1 /* 2 Include "petsctao.h" so that we can use TAO solvers. Note that this 3 file automatically includes libraries such as: 4 petsc.h - base PETSc routines petscvec.h - vectors 5 petscsys.h - system routines petscmat.h - matrices 6 petscis.h - index sets petscksp.h - Krylov subspace methods 7 petscviewer.h - viewers petscpc.h - preconditioners 8 9 This version tests correlated terms using both vector and listed forms 10 */ 11 12 #include <petsctao.h> 13 14 /* 15 Description: These data are the result of a NIST study involving 16 ultrasonic calibration. The response variable is 17 ultrasonic response, and the predictor variable is 18 metal distance. 19 20 Reference: Chwirut, D., NIST (197?). 21 Ultrasonic Reference Block Study. 22 */ 23 24 static char help[]="Finds the nonlinear least-squares solution to the model \n\ 25 y = exp[-b1*x]/(b2+b3*x) + e \n"; 26 27 /*T 28 Concepts: TAO^Solving a system of nonlinear equations, nonlinear least squares 29 Routines: TaoCreate(); 30 Routines: TaoSetType(); 31 Routines: TaoSetResidualRoutine(); 32 Routines: TaoSetJacobianRoutine(); 33 Routines: TaoSetSolution(); 34 Routines: TaoSetFromOptions(); 35 Routines: TaoSetConvergenceHistory(); TaoGetConvergenceHistory(); 36 Routines: TaoSolve(); 37 Routines: TaoView(); TaoDestroy(); 38 Processors: 1 39 T*/ 40 41 #define NOBSERVATIONS 214 42 #define NPARAMETERS 3 43 44 /* User-defined application context */ 45 typedef struct { 46 /* Working space */ 47 PetscReal t[NOBSERVATIONS]; /* array of independent variables of observation */ 48 PetscReal y[NOBSERVATIONS]; /* array of dependent variables */ 49 PetscReal j[NOBSERVATIONS][NPARAMETERS]; /* dense jacobian matrix array*/ 50 PetscInt idm[NOBSERVATIONS]; /* Matrix indices for jacobian */ 51 PetscInt idn[NPARAMETERS]; 52 } AppCtx; 53 54 /* User provided Routines */ 55 PetscErrorCode InitializeData(AppCtx *user); 56 PetscErrorCode FormStartingPoint(Vec); 57 PetscErrorCode EvaluateFunction(Tao, Vec, Vec, void *); 58 PetscErrorCode EvaluateJacobian(Tao, Vec, Mat, Mat, void *); 59 60 /*--------------------------------------------------------------------*/ 61 int main(int argc,char **argv) 62 { 63 PetscErrorCode ierr; /* used to check for functions returning nonzeros */ 64 PetscInt wtype=0; 65 Vec x, f; /* solution, function */ 66 Vec w; /* weights */ 67 Mat J; /* Jacobian matrix */ 68 Tao tao; /* Tao solver context */ 69 PetscInt i; /* iteration information */ 70 PetscReal hist[100],resid[100]; 71 PetscInt lits[100]; 72 PetscInt w_row[NOBSERVATIONS]; /* explicit weights */ 73 PetscInt w_col[NOBSERVATIONS]; 74 PetscReal w_vals[NOBSERVATIONS]; 