on TS
2624 
2625    Input Parameter:
2626 .  ts - the TS context obtained from TSCreate()
2627 
2628    Level: beginner
2629 
2630 .seealso: `TSCreate()`, `TSSetup()`, `TSDestroy()`
2631 @*/
2632 PetscErrorCode  TSReset(TS ts)
2633 {
2634   TS_RHSSplitLink ilink = ts->tsrhssplit,next;
2635 
2636   PetscFunctionBegin;
2637   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2638 
2639   PetscTryTypeMethod(ts,reset);
2640   if (ts->snes) PetscCall(SNESReset(ts->snes));
2641   if (ts->adapt) PetscCall(TSAdaptReset(ts->adapt));
2642 
2643   PetscCall(MatDestroy(&ts->Arhs));
2644   PetscCall(MatDestroy(&ts->Brhs));
2645   PetscCall(VecDestroy(&ts->Frhs));
2646   PetscCall(VecDestroy(&ts->vec_sol));
2647   PetscCall(VecDestroy(&ts->vec_dot));
2648   PetscCall(VecDestroy(&ts->vatol));
2649   PetscCall(VecDestroy(&ts->vrtol));
2650   PetscCall(VecDestroyVecs(ts->nwork,&ts->work));
2651 
2652   PetscCall(MatDestroy(&ts->Jacprhs));
2653   PetscCall(MatDestroy(&ts->Jacp));
2654   if (ts->forward_solve) PetscCall(TSForwardReset(ts));
2655   if (ts->quadraturets) {
2656     PetscCall(TSReset(ts->quadraturets));
2657     PetscCall(VecDestroy(&ts->vec_costintegrand));
2658   }
2659   while (ilink) {
2660     next = ilink->next;
2661     PetscCall(TSDestroy(&ilink->ts));
2662     PetscCall(PetscFree(ilink->splitname));
2663     PetscCall(ISDestroy(&ilink->is));
2664     PetscCall(PetscFree(ilink));
2665     ilink = next;
2666   }
2667   ts->tsrhssplit = NULL;
2668   ts->num_rhs_splits = 0;
2669   if (ts->tspan) {
2670     PetscCall(PetscFree(ts->tspan->span_times));
2671     PetscCall(VecDestroyVecs(ts->tspan->num_span_times,&ts->tspan->vecs_sol));
2672     PetscCall(PetscFree(ts->tspan));
2673   }
2674   ts->setupcalled = PETSC_FALSE;
2675   PetscFunctionReturn(0);
2676 }
2677 
2678 /*@C
2679    TSDestroy - Destroys the timestepper context that was created
2680    with TSCreate().
2681 
2682    Collective on TS
2683 
2684    Input Parameter:
2685 .  ts - the TS context obtained from TSCreate()
2686 
2687    Level: beginner
2688 
2689 .seealso: `TSCreate()`, `TSSetUp()`, `TSSolve()`
2690 @*/
2691 PetscErrorCode  TSDestroy(TS *ts)
2692 {
2693   PetscFunctionBegin;
2694   if (!*ts) PetscFunctionReturn(0);
2695   PetscValidHeaderSpecific(*ts,TS_CLASSID,1);
2696   if (--((PetscObject)(*ts))->refct > 0) {*ts = NULL; PetscFunctionReturn(0);}
2697 
2698   PetscCall(TSReset(*ts));
2699   PetscCall(TSAdjointReset(*ts));
2700   if ((*ts)->forward_solve) PetscCall(TSForwardReset(*ts));
2701 
2702   /* if memory was published with SAWs then destroy it */
2703   PetscCall(PetscObjectSAWsViewOff((PetscObject)*ts));
2704   PetscTryTypeMethod((*ts),destroy);
2705 
2706   PetscCall(TSTrajectoryDestroy(&(*ts)->trajectory));
2707 
2708   PetscCall(TSAdaptDestroy(&(*ts)->adapt));
2709   PetscCall(TSEventDestroy(&(*ts)->event));
2710 
2711   PetscCall(SNESDestroy(&(*ts)->snes));
2712   PetscCall(DMDestroy(&(*ts)->dm));
2713   PetscCall(TSMonitorCancel((*ts)));
2714   PetscCall(TSAdjointMonitorCancel((*ts)));
2715 
2716   PetscCall(TSDestroy(&(*ts)->quadraturets));
2717   PetscCall(PetscHeaderDestroy(ts));
2718   PetscFunctionReturn(0);
2719 }
2720 
2721 /*@
2722    TSGetSNES - Returns the SNES (nonlinear solver) associated with
2723    a TS (timestepper) context. Valid only for nonlinear problems.
2724 
2725    Not Collective, but SNES is parallel if TS is parallel
2726 
2727    Input Parameter:
2728 .  ts - the TS context obtained from TSCreate()
2729 
2730    Output Parameter:
2731 .  snes - the nonlinear solver context
2732 
2733    Notes:
2734    The user can then directly manipulate the SNES context to set various
2735    options, etc.  Likewise, the user can then extract and manipulate the
2736    KSP, KSP, and PC contexts as well.
2737 
2738    TSGetSNES() does not work for integrators that do not use SNES; in
2739    this case TSGetSNES() returns NULL in snes.
2740 
2741    Level: beginner
2742 
2743 @*/
2744 PetscErrorCode  TSGetSNES(TS ts,SNES *snes)
2745 {
2746   PetscFunctionBegin;
2747   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2748   PetscValidPointer(snes,2);
2749   if (!ts->snes) {
2750     PetscCall(SNESCreate(PetscObjectComm((PetscObject)ts),&ts->snes));
2751     PetscCall(PetscObjectSetOptions((PetscObject)ts->snes,((PetscObject)ts)->options));
2752     PetscCall(SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts));
2753     PetscCall(PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->snes));
2754     PetscCall(PetscObjectIncrementTabLevel((PetscObject)ts->snes,(PetscObject)ts,1));
2755     if (ts->dm) PetscCall(SNESSetDM(ts->snes,ts->dm));
2756     if (ts->problem_type == TS_LINEAR) PetscCall(SNESSetType(ts->snes,SNESKSPONLY));
2757   }
2758   *snes = ts->snes;
2759   PetscFunctionReturn(0);
2760 }
2761 
2762 /*@
2763    TSSetSNES - Set the SNES (nonlinear solver) to be used by the timestepping context
2764 
2765    Collective
2766 
2767    Input Parameters:
2768 +  ts - the TS context obtained from TSCreate()
2769 -  snes - the nonlinear solver context
2770 
2771    Notes:
2772    Most users should have the TS created by calling TSGetSNES()
2773 
2774    Level: developer
2775 
2776 @*/
2777 PetscErrorCode TSSetSNES(TS ts,SNES snes)
2778 {
2779   PetscErrorCode (*func)(SNES,Vec,Mat,Mat,void*);
2780 
2781   PetscFunctionBegin;
2782   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2783   PetscValidHeaderSpecific(snes,SNES_CLASSID,2);
2784   PetscCall(PetscObjectReference((PetscObject)snes));
2785   PetscCall(SNESDestroy(&ts->snes));
2786 
2787   ts->snes = snes;
2788 
2789   PetscCall(SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts));
2790   PetscCall(SNESGetJacobian(ts->snes,NULL,NULL,&func,NULL));
2791   if (func == SNESTSFormJacobian) PetscCall(SNESSetJacobian(ts->snes,NULL,NULL,SNESTSFormJacobian,ts));
2792   PetscFunctionReturn(0);
2793 }
2794 
2795 /*@
2796    TSGetKSP - Returns the KSP (linear solver) associated with
2797    a TS (timestepper) context.
2798 
2799    Not Collective, but KSP is parallel if TS is parallel
2800 
2801    Input Parameter:
2802 .  ts - the TS context obtained from TSCreate()
2803 
2804    Output Parameter:
2805 .  ksp - the nonlinear solver context
2806 
2807    Notes:
2808    The user can then directly manipulate the KSP context to set various
2809    options, etc.  Likewise, the user can then extract and manipulate the
2810    KSP and PC contexts as well.
2811 
2812    TSGetKSP() does not work for integrators that do not use KSP;
2813    in this case TSGetKSP() returns NULL in ksp.
2814 
2815    Level: beginner
2816 
2817 @*/
2818 PetscErrorCode  TSGetKSP(TS ts,KSP *ksp)
2819 {
2820   SNES           snes;
2821 
2822   PetscFunctionBegin;
2823   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2824   PetscValidPointer(ksp,2);
2825   PetscCheck(((PetscObject)ts)->type_name,PETSC_COMM_SELF,PETSC_ERR_ARG_NULL,"KSP is not created yet. Call TSSetType() first");
2826   PetscCheck(ts->problem_type == TS_LINEAR,PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Linear only; use TSGetSNES()");
2827   PetscCall(TSGetSNES(ts,&snes));
2828   PetscCall(SNESGetKSP(snes,ksp));
2829   PetscFunctionReturn(0);
2830 }
2831 
2832 /* ----------- Routines to set solver parameters ---------- */
2833 
2834 /*@
2835    TSSetMaxSteps - Sets the maximum number of steps to use.
2836 
2837    Logically Collective on TS
2838 
2839    Input Parameters:
2840 +  ts - the TS context obtained from TSCreate()
2841 -  maxsteps - maximum number of steps to use
2842 
2843    Options Database Keys:
2844 .  -ts_max_steps <maxsteps> - Sets maxsteps
2845 
2846    Notes:
2847    The default maximum number of steps is 5000
2848 
2849    Level: intermediate
2850 
2851 .seealso: `TSGetMaxSteps()`, `TSSetMaxTime()`, `TSSetExactFinalTime()`
2852 @*/
2853 PetscErrorCode TSSetMaxSteps(TS ts,PetscInt maxsteps)
2854 {
2855   PetscFunctionBegin;
2856   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2857   PetscValidLogicalCollectiveInt(ts,maxsteps,2);
2858   PetscCheck(maxsteps >= 0,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Maximum number of steps must be non-negative");
2859   ts->max_steps = maxsteps;
2860   PetscFunctionReturn(0);
2861 }
2862 
2863 /*@
2864    TSGetMaxSteps - Gets the maximum number of steps to use.
2865 
2866    Not Collective
2867 
2868    Input Parameters:
2869 .  ts - the TS context obtained from TSCreate()
2870 
2871    Output Parameter:
2872 .  maxsteps - maximum number of steps to use
2873 
2874    Level: advanced
2875 
2876 .seealso: `TSSetMaxSteps()`, `TSGetMaxTime()`, `TSSetMaxTime()`
2877 @*/
2878 PetscErrorCode TSGetMaxSteps(TS ts,PetscInt *maxsteps)
2879 {
2880   PetscFunctionBegin;
2881   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2882   PetscValidIntPointer(maxsteps,2);
2883   *maxsteps = ts->max_steps;
2884   PetscFunctionReturn(0);
2885 }
2886 
2887 /*@
2888    TSSetMaxTime - Sets the maximum (or final) time for timestepping.
2889 
2890    Logically Collective on TS
2891 
2892    Input Parameters:
2893 +  ts - the TS context obtained from TSCreate()
2894 -  maxtime - final time to step to
2895 
2896    Options Database Keys:
2897 .  -ts_max_time <maxtime> - Sets maxtime
2898 
2899    Notes:
2900    The default maximum time is 5.0
2901 
2902    Level: intermediate
2903 
2904 .seealso: `TSGetMaxTime()`, `TSSetMaxSteps()`, `TSSetExactFinalTime()`
2905 @*/
2906 PetscErrorCode TSSetMaxTime(TS ts,PetscReal maxtime)
2907 {
2908   PetscFunctionBegin;
2909   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2910   PetscValidLogicalCollectiveReal(ts,maxtime,2);
2911   ts->max_time = maxtime;
2912   PetscFunctionReturn(0);
2913 }
2914 
2915 /*@
2916    TSGetMaxTime - Gets the maximum (or final) time for timestepping.
2917 
2918    Not Collective
2919 
2920    Input Parameters:
2921 .  ts - the TS context obtained from TSCreate()
2922 
2923    Output Parameter:
2924 .  maxtime - final time to step to
2925 
2926    Level: advanced
2927 
2928 .seealso: `TSSetMaxTime()`, `TSGetMaxSteps()`, `TSSetMaxSteps()`
2929 @*/
2930 PetscErrorCode TSGetMaxTime(TS ts,PetscReal *maxtime)
2931 {
2932   PetscFunctionBegin;
2933   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2934   PetscValidRealPointer(maxtime,2);
2935   *maxtime = ts->max_time;
2936   PetscFunctionReturn(0);
2937 }
2938 
2939 /*@
2940    TSSetInitialTimeStep - Deprecated, use TSSetTime() and TSSetTimeStep().
2941 
2942    Level: deprecated
2943 
2944 @*/
2945 PetscErrorCode  TSSetInitialTimeStep(TS ts,PetscReal initial_time,PetscReal time_step)
2946 {
2947   PetscFunctionBegin;
2948   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2949   PetscCall(TSSetTime(ts,initial_time));
2950   PetscCall(TSSetTimeStep(ts,time_step));
2951   PetscFunctionReturn(0);
2952 }
2953 
2954 /*@
2955    TSGetDuration - Deprecated, use TSGetMaxSteps() and TSGetMaxTime().
2956 
2957    Level: deprecated
2958 
2959 @*/
2960 PetscErrorCode TSGetDuration(TS ts, PetscInt *maxsteps, PetscReal *maxtime)
2961 {
2962   PetscFunctionBegin;
2963   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
2964   if (maxsteps) {
2965     PetscValidIntPointer(maxsteps,2);
2966     *maxsteps = ts->max_steps;
2967   }
2968   if (maxtime) {
2969     PetscValidRealPointer(maxtime,3);
2970     *maxtime = ts->max_time;
2971   }
2972   PetscFunctionReturn(0);
2973 }
2974 
2975 /*@
2976    TSSetDuration - Deprecated, use TSSetMaxSteps() and TSSetMaxTime().
2977 
2978    Level: deprecated
2979 
2980 @*/
2981 PetscErrorCode TSSetDuration(TS ts,PetscInt maxsteps,PetscReal maxtime)
2982 {
2983   PetscFunctionBegin;
2984   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
2985   PetscValidLogicalCollectiveInt(ts,maxsteps,2);
2986   PetscValidLogicalCollectiveReal(ts,maxtime,3);
2987   if (maxsteps >= 0) ts->max_steps = maxsteps;
2988   if (maxtime != PETSC_DEFAULT) ts->max_time = maxtime;
2989   PetscFunctionReturn(0);
2990 }
2991 
2992 /*@
2993    TSGetTimeStepNumber - Deprecated, use TSGetStepNumber().
2994 
2995    Level: deprecated
2996 
2997 @*/
2998 PetscErrorCode TSGetTimeStepNumber(TS ts,PetscInt *steps) { return TSGetStepNumber(ts,steps); }
2999 
3000 /*@
3001    TSGetTotalSteps - Deprecated, use TSGetStepNumber().
3002 
3003    Level: deprecated
3004 
3005 @*/
3006 PetscErrorCode TSGetTotalSteps(TS ts,PetscInt *steps) { return TSGetStepNumber(ts,steps); }
3007 
3008 /*@
3009    TSSetSolution - Sets the initial solution vector
3010    for use by the TS routines.
3011 
3012    Logically Collective on TS
3013 
3014    Input Parameters:
3015 +  ts - the TS context obtained from TSCreate()
3016 -  u - the solution vector
3017 
3018    Level: beginner
3019 
3020 .seealso: `TSSetSolutionFunction()`, `TSGetSolution()`, `TSCreate()`
3021 @*/
3022 PetscErrorCode  TSSetSolution(TS ts,Vec u)
3023 {
3024   DM             dm;
3025 
3026   PetscFunctionBegin;
3027   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3028   PetscValidHeaderSpecific(u,VEC_CLASSID,2);
3029   PetscCall(PetscObjectReference((PetscObject)u));
3030   PetscCall(VecDestroy(&ts->vec_sol));
3031   ts->vec_sol = u;
3032 
3033   PetscCall(TSGetDM(ts,&dm));
3034   PetscCall(DMShellSetGlobalVector(dm,u));
3035   PetscFunctionReturn(0);
3036 }
3037 
3038 /*@C
3039   TSSetPreStep - Sets the general-purpose function
3040   called once at the beginning of each time step.
3041 
3042   Logically Collective on TS
3043 
3044   Input Parameters:
3045 + ts   - The TS context obtained from TSCreate()
3046 - func - The function
3047 
3048   Calling sequence of func:
3049 .vb
3050   PetscErrorCode func (TS ts);
3051 .ve
3052 
3053   Level: intermediate
3054 
3055 .seealso: `TSSetPreStage()`, `TSSetPostStage()`, `TSSetPostStep()`, `TSStep()`, `TSRestartStep()`
3056 @*/
3057 PetscErrorCode  TSSetPreStep(TS ts, PetscErrorCode (*func)(TS))
3058 {
3059   PetscFunctionBegin;
3060   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
3061   ts->prestep = func;
3062   PetscFunctionReturn(0);
3063 }
3064 
3065 /*@
3066   TSPreStep - Runs the user-defined pre-step function.
3067 
3068   Collective on TS
3069 
3070   Input Parameters:
3071 . ts   - The TS context obtained from TSCreate()
3072 
3073   Notes:
3074   TSPreStep() is typically used within time stepping implementations,
3075   so most users would not generally call this routine themselves.
