Actual source code: chwirut2.c
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
9: This version tests correlated terms using both vector and listed forms
10: */
12: #include <petsctao.h>
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.
20: Reference: Chwirut, D., NIST (197?).
21: Ultrasonic Reference Block Study.
22: */
24: static char help[]="Finds the nonlinear least-squares solution to the model \n\
25: y = exp[-b1*x]/(b2+b3*x) + e \n";
27: #define NOBSERVATIONS 214
28: #define NPARAMETERS 3
30: /* User-defined application context */
31: typedef struct {
32: /* Working space */
33: PetscReal t[NOBSERVATIONS]; /* array of independent variables of observation */
34: PetscReal y[NOBSERVATIONS]; /* array of dependent variables */
35: PetscReal j[NOBSERVATIONS][NPARAMETERS]; /* dense jacobian matrix array*/
36: PetscInt idm[NOBSERVATIONS]; /* Matrix indices for jacobian */
37: PetscInt idn[NPARAMETERS];
38: } AppCtx;
40: /* User provided Routines */
41: PetscErrorCode InitializeData(AppCtx *user);
42: PetscErrorCode FormStartingPoint(Vec);
43: PetscErrorCode EvaluateFunction(Tao, Vec, Vec, void *);
44: PetscErrorCode EvaluateJacobian(Tao, Vec, Mat, Mat, void *);
46: /*--------------------------------------------------------------------*/
47: int main(int argc,char **argv)
48: {
49: PetscInt wtype=0;
50: Vec x, f; /* solution, function */
51: Vec w; /* weights */
52: Mat J; /* Jacobian matrix */
53: Tao tao; /* Tao solver context */
54: PetscInt i; /* iteration information */
55: PetscReal hist[100],resid[100];
56: PetscInt lits[100];
57: PetscInt w_row[NOBSERVATIONS]; /* explicit weights */
58: PetscInt w_col[NOBSERVATIONS];
59: PetscReal w_vals[NOBSERVATIONS];
60: PetscBool flg;
61: AppCtx user; /* user-defined work context */
63: PetscInitialize(&argc,&argv,(char *)0,help);
64: PetscOptionsGetInt(NULL,NULL,"-wtype",&wtype,&flg);
65: PetscPrintf(PETSC_COMM_WORLD,"wtype=%d\n",wtype);
66: /* Allocate vectors */
67: VecCreateSeq(MPI_COMM_SELF,NPARAMETERS,&x);
68: VecCreateSeq(MPI_COMM_SELF,NOBSERVATIONS,&f);
70: VecDuplicate(f,&w);
72: /* no correlation, but set in different ways */
73: VecSet(w,1.0);
74: for (i=0;i<NOBSERVATIONS;i++) {
75: w_row[i]=i; w_col[i]=i; w_vals[i]=1.0;
76: }
78: /* Create the Jacobian matrix. */
79: MatCreateSeqDense(MPI_COMM_SELF,NOBSERVATIONS,NPARAMETERS,NULL,&J);
81: for (i=0;i<NOBSERVATIONS;i++) user.idm[i] = i;
83: for (i=0;i<NPARAMETERS;i++) user.idn[i] = i;
85: /* Create TAO solver and set desired solution method */
86: TaoCreate(PETSC_COMM_SELF,&tao);
87: TaoSetType(tao,TAOPOUNDERS);
89: /* Set the function and Jacobian routines. */
90: InitializeData(&user);
91: FormStartingPoint(x);
92: TaoSetSolution(tao,x);
93: TaoSetResidualRoutine(tao,f,EvaluateFunction,(void*)&user);
94: if (wtype == 1) {
95: TaoSetResidualWeights(tao,w,0,NULL,NULL,NULL);
96: } else if (wtype == 2) {
