Actual source code: ts.c
1: #include <petsc/private/tsimpl.h>
2: #include <petscdmda.h>
3: #include <petscdmshell.h>
4: #include <petscdmplex.h>
5: #include <petscdmswarm.h>
6: #include <petscviewer.h>
7: #include <petscdraw.h>
8: #include <petscconvest.h>
10: /* Logging support */
11: PetscClassId TS_CLASSID, DMTS_CLASSID;
12: PetscLogEvent TS_Step, TS_PseudoComputeTimeStep, TS_FunctionEval, TS_JacobianEval;
14: const char *const TSExactFinalTimeOptions[] = {"UNSPECIFIED", "STEPOVER", "INTERPOLATE", "MATCHSTEP", "TSExactFinalTimeOption", "TS_EXACTFINALTIME_", NULL};
16: static PetscErrorCode TSAdaptSetDefaultType(TSAdapt adapt, TSAdaptType default_type)
17: {
18: PetscFunctionBegin;
20: PetscAssertPointer(default_type, 2);
21: if (!((PetscObject)adapt)->type_name) PetscCall(TSAdaptSetType(adapt, default_type));
22: PetscFunctionReturn(PETSC_SUCCESS);
23: }
25: /*@
26: TSSetFromOptions - Sets various `TS` parameters from the options database
28: Collective
30: Input Parameter:
31: . ts - the `TS` context obtained from `TSCreate()`
33: Options Database Keys:
34: + -ts_type <type> - EULER, BEULER, SUNDIALS, PSEUDO, CN, RK, THETA, ALPHA, GLLE, SSP, GLEE, BSYMP, IRK, see `TSType`
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 - 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_interval <interval> - output once every interval (default=1) time steps
67: . -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)
68: - -ts_monitor_envelope - determine maximum and minimum value of each component of the solution over the solution time
70: Level: beginner
72: Notes:
73: See `SNESSetFromOptions()` and `KSPSetFromOptions()` for how to control the nonlinear and linear solves used by the time-stepper.
75: Certain `SNES` options get reset for each new nonlinear solver, for example `-snes_lag_jacobian its` and `-snes_lag_preconditioner its`, in order
76: to retain them over the multiple nonlinear solves that `TS` uses you mush also provide `-snes_lag_jacobian_persists true` and
77: `-snes_lag_preconditioner_persists true`
79: Developer Notes:
80: We should unify all the -ts_monitor options in the way that -xxx_view has been unified
82: .seealso: [](ch_ts), `TS`, `TSGetType()`
83: @*/
84: PetscErrorCode TSSetFromOptions(TS ts)
85: {
86: PetscBool opt, flg, tflg;
87: char monfilename[PETSC_MAX_PATH_LEN];
88: PetscReal time_step, tspan[100];
89: PetscInt nt = PETSC_STATIC_ARRAY_LENGTH(tspan);
90: TSExactFinalTimeOption eftopt;
91: char dir[16];
92: TSIFunction ifun;
93: const char *defaultType;
94: char typeName[256];
96: PetscFunctionBegin;
99: PetscCall(TSRegisterAll());
100: PetscCall(TSGetIFunction(ts, NULL, &ifun, NULL));
102: PetscObjectOptionsBegin((PetscObject)ts);
103: if (((PetscObject)ts)->type_name) defaultType = ((PetscObject)ts)->type_name;
104: else defaultType = ifun ? TSBEULER : TSEULER;
105: PetscCall(PetscOptionsFList("-ts_type", "TS method", "TSSetType", TSList, defaultType, typeName, 256, &opt));
106: if (opt) PetscCall(TSSetType(ts, typeName));
107: else PetscCall(TSSetType(ts, defaultType));
109: /* Handle generic TS options */
110: PetscCall(PetscOptionsDeprecated("-ts_final_time", "-ts_max_time", "3.10", NULL));
111: PetscCall(PetscOptionsReal("-ts_max_time", "Maximum time to run to", "TSSetMaxTime", ts->max_time, &ts->max_time, NULL));
112: PetscCall(PetscOptionsRealArray("-ts_time_span", "Time span", "TSSetTimeSpan", tspan, &nt, &flg));
113: if (flg) PetscCall(TSSetTimeSpan(ts, nt, tspan));
114: PetscCall(PetscOptionsInt("-ts_max_steps", "Maximum number of time steps", "TSSetMaxSteps", ts->max_steps, &ts->max_steps, NULL));
115: PetscCall(PetscOptionsReal("-ts_init_time", "Initial time", "TSSetTime", ts->ptime, &ts->ptime, NULL));
116: PetscCall(PetscOptionsReal("-ts_dt", "Initial time step", "TSSetTimeStep", ts->time_step, &time_step, &flg));
117: if (flg) PetscCall(TSSetTimeStep(ts, time_step));
118: PetscCall(PetscOptionsEnum("-ts_exact_final_time", "Option for handling of final time step", "TSSetExactFinalTime", TSExactFinalTimeOptions, (PetscEnum)ts->exact_final_time, (PetscEnum *)&eftopt, &flg));
119: if (flg) PetscCall(TSSetExactFinalTime(ts, eftopt));
120: PetscCall(PetscOptionsInt("-ts_max_snes_failures", "Maximum number of nonlinear solve failures", "TSSetMaxSNESFailures", ts->max_snes_failures, &ts->max_snes_failures, NULL));
121: PetscCall(PetscOptionsInt("-ts_max_reject", "Maximum number of step rejections before step fails", "TSSetMaxStepRejections", ts->max_reject, &ts->max_reject, NULL));
122: PetscCall(PetscOptionsBool("-ts_error_if_step_fails", "Error if no step succeeds", "TSSetErrorIfStepFails", ts->errorifstepfailed, &ts->errorifstepfailed, NULL));
123: PetscCall(PetscOptionsReal("-ts_rtol", "Relative tolerance for local truncation error", "TSSetTolerances", ts->rtol, &ts->rtol, NULL));
124: PetscCall(PetscOptionsReal("-ts_atol", "Absolute tolerance for local truncation error", "TSSetTolerances", ts->atol, &ts->atol, NULL));
126: PetscCall(PetscOptionsBool("-ts_rhs_jacobian_test_mult", "Test the RHS Jacobian for consistency with RHS at each solve ", "None", ts->testjacobian, &ts->testjacobian, NULL));
127: 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));
128: PetscCall(PetscOptionsBool("-ts_use_splitrhsfunction", "Use the split RHS function for multirate solvers ", "TSSetUseSplitRHSFunction", ts->use_splitrhsfunction, &ts->use_splitrhsfunction, NULL));
129: #if defined(PETSC_HAVE_SAWS)
130: {
131: PetscBool set;
132: flg = PETSC_FALSE;
133: PetscCall(PetscOptionsBool("-ts_saws_block", "Block for SAWs memory snooper at end of TSSolve", "PetscObjectSAWsBlock", ((PetscObject)ts)->amspublishblock, &flg, &set));
134: if (set) PetscCall(PetscObjectSAWsSetBlock((PetscObject)ts, flg));
135: }
136: #endif
138: /* Monitor options */
139: PetscCall(PetscOptionsInt("-ts_monitor_frequency", "Number of time steps between monitor output", "TSMonitorSetFrequency", ts->monitorFrequency, &ts->monitorFrequency, NULL));
140: PetscCall(TSMonitorSetFromOptions(ts, "-ts_monitor", "Monitor time and timestep size", "TSMonitorDefault", TSMonitorDefault, NULL));
141: PetscCall(TSMonitorSetFromOptions(ts, "-ts_monitor_extreme", "Monitor extreme values of the solution", "TSMonitorExtreme", TSMonitorExtreme, NULL));
142: PetscCall(TSMonitorSetFromOptions(ts, "-ts_monitor_solution", "View the solution at each timestep", "TSMonitorSolution", TSMonitorSolution, NULL));
143: PetscCall(TSMonitorSetFromOptions(ts, "-ts_dmswarm_monitor_moments", "Monitor moments of particle distribution", "TSDMSwarmMonitorMoments", TSDMSwarmMonitorMoments, NULL));
145: PetscCall(PetscOptionsString("-ts_monitor_python", "Use Python function", "TSMonitorSet", NULL, monfilename, sizeof(monfilename), &flg));
146: if (flg) PetscCall(PetscPythonMonitorSet((PetscObject)ts, monfilename));
148: PetscCall(PetscOptionsName("-ts_monitor_lg_solution", "Monitor solution graphically", "TSMonitorLGSolution", &opt));
149: if (opt) {
150: PetscInt howoften = 1;
151: DM dm;
152: PetscBool net;
154: PetscCall(PetscOptionsInt("-ts_monitor_lg_solution", "Monitor solution graphically", "TSMonitorLGSolution", howoften, &howoften, NULL));
155: PetscCall(TSGetDM(ts, &dm));
156: PetscCall(PetscObjectTypeCompare((PetscObject)dm, DMNETWORK, &net));
157: if (net) {
158: TSMonitorLGCtxNetwork ctx;
159: PetscCall(TSMonitorLGCtxNetworkCreate(ts, NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 600, 400, howoften, &ctx));
160: PetscCall(TSMonitorSet(ts, TSMonitorLGCtxNetworkSolution, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxNetworkDestroy));
161: PetscCall(PetscOptionsBool("-ts_monitor_lg_solution_semilogy", "Plot the solution with a semi-log axis", "", ctx->semilogy, &ctx->semilogy, NULL));
162: } else {
163: TSMonitorLGCtx ctx;
164: PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF, NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
165: PetscCall(TSMonitorSet(ts, TSMonitorLGSolution, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
166: }
167: }
169: PetscCall(PetscOptionsName("-ts_monitor_lg_error", "Monitor error graphically", "TSMonitorLGError", &opt));
170: if (opt) {
171: TSMonitorLGCtx ctx;
172: PetscInt howoften = 1;
174: PetscCall(PetscOptionsInt("-ts_monitor_lg_error", "Monitor error graphically", "TSMonitorLGError", howoften, &howoften, NULL));
175: PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF, NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
176: PetscCall(TSMonitorSet(ts, TSMonitorLGError, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
177: }
178: PetscCall(TSMonitorSetFromOptions(ts, "-ts_monitor_error", "View the error at each timestep", "TSMonitorError", TSMonitorError, NULL));
180: PetscCall(PetscOptionsName("-ts_monitor_lg_timestep", "Monitor timestep size graphically", "TSMonitorLGTimeStep", &opt));
181: if (opt) {
182: TSMonitorLGCtx ctx;
183: PetscInt howoften = 1;
185: PetscCall(PetscOptionsInt("-ts_monitor_lg_timestep", "Monitor timestep size graphically", "TSMonitorLGTimeStep", howoften, &howoften, NULL));
186: PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
187: PetscCall(TSMonitorSet(ts, TSMonitorLGTimeStep, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
188: }
189: PetscCall(PetscOptionsName("-ts_monitor_lg_timestep_log", "Monitor log timestep size graphically", "TSMonitorLGTimeStep", &opt));
190: if (opt) {
191: TSMonitorLGCtx ctx;
192: PetscInt howoften = 1;
194: PetscCall(PetscOptionsInt("-ts_monitor_lg_timestep_log", "Monitor log timestep size graphically", "TSMonitorLGTimeStep", howoften, &howoften, NULL));
195: PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
196: PetscCall(TSMonitorSet(ts, TSMonitorLGTimeStep, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
197: ctx->semilogy = PETSC_TRUE;
198: }
200: PetscCall(PetscOptionsName("-ts_monitor_lg_snes_iterations", "Monitor number nonlinear iterations for each timestep graphically", "TSMonitorLGSNESIterations", &opt));
201: if (opt) {
202: TSMonitorLGCtx ctx;
203: PetscInt howoften = 1;
205: PetscCall(PetscOptionsInt("-ts_monitor_lg_snes_iterations", "Monitor number nonlinear iterations for each timestep graphically", "TSMonitorLGSNESIterations", howoften, &howoften, NULL));
206: PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
207: PetscCall(TSMonitorSet(ts, TSMonitorLGSNESIterations, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
208: }
209: PetscCall(PetscOptionsName("-ts_monitor_lg_ksp_iterations", "Monitor number nonlinear iterations for each timestep graphically", "TSMonitorLGKSPIterations", &opt));
210: if (opt) {
211: TSMonitorLGCtx ctx;
212: PetscInt howoften = 1;
214: PetscCall(PetscOptionsInt("-ts_monitor_lg_ksp_iterations", "Monitor number nonlinear iterations for each timestep graphically", "TSMonitorLGKSPIterations", howoften, &howoften, NULL));
215: PetscCall(TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 400, 300, howoften, &ctx));
216: PetscCall(TSMonitorSet(ts, TSMonitorLGKSPIterations, ctx, (PetscErrorCode(*)(void **))TSMonitorLGCtxDestroy));
217: }
218: PetscCall(PetscOptionsName("-ts_monitor_sp_eig", "Monitor eigenvalues of linearized operator graphically", "TSMonitorSPEig", &opt));
219: if (opt) {
220: TSMonitorSPEigCtx ctx;
221: PetscInt howoften = 1;
223: PetscCall(PetscOptionsInt("-ts_monitor_sp_eig", "Monitor eigenvalues of linearized operator graphically", "TSMonitorSPEig", howoften, &howoften, NULL));
224: PetscCall(TSMonitorSPEigCtxCreate(PETSC_COMM_SELF, NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &ctx));
225: PetscCall(TSMonitorSet(ts, TSMonitorSPEig, ctx, (PetscErrorCode(*)(void **))TSMonitorSPEigCtxDestroy));
226: }
227: PetscCall(PetscOptionsName("-ts_monitor_sp_swarm", "Display particle phase space from the DMSwarm", "TSMonitorSPSwarm", &opt));
228: if (opt) {
229: TSMonitorSPCtx ctx;
230: PetscInt howoften = 1, retain = 0;
231: PetscBool phase = PETSC_TRUE, create = PETSC_TRUE, multispecies = PETSC_FALSE;
233: for (PetscInt i = 0; i < ts->numbermonitors; ++i)
234: if (ts->monitor[i] == TSMonitorSPSwarmSolution) {
235: create = PETSC_FALSE;
236: break;
237: }
238: if (create) {
239: PetscCall(PetscOptionsInt("-ts_monitor_sp_swarm", "Display particles phase space from the DMSwarm", "TSMonitorSPSwarm", howoften, &howoften, NULL));
240: PetscCall(PetscOptionsInt("-ts_monitor_sp_swarm_retain", "Retain n points plotted to show trajectory, -1 for all points", "TSMonitorSPSwarm", retain, &retain, NULL));
241: PetscCall(PetscOptionsBool("-ts_monitor_sp_swarm_phase", "Plot in phase space rather than coordinate space", "TSMonitorSPSwarm", phase, &phase, NULL));
242: PetscCall(PetscOptionsBool("-ts_monitor_sp_swarm_multi_species", "Color particles by particle species", "TSMonitorSPSwarm", multispecies, &multispecies, NULL));
243: PetscCall(TSMonitorSPCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, retain, phase, multispecies, &ctx));
244: PetscCall(TSMonitorSet(ts, TSMonitorSPSwarmSolution, ctx, (PetscErrorCode(*)(void **))TSMonitorSPCtxDestroy));
245: }
246: }
247: PetscCall(PetscOptionsName("-ts_monitor_hg_swarm", "Display particle histogram from the DMSwarm", "TSMonitorHGSwarm", &opt));
248: if (opt) {
249: TSMonitorHGCtx ctx;
250: PetscInt howoften = 1, Ns = 1;
251: PetscBool velocity = PETSC_FALSE, create = PETSC_TRUE;
253: for (PetscInt i = 0; i < ts->numbermonitors; ++i)
254: if (ts->monitor[i] == TSMonitorHGSwarmSolution) {
255: create = PETSC_FALSE;
256: break;
257: }
258: if (create) {
259: DM sw, dm;
260: PetscInt Nc, Nb;
262: PetscCall(TSGetDM(ts, &sw));
263: PetscCall(DMSwarmGetCellDM(sw, &dm));
264: PetscCall(DMPlexGetHeightStratum(dm, 0, NULL, &Nc));
265: Nb = PetscMin(20, PetscMax(10, Nc));
266: PetscCall(PetscOptionsInt("-ts_monitor_hg_swarm", "Display particles histogram from the DMSwarm", "TSMonitorHGSwarm", howoften, &howoften, NULL));
267: PetscCall(PetscOptionsBool("-ts_monitor_hg_swarm_velocity", "Plot in velocity space rather than coordinate space", "TSMonitorHGSwarm", velocity, &velocity, NULL));
268: PetscCall(PetscOptionsInt("-ts_monitor_hg_swarm_species", "Number of species to histogram", "TSMonitorHGSwarm", Ns, &Ns, NULL));
269: PetscCall(PetscOptionsInt("-ts_monitor_hg_swarm_bins", "Number of histogram bins", "TSMonitorHGSwarm", Nb, &Nb, NULL));
270: PetscCall(TSMonitorHGCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, Ns, Nb, velocity, &ctx));
271: PetscCall(TSMonitorSet(ts, TSMonitorHGSwarmSolution, ctx, (PetscErrorCode(*)(void **))TSMonitorHGCtxDestroy));
272: }
273: }
274: opt = PETSC_FALSE;
275: PetscCall(PetscOptionsName("-ts_monitor_draw_solution", "Monitor solution graphically", "TSMonitorDrawSolution", &opt));
276: if (opt) {
277: TSMonitorDrawCtx ctx;
278: PetscInt howoften = 1;
280: PetscCall(PetscOptionsInt("-ts_monitor_draw_solution", "Monitor solution graphically", "TSMonitorDrawSolution", howoften, &howoften, NULL));
281: PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts), NULL, "Computed Solution", PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &ctx));
282: PetscCall(TSMonitorSet(ts, TSMonitorDrawSolution, ctx, (PetscErrorCode(*)(void **))TSMonitorDrawCtxDestroy));
283: }
284: opt = PETSC_FALSE;
285: PetscCall(PetscOptionsName("-ts_monitor_draw_solution_phase", "Monitor solution graphically", "TSMonitorDrawSolutionPhase", &opt));
286: if (opt) {
287: TSMonitorDrawCtx ctx;
288: PetscReal bounds[4];
289: PetscInt n = 4;
290: PetscDraw draw;
291: PetscDrawAxis axis;
293: PetscCall(PetscOptionsRealArray("-ts_monitor_draw_solution_phase", "Monitor solution graphically", "TSMonitorDrawSolutionPhase", bounds, &n, NULL));
294: PetscCheck(n == 4, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Must provide bounding box of phase field");
295: PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, 1, &ctx));
296: PetscCall(PetscViewerDrawGetDraw(ctx->viewer, 0, &draw));
297: PetscCall(PetscViewerDrawGetDrawAxis(ctx->viewer, 0, &axis));
298: PetscCall(PetscDrawAxisSetLimits(axis, bounds[0], bounds[2], bounds[1], bounds[3]));
299: PetscCall(PetscDrawAxisSetLabels(axis, "Phase Diagram", "Variable 1", "Variable 2"));
300: PetscCall(TSMonitorSet(ts, TSMonitorDrawSolutionPhase, ctx, (PetscErrorCode(*)(void **))TSMonitorDrawCtxDestroy));
301: }
302: opt = PETSC_FALSE;
303: PetscCall(PetscOptionsName("-ts_monitor_draw_error", "Monitor error graphically", "TSMonitorDrawError", &opt));
304: if (opt) {
305: TSMonitorDrawCtx ctx;
306: PetscInt howoften = 1;
308: PetscCall(PetscOptionsInt("-ts_monitor_draw_error", "Monitor error graphically", "TSMonitorDrawError", howoften, &howoften, NULL));
309: PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts), NULL, "Error", PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &ctx));
310: PetscCall(TSMonitorSet(ts, TSMonitorDrawError, ctx, (PetscErrorCode(*)(void **))TSMonitorDrawCtxDestroy));
311: }
312: opt = PETSC_FALSE;
313: PetscCall(PetscOptionsName("-ts_monitor_draw_solution_function", "Monitor solution provided by TSMonitorSetSolutionFunction() graphically", "TSMonitorDrawSolutionFunction", &opt));
314: if (opt) {
315: TSMonitorDrawCtx ctx;
316: PetscInt howoften = 1;
318: PetscCall(PetscOptionsInt("-ts_monitor_draw_solution_function", "Monitor solution provided by TSMonitorSetSolutionFunction() graphically", "TSMonitorDrawSolutionFunction", howoften, &howoften, NULL));
319: PetscCall(TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts), NULL, "Solution provided by user function", PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &ctx));
320: PetscCall(TSMonitorSet(ts, TSMonitorDrawSolutionFunction, ctx, (PetscErrorCode(*)(void **))TSMonitorDrawCtxDestroy));
321: }
323: opt = PETSC_FALSE;
324: 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));
325: if (flg) {
326: const char *ptr = NULL, *ptr2 = NULL;
327: char *filetemplate;
328: PetscCheck(monfilename[0], PetscObjectComm((PetscObject)ts), PETSC_ERR_USER, "-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03" PetscInt_FMT ".vts");
329: /* Do some cursory validation of the input. */
330: PetscCall(PetscStrstr(monfilename, "%", (char **)&ptr));
331: PetscCheck(ptr, PetscObjectComm((PetscObject)ts), PETSC_ERR_USER, "-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03" PetscInt_FMT ".vts");
332: for (ptr++; ptr && *ptr; ptr++) {
333: PetscCall(PetscStrchr("DdiouxX", *ptr, (char **)&ptr2));
334: 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");
335: if (ptr2) break;
336: }
337: PetscCall(PetscStrallocpy(monfilename, &filetemplate));
338: PetscCall(TSMonitorSet(ts, TSMonitorSolutionVTK, filetemplate, (PetscErrorCode(*)(void **))TSMonitorSolutionVTKDestroy));
339: }
341: PetscCall(PetscOptionsString("-ts_monitor_dmda_ray", "Display a ray of the solution", "None", "y=0", dir, sizeof(dir), &flg));
342: if (flg) {
343: TSMonitorDMDARayCtx *rayctx;
344: int ray = 0;
345: DMDirection ddir;
346: DM da;
347: PetscMPIInt rank;
349: PetscCheck(dir[1] == '=', PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Unknown ray %s", dir);
350: if (dir[0] == 'x') ddir = DM_X;
351: else if (dir[0] == 'y') ddir = DM_Y;
352: else SETERRQ(PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Unknown ray %s", dir);
353: sscanf(dir + 2, "%d", &ray);
355: PetscCall(PetscInfo(((PetscObject)ts), "Displaying DMDA ray %c = %d\n", dir[0], ray));
356: PetscCall(PetscNew(&rayctx));
357: PetscCall(TSGetDM(ts, &da));
358: PetscCall(DMDAGetRay(da, ddir, ray, &rayctx->ray, &rayctx->scatter));
359: PetscCallMPI(MPI_Comm_rank(PetscObjectComm((PetscObject)ts), &rank));
360: if (rank == 0) PetscCall(PetscViewerDrawOpen(PETSC_COMM_SELF, NULL, NULL, 0, 0, 600, 300, &rayctx->viewer));
361: rayctx->lgctx = NULL;
362: PetscCall(TSMonitorSet(ts, TSMonitorDMDARay, rayctx, TSMonitorDMDARayDestroy));
363: }
364: PetscCall(PetscOptionsString("-ts_monitor_lg_dmda_ray", "Display a ray of the solution", "None", "x=0", dir, sizeof(dir), &flg));
365: if (flg) {
366: TSMonitorDMDARayCtx *rayctx;
367: int ray = 0;
368: DMDirection ddir;
369: DM da;
370: PetscInt howoften = 1;
372: PetscCheck(dir[1] == '=', PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Malformed ray %s", dir);
373: if (dir[0] == 'x') ddir = DM_X;
374: else if (dir[0] == 'y') ddir = DM_Y;
375: else SETERRQ(PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Unknown ray direction %s", dir);
376: sscanf(dir + 2, "%d", &ray);
378: PetscCall(PetscInfo(((PetscObject)ts), "Displaying LG DMDA ray %c = %d\n", dir[0], ray));
379: PetscCall(PetscNew(&rayctx));
380: PetscCall(TSGetDM(ts, &da));
381: PetscCall(DMDAGetRay(da, ddir, ray, &rayctx->ray, &rayctx->scatter));
382: PetscCall(TSMonitorLGCtxCreate(PETSC_COMM_SELF, NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 600, 400, howoften, &rayctx->lgctx));
383: PetscCall(TSMonitorSet(ts, TSMonitorLGDMDARay, rayctx, TSMonitorDMDARayDestroy));
384: }
386: PetscCall(PetscOptionsName("-ts_monitor_envelope", "Monitor maximum and minimum value of each component of the solution", "TSMonitorEnvelope", &opt));
387: if (opt) {
388: TSMonitorEnvelopeCtx ctx;
390: PetscCall(TSMonitorEnvelopeCtxCreate(ts, &ctx));
391: PetscCall(TSMonitorSet(ts, TSMonitorEnvelope, ctx, (PetscErrorCode(*)(void **))TSMonitorEnvelopeCtxDestroy));
392: }
393: flg = PETSC_FALSE;
394: PetscCall(PetscOptionsBool("-ts_monitor_cancel", "Remove all monitors", "TSMonitorCancel", flg, &flg, &opt));
395: if (opt && flg) PetscCall(TSMonitorCancel(ts));
397: flg = PETSC_FALSE;
398: PetscCall(PetscOptionsBool("-ts_fd_color", "Use finite differences with coloring to compute IJacobian", "TSComputeIJacobianDefaultColor", flg, &flg, NULL));
399: if (flg) {
400: DM dm;
402: PetscCall(TSGetDM(ts, &dm));
403: PetscCall(DMTSUnsetIJacobianContext_Internal(dm));
404: PetscCall(TSSetIJacobian(ts, NULL, NULL, TSComputeIJacobianDefaultColor, NULL));
405: PetscCall(PetscInfo(ts, "Setting default finite difference coloring Jacobian matrix\n"));
406: }
408: /* Handle specific TS options */
409: PetscTryTypeMethod(ts, setfromoptions, PetscOptionsObject);
411: /* Handle TSAdapt options */
412: PetscCall(TSGetAdapt(ts, &ts->adapt));
413: PetscCall(TSAdaptSetDefaultType(ts->adapt, ts->default_adapt_type));
414: PetscCall(TSAdaptSetFromOptions(ts->adapt, PetscOptionsObject));
416: /* TS trajectory must be set after TS, since it may use some TS options above */
417: tflg = ts->trajectory ? PETSC_TRUE : PETSC_FALSE;
418: PetscCall(PetscOptionsBool("-ts_save_trajectory", "Save the solution at each timestep", "TSSetSaveTrajectory", tflg, &tflg, NULL));
419: if (tflg) PetscCall(TSSetSaveTrajectory(ts));
421: PetscCall(TSAdjointSetFromOptions(ts, PetscOptionsObject));
423: /* process any options handlers added with PetscObjectAddOptionsHandler() */
424: PetscCall(PetscObjectProcessOptionsHandlers((PetscObject)ts, PetscOptionsObject));
425: PetscOptionsEnd();
427: if (ts->trajectory) PetscCall(TSTrajectorySetFromOptions(ts->trajectory, ts));
429: /* why do we have to do this here and not during TSSetUp? */
430: PetscCall(TSGetSNES(ts, &ts->snes));
431: if (ts->problem_type == TS_LINEAR) {
432: PetscCall(PetscObjectTypeCompareAny((PetscObject)ts->snes, &flg, SNESKSPONLY, SNESKSPTRANSPOSEONLY, ""));
433: if (!flg) PetscCall(SNESSetType(ts->snes, SNESKSPONLY));
434: }
435: PetscCall(SNESSetFromOptions(ts->snes));
436: PetscFunctionReturn(PETSC_SUCCESS);
437: }
439: /*@
440: TSGetTrajectory - Gets the trajectory from a `TS` if it exists
442: Collective
444: Input Parameter:
445: . ts - the `TS` context obtained from `TSCreate()`
447: Output Parameter:
448: . tr - the `TSTrajectory` object, if it exists
450: Level: advanced
452: Note:
453: This routine should be called after all `TS` options have been set
455: .seealso: [](ch_ts), `TS`, `TSTrajectory`, `TSAdjointSolve()`, `TSTrajectoryCreate()`
456: @*/
457: PetscErrorCode TSGetTrajectory(TS ts, TSTrajectory *tr)
458: {
459: PetscFunctionBegin;
461: *tr = ts->trajectory;
462: PetscFunctionReturn(PETSC_SUCCESS);
463: }
465: /*@
466: TSSetSaveTrajectory - Causes the `TS` to save its solutions as it iterates forward in time in a `TSTrajectory` object
468: Collective
470: Input Parameter:
471: . ts - the `TS` context obtained from `TSCreate()`
473: Options Database Keys:
474: + -ts_save_trajectory - saves the trajectory to a file
475: - -ts_trajectory_type type - set trajectory type
477: Level: intermediate
479: Notes:
480: This routine should be called after all `TS` options have been set
482: The `TSTRAJECTORYVISUALIZATION` files can be loaded into Python with $PETSC_DIR/lib/petsc/bin/PetscBinaryIOTrajectory.py and
483: MATLAB with $PETSC_DIR/share/petsc/matlab/PetscReadBinaryTrajectory.m
485: .seealso: [](ch_ts), `TS`, `TSTrajectory`, `TSGetTrajectory()`, `TSAdjointSolve()`
486: @*/
487: PetscErrorCode TSSetSaveTrajectory(TS ts)
488: {
489: PetscFunctionBegin;
491: if (!ts->trajectory) PetscCall(TSTrajectoryCreate(PetscObjectComm((PetscObject)ts), &ts->trajectory));
492: PetscFunctionReturn(PETSC_SUCCESS);
493: }
495: /*@
496: TSResetTrajectory - Destroys and recreates the internal `TSTrajectory` object
498: Collective
500: Input Parameter:
501: . ts - the `TS` context obtained from `TSCreate()`
503: Level: intermediate
505: .seealso: [](ch_ts), `TSTrajectory`, `TSGetTrajectory()`, `TSAdjointSolve()`, `TSRemoveTrajectory()`
506: @*/
507: PetscErrorCode TSResetTrajectory(TS ts)
508: {
509: PetscFunctionBegin;
511: if (ts->trajectory) {
512: PetscCall(TSTrajectoryDestroy(&ts->trajectory));
513: PetscCall(TSTrajectoryCreate(PetscObjectComm((PetscObject)ts), &ts->trajectory));
514: }
515: PetscFunctionReturn(PETSC_SUCCESS);
516: }
518: /*@
519: TSRemoveTrajectory - Destroys and removes the internal `TSTrajectory` object from a `TS`
521: Collective
523: Input Parameter:
524: . ts - the `TS` context obtained from `TSCreate()`
526: Level: intermediate
528: .seealso: [](ch_ts), `TSTrajectory`, `TSResetTrajectory()`, `TSAdjointSolve()`
529: @*/
530: PetscErrorCode TSRemoveTrajectory(TS ts)
531: {
532: PetscFunctionBegin;
534: if (ts->trajectory) PetscCall(TSTrajectoryDestroy(&ts->trajectory));
535: PetscFunctionReturn(PETSC_SUCCESS);
536: }
538: /*@
539: TSComputeRHSJacobian - Computes the Jacobian matrix that has been
540: set with `TSSetRHSJacobian()`.