3076 
3077   Level: developer
3078 
3079 .seealso: `TSSetPreStep()`, `TSPreStage()`, `TSPostStage()`, `TSPostStep()`
3080 @*/
3081 PetscErrorCode  TSPreStep(TS ts)
3082 {
3083   PetscFunctionBegin;
3084   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3085   if (ts->prestep) {
3086     Vec              U;
3087     PetscObjectId    idprev;
3088     PetscBool        sameObject;
3089     PetscObjectState sprev,spost;
3090 
3091     PetscCall(TSGetSolution(ts,&U));
3092     PetscCall(PetscObjectGetId((PetscObject)U,&idprev));
3093     PetscCall(PetscObjectStateGet((PetscObject)U,&sprev));
3094     PetscCallBack("TS callback preset",(*ts->prestep)(ts));
3095     PetscCall(TSGetSolution(ts,&U));
3096     PetscCall(PetscObjectCompareId((PetscObject)U,idprev,&sameObject));
3097     PetscCall(PetscObjectStateGet((PetscObject)U,&spost));
3098     if (!sameObject || sprev != spost) PetscCall(TSRestartStep(ts));
3099   }
3100   PetscFunctionReturn(0);
3101 }
3102 
3103 /*@C
3104   TSSetPreStage - Sets the general-purpose function
3105   called once at the beginning of each stage.
3106 
3107   Logically Collective on TS
3108 
3109   Input Parameters:
3110 + ts   - The TS context obtained from TSCreate()
3111 - func - The function
3112 
3113   Calling sequence of func:
3114 .vb
3115   PetscErrorCode func(TS ts, PetscReal stagetime);
3116 .ve
3117 
3118   Level: intermediate
3119 
3120   Note:
3121   There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3122   The time step number being computed can be queried using TSGetStepNumber() and the total size of the step being
3123   attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime().
3124 
3125 .seealso: `TSSetPostStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3126 @*/
3127 PetscErrorCode  TSSetPreStage(TS ts, PetscErrorCode (*func)(TS,PetscReal))
3128 {
3129   PetscFunctionBegin;
3130   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
3131   ts->prestage = func;
3132   PetscFunctionReturn(0);
3133 }
3134 
3135 /*@C
3136   TSSetPostStage - Sets the general-purpose function
3137   called once at the end of each stage.
3138 
3139   Logically Collective on TS
3140 
3141   Input Parameters:
3142 + ts   - The TS context obtained from TSCreate()
3143 - func - The function
3144 
3145   Calling sequence of func:
3146 .vb
3147   PetscErrorCode func(TS ts, PetscReal stagetime, PetscInt stageindex, Vec* Y);
3148 .ve
3149 
3150   Level: intermediate
3151 
3152   Note:
3153   There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3154   The time step number being computed can be queried using TSGetStepNumber() and the total size of the step being
3155   attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime().
3156 
3157 .seealso: `TSSetPreStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3158 @*/
3159 PetscErrorCode  TSSetPostStage(TS ts, PetscErrorCode (*func)(TS,PetscReal,PetscInt,Vec*))
3160 {
3161   PetscFunctionBegin;
3162   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
3163   ts->poststage = func;
3164   PetscFunctionReturn(0);
3165 }
3166 
3167 /*@C
3168   TSSetPostEvaluate - Sets the general-purpose function
3169   called once at the end of each step evaluation.
3170 
3171   Logically Collective on TS
3172 
3173   Input Parameters:
3174 + ts   - The TS context obtained from TSCreate()
3175 - func - The function
3176 
3177   Calling sequence of func:
3178 .vb
3179   PetscErrorCode func(TS ts);
3180 .ve
3181 
3182   Level: intermediate
3183 
3184   Note:
3185   Semantically, TSSetPostEvaluate() differs from TSSetPostStep() since the function it sets is called before event-handling
3186   thus guaranteeing the same solution (computed by the time-stepper) will be passed to it. On the other hand, TSPostStep()
3187   may be passed a different solution, possibly changed by the event handler. TSPostEvaluate() is called after the next step
3188   solution is evaluated allowing to modify it, if need be. The solution can be obtained with TSGetSolution(), the time step
3189   with TSGetTimeStep(), and the time at the start of the step is available via TSGetTime()
3190 
3191 .seealso: `TSSetPreStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3192 @*/
3193 PetscErrorCode  TSSetPostEvaluate(TS ts, PetscErrorCode (*func)(TS))
3194 {
3195   PetscFunctionBegin;
3196   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
3197   ts->postevaluate = func;
3198   PetscFunctionReturn(0);
3199 }
3200 
3201 /*@
3202   TSPreStage - Runs the user-defined pre-stage function set using TSSetPreStage()
3203 
3204   Collective on TS
3205 
3206   Input Parameters:
3207 . ts          - The TS context obtained from TSCreate()
3208   stagetime   - The absolute time of the current stage
3209 
3210   Notes:
3211   TSPreStage() is typically used within time stepping implementations,
3212   most users would not generally call this routine themselves.
3213 
3214   Level: developer
3215 
3216 .seealso: `TSPostStage()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3217 @*/
3218 PetscErrorCode  TSPreStage(TS ts, PetscReal stagetime)
3219 {
3220   PetscFunctionBegin;
3221   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3222   if (ts->prestage) PetscCallBack("TS callback prestage",(*ts->prestage)(ts,stagetime));
3223   PetscFunctionReturn(0);
3224 }
3225 
3226 /*@
3227   TSPostStage - Runs the user-defined post-stage function set using TSSetPostStage()
3228 
3229   Collective on TS
3230 
3231   Input Parameters:
3232 . ts          - The TS context obtained from TSCreate()
3233   stagetime   - The absolute time of the current stage
3234   stageindex  - Stage number
3235   Y           - Array of vectors (of size = total number
3236                 of stages) with the stage solutions
3237 
3238   Notes:
3239   TSPostStage() is typically used within time stepping implementations,
3240   most users would not generally call this routine themselves.
3241 
3242   Level: developer
3243 
3244 .seealso: `TSPreStage()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3245 @*/
3246 PetscErrorCode  TSPostStage(TS ts, PetscReal stagetime, PetscInt stageindex, Vec *Y)
3247 {
3248   PetscFunctionBegin;
3249   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3250   if (ts->poststage) PetscCallBack("TS callback poststage",(*ts->poststage)(ts,stagetime,stageindex,Y));
3251   PetscFunctionReturn(0);
3252 }
3253 
3254 /*@
3255   TSPostEvaluate - Runs the user-defined post-evaluate function set using TSSetPostEvaluate()
3256 
3257   Collective on TS
3258 
3259   Input Parameters:
3260 . ts          - The TS context obtained from TSCreate()
3261 
3262   Notes:
3263   TSPostEvaluate() is typically used within time stepping implementations,
3264   most users would not generally call this routine themselves.
3265 
3266   Level: developer
3267 
3268 .seealso: `TSSetPostEvaluate()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3269 @*/
3270 PetscErrorCode  TSPostEvaluate(TS ts)
3271 {
3272   PetscFunctionBegin;
3273   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3274   if (ts->postevaluate) {
3275     Vec              U;
3276     PetscObjectState sprev,spost;
3277 
3278     PetscCall(TSGetSolution(ts,&U));
3279     PetscCall(PetscObjectStateGet((PetscObject)U,&sprev));
3280     PetscCallBack("TS callback postevaluate",(*ts->postevaluate)(ts));
3281     PetscCall(PetscObjectStateGet((PetscObject)U,&spost));
3282     if (sprev != spost) PetscCall(TSRestartStep(ts));
3283   }
3284   PetscFunctionReturn(0);
3285 }
3286 
3287 /*@C
3288   TSSetPostStep - Sets the general-purpose function
3289   called once at the end of each time step.
3290 
3291   Logically Collective on TS
3292 
3293   Input Parameters:
3294 + ts   - The TS context obtained from TSCreate()
3295 - func - The function
3296 
3297   Calling sequence of func:
3298 $ func (TS ts);
3299 
3300   Notes:
3301   The function set by TSSetPostStep() is called after each successful step. The solution vector X
3302   obtained by TSGetSolution() may be different than that computed at the step end if the event handler
3303   locates an event and TSPostEvent() modifies it. Use TSSetPostEvaluate() if an unmodified solution is needed instead.
3304 
3305   Level: intermediate
3306 
3307 .seealso: `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostEvaluate()`, `TSGetTimeStep()`, `TSGetStepNumber()`, `TSGetTime()`, `TSRestartStep()`
3308 @*/
3309 PetscErrorCode  TSSetPostStep(TS ts, PetscErrorCode (*func)(TS))
3310 {
3311   PetscFunctionBegin;
3312   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
3313   ts->poststep = func;
3314   PetscFunctionReturn(0);
3315 }
3316 
3317 /*@
3318   TSPostStep - Runs the user-defined post-step function.
3319 
3320   Collective on TS
3321 
3322   Input Parameters:
3323 . ts   - The TS context obtained from TSCreate()
3324 
3325   Notes:
3326   TSPostStep() is typically used within time stepping implementations,
3327   so most users would not generally call this routine themselves.
3328 
3329   Level: developer
3330 
3331 @*/
3332 PetscErrorCode  TSPostStep(TS ts)
3333 {
3334   PetscFunctionBegin;
3335   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3336   if (ts->poststep) {
3337     Vec              U;
3338     PetscObjectId    idprev;
3339     PetscBool        sameObject;
3340     PetscObjectState sprev,spost;
3341 
3342     PetscCall(TSGetSolution(ts,&U));
3343     PetscCall(PetscObjectGetId((PetscObject)U,&idprev));
3344     PetscCall(PetscObjectStateGet((PetscObject)U,&sprev));
3345     PetscCallBack("TS callback poststep",(*ts->poststep)(ts));
3346     PetscCall(TSGetSolution(ts,&U));
3347     PetscCall(PetscObjectCompareId((PetscObject)U,idprev,&sameObject));
3348     PetscCall(PetscObjectStateGet((PetscObject)U,&spost));
3349     if (!sameObject || sprev != spost) PetscCall(TSRestartStep(ts));
3350   }
3351   PetscFunctionReturn(0);
3352 }
3353 
3354 /*@
3355    TSInterpolate - Interpolate the solution computed during the previous step to an arbitrary location in the interval
3356 
3357    Collective on TS
3358 
3359    Input Parameters:
3360 +  ts - time stepping context
3361 -  t - time to interpolate to
3362 
3363    Output Parameter:
3364 .  U - state at given time
3365 
3366    Level: intermediate
3367 
3368    Developer Notes:
3369    TSInterpolate() and the storing of previous steps/stages should be generalized to support delay differential equations and continuous adjoints.
3370 
3371 .seealso: `TSSetExactFinalTime()`, `TSSolve()`
3372 @*/
3373 PetscErrorCode TSInterpolate(TS ts,PetscReal t,Vec U)
3374 {
3375   PetscFunctionBegin;
3376   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3377   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
3378   PetscCheck(t >= ts->ptime_prev && t <= ts->ptime,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Requested time %g not in last time steps [%g,%g]",(double)t,(double)ts->ptime_prev,(double)ts->ptime);
3379   PetscUseTypeMethod(ts,interpolate ,t,U);
3380   PetscFunctionReturn(0);
3381 }
3382 
3383 /*@
3384    TSStep - Steps one time step
3385 
3386    Collective on TS
3387 
3388    Input Parameter:
3389 .  ts - the TS context obtained from TSCreate()
3390 
3391    Level: developer
3392 
3393    Notes:
3394    The public interface for the ODE/DAE solvers is TSSolve(), you should almost for sure be using that routine and not this routine.
3395 
3396    The hook set using TSSetPreStep() is called before each attempt to take the step. In general, the time step size may
3397    be changed due to adaptive error controller or solve failures. Note that steps may contain multiple stages.
3398 
3399    This may over-step the final time provided in TSSetMaxTime() depending on the time-step used. TSSolve() interpolates to exactly the
3400    time provided in TSSetMaxTime(). One can use TSInterpolate() to determine an interpolated solution within the final timestep.
3401 
3402 .seealso: `TSCreate()`, `TSSetUp()`, `TSDestroy()`, `TSSolve()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostStage()`, `TSInterpolate()`
3403 @*/
3404 PetscErrorCode  TSStep(TS ts)
3405 {
3406   static PetscBool cite = PETSC_FALSE;
3407   PetscReal        ptime;
3408 
3409   PetscFunctionBegin;
3410   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3411   PetscCall(PetscCitationsRegister("@article{tspaper,\n"
3412                                    "  title         = {{PETSc/TS}: A Modern Scalable {DAE/ODE} Solver Library},\n"
3413                                    "  author        = {Abhyankar, Shrirang and Brown, Jed and Constantinescu, Emil and Ghosh, Debojyoti and Smith, Barry F. and Zhang, Hong},\n"
3414                                    "  journal       = {arXiv e-preprints},\n"
3415                                    "  eprint        = {1806.01437},\n"
3416                                    "  archivePrefix = {arXiv},\n"
3417                                    "  year          = {2018}\n}\n",&cite));
3418   PetscCall(TSSetUp(ts));
3419   PetscCall(TSTrajectorySetUp(ts->trajectory,ts));
3420 
3421   PetscCheck(ts->max_time < PETSC_MAX_REAL || ts->max_steps != PETSC_MAX_INT,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONGSTATE,"You must call TSSetMaxTime() or TSSetMaxSteps(), or use -ts_max_time <time> or -ts_max_steps <steps>");
3422   PetscCheck(ts->exact_final_time != TS_EXACTFINALTIME_UNSPECIFIED,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONGSTATE,"You must call TSSetExactFinalTime() or use -ts_exact_final_time <stepover,interpolate,matchstep> before calling TSStep()");
3423   PetscCheck(ts->exact_final_time != TS_EXACTFINALTIME_MATCHSTEP || ts->adapt,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Since TS is not adaptive you cannot use TS_EXACTFINALTIME_MATCHSTEP, suggest TS_EXACTFINALTIME_INTERPOLATE");
3424 
3425   if (!ts->steps) ts->ptime_prev = ts->ptime;
3426   ptime = ts->ptime; ts->ptime_prev_rollback = ts->ptime_prev;
3427   ts->reason = TS_CONVERGED_ITERATING;
3428 
3429   PetscCall(PetscLogEventBegin(TS_Step,ts,0,0,0));
3430   PetscUseTypeMethod(ts,step);
3431   PetscCall(PetscLogEventEnd(TS_Step,ts,0,0,0));
3432 
3433   if (ts->tspan && PetscIsCloseAtTol(ts->ptime,ts->tspan->span_times[ts->tspan->spanctr],ts->tspan->reltol*ts->time_step+ts->tspan->abstol,0) && ts->tspan->spanctr < ts->tspan->num_span_times) PetscCall(VecCopy(ts->vec_sol,ts->tspan->vecs_sol[ts->tspan->spanctr++]));
3434   if (ts->reason >= 0) {
3435     ts->ptime_prev = ptime;
3436     ts->steps++;
3437     ts->steprollback = PETSC_FALSE;
3438     ts->steprestart  = PETSC_FALSE;
3439   }
3440   if (!ts->reason) {
3441     if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
3442     else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
3443   }
3444 
3445   PetscCheck(ts->reason >= 0 || !ts->errorifstepfailed || ts->reason != TS_DIVERGED_NONLINEAR_SOLVE,PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed due to %s, increase -ts_max_snes_failures or make negative to attempt recovery",TSConvergedReasons[ts->reason]);
3446   PetscCheck(ts->reason >= 0 || !ts->errorifstepfailed,PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed due to %s",TSConvergedReasons[ts->reason]);
3447   PetscFunctionReturn(0);
3448 }
3449 
3450 /*@
3451    TSEvaluateWLTE - Evaluate the weighted local truncation error norm
3452    at the end of a time step with a given order of accuracy.
3453 
3454    Collective on TS
3455 
3456    Input Parameters:
3457 +  ts - time stepping context
3458 -  wnormtype - norm type, either NORM_2 or NORM_INFINITY
3459 
3460    Input/Output Parameter:
3461 .  order - optional, desired order for the error evaluation or PETSC_DECIDE;
3462            on output, the actual order of the error evaluation
3463 
3464    Output Parameter:
3465 .  wlte - the weighted local truncation error norm
3466 
3467    Level: advanced
3468 
3469    Notes:
3470    If the timestepper cannot evaluate the error in a particular step
3471    (eg. in the first step or restart steps after event handling),
3472    this routine returns wlte=-1.0 .
3473 
3474 .seealso: `TSStep()`, `TSAdapt`, `TSErrorWeightedNorm()`
3475 @*/
3476 PetscErrorCode TSEvaluateWLTE(TS ts,NormType wnormtype,PetscInt *order,PetscReal *wlte)
3477 {
3478   PetscFunctionBegin;
3479   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3480   PetscValidType(ts,1);
3481   PetscValidLogicalCollectiveEnum(ts,wnormtype,2);
3482   if (order) PetscValidIntPointer(order,3);
3483   if (order) PetscValidLogicalCollectiveInt(ts,*order,3);
3484   PetscValidRealPointer(wlte,4);
3485   PetscCheck(wnormtype == NORM_2 || wnormtype == NORM_INFINITY,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]);
3486   PetscUseTypeMethod(ts,evaluatewlte ,wnormtype,order,wlte);
3487   PetscFunctionReturn(0);
3488 }
3489 
3490 /*@
3491    TSEvaluateStep - Evaluate the solution at the end of a time step with a given order of accuracy.
3492 
3493    Collective on TS
3494 
3495    Input Parameters:
3496 +  ts - time stepping context
3497 .  order - desired order of accuracy
3498 -  done - whether the step was evaluated at this order (pass NULL to generate an error if not available)
3499 
3500    Output Parameter:
3501 .  U - state at the end of the current step
3502 
3503    Level: advanced
3504 
3505    Notes:
3506    This function cannot be called until all stages have been evaluated.
3507    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.