97: TaoSetResidualWeights(tao,NULL,NOBSERVATIONS,w_row,w_col,w_vals);
98: }
99: TaoSetJacobianResidualRoutine(tao, J, J, EvaluateJacobian, (void*)&user);
100: TaoSetTolerances(tao,1e-5,0.0,PETSC_DEFAULT);
102: /* Check for any TAO command line arguments */
103: TaoSetFromOptions(tao);
105: TaoSetConvergenceHistory(tao,hist,resid,0,lits,100,PETSC_TRUE);
106: /* Perform the Solve */
107: TaoSolve(tao);
109: /* Free TAO data structures */
110: TaoDestroy(&tao);
112: /* Free PETSc data structures */
113: VecDestroy(&x);
114: VecDestroy(&w);
115: VecDestroy(&f);
116: MatDestroy(&J);
118: PetscFinalize();
119: return 0;
120: }
122: /*--------------------------------------------------------------------*/
123: PetscErrorCode EvaluateFunction(Tao tao, Vec X, Vec F, void *ptr)
124: {
125: AppCtx *user = (AppCtx *)ptr;
126: PetscInt i;
127: PetscReal *y=user->y,*f,*t=user->t;
128: const PetscReal *x;
130: VecGetArrayRead(X,&x);
131: VecGetArray(F,&f);
133: for (i=0;i<NOBSERVATIONS;i++) {
134: f[i] = y[i] - PetscExpScalar(-x[0]*t[i])/(x[1] + x[2]*t[i]);
135: }
136: VecRestoreArrayRead(X,&x);
137: VecRestoreArray(F,&f);
138: PetscLogFlops(6*NOBSERVATIONS);
139: return 0;
140: }
142: /*------------------------------------------------------------*/
143: /* J[i][j] = df[i]/dt[j] */
144: PetscErrorCode EvaluateJacobian(Tao tao, Vec X, Mat J, Mat Jpre, void *ptr)
145: {
146: AppCtx *user = (AppCtx *)ptr;
147: PetscInt i;
148: PetscReal *t=user->t;
149: const PetscReal *x;
150: PetscReal base;
152: VecGetArrayRead(X,&x);
153: for (i=0;i<NOBSERVATIONS;i++) {
154: base = PetscExpScalar(-x[0]*t[i])/(x[1] + x[2]*t[i]);
156: user->j[i][0] = t[i]*base;
157: user->j[i][1] = base/(x[1] + x[2]*t[i]);
158: user->j[i][2] = base*t[i]/(x[1] + x[2]*t[i]);
159: }
161: /* Assemble the matrix */
162: MatSetValues(J,NOBSERVATIONS,user->idm, NPARAMETERS, user->idn,(PetscReal *)user->j,INSERT_VALUES);
163: MatAssemblyBegin(J,MAT_FINAL_ASSEMBLY);
164: MatAssemblyEnd(J,MAT_FINAL_ASSEMBLY);
166: VecRestoreArrayRead(X,&x);
167: PetscLogFlops(NOBSERVATIONS * 13);
168: return 0;
169: }
171: /* ------------------------------------------------------------ */
172: PetscErrorCode FormStartingPoint(Vec X)
173: {
174: PetscReal *x;
176: VecGetArray(X,&x);
177: x[0] = 1.19;
178: x[1] = -1.86;
179: x[2] = 1.08;
180: VecRestoreArray(X,&x);
181: return 0;
182: }
184: /* ---------------------------------------------------------------------- */
185: PetscErrorCode InitializeData(AppCtx *user)
186: {
187: PetscReal *t=user->t,*y=user->y;
188: PetscInt i=0;
190: y[i] = 92.9000; t[i++] = 0.5000;
191: y[i] = 78.7000; t[i++] = 0.6250;
192: y[i] = 64.2000; t[i++] = 0.7500;
193: y[i] = 64.9000; t[i++] = 0.8750;
194: y[i] = 57.1000; t[i++] = 1.0000;
195: y[i] = 43.3000; t[i++] = 1.2500;
196: y[i] = 31.1000; t[i++] = 1.7500;
197: y[i] = 23.6000; t[i++] = 2.2500;
198: y[i] = 31.0500; t[i++] = 1.7500;