542: Collective
544: Input Parameters:
545: + ts - the `TS` context
546: . t - current timestep
547: - U - input vector
549: Output Parameters:
550: + A - Jacobian matrix
551: - B - optional preconditioning matrix
553: Level: developer
555: Note:
556: Most users should not need to explicitly call this routine, as it
557: is used internally within the nonlinear solvers.
559: .seealso: [](ch_ts), `TS`, `TSSetRHSJacobian()`, `KSPSetOperators()`
560: @*/
561: PetscErrorCode TSComputeRHSJacobian(TS ts, PetscReal t, Vec U, Mat A, Mat B)
562: {
563: PetscObjectState Ustate;
564: PetscObjectId Uid;
565: DM dm;
566: DMTS tsdm;
567: TSRHSJacobian rhsjacobianfunc;
568: void *ctx;
569: TSRHSFunction rhsfunction;
571: PetscFunctionBegin;
574: PetscCheckSameComm(ts, 1, U, 3);
575: PetscCall(TSGetDM(ts, &dm));
576: PetscCall(DMGetDMTS(dm, &tsdm));
577: PetscCall(DMTSGetRHSFunction(dm, &rhsfunction, NULL));
578: PetscCall(DMTSGetRHSJacobian(dm, &rhsjacobianfunc, &ctx));
579: PetscCall(PetscObjectStateGet((PetscObject)U, &Ustate));
580: PetscCall(PetscObjectGetId((PetscObject)U, &Uid));
582: if (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) && (rhsfunction != TSComputeRHSFunctionLinear)) PetscFunctionReturn(PETSC_SUCCESS);
584: 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);
585: if (rhsjacobianfunc) {
586: PetscCall(PetscLogEventBegin(TS_JacobianEval, ts, U, A, B));
587: PetscCallBack("TS callback Jacobian", (*rhsjacobianfunc)(ts, t, U, A, B, ctx));
588: ts->rhsjacs++;
589: PetscCall(PetscLogEventEnd(TS_JacobianEval, ts, U, A, B));
590: } else {
591: PetscCall(MatZeroEntries(A));
592: if (B && A != B) PetscCall(MatZeroEntries(B));
593: }
594: ts->rhsjacobian.time = t;
595: ts->rhsjacobian.shift = 0;
596: ts->rhsjacobian.scale = 1.;
597: PetscCall(PetscObjectGetId((PetscObject)U, &ts->rhsjacobian.Xid));
598: PetscCall(PetscObjectStateGet((PetscObject)U, &ts->rhsjacobian.Xstate));
599: PetscFunctionReturn(PETSC_SUCCESS);
600: }
602: /*@
603: TSComputeRHSFunction - Evaluates the right-hand-side function for a `TS`
605: Collective
607: Input Parameters:
608: + ts - the `TS` context
609: . t - current time
610: - U - state vector
612: Output Parameter:
613: . y - right hand side
615: Level: developer
617: Note:
618: Most users should not need to explicitly call this routine, as it
619: is used internally within the nonlinear solvers.
621: .seealso: [](ch_ts), `TS`, `TSSetRHSFunction()`, `TSComputeIFunction()`
622: @*/
623: PetscErrorCode TSComputeRHSFunction(TS ts, PetscReal t, Vec U, Vec y)
624: {
625: TSRHSFunction rhsfunction;
626: TSIFunction ifunction;
627: void *ctx;
628: DM dm;
630: PetscFunctionBegin;
634: PetscCall(TSGetDM(ts, &dm));
635: PetscCall(DMTSGetRHSFunction(dm, &rhsfunction, &ctx));
636: PetscCall(DMTSGetIFunction(dm, &ifunction, NULL));
638: PetscCheck(rhsfunction || ifunction, PetscObjectComm((PetscObject)ts), PETSC_ERR_USER, "Must call TSSetRHSFunction() and / or TSSetIFunction()");
640: if (rhsfunction) {
641: PetscCall(PetscLogEventBegin(TS_FunctionEval, ts, U, y, 0));
642: PetscCall(VecLockReadPush(U));
643: PetscCallBack("TS callback right-hand-side", (*rhsfunction)(ts, t, U, y, ctx));
644: PetscCall(VecLockReadPop(U));
645: ts->rhsfuncs++;
646: PetscCall(PetscLogEventEnd(TS_FunctionEval, ts, U, y, 0));
647: } else PetscCall(VecZeroEntries(y));
648: PetscFunctionReturn(PETSC_SUCCESS);
649: }
651: /*@
652: TSComputeSolutionFunction - Evaluates the solution function.
654: Collective
656: Input Parameters:
657: + ts - the `TS` context
658: - t - current time
660: Output Parameter:
661: . U - the solution
663: Level: developer
665: .seealso: [](ch_ts), `TS`, `TSSetSolutionFunction()`, `TSSetRHSFunction()`, `TSComputeIFunction()`
666: @*/
667: PetscErrorCode TSComputeSolutionFunction(TS ts, PetscReal t, Vec U)
668: {
669: TSSolutionFunction solutionfunction;
670: void *ctx;
671: DM dm;
673: PetscFunctionBegin;
676: PetscCall(TSGetDM(ts, &dm));
677: PetscCall(DMTSGetSolutionFunction(dm, &solutionfunction, &ctx));
678: if (solutionfunction) PetscCallBack("TS callback solution", (*solutionfunction)(ts, t, U, ctx));
679: PetscFunctionReturn(PETSC_SUCCESS);
680: }
681: /*@
682: TSComputeForcingFunction - Evaluates the forcing function.
684: Collective
686: Input Parameters:
687: + ts - the `TS` context
688: - t - current time
690: Output Parameter:
691: . U - the function value
693: Level: developer
695: .seealso: [](ch_ts), `TS`, `TSSetSolutionFunction()`, `TSSetRHSFunction()`, `TSComputeIFunction()`
696: @*/
697: PetscErrorCode TSComputeForcingFunction(TS ts, PetscReal t, Vec U)
698: {
699: void *ctx;
700: DM dm;
701: TSForcingFunction forcing;
703: PetscFunctionBegin;
706: PetscCall(TSGetDM(ts, &dm));
707: PetscCall(DMTSGetForcingFunction(dm, &forcing, &ctx));
709: if (forcing) PetscCallBack("TS callback forcing function", (*forcing)(ts, t, U, ctx));
710: PetscFunctionReturn(PETSC_SUCCESS);
711: }
713: static PetscErrorCode TSGetRHSVec_Private(TS ts, Vec *Frhs)
714: {
715: Vec F;
717: PetscFunctionBegin;
718: *Frhs = NULL;
719: PetscCall(TSGetIFunction(ts, &F, NULL, NULL));
720: if (!ts->Frhs) PetscCall(VecDuplicate(F, &ts->Frhs));
721: *Frhs = ts->Frhs;
722: PetscFunctionReturn(PETSC_SUCCESS);
723: }
725: PetscErrorCode TSGetRHSMats_Private(TS ts, Mat *Arhs, Mat *Brhs)
726: {
727: Mat A, B;
728: TSIJacobian ijacobian;
730: PetscFunctionBegin;
731: if (Arhs) *Arhs = NULL;
732: if (Brhs) *Brhs = NULL;
733: PetscCall(TSGetIJacobian(ts, &A, &B, &ijacobian, NULL));
734: if (Arhs) {
735: if (!ts->Arhs) {
736: if (ijacobian) {
737: PetscCall(MatDuplicate(A, MAT_DO_NOT_COPY_VALUES, &ts->Arhs));
738: PetscCall(TSSetMatStructure(ts, SAME_NONZERO_PATTERN));
739: } else {
740: ts->Arhs = A;
741: PetscCall(PetscObjectReference((PetscObject)A));
742: }
743: } else {
744: PetscBool flg;
745: PetscCall(SNESGetUseMatrixFree(ts->snes, NULL, &flg));
746: /* Handle case where user provided only RHSJacobian and used -snes_mf_operator */
747: if (flg && !ijacobian && ts->Arhs == ts->Brhs) {
748: PetscCall(PetscObjectDereference((PetscObject)ts->Arhs));
749: ts->Arhs = A;
750: PetscCall(PetscObjectReference((PetscObject)A));
751: }
752: }
753: *Arhs = ts->Arhs;
754: }
755: if (Brhs) {
756: if (!ts->Brhs) {
757: if (A != B) {
758: if (ijacobian) {
759: PetscCall(MatDuplicate(B, MAT_DO_NOT_COPY_VALUES, &ts->Brhs));
760: } else {
761: ts->Brhs = B;
762: PetscCall(PetscObjectReference((PetscObject)B));
763: }
764: } else {
765: PetscCall(PetscObjectReference((PetscObject)ts->Arhs));
766: ts->Brhs = ts->Arhs;
767: }
768: }
769: *Brhs = ts->Brhs;
770: }
771: PetscFunctionReturn(PETSC_SUCCESS);
772: }
774: /*@
775: TSComputeIFunction - Evaluates the DAE residual written in the implicit form F(t,U,Udot)=0
777: Collective
779: Input Parameters:
780: + ts - the `TS` context
781: . t - current time
782: . U - state vector
783: . Udot - time derivative of state vector
784: - imex - flag indicates if the method is `TSIMEX` so that the RHSFunction should be kept separate
786: Output Parameter:
787: . Y - right hand side
789: Level: developer
791: Note:
792: Most users should not need to explicitly call this routine, as it
793: is used internally within the nonlinear solvers.
795: If the user did did not write their equations in implicit form, this
796: function recasts them in implicit form.
798: .seealso: [](ch_ts), `TS`, `TSSetIFunction()`, `TSComputeRHSFunction()`
799: @*/
800: PetscErrorCode TSComputeIFunction(TS ts, PetscReal t, Vec U, Vec Udot, Vec Y, PetscBool imex)
801: {
802: TSIFunction ifunction;
803: TSRHSFunction rhsfunction;
804: void *ctx;
805: DM dm;
807: PetscFunctionBegin;
813: PetscCall(TSGetDM(ts, &dm));
814: PetscCall(DMTSGetIFunction(dm, &ifunction, &ctx));
815: PetscCall(DMTSGetRHSFunction(dm, &rhsfunction, NULL));
817: PetscCheck(rhsfunction || ifunction, PetscObjectComm((PetscObject)ts), PETSC_ERR_USER, "Must call TSSetRHSFunction() and / or TSSetIFunction()");
819: PetscCall(PetscLogEventBegin(TS_FunctionEval, ts, U, Udot, Y));
820: if (ifunction) {
821: PetscCallBack("TS callback implicit function", (*ifunction)(ts, t, U, Udot, Y, ctx));
822: ts->ifuncs++;
823: }
824: if (imex) {
825: if (!ifunction) PetscCall(VecCopy(Udot, Y));
826: } else if (rhsfunction) {
827: if (ifunction) {
828: Vec Frhs;
829: PetscCall(TSGetRHSVec_Private(ts, &Frhs));
830: PetscCall(TSComputeRHSFunction(ts, t, U, Frhs));
831: PetscCall(VecAXPY(Y, -1, Frhs));
832: } else {
833: PetscCall(TSComputeRHSFunction(ts, t, U, Y));
834: PetscCall(VecAYPX(Y, -1, Udot));
835: }
836: }
837: PetscCall(PetscLogEventEnd(TS_FunctionEval, ts, U, Udot, Y));
838: PetscFunctionReturn(PETSC_SUCCESS);
839: }
841: /*
842: TSRecoverRHSJacobian - Recover the Jacobian matrix so that one can call `TSComputeRHSJacobian()` on it.
844: Note:
845: This routine is needed when one switches from `TSComputeIJacobian()` to `TSComputeRHSJacobian()` because the Jacobian matrix may be shifted or scaled in `TSComputeIJacobian()`.
847: */
848: static PetscErrorCode TSRecoverRHSJacobian(TS ts, Mat A, Mat B)
849: {
850: PetscFunctionBegin;
852: PetscCheck(A == ts->Arhs, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "Invalid Amat");
853: PetscCheck(B == ts->Brhs, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "Invalid Bmat");
855: if (ts->rhsjacobian.shift) PetscCall(MatShift(A, -ts->rhsjacobian.shift));
856: if (ts->rhsjacobian.scale == -1.) PetscCall(MatScale(A, -1));
857: if (B && B == ts->Brhs && A != B) {
858: if (ts->rhsjacobian.shift) PetscCall(MatShift(B, -ts->rhsjacobian.shift));
859: if (ts->rhsjacobian.scale == -1.) PetscCall(MatScale(B, -1));
860: }
861: ts->rhsjacobian.shift = 0;
862: ts->rhsjacobian.scale = 1.;
863: PetscFunctionReturn(PETSC_SUCCESS);
864: }
866: /*@
867: TSComputeIJacobian - Evaluates the Jacobian of the DAE
869: Collective
871: Input Parameters:
872: + ts - the `TS` context
873: . t - current timestep
874: . U - state vector
875: . Udot - time derivative of state vector
876: . shift - shift to apply, see note below
877: - imex - flag indicates if the method is `TSIMEX` so that the RHSJacobian should be kept separate
879: Output Parameters:
880: + A - Jacobian matrix
881: - B - matrix from which the preconditioner is constructed; often the same as `A`
883: Level: developer
885: Notes:
886: If F(t,U,Udot)=0 is the DAE, the required Jacobian is
887: .vb
888: dF/dU + shift*dF/dUdot
889: .ve
890: Most users should not need to explicitly call this routine, as it
891: is used internally within the nonlinear solvers.
893: .seealso: [](ch_ts), `TS`, `TSSetIJacobian()`
894: @*/
895: PetscErrorCode TSComputeIJacobian(TS ts, PetscReal t, Vec U, Vec Udot, PetscReal shift, Mat A, Mat B, PetscBool imex)
896: {
897: TSIJacobian ijacobian;
898: TSRHSJacobian rhsjacobian;
899: DM dm;
900: void *ctx;
902: PetscFunctionBegin;
909: PetscCall(TSGetDM(ts, &dm));
910: PetscCall(DMTSGetIJacobian(dm, &ijacobian, &ctx));
911: PetscCall(DMTSGetRHSJacobian(dm, &rhsjacobian, NULL));
913: PetscCheck(rhsjacobian || ijacobian, PetscObjectComm((PetscObject)ts), PETSC_ERR_USER, "Must call TSSetRHSJacobian() and / or TSSetIJacobian()");
915: PetscCall(PetscLogEventBegin(TS_JacobianEval, ts, U, A, B));
916: if (ijacobian) {
917: PetscCallBack("TS callback implicit Jacobian", (*ijacobian)(ts, t, U, Udot, shift, A, B, ctx));
918: ts->ijacs++;
919: }
920: if (imex) {
921: if (!ijacobian) { /* system was written as Udot = G(t,U) */
922: PetscBool assembled;
923: if (rhsjacobian) {
924: Mat Arhs = NULL;
925: PetscCall(TSGetRHSMats_Private(ts, &Arhs, NULL));
926: if (A == Arhs) {
927: 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 */
928: ts->rhsjacobian.time = PETSC_MIN_REAL;
929: }
930: }
931: PetscCall(MatZeroEntries(A));
932: PetscCall(MatAssembled(A, &assembled));
933: if (!assembled) {
934: PetscCall(MatAssemblyBegin(A, MAT_FINAL_ASSEMBLY));
935: PetscCall(MatAssemblyEnd(A, MAT_FINAL_ASSEMBLY));
936: }
937: PetscCall(MatShift(A, shift));
938: if (A != B) {
939: PetscCall(MatZeroEntries(B));
940: PetscCall(MatAssembled(B, &assembled));
941: if (!assembled) {
942: PetscCall(MatAssemblyBegin(B, MAT_FINAL_ASSEMBLY));
943: PetscCall(MatAssemblyEnd(B, MAT_FINAL_ASSEMBLY));
944: }
945: PetscCall(MatShift(B, shift));
946: }
947: }
948: } else {
949: Mat Arhs = NULL, Brhs = NULL;
951: /* RHSJacobian needs to be converted to part of IJacobian if exists */
952: if (rhsjacobian) PetscCall(TSGetRHSMats_Private(ts, &Arhs, &Brhs));
953: if (Arhs == A) { /* No IJacobian matrix, so we only have the RHS matrix */
954: PetscObjectState Ustate;
955: PetscObjectId Uid;
956: TSRHSFunction rhsfunction;
958: PetscCall(DMTSGetRHSFunction(dm, &rhsfunction, NULL));
959: PetscCall(PetscObjectStateGet((PetscObject)U, &Ustate));
960: PetscCall(PetscObjectGetId((PetscObject)U, &Uid));
961: if ((rhsjacobian == TSComputeRHSJacobianConstant || (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) && rhsfunction != TSComputeRHSFunctionLinear)) &&
962: ts->rhsjacobian.scale == -1.) { /* No need to recompute RHSJacobian */
963: PetscCall(MatShift(A, shift - ts->rhsjacobian.shift)); /* revert the old shift and add the new shift with a single call to MatShift */
964: if (A != B) PetscCall(MatShift(B, shift - ts->rhsjacobian.shift));
965: } else {
966: PetscBool flg;
968: if (ts->rhsjacobian.reuse) { /* Undo the damage */
969: /* MatScale has a short path for this case.
970: However, this code path is taken the first time TSComputeRHSJacobian is called
971: and the matrices have not been assembled yet */
972: PetscCall(TSRecoverRHSJacobian(ts, A, B));
973: }
974: PetscCall(TSComputeRHSJacobian(ts, t, U, A, B));
975: PetscCall(SNESGetUseMatrixFree(ts->snes, NULL, &flg));
976: /* since -snes_mf_operator uses the full SNES function it does not need to be shifted or scaled here */
977: if (!flg) {
978: PetscCall(MatScale(A, -1));
979: PetscCall(MatShift(A, shift));
980: }
981: if (A != B) {
982: PetscCall(MatScale(B, -1));
983: PetscCall(MatShift(B, shift));
984: }
985: }
986: ts->rhsjacobian.scale = -1;
987: ts->rhsjacobian.shift = shift;
988: } else if (Arhs) { /* Both IJacobian and RHSJacobian */
989: if (!ijacobian) { /* No IJacobian provided, but we have a separate RHS matrix */
990: PetscCall(MatZeroEntries(A));
991: PetscCall(MatShift(A, shift));
992: if (A != B) {
993: PetscCall(MatZeroEntries(B));
994: PetscCall(MatShift(B, shift));
995: }
996: }
997: PetscCall(TSComputeRHSJacobian(ts, t, U, Arhs, Brhs));
998: PetscCall(MatAXPY(A, -1, Arhs, ts->axpy_pattern));
999: if (A != B) PetscCall(MatAXPY(B, -1, Brhs, ts->axpy_pattern));
1000: }
1001: }
1002: PetscCall(PetscLogEventEnd(TS_JacobianEval, ts, U, A, B));
1003: PetscFunctionReturn(PETSC_SUCCESS);
1004: }
1006: /*@C
1007: TSSetRHSFunction - Sets the routine for evaluating the function,
1008: where U_t = G(t,u).
1010: Logically Collective
1012: Input Parameters:
1013: + ts - the `TS` context obtained from `TSCreate()`
1014: . r - vector to put the computed right hand side (or `NULL` to have it created)
1015: . f - routine for evaluating the right-hand-side function
1016: - ctx - [optional] user-defined context for private data for the function evaluation routine (may be `NULL`)
1018: Level: beginner
1020: Note:
1021: You must call this function or `TSSetIFunction()` to define your ODE. You cannot use this function when solving a DAE.
1023: .seealso: [](ch_ts), `TS`, `TSRHSFunction`, `TSSetRHSJacobian()`, `TSSetIJacobian()`, `TSSetIFunction()`
1024: @*/
1025: PetscErrorCode TSSetRHSFunction(TS ts, Vec r, TSRHSFunction f, void *ctx)
1026: {
1027: SNES snes;
1028: Vec ralloc = NULL;
1029: DM dm;
1031: PetscFunctionBegin;
1035: PetscCall(TSGetDM(ts, &dm));
1036: PetscCall(DMTSSetRHSFunction(dm, f, ctx));
1037: PetscCall(TSGetSNES(ts, &snes));
1038: if (!r && !ts->dm && ts->vec_sol) {
1039: PetscCall(VecDuplicate(ts->vec_sol, &ralloc));
1040: r = ralloc;
1041: }
1042: PetscCall(SNESSetFunction(snes, r, SNESTSFormFunction, ts));
1043: PetscCall(VecDestroy(&ralloc));
1044: PetscFunctionReturn(PETSC_SUCCESS);
1045: }
1047: /*@C
1048: TSSetSolutionFunction - Provide a function that computes the solution of the ODE or DAE
1050: Logically Collective
1052: Input Parameters:
1053: + ts - the `TS` context obtained from `TSCreate()`
1054: . f - routine for evaluating the solution
1055: - ctx - [optional] user-defined context for private data for the
1056: function evaluation routine (may be `NULL`)
1058: Options Database Keys:
1059: + -ts_monitor_lg_error - create a graphical monitor of error history, requires user to have provided `TSSetSolutionFunction()`
1060: - -ts_monitor_draw_error - Monitor error graphically, requires user to have provided `TSSetSolutionFunction()`
1062: Level: intermediate
1064: Notes:
1065: This routine is used for testing accuracy of time integration schemes when you already know the solution.
1066: If analytic solutions are not known for your system, consider using the Method of Manufactured Solutions to
1067: create closed-form solutions with non-physical forcing terms.
1069: For low-dimensional problems solved in serial, such as small discrete systems, `TSMonitorLGError()` can be used to monitor the error history.
1071: .seealso: [](ch_ts), `TS`, `TSSolutionFunction`, `TSSetRHSJacobian()`, `TSSetIJacobian()`, `TSComputeSolutionFunction()`, `TSSetForcingFunction()`, `TSSetSolution()`, `TSGetSolution()`, `TSMonitorLGError()`, `TSMonitorDrawError()`
1072: @*/
1073: PetscErrorCode TSSetSolutionFunction(TS ts, TSSolutionFunction f, void *ctx)
1074: {
1075: DM dm;
1077: PetscFunctionBegin;
1079: PetscCall(TSGetDM(ts, &dm));
1080: PetscCall(DMTSSetSolutionFunction(dm, f, ctx));
1081: PetscFunctionReturn(PETSC_SUCCESS);
1082: }
1084: /*@C
1085: TSSetForcingFunction - Provide a function that computes a forcing term for a ODE or PDE
1087: Logically Collective
1089: Input Parameters:
1090: + ts - the `TS` context obtained from `TSCreate()`
1091: . func - routine for evaluating the forcing function
1092: - ctx - [optional] user-defined context for private data for the function evaluation routine
1093: (may be `NULL`)
1095: Level: intermediate
1097: Notes:
1098: This routine is useful for testing accuracy of time integration schemes when using the Method of Manufactured Solutions to
1099: create closed-form solutions with a non-physical forcing term. It allows you to use the Method of Manufactored Solution without directly editing the
1100: definition of the problem you are solving and hence possibly introducing bugs.
1102: This replaces the ODE F(u,u_t,t) = 0 the `TS` is solving with F(u,u_t,t) - func(t) = 0
1104: This forcing function does not depend on the solution to the equations, it can only depend on spatial location, time, and possibly parameters, the
1105: parameters can be passed in the ctx variable.
1107: For low-dimensional problems solved in serial, such as small discrete systems, `TSMonitorLGError()` can be used to monitor the error history.
1109: .seealso: [](ch_ts), `TS`, `TSForcingFunction`, `TSSetRHSJacobian()`, `TSSetIJacobian()`,
1110: `TSComputeSolutionFunction()`, `TSSetSolutionFunction()`
1111: @*/
1112: PetscErrorCode TSSetForcingFunction(TS ts, TSForcingFunction func, void *ctx)
1113: {
1114: DM dm;
1116: PetscFunctionBegin;
1118: PetscCall(TSGetDM(ts, &dm));
1119: PetscCall(DMTSSetForcingFunction(dm, func, ctx));
1120: PetscFunctionReturn(PETSC_SUCCESS);
1121: }
1123: /*@C
1124: TSSetRHSJacobian - Sets the function to compute the Jacobian of G,
1125: where U_t = G(U,t), as well as the location to store the matrix.
1127: Logically Collective
1129: Input Parameters:
1130: + ts - the `TS` context obtained from `TSCreate()`
1131: . Amat - (approximate) location to store Jacobian matrix entries computed by `f`
1132: . Pmat - matrix from which preconditioner is to be constructed (usually the same as `Amat`)
1133: . f - the Jacobian evaluation routine
1134: - ctx - [optional] user-defined context for private data for the Jacobian evaluation routine (may be `NULL`)
1136: Level: beginner
1138: Notes:
1139: You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1141: The `TS` solver may modify the nonzero structure and the entries of the matrices `Amat` and `Pmat` between the calls to `f()`
1142: You should not assume the values are the same in the next call to f() as you set them in the previous call.
1144: .seealso: [](ch_ts), `TS`, `TSRHSJacobian`, `SNESComputeJacobianDefaultColor()`,
1145: `TSSetRHSFunction()`, `TSRHSJacobianSetReuse()`, `TSSetIJacobian()`, `TSRHSFunction()`, `TSIFunction()`
1146: @*/
1147: PetscErrorCode TSSetRHSJacobian(TS ts, Mat Amat, Mat Pmat, TSRHSJacobian f, void *ctx)
1148: {
1149: SNES snes;
1150: DM dm;
1151: TSIJacobian ijacobian;
1153: PetscFunctionBegin;
1157: if (Amat) PetscCheckSameComm(ts, 1, Amat, 2);
1158: if (Pmat) PetscCheckSameComm(ts, 1, Pmat, 3);
1160: PetscCall(TSGetDM(ts, &dm));
1161: PetscCall(DMTSSetRHSJacobian(dm, f, ctx));
1162: PetscCall(DMTSGetIJacobian(dm, &ijacobian, NULL));
1163: PetscCall(TSGetSNES(ts, &snes));
1164: if (!ijacobian) PetscCall(SNESSetJacobian(snes, Amat, Pmat, SNESTSFormJacobian, ts));
1165: if (Amat) {
1166: PetscCall(PetscObjectReference((PetscObject)Amat));
1167: PetscCall(MatDestroy(&ts->Arhs));
1168: ts->Arhs = Amat;
1169: }
1170: if (Pmat) {
1171: PetscCall(PetscObjectReference((PetscObject)Pmat));
1172: PetscCall(MatDestroy(&ts->Brhs));
1173: ts->Brhs = Pmat;
1174: }
1175: PetscFunctionReturn(PETSC_SUCCESS);
1176: }
1178: /*@C
1179: TSSetIFunction - Set the function to compute F(t,U,U_t) where F() = 0 is the DAE to be solved.
1181: Logically Collective
1183: Input Parameters:
1184: + ts - the `TS` context obtained from `TSCreate()`
1185: . r - vector to hold the residual (or `NULL` to have it created internally)
1186: . f - the function evaluation routine
1187: - ctx - user-defined context for private data for the function evaluation routine (may be `NULL`)
1189: Level: beginner
1191: Note:
1192: The user MUST call either this routine or `TSSetRHSFunction()` to define the ODE. When solving DAEs you must use this function.
1194: .seealso: [](ch_ts), `TS`, `TSIFunction`, `TSSetRHSJacobian()`, `TSSetRHSFunction()`,
1195: `TSSetIJacobian()`
1196: @*/
1197: PetscErrorCode TSSetIFunction(TS ts, Vec r, TSIFunction f, void *ctx)
1198: {
1199: SNES snes;
1200: Vec ralloc = NULL;
1201: DM dm;
1203: PetscFunctionBegin;
1207: PetscCall(TSGetDM(ts, &dm));
1208: PetscCall(DMTSSetIFunction(dm, f, ctx));
1210: PetscCall(TSGetSNES(ts, &snes));
1211: if (!r && !ts->dm && ts->vec_sol) {
1212: PetscCall(VecDuplicate(ts->vec_sol, &ralloc));
1213: r = ralloc;
1214: }
1215: PetscCall(SNESSetFunction(snes, r, SNESTSFormFunction, ts));
1216: PetscCall(VecDestroy(&ralloc));
1217: PetscFunctionReturn(PETSC_SUCCESS);
1218: }
1220: /*@C
1221: TSGetIFunction - Returns the vector where the implicit residual is stored and the function/context to compute it.
1223: Not Collective
1225: Input Parameter:
1226: . ts - the `TS` context
1228: Output Parameters:
1229: + r - vector to hold residual (or `NULL`)
1230: . func - the function to compute residual (or `NULL`)
1231: - ctx - the function context (or `NULL`)
1233: Level: advanced
1235: .seealso: [](ch_ts), `TS`, `TSSetIFunction()`, `SNESGetFunction()`
1236: @*/
1237: PetscErrorCode TSGetIFunction(TS ts, Vec *r, TSIFunction *func, void **ctx)
1238: {
1239: SNES snes;
1240: DM dm;
1242: PetscFunctionBegin;
1244: PetscCall(TSGetSNES(ts, &snes));
1245: PetscCall(SNESGetFunction(snes, r, NULL, NULL));
1246: PetscCall(TSGetDM(ts, &dm));
1247: PetscCall(DMTSGetIFunction(dm, func, ctx));
1248: PetscFunctionReturn(PETSC_SUCCESS);
1249: }
1251: /*@C
1252: TSGetRHSFunction - Returns the vector where the right hand side is stored and the function/context to compute it.