75 PetscBool flg; 76 AppCtx user; /* user-defined work context */ 77 78 ierr = PetscInitialize(&argc,&argv,(char *)0,help);if (ierr) return ierr; 79 CHKERRQ(PetscOptionsGetInt(NULL,NULL,"-wtype",&wtype,&flg)); 80 CHKERRQ(PetscPrintf(PETSC_COMM_WORLD,"wtype=%d\n",wtype)); 81 /* Allocate vectors */ 82 CHKERRQ(VecCreateSeq(MPI_COMM_SELF,NPARAMETERS,&x)); 83 CHKERRQ(VecCreateSeq(MPI_COMM_SELF,NOBSERVATIONS,&f)); 84 85 CHKERRQ(VecDuplicate(f,&w)); 86 87 /* no correlation, but set in different ways */ 88 CHKERRQ(VecSet(w,1.0)); 89 for (i=0;i<NOBSERVATIONS;i++) { 90 w_row[i]=i; w_col[i]=i; w_vals[i]=1.0; 91 } 92 93 /* Create the Jacobian matrix. */ 94 CHKERRQ(MatCreateSeqDense(MPI_COMM_SELF,NOBSERVATIONS,NPARAMETERS,NULL,&J)); 95 96 for (i=0;i<NOBSERVATIONS;i++) user.idm[i] = i; 97 98 for (i=0;i<NPARAMETERS;i++) user.idn[i] = i; 99 100 /* Create TAO solver and set desired solution method */ 101 CHKERRQ(TaoCreate(PETSC_COMM_SELF,&tao)); 102 CHKERRQ(TaoSetType(tao,TAOPOUNDERS)); 103 104 /* Set the function and Jacobian routines. */ 105 CHKERRQ(InitializeData(&user)); 106 CHKERRQ(FormStartingPoint(x)); 107 CHKERRQ(TaoSetSolution(tao,x)); 108 CHKERRQ(TaoSetResidualRoutine(tao,f,EvaluateFunction,(void*)&user)); 109 if (wtype == 1) { 110 CHKERRQ(TaoSetResidualWeights(tao,w,0,NULL,NULL,NULL)); 111 } else if (wtype == 2) { 112 CHKERRQ(TaoSetResidualWeights(tao,NULL,NOBSERVATIONS,w_row,w_col,w_vals)); 113 } 114 CHKERRQ(TaoSetJacobianResidualRoutine(tao, J, J, EvaluateJacobian, (void*)&user)); 115 CHKERRQ(TaoSetTolerances(tao,1e-5,0.0,PETSC_DEFAULT)); 116 117 /* Check for any TAO command line arguments */ 118 CHKERRQ(TaoSetFromOptions(tao)); 119 120 CHKERRQ(TaoSetConvergenceHistory(tao,hist,resid,0,lits,100,PETSC_TRUE)); 121 /* Perform the Solve */ 122 CHKERRQ(TaoSolve(tao)); 123 124 /* Free TAO data structures */ 125 CHKERRQ(TaoDestroy(&tao)); 126 127 /* Free PETSc data structures */ 128 CHKERRQ(VecDestroy(&x)); 129 CHKERRQ(VecDestroy(&w)); 130 CHKERRQ(VecDestroy(&f)); 131 CHKERRQ(MatDestroy(&J)); 132 133 ierr = PetscFinalize(); 134 return ierr; 135 } 136 137 /*--------------------------------------------------------------------*/ 138 PetscErrorCode EvaluateFunction(Tao tao, Vec X, Vec F, void *ptr) 139 { 140 AppCtx *user = (AppCtx *)ptr; 141 PetscInt i; 142 PetscReal *y=user->y,*f,*t=user->t; 143 const PetscReal *x; 144 145 PetscFunctionBegin; 146 CHKERRQ(VecGetArrayRead(X,&x)); 147 CHKERRQ(VecGetArray(F,&f)); 148 149 for (i=0;i<NOBSERVATIONS;i++) { 150 f[i] = y[i] - PetscExpScalar(-x[0]*t[i])/(x[1] + x[2]*t[i]); 151 } 152 CHKERRQ(VecRestoreArrayRead(X,&x)); 153 