3508 
3509 .seealso: `TSStep()`, `TSAdapt`
3510 @*/
3511 PetscErrorCode TSEvaluateStep(TS ts,PetscInt order,Vec U,PetscBool *done)
3512 {
3513   PetscFunctionBegin;
3514   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3515   PetscValidType(ts,1);
3516   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
3517   PetscUseTypeMethod(ts,evaluatestep ,order,U,done);
3518   PetscFunctionReturn(0);
3519 }
3520 
3521 /*@C
3522   TSGetComputeInitialCondition - Get the function used to automatically compute an initial condition for the timestepping.
3523 
3524   Not collective
3525 
3526   Input Parameter:
3527 . ts        - time stepping context
3528 
3529   Output Parameter:
3530 . initConditions - The function which computes an initial condition
3531 
3532    Level: advanced
3533 
3534    Notes:
3535    The calling sequence for the function is
3536 $ initCondition(TS ts, Vec u)
3537 $ ts - The timestepping context
3538 $ u  - The input vector in which the initial condition is stored
3539 
3540 .seealso: `TSSetComputeInitialCondition()`, `TSComputeInitialCondition()`
3541 @*/
3542 PetscErrorCode TSGetComputeInitialCondition(TS ts, PetscErrorCode (**initCondition)(TS, Vec))
3543 {
3544   PetscFunctionBegin;
3545   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3546   PetscValidPointer(initCondition, 2);
3547   *initCondition = ts->ops->initcondition;
3548   PetscFunctionReturn(0);
3549 }
3550 
3551 /*@C
3552   TSSetComputeInitialCondition - Set the function used to automatically compute an initial condition for the timestepping.
3553 
3554   Logically collective on ts
3555 
3556   Input Parameters:
3557 + ts        - time stepping context
3558 - initCondition - The function which computes an initial condition
3559 
3560   Level: advanced
3561 
3562   Calling sequence for initCondition:
3563 $ PetscErrorCode initCondition(TS ts, Vec u)
3564 
3565 + ts - The timestepping context
3566 - u  - The input vector in which the initial condition is to be stored
3567 
3568 .seealso: `TSGetComputeInitialCondition()`, `TSComputeInitialCondition()`
3569 @*/
3570 PetscErrorCode TSSetComputeInitialCondition(TS ts, PetscErrorCode (*initCondition)(TS, Vec))
3571 {
3572   PetscFunctionBegin;
3573   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3574   PetscValidFunction(initCondition, 2);
3575   ts->ops->initcondition = initCondition;
3576   PetscFunctionReturn(0);
3577 }
3578 
3579 /*@
3580   TSComputeInitialCondition - Compute an initial condition for the timestepping using the function previously set.
3581 
3582   Collective on ts
3583 
3584   Input Parameters:
3585 + ts - time stepping context
3586 - u  - The Vec to store the condition in which will be used in TSSolve()
3587 
3588   Level: advanced
3589 
3590 .seealso: `TSGetComputeInitialCondition()`, `TSSetComputeInitialCondition()`, `TSSolve()`
3591 @*/
3592 PetscErrorCode TSComputeInitialCondition(TS ts, Vec u)
3593 {
3594   PetscFunctionBegin;
3595   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3596   PetscValidHeaderSpecific(u, VEC_CLASSID, 2);
3597   PetscTryTypeMethod(ts,initcondition, u);
3598   PetscFunctionReturn(0);
3599 }
3600 
3601 /*@C
3602   TSGetComputeExactError - Get the function used to automatically compute the exact error for the timestepping.
3603 
3604   Not collective
3605 
3606   Input Parameter:
3607 . ts         - time stepping context
3608 
3609   Output Parameter:
3610 . exactError - The function which computes the solution error
3611 
3612   Level: advanced
3613 
3614   Calling sequence for exactError:
3615 $ PetscErrorCode exactError(TS ts, Vec u)
3616 
3617 + ts - The timestepping context
3618 . u  - The approximate solution vector
3619 - e  - The input vector in which the error is stored
3620 
3621 .seealso: `TSGetComputeExactError()`, `TSComputeExactError()`
3622 @*/
3623 PetscErrorCode TSGetComputeExactError(TS ts, PetscErrorCode (**exactError)(TS, Vec, Vec))
3624 {
3625   PetscFunctionBegin;
3626   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3627   PetscValidPointer(exactError, 2);
3628   *exactError = ts->ops->exacterror;
3629   PetscFunctionReturn(0);
3630 }
3631 
3632 /*@C
3633   TSSetComputeExactError - Set the function used to automatically compute the exact error for the timestepping.
3634 
3635   Logically collective on ts
3636 
3637   Input Parameters:
3638 + ts         - time stepping context
3639 - exactError - The function which computes the solution error
3640 
3641   Level: advanced
3642 
3643   Calling sequence for exactError:
3644 $ PetscErrorCode exactError(TS ts, Vec u)
3645 
3646 + ts - The timestepping context
3647 . u  - The approximate solution vector
3648 - e  - The input vector in which the error is stored
3649 
3650 .seealso: `TSGetComputeExactError()`, `TSComputeExactError()`
3651 @*/
3652 PetscErrorCode TSSetComputeExactError(TS ts, PetscErrorCode (*exactError)(TS, Vec, Vec))
3653 {
3654   PetscFunctionBegin;
3655   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3656   PetscValidFunction(exactError, 2);
3657   ts->ops->exacterror = exactError;
3658   PetscFunctionReturn(0);
3659 }
3660 
3661 /*@
3662   TSComputeExactError - Compute the solution error for the timestepping using the function previously set.
3663 
3664   Collective on ts
3665 
3666   Input Parameters:
3667 + ts - time stepping context
3668 . u  - The approximate solution
3669 - e  - The Vec used to store the error
3670 
3671   Level: advanced
3672 
3673 .seealso: `TSGetComputeInitialCondition()`, `TSSetComputeInitialCondition()`, `TSSolve()`
3674 @*/
3675 PetscErrorCode TSComputeExactError(TS ts, Vec u, Vec e)
3676 {
3677   PetscFunctionBegin;
3678   PetscValidHeaderSpecific(ts, TS_CLASSID, 1);
3679   PetscValidHeaderSpecific(u, VEC_CLASSID, 2);
3680   PetscValidHeaderSpecific(e, VEC_CLASSID, 3);
3681   PetscTryTypeMethod(ts,exacterror, u, e);
3682   PetscFunctionReturn(0);
3683 }
3684 
3685 /*@
3686    TSSolve - Steps the requested number of timesteps.
3687 
3688    Collective on TS
3689 
3690    Input Parameters:
3691 +  ts - the TS context obtained from TSCreate()
3692 -  u - the solution vector  (can be null if TSSetSolution() was used and TSSetExactFinalTime(ts,TS_EXACTFINALTIME_MATCHSTEP) was not used,
3693                              otherwise must contain the initial conditions and will contain the solution at the final requested time
3694 
3695    Level: beginner
3696 
3697    Notes:
3698    The final time returned by this function may be different from the time of the internally
3699    held state accessible by TSGetSolution() and TSGetTime() because the method may have
3700    stepped over the final time.
3701 
3702 .seealso: `TSCreate()`, `TSSetSolution()`, `TSStep()`, `TSGetTime()`, `TSGetSolveTime()`
3703 @*/
3704 PetscErrorCode TSSolve(TS ts,Vec u)
3705 {
3706   Vec               solution;
3707 
3708   PetscFunctionBegin;
3709   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3710   if (u) PetscValidHeaderSpecific(u,VEC_CLASSID,2);
3711 
3712   PetscCall(TSSetExactFinalTimeDefault(ts));
3713   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 */
3714     if (!ts->vec_sol || u == ts->vec_sol) {
3715       PetscCall(VecDuplicate(u,&solution));
3716       PetscCall(TSSetSolution(ts,solution));
3717       PetscCall(VecDestroy(&solution)); /* grant ownership */
3718     }
3719     PetscCall(VecCopy(u,ts->vec_sol));
3720     PetscCheck(!ts->forward_solve,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Sensitivity analysis does not support the mode TS_EXACTFINALTIME_INTERPOLATE");
3721   } else if (u) PetscCall(TSSetSolution(ts,u));
3722   PetscCall(TSSetUp(ts));
3723   PetscCall(TSTrajectorySetUp(ts->trajectory,ts));
3724 
3725   PetscCheck(ts->max_time < PETSC_MAX_REAL || ts->max_steps != PETSC_MAX_INT,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONGSTATE,"You must call TSSetMaxTime() or TSSetMaxSteps(), or use -ts_max_time <time> or -ts_max_steps <steps>");
3726   PetscCheck(ts->exact_final_time != TS_EXACTFINALTIME_UNSPECIFIED,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONGSTATE,"You must call TSSetExactFinalTime() or use -ts_exact_final_time <stepover,interpolate,matchstep> before calling TSSolve()");
3727   PetscCheck(ts->exact_final_time != TS_EXACTFINALTIME_MATCHSTEP || ts->adapt,PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Since TS is not adaptive you cannot use TS_EXACTFINALTIME_MATCHSTEP, suggest TS_EXACTFINALTIME_INTERPOLATE");
3728   PetscCheck(!(ts->tspan && ts->exact_final_time != TS_EXACTFINALTIME_MATCHSTEP),PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"You must use TS_EXACTFINALTIME_MATCHSTEP when using time span");
3729 
3730   if (ts->tspan && PetscIsCloseAtTol(ts->ptime,ts->tspan->span_times[0],ts->tspan->reltol*ts->time_step+ts->tspan->abstol,0)) { /* starting point in time span */
3731     PetscCall(VecCopy(ts->vec_sol,ts->tspan->vecs_sol[0]));
3732     ts->tspan->spanctr = 1;
3733   }
3734 
3735   if (ts->forward_solve) PetscCall(TSForwardSetUp(ts));
3736 
3737   /* reset number of steps only when the step is not restarted. ARKIMEX
3738      restarts the step after an event. Resetting these counters in such case causes
3739      TSTrajectory to incorrectly save the output files
3740   */
3741   /* reset time step and iteration counters */
3742   if (!ts->steps) {
3743     ts->ksp_its           = 0;
3744     ts->snes_its          = 0;
3745     ts->num_snes_failures = 0;
3746     ts->reject            = 0;
3747     ts->steprestart       = PETSC_TRUE;
3748     ts->steprollback      = PETSC_FALSE;
3749     ts->rhsjacobian.time  = PETSC_MIN_REAL;
3750   }
3751 
3752   /* make sure initial time step does not overshoot final time or the next point in tspan */
3753   if (ts->exact_final_time == TS_EXACTFINALTIME_MATCHSTEP) {
3754     PetscReal maxdt;
3755     PetscReal dt = ts->time_step;
3756 
3757     if (ts->tspan) maxdt = ts->tspan->span_times[ts->tspan->spanctr] - ts->ptime;
3758     else maxdt = ts->max_time - ts->ptime;
3759     ts->time_step = dt >= maxdt ? maxdt : (PetscIsCloseAtTol(dt,maxdt,10*PETSC_MACHINE_EPSILON,0) ? maxdt : dt);
3760   }
3761   ts->reason = TS_CONVERGED_ITERATING;
3762 
3763   {
3764     PetscViewer       viewer;
3765     PetscViewerFormat format;
3766     PetscBool         flg;
3767     static PetscBool  incall = PETSC_FALSE;
3768 
3769     if (!incall) {
3770       /* Estimate the convergence rate of the time discretization */
3771       PetscCall(PetscOptionsGetViewer(PetscObjectComm((PetscObject) ts),((PetscObject)ts)->options, ((PetscObject) ts)->prefix, "-ts_convergence_estimate", &viewer, &format, &flg));
3772       if (flg) {
3773         PetscConvEst conv;
3774         DM           dm;
3775         PetscReal   *alpha; /* Convergence rate of the solution error for each field in the L_2 norm */
3776         PetscInt     Nf;
3777         PetscBool    checkTemporal = PETSC_TRUE;
3778 
3779         incall = PETSC_TRUE;
3780         PetscCall(PetscOptionsGetBool(((PetscObject)ts)->options, ((PetscObject) ts)->prefix, "-ts_convergence_temporal", &checkTemporal, &flg));
3781         PetscCall(TSGetDM(ts, &dm));
3782         PetscCall(DMGetNumFields(dm, &Nf));
3783         PetscCall(PetscCalloc1(PetscMax(Nf, 1), &alpha));
3784         PetscCall(PetscConvEstCreate(PetscObjectComm((PetscObject) ts), &conv));
3785         PetscCall(PetscConvEstUseTS(conv, checkTemporal));
3786         PetscCall(PetscConvEstSetSolver(conv, (PetscObject) ts));
3787         PetscCall(PetscConvEstSetFromOptions(conv));
3788         PetscCall(PetscConvEstSetUp(conv));
3789         PetscCall(PetscConvEstGetConvRate(conv, alpha));
3790         PetscCall(PetscViewerPushFormat(viewer, format));
3791         PetscCall(PetscConvEstRateView(conv, alpha, viewer));
3792         PetscCall(PetscViewerPopFormat(viewer));
3793         PetscCall(PetscViewerDestroy(&viewer));
3794         PetscCall(PetscConvEstDestroy(&conv));
3795         PetscCall(PetscFree(alpha));
3796         incall = PETSC_FALSE;
3797       }
3798     }
3799   }
3800 
3801   PetscCall(TSViewFromOptions(ts,NULL,"-ts_view_pre"));
3802 
3803   if (ts->ops->solve) { /* This private interface is transitional and should be removed when all implementations are updated. */
3804     PetscUseTypeMethod(ts,solve);
3805     if (u) PetscCall(VecCopy(ts->vec_sol,u));
3806     ts->solvetime = ts->ptime;
3807     solution = ts->vec_sol;
3808   } else { /* Step the requested number of timesteps. */
3809     if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
3810     else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
3811 
3812     if (!ts->steps) {
3813       PetscCall(TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol));
3814       PetscCall(TSEventInitialize(ts->event,ts,ts->ptime,ts->vec_sol));
3815     }
3816 
3817     while (!ts->reason) {
3818       PetscCall(TSMonitor(ts,ts->steps,ts->ptime,ts->vec_sol));
3819       if (!ts->steprollback) PetscCall(TSPreStep(ts));
3820       PetscCall(TSStep(ts));
3821       if (ts->testjacobian) PetscCall(TSRHSJacobianTest(ts,NULL));
3822       if (ts->testjacobiantranspose) PetscCall(TSRHSJacobianTestTranspose(ts,NULL));
3823       if (ts->quadraturets && ts->costintegralfwd) { /* Must evaluate the cost integral before event is handled. The cost integral value can also be rolled back. */
3824         if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
3825         PetscCall(TSForwardCostIntegral(ts));
3826         if (ts->reason >= 0) ts->steps++;
3827       }
3828       if (ts->forward_solve) { /* compute forward sensitivities before event handling because postevent() may change RHS and jump conditions may have to be applied */
3829         if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
3830         PetscCall(TSForwardStep(ts));
3831         if (ts->reason >= 0) ts->steps++;
3832       }
3833       PetscCall(TSPostEvaluate(ts));
3834       PetscCall(TSEventHandler(ts)); /* The right-hand side may be changed due to event. Be careful with Any computation using the RHS information after this point. */
3835       if (ts->steprollback) PetscCall(TSPostEvaluate(ts));
3836       if (!ts->steprollback) {
3837         PetscCall(TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol));
3838         PetscCall(TSPostStep(ts));
3839       }
3840     }
3841     PetscCall(TSMonitor(ts,ts->steps,ts->ptime,ts->vec_sol));
3842 
3843     if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE && ts->ptime > ts->max_time) {
3844       PetscCall(TSInterpolate(ts,ts->max_time,u));
3845       ts->solvetime = ts->max_time;
3846       solution = u;
3847       PetscCall(TSMonitor(ts,-1,ts->solvetime,solution));
3848     } else {
3849       if (u) PetscCall(VecCopy(ts->vec_sol,u));
3850       ts->solvetime = ts->ptime;
3851       solution = ts->vec_sol;
3852     }
3853   }
3854 
3855   PetscCall(TSViewFromOptions(ts,NULL,"-ts_view"));
3856   PetscCall(VecViewFromOptions(solution,(PetscObject)ts,"-ts_view_solution"));
3857   PetscCall(PetscObjectSAWsBlock((PetscObject)ts));
3858   if (ts->adjoint_solve) PetscCall(TSAdjointSolve(ts));
3859   PetscFunctionReturn(0);
3860 }
3861 
3862 /*@
3863    TSGetTime - Gets the time of the most recently completed step.
3864 
3865    Not Collective
3866 
3867    Input Parameter:
3868 .  ts - the TS context obtained from TSCreate()
3869 
3870    Output Parameter:
3871 .  t  - the current time. This time may not corresponds to the final time set with TSSetMaxTime(), use TSGetSolveTime().
3872 
3873    Level: beginner
3874 
3875    Note:
3876    When called during time step evaluation (e.g. during residual evaluation or via hooks set using TSSetPreStep(),
3877    TSSetPreStage(), TSSetPostStage(), or TSSetPostStep()), the time is the time at the start of the step being evaluated.