199: y[i] = 23.7750; t[i++] = 2.2500;
200: y[i] = 17.7375; t[i++] = 2.7500;
201: y[i] = 13.8000; t[i++] = 3.2500;
202: y[i] = 11.5875; t[i++] = 3.7500;
203: y[i] = 9.4125; t[i++] = 4.2500;
204: y[i] = 7.7250; t[i++] = 4.7500;
205: y[i] = 7.3500; t[i++] = 5.2500;
206: y[i] = 8.0250; t[i++] = 5.7500;
207: y[i] = 90.6000; t[i++] = 0.5000;
208: y[i] = 76.9000; t[i++] = 0.6250;
209: y[i] = 71.6000; t[i++] = 0.7500;
210: y[i] = 63.6000; t[i++] = 0.8750;
211: y[i] = 54.0000; t[i++] = 1.0000;
212: y[i] = 39.2000; t[i++] = 1.2500;
213: y[i] = 29.3000; t[i++] = 1.7500;
214: y[i] = 21.4000; t[i++] = 2.2500;
215: y[i] = 29.1750; t[i++] = 1.7500;
216: y[i] = 22.1250; t[i++] = 2.2500;
217: y[i] = 17.5125; t[i++] = 2.7500;
218: y[i] = 14.2500; t[i++] = 3.2500;
219: y[i] = 9.4500; t[i++] = 3.7500;
220: y[i] = 9.1500; t[i++] = 4.2500;
221: y[i] = 7.9125; t[i++] = 4.7500;
222: y[i] = 8.4750; t[i++] = 5.2500;
223: y[i] = 6.1125; t[i++] = 5.7500;
224: y[i] = 80.0000; t[i++] = 0.5000;
225: y[i] = 79.0000; t[i++] = 0.6250;
226: y[i] = 63.8000; t[i++] = 0.7500;
227: y[i] = 57.2000; t[i++] = 0.8750;
228: y[i] = 53.2000; t[i++] = 1.0000;
229: y[i] = 42.5000; t[i++] = 1.2500;
230: y[i] = 26.8000; t[i++] = 1.7500;
231: y[i] = 20.4000; t[i++] = 2.2500;
232: y[i] = 26.8500; t[i++] = 1.7500;
233: y[i] = 21.0000; t[i++] = 2.2500;
234: y[i] = 16.4625; t[i++] = 2.7500;
235: y[i] = 12.5250; t[i++] = 3.2500;
236: y[i] = 10.5375; t[i++] = 3.7500;
237: y[i] = 8.5875; t[i++] = 4.2500;
238: y[i] = 7.1250; t[i++] = 4.7500;
239: y[i] = 6.1125; t[i++] = 5.2500;
240: y[i] = 5.9625; t[i++] = 5.7500;
241: y[i] = 74.1000; t[i++] = 0.5000;
242: y[i] = 67.3000; t[i++] = 0.6250;
243: y[i] = 60.8000; t[i++] = 0.7500;
244: y[i] = 55.5000; t[i++] = 0.8750;
245: y[i] = 50.3000; t[i++] = 1.0000;
246: y[i] = 41.0000; t[i++] = 1.2500;
247: y[i] = 29.4000; t[i++] = 1.7500;
248: y[i] = 20.4000; t[i++] = 2.2500;
249: y[i] = 29.3625; t[i++] = 1.7500;
250: y[i] = 21.1500; t[i++] = 2.2500;
251: y[i] = 16.7625; t[i++] = 2.7500;
252: y[i] = 13.2000; t[i++] = 3.2500;
253: y[i] = 10.8750; t[i++] = 3.7500;
254: y[i] = 8.1750; t[i++] = 4.2500;
255: y[i] = 7.3500; t[i++] = 4.7500;
256: y[i] = 5.9625; t[i++] = 5.2500;
257: y[i] = 5.6250; t[i++] = 5.7500;
258: y[i] = 81.5000; t[i++] = .5000;
259: y[i] = 62.4000; t[i++] = .7500;
260: y[i] = 32.5000; t[i++] = 1.5000;
261: y[i] = 12.4100; t[i++] = 3.0000;
262: y[i] = 13.1200; t[i++] = 3.0000;
263: y[i] = 15.5600; t[i++] = 3.0000;
264: y[i] = 5.6300; t[i++] = 6.0000;
265: y[i] = 78.0000; t[i++] = .5000;
266: y[i] = 59.9000; t[i++] = .7500;
267: y[i] = 33.2000; t[i++] = 1.5000;
268: y[i] = 13.8400; t[i++] = 3.0000;
269: y[i] = 12.7500; t[i++] = 3.0000;
270: y[i] = 14.6200; t[i++] = 3.0000;
271: y[i] = 3.9400; t[i++] = 6.0000;
272: y[i] = 76.8000; t[i++] = .5000;
273: y[i] = 61.0000; t[i++] = .7500;
274: y[i] = 32.9000; t[i++] = 1.5000;