1254: Not Collective
1256: Input Parameter:
1257: . ts - the `TS` context
1259: Output Parameters:
1260: + r - vector to hold computed right hand side (or `NULL`)
1261: . func - the function to compute right hand side (or `NULL`)
1262: - ctx - the function context (or `NULL`)
1264: Level: advanced
1266: .seealso: [](ch_ts), `TS`, `TSSetRHSFunction()`, `SNESGetFunction()`
1267: @*/
1268: PetscErrorCode TSGetRHSFunction(TS ts, Vec *r, TSRHSFunction *func, void **ctx)
1269: {
1270: SNES snes;
1271: DM dm;
1273: PetscFunctionBegin;
1275: PetscCall(TSGetSNES(ts, &snes));
1276: PetscCall(SNESGetFunction(snes, r, NULL, NULL));
1277: PetscCall(TSGetDM(ts, &dm));
1278: PetscCall(DMTSGetRHSFunction(dm, func, ctx));
1279: PetscFunctionReturn(PETSC_SUCCESS);
1280: }
1282: /*@C
1283: TSSetIJacobian - Set the function to compute the matrix dF/dU + a*dF/dU_t where F(t,U,U_t) is the function
1284: provided with `TSSetIFunction()`.
1286: Logically Collective
1288: Input Parameters:
1289: + ts - the `TS` context obtained from `TSCreate()`
1290: . Amat - (approximate) matrix to store Jacobian entries computed by `f`
1291: . Pmat - matrix used to compute preconditioner (usually the same as `Amat`)
1292: . f - the Jacobian evaluation routine
1293: - ctx - user-defined context for private data for the Jacobian evaluation routine (may be `NULL`)
1295: Level: beginner
1297: Notes:
1298: The matrices `Amat` and `Pmat` are exactly the matrices that are used by `SNES` for the nonlinear solve.
1300: If you know the operator Amat has a null space you can use `MatSetNullSpace()` and `MatSetTransposeNullSpace()` to supply the null
1301: space to `Amat` and the `KSP` solvers will automatically use that null space as needed during the solution process.
1303: The matrix dF/dU + a*dF/dU_t you provide turns out to be
1304: the Jacobian of F(t,U,W+a*U) where F(t,U,U_t) = 0 is the DAE to be solved.
1305: The time integrator internally approximates U_t by W+a*U where the positive "shift"
1306: a and vector W depend on the integration method, step size, and past states. For example with
1307: the backward Euler method a = 1/dt and W = -a*U(previous timestep) so
1308: W + a*U = a*(U - U(previous timestep)) = (U - U(previous timestep))/dt
1310: You must set all the diagonal entries of the matrices, if they are zero you must still set them with a zero value
1312: The TS solver may modify the nonzero structure and the entries of the matrices `Amat` and `Pmat` between the calls to `f`
1313: You should not assume the values are the same in the next call to `f` as you set them in the previous call.
1315: .seealso: [](ch_ts), `TS`, `TSIJacobian`, `TSSetIFunction()`, `TSSetRHSJacobian()`,
1316: `SNESComputeJacobianDefaultColor()`, `SNESComputeJacobianDefault()`, `TSSetRHSFunction()`
1317: @*/
1318: PetscErrorCode TSSetIJacobian(TS ts, Mat Amat, Mat Pmat, TSIJacobian f, void *ctx)
1319: {
1320: SNES snes;
1321: DM dm;
1323: PetscFunctionBegin;
1327: if (Amat) PetscCheckSameComm(ts, 1, Amat, 2);
1328: if (Pmat) PetscCheckSameComm(ts, 1, Pmat, 3);
1330: PetscCall(TSGetDM(ts, &dm));
1331: PetscCall(DMTSSetIJacobian(dm, f, ctx));
1333: PetscCall(TSGetSNES(ts, &snes));
1334: PetscCall(SNESSetJacobian(snes, Amat, Pmat, SNESTSFormJacobian, ts));
1335: PetscFunctionReturn(PETSC_SUCCESS);
1336: }
1338: /*@
1339: TSRHSJacobianSetReuse - restore the RHS Jacobian before calling the user-provided `TSRHSJacobian()` function again
1341: Logically Collective
1343: Input Parameters:
1344: + ts - `TS` context obtained from `TSCreate()`
1345: - reuse - `PETSC_TRUE` if the RHS Jacobian
1347: Level: intermediate
1349: Notes:
1350: Without this flag, `TS` will change the sign and shift the RHS Jacobian for a
1351: finite-time-step implicit solve, in which case the user function will need to recompute the
1352: entire Jacobian. The `reuse `flag must be set if the evaluation function assumes that the
1353: matrix entries have not been changed by the `TS`.
1355: .seealso: [](ch_ts), `TS`, `TSSetRHSJacobian()`, `TSComputeRHSJacobianConstant()`
1356: @*/
1357: PetscErrorCode TSRHSJacobianSetReuse(TS ts, PetscBool reuse)
1358: {
1359: PetscFunctionBegin;
1360: ts->rhsjacobian.reuse = reuse;
1361: PetscFunctionReturn(PETSC_SUCCESS);
1362: }
1364: /*@C
1365: TSSetI2Function - Set the function to compute F(t,U,U_t,U_tt) where F = 0 is the DAE to be solved.
1367: Logically Collective
1369: Input Parameters:
1370: + ts - the `TS` context obtained from `TSCreate()`
1371: . F - vector to hold the residual (or `NULL` to have it created internally)
1372: . fun - the function evaluation routine
1373: - ctx - user-defined context for private data for the function evaluation routine (may be `NULL`)
1375: Level: beginner
1377: .seealso: [](ch_ts), `TS`, `TSI2Function`, `TSSetI2Jacobian()`, `TSSetIFunction()`,
1378: `TSCreate()`, `TSSetRHSFunction()`
1379: @*/
1380: PetscErrorCode TSSetI2Function(TS ts, Vec F, TSI2Function fun, void *ctx)
1381: {
1382: DM dm;
1384: PetscFunctionBegin;
1387: PetscCall(TSSetIFunction(ts, F, NULL, NULL));
1388: PetscCall(TSGetDM(ts, &dm));
1389: PetscCall(DMTSSetI2Function(dm, fun, ctx));
1390: PetscFunctionReturn(PETSC_SUCCESS);
1391: }
1393: /*@C
1394: TSGetI2Function - Returns the vector where the implicit residual is stored and the function/context to compute it.
1396: Not Collective
1398: Input Parameter:
1399: . ts - the `TS` context
1401: Output Parameters:
1402: + r - vector to hold residual (or `NULL`)
1403: . fun - the function to compute residual (or `NULL`)
1404: - ctx - the function context (or `NULL`)
1406: Level: advanced
1408: .seealso: [](ch_ts), `TS`, `TSSetIFunction()`, `SNESGetFunction()`, `TSCreate()`
1409: @*/
1410: PetscErrorCode TSGetI2Function(TS ts, Vec *r, TSI2Function *fun, void **ctx)
1411: {
1412: SNES snes;
1413: DM dm;
1415: PetscFunctionBegin;
1417: PetscCall(TSGetSNES(ts, &snes));
1418: PetscCall(SNESGetFunction(snes, r, NULL, NULL));
1419: PetscCall(TSGetDM(ts, &dm));
1420: PetscCall(DMTSGetI2Function(dm, fun, ctx));
1421: PetscFunctionReturn(PETSC_SUCCESS);
1422: }
1424: /*@C
1425: TSSetI2Jacobian - Set the function to compute the matrix dF/dU + v*dF/dU_t + a*dF/dU_tt
1426: where F(t,U,U_t,U_tt) is the function you provided with `TSSetI2Function()`.
1428: Logically Collective
1430: Input Parameters:
1431: + ts - the `TS` context obtained from `TSCreate()`
1432: . J - matrix to hold the Jacobian values
1433: . P - matrix for constructing the preconditioner (may be same as `J`)
1434: . jac - the Jacobian evaluation routine
1435: - ctx - user-defined context for private data for the Jacobian evaluation routine (may be `NULL`)
1437: Level: beginner
1439: Notes:
1440: The matrices `J` and `P` are exactly the matrices that are used by `SNES` for the nonlinear solve.
1442: The matrix dF/dU + v*dF/dU_t + a*dF/dU_tt you provide turns out to be
1443: 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.
1444: The time integrator internally approximates U_t by W+v*U and U_tt by W'+a*U where the positive "shift"
1445: parameters 'v' and 'a' and vectors W, W' depend on the integration method, step size, and past states.
1447: .seealso: [](ch_ts), `TS`, `TSI2Jacobian`, `TSSetI2Function()`, `TSGetI2Jacobian()`
1448: @*/
1449: PetscErrorCode TSSetI2Jacobian(TS ts, Mat J, Mat P, TSI2Jacobian jac, void *ctx)
1450: {
1451: DM dm;
1453: PetscFunctionBegin;
1457: PetscCall(TSSetIJacobian(ts, J, P, NULL, NULL));
1458: PetscCall(TSGetDM(ts, &dm));
1459: PetscCall(DMTSSetI2Jacobian(dm, jac, ctx));
1460: PetscFunctionReturn(PETSC_SUCCESS);
1461: }
1463: /*@C
1464: TSGetI2Jacobian - Returns the implicit Jacobian at the present timestep.
1466: Not Collective, but parallel objects are returned if `TS` is parallel
1468: Input Parameter:
1469: . ts - The `TS` context obtained from `TSCreate()`
1471: Output Parameters:
1472: + J - The (approximate) Jacobian of F(t,U,U_t,U_tt)
1473: . P - The matrix from which the preconditioner is constructed, often the same as `J`
1474: . jac - The function to compute the Jacobian matrices
1475: - ctx - User-defined context for Jacobian evaluation routine
1477: Level: advanced
1479: Note:
1480: You can pass in `NULL` for any return argument you do not need.
1482: .seealso: [](ch_ts), `TS`, `TSGetTimeStep()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`, `TSSetI2Jacobian()`, `TSGetI2Function()`, `TSCreate()`
1483: @*/
1484: PetscErrorCode TSGetI2Jacobian(TS ts, Mat *J, Mat *P, TSI2Jacobian *jac, void **ctx)
1485: {
1486: SNES snes;
1487: DM dm;
1489: PetscFunctionBegin;
1490: PetscCall(TSGetSNES(ts, &snes));
1491: PetscCall(SNESSetUpMatrices(snes));
1492: PetscCall(SNESGetJacobian(snes, J, P, NULL, NULL));
1493: PetscCall(TSGetDM(ts, &dm));
1494: PetscCall(DMTSGetI2Jacobian(dm, jac, ctx));
1495: PetscFunctionReturn(PETSC_SUCCESS);
1496: }
1498: /*@
1499: TSComputeI2Function - Evaluates the DAE residual written in implicit form F(t,U,U_t,U_tt) = 0
1501: Collective
1503: Input Parameters:
1504: + ts - the `TS` context
1505: . t - current time
1506: . U - state vector
1507: . V - time derivative of state vector (U_t)
1508: - A - second time derivative of state vector (U_tt)
1510: Output Parameter:
1511: . F - the residual vector
1513: Level: developer
1515: Note:
1516: Most users should not need to explicitly call this routine, as it
1517: is used internally within the nonlinear solvers.
1519: .seealso: [](ch_ts), `TS`, `TSSetI2Function()`, `TSGetI2Function()`
1520: @*/
1521: PetscErrorCode TSComputeI2Function(TS ts, PetscReal t, Vec U, Vec V, Vec A, Vec F)
1522: {
1523: DM dm;
1524: TSI2Function I2Function;
1525: void *ctx;
1526: TSRHSFunction rhsfunction;
1528: PetscFunctionBegin;
1535: PetscCall(TSGetDM(ts, &dm));
1536: PetscCall(DMTSGetI2Function(dm, &I2Function, &ctx));
1537: PetscCall(DMTSGetRHSFunction(dm, &rhsfunction, NULL));
1539: if (!I2Function) {
1540: PetscCall(TSComputeIFunction(ts, t, U, A, F, PETSC_FALSE));
1541: PetscFunctionReturn(PETSC_SUCCESS);
1542: }
1544: PetscCall(PetscLogEventBegin(TS_FunctionEval, ts, U, V, F));
1546: PetscCallBack("TS callback implicit function", I2Function(ts, t, U, V, A, F, ctx));
1548: if (rhsfunction) {
1549: Vec Frhs;
1550: PetscCall(TSGetRHSVec_Private(ts, &Frhs));
1551: PetscCall(TSComputeRHSFunction(ts, t, U, Frhs));
1552: PetscCall(VecAXPY(F, -1, Frhs));
1553: }
1555: PetscCall(PetscLogEventEnd(TS_FunctionEval, ts, U, V, F));
1556: PetscFunctionReturn(PETSC_SUCCESS);
1557: }
1559: /*@
1560: TSComputeI2Jacobian - Evaluates the Jacobian of the DAE
1562: Collective
1564: Input Parameters:
1565: + ts - the `TS` context
1566: . t - current timestep
1567: . U - state vector
1568: . V - time derivative of state vector
1569: . A - second time derivative of state vector
1570: . shiftV - shift to apply, see note below
1571: - shiftA - shift to apply, see note below
1573: Output Parameters:
1574: + J - Jacobian matrix
1575: - P - optional preconditioning matrix
1577: Level: developer
1579: Notes:
1580: If F(t,U,V,A)=0 is the DAE, the required Jacobian is
1582: dF/dU + shiftV*dF/dV + shiftA*dF/dA
1584: Most users should not need to explicitly call this routine, as it
1585: is used internally within the nonlinear solvers.
1587: .seealso: [](ch_ts), `TS`, `TSSetI2Jacobian()`
1588: @*/
1589: PetscErrorCode TSComputeI2Jacobian(TS ts, PetscReal t, Vec U, Vec V, Vec A, PetscReal shiftV, PetscReal shiftA, Mat J, Mat P)
1590: {
1591: DM dm;
1592: TSI2Jacobian I2Jacobian;
1593: void *ctx;
1594: TSRHSJacobian rhsjacobian;
1596: PetscFunctionBegin;
1604: PetscCall(TSGetDM(ts, &dm));
1605: PetscCall(DMTSGetI2Jacobian(dm, &I2Jacobian, &ctx));
1606: PetscCall(DMTSGetRHSJacobian(dm, &rhsjacobian, NULL));
1608: if (!I2Jacobian) {
1609: PetscCall(TSComputeIJacobian(ts, t, U, A, shiftA, J, P, PETSC_FALSE));
1610: PetscFunctionReturn(PETSC_SUCCESS);
1611: }
1613: PetscCall(PetscLogEventBegin(TS_JacobianEval, ts, U, J, P));
1614: PetscCallBack("TS callback implicit Jacobian", I2Jacobian(ts, t, U, V, A, shiftV, shiftA, J, P, ctx));
1615: if (rhsjacobian) {
1616: Mat Jrhs, Prhs;
1617: PetscCall(TSGetRHSMats_Private(ts, &Jrhs, &Prhs));
1618: PetscCall(TSComputeRHSJacobian(ts, t, U, Jrhs, Prhs));
1619: PetscCall(MatAXPY(J, -1, Jrhs, ts->axpy_pattern));
1620: if (P != J) PetscCall(MatAXPY(P, -1, Prhs, ts->axpy_pattern));
1621: }
1623: PetscCall(PetscLogEventEnd(TS_JacobianEval, ts, U, J, P));
1624: PetscFunctionReturn(PETSC_SUCCESS);
1625: }
1627: /*@C
1628: TSSetTransientVariable - sets function to transform from state to transient variables
1630: Logically Collective
1632: Input Parameters:
1633: + ts - time stepping context on which to change the transient variable
1634: . tvar - a function that transforms to transient variables
1635: - ctx - a context for tvar
1637: Level: advanced
1639: Notes:
1640: This is typically used to transform from primitive to conservative variables so that a time integrator (e.g., `TSBDF`)
1641: can be conservative. In this context, primitive variables P are used to model the state (e.g., because they lead to
1642: well-conditioned formulations even in limiting cases such as low-Mach or zero porosity). The transient variable is
1643: C(P), specified by calling this function. An IFunction thus receives arguments (P, Cdot) and the IJacobian must be
1644: evaluated via the chain rule, as in
1645: .vb
1646: dF/dP + shift * dF/dCdot dC/dP.
1647: .ve
1649: .seealso: [](ch_ts), `TS`, `TSBDF`, `TSTransientVariable`, `DMTSSetTransientVariable()`, `DMTSGetTransientVariable()`, `TSSetIFunction()`, `TSSetIJacobian()`
1650: @*/
1651: PetscErrorCode TSSetTransientVariable(TS ts, TSTransientVariable tvar, void *ctx)
1652: {
1653: DM dm;
1655: PetscFunctionBegin;
1657: PetscCall(TSGetDM(ts, &dm));
1658: PetscCall(DMTSSetTransientVariable(dm, tvar, ctx));
1659: PetscFunctionReturn(PETSC_SUCCESS);
1660: }
1662: /*@
1663: TSComputeTransientVariable - transforms state (primitive) variables to transient (conservative) variables
1665: Logically Collective
1667: Input Parameters:
1668: + ts - TS on which to compute
1669: - U - state vector to be transformed to transient variables
1671: Output Parameter:
1672: . C - transient (conservative) variable
1674: Level: developer
1676: Developer Notes:
1677: If `DMTSSetTransientVariable()` has not been called, then C is not modified in this routine and C = `NULL` is allowed.
1678: This makes it safe to call without a guard. One can use `TSHasTransientVariable()` to check if transient variables are
1679: being used.
1681: .seealso: [](ch_ts), `TS`, `TSBDF`, `DMTSSetTransientVariable()`, `TSComputeIFunction()`, `TSComputeIJacobian()`
1682: @*/
1683: PetscErrorCode TSComputeTransientVariable(TS ts, Vec U, Vec C)
1684: {
1685: DM dm;
1686: DMTS dmts;
1688: PetscFunctionBegin;
1691: PetscCall(TSGetDM(ts, &dm));
1692: PetscCall(DMGetDMTS(dm, &dmts));
1693: if (dmts->ops->transientvar) {
1695: PetscCall((*dmts->ops->transientvar)(ts, U, C, dmts->transientvarctx));
1696: }
1697: PetscFunctionReturn(PETSC_SUCCESS);
1698: }
1700: /*@
1701: TSHasTransientVariable - determine whether transient variables have been set
1703: Logically Collective
1705: Input Parameter:
1706: . ts - `TS` on which to compute
1708: Output Parameter:
1709: . has - `PETSC_TRUE` if transient variables have been set
1711: Level: developer
1713: .seealso: [](ch_ts), `TS`, `TSBDF`, `DMTSSetTransientVariable()`, `TSComputeTransientVariable()`
1714: @*/
1715: PetscErrorCode TSHasTransientVariable(TS ts, PetscBool *has)
1716: {
1717: DM dm;
1718: DMTS dmts;
1720: PetscFunctionBegin;
1722: PetscCall(TSGetDM(ts, &dm));
1723: PetscCall(DMGetDMTS(dm, &dmts));
1724: *has = dmts->ops->transientvar ? PETSC_TRUE : PETSC_FALSE;
1725: PetscFunctionReturn(PETSC_SUCCESS);
1726: }
1728: /*@
1729: TS2SetSolution - Sets the initial solution and time derivative vectors
1730: for use by the `TS` routines handling second order equations.
1732: Logically Collective
1734: Input Parameters:
1735: + ts - the `TS` context obtained from `TSCreate()`
1736: . u - the solution vector
1737: - v - the time derivative vector
1739: Level: beginner
1741: .seealso: [](ch_ts), `TS`
1742: @*/
1743: PetscErrorCode TS2SetSolution(TS ts, Vec u, Vec v)
1744: {
1745: PetscFunctionBegin;
1749: PetscCall(TSSetSolution(ts, u));
1750: PetscCall(PetscObjectReference((PetscObject)v));
1751: PetscCall(VecDestroy(&ts->vec_dot));
1752: ts->vec_dot = v;
1753: PetscFunctionReturn(PETSC_SUCCESS);
1754: }
1756: /*@
1757: TS2GetSolution - Returns the solution and time derivative at the present timestep
1758: for second order equations.
1760: Not Collective
1762: Input Parameter:
1763: . ts - the `TS` context obtained from `TSCreate()`
1765: Output Parameters:
1766: + u - the vector containing the solution
1767: - v - the vector containing the time derivative
1769: Level: intermediate
1771: Notes:
1772: It is valid to call this routine inside the function
1773: that you are evaluating in order to move to the new timestep. This vector not
1774: changed until the solution at the next timestep has been calculated.
1776: .seealso: [](ch_ts), `TS`, `TS2SetSolution()`, `TSGetTimeStep()`, `TSGetTime()`
1777: @*/
1778: PetscErrorCode TS2GetSolution(TS ts, Vec *u, Vec *v)
1779: {
1780: PetscFunctionBegin;
1782: if (u) PetscAssertPointer(u, 2);
1783: if (v) PetscAssertPointer(v, 3);
1784: if (u) *u = ts->vec_sol;
1785: if (v) *v = ts->vec_dot;
1786: PetscFunctionReturn(PETSC_SUCCESS);
1787: }
1789: /*@C
1790: TSLoad - Loads a `TS` that has been stored in binary with `TSView()`.
1792: Collective
1794: Input Parameters:
1795: + ts - the newly loaded `TS`, this needs to have been created with `TSCreate()` or
1796: some related function before a call to `TSLoad()`.
1797: - viewer - binary file viewer, obtained from `PetscViewerBinaryOpen()`
1799: Level: intermediate
1801: Note:
1802: The type is determined by the data in the file, any type set into the `TS` before this call is ignored.
1804: .seealso: [](ch_ts), `TS`, `PetscViewer`, `PetscViewerBinaryOpen()`, `TSView()`, `MatLoad()`, `VecLoad()`
1805: @*/
1806: PetscErrorCode TSLoad(TS ts, PetscViewer viewer)
1807: {
1808: PetscBool isbinary;
1809: PetscInt classid;
1810: char type[256];
1811: DMTS sdm;
1812: DM dm;
1814: PetscFunctionBegin;
1817: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERBINARY, &isbinary));
1818: PetscCheck(isbinary, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Invalid viewer; open viewer with PetscViewerBinaryOpen()");
1820: PetscCall(PetscViewerBinaryRead(viewer, &classid, 1, NULL, PETSC_INT));
1821: PetscCheck(classid == TS_FILE_CLASSID, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Not TS next in file");
1822: PetscCall(PetscViewerBinaryRead(viewer, type, 256, NULL, PETSC_CHAR));
1823: PetscCall(TSSetType(ts, type));
1824: PetscTryTypeMethod(ts, load, viewer);
1825: PetscCall(DMCreate(PetscObjectComm((PetscObject)ts), &dm));
1826: PetscCall(DMLoad(dm, viewer));
1827: PetscCall(TSSetDM(ts, dm));
1828: PetscCall(DMCreateGlobalVector(ts->dm, &ts->vec_sol));
1829: PetscCall(VecLoad(ts->vec_sol, viewer));
1830: PetscCall(DMGetDMTS(ts->dm, &sdm));
1831: PetscCall(DMTSLoad(sdm, viewer));
1832: PetscFunctionReturn(PETSC_SUCCESS);
1833: }
1835: #include <petscdraw.h>
1836: #if defined(PETSC_HAVE_SAWS)
1837: #include <petscviewersaws.h>
1838: #endif
1840: /*@C
1841: TSViewFromOptions - View a `TS` based on values in the options database
1843: Collective
1845: Input Parameters:
1846: + ts - the `TS` context
1847: . obj - Optional object that provides the prefix for the options database keys
1848: - name - command line option string to be passed by user
1850: Level: intermediate
1852: .seealso: [](ch_ts), `TS`, `TSView`, `PetscObjectViewFromOptions()`, `TSCreate()`
1853: @*/
1854: PetscErrorCode TSViewFromOptions(TS ts, PetscObject obj, const char name[])
1855: {
1856: PetscFunctionBegin;
1858: PetscCall(PetscObjectViewFromOptions((PetscObject)ts, obj, name));
1859: PetscFunctionReturn(PETSC_SUCCESS);
1860: }
1862: /*@C
1863: TSView - Prints the `TS` data structure.
1865: Collective
1867: Input Parameters:
1868: + ts - the `TS` context obtained from `TSCreate()`
1869: - viewer - visualization context
1871: Options Database Key:
1872: . -ts_view - calls `TSView()` at end of `TSStep()`
1874: Level: beginner
1876: Notes:
1877: The available visualization contexts include
1878: + `PETSC_VIEWER_STDOUT_SELF` - standard output (default)
1879: - `PETSC_VIEWER_STDOUT_WORLD` - synchronized standard
1880: output where only the first processor opens
1881: the file. All other processors send their
1882: data to the first processor to print.
1884: The user can open an alternative visualization context with
1885: `PetscViewerASCIIOpen()` - output to a specified file.
1887: In the debugger you can do call `TSView`(ts,0) to display the `TS` solver. (The same holds for any PETSc object viewer).
1889: .seealso: [](ch_ts), `TS`, `PetscViewer`, `PetscViewerASCIIOpen()`
1890: @*/
1891: PetscErrorCode TSView(TS ts, PetscViewer viewer)
1892: {
1893: TSType type;
1894: PetscBool iascii, isstring, isundials, isbinary, isdraw;
1895: DMTS sdm;
1896: #if defined(PETSC_HAVE_SAWS)
1897: PetscBool issaws;
1898: #endif
1900: PetscFunctionBegin;
1902: if (!viewer) PetscCall(PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)ts), &viewer));
1904: PetscCheckSameComm(ts, 1, viewer, 2);
1906: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &iascii));
1907: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERSTRING, &isstring));
1908: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERBINARY, &isbinary));
1909: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERDRAW, &isdraw));
1910: #if defined(PETSC_HAVE_SAWS)
1911: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERSAWS, &issaws));
1912: #endif
1913: if (iascii) {
1914: PetscCall(PetscObjectPrintClassNamePrefixType((PetscObject)ts, viewer));
1915: if (ts->ops->view) {
1916: PetscCall(PetscViewerASCIIPushTab(viewer));
1917: PetscUseTypeMethod(ts, view, viewer);
1918: PetscCall(PetscViewerASCIIPopTab(viewer));
1919: }
1920: if (ts->max_steps < PETSC_MAX_INT) PetscCall(PetscViewerASCIIPrintf(viewer, " maximum steps=%" PetscInt_FMT "\n", ts->max_steps));
1921: if (ts->max_time < PETSC_MAX_REAL) PetscCall(PetscViewerASCIIPrintf(viewer, " maximum time=%g\n", (double)ts->max_time));
1922: if (ts->ifuncs) PetscCall(PetscViewerASCIIPrintf(viewer, " total number of I function evaluations=%" PetscInt_FMT "\n", ts->ifuncs));
1923: if (ts->ijacs) PetscCall(PetscViewerASCIIPrintf(viewer, " total number of I Jacobian evaluations=%" PetscInt_FMT "\n", ts->ijacs));
1924: if (ts->rhsfuncs) PetscCall(PetscViewerASCIIPrintf(viewer, " total number of RHS function evaluations=%" PetscInt_FMT "\n", ts->rhsfuncs));
1925: if (ts->rhsjacs) PetscCall(PetscViewerASCIIPrintf(viewer, " total number of RHS Jacobian evaluations=%" PetscInt_FMT "\n", ts->rhsjacs));
1926: if (ts->usessnes) {
1927: PetscBool lin;
1928: if (ts->problem_type == TS_NONLINEAR) PetscCall(PetscViewerASCIIPrintf(viewer, " total number of nonlinear solver iterations=%" PetscInt_FMT "\n", ts->snes_its));
1929: PetscCall(PetscViewerASCIIPrintf(viewer, " total number of linear solver iterations=%" PetscInt_FMT "\n", ts->ksp_its));
1930: PetscCall(PetscObjectTypeCompareAny((PetscObject)ts->snes, &lin, SNESKSPONLY, SNESKSPTRANSPOSEONLY, ""));
1931: PetscCall(PetscViewerASCIIPrintf(viewer, " total number of %slinear solve failures=%" PetscInt_FMT "\n", lin ? "" : "non", ts->num_snes_failures));
1932: }
1933: PetscCall(PetscViewerASCIIPrintf(viewer, " total number of rejected steps=%" PetscInt_FMT "\n", ts->reject));
1934: if (ts->vrtol) PetscCall(PetscViewerASCIIPrintf(viewer, " using vector of relative error tolerances, "));
1935: else PetscCall(PetscViewerASCIIPrintf(viewer, " using relative error tolerance of %g, ", (double)ts->rtol));
1936: if (ts->vatol) PetscCall(PetscViewerASCIIPrintf(viewer, " using vector of absolute error tolerances\n"));
1937: else PetscCall(PetscViewerASCIIPrintf(viewer, " using absolute error tolerance of %g\n", (double)ts->atol));
1938: PetscCall(PetscViewerASCIIPushTab(viewer));
1939: PetscCall(TSAdaptView(ts->adapt, viewer));
1940: PetscCall(PetscViewerASCIIPopTab(viewer));
1941: } else if (isstring) {
1942: PetscCall(TSGetType(ts, &type));
1943: PetscCall(PetscViewerStringSPrintf(viewer, " TSType: %-7.7s", type));
1944: PetscTryTypeMethod(ts, view, viewer);
1945: } else if (isbinary) {
1946: PetscInt classid = TS_FILE_CLASSID;
1947: MPI_Comm comm;
1948: PetscMPIInt rank;
1949: char type[256];
1951: PetscCall(PetscObjectGetComm((PetscObject)ts, &comm));
1952: PetscCallMPI(MPI_Comm_rank(comm, &rank));
1953: if (rank == 0) {
1954: PetscCall(PetscViewerBinaryWrite(viewer, &classid, 1, PETSC_INT));
1955: PetscCall(PetscStrncpy(type, ((PetscObject)ts)->type_name, 256));
1956: PetscCall(PetscViewerBinaryWrite(viewer, type, 256, PETSC_CHAR));
1957: }
1958: PetscTryTypeMethod(ts, view, viewer);
1959: if (ts->adapt) PetscCall(TSAdaptView(ts->adapt, viewer));
1960: PetscCall(DMView(ts->dm, viewer));
1961: PetscCall(VecView(ts->vec_sol, viewer));
1962: PetscCall(DMGetDMTS(ts->dm, &sdm));
1963: PetscCall(DMTSView(sdm, viewer));
1964: } else if (isdraw) {
1965: PetscDraw draw;
1966: char str[36];
1967: PetscReal x, y, bottom, h;
1969: PetscCall(PetscViewerDrawGetDraw(viewer, 0, &draw));
1970: PetscCall(PetscDrawGetCurrentPoint(draw, &x, &y));
1971: PetscCall(PetscStrncpy(str, "TS: ", sizeof(str)));
1972: PetscCall(PetscStrlcat(str, ((PetscObject)ts)->type_name, sizeof(str)));
1973: PetscCall(PetscDrawStringBoxed(draw, x, y, PETSC_DRAW_BLACK, PETSC_DRAW_BLACK, str, NULL, &h));
1974: bottom = y - h;
1975: PetscCall(PetscDrawPushCurrentPoint(draw, x, bottom));
1976: PetscTryTypeMethod(ts, view, viewer);
1977: if (ts->adapt) PetscCall(TSAdaptView(ts->adapt, viewer));
1978: if (ts->snes) PetscCall(SNESView(ts->snes, viewer));
1979: PetscCall(PetscDrawPopCurrentPoint(draw));
1980: #if defined(PETSC_HAVE_SAWS)
1981: } else if (issaws) {
1982: PetscMPIInt rank;
1983: const char *name;
1985: PetscCall(PetscObjectGetName((PetscObject)ts, &name));
1986: PetscCallMPI(MPI_Comm_rank(PETSC_COMM_WORLD, &rank));
1987: if (!((PetscObject)ts)->amsmem && rank == 0) {
1988: char dir[1024];
1990: PetscCall(PetscObjectViewSAWs((PetscObject)ts, viewer));
1991: PetscCall(PetscSNPrintf(dir, 1024, "/PETSc/Objects/%s/time_step", name));
1992: PetscCallSAWs(SAWs_Register, (dir, &ts->steps, 1, SAWs_READ, SAWs_INT));
1993: PetscCall(PetscSNPrintf(dir, 1024, "/PETSc/Objects/%s/time", name));
1994: PetscCallSAWs(SAWs_Register, (dir, &ts->ptime, 1, SAWs_READ, SAWs_DOUBLE));
1995: }
1996: PetscTryTypeMethod(ts, view, viewer);
1997: #endif
1998: }
1999: if (ts->snes && ts->usessnes) {
2000: PetscCall(PetscViewerASCIIPushTab(viewer));
2001: PetscCall(SNESView(ts->snes, viewer));
2002: PetscCall(PetscViewerASCIIPopTab(viewer));
2003: }
2004: PetscCall(DMGetDMTS(ts->dm, &sdm));
2005: PetscCall(DMTSView(sdm, viewer));
2007: PetscCall(PetscViewerASCIIPushTab(viewer));
2008: PetscCall(PetscObjectTypeCompare((PetscObject)ts, TSSUNDIALS, &isundials));
2009: PetscCall(PetscViewerASCIIPopTab(viewer));
2010: PetscFunctionReturn(PETSC_SUCCESS);
2011: }
2013: /*@
2014: TSSetApplicationContext - Sets an optional user-defined context for
2015: the timesteppers.