CHKERRQ(VecRestoreArray(F,&f)); 154 PetscLogFlops(6*NOBSERVATIONS); 155 PetscFunctionReturn(0); 156 } 157 158 /*------------------------------------------------------------*/ 159 /* J[i][j] = df[i]/dt[j] */ 160 PetscErrorCode EvaluateJacobian(Tao tao, Vec X, Mat J, Mat Jpre, void *ptr) 161 { 162 AppCtx *user = (AppCtx *)ptr; 163 PetscInt i; 164 PetscReal *t=user->t; 165 const PetscReal *x; 166 PetscReal base; 167 168 PetscFunctionBegin; 169 CHKERRQ(VecGetArrayRead(X,&x)); 170 for (i=0;i<NOBSERVATIONS;i++) { 171 base = PetscExpScalar(-x[0]*t[i])/(x[1] + x[2]*t[i]); 172 173 user->j[i][0] = t[i]*base; 174 user->j[i][1] = base/(x[1] + x[2]*t[i]); 175 user->j[i][2] = base*t[i]/(x[1] + x[2]*t[i]); 176 } 177 178 /* Assemble the matrix */ 179 CHKERRQ(MatSetValues(J,NOBSERVATIONS,user->idm, NPARAMETERS, user->idn,(PetscReal *)user->j,INSERT_VALUES)); 180 CHKERRQ(MatAssemblyBegin(J,MAT_FINAL_ASSEMBLY)); 181 CHKERRQ(MatAssemblyEnd(J,MAT_FINAL_ASSEMBLY)); 182 183 CHKERRQ(VecRestoreArrayRead(X,&x)); 184 PetscLogFlops(NOBSERVATIONS * 13); 185 PetscFunctionReturn(0); 186 } 187 188 /* ------------------------------------------------------------ */ 189 PetscErrorCode FormStartingPoint(Vec X) 190 { 191 PetscReal *x; 192 193 PetscFunctionBegin; 194 CHKERRQ(VecGetArray(X,&x)); 195 x[0] = 1.19; 196 x[1] = -1.86; 197 x[2] = 1.08; 198 CHKERRQ(VecRestoreArray(X,&x)); 199 PetscFunctionReturn(0); 200 } 201 202 /* ---------------------------------------------------------------------- */ 203 PetscErrorCode InitializeData(AppCtx *user) 204 { 205 PetscReal *t=user->t,*y=user->y; 206 PetscInt i=0; 207 208 PetscFunctionBegin; 209 y[i] = 92.9000; t[i++] = 0.5000; 210 y[i] = 78.7000; t[i++] = 0.6250; 211 y[i] = 64.2000; t[i++] = 0.7500; 212 y[i] = 64.9000; t[i++] = 0.8750; 213 y[i] = 57.1000; t[i++] = 1.0000; 214 y[i] = 43.3000; t[i++] = 1.2500; 215 y[i] = 31.1000; t[i++] = 1.7500; 216 y[i] = 23.6000; t[i++] = 2.2500; 217 y[i] = 31.0500; t[i++] = 1.7500; 218 y[i] = 23.7750; t[i++] = 2.2500; 219 y[i] = 17.7375; t[i++] = 2.7500; 220 y[i] = 13.8000; t[i++] = 3.2500; 221 y[i] = 11.5875; t[i++] = 3.7500; 222 y[i] = 9.4125; t[i++] = 4.2500; 223 y[i] = 7.7250; t[i++] = 4.7500; 224 y[i] = 7.3500; t[i++] = 5.2500; 225 y[i] = 8.0250; t[i++] = 5.7500; 226 y[i] = 90.6000; t[i++] = 0.5000; 227 y[i] = 76.9000; t[i++] = 0.6250; 228 y[i] = 71.6000; t[i++] = 0.7500; 229 y[i] = 63.6000; t[i++] = 0.8750; 230 y[i] = 54.0000; t[i++] = 1.0000; 231 y[i] = 39.2000; t[i++] = 1.2500; 232 y[i] = 29.3000; t[i++] = 1.7500; 233 y[i] = 21.4000; t[i++] = 2.2500; 234 y[i] = 29.1750; t[i++] = 1.7500; 235 y[i] = 22.1250; t[i++] = 2.2500; 236 