275: y[i] = 13.8700; t[i++] = 3.0000;
276: y[i] = 11.8100; t[i++] = 3.0000;
277: y[i] = 13.3100; t[i++] = 3.0000;
278: y[i] = 5.4400; t[i++] = 6.0000;
279: y[i] = 78.0000; t[i++] = .5000;
280: y[i] = 63.5000; t[i++] = .7500;
281: y[i] = 33.8000; t[i++] = 1.5000;
282: y[i] = 12.5600; t[i++] = 3.0000;
283: y[i] = 5.6300; t[i++] = 6.0000;
284: y[i] = 12.7500; t[i++] = 3.0000;
285: y[i] = 13.1200; t[i++] = 3.0000;
286: y[i] = 5.4400; t[i++] = 6.0000;
287: y[i] = 76.8000; t[i++] = .5000;
288: y[i] = 60.0000; t[i++] = .7500;
289: y[i] = 47.8000; t[i++] = 1.0000;
290: y[i] = 32.0000; t[i++] = 1.5000;
291: y[i] = 22.2000; t[i++] = 2.0000;
292: y[i] = 22.5700; t[i++] = 2.0000;
293: y[i] = 18.8200; t[i++] = 2.5000;
294: y[i] = 13.9500; t[i++] = 3.0000;
295: y[i] = 11.2500; t[i++] = 4.0000;
296: y[i] = 9.0000; t[i++] = 5.0000;
297: y[i] = 6.6700; t[i++] = 6.0000;
298: y[i] = 75.8000; t[i++] = .5000;
299: y[i] = 62.0000; t[i++] = .7500;
300: y[i] = 48.8000; t[i++] = 1.0000;
301: y[i] = 35.2000; t[i++] = 1.5000;
302: y[i] = 20.0000; t[i++] = 2.0000;
303: y[i] = 20.3200; t[i++] = 2.0000;
304: y[i] = 19.3100; t[i++] = 2.5000;
305: y[i] = 12.7500; t[i++] = 3.0000;
306: y[i] = 10.4200; t[i++] = 4.0000;
307: y[i] = 7.3100; t[i++] = 5.0000;
308: y[i] = 7.4200; t[i++] = 6.0000;
309: y[i] = 70.5000; t[i++] = .5000;
310: y[i] = 59.5000; t[i++] = .7500;
311: y[i] = 48.5000; t[i++] = 1.0000;
312: y[i] = 35.8000; t[i++] = 1.5000;
313: y[i] = 21.0000; t[i++] = 2.0000;
314: y[i] = 21.6700; t[i++] = 2.0000;
315: y[i] = 21.0000; t[i++] = 2.5000;
316: y[i] = 15.6400; t[i++] = 3.0000;
317: y[i] = 8.1700; t[i++] = 4.0000;
318: y[i] = 8.5500; t[i++] = 5.0000;
319: y[i] = 10.1200; t[i++] = 6.0000;
320: y[i] = 78.0000; t[i++] = .5000;
321: y[i] = 66.0000; t[i++] = .6250;
322: y[i] = 62.0000; t[i++] = .7500;
323: y[i] = 58.0000; t[i++] = .8750;
324: y[i] = 47.7000; t[i++] = 1.0000;
325: y[i] = 37.8000; t[i++] = 1.2500;
326: y[i] = 20.2000; t[i++] = 2.2500;
327: y[i] = 21.0700; t[i++] = 2.2500;
328: y[i] = 13.8700; t[i++] = 2.7500;
329: y[i] = 9.6700; t[i++] = 3.2500;
330: y[i] = 7.7600; t[i++] = 3.7500;
331: y[i] = 5.4400; t[i++] = 4.2500;
332: y[i] = 4.8700; t[i++] = 4.7500;
333: y[i] = 4.0100; t[i++] = 5.2500;
334: y[i] = 3.7500; t[i++] = 5.7500;
335: y[i] = 24.1900; t[i++] = 3.0000;
336: y[i] = 25.7600; t[i++] = 3.0000;
337: y[i] = 18.0700; t[i++] = 3.0000;
338: y[i] = 11.8100; t[i++] = 3.0000;
339: y[i] = 12.0700; t[i++] = 3.0000;
340: y[i] = 16.1200; t[i++] = 3.0000;
341: y[i] = 70.8000; t[i++] = .5000;
342: y[i] = 54.7000; t[i++] = .7500;
343: y[i] = 48.0000; t[i++] = 1.0000;
344: y[i] = 39.8000; t[i++] = 1.5000;
345: y[i] = 29.8000; t[i++] = 2.0000;
346: y[i] = 23.7000; t[i++] = 2.5000;
347: y[i] = 29.6200; t[i++] = 2.0000;
348: y[i] = 23.8100; t[i++] = 2.5000;
349: y[i] = 17.7000; t[i++] = 3.0000;
350: y[i] = 11.5500; t[i++] = 4.0000;