2017: Logically Collective
2019: Input Parameters:
2020: + ts - the `TS` context obtained from `TSCreate()`
2021: - usrP - user context
2023: Level: intermediate
2025: Fortran Notes:
2026: You must write a Fortran interface definition for this
2027: function that tells Fortran the Fortran derived data type that you are passing in as the `ctx` argument.
2029: .seealso: [](ch_ts), `TS`, `TSGetApplicationContext()`
2030: @*/
2031: PetscErrorCode TSSetApplicationContext(TS ts, void *usrP)
2032: {
2033: PetscFunctionBegin;
2035: ts->user = usrP;
2036: PetscFunctionReturn(PETSC_SUCCESS);
2037: }
2039: /*@
2040: TSGetApplicationContext - Gets the user-defined context for the
2041: timestepper that was set with `TSSetApplicationContext()`
2043: Not Collective
2045: Input Parameter:
2046: . ts - the `TS` context obtained from `TSCreate()`
2048: Output Parameter:
2049: . usrP - user context
2051: Level: intermediate
2053: Fortran Notes:
2054: You must write a Fortran interface definition for this
2055: function that tells Fortran the Fortran derived data type that you are passing in as the `ctx` argument.
2057: .seealso: [](ch_ts), `TS`, `TSSetApplicationContext()`
2058: @*/
2059: PetscErrorCode TSGetApplicationContext(TS ts, void *usrP)
2060: {
2061: PetscFunctionBegin;
2063: *(void **)usrP = ts->user;
2064: PetscFunctionReturn(PETSC_SUCCESS);
2065: }
2067: /*@
2068: TSGetStepNumber - Gets the number of time steps completed.
2070: Not Collective
2072: Input Parameter:
2073: . ts - the `TS` context obtained from `TSCreate()`
2075: Output Parameter:
2076: . steps - number of steps completed so far
2078: Level: intermediate
2080: .seealso: [](ch_ts), `TS`, `TSGetTime()`, `TSGetTimeStep()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostStage()`, `TSSetPostStep()`
2081: @*/
2082: PetscErrorCode TSGetStepNumber(TS ts, PetscInt *steps)
2083: {
2084: PetscFunctionBegin;
2086: PetscAssertPointer(steps, 2);
2087: *steps = ts->steps;
2088: PetscFunctionReturn(PETSC_SUCCESS);
2089: }
2091: /*@
2092: TSSetStepNumber - Sets the number of steps completed.
2094: Logically Collective
2096: Input Parameters:
2097: + ts - the `TS` context
2098: - steps - number of steps completed so far
2100: Level: developer
2102: Note:
2103: For most uses of the `TS` solvers the user need not explicitly call
2104: `TSSetStepNumber()`, as the step counter is appropriately updated in
2105: `TSSolve()`/`TSStep()`/`TSRollBack()`. Power users may call this routine to
2106: reinitialize timestepping by setting the step counter to zero (and time
2107: to the initial time) to solve a similar problem with different initial
2108: conditions or parameters. Other possible use case is to continue
2109: timestepping from a previously interrupted run in such a way that `TS`
2110: monitors will be called with a initial nonzero step counter.
2112: .seealso: [](ch_ts), `TS`, `TSGetStepNumber()`, `TSSetTime()`, `TSSetTimeStep()`, `TSSetSolution()`
2113: @*/
2114: PetscErrorCode TSSetStepNumber(TS ts, PetscInt steps)
2115: {
2116: PetscFunctionBegin;
2119: PetscCheck(steps >= 0, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_OUTOFRANGE, "Step number must be non-negative");
2120: ts->steps = steps;
2121: PetscFunctionReturn(PETSC_SUCCESS);
2122: }
2124: /*@
2125: TSSetTimeStep - Allows one to reset the timestep at any time,
2126: useful for simple pseudo-timestepping codes.
2128: Logically Collective
2130: Input Parameters:
2131: + ts - the `TS` context obtained from `TSCreate()`
2132: - time_step - the size of the timestep
2134: Level: intermediate
2136: .seealso: [](ch_ts), `TS`, `TSPSEUDO`, `TSGetTimeStep()`, `TSSetTime()`
2137: @*/
2138: PetscErrorCode TSSetTimeStep(TS ts, PetscReal time_step)
2139: {
2140: PetscFunctionBegin;
2143: ts->time_step = time_step;
2144: PetscFunctionReturn(PETSC_SUCCESS);
2145: }
2147: /*@
2148: TSSetExactFinalTime - Determines whether to adapt the final time step to
2149: match the exact final time, interpolate solution to the exact final time,
2150: or just return at the final time `TS` computed.
2152: Logically Collective
2154: Input Parameters:
2155: + ts - the time-step context
2156: - eftopt - exact final time option
2157: .vb
2158: TS_EXACTFINALTIME_STEPOVER - Don't do anything if final time is exceeded
2159: TS_EXACTFINALTIME_INTERPOLATE - Interpolate back to final time
2160: TS_EXACTFINALTIME_MATCHSTEP - Adapt final time step to match the final time
2161: .ve
2163: Options Database Key:
2164: . -ts_exact_final_time <stepover,interpolate,matchstep> - select the final step at runtime
2166: Level: beginner
2168: Note:
2169: If you use the option `TS_EXACTFINALTIME_STEPOVER` the solution may be at a very different time
2170: then the final time you selected.
2172: .seealso: [](ch_ts), `TS`, `TSExactFinalTimeOption`, `TSGetExactFinalTime()`
2173: @*/
2174: PetscErrorCode TSSetExactFinalTime(TS ts, TSExactFinalTimeOption eftopt)
2175: {
2176: PetscFunctionBegin;
2179: ts->exact_final_time = eftopt;
2180: PetscFunctionReturn(PETSC_SUCCESS);
2181: }
2183: /*@
2184: TSGetExactFinalTime - Gets the exact final time option set with `TSSetExactFinalTime()`
2186: Not Collective
2188: Input Parameter:
2189: . ts - the `TS` context
2191: Output Parameter:
2192: . eftopt - exact final time option
2194: Level: beginner
2196: .seealso: [](ch_ts), `TS`, `TSExactFinalTimeOption`, `TSSetExactFinalTime()`
2197: @*/
2198: PetscErrorCode TSGetExactFinalTime(TS ts, TSExactFinalTimeOption *eftopt)
2199: {
2200: PetscFunctionBegin;
2202: PetscAssertPointer(eftopt, 2);
2203: *eftopt = ts->exact_final_time;
2204: PetscFunctionReturn(PETSC_SUCCESS);
2205: }
2207: /*@
2208: TSGetTimeStep - Gets the current timestep size.
2210: Not Collective
2212: Input Parameter:
2213: . ts - the `TS` context obtained from `TSCreate()`
2215: Output Parameter:
2216: . dt - the current timestep size
2218: Level: intermediate
2220: .seealso: [](ch_ts), `TS`, `TSSetTimeStep()`, `TSGetTime()`
2221: @*/
2222: PetscErrorCode TSGetTimeStep(TS ts, PetscReal *dt)
2223: {
2224: PetscFunctionBegin;
2226: PetscAssertPointer(dt, 2);
2227: *dt = ts->time_step;
2228: PetscFunctionReturn(PETSC_SUCCESS);
2229: }
2231: /*@
2232: TSGetSolution - Returns the solution at the present timestep. It
2233: is valid to call this routine inside the function that you are evaluating
2234: in order to move to the new timestep. This vector not changed until
2235: the solution at the next timestep has been calculated.
2237: Not Collective, but v returned is parallel if ts is parallel
2239: Input Parameter:
2240: . ts - the `TS` context obtained from `TSCreate()`
2242: Output Parameter:
2243: . v - the vector containing the solution
2245: Level: intermediate
2247: Note:
2248: If you used `TSSetExactFinalTime`(ts,`TS_EXACTFINALTIME_MATCHSTEP`); this does not return the solution at the requested
2249: final time. It returns the solution at the next timestep.
2251: .seealso: [](ch_ts), `TS`, `TSGetTimeStep()`, `TSGetTime()`, `TSGetSolveTime()`, `TSGetSolutionComponents()`, `TSSetSolutionFunction()`
2252: @*/
2253: PetscErrorCode TSGetSolution(TS ts, Vec *v)
2254: {
2255: PetscFunctionBegin;
2257: PetscAssertPointer(v, 2);
2258: *v = ts->vec_sol;
2259: PetscFunctionReturn(PETSC_SUCCESS);
2260: }
2262: /*@
2263: TSGetSolutionComponents - Returns any solution components at the present
2264: timestep, if available for the time integration method being used.
2265: Solution components are quantities that share the same size and
2266: structure as the solution vector.
2268: Not Collective, but v returned is parallel if ts is parallel
2270: Input Parameters:
2271: + ts - the `TS` context obtained from `TSCreate()` (input parameter).
2272: . n - If v is `NULL`, then the number of solution components is
2273: returned through n, else the n-th solution component is
2274: returned in v.
2275: - v - the vector containing the n-th solution component
2276: (may be `NULL` to use this function to find out
2277: the number of solutions components).
2279: Level: advanced
2281: .seealso: [](ch_ts), `TS`, `TSGetSolution()`
2282: @*/
2283: PetscErrorCode TSGetSolutionComponents(TS ts, PetscInt *n, Vec *v)
2284: {
2285: PetscFunctionBegin;
2287: if (!ts->ops->getsolutioncomponents) *n = 0;
2288: else PetscUseTypeMethod(ts, getsolutioncomponents, n, v);
2289: PetscFunctionReturn(PETSC_SUCCESS);
2290: }
2292: /*@
2293: TSGetAuxSolution - Returns an auxiliary solution at the present
2294: timestep, if available for the time integration method being used.
2296: Not Collective, but v returned is parallel if ts is parallel
2298: Input Parameters:
2299: + ts - the `TS` context obtained from `TSCreate()` (input parameter).
2300: - v - the vector containing the auxiliary solution
2302: Level: intermediate
2304: .seealso: [](ch_ts), `TS`, `TSGetSolution()`
2305: @*/
2306: PetscErrorCode TSGetAuxSolution(TS ts, Vec *v)
2307: {
2308: PetscFunctionBegin;
2310: if (ts->ops->getauxsolution) PetscUseTypeMethod(ts, getauxsolution, v);
2311: else PetscCall(VecZeroEntries(*v));
2312: PetscFunctionReturn(PETSC_SUCCESS);
2313: }
2315: /*@
2316: TSGetTimeError - Returns the estimated error vector, if the chosen
2317: `TSType` has an error estimation functionality and `TSSetTimeError()` was called
2319: Not Collective, but v returned is parallel if ts is parallel
2321: Input Parameters:
2322: + ts - the `TS` context obtained from `TSCreate()` (input parameter).
2323: . n - current estimate (n=0) or previous one (n=-1)
2324: - v - the vector containing the error (same size as the solution).
2326: Level: intermediate
2328: Note:
2329: MUST call after `TSSetUp()`
2331: .seealso: [](ch_ts), `TSGetSolution()`, `TSSetTimeError()`
2332: @*/
2333: PetscErrorCode TSGetTimeError(TS ts, PetscInt n, Vec *v)
2334: {
2335: PetscFunctionBegin;
2337: if (ts->ops->gettimeerror) PetscUseTypeMethod(ts, gettimeerror, n, v);
2338: else PetscCall(VecZeroEntries(*v));
2339: PetscFunctionReturn(PETSC_SUCCESS);
2340: }
2342: /*@
2343: TSSetTimeError - Sets the estimated error vector, if the chosen
2344: `TSType` has an error estimation functionality. This can be used
2345: to restart such a time integrator with a given error vector.
2347: Not Collective, but v returned is parallel if ts is parallel
2349: Input Parameters:
2350: + ts - the `TS` context obtained from `TSCreate()` (input parameter).
2351: - v - the vector containing the error (same size as the solution).
2353: Level: intermediate
2355: .seealso: [](ch_ts), `TS`, `TSSetSolution()`, `TSGetTimeError()`
2356: @*/
2357: PetscErrorCode TSSetTimeError(TS ts, Vec v)
2358: {
2359: PetscFunctionBegin;
2361: PetscCheck(ts->setupcalled, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONGSTATE, "Must call TSSetUp() first");
2362: PetscTryTypeMethod(ts, settimeerror, v);
2363: PetscFunctionReturn(PETSC_SUCCESS);
2364: }
2366: /* ----- Routines to initialize and destroy a timestepper ---- */
2367: /*@
2368: TSSetProblemType - Sets the type of problem to be solved.
2370: Not collective
2372: Input Parameters:
2373: + ts - The `TS`
2374: - type - One of `TS_LINEAR`, `TS_NONLINEAR` where these types refer to problems of the forms
2375: .vb
2376: U_t - A U = 0 (linear)
2377: U_t - A(t) U = 0 (linear)
2378: F(t,U,U_t) = 0 (nonlinear)
2379: .ve
2381: Level: beginner
2383: .seealso: [](ch_ts), `TSSetUp()`, `TSProblemType`, `TS`
2384: @*/
2385: PetscErrorCode TSSetProblemType(TS ts, TSProblemType type)
2386: {
2387: PetscFunctionBegin;
2389: ts->problem_type = type;
2390: if (type == TS_LINEAR) {
2391: SNES snes;
2392: PetscCall(TSGetSNES(ts, &snes));
2393: PetscCall(SNESSetType(snes, SNESKSPONLY));
2394: }
2395: PetscFunctionReturn(PETSC_SUCCESS);
2396: }
2398: /*@C
2399: TSGetProblemType - Gets the type of problem to be solved.
2401: Not collective
2403: Input Parameter:
2404: . ts - The `TS`
2406: Output Parameter:
2407: . type - One of `TS_LINEAR`, `TS_NONLINEAR` where these types refer to problems of the forms
2408: .vb
2409: M U_t = A U
2410: M(t) U_t = A(t) U
2411: F(t,U,U_t)
2412: .ve
2414: Level: beginner
2416: .seealso: [](ch_ts), `TSSetUp()`, `TSProblemType`, `TS`
2417: @*/
2418: PetscErrorCode TSGetProblemType(TS ts, TSProblemType *type)
2419: {
2420: PetscFunctionBegin;
2422: PetscAssertPointer(type, 2);
2423: *type = ts->problem_type;
2424: PetscFunctionReturn(PETSC_SUCCESS);
2425: }
2427: /*
2428: Attempt to check/preset a default value for the exact final time option. This is needed at the beginning of TSSolve() and in TSSetUp()
2429: */
2430: static PetscErrorCode TSSetExactFinalTimeDefault(TS ts)
2431: {
2432: PetscBool isnone;
2434: PetscFunctionBegin;
2435: PetscCall(TSGetAdapt(ts, &ts->adapt));
2436: PetscCall(TSAdaptSetDefaultType(ts->adapt, ts->default_adapt_type));
2438: PetscCall(PetscObjectTypeCompare((PetscObject)ts->adapt, TSADAPTNONE, &isnone));
2439: if (!isnone && ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) ts->exact_final_time = TS_EXACTFINALTIME_MATCHSTEP;
2440: else if (ts->exact_final_time == TS_EXACTFINALTIME_UNSPECIFIED) ts->exact_final_time = TS_EXACTFINALTIME_INTERPOLATE;
2441: PetscFunctionReturn(PETSC_SUCCESS);
2442: }
2444: /*@
2445: TSSetUp - Sets up the internal data structures for the later use of a timestepper.
2447: Collective
2449: Input Parameter:
2450: . ts - the `TS` context obtained from `TSCreate()`
2452: Level: advanced
2454: Note:
2455: For basic use of the `TS` solvers the user need not explicitly call
2456: `TSSetUp()`, since these actions will automatically occur during
2457: the call to `TSStep()` or `TSSolve()`. However, if one wishes to control this
2458: phase separately, `TSSetUp()` should be called after `TSCreate()`
2459: and optional routines of the form TSSetXXX(), but before `TSStep()` and `TSSolve()`.
2461: .seealso: [](ch_ts), `TSCreate()`, `TS`, `TSStep()`, `TSDestroy()`, `TSSolve()`
2462: @*/
2463: PetscErrorCode TSSetUp(TS ts)
2464: {
2465: DM dm;
2466: PetscErrorCode (*func)(SNES, Vec, Vec, void *);
2467: PetscErrorCode (*jac)(SNES, Vec, Mat, Mat, void *);
2468: TSIFunction ifun;
2469: TSIJacobian ijac;
2470: TSI2Jacobian i2jac;
2471: TSRHSJacobian rhsjac;
2473: PetscFunctionBegin;
2475: if (ts->setupcalled) PetscFunctionReturn(PETSC_SUCCESS);
2477: if (!((PetscObject)ts)->type_name) {
2478: PetscCall(TSGetIFunction(ts, NULL, &ifun, NULL));
2479: PetscCall(TSSetType(ts, ifun ? TSBEULER : TSEULER));
2480: }
2482: if (!ts->vec_sol) {
2483: PetscCheck(ts->dm, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONGSTATE, "Must call TSSetSolution() first");
2484: PetscCall(DMCreateGlobalVector(ts->dm, &ts->vec_sol));
2485: }
2487: if (ts->tspan) {
2488: if (!ts->tspan->vecs_sol) PetscCall(VecDuplicateVecs(ts->vec_sol, ts->tspan->num_span_times, &ts->tspan->vecs_sol));
2489: }
2490: if (!ts->Jacp && ts->Jacprhs) { /* IJacobianP shares the same matrix with RHSJacobianP if only RHSJacobianP is provided */
2491: PetscCall(PetscObjectReference((PetscObject)ts->Jacprhs));
2492: ts->Jacp = ts->Jacprhs;
2493: }
2495: if (ts->quadraturets) {
2496: PetscCall(TSSetUp(ts->quadraturets));
2497: PetscCall(VecDestroy(&ts->vec_costintegrand));
2498: PetscCall(VecDuplicate(ts->quadraturets->vec_sol, &ts->vec_costintegrand));
2499: }
2501: PetscCall(TSGetRHSJacobian(ts, NULL, NULL, &rhsjac, NULL));
2502: if (rhsjac == TSComputeRHSJacobianConstant) {
2503: Mat Amat, Pmat;
2504: SNES snes;
2505: PetscCall(TSGetSNES(ts, &snes));
2506: PetscCall(SNESGetJacobian(snes, &Amat, &Pmat, NULL, NULL));
2507: /* Matching matrices implies that an IJacobian is NOT set, because if it had been set, the IJacobian's matrix would
2508: * have displaced the RHS matrix */
2509: if (Amat && Amat == ts->Arhs) {
2510: /* 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 */
2511: PetscCall(MatDuplicate(ts->Arhs, MAT_COPY_VALUES, &Amat));
2512: PetscCall(SNESSetJacobian(snes, Amat, NULL, NULL, NULL));
2513: PetscCall(MatDestroy(&Amat));
2514: }
2515: if (Pmat && Pmat == ts->Brhs) {
2516: PetscCall(MatDuplicate(ts->Brhs, MAT_COPY_VALUES, &Pmat));
2517: PetscCall(SNESSetJacobian(snes, NULL, Pmat, NULL, NULL));
2518: PetscCall(MatDestroy(&Pmat));
2519: }
2520: }
2522: PetscCall(TSGetAdapt(ts, &ts->adapt));
2523: PetscCall(TSAdaptSetDefaultType(ts->adapt, ts->default_adapt_type));
2525: PetscTryTypeMethod(ts, setup);
2527: PetscCall(TSSetExactFinalTimeDefault(ts));
2529: /* In the case where we've set a DMTSFunction or what have you, we need the default SNESFunction
2530: to be set right but can't do it elsewhere due to the overreliance on ctx=ts.
2531: */
2532: PetscCall(TSGetDM(ts, &dm));
2533: PetscCall(DMSNESGetFunction(dm, &func, NULL));
2534: if (!func) PetscCall(DMSNESSetFunction(dm, SNESTSFormFunction, ts));
2536: /* If the SNES doesn't have a jacobian set and the TS has an ijacobian or rhsjacobian set, set the SNES to use it.
2537: Otherwise, the SNES will use coloring internally to form the Jacobian.
2538: */
2539: PetscCall(DMSNESGetJacobian(dm, &jac, NULL));
2540: PetscCall(DMTSGetIJacobian(dm, &ijac, NULL));
2541: PetscCall(DMTSGetI2Jacobian(dm, &i2jac, NULL));
2542: PetscCall(DMTSGetRHSJacobian(dm, &rhsjac, NULL));
2543: if (!jac && (ijac || i2jac || rhsjac)) PetscCall(DMSNESSetJacobian(dm, SNESTSFormJacobian, ts));
2545: /* if time integration scheme has a starting method, call it */
2546: PetscTryTypeMethod(ts, startingmethod);
2548: ts->setupcalled = PETSC_TRUE;
2549: PetscFunctionReturn(PETSC_SUCCESS);
2550: }
2552: /*@
2553: TSReset - Resets a `TS` context and removes any allocated `Vec`s and `Mat`s.
2555: Collective
2557: Input Parameter:
2558: . ts - the `TS` context obtained from `TSCreate()`
2560: Level: beginner
2562: .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetup()`, `TSDestroy()`
2563: @*/
2564: PetscErrorCode TSReset(TS ts)
2565: {
2566: TS_RHSSplitLink ilink = ts->tsrhssplit, next;
2568: PetscFunctionBegin;
2571: PetscTryTypeMethod(ts, reset);
2572: if (ts->snes) PetscCall(SNESReset(ts->snes));
2573: if (ts->adapt) PetscCall(TSAdaptReset(ts->adapt));
2575: PetscCall(MatDestroy(&ts->Arhs));
2576: PetscCall(MatDestroy(&ts->Brhs));
2577: PetscCall(VecDestroy(&ts->Frhs));
2578: PetscCall(VecDestroy(&ts->vec_sol));
2579: PetscCall(VecDestroy(&ts->vec_dot));
2580: PetscCall(VecDestroy(&ts->vatol));
2581: PetscCall(VecDestroy(&ts->vrtol));
2582: PetscCall(VecDestroyVecs(ts->nwork, &ts->work));
2584: PetscCall(MatDestroy(&ts->Jacprhs));
2585: PetscCall(MatDestroy(&ts->Jacp));
2586: if (ts->forward_solve) PetscCall(TSForwardReset(ts));
2587: if (ts->quadraturets) {
2588: PetscCall(TSReset(ts->quadraturets));
2589: PetscCall(VecDestroy(&ts->vec_costintegrand));
2590: }
2591: while (ilink) {
2592: next = ilink->next;
2593: PetscCall(TSDestroy(&ilink->ts));
2594: PetscCall(PetscFree(ilink->splitname));
2595: PetscCall(ISDestroy(&ilink->is));
2596: PetscCall(PetscFree(ilink));
2597: ilink = next;
2598: }
2599: ts->tsrhssplit = NULL;
2600: ts->num_rhs_splits = 0;
2601: if (ts->tspan) {
2602: PetscCall(PetscFree(ts->tspan->span_times));
2603: PetscCall(VecDestroyVecs(ts->tspan->num_span_times, &ts->tspan->vecs_sol));
2604: PetscCall(PetscFree(ts->tspan));
2605: }
2606: ts->setupcalled = PETSC_FALSE;
2607: PetscFunctionReturn(PETSC_SUCCESS);
2608: }
2610: static PetscErrorCode TSResizeReset(TS);
2612: /*@C
2613: TSDestroy - Destroys the timestepper context that was created
2614: with `TSCreate()`.
2616: Collective
2618: Input Parameter:
2619: . ts - the `TS` context obtained from `TSCreate()`
2621: Level: beginner
2623: .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetUp()`, `TSSolve()`
2624: @*/
2625: PetscErrorCode TSDestroy(TS *ts)
2626: {
2627: PetscFunctionBegin;
2628: if (!*ts) PetscFunctionReturn(PETSC_SUCCESS);
2630: if (--((PetscObject)(*ts))->refct > 0) {
2631: *ts = NULL;
2632: PetscFunctionReturn(PETSC_SUCCESS);
2633: }
2635: PetscCall(TSReset(*ts));
2636: PetscCall(TSAdjointReset(*ts));
2637: if ((*ts)->forward_solve) PetscCall(TSForwardReset(*ts));
2638: PetscCall(TSResizeReset(*ts));
2640: /* if memory was published with SAWs then destroy it */
2641: PetscCall(PetscObjectSAWsViewOff((PetscObject)*ts));
2642: PetscTryTypeMethod((*ts), destroy);
2644: PetscCall(TSTrajectoryDestroy(&(*ts)->trajectory));
2646: PetscCall(TSAdaptDestroy(&(*ts)->adapt));
2647: PetscCall(TSEventDestroy(&(*ts)->event));
2649: PetscCall(SNESDestroy(&(*ts)->snes));
2650: PetscCall(DMDestroy(&(*ts)->dm));
2651: PetscCall(TSMonitorCancel((*ts)));
2652: PetscCall(TSAdjointMonitorCancel((*ts)));
2654: PetscCall(TSDestroy(&(*ts)->quadraturets));
2655: PetscCall(PetscHeaderDestroy(ts));
2656: PetscFunctionReturn(PETSC_SUCCESS);
2657: }
2659: /*@
2660: TSGetSNES - Returns the `SNES` (nonlinear solver) associated with
2661: a `TS` (timestepper) context. Valid only for nonlinear problems.
2663: Not Collective, but snes is parallel if ts is parallel
2665: Input Parameter:
2666: . ts - the `TS` context obtained from `TSCreate()`
2668: Output Parameter:
2669: . snes - the nonlinear solver context
2671: Level: beginner
2673: Notes:
2674: The user can then directly manipulate the `SNES` context to set various
2675: options, etc. Likewise, the user can then extract and manipulate the
2676: `KSP`, and `PC` contexts as well.
2678: `TSGetSNES()` does not work for integrators that do not use `SNES`; in
2679: this case `TSGetSNES()` returns `NULL` in `snes`.
2681: .seealso: [](ch_ts), `TS`, `SNES`, `TSCreate()`, `TSSetUp()`, `TSSolve()`
2682: @*/
2683: PetscErrorCode TSGetSNES(TS ts, SNES *snes)
2684: {
2685: PetscFunctionBegin;
2687: PetscAssertPointer(snes, 2);
2688: if (!ts->snes) {
2689: PetscCall(SNESCreate(PetscObjectComm((PetscObject)ts), &ts->snes));
2690: PetscCall(PetscObjectSetOptions((PetscObject)ts->snes, ((PetscObject)ts)->options));
2691: PetscCall(SNESSetFunction(ts->snes, NULL, SNESTSFormFunction, ts));
2692: PetscCall(PetscObjectIncrementTabLevel((PetscObject)ts->snes, (PetscObject)ts, 1));
2693: if (ts->dm) PetscCall(SNESSetDM(ts->snes, ts->dm));
2694: if (ts->problem_type == TS_LINEAR) PetscCall(SNESSetType(ts->snes, SNESKSPONLY));
2695: }
2696: *snes = ts->snes;
2697: PetscFunctionReturn(PETSC_SUCCESS);
2698: }
2700: /*@
2701: TSSetSNES - Set the `SNES` (nonlinear solver) to be used by the timestepping context
2703: Collective
2705: Input Parameters:
2706: + ts - the `TS` context obtained from `TSCreate()`
2707: - snes - the nonlinear solver context
2709: Level: developer
2711: Note:
2712: Most users should have the `TS` created by calling `TSGetSNES()`
2714: .seealso: [](ch_ts), `TS`, `SNES`, `TSCreate()`, `TSSetUp()`, `TSSolve()`, `TSGetSNES()`
2715: @*/
2716: PetscErrorCode TSSetSNES(TS ts, SNES snes)
2717: {
2718: PetscErrorCode (*func)(SNES, Vec, Mat, Mat, void *);
2720: PetscFunctionBegin;
2723: PetscCall(PetscObjectReference((PetscObject)snes));
2724: PetscCall(SNESDestroy(&ts->snes));
2726: ts->snes = snes;
2728: PetscCall(SNESSetFunction(ts->snes, NULL, SNESTSFormFunction, ts));
2729: PetscCall(SNESGetJacobian(ts->snes, NULL, NULL, &func, NULL));
2730: if (func == SNESTSFormJacobian) PetscCall(SNESSetJacobian(ts->snes, NULL, NULL, SNESTSFormJacobian, ts));
2731: PetscFunctionReturn(PETSC_SUCCESS);
2732: }
2734: /*@
2735: TSGetKSP - Returns the `KSP` (linear solver) associated with
2736: a `TS` (timestepper) context.