y[i] = 17.5125; t[i++] = 2.7500; 237 y[i] = 14.2500; t[i++] = 3.2500; 238 y[i] = 9.4500; t[i++] = 3.7500; 239 y[i] = 9.1500; t[i++] = 4.2500; 240 y[i] = 7.9125; t[i++] = 4.7500; 241 y[i] = 8.4750; t[i++] = 5.2500; 242 y[i] = 6.1125; t[i++] = 5.7500; 243 y[i] = 80.0000; t[i++] = 0.5000; 244 y[i] = 79.0000; t[i++] = 0.6250; 245 y[i] = 63.8000; t[i++] = 0.7500; 246 y[i] = 57.2000; t[i++] = 0.8750; 247 y[i] = 53.2000; t[i++] = 1.0000; 248 y[i] = 42.5000; t[i++] = 1.2500; 249 y[i] = 26.8000; t[i++] = 1.7500; 250 y[i] = 20.4000; t[i++] = 2.2500; 251 y[i] = 26.8500; t[i++] = 1.7500; 252 y[i] = 21.0000; t[i++] = 2.2500; 253 y[i] = 16.4625; t[i++] = 2.7500; 254 y[i] = 12.5250; t[i++] = 3.2500; 255 y[i] = 10.5375; t[i++] = 3.7500; 256 y[i] = 8.5875; t[i++] = 4.2500; 257 y[i] = 7.1250; t[i++] = 4.7500; 258 y[i] = 6.1125; t[i++] = 5.2500; 259 y[i] = 5.9625; t[i++] = 5.7500; 260 y[i] = 74.1000; t[i++] = 0.5000; 261 y[i] = 67.3000; t[i++] = 0.6250; 262 y[i] = 60.8000; t[i++] = 0.7500; 263 y[i] = 55.5000; t[i++] = 0.8750; 264 y[i] = 50.3000; t[i++] = 1.0000; 265 y[i] = 41.0000; t[i++] = 1.2500; 266 y[i] = 29.4000; t[i++] = 1.7500; 267 y[i] = 20.4000; t[i++] = 2.2500; 268 y[i] = 29.3625; t[i++] = 1.7500; 269 y[i] = 21.1500; t[i++] = 2.2500; 270 y[i] = 16.7625; t[i++] = 2.7500; 271 y[i] = 13.2000; t[i++] = 3.2500; 272 y[i] = 10.8750; t[i++] = 3.7500; 273 y[i] = 8.1750; t[i++] = 4.2500; 274 y[i] = 7.3500; t[i++] = 4.7500; 275 y[i] = 5.9625; t[i++] = 5.2500; 276 y[i] = 5.6250; t[i++] = 5.7500; 277 y[i] = 81.5000; t[i++] = .5000; 278 y[i] = 62.4000; t[i++] = .7500; 279 y[i] = 32.5000; t[i++] = 1.5000; 280 y[i] = 12.4100; t[i++] = 3.0000; 281 y[i] = 13.1200; t[i++] = 3.0000; 282 y[i] = 15.5600; t[i++] = 3.0000; 283 y[i] = 5.6300; t[i++] = 6.0000; 284 y[i] = 78.0000; t[i++] = .5000; 285 y[i] = 59.9000; t[i++] = .7500; 286 y[i] = 33.2000; t[i++] = 1.5000; 287 y[i] = 13.8400; t[i++] = 3.0000; 288 y[i] = 12.7500; t[i++] = 3.0000; 289 y[i] = 14.6200; t[i++] = 3.0000; 290 y[i] = 3.9400; t[i++] = 6.0000; 291 y[i] = 76.8000; t[i++] = .5000; 292 y[i] = 61.0000; t[i++] = .7500; 293 y[i] = 32.9000; t[i++] = 1.5000; 294 y[i] = 13.8700; t[i++] = 3.0000; 295 y[i] = 11.8100; t[i++] = 3.0000; 296 y[i] = 13.3100; t[i++] = 3.0000; 297 y[i] = 5.4400; t[i++] = 6.0000; 298 y[i] = 78.0000; t[i++] = .5000; 299 y[i] = 63.5000; t[i++] = .7500; 300 y[i] = 33.8000; t[i++] = 1.5000; 301 y[i] = 12.5600; t[i++] = 3.0000; 302 y[i] = 5.6300; t[i++] = 6.0000; 303 y[i] = 12.7500; t[i++] = 3.0000; 304 y[i] = 13.1200; t[i++] = 3.0000; 305 y[i] = 5.4400; t[i++] = 6.0000; 306 y[i] = 76.8000; t[i++] = .5000; 307 y[i] = 