1 2 #include <petsc/private/tsimpl.h> /*I "petscts.h" I*/ 3 #include <petscdmshell.h> 4 #include <petscdmda.h> 5 #include <petscviewer.h> 6 #include <petscdraw.h> 7 8 /* Logging support */ 9 PetscClassId TS_CLASSID, DMTS_CLASSID; 10 PetscLogEvent TS_Step, TS_PseudoComputeTimeStep, TS_FunctionEval, TS_JacobianEval; 11 12 const char *const TSExactFinalTimeOptions[] = {"STEPOVER","INTERPOLATE","MATCHSTEP","TSExactFinalTimeOption","TS_EXACTFINALTIME_",0}; 13 14 struct _n_TSMonitorDrawCtx { 15 PetscViewer viewer; 16 PetscDrawAxis axis; 17 Vec initialsolution; 18 PetscBool showinitial; 19 PetscInt howoften; /* when > 0 uses step % howoften, when negative only final solution plotted */ 20 PetscBool showtimestepandtime; 21 int color; 22 }; 23 24 #undef __FUNCT__ 25 #define __FUNCT__ "TSSetFromOptions" 26 /*@ 27 TSSetFromOptions - Sets various TS parameters from user options. 28 29 Collective on TS 30 31 Input Parameter: 32 . ts - the TS context obtained from TSCreate() 33 34 Options Database Keys: 35 + -ts_type <type> - TSEULER, TSBEULER, TSSUNDIALS, TSPSEUDO, TSCN, TSRK, TSTHETA, TSGL, TSSSP 36 . -ts_save_trajectory - checkpoint the solution at each time-step 37 . -ts_max_steps <maxsteps> - maximum number of time-steps to take 38 . -ts_final_time <time> - maximum time to compute to 39 . -ts_dt <dt> - initial time step 40 . -ts_exact_final_time <stepover,interpolate,matchstep> whether to stop at the exact given final time and how to compute the solution at that ti,e 41 . -ts_max_snes_failures <maxfailures> - Maximum number of nonlinear solve failures allowed 42 . -ts_max_reject <maxrejects> - Maximum number of step rejections before step fails 43 . -ts_error_if_step_fails <true,false> - Error if no step succeeds 44 . -ts_rtol <rtol> - relative tolerance for local truncation error 45 . -ts_atol <atol> Absolute tolerance for local truncation error 46 . -ts_adjoint_solve <yes,no> After solving the ODE/DAE solve the adjoint problem (requires -ts_save_trajectory) 47 . -ts_fd_color - Use finite differences with coloring to compute IJacobian 48 . -ts_monitor - print information at each timestep 49 . -ts_monitor_lg_timestep - Monitor timestep size graphically 50 . -ts_monitor_lg_solution - Monitor solution graphically 51 . -ts_monitor_lg_error - Monitor error graphically 52 . -ts_monitor_lg_snes_iterations - Monitor number nonlinear iterations for each timestep graphically 53 . -ts_monitor_lg_ksp_iterations - Monitor number nonlinear iterations for each timestep graphically 54 . -ts_monitor_sp_eig - Monitor eigenvalues of linearized operator graphically 55 . -ts_monitor_draw_solution - Monitor solution graphically 56 . -ts_monitor_draw_solution_phase <xleft,yleft,xright,yright> - Monitor solution graphically with phase diagram, requires problem with exactly 2 degrees of freedom 57 . -ts_monitor_draw_error - Monitor error graphically, requires use to have provided TSSetSolutionFunction() 58 . -ts_monitor_solution_binary <filename> - Save each solution to a binary file 59 . -ts_monitor_solution_vtk <filename.vts> - Save each time step to a binary file, use filename-%%03D.vts 60 - -ts_monitor_envelope - determine maximum and minimum value of each component of the solution over the solution time 61 62 Developer Note: We should unify all the -ts_monitor options in the way that -xxx_view has been unified 63 64 Level: beginner 65 66 .keywords: TS, timestep, set, options, database 67 68 .seealso: TSGetType() 69 @*/ 70 PetscErrorCode TSSetFromOptions(TS ts) 71 { 72 PetscBool opt,flg,tflg; 73 PetscErrorCode ierr; 74 PetscViewer monviewer; 75 char monfilename[PETSC_MAX_PATH_LEN]; 76 SNES snes; 77 TSAdapt adapt; 78 PetscReal time_step; 79 TSExactFinalTimeOption eftopt; 80 char dir[16]; 81 const char *defaultType; 82 char typeName[256]; 83 84 PetscFunctionBegin; 85 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 86 ierr = PetscObjectOptionsBegin((PetscObject)ts);CHKERRQ(ierr); 87 if (((PetscObject)ts)->type_name) defaultType = ((PetscObject)ts)->type_name; 88 else defaultType = TSEULER; 89 90 ierr = TSRegisterAll();CHKERRQ(ierr); 91 ierr = PetscOptionsFList("-ts_type", "TS method"," TSSetType", TSList, defaultType, typeName, 256, &opt);CHKERRQ(ierr); 92 if (opt) { 93 ierr = TSSetType(ts, typeName);CHKERRQ(ierr); 94 } else { 95 ierr = TSSetType(ts, defaultType);CHKERRQ(ierr); 96 } 97 98 /* Handle generic TS options */ 99 if (ts->trajectory) tflg = PETSC_TRUE; 100 else tflg = PETSC_FALSE; 101 ierr = PetscOptionsBool("-ts_save_trajectory","Save the solution at each timestep","TSSetSaveTrajectory",tflg,&tflg,NULL);CHKERRQ(ierr); 102 if (tflg) {ierr = TSSetSaveTrajectory(ts);CHKERRQ(ierr);} 103 if (ts->adjoint_solve) tflg = PETSC_TRUE; 104 else tflg = PETSC_FALSE; 105 ierr = PetscOptionsBool("-ts_adjoint_solve","Solve the adjoint problem immediately after solving the forward problem","",tflg,&tflg,&flg);CHKERRQ(ierr); 106 if (flg) { 107 ierr = TSSetSaveTrajectory(ts);CHKERRQ(ierr); 108 ts->adjoint_solve = tflg; 109 } 110 ierr = PetscOptionsInt("-ts_max_steps","Maximum number of time steps","TSSetDuration",ts->max_steps,&ts->max_steps,NULL);CHKERRQ(ierr); 111 ierr = PetscOptionsReal("-ts_final_time","Time to run to","TSSetDuration",ts->max_time,&ts->max_time,NULL);CHKERRQ(ierr); 112 ierr = PetscOptionsReal("-ts_init_time","Initial time","TSSetTime",ts->ptime,&ts->ptime,NULL);CHKERRQ(ierr); 113 ierr = PetscOptionsReal("-ts_dt","Initial time step","TSSetTimeStep",ts->time_step,&time_step,&flg);CHKERRQ(ierr); 114 if (flg) { 115 ierr = TSSetTimeStep(ts,time_step);CHKERRQ(ierr); 116 } 117 ierr = PetscOptionsEnum("-ts_exact_final_time","Option for handling of final time step","TSSetExactFinalTime",TSExactFinalTimeOptions,(PetscEnum)ts->exact_final_time,(PetscEnum*)&eftopt,&flg);CHKERRQ(ierr); 118 if (flg) {ierr = TSSetExactFinalTime(ts,eftopt);CHKERRQ(ierr);} 119 ierr = PetscOptionsInt("-ts_max_snes_failures","Maximum number of nonlinear solve failures","TSSetMaxSNESFailures",ts->max_snes_failures,&ts->max_snes_failures,NULL);CHKERRQ(ierr); 120 ierr = PetscOptionsInt("-ts_max_reject","Maximum number of step rejections before step fails","TSSetMaxStepRejections",ts->max_reject,&ts->max_reject,NULL);CHKERRQ(ierr); 121 ierr = PetscOptionsBool("-ts_error_if_step_fails","Error if no step succeeds","TSSetErrorIfStepFails",ts->errorifstepfailed,&ts->errorifstepfailed,NULL);CHKERRQ(ierr); 122 ierr = PetscOptionsReal("-ts_rtol","Relative tolerance for local truncation error","TSSetTolerances",ts->rtol,&ts->rtol,NULL);CHKERRQ(ierr); 123 ierr = PetscOptionsReal("-ts_atol","Absolute tolerance for local truncation error","TSSetTolerances",ts->atol,&ts->atol,NULL);CHKERRQ(ierr); 124 125 #if defined(PETSC_HAVE_SAWS) 126 { 127 PetscBool set; 128 flg = PETSC_FALSE; 129 ierr = PetscOptionsBool("-ts_saws_block","Block for SAWs memory snooper at end of TSSolve","PetscObjectSAWsBlock",((PetscObject)ts)->amspublishblock,&flg,&set);CHKERRQ(ierr); 130 if (set) { 131 ierr = PetscObjectSAWsSetBlock((PetscObject)ts,flg);CHKERRQ(ierr); 132 } 133 } 134 #endif 135 136 /* Monitor options */ 137 ierr = PetscOptionsString("-ts_monitor","Monitor timestep size","TSMonitorDefault","stdout",monfilename,PETSC_MAX_PATH_LEN,&flg);CHKERRQ(ierr); 138 if (flg) { 139 ierr = PetscViewerASCIIOpen(PetscObjectComm((PetscObject)ts),monfilename,&monviewer);CHKERRQ(ierr); 140 ierr = TSMonitorSet(ts,TSMonitorDefault,monviewer,(PetscErrorCode (*)(void**))PetscViewerDestroy);CHKERRQ(ierr); 141 } 142 ierr = PetscOptionsString("-ts_monitor_python","Use Python function","TSMonitorSet",0,monfilename,PETSC_MAX_PATH_LEN,&flg);CHKERRQ(ierr); 143 if (flg) {ierr = PetscPythonMonitorSet((PetscObject)ts,monfilename);CHKERRQ(ierr);} 144 145 ierr = PetscOptionsName("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",&opt);CHKERRQ(ierr); 146 if (opt) { 147 TSMonitorLGCtx ctx; 148 PetscInt howoften = 1; 149 150 ierr = PetscOptionsInt("-ts_monitor_lg_timestep","Monitor timestep size graphically","TSMonitorLGTimeStep",howoften,&howoften,NULL);CHKERRQ(ierr); 151 ierr = TSMonitorLGCtxCreate(PetscObjectComm((PetscObject)ts),0,0,PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);CHKERRQ(ierr); 152 ierr = TSMonitorSet(ts,TSMonitorLGTimeStep,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);CHKERRQ(ierr); 153 } 154 ierr = PetscOptionsName("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",&opt);CHKERRQ(ierr); 155 if (opt) { 156 TSMonitorLGCtx ctx; 157 PetscInt howoften = 1; 158 159 ierr = PetscOptionsInt("-ts_monitor_lg_solution","Monitor solution graphically","TSMonitorLGSolution",howoften,&howoften,NULL);CHKERRQ(ierr); 160 ierr = TSMonitorLGCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);CHKERRQ(ierr); 161 ierr = TSMonitorSet(ts,TSMonitorLGSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);CHKERRQ(ierr); 162 } 163 ierr = PetscOptionsName("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",&opt);CHKERRQ(ierr); 164 if (opt) { 165 TSMonitorLGCtx ctx; 166 PetscInt howoften = 1; 167 168 ierr = PetscOptionsInt("-ts_monitor_lg_error","Monitor error graphically","TSMonitorLGError",howoften,&howoften,NULL);CHKERRQ(ierr); 169 ierr = TSMonitorLGCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);CHKERRQ(ierr); 170 ierr = TSMonitorSet(ts,TSMonitorLGError,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);CHKERRQ(ierr); 171 } 172 ierr = PetscOptionsName("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",&opt);CHKERRQ(ierr); 173 if (opt) { 174 TSMonitorLGCtx ctx; 175 PetscInt howoften = 1; 176 177 ierr = PetscOptionsInt("-ts_monitor_lg_snes_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGSNESIterations",howoften,&howoften,NULL);CHKERRQ(ierr); 178 ierr = TSMonitorLGCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);CHKERRQ(ierr); 179 ierr = TSMonitorSet(ts,TSMonitorLGSNESIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);CHKERRQ(ierr); 180 } 181 ierr = PetscOptionsName("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",&opt);CHKERRQ(ierr); 182 if (opt) { 183 TSMonitorLGCtx ctx; 184 PetscInt howoften = 1; 185 186 ierr = PetscOptionsInt("-ts_monitor_lg_ksp_iterations","Monitor number nonlinear iterations for each timestep graphically","TSMonitorLGKSPIterations",howoften,&howoften,NULL);CHKERRQ(ierr); 187 ierr = TSMonitorLGCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,300,300,howoften,&ctx);CHKERRQ(ierr); 188 ierr = TSMonitorSet(ts,TSMonitorLGKSPIterations,ctx,(PetscErrorCode (*)(void**))TSMonitorLGCtxDestroy);CHKERRQ(ierr); 189 } 190 ierr = PetscOptionsName("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",&opt);CHKERRQ(ierr); 191 if (opt) { 192 TSMonitorSPEigCtx ctx; 193 PetscInt howoften = 1; 194 195 ierr = PetscOptionsInt("-ts_monitor_sp_eig","Monitor eigenvalues of linearized operator graphically","TSMonitorSPEig",howoften,&howoften,NULL);CHKERRQ(ierr); 196 ierr = TSMonitorSPEigCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);CHKERRQ(ierr); 197 ierr = TSMonitorSet(ts,TSMonitorSPEig,ctx,(PetscErrorCode (*)(void**))TSMonitorSPEigCtxDestroy);CHKERRQ(ierr); 198 } 199 opt = PETSC_FALSE; 200 ierr = PetscOptionsName("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",&opt);CHKERRQ(ierr); 201 if (opt) { 202 TSMonitorDrawCtx ctx; 203 PetscInt howoften = 1; 204 205 ierr = PetscOptionsInt("-ts_monitor_draw_solution","Monitor solution graphically","TSMonitorDrawSolution",howoften,&howoften,NULL);CHKERRQ(ierr); 206 ierr = TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);CHKERRQ(ierr); 207 ierr = TSMonitorSet(ts,TSMonitorDrawSolution,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);CHKERRQ(ierr); 208 } 209 opt = PETSC_FALSE; 210 ierr = PetscOptionsName("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",&opt);CHKERRQ(ierr); 211 if (opt) { 212 TSMonitorDrawCtx ctx; 213 PetscReal bounds[4]; 214 PetscInt n = 4; 215 PetscDraw draw; 216 217 ierr = PetscOptionsRealArray("-ts_monitor_draw_solution_phase","Monitor solution graphically","TSMonitorDrawSolutionPhase",bounds,&n,NULL);CHKERRQ(ierr); 218 if (n != 4) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Must provide bounding box of phase field"); 219 ierr = TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,1,&ctx);CHKERRQ(ierr); 220 ierr = PetscViewerDrawGetDraw(ctx->viewer,0,&draw);CHKERRQ(ierr); 221 ierr = PetscDrawClear(draw);CHKERRQ(ierr); 222 ierr = PetscDrawAxisCreate(draw,&ctx->axis);CHKERRQ(ierr); 223 ierr = PetscDrawAxisSetLimits(ctx->axis,bounds[0],bounds[2],bounds[1],bounds[3]);CHKERRQ(ierr); 224 ierr = PetscDrawAxisSetLabels(ctx->axis,"Phase Diagram","Variable 1","Variable 2");CHKERRQ(ierr); 225 ierr = PetscDrawAxisDraw(ctx->axis);CHKERRQ(ierr); 226 /* ierr = PetscDrawSetCoordinates(draw,bounds[0],bounds[1],bounds[2],bounds[3]);CHKERRQ(ierr); */ 227 ierr = TSMonitorSet(ts,TSMonitorDrawSolutionPhase,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);CHKERRQ(ierr); 228 } 229 opt = PETSC_FALSE; 230 ierr = PetscOptionsName("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",&opt);CHKERRQ(ierr); 231 if (opt) { 232 TSMonitorDrawCtx ctx; 233 PetscInt howoften = 1; 234 235 ierr = PetscOptionsInt("-ts_monitor_draw_error","Monitor error graphically","TSMonitorDrawError",howoften,&howoften,NULL);CHKERRQ(ierr); 236 ierr = TSMonitorDrawCtxCreate(PetscObjectComm((PetscObject)ts),0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&ctx);CHKERRQ(ierr); 237 ierr = TSMonitorSet(ts,TSMonitorDrawError,ctx,(PetscErrorCode (*)(void**))TSMonitorDrawCtxDestroy);CHKERRQ(ierr); 238 } 239 opt = PETSC_FALSE; 240 ierr = PetscOptionsString("-ts_monitor_solution_binary","Save each solution to a binary file","TSMonitorSolutionBinary",0,monfilename,PETSC_MAX_PATH_LEN,&flg);CHKERRQ(ierr); 241 if (flg) { 242 PetscViewer ctx; 243 if (monfilename[0]) { 244 ierr = PetscViewerBinaryOpen(PetscObjectComm((PetscObject)ts),monfilename,FILE_MODE_WRITE,&ctx);CHKERRQ(ierr); 245 ierr = TSMonitorSet(ts,TSMonitorSolutionBinary,ctx,(PetscErrorCode (*)(void**))PetscViewerDestroy);CHKERRQ(ierr); 246 } else { 247 ctx = PETSC_VIEWER_BINARY_(PetscObjectComm((PetscObject)ts)); 248 ierr = TSMonitorSet(ts,TSMonitorSolutionBinary,ctx,(PetscErrorCode (*)(void**))NULL);CHKERRQ(ierr); 249 } 250 } 251 opt = PETSC_FALSE; 252 ierr = PetscOptionsString("-ts_monitor_solution_vtk","Save each time step to a binary file, use filename-%%03D.vts","TSMonitorSolutionVTK",0,monfilename,PETSC_MAX_PATH_LEN,&flg);CHKERRQ(ierr); 253 if (flg) { 254 const char *ptr,*ptr2; 255 char *filetemplate; 256 if (!monfilename[0]) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03D.vts"); 257 /* Do some cursory validation of the input. */ 258 ierr = PetscStrstr(monfilename,"%",(char**)&ptr);CHKERRQ(ierr); 259 if (!ptr) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"-ts_monitor_solution_vtk requires a file template, e.g. filename-%%03D.vts"); 260 for (ptr++; ptr && *ptr; ptr++) { 261 ierr = PetscStrchr("DdiouxX",*ptr,(char**)&ptr2);CHKERRQ(ierr); 262 if (!ptr2 && (*ptr < '0' || '9' < *ptr)) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Invalid file template argument to -ts_monitor_solution_vtk, should look like filename-%%03D.vts"); 263 if (ptr2) break; 264 } 265 ierr = PetscStrallocpy(monfilename,&filetemplate);CHKERRQ(ierr); 266 ierr = TSMonitorSet(ts,TSMonitorSolutionVTK,filetemplate,(PetscErrorCode (*)(void**))TSMonitorSolutionVTKDestroy);CHKERRQ(ierr); 267 } 268 269 ierr = PetscOptionsString("-ts_monitor_dmda_ray","Display a ray of the solution","None","y=0",dir,16,&flg);CHKERRQ(ierr); 270 if (flg) { 271 TSMonitorDMDARayCtx *rayctx; 272 int ray = 0; 273 DMDADirection ddir; 274 DM da; 275 PetscMPIInt rank; 276 277 if (dir[1] != '=') SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Unknown ray %s",dir); 278 if (dir[0] == 'x') ddir = DMDA_X; 279 else if (dir[0] == 'y') ddir = DMDA_Y; 280 else SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Unknown ray %s",dir); 281 sscanf(dir+2,"%d",&ray); 282 283 ierr = PetscInfo2(((PetscObject)ts),"Displaying DMDA ray %c = %D\n",dir[0],ray);CHKERRQ(ierr); 284 ierr = PetscNew(&rayctx);CHKERRQ(ierr); 285 ierr = TSGetDM(ts,&da);CHKERRQ(ierr); 286 ierr = DMDAGetRay(da,ddir,ray,&rayctx->ray,&rayctx->scatter);CHKERRQ(ierr); 287 ierr = MPI_Comm_rank(PetscObjectComm((PetscObject)ts),&rank);CHKERRQ(ierr); 288 if (!rank) { 289 ierr = PetscViewerDrawOpen(PETSC_COMM_SELF,0,0,0,0,600,300,&rayctx->viewer);CHKERRQ(ierr); 290 } 291 rayctx->lgctx = NULL; 292 ierr = TSMonitorSet(ts,TSMonitorDMDARay,rayctx,TSMonitorDMDARayDestroy);CHKERRQ(ierr); 293 } 294 ierr = PetscOptionsString("-ts_monitor_lg_dmda_ray","Display a ray of the solution","None","x=0",dir,16,&flg);CHKERRQ(ierr); 295 if (flg) { 296 TSMonitorDMDARayCtx *rayctx; 297 int ray = 0; 298 DMDADirection ddir; 299 DM da; 300 PetscInt howoften = 1; 301 302 if (dir[1] != '=') SETERRQ1(PetscObjectComm((PetscObject) ts), PETSC_ERR_ARG_WRONG, "Malformed ray %s", dir); 303 if (dir[0] == 'x') ddir = DMDA_X; 304 else if (dir[0] == 'y') ddir = DMDA_Y; 305 else SETERRQ1(PetscObjectComm((PetscObject) ts), PETSC_ERR_ARG_WRONG, "Unknown ray direction %s", dir); 306 sscanf(dir+2, "%d", &ray); 307 308 ierr = PetscInfo2(((PetscObject) ts),"Displaying LG DMDA ray %c = %D\n", dir[0], ray);CHKERRQ(ierr); 309 ierr = PetscNew(&rayctx);CHKERRQ(ierr); 310 ierr = TSGetDM(ts, &da);CHKERRQ(ierr); 311 ierr = DMDAGetRay(da, ddir, ray, &rayctx->ray, &rayctx->scatter);CHKERRQ(ierr); 312 ierr = TSMonitorLGCtxCreate(PETSC_COMM_SELF,0,0,PETSC_DECIDE,PETSC_DECIDE,600,400,howoften,&rayctx->lgctx);CHKERRQ(ierr); 313 ierr = TSMonitorSet(ts, TSMonitorLGDMDARay, rayctx, TSMonitorDMDARayDestroy);CHKERRQ(ierr); 314 } 315 316 ierr = PetscOptionsName("-ts_monitor_envelope","Monitor maximum and minimum value of each component of the solution","TSMonitorEnvelope",&opt);CHKERRQ(ierr); 317 if (opt) { 318 TSMonitorEnvelopeCtx ctx; 319 320 ierr = TSMonitorEnvelopeCtxCreate(ts,&ctx);CHKERRQ(ierr); 321 ierr = TSMonitorSet(ts,TSMonitorEnvelope,ctx,(PetscErrorCode (*)(void**))TSMonitorEnvelopeCtxDestroy);CHKERRQ(ierr); 322 } 323 324 flg = PETSC_FALSE; 325 ierr = PetscOptionsBool("-ts_fd_color", "Use finite differences with coloring to compute IJacobian", "TSComputeJacobianDefaultColor", flg, &flg, NULL);CHKERRQ(ierr); 326 if (flg) { 327 DM dm; 328 DMTS tdm; 329 330 ierr = TSGetDM(ts, &dm);CHKERRQ(ierr); 331 ierr = DMGetDMTS(dm, &tdm);CHKERRQ(ierr); 332 tdm->ijacobianctx = NULL; 333 ierr = TSSetIJacobian(ts, NULL, NULL, TSComputeIJacobianDefaultColor, 0);CHKERRQ(ierr); 334 ierr = PetscInfo(ts, "Setting default finite difference coloring Jacobian matrix\n");CHKERRQ(ierr); 335 } 336 337 /* 338 This code is all wrong. One is creating objects inside the TSSetFromOptions() so if run with the options gui 339 will bleed memory. Also one is using a PetscOptionsBegin() inside a PetscOptionsBegin() 340 */ 341 ierr = TSGetAdapt(ts,&adapt);CHKERRQ(ierr); 342 ierr = TSAdaptSetFromOptions(PetscOptionsObject,adapt);CHKERRQ(ierr); 343 344 /* Handle specific TS options */ 345 if (ts->ops->setfromoptions) { 346 ierr = (*ts->ops->setfromoptions)(PetscOptionsObject,ts);CHKERRQ(ierr); 347 } 348 ierr = PetscOptionsEnd();CHKERRQ(ierr); 349 350 /* process any options handlers added with PetscObjectAddOptionsHandler() */ 351 ierr = PetscObjectProcessOptionsHandlers((PetscObject)ts);CHKERRQ(ierr); 352 353 if (ts->trajectory) { 354 ierr = TSTrajectorySetFromOptions(ts->trajectory);CHKERRQ(ierr); 355 } 356 357 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 358 if (snes) { 359 if (ts->problem_type == TS_LINEAR) {ierr = SNESSetType(snes,SNESKSPONLY);CHKERRQ(ierr);} 360 ierr = SNESSetFromOptions(ts->snes);CHKERRQ(ierr); 361 } 362 PetscFunctionReturn(0); 363 } 364 365 #undef __FUNCT__ 366 #define __FUNCT__ "TSSetSaveTrajectory" 367 /*@ 368 TSSetSaveTrajectory - Causes the TS to save its solutions as it iterates forward in time in a TSTrajectory object 369 370 Collective on TS 371 372 Input Parameters: 373 . ts - the TS context obtained from TSCreate() 374 375 376 Level: intermediate 377 378 .seealso: TSGetTrajectory(), TSAdjointSolve() 379 380 .keywords: TS, set, checkpoint, 381 @*/ 382 PetscErrorCode TSSetSaveTrajectory(TS ts) 383 { 384 PetscErrorCode ierr; 385 386 PetscFunctionBegin; 387 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 388 if (!ts->trajectory) { 389 ierr = TSTrajectoryCreate(PetscObjectComm((PetscObject)ts),&ts->trajectory);CHKERRQ(ierr); 390 ierr = TSTrajectorySetType(ts->trajectory,TSTRAJECTORYBASIC);CHKERRQ(ierr); 391 } 392 PetscFunctionReturn(0); 393 } 394 395 #undef __FUNCT__ 396 #define __FUNCT__ "TSComputeRHSJacobian" 397 /*@ 398 TSComputeRHSJacobian - Computes the Jacobian matrix that has been 399 set with TSSetRHSJacobian(). 400 401 Collective on TS and Vec 402 403 Input Parameters: 404 + ts - the TS context 405 . t - current timestep 406 - U - input vector 407 408 Output Parameters: 409 + A - Jacobian matrix 410 . B - optional preconditioning matrix 411 - flag - flag indicating matrix structure 412 413 Notes: 414 Most users should not need to explicitly call this routine, as it 415 is used internally within the nonlinear solvers. 416 417 See KSPSetOperators() for important information about setting the 418 flag parameter. 419 420 Level: developer 421 422 .keywords: SNES, compute, Jacobian, matrix 423 424 .seealso: TSSetRHSJacobian(), KSPSetOperators() 425 @*/ 426 PetscErrorCode TSComputeRHSJacobian(TS ts,PetscReal t,Vec U,Mat A,Mat B) 427 { 428 PetscErrorCode ierr; 429 PetscObjectState Ustate; 430 DM dm; 431 DMTS tsdm; 432 TSRHSJacobian rhsjacobianfunc; 433 void *ctx; 434 TSIJacobian ijacobianfunc; 435 436 PetscFunctionBegin; 437 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 438 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 439 PetscCheckSameComm(ts,1,U,3); 440 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 441 ierr = DMGetDMTS(dm,&tsdm);CHKERRQ(ierr); 442 ierr = DMTSGetRHSJacobian(dm,&rhsjacobianfunc,&ctx);CHKERRQ(ierr); 443 ierr = DMTSGetIJacobian(dm,&ijacobianfunc,NULL);CHKERRQ(ierr); 444 ierr = PetscObjectStateGet((PetscObject)U,&Ustate);CHKERRQ(ierr); 445 if (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR || (ts->rhsjacobian.X == U && ts->rhsjacobian.Xstate == Ustate))) { 446 PetscFunctionReturn(0); 447 } 448 449 if (!rhsjacobianfunc && !ijacobianfunc) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSJacobian() and / or TSSetIJacobian()"); 450 451 if (ts->rhsjacobian.reuse) { 452 ierr = MatShift(A,-ts->rhsjacobian.shift);CHKERRQ(ierr); 453 ierr = MatScale(A,1./ts->rhsjacobian.scale);CHKERRQ(ierr); 454 if (A != B) { 455 ierr = MatShift(B,-ts->rhsjacobian.shift);CHKERRQ(ierr); 456 ierr = MatScale(B,1./ts->rhsjacobian.scale);CHKERRQ(ierr); 457 } 458 ts->rhsjacobian.shift = 0; 459 ts->rhsjacobian.scale = 1.; 460 } 461 462 if (rhsjacobianfunc) { 463 ierr = PetscLogEventBegin(TS_JacobianEval,ts,U,A,B);CHKERRQ(ierr); 464 PetscStackPush("TS user Jacobian function"); 465 ierr = (*rhsjacobianfunc)(ts,t,U,A,B,ctx);CHKERRQ(ierr); 466 PetscStackPop; 467 ierr = PetscLogEventEnd(TS_JacobianEval,ts,U,A,B);CHKERRQ(ierr); 468 /* make sure user returned a correct Jacobian and preconditioner */ 469 PetscValidHeaderSpecific(A,MAT_CLASSID,4); 470 PetscValidHeaderSpecific(B,MAT_CLASSID,5); 471 } else { 472 ierr = MatZeroEntries(A);CHKERRQ(ierr); 473 if (A != B) {ierr = MatZeroEntries(B);CHKERRQ(ierr);} 474 } 475 ts->rhsjacobian.time = t; 476 ts->rhsjacobian.X = U; 477 ierr = PetscObjectStateGet((PetscObject)U,&ts->rhsjacobian.Xstate);CHKERRQ(ierr); 478 PetscFunctionReturn(0); 479 } 480 481 #undef __FUNCT__ 482 #define __FUNCT__ "TSComputeRHSFunction" 483 /*@ 484 TSComputeRHSFunction - Evaluates the right-hand-side function. 485 486 Collective on TS and Vec 487 488 Input Parameters: 489 + ts - the TS context 490 . t - current time 491 - U - state vector 492 493 Output Parameter: 494 . y - right hand side 495 496 Note: 497 Most users should not need to explicitly call this routine, as it 498 is used internally within the nonlinear solvers. 499 500 Level: developer 501 502 .keywords: TS, compute 503 504 .seealso: TSSetRHSFunction(), TSComputeIFunction() 505 @*/ 506 PetscErrorCode TSComputeRHSFunction(TS ts,PetscReal t,Vec U,Vec y) 507 { 508 PetscErrorCode ierr; 509 TSRHSFunction rhsfunction; 510 TSIFunction ifunction; 511 void *ctx; 512 DM dm; 513 514 PetscFunctionBegin; 515 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 516 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 517 PetscValidHeaderSpecific(y,VEC_CLASSID,4); 518 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 519 ierr = DMTSGetRHSFunction(dm,&rhsfunction,&ctx);CHKERRQ(ierr); 520 ierr = DMTSGetIFunction(dm,&ifunction,NULL);CHKERRQ(ierr); 521 522 if (!rhsfunction && !ifunction) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSFunction() and / or TSSetIFunction()"); 523 524 ierr = PetscLogEventBegin(TS_FunctionEval,ts,U,y,0);CHKERRQ(ierr); 525 if (rhsfunction) { 526 PetscStackPush("TS user right-hand-side function"); 527 ierr = (*rhsfunction)(ts,t,U,y,ctx);CHKERRQ(ierr); 528 PetscStackPop; 529 } else { 530 ierr = VecZeroEntries(y);CHKERRQ(ierr); 531 } 532 533 ierr = PetscLogEventEnd(TS_FunctionEval,ts,U,y,0);CHKERRQ(ierr); 534 PetscFunctionReturn(0); 535 } 536 537 #undef __FUNCT__ 538 #define __FUNCT__ "TSComputeSolutionFunction" 539 /*@ 540 TSComputeSolutionFunction - Evaluates the solution function. 541 542 Collective on TS and Vec 543 544 Input Parameters: 545 + ts - the TS context 546 - t - current time 547 548 Output Parameter: 549 . U - the solution 550 551 Note: 552 Most users should not need to explicitly call this routine, as it 553 is used internally within the nonlinear solvers. 554 555 Level: developer 556 557 .keywords: TS, compute 558 559 .seealso: TSSetSolutionFunction(), TSSetRHSFunction(), TSComputeIFunction() 560 @*/ 561 PetscErrorCode TSComputeSolutionFunction(TS ts,PetscReal t,Vec U) 562 { 563 PetscErrorCode ierr; 564 TSSolutionFunction solutionfunction; 565 void *ctx; 566 DM dm; 567 568 PetscFunctionBegin; 569 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 570 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 571 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 572 ierr = DMTSGetSolutionFunction(dm,&solutionfunction,&ctx);CHKERRQ(ierr); 573 574 if (solutionfunction) { 575 PetscStackPush("TS user solution function"); 576 ierr = (*solutionfunction)(ts,t,U,ctx);CHKERRQ(ierr); 577 PetscStackPop; 578 } 579 PetscFunctionReturn(0); 580 } 581 #undef __FUNCT__ 582 #define __FUNCT__ "TSComputeForcingFunction" 583 /*@ 584 TSComputeForcingFunction - Evaluates the forcing function. 585 586 Collective on TS and Vec 587 588 Input Parameters: 589 + ts - the TS context 590 - t - current time 591 592 Output Parameter: 593 . U - the function value 594 595 Note: 596 Most users should not need to explicitly call this routine, as it 597 is used internally within the nonlinear solvers. 