2738: Not Collective, but `ksp` is parallel if `ts` is parallel
2740: Input Parameter:
2741: . ts - the `TS` context obtained from `TSCreate()`
2743: Output Parameter:
2744: . ksp - the nonlinear solver context
2746: Level: beginner
2748: Notes:
2749: The user can then directly manipulate the `KSP` context to set various
2750: options, etc. Likewise, the user can then extract and manipulate the
2751: `PC` context as well.
2753: `TSGetKSP()` does not work for integrators that do not use `KSP`;
2754: in this case `TSGetKSP()` returns `NULL` in `ksp`.
2756: .seealso: [](ch_ts), `TS`, `SNES`, `KSP`, `TSCreate()`, `TSSetUp()`, `TSSolve()`, `TSGetSNES()`
2757: @*/
2758: PetscErrorCode TSGetKSP(TS ts, KSP *ksp)
2759: {
2760: SNES snes;
2762: PetscFunctionBegin;
2764: PetscAssertPointer(ksp, 2);
2765: PetscCheck(((PetscObject)ts)->type_name, PETSC_COMM_SELF, PETSC_ERR_ARG_NULL, "KSP is not created yet. Call TSSetType() first");
2766: PetscCheck(ts->problem_type == TS_LINEAR, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Linear only; use TSGetSNES()");
2767: PetscCall(TSGetSNES(ts, &snes));
2768: PetscCall(SNESGetKSP(snes, ksp));
2769: PetscFunctionReturn(PETSC_SUCCESS);
2770: }
2772: /* ----------- Routines to set solver parameters ---------- */
2774: /*@
2775: TSSetMaxSteps - Sets the maximum number of steps to use.
2777: Logically Collective
2779: Input Parameters:
2780: + ts - the `TS` context obtained from `TSCreate()`
2781: - maxsteps - maximum number of steps to use
2783: Options Database Key:
2784: . -ts_max_steps <maxsteps> - Sets maxsteps
2786: Level: intermediate
2788: Note:
2789: The default maximum number of steps is 5000
2791: .seealso: [](ch_ts), `TS`, `TSGetMaxSteps()`, `TSSetMaxTime()`, `TSSetExactFinalTime()`
2792: @*/
2793: PetscErrorCode TSSetMaxSteps(TS ts, PetscInt maxsteps)
2794: {
2795: PetscFunctionBegin;
2798: PetscCheck(maxsteps >= 0, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_OUTOFRANGE, "Maximum number of steps must be non-negative");
2799: ts->max_steps = maxsteps;
2800: PetscFunctionReturn(PETSC_SUCCESS);
2801: }
2803: /*@
2804: TSGetMaxSteps - Gets the maximum number of steps to use.
2806: Not Collective
2808: Input Parameter:
2809: . ts - the `TS` context obtained from `TSCreate()`
2811: Output Parameter:
2812: . maxsteps - maximum number of steps to use
2814: Level: advanced
2816: .seealso: [](ch_ts), `TS`, `TSSetMaxSteps()`, `TSGetMaxTime()`, `TSSetMaxTime()`
2817: @*/
2818: PetscErrorCode TSGetMaxSteps(TS ts, PetscInt *maxsteps)
2819: {
2820: PetscFunctionBegin;
2822: PetscAssertPointer(maxsteps, 2);
2823: *maxsteps = ts->max_steps;
2824: PetscFunctionReturn(PETSC_SUCCESS);
2825: }
2827: /*@
2828: TSSetMaxTime - Sets the maximum (or final) time for timestepping.
2830: Logically Collective
2832: Input Parameters:
2833: + ts - the `TS` context obtained from `TSCreate()`
2834: - maxtime - final time to step to
2836: Options Database Key:
2837: . -ts_max_time <maxtime> - Sets maxtime
2839: Level: intermediate
2841: Notes:
2842: The default maximum time is 5.0
2844: .seealso: [](ch_ts), `TS`, `TSGetMaxTime()`, `TSSetMaxSteps()`, `TSSetExactFinalTime()`
2845: @*/
2846: PetscErrorCode TSSetMaxTime(TS ts, PetscReal maxtime)
2847: {
2848: PetscFunctionBegin;
2851: ts->max_time = maxtime;
2852: PetscFunctionReturn(PETSC_SUCCESS);
2853: }
2855: /*@
2856: TSGetMaxTime - Gets the maximum (or final) time for timestepping.
2858: Not Collective
2860: Input Parameter:
2861: . ts - the `TS` context obtained from `TSCreate()`
2863: Output Parameter:
2864: . maxtime - final time to step to
2866: Level: advanced
2868: .seealso: [](ch_ts), `TS`, `TSSetMaxTime()`, `TSGetMaxSteps()`, `TSSetMaxSteps()`
2869: @*/
2870: PetscErrorCode TSGetMaxTime(TS ts, PetscReal *maxtime)
2871: {
2872: PetscFunctionBegin;
2874: PetscAssertPointer(maxtime, 2);
2875: *maxtime = ts->max_time;
2876: PetscFunctionReturn(PETSC_SUCCESS);
2877: }
2879: // PetscClangLinter pragma disable: -fdoc-*
2880: /*@
2881: TSSetInitialTimeStep - Deprecated, use `TSSetTime()` and `TSSetTimeStep()`.
2883: Level: deprecated
2885: @*/
2886: PetscErrorCode TSSetInitialTimeStep(TS ts, PetscReal initial_time, PetscReal time_step)
2887: {
2888: PetscFunctionBegin;
2890: PetscCall(TSSetTime(ts, initial_time));
2891: PetscCall(TSSetTimeStep(ts, time_step));
2892: PetscFunctionReturn(PETSC_SUCCESS);
2893: }
2895: // PetscClangLinter pragma disable: -fdoc-*
2896: /*@
2897: TSGetDuration - Deprecated, use `TSGetMaxSteps()` and `TSGetMaxTime()`.
2899: Level: deprecated
2901: @*/
2902: PetscErrorCode TSGetDuration(TS ts, PetscInt *maxsteps, PetscReal *maxtime)
2903: {
2904: PetscFunctionBegin;
2906: if (maxsteps) {
2907: PetscAssertPointer(maxsteps, 2);
2908: *maxsteps = ts->max_steps;
2909: }
2910: if (maxtime) {
2911: PetscAssertPointer(maxtime, 3);
2912: *maxtime = ts->max_time;
2913: }
2914: PetscFunctionReturn(PETSC_SUCCESS);
2915: }
2917: // PetscClangLinter pragma disable: -fdoc-*
2918: /*@
2919: TSSetDuration - Deprecated, use `TSSetMaxSteps()` and `TSSetMaxTime()`.
2921: Level: deprecated
2923: @*/
2924: PetscErrorCode TSSetDuration(TS ts, PetscInt maxsteps, PetscReal maxtime)
2925: {
2926: PetscFunctionBegin;
2930: if (maxsteps >= 0) ts->max_steps = maxsteps;
2931: if (maxtime != (PetscReal)PETSC_DEFAULT) ts->max_time = maxtime;
2932: PetscFunctionReturn(PETSC_SUCCESS);
2933: }
2935: // PetscClangLinter pragma disable: -fdoc-*
2936: /*@
2937: TSGetTimeStepNumber - Deprecated, use `TSGetStepNumber()`.
2939: Level: deprecated
2941: @*/
2942: PetscErrorCode TSGetTimeStepNumber(TS ts, PetscInt *steps)
2943: {
2944: return TSGetStepNumber(ts, steps);
2945: }
2947: // PetscClangLinter pragma disable: -fdoc-*
2948: /*@
2949: TSGetTotalSteps - Deprecated, use `TSGetStepNumber()`.
2951: Level: deprecated
2953: @*/
2954: PetscErrorCode TSGetTotalSteps(TS ts, PetscInt *steps)
2955: {
2956: return TSGetStepNumber(ts, steps);
2957: }
2959: /*@
2960: TSSetSolution - Sets the initial solution vector
2961: for use by the `TS` routines.
2963: Logically Collective
2965: Input Parameters:
2966: + ts - the `TS` context obtained from `TSCreate()`
2967: - u - the solution vector
2969: Level: beginner
2971: .seealso: [](ch_ts), `TS`, `TSSetSolutionFunction()`, `TSGetSolution()`, `TSCreate()`
2972: @*/
2973: PetscErrorCode TSSetSolution(TS ts, Vec u)
2974: {
2975: DM dm;
2977: PetscFunctionBegin;
2980: PetscCall(PetscObjectReference((PetscObject)u));
2981: PetscCall(VecDestroy(&ts->vec_sol));
2982: ts->vec_sol = u;
2984: PetscCall(TSGetDM(ts, &dm));
2985: PetscCall(DMShellSetGlobalVector(dm, u));
2986: PetscFunctionReturn(PETSC_SUCCESS);
2987: }
2989: /*@C
2990: TSSetPreStep - Sets the general-purpose function
2991: called once at the beginning of each time step.
2993: Logically Collective
2995: Input Parameters:
2996: + ts - The `TS` context obtained from `TSCreate()`
2997: - func - The function
2999: Calling sequence of `func`:
3000: . ts - the `TS` context
3002: Level: intermediate
3004: .seealso: [](ch_ts), `TS`, `TSSetPreStage()`, `TSSetPostStage()`, `TSSetPostStep()`, `TSStep()`, `TSRestartStep()`
3005: @*/
3006: PetscErrorCode TSSetPreStep(TS ts, PetscErrorCode (*func)(TS ts))
3007: {
3008: PetscFunctionBegin;
3010: ts->prestep = func;
3011: PetscFunctionReturn(PETSC_SUCCESS);
3012: }
3014: /*@
3015: TSPreStep - Runs the user-defined pre-step function provided with `TSSetPreStep()`
3017: Collective
3019: Input Parameter:
3020: . ts - The `TS` context obtained from `TSCreate()`
3022: Level: developer
3024: Note:
3025: `TSPreStep()` is typically used within time stepping implementations,
3026: so most users would not generally call this routine themselves.
3028: .seealso: [](ch_ts), `TS`, `TSSetPreStep()`, `TSPreStage()`, `TSPostStage()`, `TSPostStep()`
3029: @*/
3030: PetscErrorCode TSPreStep(TS ts)
3031: {
3032: PetscFunctionBegin;
3034: if (ts->prestep) {
3035: Vec U;
3036: PetscObjectId idprev;
3037: PetscBool sameObject;
3038: PetscObjectState sprev, spost;
3040: PetscCall(TSGetSolution(ts, &U));
3041: PetscCall(PetscObjectGetId((PetscObject)U, &idprev));
3042: PetscCall(PetscObjectStateGet((PetscObject)U, &sprev));
3043: PetscCallBack("TS callback preset", (*ts->prestep)(ts));
3044: PetscCall(TSGetSolution(ts, &U));
3045: PetscCall(PetscObjectCompareId((PetscObject)U, idprev, &sameObject));
3046: PetscCall(PetscObjectStateGet((PetscObject)U, &spost));
3047: if (!sameObject || sprev != spost) PetscCall(TSRestartStep(ts));
3048: }
3049: PetscFunctionReturn(PETSC_SUCCESS);
3050: }
3052: /*@C
3053: TSSetPreStage - Sets the general-purpose function
3054: called once at the beginning of each stage.
3056: Logically Collective
3058: Input Parameters:
3059: + ts - The `TS` context obtained from `TSCreate()`
3060: - func - The function
3062: Calling sequence of `func`:
3063: + ts - the `TS` context
3064: - stagetime - the stage time
3066: Level: intermediate
3068: Note:
3069: There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3070: The time step number being computed can be queried using `TSGetStepNumber()` and the total size of the step being
3071: attempted can be obtained using `TSGetTimeStep()`. The time at the start of the step is available via `TSGetTime()`.
3073: .seealso: [](ch_ts), `TS`, `TSSetPostStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3074: @*/
3075: PetscErrorCode TSSetPreStage(TS ts, PetscErrorCode (*func)(TS ts, PetscReal stagetime))
3076: {
3077: PetscFunctionBegin;
3079: ts->prestage = func;
3080: PetscFunctionReturn(PETSC_SUCCESS);
3081: }
3083: /*@C
3084: TSSetPostStage - Sets the general-purpose function, provided with `TSSetPostStep()`,
3085: called once at the end of each stage.
3087: Logically Collective
3089: Input Parameters:
3090: + ts - The `TS` context obtained from `TSCreate()`
3091: - func - The function
3093: Calling sequence of `func`:
3094: + ts - the `TS` context
3095: . stagetime - the stage time
3096: . stageindex - the stage index
3097: - Y - Array of vectors (of size = total number of stages) with the stage solutions
3099: Level: intermediate
3101: Note:
3102: There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried.
3103: The time step number being computed can be queried using `TSGetStepNumber()` and the total size of the step being
3104: attempted can be obtained using `TSGetTimeStep()`. The time at the start of the step is available via `TSGetTime()`.
3106: .seealso: [](ch_ts), `TS`, `TSSetPreStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3107: @*/
3108: PetscErrorCode TSSetPostStage(TS ts, PetscErrorCode (*func)(TS ts, PetscReal stagetime, PetscInt stageindex, Vec *Y))
3109: {
3110: PetscFunctionBegin;
3112: ts->poststage = func;
3113: PetscFunctionReturn(PETSC_SUCCESS);
3114: }
3116: /*@C
3117: TSSetPostEvaluate - Sets the general-purpose function
3118: called once at the end of each step evaluation.
3120: Logically Collective
3122: Input Parameters:
3123: + ts - The `TS` context obtained from `TSCreate()`
3124: - func - The function
3126: Calling sequence of `func`:
3127: . ts - the `TS` context
3129: Level: intermediate
3131: Note:
3132: Semantically, `TSSetPostEvaluate()` differs from `TSSetPostStep()` since the function it sets is called before event-handling
3133: thus guaranteeing the same solution (computed by the time-stepper) will be passed to it. On the other hand, `TSPostStep()`
3134: may be passed a different solution, possibly changed by the event handler. `TSPostEvaluate()` is called after the next step
3135: solution is evaluated allowing to modify it, if need be. The solution can be obtained with `TSGetSolution()`, the time step
3136: with `TSGetTimeStep()`, and the time at the start of the step is available via `TSGetTime()`
3138: .seealso: [](ch_ts), `TS`, `TSSetPreStage()`, `TSSetPreStep()`, `TSSetPostStep()`, `TSGetApplicationContext()`
3139: @*/
3140: PetscErrorCode TSSetPostEvaluate(TS ts, PetscErrorCode (*func)(TS ts))
3141: {
3142: PetscFunctionBegin;
3144: ts->postevaluate = func;
3145: PetscFunctionReturn(PETSC_SUCCESS);
3146: }
3148: /*@
3149: TSPreStage - Runs the user-defined pre-stage function set using `TSSetPreStage()`
3151: Collective
3153: Input Parameters:
3154: + ts - The `TS` context obtained from `TSCreate()`
3155: - stagetime - The absolute time of the current stage
3157: Level: developer
3159: Note:
3160: `TSPreStage()` is typically used within time stepping implementations,
3161: most users would not generally call this routine themselves.
3163: .seealso: [](ch_ts), `TS`, `TSPostStage()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3164: @*/
3165: PetscErrorCode TSPreStage(TS ts, PetscReal stagetime)
3166: {
3167: PetscFunctionBegin;
3169: if (ts->prestage) PetscCallBack("TS callback prestage", (*ts->prestage)(ts, stagetime));
3170: PetscFunctionReturn(PETSC_SUCCESS);
3171: }
3173: /*@
3174: TSPostStage - Runs the user-defined post-stage function set using `TSSetPostStage()`
3176: Collective
3178: Input Parameters:
3179: + ts - The `TS` context obtained from `TSCreate()`
3180: . stagetime - The absolute time of the current stage
3181: . stageindex - Stage number
3182: - Y - Array of vectors (of size = total number of stages) with the stage solutions
3184: Level: developer
3186: Note:
3187: `TSPostStage()` is typically used within time stepping implementations,
3188: most users would not generally call this routine themselves.
3190: .seealso: [](ch_ts), `TS`, `TSPreStage()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3191: @*/
3192: PetscErrorCode TSPostStage(TS ts, PetscReal stagetime, PetscInt stageindex, Vec *Y)
3193: {
3194: PetscFunctionBegin;
3196: if (ts->poststage) PetscCallBack("TS callback poststage", (*ts->poststage)(ts, stagetime, stageindex, Y));
3197: PetscFunctionReturn(PETSC_SUCCESS);
3198: }
3200: /*@
3201: TSPostEvaluate - Runs the user-defined post-evaluate function set using `TSSetPostEvaluate()`
3203: Collective
3205: Input Parameter:
3206: . ts - The `TS` context obtained from `TSCreate()`
3208: Level: developer
3210: Note:
3211: `TSPostEvaluate()` is typically used within time stepping implementations,
3212: most users would not generally call this routine themselves.
3214: .seealso: [](ch_ts), `TS`, `TSSetPostEvaluate()`, `TSSetPreStep()`, `TSPreStep()`, `TSPostStep()`
3215: @*/
3216: PetscErrorCode TSPostEvaluate(TS ts)
3217: {
3218: PetscFunctionBegin;
3220: if (ts->postevaluate) {
3221: Vec U;
3222: PetscObjectState sprev, spost;
3224: PetscCall(TSGetSolution(ts, &U));
3225: PetscCall(PetscObjectStateGet((PetscObject)U, &sprev));
3226: PetscCallBack("TS callback postevaluate", (*ts->postevaluate)(ts));
3227: PetscCall(PetscObjectStateGet((PetscObject)U, &spost));
3228: if (sprev != spost) PetscCall(TSRestartStep(ts));
3229: }
3230: PetscFunctionReturn(PETSC_SUCCESS);
3231: }
3233: /*@C
3234: TSSetPostStep - Sets the general-purpose function
3235: called once at the end of each time step.
3237: Logically Collective
3239: Input Parameters:
3240: + ts - The `TS` context obtained from `TSCreate()`
3241: - func - The function
3243: Calling sequence of `func`:
3244: . ts - the `TS` context
3246: Level: intermediate
3248: Note:
3249: The function set by `TSSetPostStep()` is called after each successful step. The solution vector
3250: obtained by `TSGetSolution()` may be different than that computed at the step end if the event handler
3251: locates an event and `TSPostEvent()` modifies it. Use `TSSetPostEvaluate()` if an unmodified solution is needed instead.
3253: .seealso: [](ch_ts), `TS`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostEvaluate()`, `TSGetTimeStep()`, `TSGetStepNumber()`, `TSGetTime()`, `TSRestartStep()`
3254: @*/
3255: PetscErrorCode TSSetPostStep(TS ts, PetscErrorCode (*func)(TS ts))
3256: {
3257: PetscFunctionBegin;
3259: ts->poststep = func;
3260: PetscFunctionReturn(PETSC_SUCCESS);
3261: }
3263: /*@
3264: TSPostStep - Runs the user-defined post-step function that was set with `TSSetPostStep()`
3266: Collective
3268: Input Parameter:
3269: . ts - The `TS` context obtained from `TSCreate()`
3271: Note:
3272: `TSPostStep()` is typically used within time stepping implementations,
3273: so most users would not generally call this routine themselves.
3275: Level: developer
3277: .seealso: [](ch_ts), `TS`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostEvaluate()`, `TSGetTimeStep()`, `TSGetStepNumber()`, `TSGetTime()`, `TSSetPotsStep()`
3278: @*/
3279: PetscErrorCode TSPostStep(TS ts)
3280: {
3281: PetscFunctionBegin;
3283: if (ts->poststep) {
3284: Vec U;
3285: PetscObjectId idprev;
3286: PetscBool sameObject;
3287: PetscObjectState sprev, spost;
3289: PetscCall(TSGetSolution(ts, &U));
3290: PetscCall(PetscObjectGetId((PetscObject)U, &idprev));
3291: PetscCall(PetscObjectStateGet((PetscObject)U, &sprev));
3292: PetscCallBack("TS callback poststep", (*ts->poststep)(ts));
3293: PetscCall(TSGetSolution(ts, &U));
3294: PetscCall(PetscObjectCompareId((PetscObject)U, idprev, &sameObject));
3295: PetscCall(PetscObjectStateGet((PetscObject)U, &spost));
3296: if (!sameObject || sprev != spost) PetscCall(TSRestartStep(ts));
3297: }
3298: PetscFunctionReturn(PETSC_SUCCESS);
3299: }
3301: /*@
3302: TSInterpolate - Interpolate the solution computed during the previous step to an arbitrary location in the interval
3304: Collective
3306: Input Parameters:
3307: + ts - time stepping context
3308: - t - time to interpolate to
3310: Output Parameter:
3311: . U - state at given time
3313: Level: intermediate
3315: Developer Notes:
3316: `TSInterpolate()` and the storing of previous steps/stages should be generalized to support delay differential equations and continuous adjoints.
3318: .seealso: [](ch_ts), `TS`, `TSSetExactFinalTime()`, `TSSolve()`
3319: @*/
3320: PetscErrorCode TSInterpolate(TS ts, PetscReal t, Vec U)
3321: {
3322: PetscFunctionBegin;
3325: 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);
3326: PetscUseTypeMethod(ts, interpolate, t, U);
3327: PetscFunctionReturn(PETSC_SUCCESS);
3328: }
3330: /*@
3331: TSStep - Steps one time step
3333: Collective
3335: Input Parameter:
3336: . ts - the `TS` context obtained from `TSCreate()`
3338: Level: developer
3340: Notes:
3341: The public interface for the ODE/DAE solvers is `TSSolve()`, you should almost for sure be using that routine and not this routine.
3343: The hook set using `TSSetPreStep()` is called before each attempt to take the step. In general, the time step size may
3344: be changed due to adaptive error controller or solve failures. Note that steps may contain multiple stages.
3346: This may over-step the final time provided in `TSSetMaxTime()` depending on the time-step used. `TSSolve()` interpolates to exactly the
3347: time provided in `TSSetMaxTime()`. One can use `TSInterpolate()` to determine an interpolated solution within the final timestep.
3349: .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetUp()`, `TSDestroy()`, `TSSolve()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSSetPostStage()`, `TSInterpolate()`
3350: @*/
3351: PetscErrorCode TSStep(TS ts)
3352: {
3353: static PetscBool cite = PETSC_FALSE;
3354: PetscReal ptime;
3356: PetscFunctionBegin;
3358: PetscCall(PetscCitationsRegister("@article{tspaper,\n"
3359: " title = {{PETSc/TS}: A Modern Scalable {DAE/ODE} Solver Library},\n"
3360: " author = {Abhyankar, Shrirang and Brown, Jed and Constantinescu, Emil and Ghosh, Debojyoti and Smith, Barry F. and Zhang, Hong},\n"
3361: " journal = {arXiv e-preprints},\n"
3362: " eprint = {1806.01437},\n"
3363: " archivePrefix = {arXiv},\n"
3364: " year = {2018}\n}\n",
3365: &cite));
3366: PetscCall(TSSetUp(ts));
3367: PetscCall(TSTrajectorySetUp(ts->trajectory, ts));
3369: 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>");
3370: 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()");
3371: 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");
3373: if (!ts->steps) ts->ptime_prev = ts->ptime;
3374: ptime = ts->ptime;
3375: ts->ptime_prev_rollback = ts->ptime_prev;
3376: ts->reason = TS_CONVERGED_ITERATING;
3378: PetscCall(PetscLogEventBegin(TS_Step, ts, 0, 0, 0));
3379: PetscUseTypeMethod(ts, step);
3380: PetscCall(PetscLogEventEnd(TS_Step, ts, 0, 0, 0));
3382: 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)
3383: PetscCall(VecCopy(ts->vec_sol, ts->tspan->vecs_sol[ts->tspan->spanctr++]));
3384: if (ts->reason >= 0) {
3385: ts->ptime_prev = ptime;
3386: ts->steps++;
3387: ts->steprollback = PETSC_FALSE;
3388: ts->steprestart = PETSC_FALSE;
3389: }
3390: if (!ts->reason) {
3391: if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
3392: else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
3393: }
3395: if (ts->reason < 0 && ts->errorifstepfailed) {
3396: PetscCall(TSMonitorCancel(ts));
3397: PetscCheck(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]);
3398: SETERRQ(PetscObjectComm((PetscObject)ts), PETSC_ERR_NOT_CONVERGED, "TSStep has failed due to %s", TSConvergedReasons[ts->reason]);
3399: }
3400: PetscFunctionReturn(PETSC_SUCCESS);
3401: }
3403: /*@
3404: TSEvaluateWLTE - Evaluate the weighted local truncation error norm
3405: at the end of a time step with a given order of accuracy.
3407: Collective
3409: Input Parameters:
3410: + ts - time stepping context
3411: - wnormtype - norm type, either `NORM_2` or `NORM_INFINITY`
3413: Input/Output Parameter:
3414: . order - optional, desired order for the error evaluation or `PETSC_DECIDE`;
3415: on output, the actual order of the error evaluation
3417: Output Parameter:
3418: . wlte - the weighted local truncation error norm
3420: Level: advanced
3422: Note:
3423: If the timestepper cannot evaluate the error in a particular step
3424: (eg. in the first step or restart steps after event handling),
3425: this routine returns wlte=-1.0 .
3427: .seealso: [](ch_ts), `TS`, `TSStep()`, `TSAdapt`, `TSErrorWeightedNorm()`
3428: @*/
3429: PetscErrorCode TSEvaluateWLTE(TS ts, NormType wnormtype, PetscInt *order, PetscReal *wlte)
3430: {
3431: PetscFunctionBegin;
3435: if (order) PetscAssertPointer(order, 3);
3437: PetscAssertPointer(wlte, 4);
3438: PetscCheck(wnormtype == NORM_2 || wnormtype == NORM_INFINITY, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "No support for norm type %s", NormTypes[wnormtype]);
3439: PetscUseTypeMethod(ts, evaluatewlte, wnormtype, order, wlte);
3440: PetscFunctionReturn(PETSC_SUCCESS);
3441: }
3443: /*@
3444: TSEvaluateStep - Evaluate the solution at the end of a time step with a given order of accuracy.
3446: Collective
3448: Input Parameters:
3449: + ts - time stepping context
3450: . order - desired order of accuracy
3451: - done - whether the step was evaluated at this order (pass `NULL` to generate an error if not available)
3453: Output Parameter:
3454: . U - state at the end of the current step
3456: Level: advanced
3458: Notes:
3459: This function cannot be called until all stages have been evaluated.
3461: 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.
3463: .seealso: [](ch_ts), `TS`, `TSStep()`, `TSAdapt`
3464: @*/
3465: PetscErrorCode TSEvaluateStep(TS ts, PetscInt order, Vec U, PetscBool *done)
3466: {
3467: PetscFunctionBegin;
3471: PetscUseTypeMethod(ts, evaluatestep, order, U, done);
3472: PetscFunctionReturn(PETSC_SUCCESS);
3473: }
3475: /*@C
3476: TSGetComputeInitialCondition - Get the function used to automatically compute an initial condition for the timestepping.
3478: Not collective
3480: Input Parameter:
3481: . ts - time stepping context
3483: Output Parameter:
3484: . initCondition - The function which computes an initial condition
3486: Calling sequence of `initCondition`:
3487: + ts - The timestepping context
3488: - u - The input vector in which the initial condition is stored
3490: Level: advanced
3492: .seealso: [](ch_ts), `TS`, `TSSetComputeInitialCondition()`, `TSComputeInitialCondition()`
3493: @*/
3494: PetscErrorCode TSGetComputeInitialCondition(TS ts, PetscErrorCode (**initCondition)(TS ts, Vec u))
3495: {
3496: PetscFunctionBegin;
3498: PetscAssertPointer(initCondition, 2);
3499: *initCondition = ts->ops->initcondition;
3500: PetscFunctionReturn(PETSC_SUCCESS);
3501: }
3503: /*@C
3504: TSSetComputeInitialCondition - Set the function used to automatically compute an initial condition for the timestepping.
3506: Logically collective
3508: Input Parameters:
3509: + ts - time stepping context
3510: - initCondition - The function which computes an initial condition
3512: Calling sequence of `initCondition`:
3513: + ts - The timestepping context
3514: - e - The input vector in which the initial condition is to be stored
3516: Level: advanced
3518: .seealso: [](ch_ts), `TS`, `TSGetComputeInitialCondition()`, `TSComputeInitialCondition()`
3519: @*/
3520: PetscErrorCode TSSetComputeInitialCondition(TS ts, PetscErrorCode (*initCondition)(TS ts, Vec e))
3521: {
3522: PetscFunctionBegin;
3525: ts->ops->initcondition = initCondition;
3526: PetscFunctionReturn(PETSC_SUCCESS);
3527: }
3529: /*@
3530: TSComputeInitialCondition - Compute an initial condition for the timestepping using the function previously set with `TSSetComputeInitialCondition()`
3532: Collective
3534: Input Parameters:
3535: + ts - time stepping context
3536: - u - The `Vec` to store the condition in which will be used in `TSSolve()`
3538: Level: advanced
3540: .seealso: [](ch_ts), `TS`, `TSGetComputeInitialCondition()`, `TSSetComputeInitialCondition()`, `TSSolve()`
3541: @*/
3542: PetscErrorCode TSComputeInitialCondition(TS ts, Vec u)
3543: {
3544: PetscFunctionBegin;
3547: PetscTryTypeMethod(ts, initcondition, u);
3548: PetscFunctionReturn(PETSC_SUCCESS);
3549: }
3551: /*@C
3552: TSGetComputeExactError - Get the function used to automatically compute the exact error for the timestepping.