351: y[i] = 12.0700; t[i++] = 5.0000;
352: y[i] = 8.7400; t[i++] = 6.0000;
353: y[i] = 80.7000; t[i++] = .5000;
354: y[i] = 61.3000; t[i++] = .7500;
355: y[i] = 47.5000; t[i++] = 1.0000;
356: y[i] = 29.0000; t[i++] = 1.5000;
357: y[i] = 24.0000; t[i++] = 2.0000;
358: y[i] = 17.7000; t[i++] = 2.5000;
359: y[i] = 24.5600; t[i++] = 2.0000;
360: y[i] = 18.6700; t[i++] = 2.5000;
361: y[i] = 16.2400; t[i++] = 3.0000;
362: y[i] = 8.7400; t[i++] = 4.0000;
363: y[i] = 7.8700; t[i++] = 5.0000;
364: y[i] = 8.5100; t[i++] = 6.0000;
365: y[i] = 66.7000; t[i++] = .5000;
366: y[i] = 59.2000; t[i++] = .7500;
367: y[i] = 40.8000; t[i++] = 1.0000;
368: y[i] = 30.7000; t[i++] = 1.5000;
369: y[i] = 25.7000; t[i++] = 2.0000;
370: y[i] = 16.3000; t[i++] = 2.5000;
371: y[i] = 25.9900; t[i++] = 2.0000;
372: y[i] = 16.9500; t[i++] = 2.5000;
373: y[i] = 13.3500; t[i++] = 3.0000;
374: y[i] = 8.6200; t[i++] = 4.0000;
375: y[i] = 7.2000; t[i++] = 5.0000;
376: y[i] = 6.6400; t[i++] = 6.0000;
377: y[i] = 13.6900; t[i++] = 3.0000;
378: y[i] = 81.0000; t[i++] = .5000;
379: y[i] = 64.5000; t[i++] = .7500;
380: y[i] = 35.5000; t[i++] = 1.5000;
381: y[i] = 13.3100; t[i++] = 3.0000;
382: y[i] = 4.8700; t[i++] = 6.0000;
383: y[i] = 12.9400; t[i++] = 3.0000;
384: y[i] = 5.0600; t[i++] = 6.0000;
385: y[i] = 15.1900; t[i++] = 3.0000;
386: y[i] = 14.6200; t[i++] = 3.0000;
387: y[i] = 15.6400; t[i++] = 3.0000;
388: y[i] = 25.5000; t[i++] = 1.7500;
389: y[i] = 25.9500; t[i++] = 1.7500;
390: y[i] = 81.7000; t[i++] = .5000;
391: y[i] = 61.6000; t[i++] = .7500;
392: y[i] = 29.8000; t[i++] = 1.7500;
393: y[i] = 29.8100; t[i++] = 1.7500;
394: y[i] = 17.1700; t[i++] = 2.7500;
395: y[i] = 10.3900; t[i++] = 3.7500;
396: y[i] = 28.4000; t[i++] = 1.7500;
397: y[i] = 28.6900; t[i++] = 1.7500;
398: y[i] = 81.3000; t[i++] = .5000;
399: y[i] = 60.9000; t[i++] = .7500;
400: y[i] = 16.6500; t[i++] = 2.7500;
401: y[i] = 10.0500; t[i++] = 3.7500;
402: y[i] = 28.9000; t[i++] = 1.7500;
403: y[i] = 28.9500; t[i++] = 1.7500;
404: return 0;
405: }
407: /*TEST
409: build:
410: requires: !complex
412: test:
413: args: -tao_smonitor -tao_max_it 100 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5
414: requires: !single
415: TODO: produces different output for many different systems
417: test:
418: suffix: 2
419: args: -tao_smonitor -tao_max_it 100 -wtype 1 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5
420: requires: !single
421: TODO: produces different output for many different systems
423: test:
424: suffix: 3
425: args: -tao_smonitor -tao_max_it 100 -wtype 2 -tao_type pounders -tao_pounders_delta 0.05 -tao_gatol 1.e-5
426: requires: !single
427: TODO: produces different output for many different systems
429: test:
430: suffix: 4
431: 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
432: requires: !single
433: TODO: produces different output for many different systems
435: TEST*/