598 599 Level: developer 600 601 .keywords: TS, compute 602 603 .seealso: TSSetSolutionFunction(), TSSetRHSFunction(), TSComputeIFunction() 604 @*/ 605 PetscErrorCode TSComputeForcingFunction(TS ts,PetscReal t,Vec U) 606 { 607 PetscErrorCode ierr, (*forcing)(TS,PetscReal,Vec,void*); 608 void *ctx; 609 DM dm; 610 611 PetscFunctionBegin; 612 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 613 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 614 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 615 ierr = DMTSGetForcingFunction(dm,&forcing,&ctx);CHKERRQ(ierr); 616 617 if (forcing) { 618 PetscStackPush("TS user forcing function"); 619 ierr = (*forcing)(ts,t,U,ctx);CHKERRQ(ierr); 620 PetscStackPop; 621 } 622 PetscFunctionReturn(0); 623 } 624 625 #undef __FUNCT__ 626 #define __FUNCT__ "TSGetRHSVec_Private" 627 static PetscErrorCode TSGetRHSVec_Private(TS ts,Vec *Frhs) 628 { 629 Vec F; 630 PetscErrorCode ierr; 631 632 PetscFunctionBegin; 633 *Frhs = NULL; 634 ierr = TSGetIFunction(ts,&F,NULL,NULL);CHKERRQ(ierr); 635 if (!ts->Frhs) { 636 ierr = VecDuplicate(F,&ts->Frhs);CHKERRQ(ierr); 637 } 638 *Frhs = ts->Frhs; 639 PetscFunctionReturn(0); 640 } 641 642 #undef __FUNCT__ 643 #define __FUNCT__ "TSGetRHSMats_Private" 644 static PetscErrorCode TSGetRHSMats_Private(TS ts,Mat *Arhs,Mat *Brhs) 645 { 646 Mat A,B; 647 PetscErrorCode ierr; 648 649 PetscFunctionBegin; 650 if (Arhs) *Arhs = NULL; 651 if (Brhs) *Brhs = NULL; 652 ierr = TSGetIJacobian(ts,&A,&B,NULL,NULL);CHKERRQ(ierr); 653 if (Arhs) { 654 if (!ts->Arhs) { 655 ierr = MatDuplicate(A,MAT_DO_NOT_COPY_VALUES,&ts->Arhs);CHKERRQ(ierr); 656 } 657 *Arhs = ts->Arhs; 658 } 659 if (Brhs) { 660 if (!ts->Brhs) { 661 if (A != B) { 662 ierr = MatDuplicate(B,MAT_DO_NOT_COPY_VALUES,&ts->Brhs);CHKERRQ(ierr); 663 } else { 664 ts->Brhs = ts->Arhs; 665 ierr = PetscObjectReference((PetscObject)ts->Arhs);CHKERRQ(ierr); 666 } 667 } 668 *Brhs = ts->Brhs; 669 } 670 PetscFunctionReturn(0); 671 } 672 673 #undef __FUNCT__ 674 #define __FUNCT__ "TSComputeIFunction" 675 /*@ 676 TSComputeIFunction - Evaluates the DAE residual written in implicit form F(t,U,Udot)=0 677 678 Collective on TS and Vec 679 680 Input Parameters: 681 + ts - the TS context 682 . t - current time 683 . U - state vector 684 . Udot - time derivative of state vector 685 - imex - flag indicates if the method is IMEX so that the RHSFunction should be kept separate 686 687 Output Parameter: 688 . Y - right hand side 689 690 Note: 691 Most users should not need to explicitly call this routine, as it 692 is used internally within the nonlinear solvers. 693 694 If the user did did not write their equations in implicit form, this 695 function recasts them in implicit form. 696 697 Level: developer 698 699 .keywords: TS, compute 700 701 .seealso: TSSetIFunction(), TSComputeRHSFunction() 702 @*/ 703 PetscErrorCode TSComputeIFunction(TS ts,PetscReal t,Vec U,Vec Udot,Vec Y,PetscBool imex) 704 { 705 PetscErrorCode ierr; 706 TSIFunction ifunction; 707 TSRHSFunction rhsfunction; 708 void *ctx; 709 DM dm; 710 711 PetscFunctionBegin; 712 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 713 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 714 PetscValidHeaderSpecific(Udot,VEC_CLASSID,4); 715 PetscValidHeaderSpecific(Y,VEC_CLASSID,5); 716 717 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 718 ierr = DMTSGetIFunction(dm,&ifunction,&ctx);CHKERRQ(ierr); 719 ierr = DMTSGetRHSFunction(dm,&rhsfunction,NULL);CHKERRQ(ierr); 720 721 if (!rhsfunction && !ifunction) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSFunction() and / or TSSetIFunction()"); 722 723 ierr = PetscLogEventBegin(TS_FunctionEval,ts,U,Udot,Y);CHKERRQ(ierr); 724 if (ifunction) { 725 PetscStackPush("TS user implicit function"); 726 ierr = (*ifunction)(ts,t,U,Udot,Y,ctx);CHKERRQ(ierr); 727 PetscStackPop; 728 } 729 if (imex) { 730 if (!ifunction) { 731 ierr = VecCopy(Udot,Y);CHKERRQ(ierr); 732 } 733 } else if (rhsfunction) { 734 if (ifunction) { 735 Vec Frhs; 736 ierr = TSGetRHSVec_Private(ts,&Frhs);CHKERRQ(ierr); 737 ierr = TSComputeRHSFunction(ts,t,U,Frhs);CHKERRQ(ierr); 738 ierr = VecAXPY(Y,-1,Frhs);CHKERRQ(ierr); 739 } else { 740 ierr = TSComputeRHSFunction(ts,t,U,Y);CHKERRQ(ierr); 741 ierr = VecAYPX(Y,-1,Udot);CHKERRQ(ierr); 742 } 743 } 744 ierr = PetscLogEventEnd(TS_FunctionEval,ts,U,Udot,Y);CHKERRQ(ierr); 745 PetscFunctionReturn(0); 746 } 747 748 #undef __FUNCT__ 749 #define __FUNCT__ "TSComputeIJacobian" 750 /*@ 751 TSComputeIJacobian - Evaluates the Jacobian of the DAE 752 753 Collective on TS and Vec 754 755 Input 756 Input Parameters: 757 + ts - the TS context 758 . t - current timestep 759 . U - state vector 760 . Udot - time derivative of state vector 761 . shift - shift to apply, see note below 762 - imex - flag indicates if the method is IMEX so that the RHSJacobian should be kept separate 763 764 Output Parameters: 765 + A - Jacobian matrix 766 . B - optional preconditioning matrix 767 - flag - flag indicating matrix structure 768 769 Notes: 770 If F(t,U,Udot)=0 is the DAE, the required Jacobian is 771 772 dF/dU + shift*dF/dUdot 773 774 Most users should not need to explicitly call this routine, as it 775 is used internally within the nonlinear solvers. 776 777 Level: developer 778 779 .keywords: TS, compute, Jacobian, matrix 780 781 .seealso: TSSetIJacobian() 782 @*/ 783 PetscErrorCode TSComputeIJacobian(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,PetscBool imex) 784 { 785 PetscErrorCode ierr; 786 TSIJacobian ijacobian; 787 TSRHSJacobian rhsjacobian; 788 DM dm; 789 void *ctx; 790 791 PetscFunctionBegin; 792 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 793 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 794 PetscValidHeaderSpecific(Udot,VEC_CLASSID,4); 795 PetscValidPointer(A,6); 796 PetscValidHeaderSpecific(A,MAT_CLASSID,6); 797 PetscValidPointer(B,7); 798 PetscValidHeaderSpecific(B,MAT_CLASSID,7); 799 800 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 801 ierr = DMTSGetIJacobian(dm,&ijacobian,&ctx);CHKERRQ(ierr); 802 ierr = DMTSGetRHSJacobian(dm,&rhsjacobian,NULL);CHKERRQ(ierr); 803 804 if (!rhsjacobian && !ijacobian) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"Must call TSSetRHSJacobian() and / or TSSetIJacobian()"); 805 806 ierr = PetscLogEventBegin(TS_JacobianEval,ts,U,A,B);CHKERRQ(ierr); 807 if (ijacobian) { 808 PetscStackPush("TS user implicit Jacobian"); 809 ierr = (*ijacobian)(ts,t,U,Udot,shift,A,B,ctx);CHKERRQ(ierr); 810 PetscStackPop; 811 /* make sure user returned a correct Jacobian and preconditioner */ 812 PetscValidHeaderSpecific(A,MAT_CLASSID,4); 813 PetscValidHeaderSpecific(B,MAT_CLASSID,5); 814 } 815 if (imex) { 816 if (!ijacobian) { /* system was written as Udot = G(t,U) */ 817 ierr = MatZeroEntries(A);CHKERRQ(ierr); 818 ierr = MatShift(A,shift);CHKERRQ(ierr); 819 if (A != B) { 820 ierr = MatZeroEntries(B);CHKERRQ(ierr); 821 ierr = MatShift(B,shift);CHKERRQ(ierr); 822 } 823 } 824 } else { 825 Mat Arhs = NULL,Brhs = NULL; 826 if (rhsjacobian) { 827 if (ijacobian) { 828 ierr = TSGetRHSMats_Private(ts,&Arhs,&Brhs);CHKERRQ(ierr); 829 } else { 830 ierr = TSGetIJacobian(ts,&Arhs,&Brhs,NULL,NULL);CHKERRQ(ierr); 831 } 832 ierr = TSComputeRHSJacobian(ts,t,U,Arhs,Brhs);CHKERRQ(ierr); 833 } 834 if (Arhs == A) { /* No IJacobian, so we only have the RHS matrix */ 835 ts->rhsjacobian.scale = -1; 836 ts->rhsjacobian.shift = shift; 837 ierr = MatScale(A,-1);CHKERRQ(ierr); 838 ierr = MatShift(A,shift);CHKERRQ(ierr); 839 if (A != B) { 840 ierr = MatScale(B,-1);CHKERRQ(ierr); 841 ierr = MatShift(B,shift);CHKERRQ(ierr); 842 } 843 } else if (Arhs) { /* Both IJacobian and RHSJacobian */ 844 MatStructure axpy = DIFFERENT_NONZERO_PATTERN; 845 if (!ijacobian) { /* No IJacobian provided, but we have a separate RHS matrix */ 846 ierr = MatZeroEntries(A);CHKERRQ(ierr); 847 ierr = MatShift(A,shift);CHKERRQ(ierr); 848 if (A != B) { 849 ierr = MatZeroEntries(B);CHKERRQ(ierr); 850 ierr = MatShift(B,shift);CHKERRQ(ierr); 851 } 852 } 853 ierr = MatAXPY(A,-1,Arhs,axpy);CHKERRQ(ierr); 854 if (A != B) { 855 ierr = MatAXPY(B,-1,Brhs,axpy);CHKERRQ(ierr); 856 } 857 } 858 } 859 ierr = PetscLogEventEnd(TS_JacobianEval,ts,U,A,B);CHKERRQ(ierr); 860 PetscFunctionReturn(0); 861 } 862 863 #undef __FUNCT__ 864 #define __FUNCT__ "TSSetRHSFunction" 865 /*@C 866 TSSetRHSFunction - Sets the routine for evaluating the function, 867 where U_t = G(t,u). 868 869 Logically Collective on TS 870 871 Input Parameters: 872 + ts - the TS context obtained from TSCreate() 873 . r - vector to put the computed right hand side (or NULL to have it created) 874 . f - routine for evaluating the right-hand-side function 875 - ctx - [optional] user-defined context for private data for the 876 function evaluation routine (may be NULL) 877 878 Calling sequence of func: 879 $ func (TS ts,PetscReal t,Vec u,Vec F,void *ctx); 880 881 + t - current timestep 882 . u - input vector 883 . F - function vector 884 - ctx - [optional] user-defined function context 885 886 Level: beginner 887 888 Notes: You must call this function or TSSetIFunction() to define your ODE. You cannot use this function when solving a DAE. 889 890 .keywords: TS, timestep, set, right-hand-side, function 891 892 .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSSetIFunction() 893 @*/ 894 PetscErrorCode TSSetRHSFunction(TS ts,Vec r,PetscErrorCode (*f)(TS,PetscReal,Vec,Vec,void*),void *ctx) 895 { 896 PetscErrorCode ierr; 897 SNES snes; 898 Vec ralloc = NULL; 899 DM dm; 900 901 PetscFunctionBegin; 902 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 903 if (r) PetscValidHeaderSpecific(r,VEC_CLASSID,2); 904 905 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 906 ierr = DMTSSetRHSFunction(dm,f,ctx);CHKERRQ(ierr); 907 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 908 if (!r && !ts->dm && ts->vec_sol) { 909 ierr = VecDuplicate(ts->vec_sol,&ralloc);CHKERRQ(ierr); 910 r = ralloc; 911 } 912 ierr = SNESSetFunction(snes,r,SNESTSFormFunction,ts);CHKERRQ(ierr); 913 ierr = VecDestroy(&ralloc);CHKERRQ(ierr); 914 PetscFunctionReturn(0); 915 } 916 917 #undef __FUNCT__ 918 #define __FUNCT__ "TSSetSolutionFunction" 919 /*@C 920 TSSetSolutionFunction - Provide a function that computes the solution of the ODE or DAE 921 922 Logically Collective on TS 923 924 Input Parameters: 925 + ts - the TS context obtained from TSCreate() 926 . f - routine for evaluating the solution 927 - ctx - [optional] user-defined context for private data for the 928 function evaluation routine (may be NULL) 929 930 Calling sequence of func: 931 $ func (TS ts,PetscReal t,Vec u,void *ctx); 932 933 + t - current timestep 934 . u - output vector 935 - ctx - [optional] user-defined function context 936 937 Notes: 938 This routine is used for testing accuracy of time integration schemes when you already know the solution. 939 If analytic solutions are not known for your system, consider using the Method of Manufactured Solutions to 940 create closed-form solutions with non-physical forcing terms. 941 942 For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history. 943 944 Level: beginner 945 946 .keywords: TS, timestep, set, right-hand-side, function 947 948 .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSComputeSolutionFunction(), TSSetForcingFunction() 949 @*/ 950 PetscErrorCode TSSetSolutionFunction(TS ts,PetscErrorCode (*f)(TS,PetscReal,Vec,void*),void *ctx) 951 { 952 PetscErrorCode ierr; 953 DM dm; 954 955 PetscFunctionBegin; 956 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 957 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 958 ierr = DMTSSetSolutionFunction(dm,f,ctx);CHKERRQ(ierr); 959 PetscFunctionReturn(0); 960 } 961 962 #undef __FUNCT__ 963 #define __FUNCT__ "TSSetForcingFunction" 964 /*@C 965 TSSetForcingFunction - Provide a function that computes a forcing term for a ODE or PDE 966 967 Logically Collective on TS 968 969 Input Parameters: 970 + ts - the TS context obtained from TSCreate() 971 . f - routine for evaluating the forcing function 972 - ctx - [optional] user-defined context for private data for the 973 function evaluation routine (may be NULL) 974 975 Calling sequence of func: 976 $ func (TS ts,PetscReal t,Vec u,void *ctx); 977 978 + t - current timestep 979 . u - output vector 980 - ctx - [optional] user-defined function context 981 982 Notes: 983 This routine is useful for testing accuracy of time integration schemes when using the Method of Manufactured Solutions to 984 create closed-form solutions with a non-physical forcing term. 985 986 For low-dimensional problems solved in serial, such as small discrete systems, TSMonitorLGError() can be used to monitor the error history. 987 988 Level: beginner 989 990 .keywords: TS, timestep, set, right-hand-side, function 991 992 .seealso: TSSetRHSJacobian(), TSSetIJacobian(), TSComputeSolutionFunction(), TSSetSolutionFunction() 993 @*/ 994 PetscErrorCode TSSetForcingFunction(TS ts,PetscErrorCode (*f)(TS,PetscReal,Vec,void*),void *ctx) 995 { 996 PetscErrorCode ierr; 997 DM dm; 998 999 PetscFunctionBegin; 1000 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1001 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1002 ierr = DMTSSetForcingFunction(dm,f,ctx);CHKERRQ(ierr); 1003 PetscFunctionReturn(0); 1004 } 1005 1006 #undef __FUNCT__ 1007 #define __FUNCT__ "TSSetRHSJacobian" 1008 /*@C 1009 TSSetRHSJacobian - Sets the function to compute the Jacobian of G, 1010 where U_t = G(U,t), as well as the location to store the matrix. 1011 1012 Logically Collective on TS 1013 1014 Input Parameters: 1015 + ts - the TS context obtained from TSCreate() 1016 . Amat - (approximate) Jacobian matrix 1017 . Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat) 1018 . f - the Jacobian evaluation routine 1019 - ctx - [optional] user-defined context for private data for the 1020 Jacobian evaluation routine (may be NULL) 1021 1022 Calling sequence of f: 1023 $ func (TS ts,PetscReal t,Vec u,Mat A,Mat B,void *ctx); 1024 1025 + t - current timestep 1026 . u - input vector 1027 . Amat - (approximate) Jacobian matrix 1028 . Pmat - matrix from which preconditioner is to be constructed (usually the same as Amat) 1029 - ctx - [optional] user-defined context for matrix evaluation routine 1030 1031 1032 Level: beginner 1033 1034 .keywords: TS, timestep, set, right-hand-side, Jacobian 1035 1036 .seealso: SNESComputeJacobianDefaultColor(), TSSetRHSFunction(), TSRHSJacobianSetReuse(), TSSetIJacobian() 1037 1038 @*/ 1039 PetscErrorCode TSSetRHSJacobian(TS ts,Mat Amat,Mat Pmat,TSRHSJacobian f,void *ctx) 1040 { 1041 PetscErrorCode ierr; 1042 SNES snes; 1043 DM dm; 1044 TSIJacobian ijacobian; 1045 1046 PetscFunctionBegin; 1047 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1048 if (Amat) PetscValidHeaderSpecific(Amat,MAT_CLASSID,2); 1049 if (Pmat) PetscValidHeaderSpecific(Pmat,MAT_CLASSID,3); 1050 if (Amat) PetscCheckSameComm(ts,1,Amat,2); 1051 if (Pmat) PetscCheckSameComm(ts,1,Pmat,3); 1052 1053 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1054 ierr = DMTSSetRHSJacobian(dm,f,ctx);CHKERRQ(ierr); 1055 if (f == TSComputeRHSJacobianConstant) { 1056 /* Handle this case automatically for the user; otherwise user should call themselves. */ 1057 ierr = TSRHSJacobianSetReuse(ts,PETSC_TRUE);CHKERRQ(ierr); 1058 } 1059 ierr = DMTSGetIJacobian(dm,&ijacobian,NULL);CHKERRQ(ierr); 1060 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1061 if (!ijacobian) { 1062 ierr = SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts);CHKERRQ(ierr); 1063 } 1064 if (Amat) { 1065 ierr = PetscObjectReference((PetscObject)Amat);CHKERRQ(ierr); 1066 ierr = MatDestroy(&ts->Arhs);CHKERRQ(ierr); 1067 1068 ts->Arhs = Amat; 1069 } 1070 if (Pmat) { 1071 ierr = PetscObjectReference((PetscObject)Pmat);CHKERRQ(ierr); 1072 ierr = MatDestroy(&ts->Brhs);CHKERRQ(ierr); 1073 1074 ts->Brhs = Pmat; 1075 } 1076 PetscFunctionReturn(0); 1077 } 1078 1079 1080 #undef __FUNCT__ 1081 #define __FUNCT__ "TSSetIFunction" 1082 /*@C 1083 TSSetIFunction - Set the function to compute F(t,U,U_t) where F() = 0 is the DAE to be solved. 1084 1085 Logically Collective on TS 1086 1087 Input Parameters: 1088 + ts - the TS context obtained from TSCreate() 1089 . r - vector to hold the residual (or NULL to have it created internally) 1090 . f - the function evaluation routine 1091 - ctx - user-defined context for private data for the function evaluation routine (may be NULL) 1092 1093 Calling sequence of f: 1094 $ f(TS ts,PetscReal t,Vec u,Vec u_t,Vec F,ctx); 1095 1096 + t - time at step/stage being solved 1097 . u - state vector 1098 . u_t - time derivative of state vector 1099 . F - function vector 1100 - ctx - [optional] user-defined context for matrix evaluation routine 1101 1102 Important: 1103 The user MUST call either this routine or TSSetRHSFunction() to define the ODE. When solving DAEs you must use this function. 1104 1105 Level: beginner 1106 1107 .keywords: TS, timestep, set, DAE, Jacobian 1108 1109 .seealso: TSSetRHSJacobian(), TSSetRHSFunction(), TSSetIJacobian() 1110 @*/ 1111 PetscErrorCode TSSetIFunction(TS ts,Vec res,TSIFunction f,void *ctx) 1112 { 1113 PetscErrorCode ierr; 1114 SNES snes; 1115 Vec resalloc = NULL; 1116 DM dm; 1117 1118 PetscFunctionBegin; 1119 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1120 if (res) PetscValidHeaderSpecific(res,VEC_CLASSID,2); 1121 1122 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1123 ierr = DMTSSetIFunction(dm,f,ctx);CHKERRQ(ierr); 1124 1125 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1126 if (!res && !ts->dm && ts->vec_sol) { 1127 ierr = VecDuplicate(ts->vec_sol,&resalloc);CHKERRQ(ierr); 1128 res = resalloc; 1129 } 1130 ierr = SNESSetFunction(snes,res,SNESTSFormFunction,ts);CHKERRQ(ierr); 1131 ierr = VecDestroy(&resalloc);CHKERRQ(ierr); 1132 PetscFunctionReturn(0); 1133 } 1134 1135 #undef __FUNCT__ 1136 #define __FUNCT__ "TSGetIFunction" 1137 /*@C 1138 TSGetIFunction - Returns the vector where the implicit residual is stored and the function/contex to compute it. 1139 1140 Not Collective 1141 1142 Input Parameter: 1143 . ts - the TS context 1144 1145 Output Parameter: 1146 + r - vector to hold residual (or NULL) 1147 . func - the function to compute residual (or NULL) 1148 - ctx - the function context (or NULL) 1149 1150 Level: advanced 1151 1152 .keywords: TS, nonlinear, get, function 1153 1154 .seealso: TSSetIFunction(), SNESGetFunction() 1155 @*/ 1156 PetscErrorCode TSGetIFunction(TS ts,Vec *r,TSIFunction *func,void **ctx) 1157 { 1158 PetscErrorCode ierr; 1159 SNES snes; 1160 DM dm; 1161 1162 PetscFunctionBegin; 1163 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1164 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1165 ierr = SNESGetFunction(snes,r,NULL,NULL);CHKERRQ(ierr); 1166 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1167 ierr = DMTSGetIFunction(dm,func,ctx);CHKERRQ(ierr); 1168 PetscFunctionReturn(0); 1169 } 1170 1171 #undef __FUNCT__ 1172 #define __FUNCT__ "TSGetRHSFunction" 1173 /*@C 1174 TSGetRHSFunction - Returns the vector where the right hand side is stored and the function/context to compute it. 1175 1176 Not Collective 1177 1178 Input Parameter: 1179 . ts - the TS context 1180 1181 Output Parameter: 1182 + r - vector to hold computed right hand side (or NULL) 1183 . func - the function to compute right hand side (or NULL) 1184 - ctx - the function context (or NULL) 1185 1186 Level: advanced 1187 1188 .keywords: TS, nonlinear, get, function 1189 1190 .seealso: TSSetRHSFunction(), SNESGetFunction() 1191 @*/ 1192 PetscErrorCode TSGetRHSFunction(TS ts,Vec *r,TSRHSFunction *func,void **ctx) 1193 { 1194 PetscErrorCode ierr; 1195 SNES snes; 1196 DM dm; 1197 1198 PetscFunctionBegin; 1199 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1200 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1201 ierr = SNESGetFunction(snes,r,NULL,NULL);CHKERRQ(ierr); 1202 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1203 ierr = DMTSGetRHSFunction(dm,func,ctx);CHKERRQ(ierr); 1204 PetscFunctionReturn(0); 1205 } 1206 1207 #undef __FUNCT__ 1208 #define __FUNCT__ "TSSetIJacobian" 1209 /*@C 1210 TSSetIJacobian - Set the function to compute the matrix dF/dU + a*dF/dU_t where F(t,U,U_t) is the function 1211 provided with TSSetIFunction(). 1212 1213 Logically Collective on TS 1214 1215 Input Parameters: 1216 + ts - the TS context obtained from TSCreate() 1217 . Amat - (approximate) Jacobian matrix 1218 . Pmat - matrix used to compute preconditioner (usually the same as Amat) 1219 . f - the Jacobian evaluation routine 1220 - ctx - user-defined context for private data for the Jacobian evaluation routine (may be NULL) 1221 1222 Calling sequence of f: 1223 $ f(TS ts,PetscReal t,Vec U,Vec U_t,PetscReal a,Mat Amat,Mat Pmat,void *ctx); 1224 1225 + t - time at step/stage being solved 1226 . U - state vector 1227 . U_t - time derivative of state vector 1228 . a - shift 1229 . Amat - (approximate) Jacobian of F(t,U,W+a*U), equivalent to dF/dU + a*dF/dU_t 1230 . Pmat - matrix used for constructing preconditioner, usually the same as Amat 1231 - ctx - [optional] user-defined context for matrix evaluation routine 1232 1233 Notes: 1234 The matrices Amat and Pmat are exactly the matrices that are used by SNES for the nonlinear solve. 1235 1236 The matrix dF/dU + a*dF/dU_t you provide turns out to be 1237 the Jacobian of F(t,U,W+a*U) where F(t,U,U_t) = 0 is the DAE to be solved. 1238 The time integrator internally approximates U_t by W+a*U where the positive "shift" 1239 a and vector W depend on the integration method, step size, and past states. For example with 1240 the backward Euler method a = 1/dt and W = -a*U(previous timestep) so 1241 W + a*U = a*(U - U(previous timestep)) = (U - U(previous timestep))/dt 1242 1243 Level: beginner 1244 1245 .keywords: TS, timestep, DAE, Jacobian 1246 1247 .seealso: TSSetIFunction(), TSSetRHSJacobian(), SNESComputeJacobianDefaultColor(), SNESComputeJacobianDefault(), TSSetRHSFunction() 1248 1249 @*/ 1250 PetscErrorCode TSSetIJacobian(TS ts,Mat Amat,Mat Pmat,TSIJacobian f,void *ctx) 1251 { 1252 PetscErrorCode ierr; 1253 SNES snes; 1254 DM dm; 1255 1256 PetscFunctionBegin; 1257 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1258 if (Amat) PetscValidHeaderSpecific(Amat,MAT_CLASSID,2); 1259 if (Pmat) PetscValidHeaderSpecific(Pmat,MAT_CLASSID,3); 1260 if (Amat) PetscCheckSameComm(ts,1,Amat,2); 1261 if (Pmat) PetscCheckSameComm(ts,1,Pmat,3); 1262 1263 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1264 ierr = DMTSSetIJacobian(dm,f,ctx);CHKERRQ(ierr); 1265 1266 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1267 ierr = SNESSetJacobian(snes,Amat,Pmat,SNESTSFormJacobian,ts);CHKERRQ(ierr); 1268 PetscFunctionReturn(0); 1269 } 1270 1271 #undef __FUNCT__ 1272 #define __FUNCT__ "TSRHSJacobianSetReuse" 1273 /*@ 1274 TSRHSJacobianSetReuse - restore RHS Jacobian before re-evaluating. Without this flag, TS will change the sign and 1275 shift the RHS Jacobian for a finite-time-step implicit solve, in which case the user function will need to recompute 1276 the entire Jacobian. The reuse flag must be set if the evaluation function will assume that the matrix entries have 1277 not been changed by the TS. 1278 1279 Logically Collective 1280 1281 Input Arguments: 1282 + ts - TS context obtained from TSCreate() 1283 - reuse - PETSC_TRUE if the RHS Jacobian 1284 1285 Level: intermediate 1286 1287 .seealso: TSSetRHSJacobian(), TSComputeRHSJacobianConstant() 1288 @*/ 1289 PetscErrorCode TSRHSJacobianSetReuse(TS ts,PetscBool reuse) 1290 { 1291 PetscFunctionBegin; 1292 ts->rhsjacobian.reuse = reuse; 1293 PetscFunctionReturn(0); 1294 } 1295 1296 #undef __FUNCT__ 1297 #define __FUNCT__ "TSLoad" 1298 /*@C 1299 TSLoad - Loads a KSP that has been stored in binary with KSPView(). 1300 1301 Collective on PetscViewer 1302 1303 Input Parameters: 1304 + newdm - the newly loaded TS, this needs to have been created with TSCreate() or 1305 some related function before a call to TSLoad(). 1306 - viewer - binary file viewer, obtained from PetscViewerBinaryOpen() 1307 1308 Level: intermediate 1309 1310 Notes: 1311 The type is determined by the data in the file, any type set into the TS before this call is ignored. 1312 1313 Notes for advanced users: 1314 Most users should not need to know the details of the binary storage 1315 format, since TSLoad() and TSView() completely hide these details. 1316 But for anyone who's interested, the standard binary matrix storage 1317 format is 1318 .vb 1319 has not yet been determined 1320 .ve 1321 1322 .seealso: PetscViewerBinaryOpen(), TSView(), MatLoad(), VecLoad() 1323 @*/ 1324 PetscErrorCode TSLoad(TS ts, PetscViewer viewer) 1325 { 1326 PetscErrorCode ierr; 1327 PetscBool isbinary; 1328 PetscInt classid; 1329 char type[256]; 1330 DMTS sdm; 1331 DM dm; 1332 1333 PetscFunctionBegin; 1334 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1335 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2); 1336 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 1337 if (!isbinary) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Invalid viewer; open viewer with PetscViewerBinaryOpen()"); 1338 1339 ierr = PetscViewerBinaryRead(viewer,&classid,1,NULL,PETSC_INT);CHKERRQ(ierr); 1340 if (classid != TS_FILE_CLASSID) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_WRONG,"Not TS next in file"); 1341 ierr = PetscViewerBinaryRead(viewer,type,256,NULL,PETSC_CHAR);CHKERRQ(ierr); 1342 ierr = TSSetType(ts, type);CHKERRQ(ierr); 1343 if (ts->ops->load) { 1344 ierr = (*ts->ops->load)(ts,viewer);CHKERRQ(ierr); 1345 } 1346 ierr = DMCreate(PetscObjectComm((PetscObject)ts),&dm);CHKERRQ(ierr); 1347 ierr = DMLoad(dm,viewer);CHKERRQ(ierr); 1348 ierr = TSSetDM(ts,dm);CHKERRQ(ierr); 1349 ierr = DMCreateGlobalVector(ts->dm,&ts->vec_sol);CHKERRQ(ierr); 1350 ierr = VecLoad(ts->vec_sol,viewer);CHKERRQ(ierr); 1351 ierr = DMGetDMTS(ts->dm,&sdm);CHKERRQ(ierr); 1352 ierr = DMTSLoad(sdm,viewer);CHKERRQ(ierr); 1353 PetscFunctionReturn(0); 1354 } 1355 1356 #include <petscdraw.h> 1357 #if defined(PETSC_HAVE_SAWS) 1358 #include <petscviewersaws.h> 1359 #endif 1360 #undef __FUNCT__ 1361 #define __FUNCT__ "TSView" 1362 /*@C 1363 TSView - Prints the TS data structure. 1364 1365 Collective on TS 1366 1367 Input Parameters: 1368 + ts - the TS context obtained from TSCreate() 1369 - viewer - visualization context 1370 1371 Options Database Key: 1372 . -ts_view - calls TSView() at end of TSStep() 1373 1374 Notes: 1375 The available visualization contexts include 1376 + PETSC_VIEWER_STDOUT_SELF - standard output (default) 1377 - PETSC_VIEWER_STDOUT_WORLD - synchronized standard 1378 output where only the first processor opens 1379 the file. All other processors send their 1380 data to the first processor to print. 1381 1382 The user can open an alternative visualization context with 1383 PetscViewerASCIIOpen() - output to a specified file. 1384 1385 Level: beginner 1386 1387 .keywords: TS, timestep, view 1388 1389 .seealso: PetscViewerASCIIOpen() 1390 @*/ 1391 PetscErrorCode TSView(TS ts,PetscViewer viewer) 1392 { 1393 PetscErrorCode ierr; 1394 TSType type; 1395 PetscBool iascii,isstring,isundials,isbinary,isdraw; 1396 DMTS sdm; 1397 #if defined(PETSC_HAVE_SAWS) 1398 PetscBool issaws; 1399 #endif 1400 1401 PetscFunctionBegin; 1402 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1403 if (!viewer) { 1404 ierr = PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)ts),&viewer);CHKERRQ(ierr); 1405 } 1406 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2); 1407 PetscCheckSameComm(ts,1,viewer,2); 1408 1409 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);CHKERRQ(ierr); 1410 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSTRING,&isstring);CHKERRQ(ierr); 1411 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 1412 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);CHKERRQ(ierr); 1413 #if defined(PETSC_HAVE_SAWS) 1414 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERSAWS,&issaws);CHKERRQ(ierr); 1415 #endif 1416 if (iascii) { 1417 ierr = PetscObjectPrintClassNamePrefixType((PetscObject)ts,viewer);CHKERRQ(ierr); 1418 ierr = PetscViewerASCIIPrintf(viewer," maximum steps=%D\n",ts->max_steps);CHKERRQ(ierr); 1419 ierr = PetscViewerASCIIPrintf(viewer," maximum time=%g\n",(double)ts->max_time);CHKERRQ(ierr); 1420 if (ts->problem_type == TS_NONLINEAR) { 1421 ierr = PetscViewerASCIIPrintf(viewer," total number of nonlinear solver iterations=%D\n",ts->snes_its);CHKERRQ(ierr); 1422 ierr = PetscViewerASCIIPrintf(viewer," total number of nonlinear solve failures=%D\n",ts->num_snes_failures);CHKERRQ(ierr); 1423 } 1424 ierr = PetscViewerASCIIPrintf(viewer," total number of linear solver iterations=%D\n",ts->ksp_its);CHKERRQ(ierr); 1425 ierr = PetscViewerASCIIPrintf(viewer," total number of rejected steps=%D\n",ts->reject);CHKERRQ(ierr); 1426 ierr = DMGetDMTS(ts->dm,&sdm);CHKERRQ(ierr); 1427 ierr = DMTSView(sdm,viewer);CHKERRQ(ierr); 1428 if (ts->ops->view) { 1429 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 1430 ierr = (*ts->ops->view)(ts,viewer);CHKERRQ(ierr); 1431 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 1432 } 1433 } else if (isstring) { 1434 ierr = TSGetType(ts,&type);CHKERRQ(ierr); 1435 ierr = PetscViewerStringSPrintf(viewer," %-7.7s",type);CHKERRQ(ierr); 1436 } else if (isbinary) { 1437 PetscInt classid = TS_FILE_CLASSID; 1438 MPI_Comm comm; 1439 PetscMPIInt rank; 1440 char type[256]; 1441 1442 ierr = PetscObjectGetComm((PetscObject)ts,&comm);CHKERRQ(ierr); 1443 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 1444 if (!rank) { 1445 ierr = PetscViewerBinaryWrite(viewer,&classid,1,PETSC_INT,PETSC_FALSE);CHKERRQ(ierr); 1446 ierr = PetscStrncpy(type,((PetscObject)ts)->type_name,256);CHKERRQ(ierr); 1447 ierr = PetscViewerBinaryWrite(viewer,type,256,PETSC_CHAR,PETSC_FALSE);CHKERRQ(ierr); 1448 } 1449 if (ts->ops->view) { 1450 ierr = (*ts->ops->view)(ts,viewer);CHKERRQ(ierr); 1451 } 1452 ierr = DMView(ts->dm,viewer);CHKERRQ(ierr); 1453 ierr = VecView(ts->vec_sol,viewer);CHKERRQ(ierr); 1454 ierr = DMGetDMTS(ts->dm,&sdm);CHKERRQ(ierr); 1455 ierr = DMTSView(sdm,viewer);CHKERRQ(ierr); 1456 } else if (isdraw) { 1457 PetscDraw draw; 1458 char str[36]; 1459 PetscReal x,y,bottom,h; 1460 1461 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 1462 ierr = PetscDrawGetCurrentPoint(draw,&x,&y);CHKERRQ(ierr); 1463 ierr = PetscStrcpy(str,"TS: ");CHKERRQ(ierr); 1464 ierr = PetscStrcat(str,((PetscObject)ts)->type_name);CHKERRQ(ierr); 1465 ierr = PetscDrawStringBoxed(draw,x,y,PETSC_DRAW_BLACK,PETSC_DRAW_BLACK,str,NULL,&h);CHKERRQ(ierr); 1466 bottom = y - h; 1467 ierr = PetscDrawPushCurrentPoint(draw,x,bottom);CHKERRQ(ierr); 1468 if (ts->ops->view) { 1469 ierr = (*ts->ops->view)(ts,viewer);CHKERRQ(ierr); 1470 } 1471 ierr = PetscDrawPopCurrentPoint(draw);CHKERRQ(ierr); 1472 #if defined(PETSC_HAVE_SAWS) 1473 } else if (issaws) { 1474 PetscMPIInt rank; 1475 const char *name; 1476 1477 ierr = PetscObjectGetName((PetscObject)ts,&name);CHKERRQ(ierr); 1478 ierr = MPI_Comm_rank(PETSC_COMM_WORLD,&rank);CHKERRQ(ierr); 1479 if (!