3878 
3879 .seealso: `TSGetSolveTime()`, `TSSetTime()`, `TSGetTimeStep()`, `TSGetStepNumber()`
3880 
3881 @*/
3882 PetscErrorCode  TSGetTime(TS ts,PetscReal *t)
3883 {
3884   PetscFunctionBegin;
3885   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3886   PetscValidRealPointer(t,2);
3887   *t = ts->ptime;
3888   PetscFunctionReturn(0);
3889 }
3890 
3891 /*@
3892    TSGetPrevTime - Gets the starting time of the previously completed step.
3893 
3894    Not Collective
3895 
3896    Input Parameter:
3897 .  ts - the TS context obtained from TSCreate()
3898 
3899    Output Parameter:
3900 .  t  - the previous time
3901 
3902    Level: beginner
3903 
3904 .seealso: `TSGetTime()`, `TSGetSolveTime()`, `TSGetTimeStep()`
3905 
3906 @*/
3907 PetscErrorCode  TSGetPrevTime(TS ts,PetscReal *t)
3908 {
3909   PetscFunctionBegin;
3910   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3911   PetscValidRealPointer(t,2);
3912   *t = ts->ptime_prev;
3913   PetscFunctionReturn(0);
3914 }
3915 
3916 /*@
3917    TSSetTime - Allows one to reset the time.
3918 
3919    Logically Collective on TS
3920 
3921    Input Parameters:
3922 +  ts - the TS context obtained from TSCreate()
3923 -  time - the time
3924 
3925    Level: intermediate
3926 
3927 .seealso: `TSGetTime()`, `TSSetMaxSteps()`
3928 
3929 @*/
3930 PetscErrorCode  TSSetTime(TS ts, PetscReal t)
3931 {
3932   PetscFunctionBegin;
3933   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3934   PetscValidLogicalCollectiveReal(ts,t,2);
3935   ts->ptime = t;
3936   PetscFunctionReturn(0);
3937 }
3938 
3939 /*@C
3940    TSSetOptionsPrefix - Sets the prefix used for searching for all
3941    TS options in the database.
3942 
3943    Logically Collective on TS
3944 
3945    Input Parameters:
3946 +  ts     - The TS context
3947 -  prefix - The prefix to prepend to all option names
3948 
3949    Notes:
3950    A hyphen (-) must NOT be given at the beginning of the prefix name.
3951    The first character of all runtime options is AUTOMATICALLY the
3952    hyphen.
3953 
3954    Level: advanced
3955 
3956 .seealso: `TSSetFromOptions()`
3957 
3958 @*/
3959 PetscErrorCode  TSSetOptionsPrefix(TS ts,const char prefix[])
3960 {
3961   SNES           snes;
3962 
3963   PetscFunctionBegin;
3964   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3965   PetscCall(PetscObjectSetOptionsPrefix((PetscObject)ts,prefix));
3966   PetscCall(TSGetSNES(ts,&snes));
3967   PetscCall(SNESSetOptionsPrefix(snes,prefix));
3968   PetscFunctionReturn(0);
3969 }
3970 
3971 /*@C
3972    TSAppendOptionsPrefix - Appends to the prefix used for searching for all
3973    TS options in the database.
3974 
3975    Logically Collective on TS
3976 
3977    Input Parameters:
3978 +  ts     - The TS context
3979 -  prefix - The prefix to prepend to all option names
3980 
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.
3985 
3986    Level: advanced
3987 
3988 .seealso: `TSGetOptionsPrefix()`
3989 
3990 @*/
3991 PetscErrorCode  TSAppendOptionsPrefix(TS ts,const char prefix[])
3992 {
3993   SNES           snes;
3994 
3995   PetscFunctionBegin;
3996   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
3997   PetscCall(PetscObjectAppendOptionsPrefix((PetscObject)ts,prefix));
3998   PetscCall(TSGetSNES(ts,&snes));
3999   PetscCall(SNESAppendOptionsPrefix(snes,prefix));
4000   PetscFunctionReturn(0);
4001 }
4002 
4003 /*@C
4004    TSGetOptionsPrefix - Sets the prefix used for searching for all
4005    TS options in the database.
4006 
4007    Not Collective
4008 
4009    Input Parameter:
4010 .  ts - The TS context
4011 
4012    Output Parameter:
4013 .  prefix - A pointer to the prefix string used
4014 
4015    Notes:
4016     On the fortran side, the user should pass in a string 'prifix' of
4017    sufficient length to hold the prefix.
4018 
4019    Level: intermediate
4020 
4021 .seealso: `TSAppendOptionsPrefix()`
4022 @*/
4023 PetscErrorCode  TSGetOptionsPrefix(TS ts,const char *prefix[])
4024 {
4025   PetscFunctionBegin;
4026   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4027   PetscValidPointer(prefix,2);
4028   PetscCall(PetscObjectGetOptionsPrefix((PetscObject)ts,prefix));
4029   PetscFunctionReturn(0);
4030 }
4031 
4032 /*@C
4033    TSGetRHSJacobian - Returns the Jacobian J at the present timestep.
4034 
4035    Not Collective, but parallel objects are returned if TS is parallel
4036 
4037    Input Parameter:
4038 .  ts  - The TS context obtained from TSCreate()
4039 
4040    Output Parameters:
4041 +  Amat - The (approximate) Jacobian J of G, where U_t = G(U,t)  (or NULL)
4042 .  Pmat - The matrix from which the preconditioner is constructed, usually the same as Amat  (or NULL)
4043 .  func - Function to compute the Jacobian of the RHS  (or NULL)
4044 -  ctx - User-defined context for Jacobian evaluation routine  (or NULL)
4045 
4046    Notes:
4047     You can pass in NULL for any return argument you do not need.
4048 
4049    Level: intermediate
4050 
4051 .seealso: `TSGetTimeStep()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`
4052 
4053 @*/
4054 PetscErrorCode  TSGetRHSJacobian(TS ts,Mat *Amat,Mat *Pmat,TSRHSJacobian *func,void **ctx)
4055 {
4056   DM             dm;
4057 
4058   PetscFunctionBegin;
4059   if (Amat || Pmat) {
4060     SNES snes;
4061     PetscCall(TSGetSNES(ts,&snes));
4062     PetscCall(SNESSetUpMatrices(snes));
4063     PetscCall(SNESGetJacobian(snes,Amat,Pmat,NULL,NULL));
4064   }
4065   PetscCall(TSGetDM(ts,&dm));
4066   PetscCall(DMTSGetRHSJacobian(dm,func,ctx));
4067   PetscFunctionReturn(0);
4068 }
4069 
4070 /*@C
4071    TSGetIJacobian - Returns the implicit Jacobian at the present timestep.
4072 
4073    Not Collective, but parallel objects are returned if TS is parallel
4074 
4075    Input Parameter:
4076 .  ts  - The TS context obtained from TSCreate()
4077 
4078    Output Parameters:
4079 +  Amat  - The (approximate) Jacobian of F(t,U,U_t)
4080 .  Pmat - The matrix from which the preconditioner is constructed, often the same as Amat
4081 .  f   - The function to compute the matrices
4082 - ctx - User-defined context for Jacobian evaluation routine
4083 
4084    Notes:
4085     You can pass in NULL for any return argument you do not need.
4086 
4087    Level: advanced
4088 
4089 .seealso: `TSGetTimeStep()`, `TSGetRHSJacobian()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`
4090 
4091 @*/
4092 PetscErrorCode  TSGetIJacobian(TS ts,Mat *Amat,Mat *Pmat,TSIJacobian *f,void **ctx)
4093 {
4094   DM             dm;
4095 
4096   PetscFunctionBegin;
4097   if (Amat || Pmat) {
4098     SNES snes;
4099     PetscCall(TSGetSNES(ts,&snes));
4100     PetscCall(SNESSetUpMatrices(snes));
4101     PetscCall(SNESGetJacobian(snes,Amat,Pmat,NULL,NULL));
4102   }
4103   PetscCall(TSGetDM(ts,&dm));
4104   PetscCall(DMTSGetIJacobian(dm,f,ctx));
4105   PetscFunctionReturn(0);
4106 }
4107 
4108 #include <petsc/private/dmimpl.h>
4109 /*@
4110    TSSetDM - Sets the DM that may be used by some nonlinear solvers or preconditioners under the TS
4111 
4112    Logically Collective on ts
4113 
4114    Input Parameters:
4115 +  ts - the ODE integrator object
4116 -  dm - the dm, cannot be NULL
4117 
4118    Notes:
4119    A DM can only be used for solving one problem at a time because information about the problem is stored on the DM,
4120    even when not using interfaces like DMTSSetIFunction().  Use DMClone() to get a distinct DM when solving
4121    different problems using the same function space.
4122 
4123    Level: intermediate
4124 
4125 .seealso: `TSGetDM()`, `SNESSetDM()`, `SNESGetDM()`
4126 @*/
4127 PetscErrorCode  TSSetDM(TS ts,DM dm)
4128 {
4129   SNES           snes;
4130   DMTS           tsdm;
4131 
4132   PetscFunctionBegin;
4133   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4134   PetscValidHeaderSpecific(dm,DM_CLASSID,2);
4135   PetscCall(PetscObjectReference((PetscObject)dm));
4136   if (ts->dm) {               /* Move the DMTS context over to the new DM unless the new DM already has one */
4137     if (ts->dm->dmts && !dm->dmts) {
4138       PetscCall(DMCopyDMTS(ts->dm,dm));
4139       PetscCall(DMGetDMTS(ts->dm,&tsdm));
4140       /* Grant write privileges to the replacement DM */
4141       if (tsdm->originaldm == ts->dm) tsdm->originaldm = dm;
4142     }
4143     PetscCall(DMDestroy(&ts->dm));
4144   }
4145   ts->dm = dm;
4146 
4147   PetscCall(TSGetSNES(ts,&snes));
4148   PetscCall(SNESSetDM(snes,dm));
4149   PetscFunctionReturn(0);
4150 }
4151 
4152 /*@
4153    TSGetDM - Gets the DM that may be used by some preconditioners
4154 
4155    Not Collective
4156 
4157    Input Parameter:
4158 . ts - the preconditioner context
4159 
4160    Output Parameter:
4161 .  dm - the dm
4162 
4163    Level: intermediate
4164 
4165 .seealso: `TSSetDM()`, `SNESSetDM()`, `SNESGetDM()`
4166 @*/
4167 PetscErrorCode  TSGetDM(TS ts,DM *dm)
4168 {
4169   PetscFunctionBegin;
4170   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4171   if (!ts->dm) {
4172     PetscCall(DMShellCreate(PetscObjectComm((PetscObject)ts),&ts->dm));
4173     if (ts->snes) PetscCall(SNESSetDM(ts->snes,ts->dm));
4174   }
4175   *dm = ts->dm;
4176   PetscFunctionReturn(0);
4177 }
4178 
4179 /*@
4180    SNESTSFormFunction - Function to evaluate nonlinear residual
4181 
4182    Logically Collective on SNES
4183 
4184    Input Parameters:
4185 + snes - nonlinear solver
4186 . U - the current state at which to evaluate the residual
4187 - ctx - user context, must be a TS
4188 
4189    Output Parameter:
4190 . F - the nonlinear residual
4191 
4192    Notes:
4193    This function is not normally called by users and is automatically registered with the SNES used by TS.
4194    It is most frequently passed to MatFDColoringSetFunction().
4195 
4196    Level: advanced
4197 
4198 .seealso: `SNESSetFunction()`, `MatFDColoringSetFunction()`
4199 @*/
4200 PetscErrorCode  SNESTSFormFunction(SNES snes,Vec U,Vec F,void *ctx)
4201 {
4202   TS             ts = (TS)ctx;
4203 
4204   PetscFunctionBegin;
4205   PetscValidHeaderSpecific(snes,SNES_CLASSID,1);
4206   PetscValidHeaderSpecific(U,VEC_CLASSID,2);
4207   PetscValidHeaderSpecific(F,VEC_CLASSID,3);
4208   PetscValidHeaderSpecific(ts,TS_CLASSID,4);
4209   PetscCall((ts->ops->snesfunction)(snes,U,F,ts));
4210   PetscFunctionReturn(0);
4211 }
4212 
4213 /*@
4214    SNESTSFormJacobian - Function to evaluate the Jacobian
4215 
4216    Collective on SNES
4217 
4218    Input Parameters:
4219 + snes - nonlinear solver
4220 . U - the current state at which to evaluate the residual
4221 - ctx - user context, must be a TS
4222 
4223    Output Parameters:
4224 + A - the Jacobian
4225 - B - the preconditioning matrix (may be the same as A)
4226 
4227    Notes:
4228    This function is not normally called by users and is automatically registered with the SNES used by TS.
4229 
4230    Level: developer
4231 
4232 .seealso: `SNESSetJacobian()`
4233 @*/
4234 PetscErrorCode  SNESTSFormJacobian(SNES snes,Vec U,Mat A,Mat B,void *ctx)
4235 {
4236   TS             ts = (TS)ctx;
4237 
4238   PetscFunctionBegin;
4239   PetscValidHeaderSpecific(snes,SNES_CLASSID,1);
4240   PetscValidHeaderSpecific(U,VEC_CLASSID,2);
4241   PetscValidPointer(A,3);
4242   PetscValidHeaderSpecific(A,MAT_CLASSID,3);
4243   PetscValidPointer(B,4);
4244   PetscValidHeaderSpecific(B,MAT_CLASSID,4);
4245   PetscValidHeaderSpecific(ts,TS_CLASSID,5);
4246   PetscCall((ts->ops->snesjacobian)(snes,U,A,B,ts));
4247   PetscFunctionReturn(0);
4248 }
4249 
4250 /*@C
4251    TSComputeRHSFunctionLinear - Evaluate the right hand side via the user-provided Jacobian, for linear problems Udot = A U only
4252 
4253    Collective on TS
4254 
4255    Input Parameters:
4256 +  ts - time stepping context
4257 .  t - time at which to evaluate
4258 .  U - state at which to evaluate
4259 -  ctx - context
4260 
4261    Output Parameter:
4262 .  F - right hand side
4263 
4264    Level: intermediate
4265 
4266    Notes:
4267    This function is intended to be passed to TSSetRHSFunction() to evaluate the right hand side for linear problems.
4268    The matrix (and optionally the evaluation context) should be passed to TSSetRHSJacobian().
4269 
4270 .seealso: `TSSetRHSFunction()`, `TSSetRHSJacobian()`, `TSComputeRHSJacobianConstant()`
4271 @*/
4272 PetscErrorCode TSComputeRHSFunctionLinear(TS ts,PetscReal t,Vec U,Vec F,void *ctx)
4273 {
4274   Mat            Arhs,Brhs;
4275 
4276   PetscFunctionBegin;
4277   PetscCall(TSGetRHSMats_Private(ts,&Arhs,&Brhs));
4278   /* undo the damage caused by shifting */
4279   PetscCall(TSRecoverRHSJacobian(ts,Arhs,Brhs));
4280   PetscCall(TSComputeRHSJacobian(ts,t,U,Arhs,Brhs));
4281   PetscCall(MatMult(Arhs,U,F));
4282   PetscFunctionReturn(0);
4283 }
4284 
4285 /*@C
4286    TSComputeRHSJacobianConstant - Reuses a Jacobian that is time-independent.
4287 
4288    Collective on TS
4289 
4290    Input Parameters:
4291 +  ts - time stepping context
4292 .  t - time at which to evaluate
4293 .  U - state at which to evaluate
4294 -  ctx - context
4295 
4296    Output Parameters:
4297 +  A - pointer to operator
4298 -  B - pointer to preconditioning matrix
4299 
4300    Level: intermediate
4301 
4302    Notes:
4303    This function is intended to be passed to TSSetRHSJacobian() to evaluate the Jacobian for linear time-independent problems.
4304 
4305 .seealso: `TSSetRHSFunction()`, `TSSetRHSJacobian()`, `TSComputeRHSFunctionLinear()`
4306 @*/
4307 PetscErrorCode TSComputeRHSJacobianConstant(TS ts,PetscReal t,Vec U,Mat A,Mat B,void *ctx)
4308 {
4309   PetscFunctionBegin;
4310   PetscFunctionReturn(0);
4311 }
4312 
4313 /*@C
4314    TSComputeIFunctionLinear - Evaluate the left hand side via the user-provided Jacobian, for linear problems only
4315 
4316    Collective on TS
4317 
4318    Input Parameters:
4319 +  ts - time stepping context
4320 .  t - time at which to evaluate
4321 .  U - state at which to evaluate
4322 .  Udot - time derivative of state vector
4323 -  ctx - context
4324 
4325    Output Parameter:
4326 .  F - left hand side
4327 
4328    Level: intermediate
4329 
4330    Notes:
4331    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
4332    user is required to write their own TSComputeIFunction.
4333    This function is intended to be passed to TSSetIFunction() to evaluate the left hand side for linear problems.
4334    The matrix (and optionally the evaluation context) should be passed to TSSetIJacobian().
4335 
4336    Note that using this function is NOT equivalent to using TSComputeRHSFunctionLinear() since that solves Udot = A U
4337 
4338 .seealso: `TSSetIFunction()`, `TSSetIJacobian()`, `TSComputeIJacobianConstant()`, `TSComputeRHSFunctionLinear()`
4339 @*/
4340 PetscErrorCode TSComputeIFunctionLinear(TS ts,PetscReal t,Vec U,Vec Udot,Vec F,void *ctx)
4341 {
4342   Mat            A,B;
4343 
4344   PetscFunctionBegin;
4345   PetscCall(TSGetIJacobian(ts,&A,&B,NULL,NULL));
4346   PetscCall(TSComputeIJacobian(ts,t,U,Udot,1.0,A,B,PETSC_TRUE));
4347   PetscCall(MatMult(A,Udot,F));
4348   PetscFunctionReturn(0);
4349 }
4350 
4351 /*@C
4352    TSComputeIJacobianConstant - Reuses a time-independent for a semi-implicit DAE or ODE
4353 
4354    Collective on TS
4355 
4356    Input Parameters:
4357 +  ts - time stepping context
4358 .  t - time at which to evaluate
4359 .  U - state at which to evaluate
4360 .  Udot - time derivative of state vector
4361 .  shift - shift to apply
4362 -  ctx - context
4363 
4364    Output Parameters:
4365 +  A - pointer to operator
4366 -  B - pointer to preconditioning matrix
4367 
4368    Level: advanced
4369 
4370    Notes:
4371    This function is intended to be passed to TSSetIJacobian() to evaluate the Jacobian for linear time-independent problems.