3554: Not collective
3556: Input Parameter:
3557: . ts - time stepping context
3559: Output Parameter:
3560: . exactError - The function which computes the solution error
3562: Calling sequence of `exactError`:
3563: + ts - The timestepping context
3564: . u - The approximate solution vector
3565: - e - The vector in which the error is stored
3567: Level: advanced
3569: .seealso: [](ch_ts), `TS`, `TSComputeExactError()`
3570: @*/
3571: PetscErrorCode TSGetComputeExactError(TS ts, PetscErrorCode (**exactError)(TS ts, Vec u, Vec e))
3572: {
3573: PetscFunctionBegin;
3575: PetscAssertPointer(exactError, 2);
3576: *exactError = ts->ops->exacterror;
3577: PetscFunctionReturn(PETSC_SUCCESS);
3578: }
3580: /*@C
3581: TSSetComputeExactError - Set the function used to automatically compute the exact error for the timestepping.
3583: Logically collective
3585: Input Parameters:
3586: + ts - time stepping context
3587: - exactError - The function which computes the solution error
3589: Calling sequence of `exactError`:
3590: + ts - The timestepping context
3591: . u - The approximate solution vector
3592: - e - The vector in which the error is stored
3594: Level: advanced
3596: .seealso: [](ch_ts), `TS`, `TSGetComputeExactError()`, `TSComputeExactError()`
3597: @*/
3598: PetscErrorCode TSSetComputeExactError(TS ts, PetscErrorCode (*exactError)(TS ts, Vec u, Vec e))
3599: {
3600: PetscFunctionBegin;
3603: ts->ops->exacterror = exactError;
3604: PetscFunctionReturn(PETSC_SUCCESS);
3605: }
3607: /*@
3608: TSComputeExactError - Compute the solution error for the timestepping using the function previously set with `TSSetComputeExactError()`
3610: Collective
3612: Input Parameters:
3613: + ts - time stepping context
3614: . u - The approximate solution
3615: - e - The `Vec` used to store the error
3617: Level: advanced
3619: .seealso: [](ch_ts), `TS`, `TSGetComputeInitialCondition()`, `TSSetComputeInitialCondition()`, `TSSolve()`
3620: @*/
3621: PetscErrorCode TSComputeExactError(TS ts, Vec u, Vec e)
3622: {
3623: PetscFunctionBegin;
3627: PetscTryTypeMethod(ts, exacterror, u, e);
3628: PetscFunctionReturn(PETSC_SUCCESS);
3629: }
3631: /*@C
3632: TSSetResize - Sets the resize callbacks.
3634: Logically Collective
3636: Input Parameters:
3637: + ts - The `TS` context obtained from `TSCreate()`
3638: . setup - The setup function
3639: . transfer - The transfer function
3640: - ctx - [optional] The user-defined context
3642: Calling sequence of `setup`:
3643: + ts - the `TS` context
3644: . step - the current step
3645: . time - the current time
3646: . state - the current vector of state
3647: . resize - (output parameter) `PETSC_TRUE` if need resizing, `PETSC_FALSE` otherwise
3648: - ctx - user defined context
3650: Calling sequence of `transfer`:
3651: + ts - the `TS` context
3652: . nv - the number of vectors to be transferred
3653: . vecsin - array of vectors to be transferred
3654: . vecsout - array of transferred vectors
3655: - ctx - user defined context
3657: Notes:
3658: The `setup` function is called inside `TSSolve()` after `TSPostStep()` at the end of each time step
3659: to determine if the problem size has changed.
3660: If it is the case, the solver will collect the needed vectors that need to be
3661: transferred from the old to the new sizes using `transfer`. These vectors will include the current
3662: solution vector, and other vectors needed by the specific solver used.
3663: For example, `TSBDF` uses previous solutions vectors to solve for the next time step.
3664: Other application specific objects associated with the solver, i.e. Jacobian matrices and `DM`,
3665: will be automatically reset if the sizes are changed and they must be specified again by the user
3666: inside the `transfer` function.
3667: The input and output arrays passed to `transfer` are allocated by PETSc.
3668: Vectors in `vecsout` must be created by the user.
3669: Ownership of vectors in `vecsout` is transferred to PETSc.
3671: Level: advanced
3673: .seealso: [](ch_ts), `TS`, `TSSetDM()`, `TSSetIJacobian()`, `TSSetRHSJacobian()`
3674: @*/
3675: PetscErrorCode TSSetResize(TS ts, PetscErrorCode (*setup)(TS ts, PetscInt step, PetscReal time, Vec state, PetscBool *resize, void *ctx), PetscErrorCode (*transfer)(TS ts, PetscInt nv, Vec vecsin[], Vec vecsout[], void *ctx), void *ctx)
3676: {
3677: PetscFunctionBegin;
3679: ts->resizesetup = setup;
3680: ts->resizetransfer = transfer;
3681: ts->resizectx = ctx;
3682: PetscFunctionReturn(PETSC_SUCCESS);
3683: }
3685: /*
3686: TSResizeRegisterOrRetrieve - Register or import vectors transferred with `TSResize()`.
3688: Collective
3690: Input Parameters:
3691: + ts - The `TS` context obtained from `TSCreate()`
3692: - flg - If `PETSC_TRUE` each TS implementation (e.g. `TSBDF`) will register vectors to be transferred, if `PETSC_FALSE` vectors will be imported from transferred vectors.
3694: Level: developer
3696: Note:
3697: `TSResizeRegisterOrRetrieve()` is declared PETSC_INTERN since it is
3698: used within time stepping implementations,
3699: so most users would not generally call this routine themselves.
3701: .seealso: [](ch_ts), `TS`, `TSSetResize()`
3702: @*/
3703: static PetscErrorCode TSResizeRegisterOrRetrieve(TS ts, PetscBool flg)
3704: {
3705: PetscFunctionBegin;
3707: PetscTryTypeMethod(ts, resizeregister, flg);
3708: /* PetscTryTypeMethod(adapt, resizeregister, flg); */
3709: PetscFunctionReturn(PETSC_SUCCESS);
3710: }
3712: static PetscErrorCode TSResizeReset(TS ts)
3713: {
3714: PetscFunctionBegin;
3716: PetscCall(PetscObjectListDestroy(&ts->resizetransferobjs));
3717: PetscFunctionReturn(PETSC_SUCCESS);
3718: }
3720: static PetscErrorCode TSResizeTransferVecs(TS ts, PetscInt cnt, Vec vecsin[], Vec vecsout[])
3721: {
3722: PetscFunctionBegin;
3725: for (PetscInt i = 0; i < cnt; i++) PetscCall(VecLockReadPush(vecsin[i]));
3726: if (ts->resizetransfer) {
3727: PetscCall(PetscInfo(ts, "Transferring %" PetscInt_FMT " vectors\n", cnt));
3728: PetscCallBack("TS callback resize transfer", (*ts->resizetransfer)(ts, cnt, vecsin, vecsout, ts->resizectx));
3729: }
3730: for (PetscInt i = 0; i < cnt; i++) PetscCall(VecLockReadPop(vecsin[i]));
3731: PetscFunctionReturn(PETSC_SUCCESS);
3732: }
3734: /*@C
3735: TSResizeRegisterVec - Register a vector to be transferred with `TSResize()`.
3737: Collective
3739: Input Parameters:
3740: + ts - The `TS` context obtained from `TSCreate()`
3741: . name - A string identifying the vector
3742: - vec - The vector
3744: Level: developer
3746: Note:
3747: `TSResizeRegisterVec()` is typically used within time stepping implementations,
3748: so most users would not generally call this routine themselves.
3750: .seealso: [](ch_ts), `TS`, `TSSetResize()`, `TSResize()`, `TSResizeRetrieveVec()`
3751: @*/
3752: PetscErrorCode TSResizeRegisterVec(TS ts, const char *name, Vec vec)
3753: {
3754: PetscFunctionBegin;
3756: PetscAssertPointer(name, 2);
3758: PetscCall(PetscObjectListAdd(&ts->resizetransferobjs, name, (PetscObject)vec));
3759: PetscFunctionReturn(PETSC_SUCCESS);
3760: }
3762: /*@C
3763: TSResizeRetrieveVec - Retrieve a vector registered with `TSResizeRegisterVec()`.
3765: Collective
3767: Input Parameters:
3768: + ts - The `TS` context obtained from `TSCreate()`
3769: . name - A string identifying the vector
3770: - vec - The vector
3772: Level: developer
3774: Note:
3775: `TSResizeRetrieveVec()` is typically used within time stepping implementations,
3776: so most users would not generally call this routine themselves.
3778: .seealso: [](ch_ts), `TS`, `TSSetResize()`, `TSResize()`, `TSResizeRegisterVec()`
3779: @*/
3780: PetscErrorCode TSResizeRetrieveVec(TS ts, const char *name, Vec *vec)
3781: {
3782: PetscFunctionBegin;
3784: PetscAssertPointer(name, 2);
3785: PetscAssertPointer(vec, 3);
3786: PetscCall(PetscObjectListFind(ts->resizetransferobjs, name, (PetscObject *)vec));
3787: PetscFunctionReturn(PETSC_SUCCESS);
3788: }
3790: static PetscErrorCode TSResizeGetVecArray(TS ts, PetscInt *nv, const char **names[], Vec *vecs[])
3791: {
3792: PetscInt cnt;
3793: PetscObjectList tmp;
3794: Vec *vecsin = NULL;
3795: const char **namesin = NULL;
3797: PetscFunctionBegin;
3798: for (tmp = ts->resizetransferobjs, cnt = 0; tmp; tmp = tmp->next)
3799: if (tmp->obj && tmp->obj->classid == VEC_CLASSID) cnt++;
3800: if (names) PetscCall(PetscMalloc1(cnt, &namesin));
3801: if (vecs) PetscCall(PetscMalloc1(cnt, &vecsin));
3802: for (tmp = ts->resizetransferobjs, cnt = 0; tmp; tmp = tmp->next) {
3803: if (tmp->obj && tmp->obj->classid == VEC_CLASSID) {
3804: if (vecs) vecsin[cnt] = (Vec)tmp->obj;
3805: if (names) namesin[cnt] = tmp->name;
3806: cnt++;
3807: }
3808: }
3809: if (nv) *nv = cnt;
3810: if (names) *names = namesin;
3811: if (vecs) *vecs = vecsin;
3812: PetscFunctionReturn(PETSC_SUCCESS);
3813: }
3815: /*@
3816: TSResize - Runs the user-defined transfer functions provided with `TSSetResize()`
3818: Collective
3820: Input Parameter:
3821: . ts - The `TS` context obtained from `TSCreate()`
3823: Level: developer
3825: Note:
3826: `TSResize()` is typically used within time stepping implementations,
3827: so most users would not generally call this routine themselves.
3829: .seealso: [](ch_ts), `TS`, `TSSetResize()`
3830: @*/
3831: PetscErrorCode TSResize(TS ts)
3832: {
3833: PetscInt nv = 0;
3834: const char **names = NULL;
3835: Vec *vecsin = NULL;
3836: const char *solname = "ts:vec_sol";
3838: PetscFunctionBegin;
3840: if (ts->resizesetup) {
3841: PetscBool flg = PETSC_FALSE;
3843: PetscCall(VecLockReadPush(ts->vec_sol));
3844: PetscCallBack("TS callback resize setup", (*ts->resizesetup)(ts, ts->steps, ts->ptime, ts->vec_sol, &flg, ts->resizectx));
3845: PetscCall(VecLockReadPop(ts->vec_sol));
3846: if (flg) {
3847: PetscCall(TSResizeRegisterVec(ts, solname, ts->vec_sol));
3848: PetscCall(TSResizeRegisterOrRetrieve(ts, PETSC_TRUE)); /* specific impls register their own objects */
3849: }
3850: }
3852: PetscCall(TSResizeGetVecArray(ts, &nv, &names, &vecsin));
3853: if (nv) {
3854: Vec *vecsout, vecsol;
3856: /* Reset internal objects */
3857: PetscCall(TSReset(ts));
3859: /* Transfer needed vectors (users can call SetJacobian, SetDM here) */
3860: PetscCall(PetscCalloc1(nv, &vecsout));
3861: PetscCall(TSResizeTransferVecs(ts, nv, vecsin, vecsout));
3862: for (PetscInt i = 0; i < nv; i++) {
3863: PetscCall(TSResizeRegisterVec(ts, names[i], vecsout[i]));
3864: PetscCall(VecDestroy(&vecsout[i]));
3865: }
3866: PetscCall(PetscFree(vecsout));
3867: PetscCall(TSResizeRegisterOrRetrieve(ts, PETSC_FALSE)); /* specific impls import the transferred objects */
3869: PetscCall(TSResizeRetrieveVec(ts, solname, &vecsol));
3870: if (vecsol) PetscCall(TSSetSolution(ts, vecsol));
3871: PetscAssert(ts->vec_sol, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_NULL, "Missing TS solution");
3872: }
3874: PetscCall(PetscFree(names));
3875: PetscCall(PetscFree(vecsin));
3876: PetscCall(TSResizeReset(ts));
3877: PetscFunctionReturn(PETSC_SUCCESS);
3878: }
3880: /*@
3881: TSSolve - Steps the requested number of timesteps.
3883: Collective
3885: Input Parameters:
3886: + ts - the `TS` context obtained from `TSCreate()`
3887: - u - the solution vector (can be null if `TSSetSolution()` was used and `TSSetExactFinalTime`(ts,`TS_EXACTFINALTIME_MATCHSTEP`) was not used,
3888: otherwise must contain the initial conditions and will contain the solution at the final requested time
3890: Level: beginner
3892: Notes:
3893: The final time returned by this function may be different from the time of the internally
3894: held state accessible by `TSGetSolution()` and `TSGetTime()` because the method may have
3895: stepped over the final time.
3897: .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetSolution()`, `TSStep()`, `TSGetTime()`, `TSGetSolveTime()`
3898: @*/
3899: PetscErrorCode TSSolve(TS ts, Vec u)
3900: {
3901: Vec solution;
3903: PetscFunctionBegin;
3907: PetscCall(TSSetExactFinalTimeDefault(ts));
3908: 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 */
3909: if (!ts->vec_sol || u == ts->vec_sol) {
3910: PetscCall(VecDuplicate(u, &solution));
3911: PetscCall(TSSetSolution(ts, solution));
3912: PetscCall(VecDestroy(&solution)); /* grant ownership */
3913: }
3914: PetscCall(VecCopy(u, ts->vec_sol));
3915: PetscCheck(!ts->forward_solve, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "Sensitivity analysis does not support the mode TS_EXACTFINALTIME_INTERPOLATE");
3916: } else if (u) PetscCall(TSSetSolution(ts, u));
3917: PetscCall(TSSetUp(ts));
3918: PetscCall(TSTrajectorySetUp(ts->trajectory, ts));
3920: 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>");
3921: 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()");
3922: 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");
3923: 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");
3925: 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 */
3926: PetscCall(VecCopy(ts->vec_sol, ts->tspan->vecs_sol[0]));
3927: ts->tspan->spanctr = 1;
3928: }
3930: if (ts->forward_solve) PetscCall(TSForwardSetUp(ts));
3932: /* reset number of steps only when the step is not restarted. ARKIMEX
3933: restarts the step after an event. Resetting these counters in such case causes
3934: TSTrajectory to incorrectly save the output files
3935: */
3936: /* reset time step and iteration counters */
3937: if (!ts->steps) {
3938: ts->ksp_its = 0;
3939: ts->snes_its = 0;
3940: ts->num_snes_failures = 0;
3941: ts->reject = 0;
3942: ts->steprestart = PETSC_TRUE;
3943: ts->steprollback = PETSC_FALSE;
3944: ts->rhsjacobian.time = PETSC_MIN_REAL;
3945: }
3947: /* make sure initial time step does not overshoot final time or the next point in tspan */
3948: if (ts->exact_final_time == TS_EXACTFINALTIME_MATCHSTEP) {
3949: PetscReal maxdt;
3950: PetscReal dt = ts->time_step;
3952: if (ts->tspan) maxdt = ts->tspan->span_times[ts->tspan->spanctr] - ts->ptime;
3953: else maxdt = ts->max_time - ts->ptime;
3954: ts->time_step = dt >= maxdt ? maxdt : (PetscIsCloseAtTol(dt, maxdt, 10 * PETSC_MACHINE_EPSILON, 0) ? maxdt : dt);
3955: }
3956: ts->reason = TS_CONVERGED_ITERATING;
3958: {
3959: PetscViewer viewer;
3960: PetscViewerFormat format;
3961: PetscBool flg;
3962: static PetscBool incall = PETSC_FALSE;
3964: if (!incall) {
3965: /* Estimate the convergence rate of the time discretization */
3966: PetscCall(PetscOptionsGetViewer(PetscObjectComm((PetscObject)ts), ((PetscObject)ts)->options, ((PetscObject)ts)->prefix, "-ts_convergence_estimate", &viewer, &format, &flg));
3967: if (flg) {
3968: PetscConvEst conv;
3969: DM dm;
3970: PetscReal *alpha; /* Convergence rate of the solution error for each field in the L_2 norm */
3971: PetscInt Nf;
3972: PetscBool checkTemporal = PETSC_TRUE;
3974: incall = PETSC_TRUE;
3975: PetscCall(PetscOptionsGetBool(((PetscObject)ts)->options, ((PetscObject)ts)->prefix, "-ts_convergence_temporal", &checkTemporal, &flg));
3976: PetscCall(TSGetDM(ts, &dm));
3977: PetscCall(DMGetNumFields(dm, &Nf));
3978: PetscCall(PetscCalloc1(PetscMax(Nf, 1), &alpha));
3979: PetscCall(PetscConvEstCreate(PetscObjectComm((PetscObject)ts), &conv));
3980: PetscCall(PetscConvEstUseTS(conv, checkTemporal));
3981: PetscCall(PetscConvEstSetSolver(conv, (PetscObject)ts));
3982: PetscCall(PetscConvEstSetFromOptions(conv));
3983: PetscCall(PetscConvEstSetUp(conv));
3984: PetscCall(PetscConvEstGetConvRate(conv, alpha));
3985: PetscCall(PetscViewerPushFormat(viewer, format));
3986: PetscCall(PetscConvEstRateView(conv, alpha, viewer));
3987: PetscCall(PetscViewerPopFormat(viewer));
3988: PetscCall(PetscViewerDestroy(&viewer));
3989: PetscCall(PetscConvEstDestroy(&conv));
3990: PetscCall(PetscFree(alpha));
3991: incall = PETSC_FALSE;
3992: }
3993: }
3994: }
3996: PetscCall(TSViewFromOptions(ts, NULL, "-ts_view_pre"));
3998: if (ts->ops->solve) { /* This private interface is transitional and should be removed when all implementations are updated. */
3999: PetscUseTypeMethod(ts, solve);
4000: if (u) PetscCall(VecCopy(ts->vec_sol, u));
4001: ts->solvetime = ts->ptime;
4002: solution = ts->vec_sol;
4003: } else { /* Step the requested number of timesteps. */
4004: if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS;
4005: else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME;
4007: if (!ts->steps) {
4008: PetscCall(TSTrajectorySet(ts->trajectory, ts, ts->steps, ts->ptime, ts->vec_sol));
4009: PetscCall(TSEventInitialize(ts->event, ts, ts->ptime, ts->vec_sol));
4010: }
4012: while (!ts->reason) {
4013: PetscCall(TSMonitor(ts, ts->steps, ts->ptime, ts->vec_sol));
4014: if (!ts->steprollback) PetscCall(TSPreStep(ts));
4015: PetscCall(TSStep(ts));
4016: if (ts->testjacobian) PetscCall(TSRHSJacobianTest(ts, NULL));
4017: if (ts->testjacobiantranspose) PetscCall(TSRHSJacobianTestTranspose(ts, NULL));
4018: if (ts->quadraturets && ts->costintegralfwd) { /* Must evaluate the cost integral before event is handled. The cost integral value can also be rolled back. */
4019: if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
4020: PetscCall(TSForwardCostIntegral(ts));
4021: if (ts->reason >= 0) ts->steps++;
4022: }
4023: if (ts->forward_solve) { /* compute forward sensitivities before event handling because postevent() may change RHS and jump conditions may have to be applied */
4024: if (ts->reason >= 0) ts->steps--; /* Revert the step number changed by TSStep() */
4025: PetscCall(TSForwardStep(ts));
4026: if (ts->reason >= 0) ts->steps++;
4027: }
4028: PetscCall(TSPostEvaluate(ts));
4029: 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. */
4030: if (ts->steprollback) PetscCall(TSPostEvaluate(ts));
4031: if (!ts->steprollback) {
4032: PetscCall(TSTrajectorySet(ts->trajectory, ts, ts->steps, ts->ptime, ts->vec_sol));
4033: PetscCall(TSPostStep(ts));
4034: PetscCall(TSResize(ts));
4035: }
4036: }
4037: PetscCall(TSMonitor(ts, ts->steps, ts->ptime, ts->vec_sol));
4039: if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE && ts->ptime > ts->max_time) {
4040: if (!u) u = ts->vec_sol;
4041: PetscCall(TSInterpolate(ts, ts->max_time, u));
4042: ts->solvetime = ts->max_time;
4043: solution = u;
4044: PetscCall(TSMonitor(ts, -1, ts->solvetime, solution));
4045: } else {
4046: if (u) PetscCall(VecCopy(ts->vec_sol, u));
4047: ts->solvetime = ts->ptime;
4048: solution = ts->vec_sol;
4049: }
4050: }
4052: PetscCall(TSViewFromOptions(ts, NULL, "-ts_view"));
4053: PetscCall(VecViewFromOptions(solution, (PetscObject)ts, "-ts_view_solution"));
4054: PetscCall(PetscObjectSAWsBlock((PetscObject)ts));
4055: if (ts->adjoint_solve) PetscCall(TSAdjointSolve(ts));
4056: PetscFunctionReturn(PETSC_SUCCESS);
4057: }
4059: /*@
4060: TSGetTime - Gets the time of the most recently completed step.
4062: Not Collective
4064: Input Parameter:
4065: . ts - the `TS` context obtained from `TSCreate()`
4067: Output Parameter:
4068: . t - the current time. This time may not corresponds to the final time set with `TSSetMaxTime()`, use `TSGetSolveTime()`.
4070: Level: beginner
4072: Note:
4073: When called during time step evaluation (e.g. during residual evaluation or via hooks set using `TSSetPreStep()`,
4074: `TSSetPreStage()`, `TSSetPostStage()`, or `TSSetPostStep()`), the time is the time at the start of the step being evaluated.
4076: .seealso: [](ch_ts), `TS`, ``TSGetSolveTime()`, `TSSetTime()`, `TSGetTimeStep()`, `TSGetStepNumber()`
4077: @*/
4078: PetscErrorCode TSGetTime(TS ts, PetscReal *t)
4079: {
4080: PetscFunctionBegin;
4082: PetscAssertPointer(t, 2);
4083: *t = ts->ptime;
4084: PetscFunctionReturn(PETSC_SUCCESS);
4085: }
4087: /*@
4088: TSGetPrevTime - Gets the starting time of the previously completed step.
4090: Not Collective
4092: Input Parameter:
4093: . ts - the `TS` context obtained from `TSCreate()`
4095: Output Parameter:
4096: . t - the previous time
4098: Level: beginner
4100: .seealso: [](ch_ts), `TS`, ``TSGetTime()`, `TSGetSolveTime()`, `TSGetTimeStep()`
4101: @*/
4102: PetscErrorCode TSGetPrevTime(TS ts, PetscReal *t)
4103: {
4104: PetscFunctionBegin;
4106: PetscAssertPointer(t, 2);
4107: *t = ts->ptime_prev;
4108: PetscFunctionReturn(PETSC_SUCCESS);
4109: }
4111: /*@
4112: TSSetTime - Allows one to reset the time.
4114: Logically Collective
4116: Input Parameters:
4117: + ts - the `TS` context obtained from `TSCreate()`
4118: - t - the time
4120: Level: intermediate
4122: .seealso: [](ch_ts), `TS`, `TSGetTime()`, `TSSetMaxSteps()`
4123: @*/
4124: PetscErrorCode TSSetTime(TS ts, PetscReal t)
4125: {
4126: PetscFunctionBegin;
4129: ts->ptime = t;
4130: PetscFunctionReturn(PETSC_SUCCESS);
4131: }
4133: /*@C
4134: TSSetOptionsPrefix - Sets the prefix used for searching for all
4135: TS options in the database.
4137: Logically Collective
4139: Input Parameters:
4140: + ts - The `TS` context
4141: - prefix - The prefix to prepend to all option names
4143: Level: advanced
4145: Note:
4146: A hyphen (-) must NOT be given at the beginning of the prefix name.
4147: The first character of all runtime options is AUTOMATICALLY the
4148: hyphen.
4150: .seealso: [](ch_ts), `TS`, `TSSetFromOptions()`, `TSAppendOptionsPrefix()`
4151: @*/
4152: PetscErrorCode TSSetOptionsPrefix(TS ts, const char prefix[])
4153: {
4154: SNES snes;
4156: PetscFunctionBegin;
4158: PetscCall(PetscObjectSetOptionsPrefix((PetscObject)ts, prefix));
4159: PetscCall(TSGetSNES(ts, &snes));
4160: PetscCall(SNESSetOptionsPrefix(snes, prefix));
4161: PetscFunctionReturn(PETSC_SUCCESS);
4162: }
4164: /*@C
4165: TSAppendOptionsPrefix - Appends to the prefix used for searching for all
4166: TS options in the database.
4168: Logically Collective
4170: Input Parameters:
4171: + ts - The `TS` context
4172: - prefix - The prefix to prepend to all option names
4174: Level: advanced
4176: Note:
4177: A hyphen (-) must NOT be given at the beginning of the prefix name.
4178: The first character of all runtime options is AUTOMATICALLY the
4179: hyphen.
4181: .seealso: [](ch_ts), `TS`, `TSGetOptionsPrefix()`, `TSSetOptionsPrefix()`, `TSSetFromOptions()`
4182: @*/
4183: PetscErrorCode TSAppendOptionsPrefix(TS ts, const char prefix[])
4184: {
4185: SNES snes;
4187: PetscFunctionBegin;
4189: PetscCall(PetscObjectAppendOptionsPrefix((PetscObject)ts, prefix));
4190: PetscCall(TSGetSNES(ts, &snes));
4191: PetscCall(SNESAppendOptionsPrefix(snes, prefix));
4192: PetscFunctionReturn(PETSC_SUCCESS);
4193: }
4195: /*@C
4196: TSGetOptionsPrefix - Sets the prefix used for searching for all
4197: `TS` options in the database.
4199: Not Collective
4201: Input Parameter:
4202: . ts - The `TS` context
4204: Output Parameter:
4205: . prefix - A pointer to the prefix string used
4207: Level: intermediate
4209: Fortran Notes:
4210: The user should pass in a string 'prefix' of
4211: sufficient length to hold the prefix.
4213: .seealso: [](ch_ts), `TS`, `TSAppendOptionsPrefix()`, `TSSetFromOptions()`
4214: @*/
4215: PetscErrorCode TSGetOptionsPrefix(TS ts, const char *prefix[])
4216: {
4217: PetscFunctionBegin;
4219: PetscAssertPointer(prefix, 2);
4220: PetscCall(PetscObjectGetOptionsPrefix((PetscObject)ts, prefix));
4221: PetscFunctionReturn(PETSC_SUCCESS);
4222: }
4224: /*@C
4225: TSGetRHSJacobian - Returns the Jacobian J at the present timestep.
4227: Not Collective, but parallel objects are returned if ts is parallel
4229: Input Parameter:
4230: . ts - The `TS` context obtained from `TSCreate()`
4232: Output Parameters:
4233: + Amat - The (approximate) Jacobian J of G, where U_t = G(U,t) (or `NULL`)
4234: . Pmat - The matrix from which the preconditioner is constructed, usually the same as `Amat` (or `NULL`)
4235: . func - Function to compute the Jacobian of the RHS (or `NULL`)
4236: - ctx - User-defined context for Jacobian evaluation routine (or `NULL`)
4238: Level: intermediate
4240: Note:
4241: You can pass in `NULL` for any return argument you do not need.
4243: .seealso: [](ch_ts), `TS`, `TSGetTimeStep()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`
4245: @*/
4246: PetscErrorCode TSGetRHSJacobian(TS ts, Mat *Amat, Mat *Pmat, TSRHSJacobian *func, void **ctx)
4247: {
4248: DM dm;
4250: PetscFunctionBegin;
4251: if (Amat || Pmat) {
4252: SNES snes;
4253: PetscCall(TSGetSNES(ts, &snes));
4254: PetscCall(SNESSetUpMatrices(snes));
4255: PetscCall(SNESGetJacobian(snes, Amat, Pmat, NULL, NULL));
4256: }
4257: PetscCall(TSGetDM(ts, &dm));
4258: PetscCall(DMTSGetRHSJacobian(dm, func, ctx));
4259: PetscFunctionReturn(PETSC_SUCCESS);
4260: }
4262: /*@C
4263: TSGetIJacobian - Returns the implicit Jacobian at the present timestep.
4265: Not Collective, but parallel objects are returned if ts is parallel
4267: Input Parameter:
4268: . ts - The `TS` context obtained from `TSCreate()`
4270: Output Parameters:
4271: + Amat - The (approximate) Jacobian of F(t,U,U_t)
4272: . Pmat - The matrix from which the preconditioner is constructed, often the same as `Amat`
4273: . f - The function to compute the matrices
4274: - ctx - User-defined context for Jacobian evaluation routine
4276: Level: advanced
4278: Note:
4279: You can pass in `NULL` for any return argument you do not need.