((PetscObject)ts)->amsmem && !rank) { 1480 char dir[1024]; 1481 1482 ierr = PetscObjectViewSAWs((PetscObject)ts,viewer);CHKERRQ(ierr); 1483 ierr = PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time_step",name);CHKERRQ(ierr); 1484 PetscStackCallSAWs(SAWs_Register,(dir,&ts->steps,1,SAWs_READ,SAWs_INT)); 1485 ierr = PetscSNPrintf(dir,1024,"/PETSc/Objects/%s/time",name);CHKERRQ(ierr); 1486 PetscStackCallSAWs(SAWs_Register,(dir,&ts->ptime,1,SAWs_READ,SAWs_DOUBLE)); 1487 } 1488 if (ts->ops->view) { 1489 ierr = (*ts->ops->view)(ts,viewer);CHKERRQ(ierr); 1490 } 1491 #endif 1492 } 1493 1494 ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr); 1495 ierr = PetscObjectTypeCompare((PetscObject)ts,TSSUNDIALS,&isundials);CHKERRQ(ierr); 1496 ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr); 1497 PetscFunctionReturn(0); 1498 } 1499 1500 1501 #undef __FUNCT__ 1502 #define __FUNCT__ "TSSetApplicationContext" 1503 /*@ 1504 TSSetApplicationContext - Sets an optional user-defined context for 1505 the timesteppers. 1506 1507 Logically Collective on TS 1508 1509 Input Parameters: 1510 + ts - the TS context obtained from TSCreate() 1511 - usrP - optional user context 1512 1513 Level: intermediate 1514 1515 .keywords: TS, timestep, set, application, context 1516 1517 .seealso: TSGetApplicationContext() 1518 @*/ 1519 PetscErrorCode TSSetApplicationContext(TS ts,void *usrP) 1520 { 1521 PetscFunctionBegin; 1522 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1523 ts->user = usrP; 1524 PetscFunctionReturn(0); 1525 } 1526 1527 #undef __FUNCT__ 1528 #define __FUNCT__ "TSGetApplicationContext" 1529 /*@ 1530 TSGetApplicationContext - Gets the user-defined context for the 1531 timestepper. 1532 1533 Not Collective 1534 1535 Input Parameter: 1536 . ts - the TS context obtained from TSCreate() 1537 1538 Output Parameter: 1539 . usrP - user context 1540 1541 Level: intermediate 1542 1543 .keywords: TS, timestep, get, application, context 1544 1545 .seealso: TSSetApplicationContext() 1546 @*/ 1547 PetscErrorCode TSGetApplicationContext(TS ts,void *usrP) 1548 { 1549 PetscFunctionBegin; 1550 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1551 *(void**)usrP = ts->user; 1552 PetscFunctionReturn(0); 1553 } 1554 1555 #undef __FUNCT__ 1556 #define __FUNCT__ "TSGetTimeStepNumber" 1557 /*@ 1558 TSGetTimeStepNumber - Gets the number of time steps completed. 1559 1560 Not Collective 1561 1562 Input Parameter: 1563 . ts - the TS context obtained from TSCreate() 1564 1565 Output Parameter: 1566 . iter - number of steps completed so far 1567 1568 Level: intermediate 1569 1570 .keywords: TS, timestep, get, iteration, number 1571 .seealso: TSGetTime(), TSGetTimeStep(), TSSetPreStep(), TSSetPreStage(), TSSetPostStage(), TSSetPostStep() 1572 @*/ 1573 PetscErrorCode TSGetTimeStepNumber(TS ts,PetscInt *iter) 1574 { 1575 PetscFunctionBegin; 1576 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1577 PetscValidIntPointer(iter,2); 1578 *iter = ts->steps; 1579 PetscFunctionReturn(0); 1580 } 1581 1582 #undef __FUNCT__ 1583 #define __FUNCT__ "TSSetInitialTimeStep" 1584 /*@ 1585 TSSetInitialTimeStep - Sets the initial timestep to be used, 1586 as well as the initial time. 1587 1588 Logically Collective on TS 1589 1590 Input Parameters: 1591 + ts - the TS context obtained from TSCreate() 1592 . initial_time - the initial time 1593 - time_step - the size of the timestep 1594 1595 Level: intermediate 1596 1597 .seealso: TSSetTimeStep(), TSGetTimeStep() 1598 1599 .keywords: TS, set, initial, timestep 1600 @*/ 1601 PetscErrorCode TSSetInitialTimeStep(TS ts,PetscReal initial_time,PetscReal time_step) 1602 { 1603 PetscErrorCode ierr; 1604 1605 PetscFunctionBegin; 1606 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1607 ierr = TSSetTimeStep(ts,time_step);CHKERRQ(ierr); 1608 ierr = TSSetTime(ts,initial_time);CHKERRQ(ierr); 1609 PetscFunctionReturn(0); 1610 } 1611 1612 #undef __FUNCT__ 1613 #define __FUNCT__ "TSSetTimeStep" 1614 /*@ 1615 TSSetTimeStep - Allows one to reset the timestep at any time, 1616 useful for simple pseudo-timestepping codes. 1617 1618 Logically Collective on TS 1619 1620 Input Parameters: 1621 + ts - the TS context obtained from TSCreate() 1622 - time_step - the size of the timestep 1623 1624 Level: intermediate 1625 1626 .seealso: TSSetInitialTimeStep(), TSGetTimeStep() 1627 1628 .keywords: TS, set, timestep 1629 @*/ 1630 PetscErrorCode TSSetTimeStep(TS ts,PetscReal time_step) 1631 { 1632 PetscFunctionBegin; 1633 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1634 PetscValidLogicalCollectiveReal(ts,time_step,2); 1635 ts->time_step = time_step; 1636 ts->time_step_orig = time_step; 1637 PetscFunctionReturn(0); 1638 } 1639 1640 #undef __FUNCT__ 1641 #define __FUNCT__ "TSSetExactFinalTime" 1642 /*@ 1643 TSSetExactFinalTime - Determines whether to adapt the final time step to 1644 match the exact final time, interpolate solution to the exact final time, 1645 or just return at the final time TS computed. 1646 1647 Logically Collective on TS 1648 1649 Input Parameter: 1650 + ts - the time-step context 1651 - eftopt - exact final time option 1652 1653 Level: beginner 1654 1655 .seealso: TSExactFinalTimeOption 1656 @*/ 1657 PetscErrorCode TSSetExactFinalTime(TS ts,TSExactFinalTimeOption eftopt) 1658 { 1659 PetscFunctionBegin; 1660 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1661 PetscValidLogicalCollectiveEnum(ts,eftopt,2); 1662 ts->exact_final_time = eftopt; 1663 PetscFunctionReturn(0); 1664 } 1665 1666 #undef __FUNCT__ 1667 #define __FUNCT__ "TSGetTimeStep" 1668 /*@ 1669 TSGetTimeStep - Gets the current timestep size. 1670 1671 Not Collective 1672 1673 Input Parameter: 1674 . ts - the TS context obtained from TSCreate() 1675 1676 Output Parameter: 1677 . dt - the current timestep size 1678 1679 Level: intermediate 1680 1681 .seealso: TSSetInitialTimeStep(), TSGetTimeStep() 1682 1683 .keywords: TS, get, timestep 1684 @*/ 1685 PetscErrorCode TSGetTimeStep(TS ts,PetscReal *dt) 1686 { 1687 PetscFunctionBegin; 1688 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1689 PetscValidRealPointer(dt,2); 1690 *dt = ts->time_step; 1691 PetscFunctionReturn(0); 1692 } 1693 1694 #undef __FUNCT__ 1695 #define __FUNCT__ "TSGetSolution" 1696 /*@ 1697 TSGetSolution - Returns the solution at the present timestep. It 1698 is valid to call this routine inside the function that you are evaluating 1699 in order to move to the new timestep. This vector not changed until 1700 the solution at the next timestep has been calculated. 1701 1702 Not Collective, but Vec returned is parallel if TS is parallel 1703 1704 Input Parameter: 1705 . ts - the TS context obtained from TSCreate() 1706 1707 Output Parameter: 1708 . v - the vector containing the solution 1709 1710 Level: intermediate 1711 1712 .seealso: TSGetTimeStep() 1713 1714 .keywords: TS, timestep, get, solution 1715 @*/ 1716 PetscErrorCode TSGetSolution(TS ts,Vec *v) 1717 { 1718 PetscFunctionBegin; 1719 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1720 PetscValidPointer(v,2); 1721 *v = ts->vec_sol; 1722 PetscFunctionReturn(0); 1723 } 1724 1725 #undef __FUNCT__ 1726 #define __FUNCT__ "TSGetCostGradients" 1727 /*@ 1728 TSGetCostGradients - Returns the gradients from the TSAdjointSolve() 1729 1730 Not Collective, but Vec returned is parallel if TS is parallel 1731 1732 Input Parameter: 1733 . ts - the TS context obtained from TSCreate() 1734 1735 Output Parameter: 1736 + lambda - vectors containing the gradients of the cost functions with respect to the ODE/DAE solution variables 1737 - mu - vectors containing the gradients of the cost functions with respect to the problem parameters 1738 1739 Level: intermediate 1740 1741 .seealso: TSGetTimeStep() 1742 1743 .keywords: TS, timestep, get, sensitivity 1744 @*/ 1745 PetscErrorCode TSGetCostGradients(TS ts,PetscInt *numcost,Vec **lambda,Vec **mu) 1746 { 1747 PetscFunctionBegin; 1748 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1749 if (numcost) *numcost = ts->numcost; 1750 if (lambda) *lambda = ts->vecs_sensi; 1751 if (mu) *mu = ts->vecs_sensip; 1752 PetscFunctionReturn(0); 1753 } 1754 1755 /* ----- Routines to initialize and destroy a timestepper ---- */ 1756 #undef __FUNCT__ 1757 #define __FUNCT__ "TSSetProblemType" 1758 /*@ 1759 TSSetProblemType - Sets the type of problem to be solved. 1760 1761 Not collective 1762 1763 Input Parameters: 1764 + ts - The TS 1765 - type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms 1766 .vb 1767 U_t - A U = 0 (linear) 1768 U_t - A(t) U = 0 (linear) 1769 F(t,U,U_t) = 0 (nonlinear) 1770 .ve 1771 1772 Level: beginner 1773 1774 .keywords: TS, problem type 1775 .seealso: TSSetUp(), TSProblemType, TS 1776 @*/ 1777 PetscErrorCode TSSetProblemType(TS ts, TSProblemType type) 1778 { 1779 PetscErrorCode ierr; 1780 1781 PetscFunctionBegin; 1782 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 1783 ts->problem_type = type; 1784 if (type == TS_LINEAR) { 1785 SNES snes; 1786 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1787 ierr = SNESSetType(snes,SNESKSPONLY);CHKERRQ(ierr); 1788 } 1789 PetscFunctionReturn(0); 1790 } 1791 1792 #undef __FUNCT__ 1793 #define __FUNCT__ "TSGetProblemType" 1794 /*@C 1795 TSGetProblemType - Gets the type of problem to be solved. 1796 1797 Not collective 1798 1799 Input Parameter: 1800 . ts - The TS 1801 1802 Output Parameter: 1803 . type - One of TS_LINEAR, TS_NONLINEAR where these types refer to problems of the forms 1804 .vb 1805 M U_t = A U 1806 M(t) U_t = A(t) U 1807 F(t,U,U_t) 1808 .ve 1809 1810 Level: beginner 1811 1812 .keywords: TS, problem type 1813 .seealso: TSSetUp(), TSProblemType, TS 1814 @*/ 1815 PetscErrorCode TSGetProblemType(TS ts, TSProblemType *type) 1816 { 1817 PetscFunctionBegin; 1818 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 1819 PetscValidIntPointer(type,2); 1820 *type = ts->problem_type; 1821 PetscFunctionReturn(0); 1822 } 1823 1824 #undef __FUNCT__ 1825 #define __FUNCT__ "TSSetUp" 1826 /*@ 1827 TSSetUp - Sets up the internal data structures for the later use 1828 of a timestepper. 1829 1830 Collective on TS 1831 1832 Input Parameter: 1833 . ts - the TS context obtained from TSCreate() 1834 1835 Notes: 1836 For basic use of the TS solvers the user need not explicitly call 1837 TSSetUp(), since these actions will automatically occur during 1838 the call to TSStep(). However, if one wishes to control this 1839 phase separately, TSSetUp() should be called after TSCreate() 1840 and optional routines of the form TSSetXXX(), but before TSStep(). 1841 1842 Level: advanced 1843 1844 .keywords: TS, timestep, setup 1845 1846 .seealso: TSCreate(), TSStep(), TSDestroy() 1847 @*/ 1848 PetscErrorCode TSSetUp(TS ts) 1849 { 1850 PetscErrorCode ierr; 1851 DM dm; 1852 PetscErrorCode (*func)(SNES,Vec,Vec,void*); 1853 PetscErrorCode (*jac)(SNES,Vec,Mat,Mat,void*); 1854 TSIJacobian ijac; 1855 TSRHSJacobian rhsjac; 1856 1857 PetscFunctionBegin; 1858 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1859 if (ts->setupcalled) PetscFunctionReturn(0); 1860 1861 ts->total_steps = 0; 1862 if (!((PetscObject)ts)->type_name) { 1863 ierr = TSSetType(ts,TSEULER);CHKERRQ(ierr); 1864 } 1865 1866 if (!ts->vec_sol) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call TSSetSolution() first"); 1867 1868 1869 ierr = TSGetAdapt(ts,&ts->adapt);CHKERRQ(ierr); 1870 1871 if (ts->rhsjacobian.reuse) { 1872 Mat Amat,Pmat; 1873 SNES snes; 1874 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 1875 ierr = SNESGetJacobian(snes,&Amat,&Pmat,NULL,NULL);CHKERRQ(ierr); 1876 /* Matching matrices implies that an IJacobian is NOT set, because if it had been set, the IJacobian's matrix would 1877 * have displaced the RHS matrix */ 1878 if (Amat == ts->Arhs) { 1879 ierr = MatDuplicate(ts->Arhs,MAT_DO_NOT_COPY_VALUES,&Amat);CHKERRQ(ierr); 1880 ierr = SNESSetJacobian(snes,Amat,NULL,NULL,NULL);CHKERRQ(ierr); 1881 ierr = MatDestroy(&Amat);CHKERRQ(ierr); 1882 } 1883 if (Pmat == ts->Brhs) { 1884 ierr = MatDuplicate(ts->Brhs,MAT_DO_NOT_COPY_VALUES,&Pmat);CHKERRQ(ierr); 1885 ierr = SNESSetJacobian(snes,NULL,Pmat,NULL,NULL);CHKERRQ(ierr); 1886 ierr = MatDestroy(&Pmat);CHKERRQ(ierr); 1887 } 1888 } 1889 if (ts->ops->setup) { 1890 ierr = (*ts->ops->setup)(ts);CHKERRQ(ierr); 1891 } 1892 1893 /* in the case where we've set a DMTSFunction or what have you, we need the default SNESFunction 1894 to be set right but can't do it elsewhere due to the overreliance on ctx=ts. 1895 */ 1896 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 1897 ierr = DMSNESGetFunction(dm,&func,NULL);CHKERRQ(ierr); 1898 if (!func) { 1899 ierr =DMSNESSetFunction(dm,SNESTSFormFunction,ts);CHKERRQ(ierr); 1900 } 1901 /* if the SNES doesn't have a jacobian set and the TS has an ijacobian or rhsjacobian set, set the SNES to use it. 1902 Otherwise, the SNES will use coloring internally to form the Jacobian. 1903 */ 1904 ierr = DMSNESGetJacobian(dm,&jac,NULL);CHKERRQ(ierr); 1905 ierr = DMTSGetIJacobian(dm,&ijac,NULL);CHKERRQ(ierr); 1906 ierr = DMTSGetRHSJacobian(dm,&rhsjac,NULL);CHKERRQ(ierr); 1907 if (!jac && (ijac || rhsjac)) { 1908 ierr = DMSNESSetJacobian(dm,SNESTSFormJacobian,ts);CHKERRQ(ierr); 1909 } 1910 ts->setupcalled = PETSC_TRUE; 1911 PetscFunctionReturn(0); 1912 } 1913 1914 #undef __FUNCT__ 1915 #define __FUNCT__ "TSAdjointSetUp" 1916 /*@ 1917 TSAdjointSetUp - Sets up the internal data structures for the later use 1918 of an adjoint solver 1919 1920 Collective on TS 1921 1922 Input Parameter: 1923 . ts - the TS context obtained from TSCreate() 1924 1925 Notes: 1926 For basic use of the TS solvers the user need not explicitly call 1927 TSSetUp(), since these actions will automatically occur during 1928 the call to TSStep(). However, if one wishes to control this 1929 phase separately, TSSetUp() should be called after TSCreate() 1930 and optional routines of the form TSSetXXX(), but before TSStep(). 1931 1932 Level: advanced 1933 1934 .keywords: TS, timestep, setup 1935 1936 .seealso: TSCreate(), TSStep(), TSDestroy() 1937 @*/ 1938 PetscErrorCode TSAdjointSetUp(TS ts) 1939 { 1940 PetscErrorCode ierr; 1941 1942 PetscFunctionBegin; 1943 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1944 if (ts->adjointsetupcalled) PetscFunctionReturn(0); 1945 if (!ts->vecs_sensi) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call TSSetCostGradients() first"); 1946 1947 ierr = VecDuplicateVecs(ts->vecs_sensi[0],ts->numcost,&ts->vecs_drdy);CHKERRQ(ierr); 1948 if (ts->vecs_sensip){ 1949 ierr = VecDuplicateVecs(ts->vecs_sensip[0],ts->numcost,&ts->vecs_drdp);CHKERRQ(ierr); 1950 } 1951 1952 if (ts->ops->adjointsetup) { 1953 ierr = (*ts->ops->adjointsetup)(ts);CHKERRQ(ierr); 1954 } 1955 ts->adjointsetupcalled = PETSC_TRUE; 1956 PetscFunctionReturn(0); 1957 } 1958 1959 #undef __FUNCT__ 1960 #define __FUNCT__ "TSReset" 1961 /*@ 1962 TSReset - Resets a TS context and removes any allocated Vecs and Mats. 1963 1964 Collective on TS 1965 1966 Input Parameter: 1967 . ts - the TS context obtained from TSCreate() 1968 1969 Level: beginner 1970 1971 .keywords: TS, timestep, reset 1972 1973 .seealso: TSCreate(), TSSetup(), TSDestroy() 1974 @*/ 1975 PetscErrorCode TSReset(TS ts) 1976 { 1977 PetscErrorCode ierr; 1978 1979 PetscFunctionBegin; 1980 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 1981 1982 if (ts->ops->reset) { 1983 ierr = (*ts->ops->reset)(ts);CHKERRQ(ierr); 1984 } 1985 if (ts->snes) {ierr = SNESReset(ts->snes);CHKERRQ(ierr);} 1986 if (ts->adapt) {ierr = TSAdaptReset(ts->adapt);CHKERRQ(ierr);} 1987 1988 ierr = MatDestroy(&ts->Arhs);CHKERRQ(ierr); 1989 ierr = MatDestroy(&ts->Brhs);CHKERRQ(ierr); 1990 ierr = VecDestroy(&ts->Frhs);CHKERRQ(ierr); 1991 ierr = VecDestroy(&ts->vec_sol);CHKERRQ(ierr); 1992 ierr = VecDestroy(&ts->vatol);CHKERRQ(ierr); 1993 ierr = VecDestroy(&ts->vrtol);CHKERRQ(ierr); 1994 ierr = VecDestroyVecs(ts->nwork,&ts->work);CHKERRQ(ierr); 1995 ierr = ISDestroy(&ts->is_diff);CHKERRQ(ierr); 1996 1997 ierr = VecDestroyVecs(ts->numcost,&ts->vecs_drdy);CHKERRQ(ierr); 1998 if (ts->vecs_drdp){ 1999 ierr = VecDestroyVecs(ts->numcost,&ts->vecs_drdp);CHKERRQ(ierr); 2000 } 2001 ts->vecs_sensi = NULL; 2002 ts->vecs_sensip = NULL; 2003 ierr = MatDestroy(&ts->Jacp);CHKERRQ(ierr); 2004 ierr = VecDestroy(&ts->vec_costintegral);CHKERRQ(ierr); 2005 ierr = VecDestroy(&ts->vec_costintegrand);CHKERRQ(ierr); 2006 ts->setupcalled = PETSC_FALSE; 2007 PetscFunctionReturn(0); 2008 } 2009 2010 #undef __FUNCT__ 2011 #define __FUNCT__ "TSDestroy" 2012 /*@ 2013 TSDestroy - Destroys the timestepper context that was created 2014 with TSCreate(). 2015 2016 Collective on TS 2017 2018 Input Parameter: 2019 . ts - the TS context obtained from TSCreate() 2020 2021 Level: beginner 2022 2023 .keywords: TS, timestepper, destroy 2024 2025 .seealso: TSCreate(), TSSetUp(), TSSolve() 2026 @*/ 2027 PetscErrorCode TSDestroy(TS *ts) 2028 { 2029 PetscErrorCode ierr; 2030 2031 PetscFunctionBegin; 2032 if (!*ts) PetscFunctionReturn(0); 2033 PetscValidHeaderSpecific((*ts),TS_CLASSID,1); 2034 if (--((PetscObject)(*ts))->refct > 0) {*ts = 0; PetscFunctionReturn(0);} 2035 2036 ierr = TSReset((*ts));CHKERRQ(ierr); 2037 2038 /* if memory was published with SAWs then destroy it */ 2039 ierr = PetscObjectSAWsViewOff((PetscObject)*ts);CHKERRQ(ierr); 2040 if ((*ts)->ops->destroy) {ierr = (*(*ts)->ops->destroy)((*ts));CHKERRQ(ierr);} 2041 2042 ierr = TSTrajectoryDestroy(&(*ts)->trajectory);CHKERRQ(ierr); 2043 2044 ierr = TSAdaptDestroy(&(*ts)->adapt);CHKERRQ(ierr); 2045 if ((*ts)->event) { 2046 ierr = TSEventMonitorDestroy(&(*ts)->event);CHKERRQ(ierr); 2047 } 2048 ierr = SNESDestroy(&(*ts)->snes);CHKERRQ(ierr); 2049 ierr = DMDestroy(&(*ts)->dm);CHKERRQ(ierr); 2050 ierr = TSMonitorCancel((*ts));CHKERRQ(ierr); 2051 2052 ierr = PetscHeaderDestroy(ts);CHKERRQ(ierr); 2053 PetscFunctionReturn(0); 2054 } 2055 2056 #undef __FUNCT__ 2057 #define __FUNCT__ "TSGetSNES" 2058 /*@ 2059 TSGetSNES - Returns the SNES (nonlinear solver) associated with 2060 a TS (timestepper) context. Valid only for nonlinear problems. 2061 2062 Not Collective, but SNES is parallel if TS is parallel 2063 2064 Input Parameter: 2065 . ts - the TS context obtained from TSCreate() 2066 2067 Output Parameter: 2068 . snes - the nonlinear solver context 2069 2070 Notes: 2071 The user can then directly manipulate the SNES context to set various 2072 options, etc. Likewise, the user can then extract and manipulate the 2073 KSP, KSP, and PC contexts as well. 2074 2075 TSGetSNES() does not work for integrators that do not use SNES; in 2076 this case TSGetSNES() returns NULL in snes. 2077 2078 Level: beginner 2079 2080 .keywords: timestep, get, SNES 2081 @*/ 2082 PetscErrorCode TSGetSNES(TS ts,SNES *snes) 2083 { 2084 PetscErrorCode ierr; 2085 2086 PetscFunctionBegin; 2087 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2088 PetscValidPointer(snes,2); 2089 if (!ts->snes) { 2090 ierr = SNESCreate(PetscObjectComm((PetscObject)ts),&ts->snes);CHKERRQ(ierr); 2091 ierr = SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts);CHKERRQ(ierr); 2092 ierr = PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->snes);CHKERRQ(ierr); 2093 ierr = PetscObjectIncrementTabLevel((PetscObject)ts->snes,(PetscObject)ts,1);CHKERRQ(ierr); 2094 if (ts->dm) {ierr = SNESSetDM(ts->snes,ts->dm);CHKERRQ(ierr);} 2095 if (ts->problem_type == TS_LINEAR) { 2096 ierr = SNESSetType(ts->snes,SNESKSPONLY);CHKERRQ(ierr); 2097 } 2098 } 2099 *snes = ts->snes; 2100 PetscFunctionReturn(0); 2101 } 2102 2103 #undef __FUNCT__ 2104 #define __FUNCT__ "TSSetSNES" 2105 /*@ 2106 TSSetSNES - Set the SNES (nonlinear solver) to be used by the timestepping context 2107 2108 Collective 2109 2110 Input Parameter: 2111 + ts - the TS context obtained from TSCreate() 2112 - snes - the nonlinear solver context 2113 2114 Notes: 2115 Most users should have the TS created by calling TSGetSNES() 2116 2117 Level: developer 2118 2119 .keywords: timestep, set, SNES 2120 @*/ 2121 PetscErrorCode TSSetSNES(TS ts,SNES snes) 2122 { 2123 PetscErrorCode ierr; 2124 PetscErrorCode (*func)(SNES,Vec,Mat,Mat,void*); 2125 2126 PetscFunctionBegin; 2127 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2128 PetscValidHeaderSpecific(snes,SNES_CLASSID,2); 2129 ierr = PetscObjectReference((PetscObject)snes);CHKERRQ(ierr); 2130 ierr = SNESDestroy(&ts->snes);CHKERRQ(ierr); 2131 2132 ts->snes = snes; 2133 2134 ierr = SNESSetFunction(ts->snes,NULL,SNESTSFormFunction,ts);CHKERRQ(ierr); 2135 ierr = SNESGetJacobian(ts->snes,NULL,NULL,&func,NULL);CHKERRQ(ierr); 2136 if (func == SNESTSFormJacobian) { 2137 ierr = SNESSetJacobian(ts->snes,NULL,NULL,SNESTSFormJacobian,ts);CHKERRQ(ierr); 2138 } 2139 PetscFunctionReturn(0); 2140 } 2141 2142 #undef __FUNCT__ 2143 #define __FUNCT__ "TSGetKSP" 2144 /*@ 2145 TSGetKSP - Returns the KSP (linear solver) associated with 2146 a TS (timestepper) context. 2147 2148 Not Collective, but KSP is parallel if TS is parallel 2149 2150 Input Parameter: 2151 . ts - the TS context obtained from TSCreate() 2152 2153 Output Parameter: 2154 . ksp - the nonlinear solver context 2155 2156 Notes: 2157 The user can then directly manipulate the KSP context to set various 2158 options, etc. Likewise, the user can then extract and manipulate the 2159 KSP and PC contexts as well. 2160 2161 TSGetKSP() does not work for integrators that do not use KSP; 2162 in this case TSGetKSP() returns NULL in ksp. 2163 2164 Level: beginner 2165 2166 .keywords: timestep, get, KSP 2167 @*/ 2168 PetscErrorCode TSGetKSP(TS ts,KSP *ksp) 2169 { 2170 PetscErrorCode ierr; 2171 SNES snes; 2172 2173 PetscFunctionBegin; 2174 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2175 PetscValidPointer(ksp,2); 2176 if (!((PetscObject)ts)->type_name) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_NULL,"KSP is not created yet. Call TSSetType() first"); 2177 if (ts->problem_type != TS_LINEAR) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Linear only; use TSGetSNES()"); 2178 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 2179 ierr = SNESGetKSP(snes,ksp);CHKERRQ(ierr); 2180 PetscFunctionReturn(0); 2181 } 2182 2183 /* ----------- Routines to set solver parameters ---------- */ 2184 2185 #undef __FUNCT__ 2186 #define __FUNCT__ "TSGetDuration" 2187 /*@ 2188 TSGetDuration - Gets the maximum number of timesteps to use and 2189 maximum time for iteration. 2190 2191 Not Collective 2192 2193 Input Parameters: 2194 + ts - the TS context obtained from TSCreate() 2195 . maxsteps - maximum number of iterations to use, or NULL 2196 - maxtime - final time to iterate to, or NULL 2197 2198 Level: intermediate 2199 2200 .keywords: TS, timestep, get, maximum, iterations, time 2201 @*/ 2202 PetscErrorCode TSGetDuration(TS ts, PetscInt *maxsteps, PetscReal *maxtime) 2203 { 2204 PetscFunctionBegin; 2205 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2206 if (maxsteps) { 2207 PetscValidIntPointer(maxsteps,2); 2208 *maxsteps = ts->max_steps; 2209 } 2210 if (maxtime) { 2211 PetscValidScalarPointer(maxtime,3); 2212 *maxtime = ts->max_time; 2213 } 2214 PetscFunctionReturn(0); 2215 } 2216 2217 #undef __FUNCT__ 2218 #define __FUNCT__ "TSSetDuration" 2219 /*@ 2220 TSSetDuration - Sets the maximum number of timesteps to use and 2221 maximum time for iteration. 2222 2223 Logically Collective on TS 2224 2225 Input Parameters: 2226 + ts - the TS context obtained from TSCreate() 2227 . maxsteps - maximum number of iterations to use 2228 - maxtime - final time to iterate to 2229 2230 Options Database Keys: 2231 . -ts_max_steps <maxsteps> - Sets maxsteps 2232 . -ts_final_time <maxtime> - Sets maxtime 2233 2234 Notes: 2235 The default maximum number of iterations is 5000. Default time is 5.0 2236 2237 Level: intermediate 2238 2239 .keywords: TS, timestep, set, maximum, iterations 2240 2241 .seealso: TSSetExactFinalTime() 2242 @*/ 2243 PetscErrorCode TSSetDuration(TS ts,PetscInt maxsteps,PetscReal maxtime) 2244 { 2245 PetscFunctionBegin; 2246 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2247 PetscValidLogicalCollectiveInt(ts,maxsteps,2); 2248 PetscValidLogicalCollectiveReal(ts,maxtime,2); 2249 if (maxsteps >= 0) ts->max_steps = maxsteps; 2250 if (maxtime != PETSC_DEFAULT) ts->max_time = maxtime; 2251 PetscFunctionReturn(0); 2252 } 2253 2254 #undef __FUNCT__ 2255 #define __FUNCT__ "TSSetSolution" 2256 /*@ 2257 TSSetSolution - Sets the initial solution vector 2258 for use by the TS routines. 2259 2260 Logically Collective on TS and Vec 2261 2262 Input Parameters: 2263 + ts - the TS context obtained from TSCreate() 2264 - u - the solution vector 2265 2266 Level: beginner 2267 2268 .keywords: TS, timestep, set, solution, initial conditions 2269 @*/ 2270 PetscErrorCode TSSetSolution(TS ts,Vec u) 2271 { 2272 PetscErrorCode ierr; 2273 DM dm; 2274 2275 PetscFunctionBegin; 2276 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2277 PetscValidHeaderSpecific(u,VEC_CLASSID,2); 2278 ierr = PetscObjectReference((PetscObject)u);CHKERRQ(ierr); 2279 ierr = VecDestroy(&ts->vec_sol);CHKERRQ(ierr); 2280 2281 ts->vec_sol = u; 2282 2283 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 2284 ierr = DMShellSetGlobalVector(dm,u);CHKERRQ(ierr); 2285 PetscFunctionReturn(0); 2286 } 2287 2288 #undef __FUNCT__ 2289 #define __FUNCT__ "TSAdjointSetSteps" 2290 /*@ 2291 TSAdjointSetSteps - Sets the number of steps the adjoint solver should take backward in time 2292 2293 Logically Collective on TS 2294 2295 Input Parameters: 2296 + ts - the TS context obtained from TSCreate() 2297 . steps - number of steps to use 2298 2299 Level: intermediate 2300 2301 Notes: Normally one does not call this and TSAdjointSolve() integrates back to the original timestep. One can call this 2302 so as to integrate back to less than the original timestep 2303 2304 .keywords: TS, timestep, set, maximum, iterations 2305 2306 .seealso: TSSetExactFinalTime() 2307 @*/ 2308 PetscErrorCode TSAdjointSetSteps(TS ts,PetscInt steps) 2309 { 2310 PetscFunctionBegin; 2311 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2312 PetscValidLogicalCollectiveInt(ts,steps,2); 2313 if (steps < 0) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Cannot step back a negative number of steps"); 2314 if (steps > ts->total_steps) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Cannot step back more than the total number of forward steps"); 2315 ts->adjoint_max_steps = steps; 2316 PetscFunctionReturn(0); 2317 } 2318 2319 #undef __FUNCT__ 2320 #define __FUNCT__ "TSSetCostGradients" 2321 /*@ 2322 TSSetCostGradients - Sets the initial value of the gradients of the cost function w.r.t. initial conditions and w.r.t. the problem parameters 2323 for use by the TSAdjoint routines. 