4372 
4373    It is only appropriate for problems of the form
4374 
4375 $     M Udot = F(U,t)
4376 
4377   where M is constant and F is non-stiff.  The user must pass M to TSSetIJacobian().  The current implementation only
4378   works with IMEX time integration methods such as TSROSW and TSARKIMEX, since there is no support for de-constructing
4379   an implicit operator of the form
4380 
4381 $    shift*M + J
4382 
4383   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
4384   a copy of M or reassemble it when requested.
4385 
4386 .seealso: `TSSetIFunction()`, `TSSetIJacobian()`, `TSComputeIFunctionLinear()`
4387 @*/
4388 PetscErrorCode TSComputeIJacobianConstant(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,void *ctx)
4389 {
4390   PetscFunctionBegin;
4391   PetscCall(MatScale(A, shift / ts->ijacobian.shift));
4392   ts->ijacobian.shift = shift;
4393   PetscFunctionReturn(0);
4394 }
4395 
4396 /*@
4397    TSGetEquationType - Gets the type of the equation that TS is solving.
4398 
4399    Not Collective
4400 
4401    Input Parameter:
4402 .  ts - the TS context
4403 
4404    Output Parameter:
4405 .  equation_type - see TSEquationType
4406 
4407    Level: beginner
4408 
4409 .seealso: `TSSetEquationType()`, `TSEquationType`
4410 @*/
4411 PetscErrorCode  TSGetEquationType(TS ts,TSEquationType *equation_type)
4412 {
4413   PetscFunctionBegin;
4414   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4415   PetscValidPointer(equation_type,2);
4416   *equation_type = ts->equation_type;
4417   PetscFunctionReturn(0);
4418 }
4419 
4420 /*@
4421    TSSetEquationType - Sets the type of the equation that TS is solving.
4422 
4423    Not Collective
4424 
4425    Input Parameters:
4426 +  ts - the TS context
4427 -  equation_type - see TSEquationType
4428 
4429    Level: advanced
4430 
4431 .seealso: `TSGetEquationType()`, `TSEquationType`
4432 @*/
4433 PetscErrorCode  TSSetEquationType(TS ts,TSEquationType equation_type)
4434 {
4435   PetscFunctionBegin;
4436   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4437   ts->equation_type = equation_type;
4438   PetscFunctionReturn(0);
4439 }
4440 
4441 /*@
4442    TSGetConvergedReason - Gets the reason the TS iteration was stopped.
4443 
4444    Not Collective
4445 
4446    Input Parameter:
4447 .  ts - the TS context
4448 
4449    Output Parameter:
4450 .  reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the
4451             manual pages for the individual convergence tests for complete lists
4452 
4453    Level: beginner
4454 
4455    Notes:
4456    Can only be called after the call to TSSolve() is complete.
4457 
4458 .seealso: `TSSetConvergenceTest()`, `TSConvergedReason`
4459 @*/
4460 PetscErrorCode  TSGetConvergedReason(TS ts,TSConvergedReason *reason)
4461 {
4462   PetscFunctionBegin;
4463   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4464   PetscValidPointer(reason,2);
4465   *reason = ts->reason;
4466   PetscFunctionReturn(0);
4467 }
4468 
4469 /*@
4470    TSSetConvergedReason - Sets the reason for handling the convergence of TSSolve.
4471 
4472    Logically Collective; reason must contain common value
4473 
4474    Input Parameters:
4475 +  ts - the TS context
4476 -  reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the
4477             manual pages for the individual convergence tests for complete lists
4478 
4479    Level: advanced
4480 
4481    Notes:
4482    Can only be called while TSSolve() is active.
4483 
4484 .seealso: `TSConvergedReason`
4485 @*/
4486 PetscErrorCode  TSSetConvergedReason(TS ts,TSConvergedReason reason)
4487 {
4488   PetscFunctionBegin;
4489   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4490   ts->reason = reason;
4491   PetscFunctionReturn(0);
4492 }
4493 
4494 /*@
4495    TSGetSolveTime - Gets the time after a call to TSSolve()
4496 
4497    Not Collective
4498 
4499    Input Parameter:
4500 .  ts - the TS context
4501 
4502    Output Parameter:
4503 .  ftime - the final time. This time corresponds to the final time set with TSSetMaxTime()
4504 
4505    Level: beginner
4506 
4507    Notes:
4508    Can only be called after the call to TSSolve() is complete.
4509 
4510 .seealso: `TSSetConvergenceTest()`, `TSConvergedReason`
4511 @*/
4512 PetscErrorCode  TSGetSolveTime(TS ts,PetscReal *ftime)
4513 {
4514   PetscFunctionBegin;
4515   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4516   PetscValidRealPointer(ftime,2);
4517   *ftime = ts->solvetime;
4518   PetscFunctionReturn(0);
4519 }
4520 
4521 /*@
4522    TSGetSNESIterations - Gets the total number of nonlinear iterations
4523    used by the time integrator.
4524 
4525    Not Collective
4526 
4527    Input Parameter:
4528 .  ts - TS context
4529 
4530    Output Parameter:
4531 .  nits - number of nonlinear iterations
4532 
4533    Notes:
4534    This counter is reset to zero for each successive call to TSSolve().
4535 
4536    Level: intermediate
4537 
4538 .seealso: `TSGetKSPIterations()`
4539 @*/
4540 PetscErrorCode TSGetSNESIterations(TS ts,PetscInt *nits)
4541 {
4542   PetscFunctionBegin;
4543   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4544   PetscValidIntPointer(nits,2);
4545   *nits = ts->snes_its;
4546   PetscFunctionReturn(0);
4547 }
4548 
4549 /*@
4550    TSGetKSPIterations - Gets the total number of linear iterations
4551    used by the time integrator.
4552 
4553    Not Collective
4554 
4555    Input Parameter:
4556 .  ts - TS context
4557 
4558    Output Parameter:
4559 .  lits - number of linear iterations
4560 
4561    Notes:
4562    This counter is reset to zero for each successive call to TSSolve().
4563 
4564    Level: intermediate
4565 
4566 .seealso: `TSGetSNESIterations()`, `SNESGetKSPIterations()`
4567 @*/
4568 PetscErrorCode TSGetKSPIterations(TS ts,PetscInt *lits)
4569 {
4570   PetscFunctionBegin;
4571   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4572   PetscValidIntPointer(lits,2);
4573   *lits = ts->ksp_its;
4574   PetscFunctionReturn(0);
4575 }
4576 
4577 /*@
4578    TSGetStepRejections - Gets the total number of rejected steps.
4579 
4580    Not Collective
4581 
4582    Input Parameter:
4583 .  ts - TS context
4584 
4585    Output Parameter:
4586 .  rejects - number of steps rejected
4587 
4588    Notes:
4589    This counter is reset to zero for each successive call to TSSolve().
4590 
4591    Level: intermediate
4592 
4593 .seealso: `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetSNESFailures()`, `TSSetMaxSNESFailures()`, `TSSetErrorIfStepFails()`
4594 @*/
4595 PetscErrorCode TSGetStepRejections(TS ts,PetscInt *rejects)
4596 {
4597   PetscFunctionBegin;
4598   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4599   PetscValidIntPointer(rejects,2);
4600   *rejects = ts->reject;
4601   PetscFunctionReturn(0);
4602 }
4603 
4604 /*@
4605    TSGetSNESFailures - Gets the total number of failed SNES solves
4606 
4607    Not Collective
4608 
4609    Input Parameter:
4610 .  ts - TS context
4611 
4612    Output Parameter:
4613 .  fails - number of failed nonlinear solves
4614 
4615    Notes:
4616    This counter is reset to zero for each successive call to TSSolve().
4617 
4618    Level: intermediate
4619 
4620 .seealso: `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSSetMaxSNESFailures()`
4621 @*/
4622 PetscErrorCode TSGetSNESFailures(TS ts,PetscInt *fails)
4623 {
4624   PetscFunctionBegin;
4625   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4626   PetscValidIntPointer(fails,2);
4627   *fails = ts->num_snes_failures;
4628   PetscFunctionReturn(0);
4629 }
4630 
4631 /*@
4632    TSSetMaxStepRejections - Sets the maximum number of step rejections before a step fails
4633 
4634    Not Collective
4635 
4636    Input Parameters:
4637 +  ts - TS context
4638 -  rejects - maximum number of rejected steps, pass -1 for unlimited
4639 
4640    Notes:
4641    The counter is reset to zero for each step
4642 
4643    Options Database Key:
4644 .  -ts_max_reject - Maximum number of step rejections before a step fails
4645 
4646    Level: intermediate
4647 
4648 .seealso: `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxSNESFailures()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `TSSetErrorIfStepFails()`, `TSGetConvergedReason()`
4649 @*/
4650 PetscErrorCode TSSetMaxStepRejections(TS ts,PetscInt rejects)
4651 {
4652   PetscFunctionBegin;
4653   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4654   ts->max_reject = rejects;
4655   PetscFunctionReturn(0);
4656 }
4657 
4658 /*@
4659    TSSetMaxSNESFailures - Sets the maximum number of failed SNES solves
4660 
4661    Not Collective
4662 
4663    Input Parameters:
4664 +  ts - TS context
4665 -  fails - maximum number of failed nonlinear solves, pass -1 for unlimited
4666 
4667    Notes:
4668    The counter is reset to zero for each successive call to TSSolve().
4669 
4670    Options Database Key:
4671 .  -ts_max_snes_failures - Maximum number of nonlinear solve failures
4672 
4673    Level: intermediate
4674 
4675 .seealso: `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `SNESGetConvergedReason()`, `TSGetConvergedReason()`
4676 @*/
4677 PetscErrorCode TSSetMaxSNESFailures(TS ts,PetscInt fails)
4678 {
4679   PetscFunctionBegin;
4680   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4681   ts->max_snes_failures = fails;
4682   PetscFunctionReturn(0);
4683 }
4684 
4685 /*@
4686    TSSetErrorIfStepFails - Error if no step succeeds
4687 
4688    Not Collective
4689 
4690    Input Parameters:
4691 +  ts - TS context
4692 -  err - PETSC_TRUE to error if no step succeeds, PETSC_FALSE to return without failure
4693 
4694    Options Database Key:
4695 .  -ts_error_if_step_fails - Error if no step succeeds
4696 
4697    Level: intermediate
4698 
4699 .seealso: `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `TSSetErrorIfStepFails()`, `TSGetConvergedReason()`
4700 @*/
4701 PetscErrorCode TSSetErrorIfStepFails(TS ts,PetscBool err)
4702 {
4703   PetscFunctionBegin;
4704   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4705   ts->errorifstepfailed = err;
4706   PetscFunctionReturn(0);
4707 }
4708 
4709 /*@
4710    TSGetAdapt - Get the adaptive controller context for the current method
4711 
4712    Collective on TS if controller has not been created yet
4713 
4714    Input Parameter:
4715 .  ts - time stepping context
4716 
4717    Output Parameter:
4718 .  adapt - adaptive controller
4719 
4720    Level: intermediate
4721 
4722 .seealso: `TSAdapt`, `TSAdaptSetType()`, `TSAdaptChoose()`
4723 @*/
4724 PetscErrorCode TSGetAdapt(TS ts,TSAdapt *adapt)
4725 {
4726   PetscFunctionBegin;
4727   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4728   PetscValidPointer(adapt,2);
4729   if (!ts->adapt) {
4730     PetscCall(TSAdaptCreate(PetscObjectComm((PetscObject)ts),&ts->adapt));
4731     PetscCall(PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->adapt));
4732     PetscCall(PetscObjectIncrementTabLevel((PetscObject)ts->adapt,(PetscObject)ts,1));
4733   }
4734   *adapt = ts->adapt;
4735   PetscFunctionReturn(0);
4736 }
4737 
4738 /*@
4739    TSSetTolerances - Set tolerances for local truncation error when using adaptive controller
4740 
4741    Logically Collective
4742 
4743    Input Parameters:
4744 +  ts - time integration context
4745 .  atol - scalar absolute tolerances, PETSC_DECIDE to leave current value
4746 .  vatol - vector of absolute tolerances or NULL, used in preference to atol if present
4747 .  rtol - scalar relative tolerances, PETSC_DECIDE to leave current value
4748 -  vrtol - vector of relative tolerances or NULL, used in preference to atol if present
4749 
4750    Options Database keys:
4751 +  -ts_rtol <rtol> - relative tolerance for local truncation error
4752 -  -ts_atol <atol> - Absolute tolerance for local truncation error
4753 
4754    Notes:
4755    With PETSc's implicit schemes for DAE problems, the calculation of the local truncation error
4756    (LTE) includes both the differential and the algebraic variables. If one wants the LTE to be
4757    computed only for the differential or the algebraic part then this can be done using the vector of
4758    tolerances vatol. For example, by setting the tolerance vector with the desired tolerance for the
4759    differential part and infinity for the algebraic part, the LTE calculation will include only the
4760    differential variables.