4281: .seealso: [](ch_ts), `TS`, `TSGetTimeStep()`, `TSGetRHSJacobian()`, `TSGetMatrices()`, `TSGetTime()`, `TSGetStepNumber()`
4282: @*/
4283: PetscErrorCode TSGetIJacobian(TS ts, Mat *Amat, Mat *Pmat, TSIJacobian *f, void **ctx)
4284: {
4285: DM dm;
4287: PetscFunctionBegin;
4288: if (Amat || Pmat) {
4289: SNES snes;
4290: PetscCall(TSGetSNES(ts, &snes));
4291: PetscCall(SNESSetUpMatrices(snes));
4292: PetscCall(SNESGetJacobian(snes, Amat, Pmat, NULL, NULL));
4293: }
4294: PetscCall(TSGetDM(ts, &dm));
4295: PetscCall(DMTSGetIJacobian(dm, f, ctx));
4296: PetscFunctionReturn(PETSC_SUCCESS);
4297: }
4299: #include <petsc/private/dmimpl.h>
4300: /*@
4301: TSSetDM - Sets the `DM` that may be used by some nonlinear solvers or preconditioners under the `TS`
4303: Logically Collective
4305: Input Parameters:
4306: + ts - the `TS` integrator object
4307: - dm - the dm, cannot be `NULL`
4309: Level: intermediate
4311: Notes:
4312: A `DM` can only be used for solving one problem at a time because information about the problem is stored on the `DM`,
4313: even when not using interfaces like `DMTSSetIFunction()`. Use `DMClone()` to get a distinct `DM` when solving
4314: different problems using the same function space.
4316: .seealso: [](ch_ts), `TS`, `DM`, `TSGetDM()`, `SNESSetDM()`, `SNESGetDM()`
4317: @*/
4318: PetscErrorCode TSSetDM(TS ts, DM dm)
4319: {
4320: SNES snes;
4321: DMTS tsdm;
4323: PetscFunctionBegin;
4326: PetscCall(PetscObjectReference((PetscObject)dm));
4327: if (ts->dm) { /* Move the DMTS context over to the new DM unless the new DM already has one */
4328: if (ts->dm->dmts && !dm->dmts) {
4329: PetscCall(DMCopyDMTS(ts->dm, dm));
4330: PetscCall(DMGetDMTS(ts->dm, &tsdm));
4331: /* Grant write privileges to the replacement DM */
4332: if (tsdm->originaldm == ts->dm) tsdm->originaldm = dm;
4333: }
4334: PetscCall(DMDestroy(&ts->dm));
4335: }
4336: ts->dm = dm;
4338: PetscCall(TSGetSNES(ts, &snes));
4339: PetscCall(SNESSetDM(snes, dm));
4340: PetscFunctionReturn(PETSC_SUCCESS);
4341: }
4343: /*@
4344: TSGetDM - Gets the `DM` that may be used by some preconditioners
4346: Not Collective
4348: Input Parameter:
4349: . ts - the `TS`
4351: Output Parameter:
4352: . dm - the `DM`
4354: Level: intermediate
4356: .seealso: [](ch_ts), `TS`, `DM`, `TSSetDM()`, `SNESSetDM()`, `SNESGetDM()`
4357: @*/
4358: PetscErrorCode TSGetDM(TS ts, DM *dm)
4359: {
4360: PetscFunctionBegin;
4362: if (!ts->dm) {
4363: PetscCall(DMShellCreate(PetscObjectComm((PetscObject)ts), &ts->dm));
4364: if (ts->snes) PetscCall(SNESSetDM(ts->snes, ts->dm));
4365: }
4366: *dm = ts->dm;
4367: PetscFunctionReturn(PETSC_SUCCESS);
4368: }
4370: /*@
4371: SNESTSFormFunction - Function to evaluate nonlinear residual
4373: Logically Collective
4375: Input Parameters:
4376: + snes - nonlinear solver
4377: . U - the current state at which to evaluate the residual
4378: - ctx - user context, must be a TS
4380: Output Parameter:
4381: . F - the nonlinear residual
4383: Level: advanced
4385: Note:
4386: This function is not normally called by users and is automatically registered with the `SNES` used by `TS`.
4387: It is most frequently passed to `MatFDColoringSetFunction()`.
4389: .seealso: [](ch_ts), `SNESSetFunction()`, `MatFDColoringSetFunction()`
4390: @*/
4391: PetscErrorCode SNESTSFormFunction(SNES snes, Vec U, Vec F, void *ctx)
4392: {
4393: TS ts = (TS)ctx;
4395: PetscFunctionBegin;
4400: PetscCheck(ts->ops->snesfunction, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "No method snesfunction for TS of type %s", ((PetscObject)ts)->type_name);
4401: PetscCall((*ts->ops->snesfunction)(snes, U, F, ts));
4402: PetscFunctionReturn(PETSC_SUCCESS);
4403: }
4405: /*@
4406: SNESTSFormJacobian - Function to evaluate the Jacobian
4408: Collective
4410: Input Parameters:
4411: + snes - nonlinear solver
4412: . U - the current state at which to evaluate the residual
4413: - ctx - user context, must be a `TS`
4415: Output Parameters:
4416: + A - the Jacobian
4417: - B - the preconditioning matrix (may be the same as A)
4419: Level: developer
4421: Note:
4422: This function is not normally called by users and is automatically registered with the `SNES` used by `TS`.
4424: .seealso: [](ch_ts), `SNESSetJacobian()`
4425: @*/
4426: PetscErrorCode SNESTSFormJacobian(SNES snes, Vec U, Mat A, Mat B, void *ctx)
4427: {
4428: TS ts = (TS)ctx;
4430: PetscFunctionBegin;
4436: PetscCheck(ts->ops->snesjacobian, PetscObjectComm((PetscObject)ts), PETSC_ERR_SUP, "No method snesjacobian for TS of type %s", ((PetscObject)ts)->type_name);
4437: PetscCall((*ts->ops->snesjacobian)(snes, U, A, B, ts));
4438: PetscFunctionReturn(PETSC_SUCCESS);
4439: }
4441: /*@C
4442: TSComputeRHSFunctionLinear - Evaluate the right hand side via the user-provided Jacobian, for linear problems Udot = A U only
4444: Collective
4446: Input Parameters:
4447: + ts - time stepping context
4448: . t - time at which to evaluate
4449: . U - state at which to evaluate
4450: - ctx - context
4452: Output Parameter:
4453: . F - right hand side
4455: Level: intermediate
4457: Note:
4458: This function is intended to be passed to `TSSetRHSFunction()` to evaluate the right hand side for linear problems.
4459: The matrix (and optionally the evaluation context) should be passed to `TSSetRHSJacobian()`.
4461: .seealso: [](ch_ts), `TS`, `TSSetRHSFunction()`, `TSSetRHSJacobian()`, `TSComputeRHSJacobianConstant()`
4462: @*/
4463: PetscErrorCode TSComputeRHSFunctionLinear(TS ts, PetscReal t, Vec U, Vec F, void *ctx)
4464: {
4465: Mat Arhs, Brhs;
4467: PetscFunctionBegin;
4468: PetscCall(TSGetRHSMats_Private(ts, &Arhs, &Brhs));
4469: /* undo the damage caused by shifting */
4470: PetscCall(TSRecoverRHSJacobian(ts, Arhs, Brhs));
4471: PetscCall(TSComputeRHSJacobian(ts, t, U, Arhs, Brhs));
4472: PetscCall(MatMult(Arhs, U, F));
4473: PetscFunctionReturn(PETSC_SUCCESS);
4474: }
4476: /*@C
4477: TSComputeRHSJacobianConstant - Reuses a Jacobian that is time-independent.
4479: Collective
4481: Input Parameters:
4482: + ts - time stepping context
4483: . t - time at which to evaluate
4484: . U - state at which to evaluate
4485: - ctx - context
4487: Output Parameters:
4488: + A - pointer to operator
4489: - B - pointer to preconditioning matrix
4491: Level: intermediate
4493: Note:
4494: This function is intended to be passed to `TSSetRHSJacobian()` to evaluate the Jacobian for linear time-independent problems.
4496: .seealso: [](ch_ts), `TS`, `TSSetRHSFunction()`, `TSSetRHSJacobian()`, `TSComputeRHSFunctionLinear()`
4497: @*/
4498: PetscErrorCode TSComputeRHSJacobianConstant(TS ts, PetscReal t, Vec U, Mat A, Mat B, void *ctx)
4499: {
4500: PetscFunctionBegin;
4501: PetscFunctionReturn(PETSC_SUCCESS);
4502: }
4504: /*@C
4505: TSComputeIFunctionLinear - Evaluate the left hand side via the user-provided Jacobian, for linear problems only
4507: Collective
4509: Input Parameters:
4510: + ts - time stepping context
4511: . t - time at which to evaluate
4512: . U - state at which to evaluate
4513: . Udot - time derivative of state vector
4514: - ctx - context
4516: Output Parameter:
4517: . F - left hand side
4519: Level: intermediate
4521: Notes:
4522: 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
4523: user is required to write their own `TSComputeIFunction()`.
4524: This function is intended to be passed to `TSSetIFunction()` to evaluate the left hand side for linear problems.
4525: The matrix (and optionally the evaluation context) should be passed to `TSSetIJacobian()`.
4527: Note that using this function is NOT equivalent to using `TSComputeRHSFunctionLinear()` since that solves Udot = A U
4529: .seealso: [](ch_ts), `TS`, `TSSetIFunction()`, `TSSetIJacobian()`, `TSComputeIJacobianConstant()`, `TSComputeRHSFunctionLinear()`
4530: @*/
4531: PetscErrorCode TSComputeIFunctionLinear(TS ts, PetscReal t, Vec U, Vec Udot, Vec F, void *ctx)
4532: {
4533: Mat A, B;
4535: PetscFunctionBegin;
4536: PetscCall(TSGetIJacobian(ts, &A, &B, NULL, NULL));
4537: PetscCall(TSComputeIJacobian(ts, t, U, Udot, 1.0, A, B, PETSC_TRUE));
4538: PetscCall(MatMult(A, Udot, F));
4539: PetscFunctionReturn(PETSC_SUCCESS);
4540: }
4542: /*@C
4543: TSComputeIJacobianConstant - Reuses the matrix previously computed with the provided `TSIJacobian()` for a semi-implicit DAE or ODE
4545: Collective
4547: Input Parameters:
4548: + ts - time stepping context
4549: . t - time at which to evaluate
4550: . U - state at which to evaluate
4551: . Udot - time derivative of state vector
4552: . shift - shift to apply
4553: - ctx - context
4555: Output Parameters:
4556: + A - pointer to operator
4557: - B - pointer to preconditioning matrix
4559: Level: advanced
4561: Notes:
4562: This function is intended to be passed to `TSSetIJacobian()` to evaluate the Jacobian for linear time-independent problems.
4564: It is only appropriate for problems of the form
4566: $ M Udot = F(U,t)
4568: where M is constant and F is non-stiff. The user must pass M to `TSSetIJacobian()`. The current implementation only
4569: works with IMEX time integration methods such as `TSROSW` and `TSARKIMEX`, since there is no support for de-constructing
4570: an implicit operator of the form
4572: $ shift*M + J
4574: 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
4575: a copy of M or reassemble it when requested.
4577: .seealso: [](ch_ts), `TS`, `TSROSW`, `TSARKIMEX`, `TSSetIFunction()`, `TSSetIJacobian()`, `TSComputeIFunctionLinear()`
4578: @*/
4579: PetscErrorCode TSComputeIJacobianConstant(TS ts, PetscReal t, Vec U, Vec Udot, PetscReal shift, Mat A, Mat B, void *ctx)
4580: {
4581: PetscFunctionBegin;
4582: PetscCall(MatScale(A, shift / ts->ijacobian.shift));
4583: ts->ijacobian.shift = shift;
4584: PetscFunctionReturn(PETSC_SUCCESS);
4585: }
4587: /*@
4588: TSGetEquationType - Gets the type of the equation that `TS` is solving.
4590: Not Collective
4592: Input Parameter:
4593: . ts - the `TS` context
4595: Output Parameter:
4596: . equation_type - see `TSEquationType`
4598: Level: beginner
4600: .seealso: [](ch_ts), `TS`, `TSSetEquationType()`, `TSEquationType`
4601: @*/
4602: PetscErrorCode TSGetEquationType(TS ts, TSEquationType *equation_type)
4603: {
4604: PetscFunctionBegin;
4606: PetscAssertPointer(equation_type, 2);
4607: *equation_type = ts->equation_type;
4608: PetscFunctionReturn(PETSC_SUCCESS);
4609: }
4611: /*@
4612: TSSetEquationType - Sets the type of the equation that `TS` is solving.
4614: Not Collective
4616: Input Parameters:
4617: + ts - the `TS` context
4618: - equation_type - see `TSEquationType`
4620: Level: advanced
4622: .seealso: [](ch_ts), `TS`, `TSGetEquationType()`, `TSEquationType`
4623: @*/
4624: PetscErrorCode TSSetEquationType(TS ts, TSEquationType equation_type)
4625: {
4626: PetscFunctionBegin;
4628: ts->equation_type = equation_type;
4629: PetscFunctionReturn(PETSC_SUCCESS);
4630: }
4632: /*@
4633: TSGetConvergedReason - Gets the reason the `TS` iteration was stopped.
4635: Not Collective
4637: Input Parameter:
4638: . ts - the `TS` context
4640: Output Parameter:
4641: . reason - negative value indicates diverged, positive value converged, see `TSConvergedReason` or the
4642: manual pages for the individual convergence tests for complete lists
4644: Level: beginner
4646: Note:
4647: Can only be called after the call to `TSSolve()` is complete.
4649: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSSetConvergenceTest()`, `TSConvergedReason`
4650: @*/
4651: PetscErrorCode TSGetConvergedReason(TS ts, TSConvergedReason *reason)
4652: {
4653: PetscFunctionBegin;
4655: PetscAssertPointer(reason, 2);
4656: *reason = ts->reason;
4657: PetscFunctionReturn(PETSC_SUCCESS);
4658: }
4660: /*@
4661: TSSetConvergedReason - Sets the reason for handling the convergence of `TSSolve()`.
4663: Logically Collective; reason must contain common value
4665: Input Parameters:
4666: + ts - the `TS` context
4667: - reason - negative value indicates diverged, positive value converged, see `TSConvergedReason` or the
4668: manual pages for the individual convergence tests for complete lists
4670: Level: advanced
4672: Note:
4673: Can only be called while `TSSolve()` is active.
4675: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSConvergedReason`
4676: @*/
4677: PetscErrorCode TSSetConvergedReason(TS ts, TSConvergedReason reason)
4678: {
4679: PetscFunctionBegin;
4681: ts->reason = reason;
4682: PetscFunctionReturn(PETSC_SUCCESS);
4683: }
4685: /*@
4686: TSGetSolveTime - Gets the time after a call to `TSSolve()`
4688: Not Collective
4690: Input Parameter:
4691: . ts - the `TS` context
4693: Output Parameter:
4694: . ftime - the final time. This time corresponds to the final time set with `TSSetMaxTime()`
4696: Level: beginner
4698: Note:
4699: Can only be called after the call to `TSSolve()` is complete.
4701: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSSetConvergenceTest()`, `TSConvergedReason`
4702: @*/
4703: PetscErrorCode TSGetSolveTime(TS ts, PetscReal *ftime)
4704: {
4705: PetscFunctionBegin;
4707: PetscAssertPointer(ftime, 2);
4708: *ftime = ts->solvetime;
4709: PetscFunctionReturn(PETSC_SUCCESS);
4710: }
4712: /*@
4713: TSGetSNESIterations - Gets the total number of nonlinear iterations
4714: used by the time integrator.
4716: Not Collective
4718: Input Parameter:
4719: . ts - `TS` context
4721: Output Parameter:
4722: . nits - number of nonlinear iterations
4724: Level: intermediate
4726: Note:
4727: This counter is reset to zero for each successive call to `TSSolve()`.
4729: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSGetKSPIterations()`
4730: @*/
4731: PetscErrorCode TSGetSNESIterations(TS ts, PetscInt *nits)
4732: {
4733: PetscFunctionBegin;
4735: PetscAssertPointer(nits, 2);
4736: *nits = ts->snes_its;
4737: PetscFunctionReturn(PETSC_SUCCESS);
4738: }
4740: /*@
4741: TSGetKSPIterations - Gets the total number of linear iterations
4742: used by the time integrator.
4744: Not Collective
4746: Input Parameter:
4747: . ts - `TS` context
4749: Output Parameter:
4750: . lits - number of linear iterations
4752: Level: intermediate
4754: Note:
4755: This counter is reset to zero for each successive call to `TSSolve()`.
4757: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSGetSNESIterations()`, `SNESGetKSPIterations()`
4758: @*/
4759: PetscErrorCode TSGetKSPIterations(TS ts, PetscInt *lits)
4760: {
4761: PetscFunctionBegin;
4763: PetscAssertPointer(lits, 2);
4764: *lits = ts->ksp_its;
4765: PetscFunctionReturn(PETSC_SUCCESS);
4766: }
4768: /*@
4769: TSGetStepRejections - Gets the total number of rejected steps.
4771: Not Collective
4773: Input Parameter:
4774: . ts - `TS` context
4776: Output Parameter:
4777: . rejects - number of steps rejected
4779: Level: intermediate
4781: Note:
4782: This counter is reset to zero for each successive call to `TSSolve()`.
4784: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetSNESFailures()`, `TSSetMaxSNESFailures()`, `TSSetErrorIfStepFails()`
4785: @*/
4786: PetscErrorCode TSGetStepRejections(TS ts, PetscInt *rejects)
4787: {
4788: PetscFunctionBegin;
4790: PetscAssertPointer(rejects, 2);
4791: *rejects = ts->reject;
4792: PetscFunctionReturn(PETSC_SUCCESS);
4793: }
4795: /*@
4796: TSGetSNESFailures - Gets the total number of failed `SNES` solves in a `TS`
4798: Not Collective
4800: Input Parameter:
4801: . ts - `TS` context
4803: Output Parameter:
4804: . fails - number of failed nonlinear solves
4806: Level: intermediate
4808: Note:
4809: This counter is reset to zero for each successive call to `TSSolve()`.
4811: .seealso: [](ch_ts), `TS`, `TSSolve()`, `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSSetMaxSNESFailures()`
4812: @*/
4813: PetscErrorCode TSGetSNESFailures(TS ts, PetscInt *fails)
4814: {
4815: PetscFunctionBegin;
4817: PetscAssertPointer(fails, 2);
4818: *fails = ts->num_snes_failures;
4819: PetscFunctionReturn(PETSC_SUCCESS);
4820: }
4822: /*@
4823: TSSetMaxStepRejections - Sets the maximum number of step rejections before a time step fails
4825: Not Collective
4827: Input Parameters:
4828: + ts - `TS` context
4829: - rejects - maximum number of rejected steps, pass -1 for unlimited
4831: Options Database Key:
4832: . -ts_max_reject - Maximum number of step rejections before a step fails
4834: Level: intermediate
4836: .seealso: [](ch_ts), `TS`, `SNES`, `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxSNESFailures()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `TSSetErrorIfStepFails()`, `TSGetConvergedReason()`
4837: @*/
4838: PetscErrorCode TSSetMaxStepRejections(TS ts, PetscInt rejects)
4839: {
4840: PetscFunctionBegin;
4842: ts->max_reject = rejects;
4843: PetscFunctionReturn(PETSC_SUCCESS);
4844: }
4846: /*@
4847: TSSetMaxSNESFailures - Sets the maximum number of failed `SNES` solves
4849: Not Collective
4851: Input Parameters:
4852: + ts - `TS` context
4853: - fails - maximum number of failed nonlinear solves, pass -1 for unlimited
4855: Options Database Key:
4856: . -ts_max_snes_failures - Maximum number of nonlinear solve failures
4858: Level: intermediate
4860: .seealso: [](ch_ts), `TS`, `SNES`, `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `SNESGetConvergedReason()`, `TSGetConvergedReason()`
4861: @*/
4862: PetscErrorCode TSSetMaxSNESFailures(TS ts, PetscInt fails)
4863: {
4864: PetscFunctionBegin;
4866: ts->max_snes_failures = fails;
4867: PetscFunctionReturn(PETSC_SUCCESS);
4868: }
4870: /*@
4871: TSSetErrorIfStepFails - Immediately error if no step succeeds during `TSSolve()`
4873: Not Collective
4875: Input Parameters:
4876: + ts - `TS` context
4877: - err - `PETSC_TRUE` to error if no step succeeds, `PETSC_FALSE` to return without failure
4879: Options Database Key:
4880: . -ts_error_if_step_fails - Error if no step succeeds
4882: Level: intermediate
4884: .seealso: [](ch_ts), `TS`, `TSGetSNESIterations()`, `TSGetKSPIterations()`, `TSSetMaxStepRejections()`, `TSGetStepRejections()`, `TSGetSNESFailures()`, `TSGetConvergedReason()`
4885: @*/
4886: PetscErrorCode TSSetErrorIfStepFails(TS ts, PetscBool err)
4887: {
4888: PetscFunctionBegin;
4890: ts->errorifstepfailed = err;
4891: PetscFunctionReturn(PETSC_SUCCESS);
4892: }
4894: /*@
4895: TSGetAdapt - Get the adaptive controller context for the current method
4897: Collective if controller has not yet been created
4899: Input Parameter:
4900: . ts - time stepping context
4902: Output Parameter:
4903: . adapt - adaptive controller
4905: Level: intermediate
4907: .seealso: [](ch_ts), `TS`, `TSAdapt`, `TSAdaptSetType()`, `TSAdaptChoose()`
4908: @*/
4909: PetscErrorCode TSGetAdapt(TS ts, TSAdapt *adapt)
4910: {
4911: PetscFunctionBegin;
4913: PetscAssertPointer(adapt, 2);
4914: if (!ts->adapt) {
4915: PetscCall(TSAdaptCreate(PetscObjectComm((PetscObject)ts), &ts->adapt));
4916: PetscCall(PetscObjectIncrementTabLevel((PetscObject)ts->adapt, (PetscObject)ts, 1));
4917: }
4918: *adapt = ts->adapt;
4919: PetscFunctionReturn(PETSC_SUCCESS);
4920: }
4922: /*@
4923: TSSetTolerances - Set tolerances for local truncation error when using an adaptive controller
4925: Logically Collective
4927: Input Parameters:
4928: + ts - time integration context
4929: . atol - scalar absolute tolerances, `PETSC_DECIDE` to leave current value
4930: . vatol - vector of absolute tolerances or `NULL`, used in preference to atol if present
4931: . rtol - scalar relative tolerances, `PETSC_DECIDE` to leave current value
4932: - vrtol - vector of relative tolerances or `NULL`, used in preference to atol if present
4934: Options Database Keys:
4935: + -ts_rtol <rtol> - relative tolerance for local truncation error
4936: - -ts_atol <atol> - Absolute tolerance for local truncation error
4938: Level: beginner
4940: Notes:
4941: With PETSc's implicit schemes for DAE problems, the calculation of the local truncation error
4942: (LTE) includes both the differential and the algebraic variables. If one wants the LTE to be
4943: computed only for the differential or the algebraic part then this can be done using the vector of
4944: tolerances vatol. For example, by setting the tolerance vector with the desired tolerance for the
4945: differential part and infinity for the algebraic part, the LTE calculation will include only the
4946: differential variables.
4948: .seealso: [](ch_ts), `TS`, `TSAdapt`, `TSErrorWeightedNorm()`, `TSGetTolerances()`
4949: @*/
4950: PetscErrorCode TSSetTolerances(TS ts, PetscReal atol, Vec vatol, PetscReal rtol, Vec vrtol)
4951: {
4952: PetscFunctionBegin;
4953: if (atol != (PetscReal)PETSC_DECIDE && atol != (PetscReal)PETSC_DEFAULT) ts->atol = atol;
4954: if (vatol) {
4955: PetscCall(PetscObjectReference((PetscObject)vatol));
4956: PetscCall(VecDestroy(&ts->vatol));
4957: ts->vatol = vatol;
4958: }
4959: if (rtol != (PetscReal)PETSC_DECIDE && rtol != (PetscReal)PETSC_DEFAULT) ts->rtol = rtol;
4960: if (vrtol) {
4961: PetscCall(PetscObjectReference((PetscObject)vrtol));
4962: PetscCall(VecDestroy(&ts->vrtol));
4963: ts->vrtol = vrtol;
4964: }
4965: PetscFunctionReturn(PETSC_SUCCESS);
4966: }
4968: /*@
4969: TSGetTolerances - Get tolerances for local truncation error when using adaptive controller
4971: Logically Collective
4973: Input Parameter:
4974: . ts - time integration context
4976: Output Parameters:
4977: + atol - scalar absolute tolerances, `NULL` to ignore
4978: . vatol - vector of absolute tolerances, `NULL` to ignore
4979: . rtol - scalar relative tolerances, `NULL` to ignore
4980: - vrtol - vector of relative tolerances, `NULL` to ignore
4982: Level: beginner
4984: .seealso: [](ch_ts), `TS`, `TSAdapt`, `TSErrorWeightedNorm()`, `TSSetTolerances()`
4985: @*/
4986: PetscErrorCode TSGetTolerances(TS ts, PetscReal *atol, Vec *vatol, PetscReal *rtol, Vec *vrtol)
4987: {
4988: PetscFunctionBegin;
4989: if (atol) *atol = ts->atol;
4990: if (vatol) *vatol = ts->vatol;
4991: if (rtol) *rtol = ts->rtol;
4992: if (vrtol) *vrtol = ts->vrtol;
4993: PetscFunctionReturn(PETSC_SUCCESS);
4994: }
4996: /*@
4997: TSErrorWeightedNorm - compute a weighted norm of the difference between two state vectors based on supplied absolute and relative tolerances
4999: Collective
5001: Input Parameters:
5002: + ts - time stepping context
5003: . U - state vector, usually ts->vec_sol
5004: . Y - state vector to be compared to U
5005: - wnormtype - norm type, either `NORM_2` or `NORM_INFINITY`
5007: Output Parameters:
5008: + norm - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5009: . norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5010: - normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5012: Options Database Key:
5013: . -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5015: Level: developer
5017: .seealso: [](ch_ts), `TS`, `VecErrorWeightedNorms()`, `TSErrorWeightedENorm()`
5018: @*/
5019: PetscErrorCode TSErrorWeightedNorm(TS ts, Vec U, Vec Y, NormType wnormtype, PetscReal *norm, PetscReal *norma, PetscReal *normr)
5020: {
5021: PetscInt norma_loc, norm_loc, normr_loc;
5023: PetscFunctionBegin;
5025: PetscCall(VecErrorWeightedNorms(U, Y, NULL, wnormtype, ts->atol, ts->vatol, ts->rtol, ts->vrtol, ts->adapt->ignore_max, norm, &norm_loc, norma, &norma_loc, normr, &normr_loc));
5026: if (wnormtype == NORM_2) {
5027: if (norm_loc) *norm = PetscSqrtReal(PetscSqr(*norm) / norm_loc);
5028: if (norma_loc) *norma = PetscSqrtReal(PetscSqr(*norma) / norma_loc);
5029: if (normr_loc) *normr = PetscSqrtReal(PetscSqr(*normr) / normr_loc);
5030: }
5031: PetscCheck(!PetscIsInfOrNanScalar(*norm), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in norm");
5032: PetscCheck(!PetscIsInfOrNanScalar(*norma), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in norma");
5033: PetscCheck(!PetscIsInfOrNanScalar(*normr), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in normr");
5034: PetscFunctionReturn(PETSC_SUCCESS);
5035: }
5037: /*@
5038: TSErrorWeightedENorm - compute a weighted error norm based on supplied absolute and relative tolerances
5040: Collective
5042: Input Parameters:
5043: + ts - time stepping context
5044: . E - error vector
5045: . U - state vector, usually ts->vec_sol
5046: . Y - state vector, previous time step
5047: - wnormtype - norm type, either `NORM_2` or `NORM_INFINITY`
5049: Output Parameters:
5050: + norm - weighted norm, a value of 1.0 achieves a balance between absolute and relative tolerances
5051: . norma - weighted norm, a value of 1.0 means that the error meets the absolute tolerance set by the user
5052: - normr - weighted norm, a value of 1.0 means that the error meets the relative tolerance set by the user
5054: Options Database Key:
5055: . -ts_adapt_wnormtype <wnormtype> - 2, INFINITY
5057: Level: developer
5059: .seealso: [](ch_ts), `TS`, `VecErrorWeightedNorms()`, `TSErrorWeightedNorm()`
5060: @*/
5061: PetscErrorCode TSErrorWeightedENorm(TS ts, Vec E, Vec U, Vec Y, NormType wnormtype, PetscReal *norm, PetscReal *norma, PetscReal *normr)
5062: {
5063: PetscInt norma_loc, norm_loc, normr_loc;
5065: PetscFunctionBegin;
5067: PetscCall(VecErrorWeightedNorms(U, Y, E, wnormtype, ts->atol, ts->vatol, ts->rtol, ts->vrtol, ts->adapt->ignore_max, norm, &norm_loc, norma, &norma_loc, normr, &normr_loc));
5068: if (wnormtype == NORM_2) {
5069: if (norm_loc) *norm = PetscSqrtReal(PetscSqr(*norm) / norm_loc);
5070: if (norma_loc) *norma = PetscSqrtReal(PetscSqr(*norma) / norma_loc);
5071: if (normr_loc) *normr = PetscSqrtReal(PetscSqr(*normr) / normr_loc);
5072: }
5073: PetscCheck(!PetscIsInfOrNanScalar(*norm), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in norm");
5074: PetscCheck(!PetscIsInfOrNanScalar(*norma), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in norma");
5075: PetscCheck(!PetscIsInfOrNanScalar(*normr), PetscObjectComm((PetscObject)ts), PETSC_ERR_FP, "Infinite or not-a-number generated in normr");
5076: PetscFunctionReturn(PETSC_SUCCESS);
5077: }
5079: /*@
5080: TSSetCFLTimeLocal - Set the local CFL constraint relative to forward Euler
5082: Logically Collective
5084: Input Parameters:
5085: + ts - time stepping context
5086: - cfltime - maximum stable time step if using forward Euler (value can be different on each process)
5088: Note:
5089: After calling this function, the global CFL time can be obtained by calling TSGetCFLTime()
5091: Level: intermediate
5093: .seealso: [](ch_ts), `TSGetCFLTime()`, `TSADAPTCFL`
5094: @*/
5095: PetscErrorCode TSSetCFLTimeLocal(TS ts, PetscReal cfltime)
5096: {
5097: PetscFunctionBegin;
5099: ts->cfltime_local = cfltime;
5100: ts->cfltime = -1.;
5101: PetscFunctionReturn(PETSC_SUCCESS);
5102: }
5104: /*@
5105: TSGetCFLTime - Get the maximum stable time step according to CFL criteria applied to forward Euler
5107: Collective
5109: Input Parameter:
5110: . ts - time stepping context
5112: Output Parameter:
5113: . cfltime - maximum stable time step for forward Euler
5115: Level: advanced
5117: .seealso: [](ch_ts), `TSSetCFLTimeLocal()`
5118: @*/
5119: PetscErrorCode TSGetCFLTime(TS ts, PetscReal *cfltime)
5120: {
5121: PetscFunctionBegin;
5122: if (ts->cfltime < 0) PetscCall(MPIU_Allreduce(&ts->cfltime_local, &ts->cfltime, 1, MPIU_REAL, MPIU_MIN, PetscObjectComm((PetscObject)ts)));
5123: *cfltime = ts->cfltime;
5124: PetscFunctionReturn(PETSC_SUCCESS);
5125: }
5127: /*@
5128: TSVISetVariableBounds - Sets the lower and upper bounds for the solution vector. xl <= x <= xu
5130: Input Parameters:
5131: + ts - the `TS` context.