2324 2325 Logically Collective on TS and Vec 2326 2327 Input Parameters: 2328 + ts - the TS context obtained from TSCreate() 2329 . lambda - gradients with respect to the initial condition variables, the dimension and parallel layout of these vectors is the same as the ODE solution vector 2330 - mu - gradients with respect to the parameters, the number of entries in these vectors is the same as the number of parameters 2331 2332 Level: beginner 2333 2334 Notes: the entries in these vectors must be correctly initialized with the values lamda_i = df/dy|finaltime mu_i = df/dp|finaltime 2335 2336 .keywords: TS, timestep, set, sensitivity, initial conditions 2337 @*/ 2338 PetscErrorCode TSSetCostGradients(TS ts,PetscInt numcost,Vec *lambda,Vec *mu) 2339 { 2340 PetscFunctionBegin; 2341 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2342 PetscValidPointer(lambda,2); 2343 ts->vecs_sensi = lambda; 2344 ts->vecs_sensip = mu; 2345 if (ts->numcost && ts->numcost!=numcost) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"The number of cost functions (2rd parameter of TSSetCostIntegrand()) is inconsistent with the one set by TSSetCostIntegrand"); 2346 ts->numcost = numcost; 2347 PetscFunctionReturn(0); 2348 } 2349 2350 #undef __FUNCT__ 2351 #define __FUNCT__ "TSAdjointSetRHSJacobian" 2352 /*@C 2353 TSAdjointSetRHSJacobian - Sets the function that computes the Jacobian of G w.r.t. the parameters p where y_t = G(y,p,t), as well as the location to store the matrix. 2354 2355 Logically Collective on TS 2356 2357 Input Parameters: 2358 + ts - The TS context obtained from TSCreate() 2359 - func - The function 2360 2361 Calling sequence of func: 2362 $ func (TS ts,PetscReal t,Vec y,Mat A,void *ctx); 2363 + t - current timestep 2364 . y - input vector (current ODE solution) 2365 . A - output matrix 2366 - ctx - [optional] user-defined function context 2367 2368 Level: intermediate 2369 2370 Notes: Amat has the same number of rows and the same row parallel layout as u, Amat has the same number of columns and parallel layout as p 2371 2372 .keywords: TS, sensitivity 2373 .seealso: 2374 @*/ 2375 PetscErrorCode TSAdjointSetRHSJacobian(TS ts,Mat Amat,PetscErrorCode (*func)(TS,PetscReal,Vec,Mat,void*),void *ctx) 2376 { 2377 PetscErrorCode ierr; 2378 2379 PetscFunctionBegin; 2380 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2381 if (Amat) PetscValidHeaderSpecific(Amat,MAT_CLASSID,2); 2382 2383 ts->rhsjacobianp = func; 2384 ts->rhsjacobianpctx = ctx; 2385 if(Amat) { 2386 ierr = PetscObjectReference((PetscObject)Amat);CHKERRQ(ierr); 2387 ierr = MatDestroy(&ts->Jacp);CHKERRQ(ierr); 2388 ts->Jacp = Amat; 2389 } 2390 PetscFunctionReturn(0); 2391 } 2392 2393 #undef __FUNCT__ 2394 #define __FUNCT__ "TSAdjointComputeRHSJacobian" 2395 /*@C 2396 TSAdjointComputeRHSJacobian - Runs the user-defined Jacobian function. 2397 2398 Collective on TS 2399 2400 Input Parameters: 2401 . ts - The TS context obtained from TSCreate() 2402 2403 Level: developer 2404 2405 .keywords: TS, sensitivity 2406 .seealso: TSAdjointSetRHSJacobian() 2407 @*/ 2408 PetscErrorCode TSAdjointComputeRHSJacobian(TS ts,PetscReal t,Vec X,Mat Amat) 2409 { 2410 PetscErrorCode ierr; 2411 2412 PetscFunctionBegin; 2413 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2414 PetscValidHeaderSpecific(X,VEC_CLASSID,3); 2415 PetscValidPointer(Amat,4); 2416 2417 PetscStackPush("TS user JacobianP function for sensitivity analysis"); 2418 ierr = (*ts->rhsjacobianp)(ts,t,X,Amat,ts->rhsjacobianpctx); CHKERRQ(ierr); 2419 PetscStackPop; 2420 PetscFunctionReturn(0); 2421 } 2422 2423 #undef __FUNCT__ 2424 #define __FUNCT__ "TSSetCostIntegrand" 2425 /*@C 2426 TSSetCostIntegrand - Sets the routine for evaluating the integral term in one or more cost functions 2427 2428 Logically Collective on TS 2429 2430 Input Parameters: 2431 + ts - the TS context obtained from TSCreate() 2432 . numcost - number of gradients to be computed, this is the number of cost functions 2433 . rf - routine for evaluating the integrand function 2434 . drdyf - function that computes the gradients of the r's with respect to y,NULL if not a function y 2435 . drdpf - function that computes the gradients of the r's with respect to p, NULL if not a function of p 2436 - ctx - [optional] user-defined context for private data for the function evaluation routine (may be NULL) 2437 2438 Calling sequence of rf: 2439 $ rf(TS ts,PetscReal t,Vec y,Vec f[],void *ctx); 2440 2441 + t - current timestep 2442 . y - input vector 2443 . f - function result; one vector entry for each cost function 2444 - ctx - [optional] user-defined function context 2445 2446 Calling sequence of drdyf: 2447 $ PetscErroCode drdyf(TS ts,PetscReal t,Vec y,Vec *drdy,void *ctx); 2448 2449 Calling sequence of drdpf: 2450 $ PetscErroCode drdpf(TS ts,PetscReal t,Vec y,Vec *drdp,void *ctx); 2451 2452 Level: intermediate 2453 2454 Notes: For optimization there is generally a single cost function, numcost = 1. For sensitivities there may be multiple cost functions 2455 2456 .keywords: TS, sensitivity analysis, timestep, set, quadrature, function 2457 2458 .seealso: TSAdjointSetRHSJacobian(),TSGetCostGradients(), TSSetCostGradients() 2459 @*/ 2460 PetscErrorCode TSSetCostIntegrand(TS ts,PetscInt numcost, PetscErrorCode (*rf)(TS,PetscReal,Vec,Vec,void*), 2461 PetscErrorCode (*drdyf)(TS,PetscReal,Vec,Vec*,void*), 2462 PetscErrorCode (*drdpf)(TS,PetscReal,Vec,Vec*,void*),void *ctx) 2463 { 2464 PetscErrorCode ierr; 2465 2466 PetscFunctionBegin; 2467 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2468 if (ts->numcost && ts->numcost!=numcost) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_USER,"The number of cost functions (2rd parameter of TSSetCostIntegrand()) is inconsistent with the one set by TSSetCostGradients()"); 2469 if (!ts->numcost) ts->numcost=numcost; 2470 2471 ierr = VecCreateSeq(PETSC_COMM_SELF,numcost,&ts->vec_costintegral);CHKERRQ(ierr); 2472 ierr = VecDuplicate(ts->vec_costintegral,&ts->vec_costintegrand);CHKERRQ(ierr); 2473 ts->costintegrand = rf; 2474 ts->costintegrandctx = ctx; 2475 ts->drdyfunction = drdyf; 2476 ts->drdpfunction = drdpf; 2477 PetscFunctionReturn(0); 2478 } 2479 2480 #undef __FUNCT__ 2481 #define __FUNCT__ "TSGetCostIntegral" 2482 /*@ 2483 TSGetCostIntegral - Returns the values of the integral term in the cost functions. 2484 It is valid to call the routine after a backward run. 2485 2486 Not Collective 2487 2488 Input Parameter: 2489 . ts - the TS context obtained from TSCreate() 2490 2491 Output Parameter: 2492 . v - the vector containing the integrals for each cost function 2493 2494 Level: intermediate 2495 2496 .seealso: TSSetCostIntegrand() 2497 2498 .keywords: TS, sensitivity analysis 2499 @*/ 2500 PetscErrorCode TSGetCostIntegral(TS ts,Vec *v) 2501 { 2502 PetscFunctionBegin; 2503 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2504 PetscValidPointer(v,2); 2505 *v = ts->vec_costintegral; 2506 PetscFunctionReturn(0); 2507 } 2508 2509 #undef __FUNCT__ 2510 #define __FUNCT__ "TSAdjointComputeCostIntegrand" 2511 /*@ 2512 TSAdjointComputeCostIntegrand - Evaluates the integral function in the cost functions. 2513 2514 Input Parameters: 2515 + ts - the TS context 2516 . t - current time 2517 - y - state vector, i.e. current solution 2518 2519 Output Parameter: 2520 . q - vector of size numcost to hold the outputs 2521 2522 Note: 2523 Most users should not need to explicitly call this routine, as it 2524 is used internally within the sensitivity analysis context. 2525 2526 Level: developer 2527 2528 .keywords: TS, compute 2529 2530 .seealso: TSSetCostIntegrand() 2531 @*/ 2532 PetscErrorCode TSAdjointComputeCostIntegrand(TS ts,PetscReal t,Vec y,Vec q) 2533 { 2534 PetscErrorCode ierr; 2535 2536 PetscFunctionBegin; 2537 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2538 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2539 PetscValidHeaderSpecific(q,VEC_CLASSID,4); 2540 2541 ierr = PetscLogEventBegin(TS_FunctionEval,ts,y,q,0);CHKERRQ(ierr); 2542 if (ts->costintegrand) { 2543 PetscStackPush("TS user integrand in the cost function"); 2544 ierr = (*ts->costintegrand)(ts,t,y,q,ts->costintegrandctx);CHKERRQ(ierr); 2545 PetscStackPop; 2546 } else { 2547 ierr = VecZeroEntries(q);CHKERRQ(ierr); 2548 } 2549 2550 ierr = PetscLogEventEnd(TS_FunctionEval,ts,y,q,0);CHKERRQ(ierr); 2551 PetscFunctionReturn(0); 2552 } 2553 2554 #undef __FUNCT__ 2555 #define __FUNCT__ "TSAdjointComputeDRDYFunction" 2556 /*@ 2557 TSAdjointComputeDRDYFunction - Runs the user-defined DRDY function. 2558 2559 Collective on TS 2560 2561 Input Parameters: 2562 . ts - The TS context obtained from TSCreate() 2563 2564 Notes: 2565 TSAdjointComputeDRDYFunction() is typically used for sensitivity implementation, 2566 so most users would not generally call this routine themselves. 2567 2568 Level: developer 2569 2570 .keywords: TS, sensitivity 2571 .seealso: TSAdjointComputeDRDYFunction() 2572 @*/ 2573 PetscErrorCode TSAdjointComputeDRDYFunction(TS ts,PetscReal t,Vec y,Vec *drdy) 2574 { 2575 PetscErrorCode ierr; 2576 2577 PetscFunctionBegin; 2578 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2579 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2580 2581 PetscStackPush("TS user DRDY function for sensitivity analysis"); 2582 ierr = (*ts->drdyfunction)(ts,t,y,drdy,ts->costintegrandctx); CHKERRQ(ierr); 2583 PetscStackPop; 2584 PetscFunctionReturn(0); 2585 } 2586 2587 #undef __FUNCT__ 2588 #define __FUNCT__ "TSAdjointComputeDRDPFunction" 2589 /*@ 2590 TSAdjointComputeDRDPFunction - Runs the user-defined DRDP function. 2591 2592 Collective on TS 2593 2594 Input Parameters: 2595 . ts - The TS context obtained from TSCreate() 2596 2597 Notes: 2598 TSDRDPFunction() is typically used for sensitivity implementation, 2599 so most users would not generally call this routine themselves. 2600 2601 Level: developer 2602 2603 .keywords: TS, sensitivity 2604 .seealso: TSAdjointSetDRDPFunction() 2605 @*/ 2606 PetscErrorCode TSAdjointComputeDRDPFunction(TS ts,PetscReal t,Vec y,Vec *drdp) 2607 { 2608 PetscErrorCode ierr; 2609 2610 PetscFunctionBegin; 2611 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2612 PetscValidHeaderSpecific(y,VEC_CLASSID,3); 2613 2614 PetscStackPush("TS user DRDP function for sensitivity analysis"); 2615 ierr = (*ts->drdpfunction)(ts,t,y,drdp,ts->costintegrandctx); CHKERRQ(ierr); 2616 PetscStackPop; 2617 PetscFunctionReturn(0); 2618 } 2619 2620 #undef __FUNCT__ 2621 #define __FUNCT__ "TSSetPreStep" 2622 /*@C 2623 TSSetPreStep - Sets the general-purpose function 2624 called once at the beginning of each time step. 2625 2626 Logically Collective on TS 2627 2628 Input Parameters: 2629 + ts - The TS context obtained from TSCreate() 2630 - func - The function 2631 2632 Calling sequence of func: 2633 . func (TS ts); 2634 2635 Level: intermediate 2636 2637 Note: 2638 If a step is rejected, TSStep() will call this routine again before each attempt. 2639 The last completed time step number can be queried using TSGetTimeStepNumber(), the 2640 size of the step being attempted can be obtained using TSGetTimeStep(). 2641 2642 .keywords: TS, timestep 2643 .seealso: TSSetPreStage(), TSSetPostStage(), TSSetPostStep(), TSStep() 2644 @*/ 2645 PetscErrorCode TSSetPreStep(TS ts, PetscErrorCode (*func)(TS)) 2646 { 2647 PetscFunctionBegin; 2648 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2649 ts->prestep = func; 2650 PetscFunctionReturn(0); 2651 } 2652 2653 #undef __FUNCT__ 2654 #define __FUNCT__ "TSPreStep" 2655 /*@ 2656 TSPreStep - Runs the user-defined pre-step function. 2657 2658 Collective on TS 2659 2660 Input Parameters: 2661 . ts - The TS context obtained from TSCreate() 2662 2663 Notes: 2664 TSPreStep() is typically used within time stepping implementations, 2665 so most users would not generally call this routine themselves. 2666 2667 Level: developer 2668 2669 .keywords: TS, timestep 2670 .seealso: TSSetPreStep(), TSPreStage(), TSPostStage(), TSPostStep() 2671 @*/ 2672 PetscErrorCode TSPreStep(TS ts) 2673 { 2674 PetscErrorCode ierr; 2675 2676 PetscFunctionBegin; 2677 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2678 if (ts->prestep) { 2679 PetscStackCallStandard((*ts->prestep),(ts)); 2680 } 2681 PetscFunctionReturn(0); 2682 } 2683 2684 #undef __FUNCT__ 2685 #define __FUNCT__ "TSSetPreStage" 2686 /*@C 2687 TSSetPreStage - Sets the general-purpose function 2688 called once at the beginning of each stage. 2689 2690 Logically Collective on TS 2691 2692 Input Parameters: 2693 + ts - The TS context obtained from TSCreate() 2694 - func - The function 2695 2696 Calling sequence of func: 2697 . PetscErrorCode func(TS ts, PetscReal stagetime); 2698 2699 Level: intermediate 2700 2701 Note: 2702 There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried. 2703 The time step number being computed can be queried using TSGetTimeStepNumber() and the total size of the step being 2704 attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime(). 2705 2706 .keywords: TS, timestep 2707 .seealso: TSSetPostStage(), TSSetPreStep(), TSSetPostStep(), TSGetApplicationContext() 2708 @*/ 2709 PetscErrorCode TSSetPreStage(TS ts, PetscErrorCode (*func)(TS,PetscReal)) 2710 { 2711 PetscFunctionBegin; 2712 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2713 ts->prestage = func; 2714 PetscFunctionReturn(0); 2715 } 2716 2717 #undef __FUNCT__ 2718 #define __FUNCT__ "TSSetPostStage" 2719 /*@C 2720 TSSetPostStage - Sets the general-purpose function 2721 called once at the end of each stage. 2722 2723 Logically Collective on TS 2724 2725 Input Parameters: 2726 + ts - The TS context obtained from TSCreate() 2727 - func - The function 2728 2729 Calling sequence of func: 2730 . PetscErrorCode func(TS ts, PetscReal stagetime, PetscInt stageindex, Vec* Y); 2731 2732 Level: intermediate 2733 2734 Note: 2735 There may be several stages per time step. If the solve for a given stage fails, the step may be rejected and retried. 2736 The time step number being computed can be queried using TSGetTimeStepNumber() and the total size of the step being 2737 attempted can be obtained using TSGetTimeStep(). The time at the start of the step is available via TSGetTime(). 2738 2739 .keywords: TS, timestep 2740 .seealso: TSSetPreStage(), TSSetPreStep(), TSSetPostStep(), TSGetApplicationContext() 2741 @*/ 2742 PetscErrorCode TSSetPostStage(TS ts, PetscErrorCode (*func)(TS,PetscReal,PetscInt,Vec*)) 2743 { 2744 PetscFunctionBegin; 2745 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2746 ts->poststage = func; 2747 PetscFunctionReturn(0); 2748 } 2749 2750 #undef __FUNCT__ 2751 #define __FUNCT__ "TSPreStage" 2752 /*@ 2753 TSPreStage - Runs the user-defined pre-stage function set using TSSetPreStage() 2754 2755 Collective on TS 2756 2757 Input Parameters: 2758 . ts - The TS context obtained from TSCreate() 2759 stagetime - The absolute time of the current stage 2760 2761 Notes: 2762 TSPreStage() is typically used within time stepping implementations, 2763 most users would not generally call this routine themselves. 2764 2765 Level: developer 2766 2767 .keywords: TS, timestep 2768 .seealso: TSPostStage(), TSSetPreStep(), TSPreStep(), TSPostStep() 2769 @*/ 2770 PetscErrorCode TSPreStage(TS ts, PetscReal stagetime) 2771 { 2772 PetscErrorCode ierr; 2773 2774 PetscFunctionBegin; 2775 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2776 if (ts->prestage) { 2777 PetscStackCallStandard((*ts->prestage),(ts,stagetime)); 2778 } 2779 PetscFunctionReturn(0); 2780 } 2781 2782 #undef __FUNCT__ 2783 #define __FUNCT__ "TSPostStage" 2784 /*@ 2785 TSPostStage - Runs the user-defined post-stage function set using TSSetPostStage() 2786 2787 Collective on TS 2788 2789 Input Parameters: 2790 . ts - The TS context obtained from TSCreate() 2791 stagetime - The absolute time of the current stage 2792 stageindex - Stage number 2793 Y - Array of vectors (of size = total number 2794 of stages) with the stage solutions 2795 2796 Notes: 2797 TSPostStage() is typically used within time stepping implementations, 2798 most users would not generally call this routine themselves. 2799 2800 Level: developer 2801 2802 .keywords: TS, timestep 2803 .seealso: TSPreStage(), TSSetPreStep(), TSPreStep(), TSPostStep() 2804 @*/ 2805 PetscErrorCode TSPostStage(TS ts, PetscReal stagetime, PetscInt stageindex, Vec *Y) 2806 { 2807 PetscErrorCode ierr; 2808 2809 PetscFunctionBegin; 2810 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2811 if (ts->poststage) { 2812 PetscStackCallStandard((*ts->poststage),(ts,stagetime,stageindex,Y)); 2813 } 2814 PetscFunctionReturn(0); 2815 } 2816 2817 #undef __FUNCT__ 2818 #define __FUNCT__ "TSSetPostStep" 2819 /*@C 2820 TSSetPostStep - Sets the general-purpose function 2821 called once at the end of each time step. 2822 2823 Logically Collective on TS 2824 2825 Input Parameters: 2826 + ts - The TS context obtained from TSCreate() 2827 - func - The function 2828 2829 Calling sequence of func: 2830 $ func (TS ts); 2831 2832 Level: intermediate 2833 2834 .keywords: TS, timestep 2835 .seealso: TSSetPreStep(), TSSetPreStage(), TSGetTimeStep(), TSGetTimeStepNumber(), TSGetTime() 2836 @*/ 2837 PetscErrorCode TSSetPostStep(TS ts, PetscErrorCode (*func)(TS)) 2838 { 2839 PetscFunctionBegin; 2840 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 2841 ts->poststep = func; 2842 PetscFunctionReturn(0); 2843 } 2844 2845 #undef __FUNCT__ 2846 #define __FUNCT__ "TSPostStep" 2847 /*@ 2848 TSPostStep - Runs the user-defined post-step function. 2849 2850 Collective on TS 2851 2852 Input Parameters: 2853 . ts - The TS context obtained from TSCreate() 2854 2855 Notes: 2856 TSPostStep() is typically used within time stepping implementations, 2857 so most users would not generally call this routine themselves. 2858 2859 Level: developer 2860 2861 .keywords: TS, timestep 2862 @*/ 2863 PetscErrorCode TSPostStep(TS ts) 2864 { 2865 PetscErrorCode ierr; 2866 2867 PetscFunctionBegin; 2868 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2869 if (ts->poststep) { 2870 PetscStackCallStandard((*ts->poststep),(ts)); 2871 } 2872 PetscFunctionReturn(0); 2873 } 2874 2875 /* ------------ Routines to set performance monitoring options ----------- */ 2876 2877 #undef __FUNCT__ 2878 #define __FUNCT__ "TSMonitorSet" 2879 /*@C 2880 TSMonitorSet - Sets an ADDITIONAL function that is to be used at every 2881 timestep to display the iteration's progress. 2882 2883 Logically Collective on TS 2884 2885 Input Parameters: 2886 + ts - the TS context obtained from TSCreate() 2887 . monitor - monitoring routine 2888 . mctx - [optional] user-defined context for private data for the 2889 monitor routine (use NULL if no context is desired) 2890 - monitordestroy - [optional] routine that frees monitor context 2891 (may be NULL) 2892 2893 Calling sequence of monitor: 2894 $ int monitor(TS ts,PetscInt steps,PetscReal time,Vec u,void *mctx) 2895 2896 + ts - the TS context 2897 . steps - iteration number (after the final time step the monitor routine is called with a step of -1, this is at the final time which may have 2898 been interpolated to) 2899 . time - current time 2900 . u - current iterate 2901 - mctx - [optional] monitoring context 2902 2903 Notes: 2904 This routine adds an additional monitor to the list of monitors that 2905 already has been loaded. 2906 2907 Fortran notes: Only a single monitor function can be set for each TS object 2908 2909 Level: intermediate 2910 2911 .keywords: TS, timestep, set, monitor 2912 2913 .seealso: TSMonitorDefault(), TSMonitorCancel() 2914 @*/ 2915 PetscErrorCode TSMonitorSet(TS ts,PetscErrorCode (*monitor)(TS,PetscInt,PetscReal,Vec,void*),void *mctx,PetscErrorCode (*mdestroy)(void**)) 2916 { 2917 PetscFunctionBegin; 2918 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2919 if (ts->numbermonitors >= MAXTSMONITORS) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Too many monitors set"); 2920 ts->monitor[ts->numbermonitors] = monitor; 2921 ts->monitordestroy[ts->numbermonitors] = mdestroy; 2922 ts->monitorcontext[ts->numbermonitors++] = (void*)mctx; 2923 PetscFunctionReturn(0); 2924 } 2925 2926 #undef __FUNCT__ 2927 #define __FUNCT__ "TSMonitorCancel" 2928 /*@C 2929 TSMonitorCancel - Clears all the monitors that have been set on a time-step object. 2930 2931 Logically Collective on TS 2932 2933 Input Parameters: 2934 . ts - the TS context obtained from TSCreate() 2935 2936 Notes: 2937 There is no way to remove a single, specific monitor. 2938 2939 Level: intermediate 2940 2941 .keywords: TS, timestep, set, monitor 2942 2943 .seealso: TSMonitorDefault(), TSMonitorSet() 2944 @*/ 2945 PetscErrorCode TSMonitorCancel(TS ts) 2946 { 2947 PetscErrorCode ierr; 2948 PetscInt i; 2949 2950 PetscFunctionBegin; 2951 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 2952 for (i=0; i<ts->numbermonitors; i++) { 2953 if (ts->monitordestroy[i]) { 2954 ierr = (*ts->monitordestroy[i])(&ts->monitorcontext[i]);CHKERRQ(ierr); 2955 } 2956 } 2957 ts->numbermonitors = 0; 2958 PetscFunctionReturn(0); 2959 } 2960 2961 #undef __FUNCT__ 2962 #define __FUNCT__ "TSMonitorDefault" 2963 /*@ 2964 TSMonitorDefault - Sets the Default monitor 2965 2966 Level: intermediate 2967 2968 .keywords: TS, set, monitor 2969 2970 .seealso: TSMonitorDefault(), TSMonitorSet() 2971 @*/ 2972 PetscErrorCode TSMonitorDefault(TS ts,PetscInt step,PetscReal ptime,Vec v,void *dummy) 2973 { 2974 PetscErrorCode ierr; 2975 PetscViewer viewer = dummy ? (PetscViewer) dummy : PETSC_VIEWER_STDOUT_(PetscObjectComm((PetscObject)ts)); 2976 2977 PetscFunctionBegin; 2978 ierr = PetscViewerASCIIAddTab(viewer,((PetscObject)ts)->tablevel);CHKERRQ(ierr); 2979 ierr = PetscViewerASCIIPrintf(viewer,"%D TS dt %g time %g %s",step,(double)ts->time_step,(double)ptime,ts->steprollback ? "(r)\n" : "\n");CHKERRQ(ierr); 2980 ierr = PetscViewerASCIISubtractTab(viewer,((PetscObject)ts)->tablevel);CHKERRQ(ierr); 2981 PetscFunctionReturn(0); 2982 } 2983 2984 #undef __FUNCT__ 2985 #define __FUNCT__ "TSSetRetainStages" 2986 /*@ 2987 TSSetRetainStages - Request that all stages in the upcoming step be stored so that interpolation will be available. 2988 2989 Logically Collective on TS 2990 2991 Input Argument: 2992 . ts - time stepping context 2993 2994 Output Argument: 2995 . flg - PETSC_TRUE or PETSC_FALSE 2996 2997 Level: intermediate 2998 2999 .keywords: TS, set 3000 3001 .seealso: TSInterpolate(), TSSetPostStep() 3002 @*/ 3003 PetscErrorCode TSSetRetainStages(TS ts,PetscBool flg) 3004 { 3005 PetscFunctionBegin; 3006 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3007 ts->retain_stages = flg; 3008 PetscFunctionReturn(0); 3009 } 3010 3011 #undef __FUNCT__ 3012 #define __FUNCT__ "TSInterpolate" 3013 /*@ 3014 TSInterpolate - Interpolate the solution computed during the previous step to an arbitrary location in the interval 3015 3016 Collective on TS 3017 3018 Input Argument: 3019 + ts - time stepping context 3020 - t - time to interpolate to 3021 3022 Output Argument: 3023 . U - state at given time 3024 3025 Notes: 3026 The user should call TSSetRetainStages() before taking a step in which interpolation will be requested. 3027 3028 Level: intermediate 3029 3030 Developer Notes: 3031 TSInterpolate() and the storing of previous steps/stages should be generalized to support delay differential equations and continuous adjoints. 3032 3033 .keywords: TS, set 3034 3035 .seealso: TSSetRetainStages(), TSSetPostStep() 3036 @*/ 3037 PetscErrorCode TSInterpolate(TS ts,PetscReal t,Vec U) 3038 { 3039 PetscErrorCode ierr; 3040 3041 PetscFunctionBegin; 3042 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3043 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 3044 if (t < ts->ptime - ts->time_step_prev || t > ts->ptime) SETERRQ3(PetscObjectComm((PetscObject)ts),PETSC_ERR_ARG_OUTOFRANGE,"Requested time %g not in last time steps [%g,%g]",t,(double)(ts->ptime-ts->time_step_prev),(double)ts->ptime); 3045 if (!ts->ops->interpolate) SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"%s does not provide interpolation",((PetscObject)ts)->type_name); 3046 ierr = (*ts->ops->interpolate)(ts,t,U);CHKERRQ(ierr); 3047 PetscFunctionReturn(0); 3048 } 3049 3050 #undef __FUNCT__ 3051 #define __FUNCT__ "TSStep" 3052 /*@ 3053 TSStep - Steps one time step 3054 3055 Collective on TS 3056 3057 Input Parameter: 3058 . ts - the TS context obtained from TSCreate() 3059 3060 Level: developer 3061 3062 Notes: 3063 The public interface for the ODE/DAE solvers is TSSolve(), you should almost for sure be using that routine and not this routine. 3064 3065 The hook set using TSSetPreStep() is called before each attempt to take the step. In general, the time step size may 3066 be changed due to adaptive error controller or solve failures. Note that steps may contain multiple stages. 3067 3068 This may over-step the final time provided in TSSetDuration() depending on the time-step used. TSSolve() interpolates to exactly the 3069 time provided in TSSetDuration(). One can use TSInterpolate() to determine an interpolated solution within the final timestep. 3070 3071 .keywords: TS, timestep, solve 3072 3073 .seealso: TSCreate(), TSSetUp(), TSDestroy(), TSSolve(), TSSetPreStep(), TSSetPreStage(), TSSetPostStage(), TSInterpolate() 3074 @*/ 3075 PetscErrorCode TSStep(TS ts) 3076 { 3077 DM dm; 3078 PetscErrorCode ierr; 3079 static PetscBool cite = PETSC_FALSE; 3080 3081 PetscFunctionBegin; 3082 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 3083 ierr = PetscCitationsRegister("@techreport{tspaper,\n" 3084 " title = {{PETSc/TS}: A Modern Scalable {DAE/ODE} Solver Library},\n" 3085 " author = {Shrirang Abhyankar and Jed Brown and Emil Constantinescu and Debojyoti Ghosh and Barry F. Smith},\n" 3086 " type = {Preprint},\n" 3087 " number = {ANL/MCS-P5061-0114},\n" 3088 " institution = {Argonne National Laboratory},\n" 3089 " year = {2014}\n}\n",&cite);CHKERRQ(ierr); 3090 3091 ierr = TSGetDM(ts, &dm);CHKERRQ(ierr); 3092 ierr = TSSetUp(ts);CHKERRQ(ierr); 3093 3094 ts->reason = TS_CONVERGED_ITERATING; 3095 ts->ptime_prev = ts->ptime; 3096 ierr = DMSetOutputSequenceNumber(dm, ts->steps, ts->ptime);CHKERRQ(ierr); 3097 3098 if (!ts->ops->step) SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"TSStep not implemented for type '%s'",((PetscObject)ts)->type_name); 3099 ierr = PetscLogEventBegin(TS_Step,ts,0,0,0);CHKERRQ(ierr); 3100 ierr = (*ts->ops->step)(ts);CHKERRQ(ierr); 3101 ierr = PetscLogEventEnd(TS_Step,ts,0,0,0);CHKERRQ(ierr); 3102 3103 ts->time_step_prev = ts->ptime - ts->ptime_prev; 3104 ierr = DMSetOutputSequenceNumber(dm, ts->steps, ts->ptime);CHKERRQ(ierr); 3105 3106 if (ts->reason < 0) { 3107 if (ts->errorifstepfailed) { 3108 if (ts->reason == TS_DIVERGED_NONLINEAR_SOLVE) SETERRQ1(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]); 3109 else SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed due to %s",TSConvergedReasons[ts->reason]); 3110 } 3111 } else if (!ts->reason) { 3112 if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS; 3113 else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME; 3114 } 3115 ts->total_steps++; 3116 ts->steprollback = PETSC_FALSE; 3117 PetscFunctionReturn(0); 3118 } 3119 3120 #undef __FUNCT__ 3121 #define __FUNCT__ "TSAdjointStep" 3122 /*@ 3123 TSAdjointStep - Steps one time step 3124 3125 Collective on TS 3126 3127 Input Parameter: 3128 . ts - the TS context obtained from TSCreate() 3129 3130 Level: intermediate 3131 3132 Notes: 3133 The hook set using TSSetPreStep() is called before each attempt to take the step. In general, the time step size may 3134 be changed due to adaptive error controller or solve failures. Note that steps may contain multiple stages. 