4761 
4762    Level: beginner
4763 
4764 .seealso: `TS`, `TSAdapt`, `TSErrorWeightedNorm()`, `TSGetTolerances()`
4765 @*/
4766 PetscErrorCode TSSetTolerances(TS ts,PetscReal atol,Vec vatol,PetscReal rtol,Vec vrtol)
4767 {
4768   PetscFunctionBegin;
4769   if (atol != PETSC_DECIDE && atol != PETSC_DEFAULT) ts->atol = atol;
4770   if (vatol) {
4771     PetscCall(PetscObjectReference((PetscObject)vatol));
4772     PetscCall(VecDestroy(&ts->vatol));
4773     ts->vatol = vatol;
4774   }
4775   if (rtol != PETSC_DECIDE && rtol != PETSC_DEFAULT) ts->rtol = rtol;
4776   if (vrtol) {
4777     PetscCall(PetscObjectReference((PetscObject)vrtol));
4778     PetscCall(VecDestroy(&ts->vrtol));
4779     ts->vrtol = vrtol;
4780   }
4781   PetscFunctionReturn(0);
4782 }
4783 
4784 /*@
4785    TSGetTolerances - Get tolerances for local truncation error when using adaptive controller
4786 
4787    Logically Collective
4788 
4789    Input Parameter:
4790 .  ts - time integration context
4791 
4792    Output Parameters:
4793 +  atol - scalar absolute tolerances, NULL to ignore
4794 .  vatol - vector of absolute tolerances, NULL to ignore
4795 .  rtol - scalar relative tolerances, NULL to ignore
4796 -  vrtol - vector of relative tolerances, NULL to ignore
4797 
4798    Level: beginner
4799 
4800 .seealso: `TS`, `TSAdapt`, `TSErrorWeightedNorm()`, `TSSetTolerances()`
4801 @*/
4802 PetscErrorCode TSGetTolerances(TS ts,PetscReal *atol,Vec *vatol,PetscReal *rtol,Vec *vrtol)
4803 {
4804   PetscFunctionBegin;
4805   if (atol)  *atol  = ts->atol;
4806   if (vatol) *vatol = ts->vatol;
4807   if (rtol)  *rtol  = ts->rtol;
4808   if (vrtol) *vrtol = ts->vrtol;
4809   PetscFunctionReturn(0);
4810 }
4811 
4812 /*@
4813    TSErrorWeightedNorm2 - compute a weighted 2-norm of the difference between two state vectors
4814 
4815    Collective on TS
4816 
4817    Input Parameters:
4818 +  ts - time stepping context
4819 .  U - state vector, usually ts->vec_sol
4820 -  Y - state vector to be compared to U
4821 
4822    Output Parameters:
4823 +  norm - weighted norm, a value of 1.0 means that the error matches the tolerances
4824 .  norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
4825 -  normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
4826 
4827    Level: developer
4828 
4829 .seealso: `TSErrorWeightedNorm()`, `TSErrorWeightedNormInfinity()`
4830 @*/
4831 PetscErrorCode TSErrorWeightedNorm2(TS ts,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
4832 {
4833   PetscInt          i,n,N,rstart;
4834   PetscInt          n_loc,na_loc,nr_loc;
4835   PetscReal         n_glb,na_glb,nr_glb;
4836   const PetscScalar *u,*y;
4837   PetscReal         sum,suma,sumr,gsum,gsuma,gsumr,diff;
4838   PetscReal         tol,tola,tolr;
4839   PetscReal         err_loc[6],err_glb[6];
4840 
4841   PetscFunctionBegin;
4842   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
4843   PetscValidHeaderSpecific(U,VEC_CLASSID,2);
4844   PetscValidHeaderSpecific(Y,VEC_CLASSID,3);
4845   PetscValidType(U,2);
4846   PetscValidType(Y,3);
4847   PetscCheckSameComm(U,2,Y,3);
4848   PetscValidRealPointer(norm,4);
4849   PetscValidRealPointer(norma,5);
4850   PetscValidRealPointer(normr,6);
4851   PetscCheck(U != Y,PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_IDN,"U and Y cannot be the same vector");
4852 
4853   PetscCall(VecGetSize(U,&N));
4854   PetscCall(VecGetLocalSize(U,&n));
4855   PetscCall(VecGetOwnershipRange(U,&rstart,NULL));
4856   PetscCall(VecGetArrayRead(U,&u));
4857   PetscCall(VecGetArrayRead(Y,&y));
4858   sum  = 0.; n_loc  = 0;
4859   suma = 0.; na_loc = 0;
4860   sumr = 0.; nr_loc = 0;
4861   if (ts->vatol && ts->vrtol) {
4862     const PetscScalar *atol,*rtol;
4863     PetscCall(VecGetArrayRead(ts->vatol,&atol));
4864     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
4865     for (i=0; i<n; i++) {
4866       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4867       diff = PetscAbsScalar(y[i] - u[i]);
4868       tola = PetscRealPart(atol[i]);
4869       if (tola>0.) {
4870         suma  += PetscSqr(diff/tola);
4871         na_loc++;
4872       }
4873       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4874       if (tolr>0.) {
4875         sumr  += PetscSqr(diff/tolr);
4876         nr_loc++;
4877       }
4878       tol=tola+tolr;
4879       if (tol>0.) {
4880         sum  += PetscSqr(diff/tol);
4881         n_loc++;
4882       }
4883     }
4884     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
4885     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
4886   } else if (ts->vatol) {       /* vector atol, scalar rtol */
4887     const PetscScalar *atol;
4888     PetscCall(VecGetArrayRead(ts->vatol,&atol));
4889     for (i=0; i<n; i++) {
4890       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4891       diff = PetscAbsScalar(y[i] - u[i]);
4892       tola = PetscRealPart(atol[i]);
4893       if (tola>0.) {
4894         suma  += PetscSqr(diff/tola);
4895         na_loc++;
4896       }
4897       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4898       if (tolr>0.) {
4899         sumr  += PetscSqr(diff/tolr);
4900         nr_loc++;
4901       }
4902       tol=tola+tolr;
4903       if (tol>0.) {
4904         sum  += PetscSqr(diff/tol);
4905         n_loc++;
4906       }
4907     }
4908     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
4909   } else if (ts->vrtol) {       /* scalar atol, vector rtol */
4910     const PetscScalar *rtol;
4911     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
4912     for (i=0; i<n; i++) {
4913       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4914       diff = PetscAbsScalar(y[i] - u[i]);
4915       tola = ts->atol;
4916       if (tola>0.) {
4917         suma  += PetscSqr(diff/tola);
4918         na_loc++;
4919       }
4920       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4921       if (tolr>0.) {
4922         sumr  += PetscSqr(diff/tolr);
4923         nr_loc++;
4924       }
4925       tol=tola+tolr;
4926       if (tol>0.) {
4927         sum  += PetscSqr(diff/tol);
4928         n_loc++;
4929       }
4930     }
4931     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
4932   } else {                      /* scalar atol, scalar rtol */
4933     for (i=0; i<n; i++) {
4934       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
4935       diff = PetscAbsScalar(y[i] - u[i]);
4936       tola = ts->atol;
4937       if (tola>0.) {
4938         suma  += PetscSqr(diff/tola);
4939         na_loc++;
4940       }
4941       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
4942       if (tolr>0.) {
4943         sumr  += PetscSqr(diff/tolr);
4944         nr_loc++;
4945       }
4946       tol=tola+tolr;
4947       if (tol>0.) {
4948         sum  += PetscSqr(diff/tol);
4949         n_loc++;
4950       }
4951     }
4952   }
4953   PetscCall(VecRestoreArrayRead(U,&u));
4954   PetscCall(VecRestoreArrayRead(Y,&y));
4955 
4956   err_loc[0] = sum;
4957   err_loc[1] = suma;
4958   err_loc[2] = sumr;
4959   err_loc[3] = (PetscReal)n_loc;
4960   err_loc[4] = (PetscReal)na_loc;
4961   err_loc[5] = (PetscReal)nr_loc;
4962 
4963   PetscCall(MPIU_Allreduce(err_loc,err_glb,6,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)ts)));
4964 
4965   gsum   = err_glb[0];
4966   gsuma  = err_glb[1];
4967   gsumr  = err_glb[2];
4968   n_glb  = err_glb[3];
4969   na_glb = err_glb[4];
4970   nr_glb = err_glb[5];
4971 
4972   *norm  = 0.;
4973   if (n_glb>0.) *norm  = PetscSqrtReal(gsum  / n_glb);
4974   *norma = 0.;
4975   if (na_glb>0.) *norma = PetscSqrtReal(gsuma / na_glb);
4976   *normr = 0.;
4977   if (nr_glb>0.) *normr = PetscSqrtReal(gsumr / nr_glb);
4978 
4979   PetscCheck(!PetscIsInfOrNanScalar(*norm),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm");
4980   PetscCheck(!PetscIsInfOrNanScalar(*norma),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norma");
4981   PetscCheck(!PetscIsInfOrNanScalar(*normr),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in normr");
4982   PetscFunctionReturn(0);
4983 }
4984 
4985 /*@
4986    TSErrorWeightedNormInfinity - compute a weighted infinity-norm of the difference between two state vectors
4987 
4988    Collective on TS
4989 
4990    Input Parameters:
4991 +  ts - time stepping context
4992 .  U - state vector, usually ts->vec_sol
4993 -  Y - state vector to be compared to U
4994 
4995    Output Parameters:
4996 +  norm - weighted norm, a value of 1.0 means that the error matches the tolerances
4997 .  norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
4998 -  normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
4999 
5000    Level: developer
5001 
5002 .seealso: `TSErrorWeightedNorm()`, `TSErrorWeightedNorm2()`
5003 @*/
5004 PetscErrorCode TSErrorWeightedNormInfinity(TS ts,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5005 {
5006   PetscInt          i,n,N,rstart;
5007   const PetscScalar *u,*y;
5008   PetscReal         max,gmax,maxa,gmaxa,maxr,gmaxr;
5009   PetscReal         tol,tola,tolr,diff;
5010   PetscReal         err_loc[3],err_glb[3];
5011 
5012   PetscFunctionBegin;
5013   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5014   PetscValidHeaderSpecific(U,VEC_CLASSID,2);
5015   PetscValidHeaderSpecific(Y,VEC_CLASSID,3);
5016   PetscValidType(U,2);
5017   PetscValidType(Y,3);
5018   PetscCheckSameComm(U,2,Y,3);
5019   PetscValidRealPointer(norm,4);
5020   PetscValidRealPointer(norma,5);
5021   PetscValidRealPointer(normr,6);
5022   PetscCheck(U != Y,PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_IDN,"U and Y cannot be the same vector");
5023 
5024   PetscCall(VecGetSize(U,&N));
5025   PetscCall(VecGetLocalSize(U,&n));
5026   PetscCall(VecGetOwnershipRange(U,&rstart,NULL));
5027   PetscCall(VecGetArrayRead(U,&u));
5028   PetscCall(VecGetArrayRead(Y,&y));
5029 
5030   max=0.;
5031   maxa=0.;
5032   maxr=0.;
5033 
5034   if (ts->vatol && ts->vrtol) {     /* vector atol, vector rtol */
5035     const PetscScalar *atol,*rtol;
5036     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5037     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5038 
5039     for (i=0; i<n; i++) {
5040       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5041       diff = PetscAbsScalar(y[i] - u[i]);
5042       tola = PetscRealPart(atol[i]);
5043       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5044       tol  = tola+tolr;
5045       if (tola>0.) maxa = PetscMax(maxa,diff / tola);
5046       if (tolr>0.) maxr = PetscMax(maxr,diff / tolr);
5047       if (tol>0.)  max = PetscMax(max,diff / tol);
5048     }
5049     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5050     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5051   } else if (ts->vatol) {       /* vector atol, scalar rtol */
5052     const PetscScalar *atol;
5053     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5054     for (i=0; i<n; i++) {
5055       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5056       diff = PetscAbsScalar(y[i] - u[i]);
5057       tola = PetscRealPart(atol[i]);
5058       tolr = ts->rtol  * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5059       tol  = tola+tolr;
5060       if (tola>0.) maxa = PetscMax(maxa,diff / tola);
5061       if (tolr>0.) maxr = PetscMax(maxr,diff / tolr);
5062       if (tol>0.)  max = PetscMax(max,diff / tol);
5063     }
5064     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5065   } else if (ts->vrtol) {       /* scalar atol, vector rtol */
5066     const PetscScalar *rtol;
5067     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5068 
5069     for (i=0; i<n; i++) {
5070       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5071       diff = PetscAbsScalar(y[i] - u[i]);
5072       tola = ts->atol;
5073       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5074       tol  = tola+tolr;
5075       if (tola>0.) maxa = PetscMax(maxa,diff / tola);
5076       if (tolr>0.) maxr = PetscMax(maxr,diff / tolr);
5077       if (tol>0.)  max = PetscMax(max,diff / tol);
5078     }
5079     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5080   } else {                      /* scalar atol, scalar rtol */
5081 
5082     for (i=0; i<n; i++) {
5083       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5084       diff = PetscAbsScalar(y[i] - u[i]);
5085       tola = ts->atol;
5086       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5087       tol  = tola+tolr;
5088       if (tola>0.) maxa = PetscMax(maxa,diff / tola);
5089       if (tolr>0.) maxr = PetscMax(maxr,diff / tolr);
5090       if (tol>0.) max = PetscMax(max,diff / tol);
5091     }
5092   }
5093   PetscCall(VecRestoreArrayRead(U,&u));
5094   PetscCall(VecRestoreArrayRead(Y,&y));
5095   err_loc[0] = max;
5096   err_loc[1] = maxa;
5097   err_loc[2] = maxr;
5098   PetscCall(MPIU_Allreduce(err_loc,err_glb,3,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)ts)));
5099   gmax   = err_glb[0];
5100   gmaxa  = err_glb[1];
5101   gmaxr  = err_glb[2];
5102 
5103   *norm = gmax;
5104   *norma = gmaxa;
5105   *normr = gmaxr;
5106   PetscCheck(!PetscIsInfOrNanScalar(*norm),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm");
5107   PetscCheck(!PetscIsInfOrNanScalar(*norma),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norma");
5108   PetscCheck(!PetscIsInfOrNanScalar(*normr),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in normr");
5109   PetscFunctionReturn(0);
5110 }
5111 
5112 /*@
5113    TSErrorWeightedNorm - compute a weighted norm of the difference between two state vectors based on supplied absolute and relative tolerances
5114 
5115    Collective on TS
5116 
5117    Input Parameters:
5118 +  ts - time stepping context
5119 .  U - state vector, usually ts->vec_sol
5120 .  Y - state vector to be compared to U
5121 -  wnormtype - norm type, either NORM_2 or NORM_INFINITY
5122 
5123    Output Parameters:
5124 +  norm  - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5125 .  norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5126 -  normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5127 
5128    Options Database Keys:
5129 .  -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5130 
5131    Level: developer
5132 
5133 .seealso: `TSErrorWeightedNormInfinity()`, `TSErrorWeightedNorm2()`, `TSErrorWeightedENorm`
5134 @*/
5135 PetscErrorCode TSErrorWeightedNorm(TS ts,Vec U,Vec Y,NormType wnormtype,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5136 {
5137   PetscFunctionBegin;
5138   if (wnormtype == NORM_2) PetscCall(TSErrorWeightedNorm2(ts,U,Y,norm,norma,normr));
5139   else if (wnormtype == NORM_INFINITY) PetscCall(TSErrorWeightedNormInfinity(ts,U,Y,norm,norma,normr));
5140   else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]);
5141   PetscFunctionReturn(0);
5142 }
5143 
5144 /*@
5145    TSErrorWeightedENorm2 - compute a weighted 2 error norm based on supplied absolute and relative tolerances
5146 
5147    Collective on TS
5148 
5149    Input Parameters:
5150 +  ts - time stepping context
5151 .  E - error vector
5152 .  U - state vector, usually ts->vec_sol
5153 -  Y - state vector, previous time step
5154 
5155    Output Parameters:
5156 +  norm - weighted norm, a value of 1.0 means that the error matches the tolerances
5157 .  norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
5158 -  normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
5159 
5160    Level: developer
5161 
5162 .seealso: `TSErrorWeightedENorm()`, `TSErrorWeightedENormInfinity()`
5163 @*/
5164 PetscErrorCode TSErrorWeightedENorm2(TS ts,Vec E,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5165 {
5166   PetscInt          i,n,N,rstart;
5167   PetscInt          n_loc,na_loc,nr_loc;
5168   PetscReal         n_glb,na_glb,nr_glb;
5169   const PetscScalar *e,*u,*y;
5170   PetscReal         err,sum,suma,sumr,gsum,gsuma,gsumr;
5171   PetscReal         tol,tola,tolr;
5172   PetscReal         err_loc[6],err_glb[6];
5173 
5174   PetscFunctionBegin;
5175   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5176   PetscValidHeaderSpecific(E,VEC_CLASSID,2);
5177   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
5178   PetscValidHeaderSpecific(Y,VEC_CLASSID,4);
5179   PetscValidType(E,2);
5180   PetscValidType(U,3);
5181   PetscValidType(Y,4);
5182   PetscCheckSameComm(E,2,U,3);
5183   PetscCheckSameComm(U,3,Y,4);
5184   PetscValidRealPointer(norm,5);
5185   PetscValidRealPointer(norma,6);
5186   PetscValidRealPointer(normr,7);
5187 
5188   PetscCall(VecGetSize(E,&N));
5189   PetscCall(VecGetLocalSize(E,&n));
5190   PetscCall(VecGetOwnershipRange(E,&rstart,NULL));
5191   PetscCall(VecGetArrayRead(E,&e));
5192   PetscCall(VecGetArrayRead(U,&u));
5193   PetscCall(VecGetArrayRead(Y,&y));
5194   sum  = 0.; n_loc  = 0;
5195   suma = 0.; na_loc = 0;
5196   sumr = 0.; nr_loc = 0;
5197   if (ts->vatol && ts->vrtol) {
5198     const PetscScalar *atol,*rtol;
5199     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5200     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5201     for (i=0; i<n; i++) {
5202       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5203       err = PetscAbsScalar(e[i]);
5204       tola = PetscRealPart(atol[i]);
5205       if (tola>0.) {
5206         suma  += PetscSqr(err/tola);
5207         na_loc++;
5208       }
5209       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5210       if (tolr>0.) {
5211         sumr  += PetscSqr(err/tolr);
5212         nr_loc++;
5213       }
5214       tol=tola+tolr;
5215       if (tol>0.) {
5216         sum  += PetscSqr(err/tol);
5217         n_loc++;
5218       }
5219     }
5220     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5221     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5222   } else if (ts->vatol) {       /* vector atol, scalar rtol */
5223     const PetscScalar *atol;
5224     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5225     for (i=0; i<n; i++) {
5226       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5227       err = PetscAbsScalar(e[i]);
5228       tola = PetscRealPart(atol[i]);
5229       if (tola>0.) {
5230         suma  += PetscSqr(err/tola);
5231         na_loc++;
5232       }
5233       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5234       if (tolr>0.) {
5235         sumr  += PetscSqr(err/tolr);
5236         nr_loc++;
5237       }
5238       tol=tola+tolr;
5239       if (tol>0.) {
5240         sum  += PetscSqr(err/tol);
5241         n_loc++;
5242       }
5243     }
5244     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5245   } else if (ts->vrtol) {       /* scalar atol, vector rtol */
5246     const PetscScalar *rtol;
5247     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5248     for (i=0; i<n; i++) {
5249       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5250       err = PetscAbsScalar(e[i]);
5251       tola = ts->atol;
5252       if (tola>0.) {
5253         suma  += PetscSqr(err/tola);
5254         na_loc++;
5255       }
5256       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5257       if (tolr>0.) {
5258         sumr  += PetscSqr(err/tolr);
5259         nr_loc++;
5260       }
5261       tol=tola+tolr;
5262       if (tol>0.) {
5263         sum  += PetscSqr(err/tol);
5264         n_loc++;
5265       }
