Actual source code: adaptbasic.c
1: #include <petsc/private/tsimpl.h>
2: #include <petscdm.h>
4: static PetscErrorCode TSAdaptChoose_Basic(TSAdapt adapt, TS ts, PetscReal h, PetscInt *next_sc, PetscReal *next_h, PetscBool *accept, PetscReal *wlte, PetscReal *wltea, PetscReal *wlter)
5: {
6: Vec Y;
7: DM dm;
8: PetscInt order = PETSC_DECIDE;
9: PetscReal enorm = -1;
10: PetscReal enorma, enormr;
11: PetscReal safety = adapt->safety;
12: PetscReal hfac_lte, h_lte;
14: PetscFunctionBegin;
15: *next_sc = 0; /* Reuse the same order scheme */
16: *wltea = -1; /* Weighted absolute local truncation error is not used */
17: *wlter = -1; /* Weighted relative local truncation error is not used */
19: if (ts->ops->evaluatewlte) {
20: PetscCall(TSEvaluateWLTE(ts, adapt->wnormtype, &order, &enorm));
21: PetscCheck(enorm < 0 || order >= 1, PetscObjectComm((PetscObject)adapt), PETSC_ERR_ARG_OUTOFRANGE, "Computed error order %" PetscInt_FMT " must be positive", order);
22: } else if (ts->ops->evaluatestep) {
23: PetscCheck(adapt->candidates.n >= 1, PetscObjectComm((PetscObject)adapt), PETSC_ERR_ARG_WRONGSTATE, "No candidate has been registered");
24: PetscCheck(adapt->candidates.inuse_set, PetscObjectComm((PetscObject)adapt), PETSC_ERR_ARG_WRONGSTATE, "The current in-use scheme is not among the %" PetscInt_FMT " candidates", adapt->candidates.n);
25: order = adapt->candidates.order[0];
26: PetscCall(TSGetDM(ts, &dm));
27: PetscCall(DMGetGlobalVector(dm, &Y));
28: PetscCall(TSEvaluateStep(ts, order - 1, Y, NULL));
29: PetscCall(TSErrorWeightedNorm(ts, ts->vec_sol, Y, adapt->wnormtype, &enorm, &enorma, &enormr));
30: PetscCall(DMRestoreGlobalVector(dm, &Y));
31: }
33: if (enorm < 0) {
34: *accept = PETSC_TRUE;
35: *next_h = h; /* Reuse the old step */
36: *wlte = -1; /* Weighted local truncation error was not evaluated */
37: PetscFunctionReturn(PETSC_SUCCESS);
38: }
40: /* Determine whether the step is accepted of rejected */
41: if (enorm > 1) {
42: if (!*accept) safety *= adapt->reject_safety; /* The last attempt also failed, shorten more aggressively */
43: if (h < (1 + PETSC_SQRT_MACHINE_EPSILON) * adapt->dt_min) {
44: PetscCall(PetscInfo(adapt, "Estimated scaled local truncation error %g, accepting because step size %g is at minimum\n", (double)enorm, (double)h));
45: *accept = PETSC_TRUE;
46: } else if (adapt->always_accept) {
47: PetscCall(PetscInfo(adapt, "Estimated scaled local truncation error %g, accepting step of size %g because always_accept is set\n", (double)enorm, (double)h));
48: *accept = PETSC_TRUE;
49: } else {
50: PetscCall(PetscInfo(adapt, "Estimated scaled local truncation error %g, rejecting step of size %g\n", (double)enorm, (double)h));
51: *accept = PETSC_FALSE;
52: }
53: } else {
54: PetscCall(PetscInfo(adapt, "Estimated scaled local truncation error %g, accepting step of size %g\n", (double)enorm, (double)h));
55: *accept = PETSC_TRUE;
56: }
58: /* The optimal new step based purely on local truncation error for this step. */
59: if (enorm > 0) hfac_lte = safety * PetscPowReal(enorm, ((PetscReal)-1) / order);
60: else hfac_lte = safety * PETSC_INFINITY;
61: if (adapt->timestepjustdecreased) {
62: hfac_lte = PetscMin(hfac_lte, 1.0);
63: adapt->timestepjustdecreased--;
64: }
65: h_lte = h * PetscClipInterval(hfac_lte, adapt->clip[0], adapt->clip[1]);
67: *next_h = PetscClipInterval(h_lte, adapt->dt_min, adapt->dt_max);
68: *wlte = enorm;
69: PetscFunctionReturn(PETSC_SUCCESS);
70: }
72: /*MC
73: TSADAPTBASIC - Basic adaptive controller for time stepping
75: Level: intermediate
77: .seealso: [](ch_ts), `TS`, `TSAdapt`, `TSGetAdapt()`, `TSAdaptType`
78: M*/
79: PETSC_EXTERN PetscErrorCode TSAdaptCreate_Basic(TSAdapt adapt)
80: {
81: PetscFunctionBegin;
82: adapt->ops->choose = TSAdaptChoose_Basic;
83: PetscFunctionReturn(PETSC_SUCCESS);
84: }