60.0000; t[i++] = .7500; 308 y[i] = 47.8000; t[i++] = 1.0000; 309 y[i] = 32.0000; t[i++] = 1.5000; 310 y[i] = 22.2000; t[i++] = 2.0000; 311 y[i] = 22.5700; t[i++] = 2.0000; 312 y[i] = 18.8200; t[i++] = 2.5000; 313 y[i] = 13.9500; t[i++] = 3.0000; 314 y[i] = 11.2500; t[i++] = 4.0000; 315 y[i] = 9.0000; t[i++] = 5.0000; 316 y[i] = 6.6700; t[i++] = 6.0000; 317 y[i] = 75.8000; t[i++] = .5000; 318 y[i] = 62.0000; t[i++] = .7500; 319 y[i] = 48.8000; t[i++] = 1.0000; 320 y[i] = 35.2000; t[i++] = 1.5000; 321 y[i] = 20.0000; t[i++] = 2.0000; 322 y[i] = 20.3200; t[i++] = 2.0000; 323 y[i] = 19.3100; t[i++] = 2.5000; 324 y[i] = 12.7500; t[i++] = 3.0000; 325 y[i] = 10.4200; t[i++] = 4.0000; 326 y[i] = 7.3100; t[i++] = 5.0000; 327 y[i] = 7.4200; t[i++] = 6.0000; 328 y[i] = 70.5000; t[i++] = .5000; 329 y[i] = 59.5000; t[i++] = .7500; 330 y[i] = 48.5000; t[i++] = 1.0000; 331 y[i] = 35.8000; t[i++] = 1.5000; 332 y[i] = 21.0000; t[i++] = 2.0000; 333 y[i] = 21.6700; t[i++] = 2.0000; 334 y[i] = 21.0000; t[i++] = 2.5000; 335 y[i] = 15.6400; t[i++] = 3.0000; 336 y[i] = 8.1700; t[i++] = 4.0000; 337 y[i] = 8.5500; t[i++] = 5.0000; 338 y[i] = 10.1200; t[i++] = 6.0000; 339 y[i] = 78.0000; t[i++] = .5000; 340 y[i] = 66.0000; t[i++] = .6250; 341 y[i] = 62.0000; t[i++] = .7500; 342 y[i] = 58.0000; t[i++] = .8750; 343 y[i] = 47.7000; t[i++] = 1.0000; 344 y[i] = 37.8000; t[i++] = 1.2500; 345 y[i] = 20.2000; t[i++] = 2.2500; 346 y[i] = 21.0700; t[i++] = 2.2500; 347 y[i] = 13.8700; t[i++] = 2.7500; 348 y[i] = 9.6700; t[i++] = 3.2500; 349 y[i] = 7.7600; t[i++] = 3.7500; 350 y[i] = 5.4400; t[i++] = 4.2500; 351 y[i] = 4.8700; t[i++] = 4.7500; 352 y[i] = 4.0100; t[i++] = 5.2500; 353 y[i] = 3.7500; t[i++] = 5.7500; 354 y[i] = 24.1900; t[i++] = 3.0000; 355 y[i] = 25.7600; t[i++] = 3.0000; 356 y[i] = 18.0700; t[i++] = 3.0000; 357 y[i] = 11.8100; t[i++] = 3.0000; 358 y[i] = 12.0700; t[i++] = 3.0000; 359 y[i] = 16.1200; t[i++] = 3.0000; 360 y[i] = 70.8000; t[i++] = .5000; 361 y[i] = 54.7000; t[i++] = .7500; 362 y[i] = 48.0000; t[i++] = 1.0000; 363 y[i] = 39.8000; t[i++] = 1.5000; 364 y[i] = 29.8000; t[i++] = 2.0000; 365 y[i] = 23.7000; t[i++] = 2.5000; 366 y[i] = 29.6200; t[i++] = 2.0000; 367 y[i] = 23.8100; t[i++] = 2.5000; 368 y[i] = 17.7000; t[i++] = 3.0000; 369 y[i] = 11.5500; t[i++] = 4.0000; 370 y[i] = 12.0700; t[i++] = 5.0000; 371 y[i] = 8.7400; t[i++] = 6.0000; 372 y[i] = 80.7000; t[i++] = .5000; 373 y[i] = 61.3000; t[i++] = .7500; 374 y[i] = 47.5000; t[i++] = 1.0000; 375 y[i] = 29.0000; t[i++] = 1.5000; 376 y[i] = 24.0000; t[i++] = 2.0000; 377 y[i] = 17.7000; t[i++] = 2.5000; 378 y[i] = 24.5600; t[i++] = 