5266     }
5267     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5268   } else {                      /* scalar atol, scalar rtol */
5269     for (i=0; i<n; i++) {
5270       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5271       err = PetscAbsScalar(e[i]);
5272       tola = ts->atol;
5273       if (tola>0.) {
5274         suma  += PetscSqr(err/tola);
5275         na_loc++;
5276       }
5277       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5278       if (tolr>0.) {
5279         sumr  += PetscSqr(err/tolr);
5280         nr_loc++;
5281       }
5282       tol=tola+tolr;
5283       if (tol>0.) {
5284         sum  += PetscSqr(err/tol);
5285         n_loc++;
5286       }
5287     }
5288   }
5289   PetscCall(VecRestoreArrayRead(E,&e));
5290   PetscCall(VecRestoreArrayRead(U,&u));
5291   PetscCall(VecRestoreArrayRead(Y,&y));
5292 
5293   err_loc[0] = sum;
5294   err_loc[1] = suma;
5295   err_loc[2] = sumr;
5296   err_loc[3] = (PetscReal)n_loc;
5297   err_loc[4] = (PetscReal)na_loc;
5298   err_loc[5] = (PetscReal)nr_loc;
5299 
5300   PetscCall(MPIU_Allreduce(err_loc,err_glb,6,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)ts)));
5301 
5302   gsum   = err_glb[0];
5303   gsuma  = err_glb[1];
5304   gsumr  = err_glb[2];
5305   n_glb  = err_glb[3];
5306   na_glb = err_glb[4];
5307   nr_glb = err_glb[5];
5308 
5309   *norm  = 0.;
5310   if (n_glb>0.) *norm  = PetscSqrtReal(gsum  / n_glb);
5311   *norma = 0.;
5312   if (na_glb>0.) *norma = PetscSqrtReal(gsuma / na_glb);
5313   *normr = 0.;
5314   if (nr_glb>0.) *normr = PetscSqrtReal(gsumr / nr_glb);
5315 
5316   PetscCheck(!PetscIsInfOrNanScalar(*norm),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm");
5317   PetscCheck(!PetscIsInfOrNanScalar(*norma),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norma");
5318   PetscCheck(!PetscIsInfOrNanScalar(*normr),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in normr");
5319   PetscFunctionReturn(0);
5320 }
5321 
5322 /*@
5323    TSErrorWeightedENormInfinity - compute a weighted infinity error norm based on supplied absolute and relative tolerances
5324    Collective on TS
5325 
5326    Input Parameters:
5327 +  ts - time stepping context
5328 .  E - error vector
5329 .  U - state vector, usually ts->vec_sol
5330 -  Y - state vector, previous time step
5331 
5332    Output Parameters:
5333 +  norm - weighted norm, a value of 1.0 means that the error matches the tolerances
5334 .  norma - weighted norm based on the absolute tolerance, a value of 1.0 means that the error matches the tolerances
5335 -  normr - weighted norm based on the relative tolerance, a value of 1.0 means that the error matches the tolerances
5336 
5337    Level: developer
5338 
5339 .seealso: `TSErrorWeightedENorm()`, `TSErrorWeightedENorm2()`
5340 @*/
5341 PetscErrorCode TSErrorWeightedENormInfinity(TS ts,Vec E,Vec U,Vec Y,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5342 {
5343   PetscInt          i,n,N,rstart;
5344   const PetscScalar *e,*u,*y;
5345   PetscReal         err,max,gmax,maxa,gmaxa,maxr,gmaxr;
5346   PetscReal         tol,tola,tolr;
5347   PetscReal         err_loc[3],err_glb[3];
5348 
5349   PetscFunctionBegin;
5350   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5351   PetscValidHeaderSpecific(E,VEC_CLASSID,2);
5352   PetscValidHeaderSpecific(U,VEC_CLASSID,3);
5353   PetscValidHeaderSpecific(Y,VEC_CLASSID,4);
5354   PetscValidType(E,2);
5355   PetscValidType(U,3);
5356   PetscValidType(Y,4);
5357   PetscCheckSameComm(E,2,U,3);
5358   PetscCheckSameComm(U,3,Y,4);
5359   PetscValidRealPointer(norm,5);
5360   PetscValidRealPointer(norma,6);
5361   PetscValidRealPointer(normr,7);
5362 
5363   PetscCall(VecGetSize(E,&N));
5364   PetscCall(VecGetLocalSize(E,&n));
5365   PetscCall(VecGetOwnershipRange(E,&rstart,NULL));
5366   PetscCall(VecGetArrayRead(E,&e));
5367   PetscCall(VecGetArrayRead(U,&u));
5368   PetscCall(VecGetArrayRead(Y,&y));
5369 
5370   max=0.;
5371   maxa=0.;
5372   maxr=0.;
5373 
5374   if (ts->vatol && ts->vrtol) {     /* vector atol, vector rtol */
5375     const PetscScalar *atol,*rtol;
5376     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5377     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5378 
5379     for (i=0; i<n; i++) {
5380       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5381       err = PetscAbsScalar(e[i]);
5382       tola = PetscRealPart(atol[i]);
5383       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5384       tol  = tola+tolr;
5385       if (tola>0.) maxa = PetscMax(maxa,err / tola);
5386       if (tolr>0.) maxr = PetscMax(maxr,err / tolr);
5387       if (tol>0.)  max = PetscMax(max,err / tol);
5388     }
5389     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5390     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5391   } else if (ts->vatol) {       /* vector atol, scalar rtol */
5392     const PetscScalar *atol;
5393     PetscCall(VecGetArrayRead(ts->vatol,&atol));
5394     for (i=0; i<n; i++) {
5395       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5396       err = PetscAbsScalar(e[i]);
5397       tola = PetscRealPart(atol[i]);
5398       tolr = ts->rtol  * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5399       tol  = tola+tolr;
5400       if (tola>0.) maxa = PetscMax(maxa,err / tola);
5401       if (tolr>0.) maxr = PetscMax(maxr,err / tolr);
5402       if (tol>0.)  max = PetscMax(max,err / tol);
5403     }
5404     PetscCall(VecRestoreArrayRead(ts->vatol,&atol));
5405   } else if (ts->vrtol) {       /* scalar atol, vector rtol */
5406     const PetscScalar *rtol;
5407     PetscCall(VecGetArrayRead(ts->vrtol,&rtol));
5408 
5409     for (i=0; i<n; i++) {
5410       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5411       err = PetscAbsScalar(e[i]);
5412       tola = ts->atol;
5413       tolr = PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5414       tol  = tola+tolr;
5415       if (tola>0.) maxa = PetscMax(maxa,err / tola);
5416       if (tolr>0.) maxr = PetscMax(maxr,err / tolr);
5417       if (tol>0.) max = PetscMax(max,err / tol);
5418     }
5419     PetscCall(VecRestoreArrayRead(ts->vrtol,&rtol));
5420   } else {                      /* scalar atol, scalar rtol */
5421 
5422     for (i=0; i<n; i++) {
5423       SkipSmallValue(y[i],u[i],ts->adapt->ignore_max);
5424       err = PetscAbsScalar(e[i]);
5425       tola = ts->atol;
5426       tolr = ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i]));
5427       tol  = tola+tolr;
5428       if (tola>0.) maxa = PetscMax(maxa,err / tola);
5429       if (tolr>0.) maxr = PetscMax(maxr,err / tolr);
5430       if (tol>0.)  max = PetscMax(max,err / tol);
5431     }
5432   }
5433   PetscCall(VecRestoreArrayRead(E,&e));
5434   PetscCall(VecRestoreArrayRead(U,&u));
5435   PetscCall(VecRestoreArrayRead(Y,&y));
5436   err_loc[0] = max;
5437   err_loc[1] = maxa;
5438   err_loc[2] = maxr;
5439   PetscCall(MPIU_Allreduce(err_loc,err_glb,3,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)ts)));
5440   gmax   = err_glb[0];
5441   gmaxa  = err_glb[1];
5442   gmaxr  = err_glb[2];
5443 
5444   *norm = gmax;
5445   *norma = gmaxa;
5446   *normr = gmaxr;
5447   PetscCheck(!PetscIsInfOrNanScalar(*norm),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm");
5448   PetscCheck(!PetscIsInfOrNanScalar(*norma),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norma");
5449   PetscCheck(!PetscIsInfOrNanScalar(*normr),PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in normr");
5450   PetscFunctionReturn(0);
5451 }
5452 
5453 /*@
5454    TSErrorWeightedENorm - compute a weighted error norm based on supplied absolute and relative tolerances
5455 
5456    Collective on TS
5457 
5458    Input Parameters:
5459 +  ts - time stepping context
5460 .  E - error vector
5461 .  U - state vector, usually ts->vec_sol
5462 .  Y - state vector, previous time step
5463 -  wnormtype - norm type, either NORM_2 or NORM_INFINITY
5464 
5465    Output Parameters:
5466 +  norm  - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5467 .  norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5468 -  normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5469 
5470    Options Database Keys:
5471 .  -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5472 
5473    Level: developer
5474 
5475 .seealso: `TSErrorWeightedENormInfinity()`, `TSErrorWeightedENorm2()`, `TSErrorWeightedNormInfinity()`, `TSErrorWeightedNorm2()`
5476 @*/
5477 PetscErrorCode TSErrorWeightedENorm(TS ts,Vec E,Vec U,Vec Y,NormType wnormtype,PetscReal *norm,PetscReal *norma,PetscReal *normr)
5478 {
5479   PetscFunctionBegin;
5480   if (wnormtype == NORM_2)PetscCall(TSErrorWeightedENorm2(ts,E,U,Y,norm,norma,normr));
5481   else if (wnormtype == NORM_INFINITY) PetscCall(TSErrorWeightedENormInfinity(ts,E,U,Y,norm,norma,normr));
5482   else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]);
5483   PetscFunctionReturn(0);
5484 }
5485 
5486 /*@
5487    TSSetCFLTimeLocal - Set the local CFL constraint relative to forward Euler
5488 
5489    Logically Collective on TS
5490 
5491    Input Parameters:
5492 +  ts - time stepping context
5493 -  cfltime - maximum stable time step if using forward Euler (value can be different on each process)
5494 
5495    Note:
5496    After calling this function, the global CFL time can be obtained by calling TSGetCFLTime()
5497 
5498    Level: intermediate
5499 
5500 .seealso: `TSGetCFLTime()`, `TSADAPTCFL`
5501 @*/
5502 PetscErrorCode TSSetCFLTimeLocal(TS ts,PetscReal cfltime)
5503 {
5504   PetscFunctionBegin;
5505   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5506   ts->cfltime_local = cfltime;
5507   ts->cfltime       = -1.;
5508   PetscFunctionReturn(0);
5509 }
5510 
5511 /*@
5512    TSGetCFLTime - Get the maximum stable time step according to CFL criteria applied to forward Euler
5513 
5514    Collective on TS
5515 
5516    Input Parameter:
5517 .  ts - time stepping context
5518 
5519    Output Parameter:
5520 .  cfltime - maximum stable time step for forward Euler
5521 
5522    Level: advanced
5523 
5524 .seealso: `TSSetCFLTimeLocal()`
5525 @*/
5526 PetscErrorCode TSGetCFLTime(TS ts,PetscReal *cfltime)
5527 {
5528   PetscFunctionBegin;
5529   if (ts->cfltime < 0) PetscCall(MPIU_Allreduce(&ts->cfltime_local,&ts->cfltime,1,MPIU_REAL,MPIU_MIN,PetscObjectComm((PetscObject)ts)));
5530   *cfltime = ts->cfltime;
5531   PetscFunctionReturn(0);
5532 }
5533 
5534 /*@
5535    TSVISetVariableBounds - Sets the lower and upper bounds for the solution vector. xl <= x <= xu
5536 
5537    Input Parameters:
5538 +  ts   - the TS context.
5539 .  xl   - lower bound.
5540 -  xu   - upper bound.
5541 
5542    Notes:
5543    If this routine is not called then the lower and upper bounds are set to
5544    PETSC_NINFINITY and PETSC_INFINITY respectively during SNESSetUp().
5545 
5546    Level: advanced
5547 
5548 @*/
5549 PetscErrorCode TSVISetVariableBounds(TS ts, Vec xl, Vec xu)
5550 {
5551   SNES           snes;
5552 
5553   PetscFunctionBegin;
5554   PetscCall(TSGetSNES(ts,&snes));
5555   PetscCall(SNESVISetVariableBounds(snes,xl,xu));
5556   PetscFunctionReturn(0);
5557 }
5558 
5559 /*@
5560    TSComputeLinearStability - computes the linear stability function at a point
5561 
5562    Collective on TS
5563 
5564    Input Parameters:
5565 +  ts - the TS context
5566 -  xr,xi - real and imaginary part of input arguments
5567 
5568    Output Parameters:
5569 .  yr,yi - real and imaginary part of function value
5570 
5571    Level: developer
5572 
5573 .seealso: `TSSetRHSFunction()`, `TSComputeIFunction()`
5574 @*/
5575 PetscErrorCode TSComputeLinearStability(TS ts,PetscReal xr,PetscReal xi,PetscReal *yr,PetscReal *yi)
5576 {
5577   PetscFunctionBegin;
5578   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5579   PetscUseTypeMethod(ts,linearstability ,xr,xi,yr,yi);
5580   PetscFunctionReturn(0);
5581 }
5582 
5583 /*@
5584    TSRestartStep - Flags the solver to restart the next step
5585 
5586    Collective on TS
5587 
5588    Input Parameter:
5589 .  ts - the TS context obtained from TSCreate()
5590 
5591    Level: advanced
5592 
5593    Notes:
5594    Multistep methods like BDF or Runge-Kutta methods with FSAL property require restarting the solver in the event of
5595    discontinuities. These discontinuities may be introduced as a consequence of explicitly modifications to the solution
5596    vector (which PETSc attempts to detect and handle) or problem coefficients (which PETSc is not able to detect). For
5597    the sake of correctness and maximum safety, users are expected to call TSRestart() whenever they introduce
5598    discontinuities in callback routines (e.g. prestep and poststep routines, or implicit/rhs function routines with
5599    discontinuous source terms).
5600 
5601 .seealso: `TSSolve()`, `TSSetPreStep()`, `TSSetPostStep()`
5602 @*/
5603 PetscErrorCode TSRestartStep(TS ts)
5604 {
5605   PetscFunctionBegin;
5606   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5607   ts->steprestart = PETSC_TRUE;
5608   PetscFunctionReturn(0);
5609 }
5610 
5611 /*@
5612    TSRollBack - Rolls back one time step
5613 
5614    Collective on TS
5615 
5616    Input Parameter:
5617 .  ts - the TS context obtained from TSCreate()
5618 
5619    Level: advanced
5620 
5621 .seealso: `TSCreate()`, `TSSetUp()`, `TSDestroy()`, `TSSolve()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSInterpolate()`
5622 @*/
5623 PetscErrorCode  TSRollBack(TS ts)
5624 {
5625   PetscFunctionBegin;
5626   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
5627   PetscCheck(!ts->steprollback,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONGSTATE,"TSRollBack already called");
5628   PetscUseTypeMethod(ts,rollback);
5629   ts->time_step = ts->ptime - ts->ptime_prev;
5630   ts->ptime = ts->ptime_prev;
5631   ts->ptime_prev = ts->ptime_prev_rollback;
5632   ts->steps--;
5633   ts->steprollback = PETSC_TRUE;
5634   PetscFunctionReturn(0);
5635 }
5636 
5637 /*@
5638    TSGetStages - Get the number of stages and stage values
5639 
5640    Input Parameter:
5641 .  ts - the TS context obtained from TSCreate()
5642 
5643    Output Parameters:
5644 +  ns - the number of stages
5645 -  Y - the current stage vectors
5646 
5647    Level: advanced
5648 
5649    Notes: Both ns and Y can be NULL.
5650 
5651 .seealso: `TSCreate()`
5652 @*/
5653 PetscErrorCode  TSGetStages(TS ts,PetscInt *ns,Vec **Y)
5654 {
5655   PetscFunctionBegin;
5656   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
5657   if (ns) PetscValidIntPointer(ns,2);
5658   if (Y) PetscValidPointer(Y,3);
5659   if (!ts->ops->getstages) {
5660     if (ns) *ns = 0;
5661     if (Y) *Y = NULL;
5662   } else PetscUseTypeMethod(ts,getstages ,ns,Y);
5663   PetscFunctionReturn(0);
5664 }
5665 
5666 /*@C
5667   TSComputeIJacobianDefaultColor - Computes the Jacobian using finite differences and coloring to exploit matrix sparsity.
5668 
5669   Collective on SNES
5670 
5671   Input Parameters:
5672 + ts - the TS context
5673 . t - current timestep
5674 . U - state vector
5675 . Udot - time derivative of state vector
5676 . shift - shift to apply, see note below
5677 - ctx - an optional user context
5678 
5679   Output Parameters:
5680 + J - Jacobian matrix (not altered in this routine)
5681 - B - newly computed Jacobian matrix to use with preconditioner (generally the same as J)
5682 
5683   Level: intermediate
5684 
5685   Notes:
5686   If F(t,U,Udot)=0 is the DAE, the required Jacobian is
5687 
5688   dF/dU + shift*dF/dUdot
5689 
5690   Most users should not need to explicitly call this routine, as it
5691   is used internally within the nonlinear solvers.
5692 
5693   This will first try to get the coloring from the DM.  If the DM type has no coloring
5694   routine, then it will try to get the coloring from the matrix.  This requires that the
5695   matrix have nonzero entries precomputed.
5696 
5697 .seealso: `TSSetIJacobian()`, `MatFDColoringCreate()`, `MatFDColoringSetFunction()`
5698 @*/
5699 PetscErrorCode TSComputeIJacobianDefaultColor(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat J,Mat B,void *ctx)
5700 {
5701   SNES           snes;
5702   MatFDColoring  color;
5703   PetscBool      hascolor, matcolor = PETSC_FALSE;
5704 
5705   PetscFunctionBegin;
5706   PetscCall(PetscOptionsGetBool(((PetscObject)ts)->options,((PetscObject) ts)->prefix, "-ts_fd_color_use_mat", &matcolor, NULL));
5707   PetscCall(PetscObjectQuery((PetscObject) B, "TSMatFDColoring", (PetscObject *) &color));
5708   if (!color) {
5709     DM         dm;
5710     ISColoring iscoloring;
5711 
5712     PetscCall(TSGetDM(ts, &dm));
5713     PetscCall(DMHasColoring(dm, &hascolor));
5714     if (hascolor && !matcolor) {
5715       PetscCall(DMCreateColoring(dm, IS_COLORING_GLOBAL, &iscoloring));
5716       PetscCall(MatFDColoringCreate(B, iscoloring, &color));
5717       PetscCall(MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts));
5718       PetscCall(MatFDColoringSetFromOptions(color));
5719       PetscCall(MatFDColoringSetUp(B, iscoloring, color));
5720       PetscCall(ISColoringDestroy(&iscoloring));
5721     } else {
5722       MatColoring mc;
5723 
5724       PetscCall(MatColoringCreate(B, &mc));
5725       PetscCall(MatColoringSetDistance(mc, 2));
5726       PetscCall(MatColoringSetType(mc, MATCOLORINGSL));
5727       PetscCall(MatColoringSetFromOptions(mc));
5728       PetscCall(MatColoringApply(mc, &iscoloring));
5729       PetscCall(MatColoringDestroy(&mc));
5730       PetscCall(MatFDColoringCreate(B, iscoloring, &color));
5731       PetscCall(MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts));
5732       PetscCall(MatFDColoringSetFromOptions(color));
5733       PetscCall(MatFDColoringSetUp(B, iscoloring, color));
5734       PetscCall(ISColoringDestroy(&iscoloring));
5735     }
5736     PetscCall(PetscObjectCompose((PetscObject) B, "TSMatFDColoring", (PetscObject) color));
5737     PetscCall(PetscObjectDereference((PetscObject) color));
5738   }
5739   PetscCall(TSGetSNES(ts, &snes));
5740   PetscCall(MatFDColoringApply(B, color, U, snes));
5741   if (J != B) {
5742     PetscCall(MatAssemblyBegin(J, MAT_FINAL_ASSEMBLY));
5743     PetscCall(MatAssemblyEnd(J, MAT_FINAL_ASSEMBLY));
5744   }
5745   PetscFunctionReturn(0);
5746 }
5747 
5748 /*@
5749     TSSetFunctionDomainError - Set a function that tests if the current state vector is valid
5750 
5751     Input Parameters:
5752 +    ts - the TS context
5753 -    func - function called within TSFunctionDomainError
5754 
5755     Calling sequence of func:
5756 $     PetscErrorCode func(TS ts,PetscReal time,Vec state,PetscBool reject)
5757 
5758 +   ts - the TS context
5759 .   time - the current time (of the stage)
5760 .   state - the state to check if it is valid
5761 -   reject - (output parameter) PETSC_FALSE if the state is acceptable, PETSC_TRUE if not acceptable
5762 
5763     Level: intermediate
5764 
5765     Notes:
5766       If an implicit ODE solver is being used then, in addition to providing this routine, the
5767       user's code should call SNESSetFunctionDomainError() when domain errors occur during
5768       function evaluations where the functions are provided by TSSetIFunction() or TSSetRHSFunction().