3135 3136 This may over-step the final time provided in TSSetDuration() depending on the time-step used. TSSolve() interpolates to exactly the 3137 time provided in TSSetDuration(). One can use TSInterpolate() to determine an interpolated solution within the final timestep. 3138 3139 .keywords: TS, timestep, solve 3140 3141 .seealso: TSCreate(), TSSetUp(), TSDestroy(), TSSolve(), TSSetPreStep(), TSSetPreStage(), TSSetPostStage(), TSInterpolate() 3142 @*/ 3143 PetscErrorCode TSAdjointStep(TS ts) 3144 { 3145 DM dm; 3146 PetscErrorCode ierr; 3147 3148 PetscFunctionBegin; 3149 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 3150 ierr = TSGetDM(ts, &dm);CHKERRQ(ierr); 3151 ierr = TSAdjointSetUp(ts);CHKERRQ(ierr); 3152 3153 ts->reason = TS_CONVERGED_ITERATING; 3154 ts->ptime_prev = ts->ptime; 3155 ierr = DMSetOutputSequenceNumber(dm, ts->steps, ts->ptime);CHKERRQ(ierr); 3156 ierr = VecViewFromOptions(ts->vec_sol, ((PetscObject) ts)->prefix, "-ts_view_solution");CHKERRQ(ierr); 3157 3158 ierr = PetscLogEventBegin(TS_Step,ts,0,0,0);CHKERRQ(ierr); 3159 if (!ts->ops->adjointstep) SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed because the adjoint of %s has not been implemented, try other time stepping methods for adjoint sensitivity analysis",((PetscObject)ts)->type_name); 3160 ierr = (*ts->ops->adjointstep)(ts);CHKERRQ(ierr); 3161 ierr = PetscLogEventEnd(TS_Step,ts,0,0,0);CHKERRQ(ierr); 3162 3163 ts->time_step_prev = ts->ptime - ts->ptime_prev; 3164 ierr = DMSetOutputSequenceNumber(dm, ts->steps, ts->ptime);CHKERRQ(ierr); 3165 3166 if (ts->reason < 0) { 3167 if (ts->errorifstepfailed) { 3168 if (ts->reason == TS_DIVERGED_NONLINEAR_SOLVE) { 3169 SETERRQ1(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]); 3170 } else if (ts->reason == TS_DIVERGED_STEP_REJECTED) { 3171 SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed due to %s, increase -ts_max_reject or make negative to attempt recovery",TSConvergedReasons[ts->reason]); 3172 } else SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_NOT_CONVERGED,"TSStep has failed due to %s",TSConvergedReasons[ts->reason]); 3173 } 3174 } else if (!ts->reason) { 3175 if (ts->steps >= ts->adjoint_max_steps) ts->reason = TS_CONVERGED_ITS; 3176 else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME; 3177 } 3178 ts->total_steps--; 3179 PetscFunctionReturn(0); 3180 } 3181 3182 #undef __FUNCT__ 3183 #define __FUNCT__ "TSEvaluateStep" 3184 /*@ 3185 TSEvaluateStep - Evaluate the solution at the end of a time step with a given order of accuracy. 3186 3187 Collective on TS 3188 3189 Input Arguments: 3190 + ts - time stepping context 3191 . order - desired order of accuracy 3192 - done - whether the step was evaluated at this order (pass NULL to generate an error if not available) 3193 3194 Output Arguments: 3195 . U - state at the end of the current step 3196 3197 Level: advanced 3198 3199 Notes: 3200 This function cannot be called until all stages have been evaluated. 3201 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. 3202 3203 .seealso: TSStep(), TSAdapt 3204 @*/ 3205 PetscErrorCode TSEvaluateStep(TS ts,PetscInt order,Vec U,PetscBool *done) 3206 { 3207 PetscErrorCode ierr; 3208 3209 PetscFunctionBegin; 3210 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3211 PetscValidType(ts,1); 3212 PetscValidHeaderSpecific(U,VEC_CLASSID,3); 3213 if (!ts->ops->evaluatestep) SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"TSEvaluateStep not implemented for type '%s'",((PetscObject)ts)->type_name); 3214 ierr = (*ts->ops->evaluatestep)(ts,order,U,done);CHKERRQ(ierr); 3215 PetscFunctionReturn(0); 3216 } 3217 3218 3219 #undef __FUNCT__ 3220 #define __FUNCT__ "TSSolve" 3221 /*@ 3222 TSSolve - Steps the requested number of timesteps. 3223 3224 Collective on TS 3225 3226 Input Parameter: 3227 + ts - the TS context obtained from TSCreate() 3228 - u - the solution vector (can be null if TSSetSolution() was used, otherwise must contain the initial conditions) 3229 3230 Level: beginner 3231 3232 Notes: 3233 The final time returned by this function may be different from the time of the internally 3234 held state accessible by TSGetSolution() and TSGetTime() because the method may have 3235 stepped over the final time. 3236 3237 .keywords: TS, timestep, solve 3238 3239 .seealso: TSCreate(), TSSetSolution(), TSStep() 3240 @*/ 3241 PetscErrorCode TSSolve(TS ts,Vec u) 3242 { 3243 Vec solution; 3244 PetscErrorCode ierr; 3245 3246 PetscFunctionBegin; 3247 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3248 if (u) PetscValidHeaderSpecific(u,VEC_CLASSID,2); 3249 if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE) { /* Need ts->vec_sol to be distinct so it is not overwritten when we interpolate at the end */ 3250 PetscValidHeaderSpecific(u,VEC_CLASSID,2); 3251 if (!ts->vec_sol || u == ts->vec_sol) { 3252 ierr = VecDuplicate(u,&solution);CHKERRQ(ierr); 3253 ierr = TSSetSolution(ts,solution);CHKERRQ(ierr); 3254 ierr = VecDestroy(&solution);CHKERRQ(ierr); /* grant ownership */ 3255 } 3256 ierr = VecCopy(u,ts->vec_sol);CHKERRQ(ierr); 3257 } else if (u) { 3258 ierr = TSSetSolution(ts,u);CHKERRQ(ierr); 3259 } 3260 ierr = TSSetUp(ts);CHKERRQ(ierr); 3261 /* reset time step and iteration counters */ 3262 ts->steps = 0; 3263 ts->ksp_its = 0; 3264 ts->snes_its = 0; 3265 ts->num_snes_failures = 0; 3266 ts->reject = 0; 3267 ts->reason = TS_CONVERGED_ITERATING; 3268 3269 ierr = TSViewFromOptions(ts,NULL,"-ts_view_pre");CHKERRQ(ierr); 3270 { 3271 DM dm; 3272 ierr = TSGetDM(ts, &dm);CHKERRQ(ierr); 3273 ierr = DMSetOutputSequenceNumber(dm, ts->steps, ts->ptime);CHKERRQ(ierr); 3274 } 3275 3276 if (ts->ops->solve) { /* This private interface is transitional and should be removed when all implementations are updated. */ 3277 ierr = (*ts->ops->solve)(ts);CHKERRQ(ierr); 3278 ierr = VecCopy(ts->vec_sol,u);CHKERRQ(ierr); 3279 ts->solvetime = ts->ptime; 3280 } else { 3281 /* steps the requested number of timesteps. */ 3282 if (ts->steps >= ts->max_steps) ts->reason = TS_CONVERGED_ITS; 3283 else if (ts->ptime >= ts->max_time) ts->reason = TS_CONVERGED_TIME; 3284 ierr = TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol);CHKERRQ(ierr); 3285 if(ts->event) { 3286 ierr = TSEventMonitorInitialize(ts);CHKERRQ(ierr); 3287 } 3288 while (!ts->reason) { 3289 ierr = TSMonitor(ts,ts->steps,ts->ptime,ts->vec_sol);CHKERRQ(ierr); 3290 ierr = TSStep(ts);CHKERRQ(ierr); 3291 if (ts->event) { 3292 ierr = TSEventMonitor(ts);CHKERRQ(ierr); 3293 } 3294 if(!ts->steprollback) { 3295 ierr = TSTrajectorySet(ts->trajectory,ts,ts->steps,ts->ptime,ts->vec_sol);CHKERRQ(ierr); 3296 ierr = TSPostStep(ts);CHKERRQ(ierr); 3297 } 3298 } 3299 if (ts->exact_final_time == TS_EXACTFINALTIME_INTERPOLATE && ts->ptime > ts->max_time) { 3300 ierr = TSInterpolate(ts,ts->max_time,u);CHKERRQ(ierr); 3301 ts->solvetime = ts->max_time; 3302 solution = u; 3303 } else { 3304 if (u) {ierr = VecCopy(ts->vec_sol,u);CHKERRQ(ierr);} 3305 ts->solvetime = ts->ptime; 3306 solution = ts->vec_sol; 3307 } 3308 ierr = TSMonitor(ts,ts->steps,ts->solvetime,solution);CHKERRQ(ierr); 3309 ierr = VecViewFromOptions(solution, ((PetscObject) ts)->prefix, "-ts_view_solution");CHKERRQ(ierr); 3310 } 3311 3312 ierr = TSViewFromOptions(ts,NULL,"-ts_view");CHKERRQ(ierr); 3313 ierr = PetscObjectSAWsBlock((PetscObject)ts);CHKERRQ(ierr); 3314 if (ts->vec_costintegral) ts->costintegralfwd=PETSC_TRUE; 3315 if (ts->adjoint_solve) { 3316 ierr = TSAdjointSolve(ts);CHKERRQ(ierr); 3317 } 3318 PetscFunctionReturn(0); 3319 } 3320 3321 #undef __FUNCT__ 3322 #define __FUNCT__ "TSAdjointSolve" 3323 /*@ 3324 TSAdjointSolve - Solves the discrete ajoint problem for an ODE/DAE 3325 3326 Collective on TS 3327 3328 Input Parameter: 3329 . ts - the TS context obtained from TSCreate() 3330 3331 Options Database: 3332 . -ts_adjoint_view_solution <viewerinfo> - views the first gradient with respect to the initial conditions 3333 3334 Level: intermediate 3335 3336 Notes: 3337 This must be called after a call to TSSolve() that solves the forward problem 3338 3339 By default this will integrate back to the initial time, one can use TSAdjointSetSteps() to step back to a later time 3340 3341 .keywords: TS, timestep, solve 3342 3343 .seealso: TSCreate(), TSSetSolution(), TSStep() 3344 @*/ 3345 PetscErrorCode TSAdjointSolve(TS ts) 3346 { 3347 PetscErrorCode ierr; 3348 3349 PetscFunctionBegin; 3350 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3351 ierr = TSAdjointSetUp(ts);CHKERRQ(ierr); 3352 /* reset time step and iteration counters */ 3353 ts->steps = 0; 3354 ts->ksp_its = 0; 3355 ts->snes_its = 0; 3356 ts->num_snes_failures = 0; 3357 ts->reject = 0; 3358 ts->reason = TS_CONVERGED_ITERATING; 3359 3360 if (!ts->adjoint_max_steps) ts->adjoint_max_steps = ts->total_steps; 3361 3362 if (ts->steps >= ts->adjoint_max_steps) ts->reason = TS_CONVERGED_ITS; 3363 while (!ts->reason) { 3364 ierr = TSTrajectoryGet(ts->trajectory,ts,ts->adjoint_max_steps-ts->steps,ts->ptime);CHKERRQ(ierr); 3365 ierr = TSMonitor(ts,ts->adjoint_max_steps-ts->steps,ts->ptime,ts->vec_sol);CHKERRQ(ierr); 3366 ierr = TSAdjointStep(ts);CHKERRQ(ierr); 3367 if (ts->event) { 3368 ierr = TSAdjointEventMonitor(ts);CHKERRQ(ierr); 3369 } 3370 3371 #if 0 /* I don't think PostStep is needed in AdjointSolve */ 3372 if (ts->event->status != TSEVENT_PROCESSING) { 3373 ierr = TSPostStep(ts);CHKERRQ(ierr); 3374 } 3375 } else { 3376 ierr = TSPostStep(ts);CHKERRQ(ierr); 3377 } 3378 #endif 3379 } 3380 ts->solvetime = ts->ptime; 3381 ierr = VecViewFromOptions(ts->vecs_sensi[0], ((PetscObject) ts)->prefix, "-ts_adjoint_view_solution");CHKERRQ(ierr); 3382 PetscFunctionReturn(0); 3383 } 3384 3385 #undef __FUNCT__ 3386 #define __FUNCT__ "TSMonitor" 3387 /*@ 3388 TSMonitor - Runs all user-provided monitor routines set using TSMonitorSet() 3389 3390 Collective on TS 3391 3392 Input Parameters: 3393 + ts - time stepping context obtained from TSCreate() 3394 . step - step number that has just completed 3395 . ptime - model time of the state 3396 - u - state at the current model time 3397 3398 Notes: 3399 TSMonitor() is typically used within the time stepping implementations. 3400 Users might call this function when using the TSStep() interface instead of TSSolve(). 3401 3402 Level: advanced 3403 3404 .keywords: TS, timestep 3405 @*/ 3406 PetscErrorCode TSMonitor(TS ts,PetscInt step,PetscReal ptime,Vec u) 3407 { 3408 PetscErrorCode ierr; 3409 PetscInt i,n = ts->numbermonitors; 3410 3411 PetscFunctionBegin; 3412 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3413 PetscValidHeaderSpecific(u,VEC_CLASSID,4); 3414 ierr = VecLockPush(u);CHKERRQ(ierr); 3415 for (i=0; i<n; i++) { 3416 ierr = (*ts->monitor[i])(ts,step,ptime,u,ts->monitorcontext[i]);CHKERRQ(ierr); 3417 } 3418 ierr = VecLockPop(u);CHKERRQ(ierr); 3419 PetscFunctionReturn(0); 3420 } 3421 3422 /* ------------------------------------------------------------------------*/ 3423 #undef __FUNCT__ 3424 #define __FUNCT__ "TSMonitorLGCtxCreate" 3425 /*@C 3426 TSMonitorLGCtxCreate - Creates a line graph context for use with 3427 TS to monitor the solution process graphically in various ways 3428 3429 Collective on TS 3430 3431 Input Parameters: 3432 + host - the X display to open, or null for the local machine 3433 . label - the title to put in the title bar 3434 . x, y - the screen coordinates of the upper left coordinate of the window 3435 . m, n - the screen width and height in pixels 3436 - howoften - if positive then determines the frequency of the plotting, if -1 then only at the final time 3437 3438 Output Parameter: 3439 . ctx - the context 3440 3441 Options Database Key: 3442 + -ts_monitor_lg_timestep - automatically sets line graph monitor 3443 . -ts_monitor_lg_solution - 3444 . -ts_monitor_lg_error - 3445 . -ts_monitor_lg_ksp_iterations - 3446 . -ts_monitor_lg_snes_iterations - 3447 - -lg_use_markers <true,false> - mark the data points (at each time step) on the plot; default is true 3448 3449 Notes: 3450 Use TSMonitorLGCtxDestroy() to destroy. 3451 3452 Level: intermediate 3453 3454 .keywords: TS, monitor, line graph, residual, seealso 3455 3456 .seealso: TSMonitorLGTimeStep(), TSMonitorSet(), TSMonitorLGSolution(), TSMonitorLGError() 3457 3458 @*/ 3459 PetscErrorCode TSMonitorLGCtxCreate(MPI_Comm comm,const char host[],const char label[],int x,int y,int m,int n,PetscInt howoften,TSMonitorLGCtx *ctx) 3460 { 3461 PetscDraw win; 3462 PetscErrorCode ierr; 3463 3464 PetscFunctionBegin; 3465 ierr = PetscNew(ctx);CHKERRQ(ierr); 3466 ierr = PetscDrawCreate(comm,host,label,x,y,m,n,&win);CHKERRQ(ierr); 3467 ierr = PetscDrawSetFromOptions(win);CHKERRQ(ierr); 3468 ierr = PetscDrawLGCreate(win,1,&(*ctx)->lg);CHKERRQ(ierr); 3469 ierr = PetscLogObjectParent((PetscObject)(*ctx)->lg,(PetscObject)win);CHKERRQ(ierr); 3470 ierr = PetscDrawLGSetUseMarkers((*ctx)->lg,PETSC_TRUE);CHKERRQ(ierr); 3471 ierr = PetscDrawLGSetFromOptions((*ctx)->lg);CHKERRQ(ierr); 3472 (*ctx)->howoften = howoften; 3473 PetscFunctionReturn(0); 3474 } 3475 3476 #undef __FUNCT__ 3477 #define __FUNCT__ "TSMonitorLGTimeStep" 3478 PetscErrorCode TSMonitorLGTimeStep(TS ts,PetscInt step,PetscReal ptime,Vec v,void *monctx) 3479 { 3480 TSMonitorLGCtx ctx = (TSMonitorLGCtx) monctx; 3481 PetscReal x = ptime,y; 3482 PetscErrorCode ierr; 3483 3484 PetscFunctionBegin; 3485 if (!step) { 3486 PetscDrawAxis axis; 3487 ierr = PetscDrawLGGetAxis(ctx->lg,&axis);CHKERRQ(ierr); 3488 ierr = PetscDrawAxisSetLabels(axis,"Timestep as function of time","Time","Time step");CHKERRQ(ierr); 3489 ierr = PetscDrawLGReset(ctx->lg);CHKERRQ(ierr); 3490 } 3491 ierr = TSGetTimeStep(ts,&y);CHKERRQ(ierr); 3492 ierr = PetscDrawLGAddPoint(ctx->lg,&x,&y);CHKERRQ(ierr); 3493 if (((ctx->howoften > 0) && (!(step % ctx->howoften))) || ((ctx->howoften == -1) && ts->reason)) { 3494 ierr = PetscDrawLGDraw(ctx->lg);CHKERRQ(ierr); 3495 } 3496 PetscFunctionReturn(0); 3497 } 3498 3499 #undef __FUNCT__ 3500 #define __FUNCT__ "TSMonitorLGCtxDestroy" 3501 /*@C 3502 TSMonitorLGCtxDestroy - Destroys a line graph context that was created 3503 with TSMonitorLGCtxCreate(). 3504 3505 Collective on TSMonitorLGCtx 3506 3507 Input Parameter: 3508 . ctx - the monitor context 3509 3510 Level: intermediate 3511 3512 .keywords: TS, monitor, line graph, destroy 3513 3514 .seealso: TSMonitorLGCtxCreate(), TSMonitorSet(), TSMonitorLGTimeStep(); 3515 @*/ 3516 PetscErrorCode TSMonitorLGCtxDestroy(TSMonitorLGCtx *ctx) 3517 { 3518 PetscDraw draw; 3519 PetscErrorCode ierr; 3520 3521 PetscFunctionBegin; 3522 if ((*ctx)->transformdestroy) { 3523 ierr = ((*ctx)->transformdestroy)((*ctx)->transformctx);CHKERRQ(ierr); 3524 } 3525 ierr = PetscDrawLGGetDraw((*ctx)->lg,&draw);CHKERRQ(ierr); 3526 ierr = PetscDrawDestroy(&draw);CHKERRQ(ierr); 3527 ierr = PetscDrawLGDestroy(&(*ctx)->lg);CHKERRQ(ierr); 3528 ierr = PetscStrArrayDestroy(&(*ctx)->names);CHKERRQ(ierr); 3529 ierr = PetscStrArrayDestroy(&(*ctx)->displaynames);CHKERRQ(ierr); 3530 ierr = PetscFree((*ctx)->displayvariables);CHKERRQ(ierr); 3531 ierr = PetscFree((*ctx)->displayvalues);CHKERRQ(ierr); 3532 ierr = PetscFree(*ctx);CHKERRQ(ierr); 3533 PetscFunctionReturn(0); 3534 } 3535 3536 #undef __FUNCT__ 3537 #define __FUNCT__ "TSGetTime" 3538 /*@ 3539 TSGetTime - Gets the time of the most recently completed step. 3540 3541 Not Collective 3542 3543 Input Parameter: 3544 . ts - the TS context obtained from TSCreate() 3545 3546 Output Parameter: 3547 . t - the current time 3548 3549 Level: beginner 3550 3551 Note: 3552 When called during time step evaluation (e.g. during residual evaluation or via hooks set using TSSetPreStep(), 3553 TSSetPreStage(), TSSetPostStage(), or TSSetPostStep()), the time is the time at the start of the step being evaluated. 3554 3555 .seealso: TSSetInitialTimeStep(), TSGetTimeStep() 3556 3557 .keywords: TS, get, time 3558 @*/ 3559 PetscErrorCode TSGetTime(TS ts,PetscReal *t) 3560 { 3561 PetscFunctionBegin; 3562 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3563 PetscValidRealPointer(t,2); 3564 *t = ts->ptime; 3565 PetscFunctionReturn(0); 3566 } 3567 3568 #undef __FUNCT__ 3569 #define __FUNCT__ "TSGetPrevTime" 3570 /*@ 3571 TSGetPrevTime - Gets the starting time of the previously completed step. 3572 3573 Not Collective 3574 3575 Input Parameter: 3576 . ts - the TS context obtained from TSCreate() 3577 3578 Output Parameter: 3579 . t - the previous time 3580 3581 Level: beginner 3582 3583 .seealso: TSSetInitialTimeStep(), TSGetTimeStep() 3584 3585 .keywords: TS, get, time 3586 @*/ 3587 PetscErrorCode TSGetPrevTime(TS ts,PetscReal *t) 3588 { 3589 PetscFunctionBegin; 3590 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3591 PetscValidRealPointer(t,2); 3592 *t = ts->ptime_prev; 3593 PetscFunctionReturn(0); 3594 } 3595 3596 #undef __FUNCT__ 3597 #define __FUNCT__ "TSSetTime" 3598 /*@ 3599 TSSetTime - Allows one to reset the time. 3600 3601 Logically Collective on TS 3602 3603 Input Parameters: 3604 + ts - the TS context obtained from TSCreate() 3605 - time - the time 3606 3607 Level: intermediate 3608 3609 .seealso: TSGetTime(), TSSetDuration() 3610 3611 .keywords: TS, set, time 3612 @*/ 3613 PetscErrorCode TSSetTime(TS ts, PetscReal t) 3614 { 3615 PetscFunctionBegin; 3616 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3617 PetscValidLogicalCollectiveReal(ts,t,2); 3618 ts->ptime = t; 3619 PetscFunctionReturn(0); 3620 } 3621 3622 #undef __FUNCT__ 3623 #define __FUNCT__ "TSSetOptionsPrefix" 3624 /*@C 3625 TSSetOptionsPrefix - Sets the prefix used for searching for all 3626 TS options in the database. 3627 3628 Logically Collective on TS 3629 3630 Input Parameter: 3631 + ts - The TS context 3632 - prefix - The prefix to prepend to all option names 3633 3634 Notes: 3635 A hyphen (-) must NOT be given at the beginning of the prefix name. 3636 The first character of all runtime options is AUTOMATICALLY the 3637 hyphen. 3638 3639 Level: advanced 3640 3641 .keywords: TS, set, options, prefix, database 3642 3643 .seealso: TSSetFromOptions() 3644 3645 @*/ 3646 PetscErrorCode TSSetOptionsPrefix(TS ts,const char prefix[]) 3647 { 3648 PetscErrorCode ierr; 3649 SNES snes; 3650 3651 PetscFunctionBegin; 3652 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3653 ierr = PetscObjectSetOptionsPrefix((PetscObject)ts,prefix);CHKERRQ(ierr); 3654 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 3655 ierr = SNESSetOptionsPrefix(snes,prefix);CHKERRQ(ierr); 3656 PetscFunctionReturn(0); 3657 } 3658 3659 3660 #undef __FUNCT__ 3661 #define __FUNCT__ "TSAppendOptionsPrefix" 3662 /*@C 3663 TSAppendOptionsPrefix - Appends to the prefix used for searching for all 3664 TS options in the database. 3665 3666 Logically Collective on TS 3667 3668 Input Parameter: 3669 + ts - The TS context 3670 - prefix - The prefix to prepend to all option names 3671 3672 Notes: 3673 A hyphen (-) must NOT be given at the beginning of the prefix name. 3674 The first character of all runtime options is AUTOMATICALLY the 3675 hyphen. 3676 3677 Level: advanced 3678 3679 .keywords: TS, append, options, prefix, database 3680 3681 .seealso: TSGetOptionsPrefix() 3682 3683 @*/ 3684 PetscErrorCode TSAppendOptionsPrefix(TS ts,const char prefix[]) 3685 { 3686 PetscErrorCode ierr; 3687 SNES snes; 3688 3689 PetscFunctionBegin; 3690 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3691 ierr = PetscObjectAppendOptionsPrefix((PetscObject)ts,prefix);CHKERRQ(ierr); 3692 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 3693 ierr = SNESAppendOptionsPrefix(snes,prefix);CHKERRQ(ierr); 3694 PetscFunctionReturn(0); 3695 } 3696 3697 #undef __FUNCT__ 3698 #define __FUNCT__ "TSGetOptionsPrefix" 3699 /*@C 3700 TSGetOptionsPrefix - Sets the prefix used for searching for all 3701 TS options in the database. 3702 3703 Not Collective 3704 3705 Input Parameter: 3706 . ts - The TS context 3707 3708 Output Parameter: 3709 . prefix - A pointer to the prefix string used 3710 3711 Notes: On the fortran side, the user should pass in a string 'prifix' of 3712 sufficient length to hold the prefix. 3713 3714 Level: intermediate 3715 3716 .keywords: TS, get, options, prefix, database 3717 3718 .seealso: TSAppendOptionsPrefix() 3719 @*/ 3720 PetscErrorCode TSGetOptionsPrefix(TS ts,const char *prefix[]) 3721 { 3722 PetscErrorCode ierr; 3723 3724 PetscFunctionBegin; 3725 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 3726 PetscValidPointer(prefix,2); 3727 ierr = PetscObjectGetOptionsPrefix((PetscObject)ts,prefix);CHKERRQ(ierr); 3728 PetscFunctionReturn(0); 3729 } 3730 3731 #undef __FUNCT__ 3732 #define __FUNCT__ "TSGetRHSJacobian" 3733 /*@C 3734 TSGetRHSJacobian - Returns the Jacobian J at the present timestep. 3735 3736 Not Collective, but parallel objects are returned if TS is parallel 3737 3738 Input Parameter: 3739 . ts - The TS context obtained from TSCreate() 3740 3741 Output Parameters: 3742 + Amat - The (approximate) Jacobian J of G, where U_t = G(U,t) (or NULL) 3743 . Pmat - The matrix from which the preconditioner is constructed, usually the same as Amat (or NULL) 3744 . func - Function to compute the Jacobian of the RHS (or NULL) 3745 - ctx - User-defined context for Jacobian evaluation routine (or NULL) 3746 3747 Notes: You can pass in NULL for any return argument you do not need. 3748 3749 Level: intermediate 3750 3751 .seealso: TSGetTimeStep(), TSGetMatrices(), TSGetTime(), TSGetTimeStepNumber() 3752 3753 .keywords: TS, timestep, get, matrix, Jacobian 3754 @*/ 3755 PetscErrorCode TSGetRHSJacobian(TS ts,Mat *Amat,Mat *Pmat,TSRHSJacobian *func,void **ctx) 3756 { 3757 PetscErrorCode ierr; 3758 SNES snes; 3759 DM dm; 3760 3761 PetscFunctionBegin; 3762 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 3763 ierr = SNESGetJacobian(snes,Amat,Pmat,NULL,NULL);CHKERRQ(ierr); 3764 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 3765 ierr = DMTSGetRHSJacobian(dm,func,ctx);CHKERRQ(ierr); 3766 PetscFunctionReturn(0); 3767 } 3768 3769 #undef __FUNCT__ 3770 #define __FUNCT__ "TSGetIJacobian" 3771 /*@C 3772 TSGetIJacobian - Returns the implicit Jacobian at the present timestep. 3773 3774 Not Collective, but parallel objects are returned if TS is parallel 3775 3776 Input Parameter: 3777 . ts - The TS context obtained from TSCreate() 3778 3779 Output Parameters: 3780 + Amat - The (approximate) Jacobian of F(t,U,U_t) 3781 . Pmat - The matrix from which the preconditioner is constructed, often the same as Amat 3782 . f - The function to compute the matrices 3783 - ctx - User-defined context for Jacobian evaluation routine 3784 3785 Notes: You can pass in NULL for any return argument you do not need. 3786 3787 Level: advanced 3788 3789 .seealso: TSGetTimeStep(), TSGetRHSJacobian(), TSGetMatrices(), TSGetTime(), TSGetTimeStepNumber() 3790 3791 .keywords: TS, timestep, get, matrix, Jacobian 3792 @*/ 3793 PetscErrorCode TSGetIJacobian(TS ts,Mat *Amat,Mat *Pmat,TSIJacobian *f,void **ctx) 3794 { 3795 PetscErrorCode ierr; 3796 SNES snes; 3797 DM dm; 3798 3799 PetscFunctionBegin; 3800 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 3801 ierr = SNESSetUpMatrices(snes);CHKERRQ(ierr); 3802 ierr = SNESGetJacobian(snes,Amat,Pmat,NULL,NULL);CHKERRQ(ierr); 3803 ierr = TSGetDM(ts,&dm);CHKERRQ(ierr); 3804 ierr = DMTSGetIJacobian(dm,f,ctx);CHKERRQ(ierr); 3805 PetscFunctionReturn(0); 3806 } 3807 3808 3809 #undef __FUNCT__ 3810 #define __FUNCT__ "TSMonitorDrawSolution" 3811 /*@C 3812 TSMonitorDrawSolution - Monitors progress of the TS solvers by calling 3813 VecView() for the solution at each timestep 3814 3815 Collective on TS 3816 3817 Input Parameters: 3818 + ts - the TS context 3819 . step - current time-step 3820 . ptime - current time 3821 - dummy - either a viewer or NULL 3822 3823 Options Database: 3824 . -ts_monitor_draw_solution_initial - show initial solution as well as current solution 3825 3826 Notes: the initial solution and current solution are not display with a common axis scaling so generally the option -ts_monitor_draw_solution_initial 3827 will look bad 3828 3829 Level: intermediate 3830 3831 .keywords: TS, vector, monitor, view 3832 3833 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 3834 @*/ 3835 PetscErrorCode TSMonitorDrawSolution(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 3836 { 3837 PetscErrorCode ierr; 3838 TSMonitorDrawCtx ictx = (TSMonitorDrawCtx)dummy; 3839 PetscDraw draw; 3840 3841 PetscFunctionBegin; 3842 if (!step && ictx->showinitial) { 3843 if (!ictx->initialsolution) { 3844 ierr = VecDuplicate(u,&ictx->initialsolution);CHKERRQ(ierr); 3845 } 3846 ierr = VecCopy(u,ictx->initialsolution);CHKERRQ(ierr); 3847 } 3848 if (!(((ictx->howoften > 0) && (!