5132: . xl - lower bound.
5133: - xu - upper bound.
5135: Level: advanced
5137: Note:
5138: If this routine is not called then the lower and upper bounds are set to
5139: `PETSC_NINFINITY` and `PETSC_INFINITY` respectively during `SNESSetUp()`.
5141: .seealso: [](ch_ts), `TS`
5142: @*/
5143: PetscErrorCode TSVISetVariableBounds(TS ts, Vec xl, Vec xu)
5144: {
5145: SNES snes;
5147: PetscFunctionBegin;
5148: PetscCall(TSGetSNES(ts, &snes));
5149: PetscCall(SNESVISetVariableBounds(snes, xl, xu));
5150: PetscFunctionReturn(PETSC_SUCCESS);
5151: }
5153: /*@
5154: TSComputeLinearStability - computes the linear stability function at a point
5156: Collective
5158: Input Parameters:
5159: + ts - the `TS` context
5160: . xr - real part of input argument
5161: - xi - imaginary part of input argument
5163: Output Parameters:
5164: + yr - real part of function value
5165: - yi - imaginary part of function value
5167: Level: developer
5169: .seealso: [](ch_ts), `TS`, `TSSetRHSFunction()`, `TSComputeIFunction()`
5170: @*/
5171: PetscErrorCode TSComputeLinearStability(TS ts, PetscReal xr, PetscReal xi, PetscReal *yr, PetscReal *yi)
5172: {
5173: PetscFunctionBegin;
5175: PetscUseTypeMethod(ts, linearstability, xr, xi, yr, yi);
5176: PetscFunctionReturn(PETSC_SUCCESS);
5177: }
5179: /*@
5180: TSRestartStep - Flags the solver to restart the next step
5182: Collective
5184: Input Parameter:
5185: . ts - the `TS` context obtained from `TSCreate()`
5187: Level: advanced
5189: Notes:
5190: Multistep methods like `TSBDF` or Runge-Kutta methods with FSAL property require restarting the solver in the event of
5191: discontinuities. These discontinuities may be introduced as a consequence of explicitly modifications to the solution
5192: vector (which PETSc attempts to detect and handle) or problem coefficients (which PETSc is not able to detect). For
5193: the sake of correctness and maximum safety, users are expected to call `TSRestart()` whenever they introduce
5194: discontinuities in callback routines (e.g. prestep and poststep routines, or implicit/rhs function routines with
5195: discontinuous source terms).
5197: .seealso: [](ch_ts), `TS`, `TSBDF`, `TSSolve()`, `TSSetPreStep()`, `TSSetPostStep()`
5198: @*/
5199: PetscErrorCode TSRestartStep(TS ts)
5200: {
5201: PetscFunctionBegin;
5203: ts->steprestart = PETSC_TRUE;
5204: PetscFunctionReturn(PETSC_SUCCESS);
5205: }
5207: /*@
5208: TSRollBack - Rolls back one time step
5210: Collective
5212: Input Parameter:
5213: . ts - the `TS` context obtained from `TSCreate()`
5215: Level: advanced
5217: .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetUp()`, `TSDestroy()`, `TSSolve()`, `TSSetPreStep()`, `TSSetPreStage()`, `TSInterpolate()`
5218: @*/
5219: PetscErrorCode TSRollBack(TS ts)
5220: {
5221: PetscFunctionBegin;
5223: PetscCheck(!ts->steprollback, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONGSTATE, "TSRollBack already called");
5224: PetscUseTypeMethod(ts, rollback);
5225: ts->time_step = ts->ptime - ts->ptime_prev;
5226: ts->ptime = ts->ptime_prev;
5227: ts->ptime_prev = ts->ptime_prev_rollback;
5228: ts->steps--;
5229: ts->steprollback = PETSC_TRUE;
5230: PetscFunctionReturn(PETSC_SUCCESS);
5231: }
5233: /*@
5234: TSGetStages - Get the number of stages and stage values
5236: Input Parameter:
5237: . ts - the `TS` context obtained from `TSCreate()`
5239: Output Parameters:
5240: + ns - the number of stages
5241: - Y - the current stage vectors
5243: Level: advanced
5245: Note:
5246: Both `ns` and `Y` can be `NULL`.
5248: .seealso: [](ch_ts), `TS`, `TSCreate()`
5249: @*/
5250: PetscErrorCode TSGetStages(TS ts, PetscInt *ns, Vec **Y)
5251: {
5252: PetscFunctionBegin;
5254: if (ns) PetscAssertPointer(ns, 2);
5255: if (Y) PetscAssertPointer(Y, 3);
5256: if (!ts->ops->getstages) {
5257: if (ns) *ns = 0;
5258: if (Y) *Y = NULL;
5259: } else PetscUseTypeMethod(ts, getstages, ns, Y);
5260: PetscFunctionReturn(PETSC_SUCCESS);
5261: }
5263: /*@C
5264: TSComputeIJacobianDefaultColor - Computes the Jacobian using finite differences and coloring to exploit matrix sparsity.
5266: Collective
5268: Input Parameters:
5269: + ts - the `TS` context
5270: . t - current timestep
5271: . U - state vector
5272: . Udot - time derivative of state vector
5273: . shift - shift to apply, see note below
5274: - ctx - an optional user context
5276: Output Parameters:
5277: + J - Jacobian matrix (not altered in this routine)
5278: - B - newly computed Jacobian matrix to use with preconditioner (generally the same as `J`)
5280: Level: intermediate
5282: Notes:
5283: If F(t,U,Udot)=0 is the DAE, the required Jacobian is
5285: dF/dU + shift*dF/dUdot
5287: Most users should not need to explicitly call this routine, as it
5288: is used internally within the nonlinear solvers.
5290: This will first try to get the coloring from the `DM`. If the `DM` type has no coloring
5291: routine, then it will try to get the coloring from the matrix. This requires that the
5292: matrix have nonzero entries precomputed.
5294: .seealso: [](ch_ts), `TS`, `TSSetIJacobian()`, `MatFDColoringCreate()`, `MatFDColoringSetFunction()`
5295: @*/
5296: PetscErrorCode TSComputeIJacobianDefaultColor(TS ts, PetscReal t, Vec U, Vec Udot, PetscReal shift, Mat J, Mat B, void *ctx)
5297: {
5298: SNES snes;
5299: MatFDColoring color;
5300: PetscBool hascolor, matcolor = PETSC_FALSE;
5302: PetscFunctionBegin;
5303: PetscCall(PetscOptionsGetBool(((PetscObject)ts)->options, ((PetscObject)ts)->prefix, "-ts_fd_color_use_mat", &matcolor, NULL));
5304: PetscCall(PetscObjectQuery((PetscObject)B, "TSMatFDColoring", (PetscObject *)&color));
5305: if (!color) {
5306: DM dm;
5307: ISColoring iscoloring;
5309: PetscCall(TSGetDM(ts, &dm));
5310: PetscCall(DMHasColoring(dm, &hascolor));
5311: if (hascolor && !matcolor) {
5312: PetscCall(DMCreateColoring(dm, IS_COLORING_GLOBAL, &iscoloring));
5313: PetscCall(MatFDColoringCreate(B, iscoloring, &color));
5314: PetscCall(MatFDColoringSetFunction(color, (PetscErrorCode(*)(void))SNESTSFormFunction, (void *)ts));
5315: PetscCall(MatFDColoringSetFromOptions(color));
5316: PetscCall(MatFDColoringSetUp(B, iscoloring, color));
5317: PetscCall(ISColoringDestroy(&iscoloring));
5318: } else {
5319: MatColoring mc;
5321: PetscCall(MatColoringCreate(B, &mc));
5322: PetscCall(MatColoringSetDistance(mc, 2));
5323: PetscCall(MatColoringSetType(mc, MATCOLORINGSL));
5324: PetscCall(MatColoringSetFromOptions(mc));
5325: PetscCall(MatColoringApply(mc, &iscoloring));
5326: PetscCall(MatColoringDestroy(&mc));
5327: PetscCall(MatFDColoringCreate(B, iscoloring, &color));
5328: PetscCall(MatFDColoringSetFunction(color, (PetscErrorCode(*)(void))SNESTSFormFunction, (void *)ts));
5329: PetscCall(MatFDColoringSetFromOptions(color));
5330: PetscCall(MatFDColoringSetUp(B, iscoloring, color));
5331: PetscCall(ISColoringDestroy(&iscoloring));
5332: }
5333: PetscCall(PetscObjectCompose((PetscObject)B, "TSMatFDColoring", (PetscObject)color));
5334: PetscCall(PetscObjectDereference((PetscObject)color));
5335: }
5336: PetscCall(TSGetSNES(ts, &snes));
5337: PetscCall(MatFDColoringApply(B, color, U, snes));
5338: if (J != B) {
5339: PetscCall(MatAssemblyBegin(J, MAT_FINAL_ASSEMBLY));
5340: PetscCall(MatAssemblyEnd(J, MAT_FINAL_ASSEMBLY));
5341: }
5342: PetscFunctionReturn(PETSC_SUCCESS);
5343: }
5345: /*@C
5346: TSSetFunctionDomainError - Set a function that tests if the current state vector is valid
5348: Input Parameters:
5349: + ts - the `TS` context
5350: - func - function called within `TSFunctionDomainError()`
5352: Calling sequence of `func`:
5353: + ts - the `TS` context
5354: . time - the current time (of the stage)
5355: . state - the state to check if it is valid
5356: - accept - (output parameter) `PETSC_FALSE` if the state is not acceptable, `PETSC_TRUE` if acceptable
5358: Level: intermediate
5360: Notes:
5361: If an implicit ODE solver is being used then, in addition to providing this routine, the
5362: user's code should call `SNESSetFunctionDomainError()` when domain errors occur during
5363: function evaluations where the functions are provided by `TSSetIFunction()` or `TSSetRHSFunction()`.
5364: Use `TSGetSNES()` to obtain the `SNES` object
5366: Developer Notes:
5367: The naming of this function is inconsistent with the `SNESSetFunctionDomainError()`
5368: since one takes a function pointer and the other does not.
5370: .seealso: [](ch_ts), `TSAdaptCheckStage()`, `TSFunctionDomainError()`, `SNESSetFunctionDomainError()`, `TSGetSNES()`
5371: @*/
5372: PetscErrorCode TSSetFunctionDomainError(TS ts, PetscErrorCode (*func)(TS ts, PetscReal time, Vec state, PetscBool *accept))
5373: {
5374: PetscFunctionBegin;
5376: ts->functiondomainerror = func;
5377: PetscFunctionReturn(PETSC_SUCCESS);
5378: }
5380: /*@
5381: TSFunctionDomainError - Checks if the current state is valid
5383: Input Parameters:
5384: + ts - the `TS` context
5385: . stagetime - time of the simulation
5386: - Y - state vector to check.
5388: Output Parameter:
5389: . accept - Set to `PETSC_FALSE` if the current state vector is valid.
5391: Level: developer
5393: Note:
5394: This function is called by the `TS` integration routines and calls the user provided function (set with `TSSetFunctionDomainError()`)
5395: to check if the current state is valid.
5397: .seealso: [](ch_ts), `TS`, `TSSetFunctionDomainError()`
5398: @*/
5399: PetscErrorCode TSFunctionDomainError(TS ts, PetscReal stagetime, Vec Y, PetscBool *accept)
5400: {
5401: PetscFunctionBegin;
5403: *accept = PETSC_TRUE;
5404: if (ts->functiondomainerror) PetscCall((*ts->functiondomainerror)(ts, stagetime, Y, accept));
5405: PetscFunctionReturn(PETSC_SUCCESS);
5406: }
5408: /*@C
5409: TSClone - This function clones a time step `TS` object.
5411: Collective
5413: Input Parameter:
5414: . tsin - The input `TS`
5416: Output Parameter:
5417: . tsout - The output `TS` (cloned)
5419: Level: developer
5421: Notes:
5422: This function is used to create a clone of a `TS` object. It is used in `TSARKIMEX` for initializing the slope for first stage explicit methods.
5423: It will likely be replaced in the future with a mechanism of switching methods on the fly.
5425: 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
5426: .vb
5427: SNES snes_dup = NULL;
5428: TSGetSNES(ts,&snes_dup);
5429: TSSetSNES(ts,snes_dup);
5430: .ve
5432: .seealso: [](ch_ts), `TS`, `SNES`, `TSCreate()`, `TSSetType()`, `TSSetUp()`, `TSDestroy()`, `TSSetProblemType()`
5433: @*/
5434: PetscErrorCode TSClone(TS tsin, TS *tsout)
5435: {
5436: TS t;
5437: SNES snes_start;
5438: DM dm;
5439: TSType type;
5441: PetscFunctionBegin;
5442: PetscAssertPointer(tsin, 1);
5443: *tsout = NULL;
5445: PetscCall(PetscHeaderCreate(t, TS_CLASSID, "TS", "Time stepping", "TS", PetscObjectComm((PetscObject)tsin), TSDestroy, TSView));
5447: /* General TS description */
5448: t->numbermonitors = 0;
5449: t->monitorFrequency = 1;
5450: t->setupcalled = 0;
5451: t->ksp_its = 0;
5452: t->snes_its = 0;
5453: t->nwork = 0;
5454: t->rhsjacobian.time = PETSC_MIN_REAL;
5455: t->rhsjacobian.scale = 1.;
5456: t->ijacobian.shift = 1.;
5458: PetscCall(TSGetSNES(tsin, &snes_start));
5459: PetscCall(TSSetSNES(t, snes_start));
5461: PetscCall(TSGetDM(tsin, &dm));
5462: PetscCall(TSSetDM(t, dm));
5464: t->adapt = tsin->adapt;
5465: PetscCall(PetscObjectReference((PetscObject)t->adapt));
5467: t->trajectory = tsin->trajectory;
5468: PetscCall(PetscObjectReference((PetscObject)t->trajectory));
5470: t->event = tsin->event;
5471: if (t->event) t->event->refct++;
5473: t->problem_type = tsin->problem_type;
5474: t->ptime = tsin->ptime;
5475: t->ptime_prev = tsin->ptime_prev;
5476: t->time_step = tsin->time_step;
5477: t->max_time = tsin->max_time;
5478: t->steps = tsin->steps;
5479: t->max_steps = tsin->max_steps;
5480: t->equation_type = tsin->equation_type;
5481: t->atol = tsin->atol;
5482: t->rtol = tsin->rtol;
5483: t->max_snes_failures = tsin->max_snes_failures;
5484: t->max_reject = tsin->max_reject;
5485: t->errorifstepfailed = tsin->errorifstepfailed;
5487: PetscCall(TSGetType(tsin, &type));
5488: PetscCall(TSSetType(t, type));
5490: t->vec_sol = NULL;
5492: t->cfltime = tsin->cfltime;
5493: t->cfltime_local = tsin->cfltime_local;
5494: t->exact_final_time = tsin->exact_final_time;
5496: t->ops[0] = tsin->ops[0];
5498: if (((PetscObject)tsin)->fortran_func_pointers) {
5499: PetscInt i;
5500: PetscCall(PetscMalloc((10) * sizeof(void (*)(void)), &((PetscObject)t)->fortran_func_pointers));
5501: for (i = 0; i < 10; i++) ((PetscObject)t)->fortran_func_pointers[i] = ((PetscObject)tsin)->fortran_func_pointers[i];
5502: }
5503: *tsout = t;
5504: PetscFunctionReturn(PETSC_SUCCESS);
5505: }
5507: static PetscErrorCode RHSWrapperFunction_TSRHSJacobianTest(void *ctx, Vec x, Vec y)
5508: {
5509: TS ts = (TS)ctx;
5511: PetscFunctionBegin;
5512: PetscCall(TSComputeRHSFunction(ts, 0, x, y));
5513: PetscFunctionReturn(PETSC_SUCCESS);
5514: }
5516: /*@
5517: TSRHSJacobianTest - Compares the multiply routine provided to the `MATSHELL` with differencing on the `TS` given RHS function.
5519: Logically Collective
5521: Input Parameter:
5522: . ts - the time stepping routine
5524: Output Parameter:
5525: . flg - `PETSC_TRUE` if the multiply is likely correct
5527: Options Database Key:
5528: . -ts_rhs_jacobian_test_mult -mat_shell_test_mult_view - run the test at each timestep of the integrator
5530: Level: advanced
5532: Note:
5533: This only works for problems defined using `TSSetRHSFunction()` and Jacobian NOT `TSSetIFunction()` and Jacobian
5535: .seealso: [](ch_ts), `TS`, `Mat`, `MATSHELL`, `MatCreateShell()`, `MatShellGetContext()`, `MatShellGetOperation()`, `MatShellTestMultTranspose()`, `TSRHSJacobianTestTranspose()`
5536: @*/
5537: PetscErrorCode TSRHSJacobianTest(TS ts, PetscBool *flg)
5538: {
5539: Mat J, B;
5540: TSRHSJacobian func;
5541: void *ctx;
5543: PetscFunctionBegin;
5544: PetscCall(TSGetRHSJacobian(ts, &J, &B, &func, &ctx));
5545: PetscCall((*func)(ts, 0.0, ts->vec_sol, J, B, ctx));
5546: PetscCall(MatShellTestMult(J, RHSWrapperFunction_TSRHSJacobianTest, ts->vec_sol, ts, flg));
5547: PetscFunctionReturn(PETSC_SUCCESS);
5548: }
5550: /*@C
5551: TSRHSJacobianTestTranspose - Compares the multiply transpose routine provided to the `MATSHELL` with differencing on the `TS` given RHS function.
5553: Logically Collective
5555: Input Parameter:
5556: . ts - the time stepping routine
5558: Output Parameter:
5559: . flg - `PETSC_TRUE` if the multiply is likely correct
5561: Options Database Key:
5562: . -ts_rhs_jacobian_test_mult_transpose -mat_shell_test_mult_transpose_view - run the test at each timestep of the integrator
5564: Level: advanced
5566: Notes:
5567: This only works for problems defined using `TSSetRHSFunction()` and Jacobian NOT `TSSetIFunction()` and Jacobian
5569: .seealso: [](ch_ts), `TS`, `Mat`, `MatCreateShell()`, `MatShellGetContext()`, `MatShellGetOperation()`, `MatShellTestMultTranspose()`, `TSRHSJacobianTest()`
5570: @*/
5571: PetscErrorCode TSRHSJacobianTestTranspose(TS ts, PetscBool *flg)
5572: {
5573: Mat J, B;
5574: void *ctx;
5575: TSRHSJacobian func;
5577: PetscFunctionBegin;
5578: PetscCall(TSGetRHSJacobian(ts, &J, &B, &func, &ctx));
5579: PetscCall((*func)(ts, 0.0, ts->vec_sol, J, B, ctx));
5580: PetscCall(MatShellTestMultTranspose(J, RHSWrapperFunction_TSRHSJacobianTest, ts->vec_sol, ts, flg));
5581: PetscFunctionReturn(PETSC_SUCCESS);
5582: }
5584: /*@
5585: TSSetUseSplitRHSFunction - Use the split RHSFunction when a multirate method is used.
5587: Logically Collective
5589: Input Parameters:
5590: + ts - timestepping context
5591: - use_splitrhsfunction - `PETSC_TRUE` indicates that the split RHSFunction will be used
5593: Options Database Key:
5594: . -ts_use_splitrhsfunction - <true,false>
5596: Level: intermediate
5598: Note:
5599: This is only for multirate methods
5601: .seealso: [](ch_ts), `TS`, `TSGetUseSplitRHSFunction()`
5602: @*/
5603: PetscErrorCode TSSetUseSplitRHSFunction(TS ts, PetscBool use_splitrhsfunction)
5604: {
5605: PetscFunctionBegin;
5607: ts->use_splitrhsfunction = use_splitrhsfunction;
5608: PetscFunctionReturn(PETSC_SUCCESS);
5609: }
5611: /*@
5612: TSGetUseSplitRHSFunction - Gets whether to use the split RHSFunction when a multirate method is used.
5614: Not Collective
5616: Input Parameter:
5617: . ts - timestepping context
5619: Output Parameter:
5620: . use_splitrhsfunction - `PETSC_TRUE` indicates that the split RHSFunction will be used
5622: Level: intermediate
5624: .seealso: [](ch_ts), `TS`, `TSSetUseSplitRHSFunction()`
5625: @*/
5626: PetscErrorCode TSGetUseSplitRHSFunction(TS ts, PetscBool *use_splitrhsfunction)
5627: {
5628: PetscFunctionBegin;
5630: *use_splitrhsfunction = ts->use_splitrhsfunction;
5631: PetscFunctionReturn(PETSC_SUCCESS);
5632: }
5634: /*@
5635: TSSetMatStructure - sets the relationship between the nonzero structure of the RHS Jacobian matrix to the IJacobian matrix.
5637: Logically Collective
5639: Input Parameters:
5640: + ts - the time-stepper
5641: - str - the structure (the default is `UNKNOWN_NONZERO_PATTERN`)
5643: Level: intermediate
5645: Note:
5646: When the relationship between the nonzero structures is known and supplied the solution process can be much faster
5648: .seealso: [](ch_ts), `TS`, `MatAXPY()`, `MatStructure`
5649: @*/
5650: PetscErrorCode TSSetMatStructure(TS ts, MatStructure str)
5651: {
5652: PetscFunctionBegin;
5654: ts->axpy_pattern = str;
5655: PetscFunctionReturn(PETSC_SUCCESS);
5656: }
5658: /*@
5659: TSSetTimeSpan - sets the time span. The solution will be computed and stored for each time requested in the span
5661: Collective
5663: Input Parameters:
5664: + ts - the time-stepper
5665: . n - number of the time points (>=2)
5666: - span_times - array of the time points. The first element and the last element are the initial time and the final time respectively.
5668: Options Database Key:
5669: . -ts_time_span <t0,...tf> - Sets the time span
5671: Level: intermediate
5673: Notes:
5674: The elements in tspan must be all increasing. They correspond to the intermediate points for time integration.
5675: `TS_EXACTFINALTIME_MATCHSTEP` must be used to make the last time step in each sub-interval match the intermediate points specified.
5676: The intermediate solutions are saved in a vector array that can be accessed with `TSGetTimeSpanSolutions()`. Thus using time span may
5677: pressure the memory system when using a large number of span points.
5679: .seealso: [](ch_ts), `TS`, `TSGetTimeSpan()`, `TSGetTimeSpanSolutions()`
5680: @*/
5681: PetscErrorCode TSSetTimeSpan(TS ts, PetscInt n, PetscReal *span_times)
5682: {
5683: PetscFunctionBegin;
5685: PetscCheck(n >= 2, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_WRONG, "Minimum time span size is 2 but %" PetscInt_FMT " is provided", n);
5686: if (ts->tspan && n != ts->tspan->num_span_times) {
5687: PetscCall(PetscFree(ts->tspan->span_times));
5688: PetscCall(VecDestroyVecs(ts->tspan->num_span_times, &ts->tspan->vecs_sol));
5689: PetscCall(PetscMalloc1(n, &ts->tspan->span_times));
5690: }
5691: if (!ts->tspan) {
5692: TSTimeSpan tspan;
5693: PetscCall(PetscNew(&tspan));
5694: PetscCall(PetscMalloc1(n, &tspan->span_times));
5695: tspan->reltol = 1e-6;
5696: tspan->abstol = 10 * PETSC_MACHINE_EPSILON;
5697: ts->tspan = tspan;
5698: }
5699: ts->tspan->num_span_times = n;
5700: PetscCall(PetscArraycpy(ts->tspan->span_times, span_times, n));
5701: PetscCall(TSSetTime(ts, ts->tspan->span_times[0]));
5702: PetscCall(TSSetMaxTime(ts, ts->tspan->span_times[n - 1]));
5703: PetscFunctionReturn(PETSC_SUCCESS);
5704: }
5706: /*@C
5707: TSGetTimeSpan - gets the time span set with `TSSetTimeSpan()`
5709: Not Collective
5711: Input Parameter:
5712: . ts - the time-stepper
5714: Output Parameters:
5715: + n - number of the time points (>=2)
5716: - span_times - array of the time points. The first element and the last element are the initial time and the final time respectively.
5718: Level: beginner
5720: Note:
5721: The values obtained are valid until the `TS` object is destroyed.
5723: Both `n` and `span_times` can be `NULL`.
5725: .seealso: [](ch_ts), `TS`, `TSSetTimeSpan()`, `TSGetTimeSpanSolutions()`
5726: @*/
5727: PetscErrorCode TSGetTimeSpan(TS ts, PetscInt *n, const PetscReal **span_times)
5728: {
5729: PetscFunctionBegin;
5731: if (n) PetscAssertPointer(n, 2);
5732: if (span_times) PetscAssertPointer(span_times, 3);
5733: if (!ts->tspan) {
5734: if (n) *n = 0;
5735: if (span_times) *span_times = NULL;
5736: } else {
5737: if (n) *n = ts->tspan->num_span_times;
5738: if (span_times) *span_times = ts->tspan->span_times;
5739: }
5740: PetscFunctionReturn(PETSC_SUCCESS);
5741: }
5743: /*@
5744: TSGetTimeSpanSolutions - Get the number of solutions and the solutions at the time points specified by the time span.
5746: Input Parameter:
5747: . ts - the `TS` context obtained from `TSCreate()`
5749: Output Parameters:
5750: + nsol - the number of solutions
5751: - Sols - the solution vectors
5753: Level: intermediate
5755: Notes:
5756: Both `nsol` and `Sols` can be `NULL`.
5758: 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()`.
5759: For example, manipulating the step size, especially with a reduced precision, may cause `TS` to step over certain points in the span.
5761: .seealso: [](ch_ts), `TS`, `TSSetTimeSpan()`
5762: @*/
5763: PetscErrorCode TSGetTimeSpanSolutions(TS ts, PetscInt *nsol, Vec **Sols)
5764: {
5765: PetscFunctionBegin;
5767: if (nsol) PetscAssertPointer(nsol, 2);
5768: if (Sols) PetscAssertPointer(Sols, 3);
5769: if (!ts->tspan) {
5770: if (nsol) *nsol = 0;
5771: if (Sols) *Sols = NULL;
5772: } else {
5773: if (nsol) *nsol = ts->tspan->spanctr;
5774: if (Sols) *Sols = ts->tspan->vecs_sol;
5775: }
5776: PetscFunctionReturn(PETSC_SUCCESS);
5777: }
5779: /*@C
5780: TSPruneIJacobianColor - Remove nondiagonal zeros in the Jacobian matrix and update the `MatMFFD` coloring information.
5782: Collective
5784: Input Parameters:
5785: + ts - the `TS` context
5786: . J - Jacobian matrix (not altered in this routine)
5787: - B - newly computed Jacobian matrix to use with preconditioner
5789: Level: intermediate
5791: Notes:
5792: This function improves the `MatFDColoring` performance when the Jacobian matrix was over-allocated or contains
5793: many constant zeros entries, which is typically the case when the matrix is generated by a `DM`
5794: and multiple fields are involved.
5796: Users need to make sure that the Jacobian matrix is properly filled to reflect the sparsity
5797: structure. For `MatFDColoring`, the values of nonzero entries are not important. So one can
5798: usually call `TSComputeIJacobian()` with randomized input vectors to generate a dummy Jacobian.
5799: `TSComputeIJacobian()` should be called before `TSSolve()` but after `TSSetUp()`.
5801: .seealso: [](ch_ts), `TS`, `MatFDColoring`, `TSComputeIJacobianDefaultColor()`, `MatEliminateZeros()`, `MatFDColoringCreate()`, `MatFDColoringSetFunction()`
5802: @*/
5803: PetscErrorCode TSPruneIJacobianColor(TS ts, Mat J, Mat B)
5804: {
5805: MatColoring mc = NULL;
5806: ISColoring iscoloring = NULL;
5807: MatFDColoring matfdcoloring = NULL;
5809: PetscFunctionBegin;
5810: /* Generate new coloring after eliminating zeros in the matrix */
5811: PetscCall(MatEliminateZeros(B, PETSC_TRUE));
5812: PetscCall(MatColoringCreate(B, &mc));
5813: PetscCall(MatColoringSetDistance(mc, 2));
5814: PetscCall(MatColoringSetType(mc, MATCOLORINGSL));
5815: PetscCall(MatColoringSetFromOptions(mc));
5816: PetscCall(MatColoringApply(mc, &iscoloring));
5817: PetscCall(MatColoringDestroy(&mc));
5818: /* Replace the old coloring with the new one */
5819: PetscCall(MatFDColoringCreate(B, iscoloring, &matfdcoloring));
5820: PetscCall(MatFDColoringSetFunction(matfdcoloring, (PetscErrorCode(*)(void))SNESTSFormFunction, (void *)ts));
5821: PetscCall(MatFDColoringSetFromOptions(matfdcoloring));
5822: PetscCall(MatFDColoringSetUp(B, iscoloring, matfdcoloring));
5823: PetscCall(PetscObjectCompose((PetscObject)B, "TSMatFDColoring", (PetscObject)matfdcoloring));
5824: PetscCall(PetscObjectDereference((PetscObject)matfdcoloring));
5825: PetscCall(ISColoringDestroy(&iscoloring));
5826: PetscFunctionReturn(PETSC_SUCCESS);
5827: }