5769       Use TSGetSNES() to obtain the SNES object
5770 
5771     Developer Notes:
5772       The naming of this function is inconsistent with the SNESSetFunctionDomainError()
5773       since one takes a function pointer and the other does not.
5774 
5775 .seealso: `TSAdaptCheckStage()`, `TSFunctionDomainError()`, `SNESSetFunctionDomainError()`, `TSGetSNES()`
5776 @*/
5777 
5778 PetscErrorCode TSSetFunctionDomainError(TS ts, PetscErrorCode (*func)(TS,PetscReal,Vec,PetscBool*))
5779 {
5780   PetscFunctionBegin;
5781   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
5782   ts->functiondomainerror = func;
5783   PetscFunctionReturn(0);
5784 }
5785 
5786 /*@
5787     TSFunctionDomainError - Checks if the current state is valid
5788 
5789     Input Parameters:
5790 +    ts - the TS context
5791 .    stagetime - time of the simulation
5792 -    Y - state vector to check.
5793 
5794     Output Parameter:
5795 .    accept - Set to PETSC_FALSE if the current state vector is valid.
5796 
5797     Note:
5798     This function is called by the TS integration routines and calls the user provided function (set with TSSetFunctionDomainError())
5799     to check if the current state is valid.
5800 
5801     Level: developer
5802 
5803 .seealso: `TSSetFunctionDomainError()`
5804 @*/
5805 PetscErrorCode TSFunctionDomainError(TS ts,PetscReal stagetime,Vec Y,PetscBool* accept)
5806 {
5807   PetscFunctionBegin;
5808   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
5809   *accept = PETSC_TRUE;
5810   if (ts->functiondomainerror) PetscCall((*ts->functiondomainerror)(ts,stagetime,Y,accept));
5811   PetscFunctionReturn(0);
5812 }
5813 
5814 /*@C
5815   TSClone - This function clones a time step object.
5816 
5817   Collective
5818 
5819   Input Parameter:
5820 . tsin    - The input TS
5821 
5822   Output Parameter:
5823 . tsout   - The output TS (cloned)
5824 
5825   Notes:
5826   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.
5827 
5828   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);
5829 
5830   Level: developer
5831 
5832 .seealso: `TSCreate()`, `TSSetType()`, `TSSetUp()`, `TSDestroy()`, `TSSetProblemType()`
5833 @*/
5834 PetscErrorCode  TSClone(TS tsin, TS *tsout)
5835 {
5836   TS             t;
5837   SNES           snes_start;
5838   DM             dm;
5839   TSType         type;
5840 
5841   PetscFunctionBegin;
5842   PetscValidPointer(tsin,1);
5843   *tsout = NULL;
5844 
5845   PetscCall(PetscHeaderCreate(t, TS_CLASSID, "TS", "Time stepping", "TS", PetscObjectComm((PetscObject)tsin), TSDestroy, TSView));
5846 
5847   /* General TS description */
5848   t->numbermonitors    = 0;
5849   t->monitorFrequency  = 1;
5850   t->setupcalled       = 0;
5851   t->ksp_its           = 0;
5852   t->snes_its          = 0;
5853   t->nwork             = 0;
5854   t->rhsjacobian.time  = PETSC_MIN_REAL;
5855   t->rhsjacobian.scale = 1.;
5856   t->ijacobian.shift   = 1.;
5857 
5858   PetscCall(TSGetSNES(tsin,&snes_start));
5859   PetscCall(TSSetSNES(t,snes_start));
5860 
5861   PetscCall(TSGetDM(tsin,&dm));
5862   PetscCall(TSSetDM(t,dm));
5863 
5864   t->adapt = tsin->adapt;
5865   PetscCall(PetscObjectReference((PetscObject)t->adapt));
5866 
5867   t->trajectory = tsin->trajectory;
5868   PetscCall(PetscObjectReference((PetscObject)t->trajectory));
5869 
5870   t->event = tsin->event;
5871   if (t->event) t->event->refct++;
5872 
5873   t->problem_type      = tsin->problem_type;
5874   t->ptime             = tsin->ptime;
5875   t->ptime_prev        = tsin->ptime_prev;
5876   t->time_step         = tsin->time_step;
5877   t->max_time          = tsin->max_time;
5878   t->steps             = tsin->steps;
5879   t->max_steps         = tsin->max_steps;
5880   t->equation_type     = tsin->equation_type;
5881   t->atol              = tsin->atol;
5882   t->rtol              = tsin->rtol;
5883   t->max_snes_failures = tsin->max_snes_failures;
5884   t->max_reject        = tsin->max_reject;
5885   t->errorifstepfailed = tsin->errorifstepfailed;
5886 
5887   PetscCall(TSGetType(tsin,&type));
5888   PetscCall(TSSetType(t,type));
5889 
5890   t->vec_sol           = NULL;
5891 
5892   t->cfltime          = tsin->cfltime;
5893   t->cfltime_local    = tsin->cfltime_local;
5894   t->exact_final_time = tsin->exact_final_time;
5895 
5896   PetscCall(PetscMemcpy(t->ops,tsin->ops,sizeof(struct _TSOps)));
5897 
5898   if (((PetscObject)tsin)->fortran_func_pointers) {
5899     PetscInt i;
5900     PetscCall(PetscMalloc((10)*sizeof(void(*)(void)),&((PetscObject)t)->fortran_func_pointers));
5901     for (i=0; i<10; i++) {
5902       ((PetscObject)t)->fortran_func_pointers[i] = ((PetscObject)tsin)->fortran_func_pointers[i];
5903     }
5904   }
5905   *tsout = t;
5906   PetscFunctionReturn(0);
5907 }
5908 
5909 static PetscErrorCode RHSWrapperFunction_TSRHSJacobianTest(void* ctx,Vec x,Vec y)
5910 {
5911   TS             ts = (TS) ctx;
5912 
5913   PetscFunctionBegin;
5914   PetscCall(TSComputeRHSFunction(ts,0,x,y));
5915   PetscFunctionReturn(0);
5916 }
5917 
5918 /*@
5919     TSRHSJacobianTest - Compares the multiply routine provided to the MATSHELL with differencing on the TS given RHS function.
5920 
5921    Logically Collective on TS
5922 
5923     Input Parameters:
5924     TS - the time stepping routine
5925 
5926    Output Parameter:
5927 .   flg - PETSC_TRUE if the multiply is likely correct
5928 
5929    Options Database:
5930  .   -ts_rhs_jacobian_test_mult -mat_shell_test_mult_view - run the test at each timestep of the integrator
5931 
5932    Level: advanced
5933 
5934    Notes:
5935     This only works for problems defined only the RHS function and Jacobian NOT IFunction and IJacobian
5936 
5937 .seealso: `MatCreateShell()`, `MatShellGetContext()`, `MatShellGetOperation()`, `MatShellTestMultTranspose()`, `TSRHSJacobianTestTranspose()`
5938 @*/
5939 PetscErrorCode  TSRHSJacobianTest(TS ts,PetscBool *flg)
5940 {
5941   Mat            J,B;
5942   TSRHSJacobian  func;
5943   void*          ctx;
5944 
5945   PetscFunctionBegin;
5946   PetscCall(TSGetRHSJacobian(ts,&J,&B,&func,&ctx));
5947   PetscCall((*func)(ts,0.0,ts->vec_sol,J,B,ctx));
5948   PetscCall(MatShellTestMult(J,RHSWrapperFunction_TSRHSJacobianTest,ts->vec_sol,ts,flg));
5949   PetscFunctionReturn(0);
5950 }
5951 
5952 /*@C
5953     TSRHSJacobianTestTranspose - Compares the multiply transpose routine provided to the MATSHELL with differencing on the TS given RHS function.
5954 
5955    Logically Collective on TS
5956 
5957     Input Parameters:
5958     TS - the time stepping routine
5959 
5960    Output Parameter:
5961 .   flg - PETSC_TRUE if the multiply is likely correct
5962 
5963    Options Database:
5964 .   -ts_rhs_jacobian_test_mult_transpose -mat_shell_test_mult_transpose_view - run the test at each timestep of the integrator
5965 
5966    Notes:
5967     This only works for problems defined only the RHS function and Jacobian NOT IFunction and IJacobian
5968 
5969    Level: advanced
5970 
5971 .seealso: `MatCreateShell()`, `MatShellGetContext()`, `MatShellGetOperation()`, `MatShellTestMultTranspose()`, `TSRHSJacobianTest()`
5972 @*/
5973 PetscErrorCode  TSRHSJacobianTestTranspose(TS ts,PetscBool *flg)
5974 {
5975   Mat            J,B;
5976   void           *ctx;
5977   TSRHSJacobian  func;
5978 
5979   PetscFunctionBegin;
5980   PetscCall(TSGetRHSJacobian(ts,&J,&B,&func,&ctx));
5981   PetscCall((*func)(ts,0.0,ts->vec_sol,J,B,ctx));
5982   PetscCall(MatShellTestMultTranspose(J,RHSWrapperFunction_TSRHSJacobianTest,ts->vec_sol,ts,flg));
5983   PetscFunctionReturn(0);
5984 }
5985 
5986 /*@
5987   TSSetUseSplitRHSFunction - Use the split RHSFunction when a multirate method is used.
5988 
5989   Logically collective
5990 
5991   Input Parameters:
5992 +  ts - timestepping context
5993 -  use_splitrhsfunction - PETSC_TRUE indicates that the split RHSFunction will be used
5994 
5995   Options Database:
5996 .   -ts_use_splitrhsfunction - <true,false>
5997 
5998   Notes:
5999     This is only useful for multirate methods
6000 
6001   Level: intermediate
6002 
6003 .seealso: `TSGetUseSplitRHSFunction()`
6004 @*/
6005 PetscErrorCode TSSetUseSplitRHSFunction(TS ts, PetscBool use_splitrhsfunction)
6006 {
6007   PetscFunctionBegin;
6008   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
6009   ts->use_splitrhsfunction = use_splitrhsfunction;
6010   PetscFunctionReturn(0);
6011 }
6012 
6013 /*@
6014   TSGetUseSplitRHSFunction - Gets whether to use the split RHSFunction when a multirate method is used.
6015 
6016   Not collective
6017 
6018   Input Parameter:
6019 .  ts - timestepping context
6020 
6021   Output Parameter:
6022 .  use_splitrhsfunction - PETSC_TRUE indicates that the split RHSFunction will be used
6023 
6024   Level: intermediate
6025 
6026 .seealso: `TSSetUseSplitRHSFunction()`
6027 @*/
6028 PetscErrorCode TSGetUseSplitRHSFunction(TS ts, PetscBool *use_splitrhsfunction)
6029 {
6030   PetscFunctionBegin;
6031   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
6032   *use_splitrhsfunction = ts->use_splitrhsfunction;
6033   PetscFunctionReturn(0);
6034 }
6035 
6036 /*@
6037     TSSetMatStructure - sets the relationship between the nonzero structure of the RHS Jacobian matrix to the IJacobian matrix.
6038 
6039    Logically  Collective on ts
6040 
6041    Input Parameters:
6042 +  ts - the time-stepper
6043 -  str - the structure (the default is UNKNOWN_NONZERO_PATTERN)
6044 
6045    Level: intermediate
6046 
6047    Notes:
6048      When the relationship between the nonzero structures is known and supplied the solution process can be much faster
6049 
6050 .seealso: `MatAXPY()`, `MatStructure`
6051  @*/
6052 PetscErrorCode TSSetMatStructure(TS ts,MatStructure str)
6053 {
6054   PetscFunctionBegin;
6055   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
6056   ts->axpy_pattern = str;
6057   PetscFunctionReturn(0);
6058 }
6059 
6060 /*@
6061   TSSetTimeSpan - sets the time span. The solution will be computed and stored for each time requested.
6062 
6063   Collective on ts
6064 
6065   Input Parameters:
6066 + ts - the time-stepper
6067 . n - number of the time points (>=2)
6068 - span_times - array of the time points. The first element and the last element are the initial time and the final time respectively.
6069 
6070   Options Database Keys:
6071 . -ts_time_span <t0,...tf> - Sets the time span
6072 
6073   Level: beginner
6074 
6075   Notes:
6076   The elements in tspan must be all increasing. They correspond to the intermediate points for time integration.
6077   TS_EXACTFINALTIME_MATCHSTEP must be used to make the last time step in each sub-interval match the intermediate points specified.
6078   The intermediate solutions are saved in a vector array that can be accessed with TSGetSolutions(). Thus using time span may
6079   pressure the memory system when using a large number of span points.
6080 
6081 .seealso: `TSGetTimeSpan()`, `TSGetSolutions()`
6082  @*/
6083 PetscErrorCode TSSetTimeSpan(TS ts,PetscInt n,PetscReal *span_times)
6084 {
6085   PetscFunctionBegin;
6086   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
6087   PetscCheck(n >= 2,PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Minimum time span size is 2 but %" PetscInt_FMT " is provided",n);
6088   if (ts->tspan && n != ts->tspan->num_span_times) {
6089     PetscCall(PetscFree(ts->tspan->span_times));
6090     PetscCall(VecDestroyVecs(ts->tspan->num_span_times,&ts->tspan->vecs_sol));
6091     PetscCall(PetscMalloc1(n,&ts->tspan->span_times));
6092   }
6093   if (!ts->tspan) {
6094     TSTimeSpan tspan;
6095     PetscCall(PetscNew(&tspan));
6096     PetscCall(PetscMalloc1(n,&tspan->span_times));
6097     tspan->reltol = 1e-6;
6098     tspan->abstol = 10*PETSC_MACHINE_EPSILON;
6099     ts->tspan = tspan;
6100   }
6101   ts->tspan->num_span_times = n;
6102   PetscCall(PetscArraycpy(ts->tspan->span_times,span_times,n));
6103   PetscCall(TSSetTime(ts,ts->tspan->span_times[0]));
6104   PetscCall(TSSetMaxTime(ts,ts->tspan->span_times[n-1]));
6105   PetscFunctionReturn(0);
6106 }
6107 
6108 /*@C
6109   TSGetTimeSpan - gets the time span.
6110 
6111   Not Collective
6112 
6113   Input Parameter:
6114 . ts - the time-stepper
6115 
6116   Output Parameters:
6117 + n - number of the time points (>=2)
6118 - span_times - array of the time points. The first element and the last element are the initial time and the final time respectively. The values are valid until the TS object is destroyed.
6119 
6120   Level: beginner
6121   Notes: Both n and span_times can be NULL.
6122 
6123 .seealso: `TSSetTimeSpan()`, `TSGetSolutions()`
6124  @*/
6125 PetscErrorCode TSGetTimeSpan(TS ts,PetscInt *n,const PetscReal **span_times)
6126 {
6127   PetscFunctionBegin;
6128   PetscValidHeaderSpecific(ts,TS_CLASSID,1);
6129   if (n) PetscValidIntPointer(n,2);
6130   if (span_times) PetscValidPointer(span_times,3);
6131   if (!ts->tspan) {
6132     if (n) *n = 0;
6133     if (span_times) *span_times = NULL;
6134   } else {
6135     if (n) *n = ts->tspan->num_span_times;
6136     if (span_times) *span_times = ts->tspan->span_times;
6137   }
6138   PetscFunctionReturn(0);
6139 }
6140 
6141 /*@
6142    TSGetTimeSpanSolutions - Get the number of solutions and the solutions at the time points specified by the time span.
6143 
6144    Input Parameter:
6145 .  ts - the TS context obtained from TSCreate()
6146 
6147    Output Parameters:
6148 +  nsol - the number of solutions
6149 -  Sols - the solution vectors
6150 
6151    Level: beginner
6152 
6153    Notes:
6154     Both nsol and Sols can be NULL.
6155     Some time points in the time span may be skipped by TS so that nsol is less than the number of points specified by TSSetTimeSpan(). For example, manipulating the step size, especially with a reduced precision, may cause TS to step over certain points in the span.
6156 
6157 .seealso: `TSSetTimeSpan()`
6158 @*/
6159 PetscErrorCode TSGetTimeSpanSolutions(TS ts,PetscInt *nsol,Vec **Sols)
6160 {
6161   PetscFunctionBegin;
6162   PetscValidHeaderSpecific(ts, TS_CLASSID,1);
6163   if (nsol) PetscValidIntPointer(nsol,2);
6164   if (Sols) PetscValidPointer(Sols,3);
6165   if (!ts->tspan) {
6166     if (nsol) *nsol = 0;
6167     if (Sols) *Sols = NULL;
6168   } else {
6169     if (nsol) *nsol = ts->tspan->spanctr;
6170     if (Sols) *Sols = ts->tspan->vecs_sol;
6171   }
6172   PetscFunctionReturn(0);
6173 }
6174