(step % ictx->howoften))) || ((ictx->howoften == -1) && ts->reason))) PetscFunctionReturn(0); 3849 3850 if (ictx->showinitial) { 3851 PetscReal pause; 3852 ierr = PetscViewerDrawGetPause(ictx->viewer,&pause);CHKERRQ(ierr); 3853 ierr = PetscViewerDrawSetPause(ictx->viewer,0.0);CHKERRQ(ierr); 3854 ierr = VecView(ictx->initialsolution,ictx->viewer);CHKERRQ(ierr); 3855 ierr = PetscViewerDrawSetPause(ictx->viewer,pause);CHKERRQ(ierr); 3856 ierr = PetscViewerDrawSetHold(ictx->viewer,PETSC_TRUE);CHKERRQ(ierr); 3857 } 3858 ierr = VecView(u,ictx->viewer);CHKERRQ(ierr); 3859 if (ictx->showtimestepandtime) { 3860 PetscReal xl,yl,xr,yr,h; 3861 char time[32]; 3862 3863 ierr = PetscViewerDrawGetDraw(ictx->viewer,0,&draw);CHKERRQ(ierr); 3864 ierr = PetscSNPrintf(time,32,"Timestep %d Time %g",(int)step,(double)ptime);CHKERRQ(ierr); 3865 ierr = PetscDrawGetCoordinates(draw,&xl,&yl,&xr,&yr);CHKERRQ(ierr); 3866 h = yl + .95*(yr - yl); 3867 ierr = PetscDrawStringCentered(draw,.5*(xl+xr),h,PETSC_DRAW_BLACK,time);CHKERRQ(ierr); 3868 ierr = PetscDrawFlush(draw);CHKERRQ(ierr); 3869 } 3870 3871 if (ictx->showinitial) { 3872 ierr = PetscViewerDrawSetHold(ictx->viewer,PETSC_FALSE);CHKERRQ(ierr); 3873 } 3874 PetscFunctionReturn(0); 3875 } 3876 3877 #undef __FUNCT__ 3878 #define __FUNCT__ "TSMonitorDrawSolutionPhase" 3879 /*@C 3880 TSMonitorDrawSolutionPhase - Monitors progress of the TS solvers by plotting the solution as a phase diagram 3881 3882 Collective on TS 3883 3884 Input Parameters: 3885 + ts - the TS context 3886 . step - current time-step 3887 . ptime - current time 3888 - dummy - either a viewer or NULL 3889 3890 Level: intermediate 3891 3892 .keywords: TS, vector, monitor, view 3893 3894 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 3895 @*/ 3896 PetscErrorCode TSMonitorDrawSolutionPhase(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 3897 { 3898 PetscErrorCode ierr; 3899 TSMonitorDrawCtx ictx = (TSMonitorDrawCtx)dummy; 3900 PetscDraw draw; 3901 MPI_Comm comm; 3902 PetscInt n; 3903 PetscMPIInt size; 3904 PetscReal xl,yl,xr,yr,h; 3905 char time[32]; 3906 const PetscScalar *U; 3907 3908 PetscFunctionBegin; 3909 ierr = PetscObjectGetComm((PetscObject)ts,&comm);CHKERRQ(ierr); 3910 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 3911 if (size != 1) SETERRQ(comm,PETSC_ERR_SUP,"Only allowed for sequential runs"); 3912 ierr = VecGetSize(u,&n);CHKERRQ(ierr); 3913 if (n != 2) SETERRQ(comm,PETSC_ERR_SUP,"Only for ODEs with two unknowns"); 3914 3915 ierr = PetscViewerDrawGetDraw(ictx->viewer,0,&draw);CHKERRQ(ierr); 3916 3917 ierr = VecGetArrayRead(u,&U);CHKERRQ(ierr); 3918 ierr = PetscDrawAxisGetLimits(ictx->axis,&xl,&xr,&yl,&yr);CHKERRQ(ierr); 3919 if ((PetscRealPart(U[0]) < xl) || (PetscRealPart(U[1]) < yl) || (PetscRealPart(U[0]) > xr) || (PetscRealPart(U[1]) > yr)) { 3920 ierr = VecRestoreArrayRead(u,&U);CHKERRQ(ierr); 3921 PetscFunctionReturn(0); 3922 } 3923 if (!step) ictx->color++; 3924 ierr = PetscDrawPoint(draw,PetscRealPart(U[0]),PetscRealPart(U[1]),ictx->color);CHKERRQ(ierr); 3925 ierr = VecRestoreArrayRead(u,&U);CHKERRQ(ierr); 3926 3927 if (ictx->showtimestepandtime) { 3928 ierr = PetscDrawGetCoordinates(draw,&xl,&yl,&xr,&yr);CHKERRQ(ierr); 3929 ierr = PetscSNPrintf(time,32,"Timestep %d Time %g",(int)step,(double)ptime);CHKERRQ(ierr); 3930 h = yl + .95*(yr - yl); 3931 ierr = PetscDrawStringCentered(draw,.5*(xl+xr),h,PETSC_DRAW_BLACK,time);CHKERRQ(ierr); 3932 } 3933 ierr = PetscDrawFlush(draw);CHKERRQ(ierr); 3934 PetscFunctionReturn(0); 3935 } 3936 3937 3938 #undef __FUNCT__ 3939 #define __FUNCT__ "TSMonitorDrawCtxDestroy" 3940 /*@C 3941 TSMonitorDrawCtxDestroy - Destroys the monitor context for TSMonitorDrawSolution() 3942 3943 Collective on TS 3944 3945 Input Parameters: 3946 . ctx - the monitor context 3947 3948 Level: intermediate 3949 3950 .keywords: TS, vector, monitor, view 3951 3952 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorDrawSolution(), TSMonitorDrawError() 3953 @*/ 3954 PetscErrorCode TSMonitorDrawCtxDestroy(TSMonitorDrawCtx *ictx) 3955 { 3956 PetscErrorCode ierr; 3957 3958 PetscFunctionBegin; 3959 ierr = PetscDrawAxisDestroy(&(*ictx)->axis);CHKERRQ(ierr); 3960 ierr = PetscViewerDestroy(&(*ictx)->viewer);CHKERRQ(ierr); 3961 ierr = VecDestroy(&(*ictx)->initialsolution);CHKERRQ(ierr); 3962 ierr = PetscFree(*ictx);CHKERRQ(ierr); 3963 PetscFunctionReturn(0); 3964 } 3965 3966 #undef __FUNCT__ 3967 #define __FUNCT__ "TSMonitorDrawCtxCreate" 3968 /*@C 3969 TSMonitorDrawCtxCreate - Creates the monitor context for TSMonitorDrawCtx 3970 3971 Collective on TS 3972 3973 Input Parameter: 3974 . ts - time-step context 3975 3976 Output Patameter: 3977 . ctx - the monitor context 3978 3979 Options Database: 3980 . -ts_monitor_draw_solution_initial - show initial solution as well as current solution 3981 3982 Level: intermediate 3983 3984 .keywords: TS, vector, monitor, view 3985 3986 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorDrawCtx() 3987 @*/ 3988 PetscErrorCode TSMonitorDrawCtxCreate(MPI_Comm comm,const char host[],const char label[],int x,int y,int m,int n,PetscInt howoften,TSMonitorDrawCtx *ctx) 3989 { 3990 PetscErrorCode ierr; 3991 3992 PetscFunctionBegin; 3993 ierr = PetscNew(ctx);CHKERRQ(ierr); 3994 ierr = PetscViewerDrawOpen(comm,host,label,x,y,m,n,&(*ctx)->viewer);CHKERRQ(ierr); 3995 ierr = PetscViewerSetFromOptions((*ctx)->viewer);CHKERRQ(ierr); 3996 3997 (*ctx)->howoften = howoften; 3998 (*ctx)->showinitial = PETSC_FALSE; 3999 ierr = PetscOptionsGetBool(NULL,"-ts_monitor_draw_solution_initial",&(*ctx)->showinitial,NULL);CHKERRQ(ierr); 4000 4001 (*ctx)->showtimestepandtime = PETSC_FALSE; 4002 ierr = PetscOptionsGetBool(NULL,"-ts_monitor_draw_solution_show_time",&(*ctx)->showtimestepandtime,NULL);CHKERRQ(ierr); 4003 (*ctx)->color = PETSC_DRAW_WHITE; 4004 PetscFunctionReturn(0); 4005 } 4006 4007 #undef __FUNCT__ 4008 #define __FUNCT__ "TSMonitorDrawError" 4009 /*@C 4010 TSMonitorDrawError - Monitors progress of the TS solvers by calling 4011 VecView() for the error at each timestep 4012 4013 Collective on TS 4014 4015 Input Parameters: 4016 + ts - the TS context 4017 . step - current time-step 4018 . ptime - current time 4019 - dummy - either a viewer or NULL 4020 4021 Level: intermediate 4022 4023 .keywords: TS, vector, monitor, view 4024 4025 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 4026 @*/ 4027 PetscErrorCode TSMonitorDrawError(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 4028 { 4029 PetscErrorCode ierr; 4030 TSMonitorDrawCtx ctx = (TSMonitorDrawCtx)dummy; 4031 PetscViewer viewer = ctx->viewer; 4032 Vec work; 4033 4034 PetscFunctionBegin; 4035 if (!(((ctx->howoften > 0) && (!(step % ctx->howoften))) || ((ctx->howoften == -1) && ts->reason))) PetscFunctionReturn(0); 4036 ierr = VecDuplicate(u,&work);CHKERRQ(ierr); 4037 ierr = TSComputeSolutionFunction(ts,ptime,work);CHKERRQ(ierr); 4038 ierr = VecAXPY(work,-1.0,u);CHKERRQ(ierr); 4039 ierr = VecView(work,viewer);CHKERRQ(ierr); 4040 ierr = VecDestroy(&work);CHKERRQ(ierr); 4041 PetscFunctionReturn(0); 4042 } 4043 4044 #include <petsc/private/dmimpl.h> 4045 #undef __FUNCT__ 4046 #define __FUNCT__ "TSSetDM" 4047 /*@ 4048 TSSetDM - Sets the DM that may be used by some preconditioners 4049 4050 Logically Collective on TS and DM 4051 4052 Input Parameters: 4053 + ts - the preconditioner context 4054 - dm - the dm 4055 4056 Level: intermediate 4057 4058 4059 .seealso: TSGetDM(), SNESSetDM(), SNESGetDM() 4060 @*/ 4061 PetscErrorCode TSSetDM(TS ts,DM dm) 4062 { 4063 PetscErrorCode ierr; 4064 SNES snes; 4065 DMTS tsdm; 4066 4067 PetscFunctionBegin; 4068 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4069 ierr = PetscObjectReference((PetscObject)dm);CHKERRQ(ierr); 4070 if (ts->dm) { /* Move the DMTS context over to the new DM unless the new DM already has one */ 4071 if (ts->dm->dmts && !dm->dmts) { 4072 ierr = DMCopyDMTS(ts->dm,dm);CHKERRQ(ierr); 4073 ierr = DMGetDMTS(ts->dm,&tsdm);CHKERRQ(ierr); 4074 if (tsdm->originaldm == ts->dm) { /* Grant write privileges to the replacement DM */ 4075 tsdm->originaldm = dm; 4076 } 4077 } 4078 ierr = DMDestroy(&ts->dm);CHKERRQ(ierr); 4079 } 4080 ts->dm = dm; 4081 4082 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 4083 ierr = SNESSetDM(snes,dm);CHKERRQ(ierr); 4084 PetscFunctionReturn(0); 4085 } 4086 4087 #undef __FUNCT__ 4088 #define __FUNCT__ "TSGetDM" 4089 /*@ 4090 TSGetDM - Gets the DM that may be used by some preconditioners 4091 4092 Not Collective 4093 4094 Input Parameter: 4095 . ts - the preconditioner context 4096 4097 Output Parameter: 4098 . dm - the dm 4099 4100 Level: intermediate 4101 4102 4103 .seealso: TSSetDM(), SNESSetDM(), SNESGetDM() 4104 @*/ 4105 PetscErrorCode TSGetDM(TS ts,DM *dm) 4106 { 4107 PetscErrorCode ierr; 4108 4109 PetscFunctionBegin; 4110 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4111 if (!ts->dm) { 4112 ierr = DMShellCreate(PetscObjectComm((PetscObject)ts),&ts->dm);CHKERRQ(ierr); 4113 if (ts->snes) {ierr = SNESSetDM(ts->snes,ts->dm);CHKERRQ(ierr);} 4114 } 4115 *dm = ts->dm; 4116 PetscFunctionReturn(0); 4117 } 4118 4119 #undef __FUNCT__ 4120 #define __FUNCT__ "SNESTSFormFunction" 4121 /*@ 4122 SNESTSFormFunction - Function to evaluate nonlinear residual 4123 4124 Logically Collective on SNES 4125 4126 Input Parameter: 4127 + snes - nonlinear solver 4128 . U - the current state at which to evaluate the residual 4129 - ctx - user context, must be a TS 4130 4131 Output Parameter: 4132 . F - the nonlinear residual 4133 4134 Notes: 4135 This function is not normally called by users and is automatically registered with the SNES used by TS. 4136 It is most frequently passed to MatFDColoringSetFunction(). 4137 4138 Level: advanced 4139 4140 .seealso: SNESSetFunction(), MatFDColoringSetFunction() 4141 @*/ 4142 PetscErrorCode SNESTSFormFunction(SNES snes,Vec U,Vec F,void *ctx) 4143 { 4144 TS ts = (TS)ctx; 4145 PetscErrorCode ierr; 4146 4147 PetscFunctionBegin; 4148 PetscValidHeaderSpecific(snes,SNES_CLASSID,1); 4149 PetscValidHeaderSpecific(U,VEC_CLASSID,2); 4150 PetscValidHeaderSpecific(F,VEC_CLASSID,3); 4151 PetscValidHeaderSpecific(ts,TS_CLASSID,4); 4152 ierr = (ts->ops->snesfunction)(snes,U,F,ts);CHKERRQ(ierr); 4153 PetscFunctionReturn(0); 4154 } 4155 4156 #undef __FUNCT__ 4157 #define __FUNCT__ "SNESTSFormJacobian" 4158 /*@ 4159 SNESTSFormJacobian - Function to evaluate the Jacobian 4160 4161 Collective on SNES 4162 4163 Input Parameter: 4164 + snes - nonlinear solver 4165 . U - the current state at which to evaluate the residual 4166 - ctx - user context, must be a TS 4167 4168 Output Parameter: 4169 + A - the Jacobian 4170 . B - the preconditioning matrix (may be the same as A) 4171 - flag - indicates any structure change in the matrix 4172 4173 Notes: 4174 This function is not normally called by users and is automatically registered with the SNES used by TS. 4175 4176 Level: developer 4177 4178 .seealso: SNESSetJacobian() 4179 @*/ 4180 PetscErrorCode SNESTSFormJacobian(SNES snes,Vec U,Mat A,Mat B,void *ctx) 4181 { 4182 TS ts = (TS)ctx; 4183 PetscErrorCode ierr; 4184 4185 PetscFunctionBegin; 4186 PetscValidHeaderSpecific(snes,SNES_CLASSID,1); 4187 PetscValidHeaderSpecific(U,VEC_CLASSID,2); 4188 PetscValidPointer(A,3); 4189 PetscValidHeaderSpecific(A,MAT_CLASSID,3); 4190 PetscValidPointer(B,4); 4191 PetscValidHeaderSpecific(B,MAT_CLASSID,4); 4192 PetscValidHeaderSpecific(ts,TS_CLASSID,6); 4193 ierr = (ts->ops->snesjacobian)(snes,U,A,B,ts);CHKERRQ(ierr); 4194 PetscFunctionReturn(0); 4195 } 4196 4197 #undef __FUNCT__ 4198 #define __FUNCT__ "TSComputeRHSFunctionLinear" 4199 /*@C 4200 TSComputeRHSFunctionLinear - Evaluate the right hand side via the user-provided Jacobian, for linear problems only 4201 4202 Collective on TS 4203 4204 Input Arguments: 4205 + ts - time stepping context 4206 . t - time at which to evaluate 4207 . U - state at which to evaluate 4208 - ctx - context 4209 4210 Output Arguments: 4211 . F - right hand side 4212 4213 Level: intermediate 4214 4215 Notes: 4216 This function is intended to be passed to TSSetRHSFunction() to evaluate the right hand side for linear problems. 4217 The matrix (and optionally the evaluation context) should be passed to TSSetRHSJacobian(). 4218 4219 .seealso: TSSetRHSFunction(), TSSetRHSJacobian(), TSComputeRHSJacobianConstant() 4220 @*/ 4221 PetscErrorCode TSComputeRHSFunctionLinear(TS ts,PetscReal t,Vec U,Vec F,void *ctx) 4222 { 4223 PetscErrorCode ierr; 4224 Mat Arhs,Brhs; 4225 4226 PetscFunctionBegin; 4227 ierr = TSGetRHSMats_Private(ts,&Arhs,&Brhs);CHKERRQ(ierr); 4228 ierr = TSComputeRHSJacobian(ts,t,U,Arhs,Brhs);CHKERRQ(ierr); 4229 ierr = MatMult(Arhs,U,F);CHKERRQ(ierr); 4230 PetscFunctionReturn(0); 4231 } 4232 4233 #undef __FUNCT__ 4234 #define __FUNCT__ "TSComputeRHSJacobianConstant" 4235 /*@C 4236 TSComputeRHSJacobianConstant - Reuses a Jacobian that is time-independent. 4237 4238 Collective on TS 4239 4240 Input Arguments: 4241 + ts - time stepping context 4242 . t - time at which to evaluate 4243 . U - state at which to evaluate 4244 - ctx - context 4245 4246 Output Arguments: 4247 + A - pointer to operator 4248 . B - pointer to preconditioning matrix 4249 - flg - matrix structure flag 4250 4251 Level: intermediate 4252 4253 Notes: 4254 This function is intended to be passed to TSSetRHSJacobian() to evaluate the Jacobian for linear time-independent problems. 4255 4256 .seealso: TSSetRHSFunction(), TSSetRHSJacobian(), TSComputeRHSFunctionLinear() 4257 @*/ 4258 PetscErrorCode TSComputeRHSJacobianConstant(TS ts,PetscReal t,Vec U,Mat A,Mat B,void *ctx) 4259 { 4260 PetscFunctionBegin; 4261 PetscFunctionReturn(0); 4262 } 4263 4264 #undef __FUNCT__ 4265 #define __FUNCT__ "TSComputeIFunctionLinear" 4266 /*@C 4267 TSComputeIFunctionLinear - Evaluate the left hand side via the user-provided Jacobian, for linear problems only 4268 4269 Collective on TS 4270 4271 Input Arguments: 4272 + ts - time stepping context 4273 . t - time at which to evaluate 4274 . U - state at which to evaluate 4275 . Udot - time derivative of state vector 4276 - ctx - context 4277 4278 Output Arguments: 4279 . F - left hand side 4280 4281 Level: intermediate 4282 4283 Notes: 4284 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 4285 user is required to write their own TSComputeIFunction. 4286 This function is intended to be passed to TSSetIFunction() to evaluate the left hand side for linear problems. 4287 The matrix (and optionally the evaluation context) should be passed to TSSetIJacobian(). 4288 4289 .seealso: TSSetIFunction(), TSSetIJacobian(), TSComputeIJacobianConstant() 4290 @*/ 4291 PetscErrorCode TSComputeIFunctionLinear(TS ts,PetscReal t,Vec U,Vec Udot,Vec F,void *ctx) 4292 { 4293 PetscErrorCode ierr; 4294 Mat A,B; 4295 4296 PetscFunctionBegin; 4297 ierr = TSGetIJacobian(ts,&A,&B,NULL,NULL);CHKERRQ(ierr); 4298 ierr = TSComputeIJacobian(ts,t,U,Udot,1.0,A,B,PETSC_TRUE);CHKERRQ(ierr); 4299 ierr = MatMult(A,Udot,F);CHKERRQ(ierr); 4300 PetscFunctionReturn(0); 4301 } 4302 4303 #undef __FUNCT__ 4304 #define __FUNCT__ "TSComputeIJacobianConstant" 4305 /*@C 4306 TSComputeIJacobianConstant - Reuses a time-independent for a semi-implicit DAE or ODE 4307 4308 Collective on TS 4309 4310 Input Arguments: 4311 + ts - time stepping context 4312 . t - time at which to evaluate 4313 . U - state at which to evaluate 4314 . Udot - time derivative of state vector 4315 . shift - shift to apply 4316 - ctx - context 4317 4318 Output Arguments: 4319 + A - pointer to operator 4320 . B - pointer to preconditioning matrix 4321 - flg - matrix structure flag 4322 4323 Level: advanced 4324 4325 Notes: 4326 This function is intended to be passed to TSSetIJacobian() to evaluate the Jacobian for linear time-independent problems. 4327 4328 It is only appropriate for problems of the form 4329 4330 $ M Udot = F(U,t) 4331 4332 where M is constant and F is non-stiff. The user must pass M to TSSetIJacobian(). The current implementation only 4333 works with IMEX time integration methods such as TSROSW and TSARKIMEX, since there is no support for de-constructing 4334 an implicit operator of the form 4335 4336 $ shift*M + J 4337 4338 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 4339 a copy of M or reassemble it when requested. 4340 4341 .seealso: TSSetIFunction(), TSSetIJacobian(), TSComputeIFunctionLinear() 4342 @*/ 4343 PetscErrorCode TSComputeIJacobianConstant(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat A,Mat B,void *ctx) 4344 { 4345 PetscErrorCode ierr; 4346 4347 PetscFunctionBegin; 4348 ierr = MatScale(A, shift / ts->ijacobian.shift);CHKERRQ(ierr); 4349 ts->ijacobian.shift = shift; 4350 PetscFunctionReturn(0); 4351 } 4352 4353 #undef __FUNCT__ 4354 #define __FUNCT__ "TSGetEquationType" 4355 /*@ 4356 TSGetEquationType - Gets the type of the equation that TS is solving. 4357 4358 Not Collective 4359 4360 Input Parameter: 4361 . ts - the TS context 4362 4363 Output Parameter: 4364 . equation_type - see TSEquationType 4365 4366 Level: beginner 4367 4368 .keywords: TS, equation type 4369 4370 .seealso: TSSetEquationType(), TSEquationType 4371 @*/ 4372 PetscErrorCode TSGetEquationType(TS ts,TSEquationType *equation_type) 4373 { 4374 PetscFunctionBegin; 4375 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4376 PetscValidPointer(equation_type,2); 4377 *equation_type = ts->equation_type; 4378 PetscFunctionReturn(0); 4379 } 4380 4381 #undef __FUNCT__ 4382 #define __FUNCT__ "TSSetEquationType" 4383 /*@ 4384 TSSetEquationType - Sets the type of the equation that TS is solving. 4385 4386 Not Collective 4387 4388 Input Parameter: 4389 + ts - the TS context 4390 . equation_type - see TSEquationType 4391 4392 Level: advanced 4393 4394 .keywords: TS, equation type 4395 4396 .seealso: TSGetEquationType(), TSEquationType 4397 @*/ 4398 PetscErrorCode TSSetEquationType(TS ts,TSEquationType equation_type) 4399 { 4400 PetscFunctionBegin; 4401 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4402 ts->equation_type = equation_type; 4403 PetscFunctionReturn(0); 4404 } 4405 4406 #undef __FUNCT__ 4407 #define __FUNCT__ "TSGetConvergedReason" 4408 /*@ 4409 TSGetConvergedReason - Gets the reason the TS iteration was stopped. 4410 4411 Not Collective 4412 4413 Input Parameter: 4414 . ts - the TS context 4415 4416 Output Parameter: 4417 . reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the 4418 manual pages for the individual convergence tests for complete lists 4419 4420 Level: beginner 4421 4422 Notes: 4423 Can only be called after the call to TSSolve() is complete. 4424 4425 .keywords: TS, nonlinear, set, convergence, test 4426 4427 .seealso: TSSetConvergenceTest(), TSConvergedReason 4428 @*/ 4429 PetscErrorCode TSGetConvergedReason(TS ts,TSConvergedReason *reason) 4430 { 4431 PetscFunctionBegin; 4432 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4433 PetscValidPointer(reason,2); 4434 *reason = ts->reason; 4435 PetscFunctionReturn(0); 4436 } 4437 4438 #undef __FUNCT__ 4439 #define __FUNCT__ "TSSetConvergedReason" 4440 /*@ 4441 TSSetConvergedReason - Sets the reason for handling the convergence of TSSolve. 4442 4443 Not Collective 4444 4445 Input Parameter: 4446 + ts - the TS context 4447 . reason - negative value indicates diverged, positive value converged, see TSConvergedReason or the 4448 manual pages for the individual convergence tests for complete lists 4449 4450 Level: advanced 4451 4452 Notes: 4453 Can only be called during TSSolve() is active. 4454 4455 .keywords: TS, nonlinear, set, convergence, test 4456 4457 .seealso: TSConvergedReason 4458 @*/ 4459 PetscErrorCode TSSetConvergedReason(TS ts,TSConvergedReason reason) 4460 { 4461 PetscFunctionBegin; 4462 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4463 ts->reason = reason; 4464 PetscFunctionReturn(0); 4465 } 4466 4467 #undef __FUNCT__ 4468 #define __FUNCT__ "TSGetSolveTime" 4469 /*@ 4470 TSGetSolveTime - Gets the time after a call to TSSolve() 4471 4472 Not Collective 4473 4474 Input Parameter: 4475 . ts - the TS context 4476 4477 Output Parameter: 4478 . ftime - the final time. This time should correspond to the final time set with TSSetDuration() 4479 4480 Level: beginner 4481 4482 Notes: 4483 Can only be called after the call to TSSolve() is complete. 4484 4485 .keywords: TS, nonlinear, set, convergence, test 4486 4487 .seealso: TSSetConvergenceTest(), TSConvergedReason 4488 @*/ 4489 PetscErrorCode TSGetSolveTime(TS ts,PetscReal *ftime) 4490 { 4491 PetscFunctionBegin; 4492 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4493 PetscValidPointer(ftime,2); 4494 *ftime = ts->solvetime; 4495 PetscFunctionReturn(0); 4496 } 4497 4498 #undef __FUNCT__ 4499 #define __FUNCT__ "TSGetTotalSteps" 4500 /*@ 4501 TSGetTotalSteps - Gets the total number of steps done since the last call to TSSetUp() or TSCreate() 4502 4503 Not Collective 4504 4505 Input Parameter: 4506 . ts - the TS context 4507 4508 Output Parameter: 4509 . steps - the number of steps 4510 4511 Level: beginner 4512 4513 Notes: 4514 Includes the number of steps for all calls to TSSolve() since TSSetUp() was called 4515 4516 .keywords: TS, nonlinear, set, convergence, test 4517 4518 .seealso: TSSetConvergenceTest(), TSConvergedReason 4519 @*/ 4520 PetscErrorCode TSGetTotalSteps(TS ts,PetscInt *steps) 4521 { 4522 PetscFunctionBegin; 4523 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4524 PetscValidPointer(steps,2); 4525 *steps = ts->total_steps; 4526 PetscFunctionReturn(0); 4527 } 4528 4529 #undef __FUNCT__ 4530 #define __FUNCT__ "TSGetSNESIterations" 4531 /*@ 4532 TSGetSNESIterations - Gets the total number of nonlinear iterations 4533 used by the time integrator. 4534 4535 Not Collective 4536 4537 Input Parameter: 4538 . ts - TS context 4539 4540 Output Parameter: 4541 . nits - number of nonlinear iterations 4542 4543 Notes: 4544 This counter is reset to zero for each successive call to TSSolve(). 4545 4546 Level: intermediate 4547 4548 .keywords: TS, get, number, nonlinear, iterations 4549 4550 .seealso: TSGetKSPIterations() 4551 @*/ 4552 PetscErrorCode TSGetSNESIterations(TS ts,PetscInt *nits) 4553 { 4554 PetscFunctionBegin; 4555 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4556 PetscValidIntPointer(nits,2); 4557 *nits = ts->snes_its; 4558 PetscFunctionReturn(0); 4559 } 4560 4561 #undef __FUNCT__ 4562 #define __FUNCT__ "TSGetKSPIterations" 4563 /*@ 4564 TSGetKSPIterations - Gets the total number of linear iterations 4565 used by the time integrator. 4566 4567 Not Collective 4568 4569 Input Parameter: 4570 . ts - TS context 4571 4572 Output Parameter: 4573 . lits - number of linear iterations 4574 4575 Notes: 4576 This counter is reset to zero for each successive call to TSSolve(). 4577 4578 Level: intermediate 4579 4580 .keywords: TS, get, number, linear, iterations 4581 4582 .seealso: TSGetSNESIterations(), SNESGetKSPIterations() 4583 @*/ 4584 PetscErrorCode TSGetKSPIterations(TS ts,PetscInt *lits) 4585 { 4586 PetscFunctionBegin; 4587 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4588 PetscValidIntPointer(lits,2); 4589 *lits = ts->ksp_its; 4590 PetscFunctionReturn(0); 4591 } 4592 4593 #undef __FUNCT__ 4594 #define __FUNCT__ "TSGetStepRejections" 4595 /*@ 4596 TSGetStepRejections - Gets the total number of rejected steps. 4597 4598 Not Collective 4599 4600 Input Parameter: 4601 . ts - TS context 4602 4603 Output Parameter: 4604 . rejects - number of steps rejected 4605 4606 Notes: 4607 This counter is reset to zero for each successive call to TSSolve(). 4608 4609 Level: intermediate 4610 4611 .keywords: TS, get, number 4612 4613 .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetSNESFailures(), TSSetMaxSNESFailures(), TSSetErrorIfStepFails() 4614 @*/ 4615 PetscErrorCode TSGetStepRejections(TS ts,PetscInt *rejects) 4616 { 4617 PetscFunctionBegin; 4618 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4619 PetscValidIntPointer(rejects,2); 4620 *rejects = ts->reject; 4621 PetscFunctionReturn(0); 4622 } 4623 4624 #undef __FUNCT__ 4625 #define __FUNCT__ "TSGetSNESFailures" 4626 /*@ 4627 TSGetSNESFailures - Gets the total number of failed SNES solves 4628 4629 Not Collective 4630 4631 Input Parameter: 4632 . ts - TS context 4633 4634 Output Parameter: 4635 . fails - number of failed nonlinear solves 4636 4637 Notes: 4638 This counter is reset to zero for each successive call to TSSolve(). 4639 4640 Level: intermediate 4641 4642 .keywords: TS, get, number 4643 4644 .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSSetMaxSNESFailures() 4645 @*/ 4646 PetscErrorCode TSGetSNESFailures(TS ts,PetscInt *fails) 4647 { 4648 PetscFunctionBegin; 4649 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4650 PetscValidIntPointer(fails,2); 4651 *fails = ts->num_snes_failures; 4652 PetscFunctionReturn(0); 4653 } 4654 4655 #undef __FUNCT__ 4656 #define __FUNCT__ "TSSetMaxStepRejections" 4657 /*@ 4658 TSSetMaxStepRejections - Sets the maximum number of step rejections before a step fails 4659 4660 Not Collective 4661 4662 Input Parameter: 4663 + ts - TS context 4664 - rejects - maximum number of rejected steps, pass -1 for unlimited 4665 4666 Notes: 4667 The counter is reset to zero for each step 4668 4669 Options Database Key: 4670 . -ts_max_reject - Maximum number of step rejections before a step fails 4671 4672 Level: intermediate 4673 4674 .keywords: TS, set, maximum, number 4675 4676 .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxSNESFailures(), TSGetStepRejections(), TSGetSNESFailures(), TSSetErrorIfStepFails(), TSGetConvergedReason() 4677 @*/ 4678 PetscErrorCode TSSetMaxStepRejections(TS ts,PetscInt rejects) 4679 { 4680 PetscFunctionBegin; 4681 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4682 ts->max_reject = rejects; 4683 PetscFunctionReturn(0); 4684 } 4685 4686 #undef __FUNCT__ 4687 #define __FUNCT__ "TSSetMaxSNESFailures" 4688 /*@ 4689 TSSetMaxSNESFailures - Sets the maximum number of failed SNES solves 4690 4691 Not Collective 4692 4693 Input Parameter: 4694 + ts - TS context 4695 - fails - maximum number of failed nonlinear solves, pass -1 for unlimited 4696 4697 Notes: 4698 The counter is reset to zero for each successive call to TSSolve(). 4699 4700 Options Database Key: 4701 . -ts_max_snes_failures - Maximum number of nonlinear solve failures 4702 4703 Level: intermediate 4704 4705 .keywords: TS, set, maximum, number 4706 4707 .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSGetSNESFailures(), SNESGetConvergedReason(), TSGetConvergedReason() 4708 @*/ 4709 PetscErrorCode TSSetMaxSNESFailures(TS ts,PetscInt fails) 4710 { 4711 PetscFunctionBegin; 4712 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4713 ts->max_snes_failures = fails; 4714 PetscFunctionReturn(0); 4715 } 4716 4717 #undef __FUNCT__ 4718 #define __FUNCT__ "TSSetErrorIfStepFails" 4719 /*@ 4720 TSSetErrorIfStepFails - Error if no step succeeds 4721 4722 Not Collective 4723 4724 Input Parameter: 4725 + ts - TS context 4726 - err - PETSC_TRUE to error if no step succeeds, PETSC_FALSE to return without failure 4727 4728 Options Database Key: 4729 . -ts_error_if_step_fails - Error if no step succeeds 4730 4731 Level: intermediate 4732 4733 .keywords: TS, set, error 4734 4735 .seealso: TSGetSNESIterations(), TSGetKSPIterations(), TSSetMaxStepRejections(), TSGetStepRejections(), TSGetSNESFailures(), TSSetErrorIfStepFails(), TSGetConvergedReason() 4736 @*/ 4737 PetscErrorCode TSSetErrorIfStepFails(TS ts,PetscBool err) 4738 { 4739 PetscFunctionBegin; 4740 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4741 ts->errorifstepfailed = err; 4742 PetscFunctionReturn(0); 4743 } 4744 4745 #undef __FUNCT__ 4746 #define __FUNCT__ "TSMonitorSolutionBinary" 4747 /*@C 4748 TSMonitorSolutionBinary - Monitors progress of the TS solvers by VecView() for the solution at each timestep. Normally the viewer is a binary file 4749 4750 Collective on TS 4751 4752 Input Parameters: 4753 + ts - the TS context 4754 . step - current time-step 4755 . ptime - current time 4756 . u - current state 4757 - viewer - binary viewer 4758 4759 Level: intermediate 4760 4761 .keywords: TS, vector, monitor, view 4762 4763 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 4764 @*/ 4765 PetscErrorCode TSMonitorSolutionBinary(TS ts,PetscInt step,PetscReal ptime,Vec u,void *viewer) 4766 { 4767 PetscErrorCode ierr; 4768 PetscViewer v = (PetscViewer)viewer; 4769 4770 PetscFunctionBegin; 4771 ierr = VecView(u,v);CHKERRQ(ierr); 4772 PetscFunctionReturn(0); 4773 } 4774 4775 #undef __FUNCT__ 4776 #define __FUNCT__ "TSMonitorSolutionVTK" 4777 /*@C 4778 TSMonitorSolutionVTK - Monitors progress of the TS solvers by VecView() for the solution at each timestep. 4779 4780 Collective on TS 4781 4782 Input Parameters: 4783 + ts - the TS context 4784 . step - current time-step 4785 . ptime - current time 4786 . u - current state 4787 - filenametemplate - string containing a format specifier for the integer time step (e.g. %03D) 4788 4789 Level: intermediate 4790 4791 Notes: 4792 The VTK format does not allow writing multiple time steps in the same file, therefore a different file will be written for each time step. 4793 These are named according to the file name template. 4794 4795 This function is normally passed as an argument to TSMonitorSet() along with TSMonitorSolutionVTKDestroy(). 4796 4797 .keywords: TS, vector, monitor, view 4798 4799 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 4800 @*/ 4801 PetscErrorCode TSMonitorSolutionVTK(TS ts,PetscInt step,PetscReal ptime,Vec u,void *filenametemplate) 4802 { 4803 PetscErrorCode ierr; 4804 char filename[PETSC_MAX_PATH_LEN]; 4805 PetscViewer viewer; 4806 4807 PetscFunctionBegin; 4808 ierr = PetscSNPrintf(filename,sizeof(filename),(const char*)filenametemplate,step);CHKERRQ(ierr); 4809 ierr = PetscViewerVTKOpen(PetscObjectComm((PetscObject)ts),filename,FILE_MODE_WRITE,&viewer);CHKERRQ(ierr); 4810 ierr = VecView(u,viewer);CHKERRQ(ierr); 4811 ierr = PetscViewerDestroy(&viewer);CHKERRQ(ierr); 4812 PetscFunctionReturn(0); 4813 } 4814 4815 #undef __FUNCT__ 4816 #define __FUNCT__ "TSMonitorSolutionVTKDestroy" 4817 /*@C 4818 TSMonitorSolutionVTKDestroy - Destroy context for monitoring 4819 4820 Collective on TS 4821 4822 Input Parameters: 4823 . filenametemplate - string containing a format specifier for the integer time step (e.g. %03D) 4824 4825 Level: intermediate 4826 4827 Note: 4828 This function is normally passed to TSMonitorSet() along with TSMonitorSolutionVTK(). 