2.0000; 379 y[i] = 18.6700; t[i++] = 2.5000; 380 y[i] = 16.2400; t[i++] = 3.0000; 381 y[i] = 8.7400; t[i++] = 4.0000; 382 y[i] = 7.8700; t[i++] = 5.0000; 383 y[i] = 8.5100; t[i++] = 6.0000; 384 y[i] = 66.7000; t[i++] = .5000; 385 y[i] = 59.2000; t[i++] = .7500; 386 y[i] = 40.8000; t[i++] = 1.0000; 387 y[i] = 30.7000; t[i++] = 1.5000; 388 y[i] = 25.7000; t[i++] = 2.0000; 389 y[i] = 16.3000; t[i++] = 2.5000; 390 y[i] = 25.9900; t[i++] = 2.0000; 391 y[i] = 16.9500; t[i++] = 2.5000; 392 y[i] = 13.3500; t[i++] = 3.0000; 393 y[i] = 8.6200; t[i++] = 4.0000; 394 y[i] = 7.2000; t[i++] = 5.0000; 395 y[i] = 6.6400; t[i++] = 6.0000; 396 y[i] = 13.6900; t[i++] = 3.0000; 397 y[i] = 81.0000; t[i++] = .5000; 398 y[i] = 64.5000; t[i++] = .7500; 399 y[i] = 35.5000; t[i++] = 1.5000; 400 y[i] = 13.3100; t[i++] = 3.0000; 401 y[i] = 4.8700; t[i++] = 6.0000; 402 y[i] = 12.9400; t[i++] = 3.0000; 403 y[i] = 5.0600; t[i++] = 6.0000; 404 y[i] = 15.1900; t[i++] = 3.0000; 405 y[i] = 14.6200; t[i++] = 3.0000; 406 y[i] = 15.6400; t[i++] = 3.0000; 407 y[i] = 25.5000; t[i++] = 1.7500; 408 y[i] = 25.9500; t[i++] = 1.7500; 409 y[i] = 81.7000; t[i++] = .5000; 410 y[i] = 61.6000; t[i++] = .7500; 411 y[i] = 29.8000; t[i++] = 1.7500; 412 y[i] = 29.8100; t[i++] = 1.7500; 413 y[i] = 17.1700; t[i++] = 2.7500; 414 y[i] = 10.3900; t[i++] = 3.7500; 415 y[i] = 28.4000; t[i++] = 1.7500; 416 y[i] = 28.6900; t[i++] = 1.7500; 417 y[i] = 81.3000; t[i++] = .5000; 418 y[i] = 60.9000; t[i++] = .7500; 419 y[i] = 16.6500; t[i++] = 2.7500; 420 y[i] = 10.0500; t[i++] = 3.7500; 421 y[i] = 28.9000; t[i++] = 1.7500; 422 y[i] = 28.9500; t[i++] = 1.7500; 423 PetscFunctionReturn(0); 424 } 425 426 /*TEST 427 428 build: 429 requires: !complex 430 431 test: 432 args: -tao_smonitor -tao_max_it 100 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5 433 requires: !single 434 TODO: produces different output for many different systems 435 436 test: 437 suffix: 2 438 args: -tao_smonitor -tao_max_it 100 -wtype 1 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5 439 requires: !single 440 TODO: produces different output for many different systems 441 442 test: 443 suffix: 3 444 args: -tao_smonitor -tao_max_it 100 -wtype 2 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5 445 requires: !single 446 TODO: produces different output for many different systems 447 448 test: 449 suffix: 4 450 args: -tao_smonitor -tao_max_it 100 -tao_type pounders -tao_pounders_delta 0.05 -pounders_subsolver_tao_type blmvm -tao_gatol 1.e-5 451 requires: !single 452 TODO: produces different output for many different systems 453 454 TEST*/ 455