4829 4830 .keywords: TS, vector, monitor, view 4831 4832 .seealso: TSMonitorSet(), TSMonitorSolutionVTK() 4833 @*/ 4834 PetscErrorCode TSMonitorSolutionVTKDestroy(void *filenametemplate) 4835 { 4836 PetscErrorCode ierr; 4837 4838 PetscFunctionBegin; 4839 ierr = PetscFree(*(char**)filenametemplate);CHKERRQ(ierr); 4840 PetscFunctionReturn(0); 4841 } 4842 4843 #undef __FUNCT__ 4844 #define __FUNCT__ "TSGetAdapt" 4845 /*@ 4846 TSGetAdapt - Get the adaptive controller context for the current method 4847 4848 Collective on TS if controller has not been created yet 4849 4850 Input Arguments: 4851 . ts - time stepping context 4852 4853 Output Arguments: 4854 . adapt - adaptive controller 4855 4856 Level: intermediate 4857 4858 .seealso: TSAdapt, TSAdaptSetType(), TSAdaptChoose() 4859 @*/ 4860 PetscErrorCode TSGetAdapt(TS ts,TSAdapt *adapt) 4861 { 4862 PetscErrorCode ierr; 4863 4864 PetscFunctionBegin; 4865 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4866 PetscValidPointer(adapt,2); 4867 if (!ts->adapt) { 4868 ierr = TSAdaptCreate(PetscObjectComm((PetscObject)ts),&ts->adapt);CHKERRQ(ierr); 4869 ierr = PetscLogObjectParent((PetscObject)ts,(PetscObject)ts->adapt);CHKERRQ(ierr); 4870 ierr = PetscObjectIncrementTabLevel((PetscObject)ts->adapt,(PetscObject)ts,1);CHKERRQ(ierr); 4871 } 4872 *adapt = ts->adapt; 4873 PetscFunctionReturn(0); 4874 } 4875 4876 #undef __FUNCT__ 4877 #define __FUNCT__ "TSSetTolerances" 4878 /*@ 4879 TSSetTolerances - Set tolerances for local truncation error when using adaptive controller 4880 4881 Logically Collective 4882 4883 Input Arguments: 4884 + ts - time integration context 4885 . atol - scalar absolute tolerances, PETSC_DECIDE to leave current value 4886 . vatol - vector of absolute tolerances or NULL, used in preference to atol if present 4887 . rtol - scalar relative tolerances, PETSC_DECIDE to leave current value 4888 - vrtol - vector of relative tolerances or NULL, used in preference to atol if present 4889 4890 Options Database keys: 4891 + -ts_rtol <rtol> - relative tolerance for local truncation error 4892 - -ts_atol <atol> Absolute tolerance for local truncation error 4893 4894 Level: beginner 4895 4896 .seealso: TS, TSAdapt, TSVecNormWRMS(), TSGetTolerances() 4897 @*/ 4898 PetscErrorCode TSSetTolerances(TS ts,PetscReal atol,Vec vatol,PetscReal rtol,Vec vrtol) 4899 { 4900 PetscErrorCode ierr; 4901 4902 PetscFunctionBegin; 4903 if (atol != PETSC_DECIDE && atol != PETSC_DEFAULT) ts->atol = atol; 4904 if (vatol) { 4905 ierr = PetscObjectReference((PetscObject)vatol);CHKERRQ(ierr); 4906 ierr = VecDestroy(&ts->vatol);CHKERRQ(ierr); 4907 4908 ts->vatol = vatol; 4909 } 4910 if (rtol != PETSC_DECIDE && rtol != PETSC_DEFAULT) ts->rtol = rtol; 4911 if (vrtol) { 4912 ierr = PetscObjectReference((PetscObject)vrtol);CHKERRQ(ierr); 4913 ierr = VecDestroy(&ts->vrtol);CHKERRQ(ierr); 4914 4915 ts->vrtol = vrtol; 4916 } 4917 PetscFunctionReturn(0); 4918 } 4919 4920 #undef __FUNCT__ 4921 #define __FUNCT__ "TSGetTolerances" 4922 /*@ 4923 TSGetTolerances - Get tolerances for local truncation error when using adaptive controller 4924 4925 Logically Collective 4926 4927 Input Arguments: 4928 . ts - time integration context 4929 4930 Output Arguments: 4931 + atol - scalar absolute tolerances, NULL to ignore 4932 . vatol - vector of absolute tolerances, NULL to ignore 4933 . rtol - scalar relative tolerances, NULL to ignore 4934 - vrtol - vector of relative tolerances, NULL to ignore 4935 4936 Level: beginner 4937 4938 .seealso: TS, TSAdapt, TSVecNormWRMS(), TSSetTolerances() 4939 @*/ 4940 PetscErrorCode TSGetTolerances(TS ts,PetscReal *atol,Vec *vatol,PetscReal *rtol,Vec *vrtol) 4941 { 4942 PetscFunctionBegin; 4943 if (atol) *atol = ts->atol; 4944 if (vatol) *vatol = ts->vatol; 4945 if (rtol) *rtol = ts->rtol; 4946 if (vrtol) *vrtol = ts->vrtol; 4947 PetscFunctionReturn(0); 4948 } 4949 4950 #undef __FUNCT__ 4951 #define __FUNCT__ "TSSetDifferentialEquationsIS" 4952 /*@ 4953 TSSetDifferentialEquationsIS - Sets an IS containing locations of differential equations in a DAE 4954 4955 Not Collective 4956 4957 Input Arguments: 4958 + ts - time stepping context 4959 - is_diff - Index set for differential equations 4960 4961 Level: advanced 4962 4963 @*/ 4964 PetscErrorCode TSSetDifferentialEquationsIS(TS ts, IS is_diff) 4965 { 4966 PetscErrorCode ierr; 4967 4968 PetscFunctionBegin; 4969 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 4970 PetscValidHeaderSpecific(is_diff,IS_CLASSID,2); 4971 PetscCheckSameComm(ts,1,is_diff,2); 4972 ierr = PetscObjectReference((PetscObject)is_diff);CHKERRQ(ierr); 4973 ierr = ISDestroy(&ts->is_diff);CHKERRQ(ierr); 4974 ts->is_diff = is_diff; 4975 PetscFunctionReturn(0); 4976 } 4977 4978 #undef __FUNCT__ 4979 #define __FUNCT__ "TSErrorWeightedNorm2" 4980 /*@ 4981 TSErrorWeightedNorm2 - compute a weighted 2-norm of the difference between two state vectors 4982 4983 Collective on TS 4984 4985 Input Arguments: 4986 + ts - time stepping context 4987 . U - state vector, usually ts->vec_sol 4988 - Y - state vector to be compared to U 4989 4990 Output Arguments: 4991 . norm - weighted norm, a value of 1.0 is considered small 4992 4993 Level: developer 4994 4995 .seealso: TSErrorWeightedNorm(), TSErrorWeightedNormInfinity() 4996 @*/ 4997 PetscErrorCode TSErrorWeightedNorm2(TS ts,Vec U,Vec Y,PetscReal *norm) 4998 { 4999 PetscErrorCode ierr; 5000 PetscInt i,n,N,rstart; 5001 const PetscScalar *u,*y; 5002 PetscReal sum,gsum; 5003 PetscReal tol; 5004 5005 PetscFunctionBegin; 5006 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 5007 PetscValidHeaderSpecific(U,VEC_CLASSID,2); 5008 PetscValidHeaderSpecific(Y,VEC_CLASSID,3); 5009 PetscValidType(U,2); 5010 PetscValidType(Y,3); 5011 PetscCheckSameComm(U,2,Y,3); 5012 PetscValidPointer(norm,4); 5013 if (U == Y) SETERRQ(PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_IDN,"U and Y cannot be the same vector"); 5014 5015 ierr = VecGetSize(U,&N);CHKERRQ(ierr); 5016 ierr = VecGetLocalSize(U,&n);CHKERRQ(ierr); 5017 ierr = VecGetOwnershipRange(U,&rstart,NULL);CHKERRQ(ierr); 5018 ierr = VecGetArrayRead(U,&u);CHKERRQ(ierr); 5019 ierr = VecGetArrayRead(Y,&y);CHKERRQ(ierr); 5020 sum = 0.; 5021 if (ts->vatol && ts->vrtol) { 5022 const PetscScalar *atol,*rtol; 5023 ierr = VecGetArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5024 ierr = VecGetArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5025 if(ts->is_diff) { 5026 const PetscInt *idx; 5027 PetscInt k; 5028 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5029 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5030 for(i=0; i < n; i++) { 5031 k = idx[i] - rstart; 5032 tol = PetscRealPart(atol[k]) + PetscRealPart(rtol[k])*PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5033 sum += PetscSqr(PetscAbsScalar(y[k] - u[k]) / tol); 5034 } 5035 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5036 } else { 5037 for (i=0; i<n; i++) { 5038 tol = PetscRealPart(atol[i]) + PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5039 sum += PetscSqr(PetscAbsScalar(y[i] - u[i]) / tol); 5040 } 5041 } 5042 ierr = VecRestoreArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5043 ierr = VecRestoreArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5044 } else if (ts->vatol) { /* vector atol, scalar rtol */ 5045 const PetscScalar *atol; 5046 ierr = VecGetArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5047 if(ts->is_diff) { 5048 const PetscInt *idx; 5049 PetscInt k; 5050 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5051 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5052 for(i=0; i < n; i++) { 5053 k = idx[i] - rstart; 5054 tol = PetscRealPart(atol[k]) + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5055 sum += PetscSqr(PetscAbsScalar(y[k] - u[k]) / tol); 5056 } 5057 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5058 } else { 5059 for (i=0; i<n; i++) { 5060 tol = PetscRealPart(atol[i]) + ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5061 sum += PetscSqr(PetscAbsScalar(y[i] - u[i]) / tol); 5062 } 5063 } 5064 ierr = VecRestoreArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5065 } else if (ts->vrtol) { /* scalar atol, vector rtol */ 5066 const PetscScalar *rtol; 5067 ierr = VecGetArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5068 if(ts->is_diff) { 5069 const PetscInt *idx; 5070 PetscInt k; 5071 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5072 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5073 for(i=0; i < n; i++) { 5074 k = idx[i] - rstart; 5075 tol = ts->atol + PetscRealPart(rtol[k]) * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5076 sum += PetscSqr(PetscAbsScalar(y[k] - u[k]) / tol); 5077 } 5078 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5079 } else { 5080 for (i=0; i<n; i++) { 5081 tol = ts->atol + PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5082 sum += PetscSqr(PetscAbsScalar(y[i] - u[i]) / tol); 5083 } 5084 } 5085 ierr = VecRestoreArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5086 } else { /* scalar atol, scalar rtol */ 5087 if (ts->is_diff) { 5088 const PetscInt *idx; 5089 PetscInt k; 5090 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5091 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5092 for (i=0; i<n; i++) { 5093 k = idx[i] - rstart; 5094 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5095 sum += PetscSqr(PetscAbsScalar(y[k] - u[k]) / tol); 5096 } 5097 } else { 5098 for (i=0; i<n; i++) { 5099 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5100 sum += PetscSqr(PetscAbsScalar(y[i] - u[i]) / tol); 5101 } 5102 } 5103 } 5104 ierr = VecRestoreArrayRead(U,&u);CHKERRQ(ierr); 5105 ierr = VecRestoreArrayRead(Y,&y);CHKERRQ(ierr); 5106 5107 ierr = MPI_Allreduce(&sum,&gsum,1,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)ts));CHKERRQ(ierr); 5108 *norm = PetscSqrtReal(gsum / N); 5109 5110 if (PetscIsInfOrNanScalar(*norm)) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm"); 5111 PetscFunctionReturn(0); 5112 } 5113 5114 #undef __FUNCT__ 5115 #define __FUNCT__ "TSErrorWeightedNormInfinity" 5116 /*@ 5117 TSErrorWeightedNormInfinity - compute a weighted infinity-norm of the difference between two state vectors 5118 5119 Collective on TS 5120 5121 Input Arguments: 5122 + ts - time stepping context 5123 . U - state vector, usually ts->vec_sol 5124 - Y - state vector to be compared to U 5125 5126 Output Arguments: 5127 . norm - weighted norm, a value of 1.0 is considered small 5128 5129 Level: developer 5130 5131 .seealso: TSErrorWeightedNorm(), TSErrorWeightedNorm2() 5132 @*/ 5133 PetscErrorCode TSErrorWeightedNormInfinity(TS ts,Vec U,Vec Y,PetscReal *norm) 5134 { 5135 PetscErrorCode ierr; 5136 PetscInt i,n,N,rstart,k; 5137 const PetscScalar *u,*y; 5138 PetscReal max,gmax; 5139 PetscReal tol; 5140 5141 PetscFunctionBegin; 5142 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 5143 PetscValidHeaderSpecific(U,VEC_CLASSID,2); 5144 PetscValidHeaderSpecific(Y,VEC_CLASSID,3); 5145 PetscValidType(U,2); 5146 PetscValidType(Y,3); 5147 PetscCheckSameComm(U,2,Y,3); 5148 PetscValidPointer(norm,4); 5149 if (U == Y) SETERRQ(PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_IDN,"U and Y cannot be the same vector"); 5150 5151 ierr = VecGetSize(U,&N);CHKERRQ(ierr); 5152 ierr = VecGetLocalSize(U,&n);CHKERRQ(ierr); 5153 ierr = VecGetOwnershipRange(U,&rstart,NULL);CHKERRQ(ierr); 5154 ierr = VecGetArrayRead(U,&u);CHKERRQ(ierr); 5155 ierr = VecGetArrayRead(Y,&y);CHKERRQ(ierr); 5156 if (ts->vatol && ts->vrtol) { 5157 const PetscScalar *atol,*rtol; 5158 ierr = VecGetArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5159 ierr = VecGetArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5160 if(ts->is_diff) { 5161 const PetscInt *idx; 5162 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5163 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5164 5165 k = idx[0]; 5166 tol = PetscRealPart(atol[k]) + PetscRealPart(rtol[k]) * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5167 max = PetscAbsScalar(y[k] - u[k]) / tol; 5168 for(i=1; i < n; i++) { 5169 k = idx[i] - rstart; 5170 tol = PetscRealPart(atol[k]) + PetscRealPart(rtol[k])*PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5171 max = PetscMax(max,PetscAbsScalar(y[k] - u[k]) / tol); 5172 } 5173 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5174 } else { 5175 k = 0; 5176 tol = PetscRealPart(atol[k]) + PetscRealPart(rtol[k]) * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5177 max = PetscAbsScalar(y[k] - u[k]) / tol; 5178 for (i=1; i<n; i++) { 5179 tol = PetscRealPart(atol[i]) + PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5180 max = PetscMax(max,PetscAbsScalar(y[i] - u[i]) / tol); 5181 } 5182 } 5183 ierr = VecRestoreArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5184 ierr = VecRestoreArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5185 } else if (ts->vatol) { /* vector atol, scalar rtol */ 5186 const PetscScalar *atol; 5187 ierr = VecGetArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5188 if(ts->is_diff) { 5189 const PetscInt *idx; 5190 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5191 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5192 5193 k = idx[0]; 5194 tol = PetscRealPart(atol[k]) + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5195 max = PetscAbsScalar(y[k] - u[k]) / tol; 5196 for(i=1; i < n; i++) { 5197 k = idx[i] - rstart; 5198 tol = PetscRealPart(atol[k]) + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5199 max = PetscMax(max,PetscAbsScalar(y[k] - u[k]) / tol); 5200 } 5201 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5202 } else { 5203 k = 0; 5204 tol = PetscRealPart(atol[k]) + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5205 max = PetscAbsScalar(y[k] - u[k]) / tol; 5206 for (i=1; i<n; i++) { 5207 tol = PetscRealPart(atol[i]) + ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5208 max = PetscMax(max,PetscAbsScalar(y[i] - u[i]) / tol); 5209 } 5210 } 5211 ierr = VecRestoreArrayRead(ts->vatol,&atol);CHKERRQ(ierr); 5212 } else if (ts->vrtol) { /* scalar atol, vector rtol */ 5213 const PetscScalar *rtol; 5214 ierr = VecGetArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5215 if(ts->is_diff) { 5216 const PetscInt *idx; 5217 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5218 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5219 5220 k = idx[0]; 5221 tol = ts->atol + PetscRealPart(rtol[k])*PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5222 max = PetscAbsScalar(y[k] - u[k]) / tol; 5223 for(i=1; i < n; i++) { 5224 k = idx[i] - rstart; 5225 tol = ts->atol + PetscRealPart(rtol[k]) * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5226 max = PetscMax(max,PetscAbsScalar(y[k] - u[k]) / tol); 5227 } 5228 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5229 } else { 5230 k = 0; 5231 tol = ts->atol + PetscRealPart(rtol[k]) * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5232 max = PetscAbsScalar(y[k] - u[k]) / tol; 5233 for (i=1; i<n; i++) { 5234 tol = ts->atol + PetscRealPart(rtol[i]) * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5235 max = PetscMax(max,PetscAbsScalar(y[i] - u[i]) / tol); 5236 } 5237 } 5238 ierr = VecRestoreArrayRead(ts->vrtol,&rtol);CHKERRQ(ierr); 5239 } else { /* scalar atol, scalar rtol */ 5240 if (ts->is_diff) { 5241 const PetscInt *idx; 5242 ierr = ISGetIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5243 ierr = ISGetLocalSize(ts->is_diff,&n);CHKERRQ(ierr); 5244 5245 k = idx[0]; 5246 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5247 max = PetscAbsScalar(y[k] - u[k]) / tol; 5248 for (i=1; i<n; i++) { 5249 k = idx[i] - rstart; 5250 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5251 max = PetscMax(max,PetscAbsScalar(y[k] - u[k]) / tol); 5252 } 5253 ierr = ISRestoreIndices(ts->is_diff,&idx);CHKERRQ(ierr); 5254 } else { 5255 k = 0; 5256 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[k]),PetscAbsScalar(y[k])); 5257 max = PetscAbsScalar(y[k] - u[k]) / tol; 5258 for (i=1; i<n; i++) { 5259 tol = ts->atol + ts->rtol * PetscMax(PetscAbsScalar(u[i]),PetscAbsScalar(y[i])); 5260 max = PetscMax(max,PetscAbsScalar(y[i] - u[i]) / tol); 5261 } 5262 } 5263 } 5264 ierr = VecRestoreArrayRead(U,&u);CHKERRQ(ierr); 5265 ierr = VecRestoreArrayRead(Y,&y);CHKERRQ(ierr); 5266 5267 ierr = MPI_Allreduce(&max,&gmax,1,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)ts));CHKERRQ(ierr); 5268 *norm = gmax; 5269 5270 if (PetscIsInfOrNanScalar(*norm)) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_FP,"Infinite or not-a-number generated in norm"); 5271 PetscFunctionReturn(0); 5272 } 5273 5274 #undef __FUNCT__ 5275 #define __FUNCT__ "TSErrorWeightedNorm" 5276 /*@ 5277 TSErrorWeightedNorm - compute a weighted norm of the difference between two state vectors 5278 5279 Collective on TS 5280 5281 Input Arguments: 5282 + ts - time stepping context 5283 . U - state vector, usually ts->vec_sol 5284 . Y - state vector to be compared to U 5285 - wnormtype - norm type, either NORM_2 or NORM_INFINITY 5286 5287 Output Arguments: 5288 . norm - weighted norm, a value of 1.0 is considered small 5289 5290 5291 Options Database Keys: 5292 . -ts_adapt_wnormtype <wnormtype> - 2, INFINITY 5293 5294 Level: developer 5295 5296 .seealso: TSErrorWeightedNormInfinity(), TSErrorWeightedNorm2() 5297 @*/ 5298 PetscErrorCode TSErrorWeightedNorm(TS ts,Vec U,Vec Y,NormType wnormtype,PetscReal *norm) 5299 { 5300 PetscErrorCode ierr; 5301 5302 PetscFunctionBegin; 5303 if (wnormtype == NORM_2) { 5304 ierr = TSErrorWeightedNorm2(ts,U,Y,norm);CHKERRQ(ierr); 5305 } else if(wnormtype == NORM_INFINITY) { 5306 ierr = TSErrorWeightedNormInfinity(ts,U,Y,norm);CHKERRQ(ierr); 5307 } else SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for norm type %s",NormTypes[wnormtype]); 5308 PetscFunctionReturn(0); 5309 } 5310 5311 #undef __FUNCT__ 5312 #define __FUNCT__ "TSSetCFLTimeLocal" 5313 /*@ 5314 TSSetCFLTimeLocal - Set the local CFL constraint relative to forward Euler 5315 5316 Logically Collective on TS 5317 5318 Input Arguments: 5319 + ts - time stepping context 5320 - cfltime - maximum stable time step if using forward Euler (value can be different on each process) 5321 5322 Note: 5323 After calling this function, the global CFL time can be obtained by calling TSGetCFLTime() 5324 5325 Level: intermediate 5326 5327 .seealso: TSGetCFLTime(), TSADAPTCFL 5328 @*/ 5329 PetscErrorCode TSSetCFLTimeLocal(TS ts,PetscReal cfltime) 5330 { 5331 PetscFunctionBegin; 5332 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 5333 ts->cfltime_local = cfltime; 5334 ts->cfltime = -1.; 5335 PetscFunctionReturn(0); 5336 } 5337 5338 #undef __FUNCT__ 5339 #define __FUNCT__ "TSGetCFLTime" 5340 /*@ 5341 TSGetCFLTime - Get the maximum stable time step according to CFL criteria applied to forward Euler 5342 5343 Collective on TS 5344 5345 Input Arguments: 5346 . ts - time stepping context 5347 5348 Output Arguments: 5349 . cfltime - maximum stable time step for forward Euler 5350 5351 Level: advanced 5352 5353 .seealso: TSSetCFLTimeLocal() 5354 @*/ 5355 PetscErrorCode TSGetCFLTime(TS ts,PetscReal *cfltime) 5356 { 5357 PetscErrorCode ierr; 5358 5359 PetscFunctionBegin; 5360 if (ts->cfltime < 0) { 5361 ierr = MPI_Allreduce(&ts->cfltime_local,&ts->cfltime,1,MPIU_REAL,MPIU_MIN,PetscObjectComm((PetscObject)ts));CHKERRQ(ierr); 5362 } 5363 *cfltime = ts->cfltime; 5364 PetscFunctionReturn(0); 5365 } 5366 5367 #undef __FUNCT__ 5368 #define __FUNCT__ "TSVISetVariableBounds" 5369 /*@ 5370 TSVISetVariableBounds - Sets the lower and upper bounds for the solution vector. xl <= x <= xu 5371 5372 Input Parameters: 5373 . ts - the TS context. 5374 . xl - lower bound. 5375 . xu - upper bound. 5376 5377 Notes: 5378 If this routine is not called then the lower and upper bounds are set to 5379 PETSC_NINFINITY and PETSC_INFINITY respectively during SNESSetUp(). 5380 5381 Level: advanced 5382 5383 @*/ 5384 PetscErrorCode TSVISetVariableBounds(TS ts, Vec xl, Vec xu) 5385 { 5386 PetscErrorCode ierr; 5387 SNES snes; 5388 5389 PetscFunctionBegin; 5390 ierr = TSGetSNES(ts,&snes);CHKERRQ(ierr); 5391 ierr = SNESVISetVariableBounds(snes,xl,xu);CHKERRQ(ierr); 5392 PetscFunctionReturn(0); 5393 } 5394 5395 #if defined(PETSC_HAVE_MATLAB_ENGINE) 5396 #include <mex.h> 5397 5398 typedef struct {char *funcname; mxArray *ctx;} TSMatlabContext; 5399 5400 #undef __FUNCT__ 5401 #define __FUNCT__ "TSComputeFunction_Matlab" 5402 /* 5403 TSComputeFunction_Matlab - Calls the function that has been set with 5404 TSSetFunctionMatlab(). 5405 5406 Collective on TS 5407 5408 Input Parameters: 5409 + snes - the TS context 5410 - u - input vector 5411 5412 Output Parameter: 5413 . y - function vector, as set by TSSetFunction() 5414 5415 Notes: 5416 TSComputeFunction() is typically used within nonlinear solvers 5417 implementations, so most users would not generally call this routine 5418 themselves. 5419 5420 Level: developer 5421 5422 .keywords: TS, nonlinear, compute, function 5423 5424 .seealso: TSSetFunction(), TSGetFunction() 5425 */ 5426 PetscErrorCode TSComputeFunction_Matlab(TS snes,PetscReal time,Vec u,Vec udot,Vec y, void *ctx) 5427 { 5428 PetscErrorCode ierr; 5429 TSMatlabContext *sctx = (TSMatlabContext*)ctx; 5430 int nlhs = 1,nrhs = 7; 5431 mxArray *plhs[1],*prhs[7]; 5432 long long int lx = 0,lxdot = 0,ly = 0,ls = 0; 5433 5434 PetscFunctionBegin; 5435 PetscValidHeaderSpecific(snes,TS_CLASSID,1); 5436 PetscValidHeaderSpecific(u,VEC_CLASSID,3); 5437 PetscValidHeaderSpecific(udot,VEC_CLASSID,4); 5438 PetscValidHeaderSpecific(y,VEC_CLASSID,5); 5439 PetscCheckSameComm(snes,1,u,3); 5440 PetscCheckSameComm(snes,1,y,5); 5441 5442 ierr = PetscMemcpy(&ls,&snes,sizeof(snes));CHKERRQ(ierr); 5443 ierr = PetscMemcpy(&lx,&u,sizeof(u));CHKERRQ(ierr); 5444 ierr = PetscMemcpy(&lxdot,&udot,sizeof(udot));CHKERRQ(ierr); 5445 ierr = PetscMemcpy(&ly,&y,sizeof(u));CHKERRQ(ierr); 5446 5447 prhs[0] = mxCreateDoubleScalar((double)ls); 5448 prhs[1] = mxCreateDoubleScalar(time); 5449 prhs[2] = mxCreateDoubleScalar((double)lx); 5450 prhs[3] = mxCreateDoubleScalar((double)lxdot); 5451 prhs[4] = mxCreateDoubleScalar((double)ly); 5452 prhs[5] = mxCreateString(sctx->funcname); 5453 prhs[6] = sctx->ctx; 5454 ierr = mexCallMATLAB(nlhs,plhs,nrhs,prhs,"PetscTSComputeFunctionInternal");CHKERRQ(ierr); 5455 ierr = mxGetScalar(plhs[0]);CHKERRQ(ierr); 5456 mxDestroyArray(prhs[0]); 5457 mxDestroyArray(prhs[1]); 5458 mxDestroyArray(prhs[2]); 5459 mxDestroyArray(prhs[3]); 5460 mxDestroyArray(prhs[4]); 5461 mxDestroyArray(prhs[5]); 5462 mxDestroyArray(plhs[0]); 5463 PetscFunctionReturn(0); 5464 } 5465 5466 5467 #undef __FUNCT__ 5468 #define __FUNCT__ "TSSetFunctionMatlab" 5469 /* 5470 TSSetFunctionMatlab - Sets the function evaluation routine and function 5471 vector for use by the TS routines in solving ODEs 5472 equations from MATLAB. Here the function is a string containing the name of a MATLAB function 5473 5474 Logically Collective on TS 5475 5476 Input Parameters: 5477 + ts - the TS context 5478 - func - function evaluation routine 5479 5480 Calling sequence of func: 5481 $ func (TS ts,PetscReal time,Vec u,Vec udot,Vec f,void *ctx); 5482 5483 Level: beginner 5484 5485 .keywords: TS, nonlinear, set, function 5486 5487 .seealso: TSGetFunction(), TSComputeFunction(), TSSetJacobian(), TSSetFunction() 5488 */ 5489 PetscErrorCode TSSetFunctionMatlab(TS ts,const char *func,mxArray *ctx) 5490 { 5491 PetscErrorCode ierr; 5492 TSMatlabContext *sctx; 5493 5494 PetscFunctionBegin; 5495 /* currently sctx is memory bleed */ 5496 ierr = PetscMalloc(sizeof(TSMatlabContext),&sctx);CHKERRQ(ierr); 5497 ierr = PetscStrallocpy(func,&sctx->funcname);CHKERRQ(ierr); 5498 /* 5499 This should work, but it doesn't 5500 sctx->ctx = ctx; 5501 mexMakeArrayPersistent(sctx->ctx); 5502 */ 5503 sctx->ctx = mxDuplicateArray(ctx); 5504 5505 ierr = TSSetIFunction(ts,NULL,TSComputeFunction_Matlab,sctx);CHKERRQ(ierr); 5506 PetscFunctionReturn(0); 5507 } 5508 5509 #undef __FUNCT__ 5510 #define __FUNCT__ "TSComputeJacobian_Matlab" 5511 /* 5512 TSComputeJacobian_Matlab - Calls the function that has been set with 5513 TSSetJacobianMatlab(). 5514 5515 Collective on TS 5516 5517 Input Parameters: 5518 + ts - the TS context 5519 . u - input vector 5520 . A, B - the matrices 5521 - ctx - user context 5522 5523 Level: developer 5524 5525 .keywords: TS, nonlinear, compute, function 5526 5527 .seealso: TSSetFunction(), TSGetFunction() 5528 @*/ 5529 PetscErrorCode TSComputeJacobian_Matlab(TS ts,PetscReal time,Vec u,Vec udot,PetscReal shift,Mat A,Mat B,void *ctx) 5530 { 5531 PetscErrorCode ierr; 5532 TSMatlabContext *sctx = (TSMatlabContext*)ctx; 5533 int nlhs = 2,nrhs = 9; 5534 mxArray *plhs[2],*prhs[9]; 5535 long long int lx = 0,lxdot = 0,lA = 0,ls = 0, lB = 0; 5536 5537 PetscFunctionBegin; 5538 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 5539 PetscValidHeaderSpecific(u,VEC_CLASSID,3); 5540 5541 /* call Matlab function in ctx with arguments u and y */ 5542 5543 ierr = PetscMemcpy(&ls,&ts,sizeof(ts));CHKERRQ(ierr); 5544 ierr = PetscMemcpy(&lx,&u,sizeof(u));CHKERRQ(ierr); 5545 ierr = PetscMemcpy(&lxdot,&udot,sizeof(u));CHKERRQ(ierr); 5546 ierr = PetscMemcpy(&lA,A,sizeof(u));CHKERRQ(ierr); 5547 ierr = PetscMemcpy(&lB,B,sizeof(u));CHKERRQ(ierr); 5548 5549 prhs[0] = mxCreateDoubleScalar((double)ls); 5550 prhs[1] = mxCreateDoubleScalar((double)time); 5551 prhs[2] = mxCreateDoubleScalar((double)lx); 5552 prhs[3] = mxCreateDoubleScalar((double)lxdot); 5553 prhs[4] = mxCreateDoubleScalar((double)shift); 5554 prhs[5] = mxCreateDoubleScalar((double)lA); 5555 prhs[6] = mxCreateDoubleScalar((double)lB); 5556 prhs[7] = mxCreateString(sctx->funcname); 5557 prhs[8] = sctx->ctx; 5558 ierr = mexCallMATLAB(nlhs,plhs,nrhs,prhs,"PetscTSComputeJacobianInternal");CHKERRQ(ierr); 5559 ierr = mxGetScalar(plhs[0]);CHKERRQ(ierr); 5560 mxDestroyArray(prhs[0]); 5561 mxDestroyArray(prhs[1]); 5562 mxDestroyArray(prhs[2]); 5563 mxDestroyArray(prhs[3]); 5564 mxDestroyArray(prhs[4]); 5565 mxDestroyArray(prhs[5]); 5566 mxDestroyArray(prhs[6]); 5567 mxDestroyArray(prhs[7]); 5568 mxDestroyArray(plhs[0]); 5569 mxDestroyArray(plhs[1]); 5570 PetscFunctionReturn(0); 5571 } 5572 5573 5574 #undef __FUNCT__ 5575 #define __FUNCT__ "TSSetJacobianMatlab" 5576 /* 5577 TSSetJacobianMatlab - Sets the Jacobian function evaluation routine and two empty Jacobian matrices 5578 vector for use by the TS routines in solving ODEs from MATLAB. Here the function is a string containing the name of a MATLAB function 5579 5580 Logically Collective on TS 5581 5582 Input Parameters: 5583 + ts - the TS context 5584 . A,B - Jacobian matrices 5585 . func - function evaluation routine 5586 - ctx - user context 5587 5588 Calling sequence of func: 5589 $ flag = func (TS ts,PetscReal time,Vec u,Vec udot,Mat A,Mat B,void *ctx); 5590 5591 5592 Level: developer 5593 5594 .keywords: TS, nonlinear, set, function 5595 5596 .seealso: TSGetFunction(), TSComputeFunction(), TSSetJacobian(), TSSetFunction() 5597 */ 5598 PetscErrorCode TSSetJacobianMatlab(TS ts,Mat A,Mat B,const char *func,mxArray *ctx) 5599 { 5600 PetscErrorCode ierr; 5601 TSMatlabContext *sctx; 5602 5603 PetscFunctionBegin; 5604 /* currently sctx is memory bleed */ 5605 ierr = PetscMalloc(sizeof(TSMatlabContext),&sctx);CHKERRQ(ierr); 5606 ierr = PetscStrallocpy(func,&sctx->funcname);CHKERRQ(ierr); 5607 /* 5608 This should work, but it doesn't 5609 sctx->ctx = ctx; 5610 mexMakeArrayPersistent(sctx->ctx); 5611 */ 5612 sctx->ctx = mxDuplicateArray(ctx); 5613 5614 ierr = TSSetIJacobian(ts,A,B,TSComputeJacobian_Matlab,sctx);CHKERRQ(ierr); 5615 PetscFunctionReturn(0); 5616 } 5617 5618 #undef __FUNCT__ 5619 #define __FUNCT__ "TSMonitor_Matlab" 5620 /* 5621 TSMonitor_Matlab - Calls the function that has been set with TSMonitorSetMatlab(). 5622 5623 Collective on TS 5624 5625 .seealso: TSSetFunction(), TSGetFunction() 5626 @*/ 5627 PetscErrorCode TSMonitor_Matlab(TS ts,PetscInt it, PetscReal time,Vec u, void *ctx) 5628 { 5629 PetscErrorCode ierr; 5630 TSMatlabContext *sctx = (TSMatlabContext*)ctx; 5631 int nlhs = 1,nrhs = 6; 5632 mxArray *plhs[1],*prhs[6]; 5633 long long int lx = 0,ls = 0; 5634 5635 PetscFunctionBegin; 5636 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 5637 PetscValidHeaderSpecific(u,VEC_CLASSID,4); 5638 5639 ierr = PetscMemcpy(&ls,&ts,sizeof(ts));CHKERRQ(ierr); 5640 ierr = PetscMemcpy(&lx,&u,sizeof(u));CHKERRQ(ierr); 5641 5642 prhs[0] = mxCreateDoubleScalar((double)ls); 5643 prhs[1] = mxCreateDoubleScalar((double)it); 5644 prhs[2] = mxCreateDoubleScalar((double)time); 5645 prhs[3] = mxCreateDoubleScalar((double)lx); 5646 prhs[4] = mxCreateString(sctx->funcname); 5647 prhs[5] = sctx->ctx; 5648 ierr = mexCallMATLAB(nlhs,plhs,nrhs,prhs,"PetscTSMonitorInternal");CHKERRQ(ierr); 5649 ierr = mxGetScalar(plhs[0]);CHKERRQ(ierr); 5650 mxDestroyArray(prhs[0]); 5651 mxDestroyArray(prhs[1]); 5652 mxDestroyArray(prhs[2]); 5653 mxDestroyArray(prhs[3]); 5654 mxDestroyArray(prhs[4]); 5655 mxDestroyArray(plhs[0]); 5656 PetscFunctionReturn(0); 5657 } 5658 5659 5660 #undef __FUNCT__ 5661 #define __FUNCT__ "TSMonitorSetMatlab" 5662 /* 5663 TSMonitorSetMatlab - Sets the monitor function from Matlab 5664 5665 Level: developer 5666 5667 .keywords: TS, nonlinear, set, function 5668 5669 .seealso: TSGetFunction(), TSComputeFunction(), TSSetJacobian(), TSSetFunction() 5670 */ 5671 PetscErrorCode TSMonitorSetMatlab(TS ts,const char *func,mxArray *ctx) 5672 { 5673 PetscErrorCode ierr; 5674 TSMatlabContext *sctx; 5675 5676 PetscFunctionBegin; 5677 /* currently sctx is memory bleed */ 5678 ierr = PetscMalloc(sizeof(TSMatlabContext),&sctx);CHKERRQ(ierr); 5679 ierr = PetscStrallocpy(func,&sctx->funcname);CHKERRQ(ierr); 5680 /* 5681 This should work, but it doesn't 5682 sctx->ctx = ctx; 5683 mexMakeArrayPersistent(sctx->ctx); 5684 */ 5685 sctx->ctx = mxDuplicateArray(ctx); 5686 5687 ierr = TSMonitorSet(ts,TSMonitor_Matlab,sctx,NULL);CHKERRQ(ierr); 5688 PetscFunctionReturn(0); 5689 } 5690 #endif 5691 5692 #undef __FUNCT__ 5693 #define __FUNCT__ "TSMonitorLGSolution" 5694 /*@C 5695 TSMonitorLGSolution - Monitors progress of the TS solvers by plotting each component of the solution vector 5696 in a time based line graph 5697 5698 Collective on TS 5699 5700 Input Parameters: 5701 + ts - the TS context 5702 . step - current time-step 5703 . ptime - current time 5704 - lg - a line graph object 5705 5706 Options Database: 5707 . -ts_monitor_lg_solution_variables 5708 5709 Level: intermediate 5710 5711 Notes: each process in a parallel run displays its component solutions in a separate window 5712 5713 .keywords: TS, vector, monitor, view 5714 5715 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView() 5716 @*/ 5717 PetscErrorCode TSMonitorLGSolution(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 5718 { 5719 PetscErrorCode ierr; 5720 TSMonitorLGCtx ctx = (TSMonitorLGCtx)dummy; 5721 const PetscScalar *yy; 5722 PetscInt dim; 5723 Vec v; 5724 5725 PetscFunctionBegin; 5726 if (!step) { 5727 PetscDrawAxis axis; 5728 ierr = PetscDrawLGGetAxis(ctx->lg,&axis);CHKERRQ(ierr); 5729 ierr = PetscDrawAxisSetLabels(axis,"Solution as function of time","Time","Solution");CHKERRQ(ierr); 5730 if (ctx->names && !ctx->displaynames) { 5731 char **displaynames; 5732 PetscBool flg; 5733 5734 ierr = VecGetLocalSize(u,&dim);CHKERRQ(ierr); 5735 ierr = PetscMalloc((dim+1)*sizeof(char*),&displaynames);CHKERRQ(ierr); 5736 ierr = PetscMemzero(displaynames,(dim+1)*sizeof(char*));CHKERRQ(ierr); 5737 ierr = PetscOptionsGetStringArray(((PetscObject)ts)->prefix,"-ts_monitor_lg_solution_variables",displaynames,&dim,&flg);CHKERRQ(ierr); 5738 if (flg) { 5739 ierr = TSMonitorLGCtxSetDisplayVariables(ctx,(const char *const *)displaynames);CHKERRQ(ierr); 5740 } 5741 ierr = PetscStrArrayDestroy(&displaynames);CHKERRQ(ierr); 5742 } 5743 if (ctx->displaynames) { 5744 ierr = PetscDrawLGSetDimension(ctx->lg,ctx->ndisplayvariables);CHKERRQ(ierr); 5745 ierr = PetscDrawLGSetLegend(ctx->lg,(const char *const *)ctx->displaynames);CHKERRQ(ierr); 5746 } else if (ctx->names) { 5747 ierr = VecGetLocalSize(u,&dim);CHKERRQ(ierr); 5748 ierr = PetscDrawLGSetDimension(ctx->lg,dim);CHKERRQ(ierr); 5749 ierr = PetscDrawLGSetLegend(ctx->lg,(const char *const *)ctx->names);CHKERRQ(ierr); 5750 } 5751 ierr = PetscDrawLGReset(ctx->lg);CHKERRQ(ierr); 5752 } 5753 if (ctx->transform) { 5754 ierr = (*ctx->transform)(ctx->transformctx,u,&v);CHKERRQ(ierr); 5755 } else { 5756 v = u; 5757 } 5758 ierr = VecGetArrayRead(v,&yy);CHKERRQ(ierr); 5759 #if defined(PETSC_USE_COMPLEX) 5760 { 5761 PetscReal *yreal; 5762 PetscInt i,n; 5763 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 5764 ierr = PetscMalloc1(n,&yreal);CHKERRQ(ierr); 5765 for (i=0; i<n; i++) yreal[i] = PetscRealPart(yy[i]); 5766 ierr = PetscDrawLGAddCommonPoint(ctx->lg,ptime,yreal);CHKERRQ(ierr); 5767 ierr = PetscFree(yreal);CHKERRQ(ierr); 5768 } 5769 #else 5770 if (ctx->displaynames) { 5771 PetscInt i; 5772 for (i=0; i<ctx->ndisplayvariables; i++) { 5773 ctx->displayvalues[i] = yy[ctx->displayvariables[i]]; 5774 } 5775 ierr = PetscDrawLGAddCommonPoint(ctx->lg,ptime,ctx->displayvalues);CHKERRQ(ierr); 5776 } else { 5777 ierr = PetscDrawLGAddCommonPoint(ctx->lg,ptime,yy);CHKERRQ(ierr); 5778 } 5779 #endif 5780 ierr = VecRestoreArrayRead(v,&yy);CHKERRQ(ierr); 5781 if (ctx->transform) { 5782 ierr = VecDestroy(&v);CHKERRQ(ierr); 5783 } 5784 if (((ctx->howoften > 0) && (!(step % ctx->howoften))) || ((ctx->howoften == -1) && ts->reason)) { 5785 ierr = PetscDrawLGDraw(ctx->lg);CHKERRQ(ierr); 5786 } 5787 PetscFunctionReturn(0); 5788 } 5789 5790 5791 #undef __FUNCT__ 5792 #define __FUNCT__ "TSMonitorLGSetVariableNames" 5793 /*@C 5794 TSMonitorLGSetVariableNames - Sets the name of each component in the solution vector so that it may be displayed in the plot 5795 5796 Collective on TS 5797 5798 Input Parameters: 5799 + ts - the TS context 5800 - names - the names of the components, final string must be NULL 5801 5802 Level: intermediate 5803 5804 .keywords: TS, vector, monitor, view 5805 5806 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetDisplayVariables(), TSMonitorLGCtxSetVariableNames() 5807 @*/ 5808 PetscErrorCode TSMonitorLGSetVariableNames(TS ts,const char * const *names) 5809 { 5810 PetscErrorCode ierr; 5811 PetscInt i; 5812 5813 PetscFunctionBegin; 5814 for (i=0; i<ts->numbermonitors; i++) { 5815 if (ts->monitor[i] == TSMonitorLGSolution) { 5816 ierr = TSMonitorLGCtxSetVariableNames((TSMonitorLGCtx)ts->monitorcontext[i],names);CHKERRQ(ierr); 5817 break; 5818 } 5819 } 5820 PetscFunctionReturn(0); 5821 } 5822 5823 #undef __FUNCT__ 5824 #define __FUNCT__ "TSMonitorLGCtxSetVariableNames" 5825 /*@C 5826 TSMonitorLGCtxSetVariableNames - Sets the name of each component in the solution vector so that it may be displayed in the plot 5827 5828 Collective on TS 5829 5830 Input Parameters: 5831 + ts - the TS context 5832 - names - the names of the components, final string must be NULL 5833 5834 Level: intermediate 5835 5836 .keywords: TS, vector, monitor, view 5837 5838 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetDisplayVariables(), TSMonitorLGSetVariableNames() 5839 @*/ 5840 PetscErrorCode TSMonitorLGCtxSetVariableNames(TSMonitorLGCtx ctx,const char * const *names) 5841 { 5842 PetscErrorCode ierr; 5843 5844 PetscFunctionBegin; 5845 ierr = PetscStrArrayDestroy(&ctx->names);CHKERRQ(ierr); 5846 ierr = PetscStrArrayallocpy(names,&ctx->names);CHKERRQ(ierr); 5847 PetscFunctionReturn(0); 5848 } 5849 5850 #undef __FUNCT__ 5851 #define __FUNCT__ "TSMonitorLGGetVariableNames" 5852 /*@C 5853 TSMonitorLGGetVariableNames - Gets the name of each component in the solution vector so that it may be displayed in the plot 5854 5855 Collective on TS 5856 5857 Input Parameter: 5858 . ts - the TS context 5859 5860 Output Parameter: 5861 . names - the names of the components, final string must be NULL 5862 5863 Level: intermediate 5864 5865 .keywords: TS, vector, monitor, view 5866 5867 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetDisplayVariables() 5868 @*/ 5869 PetscErrorCode TSMonitorLGGetVariableNames(TS ts,const char *const **names) 5870 { 5871 PetscInt i; 5872 5873 PetscFunctionBegin; 5874 *names = NULL; 5875 for (i=0; i<ts->numbermonitors; i++) { 5876 if (ts->monitor[i] == TSMonitorLGSolution) { 5877 TSMonitorLGCtx ctx = (TSMonitorLGCtx) ts->monitorcontext[i]; 5878 *names = (const char *const *)ctx->names; 5879 break; 5880 } 5881 } 5882 PetscFunctionReturn(0); 5883 } 5884 5885 #undef __FUNCT__ 5886 #define __FUNCT__ "TSMonitorLGCtxSetDisplayVariables" 5887 /*@C 5888 TSMonitorLGCtxSetDisplayVariables - Sets the variables that are to be display in the monitor 5889 5890 Collective on TS 5891 5892 Input Parameters: 5893 + ctx - the TSMonitorLG context 5894 . displaynames - the names of the components, final string must be NULL 5895 5896 Level: intermediate 5897 5898 .keywords: TS, vector, monitor, view 5899 5900 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetVariableNames() 5901 @*/ 5902 PetscErrorCode TSMonitorLGCtxSetDisplayVariables(TSMonitorLGCtx ctx,const char * const *displaynames) 5903 { 5904 PetscInt j = 0,k; 5905 PetscErrorCode ierr; 5906 5907 PetscFunctionBegin; 5908 if (!ctx->names) PetscFunctionReturn(0); 5909 ierr = PetscStrArrayDestroy(&ctx->displaynames);CHKERRQ(ierr); 5910 ierr = PetscStrArrayallocpy(displaynames,&ctx->displaynames);CHKERRQ(ierr); 5911 while (displaynames[j]) j++; 5912 ctx->ndisplayvariables = j; 5913 ierr = PetscMalloc1(ctx->ndisplayvariables,&ctx->displayvariables);CHKERRQ(ierr); 5914 ierr = PetscMalloc1(ctx->ndisplayvariables,&ctx->displayvalues);CHKERRQ(ierr); 5915 j = 0; 5916 while (displaynames[j]) { 5917 k = 0; 5918 while (ctx->names[k]) { 5919 PetscBool flg; 5920 ierr = PetscStrcmp(displaynames[j],ctx->names[k],&flg);CHKERRQ(ierr); 5921 if (flg) { 5922 ctx->displayvariables[j] = k; 5923 break; 5924 } 5925 k++; 5926 } 5927 j++; 5928 } 5929 PetscFunctionReturn(0); 5930 } 5931 5932 5933 #undef __FUNCT__ 5934 #define __FUNCT__ "TSMonitorLGSetDisplayVariables" 5935 /*@C 5936 TSMonitorLGSetDisplayVariables - Sets the variables that are to be display in the monitor 5937 5938 Collective on TS 5939 5940 Input Parameters: 5941 + ts - the TS context 5942 . displaynames - the names of the components, final string must be NULL 5943 5944 Level: intermediate 5945 5946 .keywords: TS, vector, monitor, view 5947 5948 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetVariableNames() 5949 @*/ 5950 PetscErrorCode TSMonitorLGSetDisplayVariables(TS ts,const char * const *displaynames) 5951 { 5952 PetscInt i; 5953 PetscErrorCode ierr; 5954 5955 PetscFunctionBegin; 5956 for (i=0; i<ts->numbermonitors; i++) { 5957 if (ts->monitor[i] == TSMonitorLGSolution) { 5958 ierr = TSMonitorLGCtxSetDisplayVariables((TSMonitorLGCtx)ts->monitorcontext[i],displaynames);CHKERRQ(ierr); 5959 break; 5960 } 5961 } 5962 PetscFunctionReturn(0); 5963 } 5964 5965 #undef __FUNCT__ 5966 #define __FUNCT__ "TSMonitorLGSetTransform" 5967 /*@C 5968 TSMonitorLGSetTransform - Solution vector will be transformed by provided function before being displayed 5969 5970 Collective on TS 5971 5972 Input Parameters: 5973 + ts - the TS context 5974 . transform - the transform function 5975 . destroy - function to destroy the optional context 5976 - ctx - optional context used by transform function 5977 5978 Level: intermediate 5979 5980 .keywords: TS, vector, monitor, view 5981 5982 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetVariableNames(), TSMonitorLGCtxSetTransform() 5983 @*/ 5984 PetscErrorCode TSMonitorLGSetTransform(TS ts,PetscErrorCode (*transform)(void*,Vec,Vec*),PetscErrorCode (*destroy)(void*),void *tctx) 5985 { 5986 PetscInt i; 5987 PetscErrorCode ierr; 5988 5989 PetscFunctionBegin; 5990 for (i=0; i<ts->numbermonitors; i++) { 5991 if (ts->monitor[i] == TSMonitorLGSolution) { 5992 ierr = TSMonitorLGCtxSetTransform((TSMonitorLGCtx)ts->monitorcontext[i],transform,destroy,tctx);CHKERRQ(ierr); 5993 } 5994 } 5995 PetscFunctionReturn(0); 5996 } 5997 5998 #undef __FUNCT__ 5999 #define __FUNCT__ "TSMonitorLGCtxSetTransform" 6000 /*@C 6001 TSMonitorLGCtxSetTransform - Solution vector will be transformed by provided function before being displayed 6002 6003 Collective on TSLGCtx 6004 6005 Input Parameters: 6006 + ts - the TS context 6007 . transform - the transform function 6008 . destroy - function to destroy the optional context 6009 - ctx - optional context used by transform function 6010 6011 Level: intermediate 6012 6013 .keywords: TS, vector, monitor, view 6014 6015 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetVariableNames(), TSMonitorLGSetTransform() 6016 @*/ 6017 PetscErrorCode TSMonitorLGCtxSetTransform(TSMonitorLGCtx ctx,PetscErrorCode (*transform)(void*,Vec,Vec*),PetscErrorCode (*destroy)(void*),void *tctx) 6018 { 6019 PetscFunctionBegin; 6020 ctx->transform = transform; 6021 ctx->transformdestroy = destroy; 6022 ctx->transformctx = tctx; 6023 PetscFunctionReturn(0); 6024 } 6025 6026 #undef __FUNCT__ 6027 #define __FUNCT__ "TSMonitorLGError" 6028 /*@C 6029 TSMonitorLGError - Monitors progress of the TS solvers by plotting each component of the solution vector 6030 in a time based line graph 6031 6032 Collective on TS 6033 6034 Input Parameters: 6035 + ts - the TS context 6036 . step - current time-step 6037 . ptime - current time 6038 - lg - a line graph object 6039 6040 Level: intermediate 6041 6042 Notes: 6043 Only for sequential solves. 6044 6045 The user must provide the solution using TSSetSolutionFunction() to use this monitor. 6046 6047 Options Database Keys: 6048 . -ts_monitor_lg_error - create a graphical monitor of error history 6049 6050 .keywords: TS, vector, monitor, view 6051 6052 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSSetSolutionFunction() 6053 @*/ 6054 PetscErrorCode TSMonitorLGError(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 6055 { 6056 PetscErrorCode ierr; 6057 TSMonitorLGCtx ctx = (TSMonitorLGCtx)dummy; 6058 const PetscScalar *yy; 6059 Vec y; 6060 PetscInt dim; 6061 6062 PetscFunctionBegin; 6063 if (!step) { 6064 PetscDrawAxis axis; 6065 ierr = PetscDrawLGGetAxis(ctx->lg,&axis);CHKERRQ(ierr); 6066 ierr = PetscDrawAxisSetLabels(axis,"Error in solution as function of time","Time","Solution");CHKERRQ(ierr); 6067 ierr = VecGetLocalSize(u,&dim);CHKERRQ(ierr); 6068 ierr = PetscDrawLGSetDimension(ctx->lg,dim);CHKERRQ(ierr); 6069 ierr = PetscDrawLGReset(ctx->lg);CHKERRQ(ierr); 6070 } 6071 ierr = VecDuplicate(u,&y);CHKERRQ(ierr); 6072 ierr = TSComputeSolutionFunction(ts,ptime,y);CHKERRQ(ierr); 6073 ierr = VecAXPY(y,-1.0,u);CHKERRQ(ierr); 6074 ierr = VecGetArrayRead(y,&yy);CHKERRQ(ierr); 6075 #if defined(PETSC_USE_COMPLEX) 6076 { 6077 PetscReal *yreal; 6078 PetscInt i,n; 6079 ierr = VecGetLocalSize(y,&n);CHKERRQ(ierr); 6080 ierr = PetscMalloc1(n,&yreal);CHKERRQ(ierr); 6081 for (i=0; i<n; i++) yreal[i] = PetscRealPart(yy[i]); 6082 ierr = PetscDrawLGAddCommonPoint(ctx->lg,ptime,yreal);CHKERRQ(ierr); 6083 ierr = PetscFree(yreal);CHKERRQ(ierr); 6084 } 6085 #else 6086 ierr = PetscDrawLGAddCommonPoint(ctx->lg,ptime,yy);CHKERRQ(ierr); 6087 #endif 6088 ierr = VecRestoreArrayRead(y,&yy);CHKERRQ(ierr); 6089 ierr = VecDestroy(&y);CHKERRQ(ierr); 6090 if (((ctx->howoften > 0) && (!(step % ctx->howoften))) || ((ctx->howoften == -1) && ts->reason)) { 6091 ierr = PetscDrawLGDraw(ctx->lg);CHKERRQ(ierr); 6092 } 6093 PetscFunctionReturn(0); 6094 } 6095 6096 #undef __FUNCT__ 6097 #define __FUNCT__ "TSMonitorLGSNESIterations" 6098 PetscErrorCode TSMonitorLGSNESIterations(TS ts,PetscInt n,PetscReal ptime,Vec v,void *monctx) 6099 { 6100 TSMonitorLGCtx ctx = (TSMonitorLGCtx) monctx; 6101 PetscReal x = ptime,y; 6102 PetscErrorCode ierr; 6103 PetscInt its; 6104 6105 PetscFunctionBegin; 6106 if (!n) { 6107 PetscDrawAxis axis; 6108 6109 ierr = PetscDrawLGGetAxis(ctx->lg,&axis);CHKERRQ(ierr); 6110 ierr = PetscDrawAxisSetLabels(axis,"Nonlinear iterations as function of time","Time","SNES Iterations");CHKERRQ(ierr); 6111 ierr = PetscDrawLGReset(ctx->lg);CHKERRQ(ierr); 6112 6113 ctx->snes_its = 0; 6114 } 6115 ierr = TSGetSNESIterations(ts,&its);CHKERRQ(ierr); 6116 y = its - ctx->snes_its; 6117 ierr = PetscDrawLGAddPoint(ctx->lg,&x,&y);CHKERRQ(ierr); 6118 if (((ctx->howoften > 0) && (!(n % ctx->howoften)) && (n > -1)) || ((ctx->howoften == -1) && (n == -1))) { 6119 ierr = PetscDrawLGDraw(ctx->lg);CHKERRQ(ierr); 6120 } 6121 ctx->snes_its = its; 6122 PetscFunctionReturn(0); 6123 } 6124 6125 #undef __FUNCT__ 6126 #define __FUNCT__ "TSMonitorLGKSPIterations" 6127 PetscErrorCode TSMonitorLGKSPIterations(TS ts,PetscInt n,PetscReal ptime,Vec v,void *monctx) 6128 { 6129 TSMonitorLGCtx ctx = (TSMonitorLGCtx) monctx; 6130 PetscReal x = ptime,y; 6131 PetscErrorCode ierr; 6132 PetscInt its; 6133 6134 PetscFunctionBegin; 6135 if (!n) { 6136 PetscDrawAxis axis; 6137 6138 ierr = PetscDrawLGGetAxis(ctx->lg,&axis);CHKERRQ(ierr); 6139 ierr = PetscDrawAxisSetLabels(axis,"Linear iterations as function of time","Time","KSP Iterations");CHKERRQ(ierr); 6140 ierr = PetscDrawLGReset(ctx->lg);CHKERRQ(ierr); 6141 6142 ctx->ksp_its = 0; 6143 } 6144 ierr = TSGetKSPIterations(ts,&its);CHKERRQ(ierr); 6145 y = its - ctx->ksp_its; 6146 ierr = PetscDrawLGAddPoint(ctx->lg,&x,&y);CHKERRQ(ierr); 6147 if (((ctx->howoften > 0) && (!(n % ctx->howoften)) && (n > -1)) || ((ctx->howoften == -1) && (n == -1))) { 6148 ierr = PetscDrawLGDraw(ctx->lg);CHKERRQ(ierr); 6149 } 6150 ctx->ksp_its = its; 6151 PetscFunctionReturn(0); 6152 } 6153 6154 #undef __FUNCT__ 6155 #define __FUNCT__ "TSComputeLinearStability" 6156 /*@ 6157 TSComputeLinearStability - computes the linear stability function at a point 6158 6159 Collective on TS and Vec 6160 6161 Input Parameters: 6162 + ts - the TS context 6163 - xr,xi - real and imaginary part of input arguments 6164 6165 Output Parameters: 6166 . yr,yi - real and imaginary part of function value 6167 6168 Level: developer 6169 6170 .keywords: TS, compute 6171 6172 .seealso: TSSetRHSFunction(), TSComputeIFunction() 6173 @*/ 6174 PetscErrorCode TSComputeLinearStability(TS ts,PetscReal xr,PetscReal xi,PetscReal *yr,PetscReal *yi) 6175 { 6176 PetscErrorCode ierr; 6177 6178 PetscFunctionBegin; 6179 PetscValidHeaderSpecific(ts,TS_CLASSID,1); 6180 if (!ts->ops->linearstability) SETERRQ(PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"Linearized stability function not provided for this method"); 6181 ierr = (*ts->ops->linearstability)(ts,xr,xi,yr,yi);CHKERRQ(ierr); 6182 PetscFunctionReturn(0); 6183 } 6184 6185 /* ------------------------------------------------------------------------*/ 6186 #undef __FUNCT__ 6187 #define __FUNCT__ "TSMonitorEnvelopeCtxCreate" 6188 /*@C 6189 TSMonitorEnvelopeCtxCreate - Creates a context for use with TSMonitorEnvelope() 6190 6191 Collective on TS 6192 6193 Input Parameters: 6194 . ts - the ODE solver object 6195 6196 Output Parameter: 6197 . ctx - the context 6198 6199 Level: intermediate 6200 6201 .keywords: TS, monitor, line graph, residual, seealso 6202 6203 .seealso: TSMonitorLGTimeStep(), TSMonitorSet(), TSMonitorLGSolution(), TSMonitorLGError() 6204 6205 @*/ 6206 PetscErrorCode TSMonitorEnvelopeCtxCreate(TS ts,TSMonitorEnvelopeCtx *ctx) 6207 { 6208 PetscErrorCode ierr; 6209 6210 PetscFunctionBegin; 6211 ierr = PetscNew(ctx);CHKERRQ(ierr); 6212 PetscFunctionReturn(0); 6213 } 6214 6215 #undef __FUNCT__ 6216 #define __FUNCT__ "TSMonitorEnvelope" 6217 /*@C 6218 TSMonitorEnvelope - Monitors the maximum and minimum value of each component of the solution 6219 6220 Collective on TS 6221 6222 Input Parameters: 6223 + ts - the TS context 6224 . step - current time-step 6225 . ptime - current time 6226 - ctx - the envelope context 6227 6228 Options Database: 6229 . -ts_monitor_envelope 6230 6231 Level: intermediate 6232 6233 Notes: after a solve you can use TSMonitorEnvelopeGetBounds() to access the envelope 6234 6235 .keywords: TS, vector, monitor, view 6236 6237 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorEnvelopeGetBounds() 6238 @*/ 6239 PetscErrorCode TSMonitorEnvelope(TS ts,PetscInt step,PetscReal ptime,Vec u,void *dummy) 6240 { 6241 PetscErrorCode ierr; 6242 TSMonitorEnvelopeCtx ctx = (TSMonitorEnvelopeCtx)dummy; 6243 6244 PetscFunctionBegin; 6245 if (!ctx->max) { 6246 ierr = VecDuplicate(u,&ctx->max);CHKERRQ(ierr); 6247 ierr = VecDuplicate(u,&ctx->min);CHKERRQ(ierr); 6248 ierr = VecCopy(u,ctx->max);CHKERRQ(ierr); 6249 ierr = VecCopy(u,ctx->min);CHKERRQ(ierr); 6250 } else { 6251 ierr = VecPointwiseMax(ctx->max,u,ctx->max);CHKERRQ(ierr); 6252 ierr = VecPointwiseMin(ctx->min,u,ctx->min);CHKERRQ(ierr); 6253 } 6254 PetscFunctionReturn(0); 6255 } 6256 6257 6258 #undef __FUNCT__ 6259 #define __FUNCT__ "TSMonitorEnvelopeGetBounds" 6260 /*@C 6261 TSMonitorEnvelopeGetBounds - Gets the bounds for the components of the solution 6262 6263 Collective on TS 6264 6265 Input Parameter: 6266 . ts - the TS context 6267 6268 Output Parameter: 6269 + max - the maximum values 6270 - min - the minimum values 6271 6272 Level: intermediate 6273 6274 .keywords: TS, vector, monitor, view 6275 6276 .seealso: TSMonitorSet(), TSMonitorDefault(), VecView(), TSMonitorLGSetDisplayVariables() 6277 @*/ 6278 PetscErrorCode TSMonitorEnvelopeGetBounds(TS ts,Vec *max,Vec *min) 6279 { 6280 PetscInt i; 6281 6282 PetscFunctionBegin; 6283 if (max) *max = NULL; 6284 if (min) *min = NULL; 6285 for (i=0; i<ts->numbermonitors; i++) { 6286 if (ts->monitor[i] == TSMonitorEnvelope) { 6287 TSMonitorEnvelopeCtx ctx = (TSMonitorEnvelopeCtx) ts->monitorcontext[i]; 6288 if (max) *max = ctx->max; 6289 if (min) *min = ctx->min; 6290 break; 6291 } 6292 } 6293 PetscFunctionReturn(0); 6294 } 6295 6296 #undef __FUNCT__ 6297 #define __FUNCT__ "TSMonitorEnvelopeCtxDestroy" 6298 /*@C 6299 TSMonitorEnvelopeCtxDestroy - Destroys a context that was created with TSMonitorEnvelopeCtxCreate(). 6300 6301 Collective on TSMonitorEnvelopeCtx 6302 6303 Input Parameter: 6304 . ctx - the monitor context 6305 6306 Level: intermediate 6307 6308 .keywords: TS, monitor, line graph, destroy 6309 6310 .seealso: TSMonitorLGCtxCreate(), TSMonitorSet(), TSMonitorLGTimeStep(); 6311 @*/ 6312 PetscErrorCode TSMonitorEnvelopeCtxDestroy(TSMonitorEnvelopeCtx *ctx) 6313 { 6314 PetscErrorCode ierr; 6315 6316 PetscFunctionBegin; 6317 ierr = VecDestroy(&(*ctx)->min);CHKERRQ(ierr); 6318 ierr = VecDestroy(&(*ctx)->max);CHKERRQ(ierr); 6319 ierr = PetscFree(*ctx);CHKERRQ(ierr); 6320 PetscFunctionReturn(0); 6321 } 6322 6323 #undef __FUNCT__ 6324 #define __FUNCT__ "TSRollBack" 6325 /*@ 6326 TSRollBack - Rolls back one time step 6327 6328 Collective on TS 6329 6330 Input Parameter: 6331 . ts - the TS context obtained from TSCreate() 6332 6333 Level: advanced 6334 6335 .keywords: TS, timestep, rollback 6336 6337 .seealso: TSCreate(), TSSetUp(), TSDestroy(), TSSolve(), TSSetPreStep(), TSSetPreStage(), TSInterpolate() 6338 @*/ 6339 PetscErrorCode TSRollBack(TS ts) 6340 { 6341 PetscErrorCode ierr; 6342 6343 PetscFunctionBegin; 6344 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 6345 6346 if (!ts->ops->rollback) SETERRQ1(PetscObjectComm((PetscObject)ts),PETSC_ERR_SUP,"TSRollBack not implemented for type '%s'",((PetscObject)ts)->type_name); 6347 ierr = (*ts->ops->rollback)(ts);CHKERRQ(ierr); 6348 ts->time_step = ts->ptime - ts->ptime_prev; 6349 ts->ptime = ts->ptime_prev; 6350 ts->steprollback = PETSC_TRUE; /* Flag to indicate that the step is rollbacked */ 6351 PetscFunctionReturn(0); 6352 } 6353 6354 #undef __FUNCT__ 6355 #define __FUNCT__ "TSGetStages" 6356 /*@ 6357 TSGetStages - Get the number of stages and stage values 6358 6359 Input Parameter: 6360 . ts - the TS context obtained from TSCreate() 6361 6362 Level: advanced 6363 6364 .keywords: TS, getstages 6365 6366 .seealso: TSCreate() 6367 @*/ 6368 PetscErrorCode TSGetStages(TS ts,PetscInt *ns, Vec **Y) 6369 { 6370 PetscErrorCode ierr; 6371 6372 PetscFunctionBegin; 6373 PetscValidHeaderSpecific(ts, TS_CLASSID,1); 6374 PetscValidPointer(ns,2); 6375 6376 if (!ts->ops->getstages) *ns=0; 6377 else { 6378 ierr = (*ts->ops->getstages)(ts,ns,Y);CHKERRQ(ierr); 6379 } 6380 PetscFunctionReturn(0); 6381 } 6382 6383 6384 #undef __FUNCT__ 6385 #define __FUNCT__ "TSComputeIJacobianDefaultColor" 6386 /*@C 6387 TSComputeIJacobianDefaultColor - Computes the Jacobian using finite differences and coloring to exploit matrix sparsity. 6388 6389 Collective on SNES 6390 6391 Input Parameters: 6392 + ts - the TS context 6393 . t - current timestep 6394 . U - state vector 6395 . Udot - time derivative of state vector 6396 . shift - shift to apply, see note below 6397 - ctx - an optional user context 6398 6399 Output Parameters: 6400 + J - Jacobian matrix (not altered in this routine) 6401 - B - newly computed Jacobian matrix to use with preconditioner (generally the same as J) 6402 6403 Level: intermediate 6404 6405 Notes: 6406 If F(t,U,Udot)=0 is the DAE, the required Jacobian is 6407 6408 dF/dU + shift*dF/dUdot 6409 6410 Most users should not need to explicitly call this routine, as it 6411 is used internally within the nonlinear solvers. 6412 6413 This will first try to get the coloring from the DM. If the DM type has no coloring 6414 routine, then it will try to get the coloring from the matrix. This requires that the 6415 matrix have nonzero entries precomputed. 6416 6417 .keywords: TS, finite differences, Jacobian, coloring, sparse 6418 .seealso: TSSetIJacobian(), MatFDColoringCreate(), MatFDColoringSetFunction() 6419 @*/ 6420 PetscErrorCode TSComputeIJacobianDefaultColor(TS ts,PetscReal t,Vec U,Vec Udot,PetscReal shift,Mat J,Mat B,void *ctx) 6421 { 6422 SNES snes; 6423 MatFDColoring color; 6424 PetscBool hascolor, matcolor = PETSC_FALSE; 6425 PetscErrorCode ierr; 6426 6427 PetscFunctionBegin; 6428 ierr = PetscOptionsGetBool(((PetscObject) ts)->prefix, "-ts_fd_color_use_mat", &matcolor, NULL);CHKERRQ(ierr); 6429 ierr = PetscObjectQuery((PetscObject) B, "TSMatFDColoring", (PetscObject *) &color);CHKERRQ(ierr); 6430 if (!color) { 6431 DM dm; 6432 ISColoring iscoloring; 6433 6434 ierr = TSGetDM(ts, &dm);CHKERRQ(ierr); 6435 ierr = DMHasColoring(dm, &hascolor);CHKERRQ(ierr); 6436 if (hascolor && !matcolor) { 6437 ierr = DMCreateColoring(dm, IS_COLORING_GLOBAL, &iscoloring);CHKERRQ(ierr); 6438 ierr = MatFDColoringCreate(B, iscoloring, &color);CHKERRQ(ierr); 6439 ierr = MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts);CHKERRQ(ierr); 6440 ierr = MatFDColoringSetFromOptions(color);CHKERRQ(ierr); 6441 ierr = MatFDColoringSetUp(B, iscoloring, color);CHKERRQ(ierr); 6442 ierr = ISColoringDestroy(&iscoloring);CHKERRQ(ierr); 6443 } else { 6444 MatColoring mc; 6445 6446 ierr = MatColoringCreate(B, &mc);CHKERRQ(ierr); 6447 ierr = MatColoringSetDistance(mc, 2);CHKERRQ(ierr); 6448 ierr = MatColoringSetType(mc, MATCOLORINGSL);CHKERRQ(ierr); 6449 ierr = MatColoringSetFromOptions(mc);CHKERRQ(ierr); 6450 ierr = MatColoringApply(mc, &iscoloring);CHKERRQ(ierr); 6451 ierr = MatColoringDestroy(&mc);CHKERRQ(ierr); 6452 ierr = MatFDColoringCreate(B, iscoloring, &color);CHKERRQ(ierr); 6453 ierr = MatFDColoringSetFunction(color, (PetscErrorCode (*)(void)) SNESTSFormFunction, (void *) ts);CHKERRQ(ierr); 6454 ierr = MatFDColoringSetFromOptions(color);CHKERRQ(ierr); 6455 ierr = MatFDColoringSetUp(B, iscoloring, color);CHKERRQ(ierr); 6456 ierr = ISColoringDestroy(&iscoloring);CHKERRQ(ierr); 6457 } 6458 ierr = PetscObjectCompose((PetscObject) B, "TSMatFDColoring", (PetscObject) color);CHKERRQ(ierr); 6459 ierr = PetscObjectDereference((PetscObject) color);CHKERRQ(ierr); 6460 } 6461 ierr = TSGetSNES(ts, &snes);CHKERRQ(ierr); 6462 ierr = MatFDColoringApply(B, color, U, snes);CHKERRQ(ierr); 6463 if (J != B) { 6464 ierr = MatAssemblyBegin(J, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 6465 ierr = MatAssemblyEnd(J, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 6466 } 6467 PetscFunctionReturn(0); 6468 } 6469