| 545 |
dsic.upv.es!jroman |
1 |
/*
|
|
|
2 |
EPS routines related to the solution process.
|
|
|
3 |
*/
|
| 528 |
dsic.upv.es!antodo |
4 |
#include "src/eps/epsimpl.h" /*I "slepceps.h" I*/
|
|
|
5 |
|
|
|
6 |
#undef __FUNCT__
|
|
|
7 |
#define __FUNCT__ "EPSSolve"
|
|
|
8 |
/*@
|
|
|
9 |
EPSSolve - Solves the eigensystem.
|
|
|
10 |
|
|
|
11 |
Collective on EPS
|
|
|
12 |
|
|
|
13 |
Input Parameter:
|
|
|
14 |
. eps - eigensolver context obtained from EPSCreate()
|
|
|
15 |
|
|
|
16 |
Options Database:
|
|
|
17 |
+ -eps_view - print information about the solver used
|
|
|
18 |
. -eps_view_binary - save the matrices to the default binary file
|
|
|
19 |
- -eps_plot_eigs - plot computed eigenvalues
|
|
|
20 |
|
|
|
21 |
Level: beginner
|
|
|
22 |
|
|
|
23 |
.seealso: EPSCreate(), EPSSetUp(), EPSDestroy(), EPSSetTolerances()
|
|
|
24 |
@*/
|
|
|
25 |
PetscErrorCode EPSSolve(EPS eps)
|
|
|
26 |
{
|
|
|
27 |
PetscErrorCode ierr;
|
|
|
28 |
int i;
|
|
|
29 |
PetscReal re,im;
|
|
|
30 |
PetscTruth flg;
|
|
|
31 |
PetscViewer viewer;
|
|
|
32 |
PetscDraw draw;
|
|
|
33 |
PetscDrawSP drawsp;
|
|
|
34 |
|
|
|
35 |
PetscFunctionBegin;
|
|
|
36 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
37 |
|
|
|
38 |
ierr = PetscOptionsHasName(eps->prefix,"-eps_view_binary",&flg);CHKERRQ(ierr);
|
|
|
39 |
if (flg) {
|
|
|
40 |
Mat A,B;
|
|
|
41 |
ierr = STGetOperators(eps->OP,&A,&B);CHKERRQ(ierr);
|
|
|
42 |
ierr = MatView(A,PETSC_VIEWER_BINARY_(eps->comm));CHKERRQ(ierr);
|
|
|
43 |
if (B) ierr = MatView(B,PETSC_VIEWER_BINARY_(eps->comm));CHKERRQ(ierr);
|
|
|
44 |
}
|
|
|
45 |
|
|
|
46 |
/* reset the convergence flag from the previous solves */
|
|
|
47 |
eps->reason = EPS_CONVERGED_ITERATING;
|
|
|
48 |
|
|
|
49 |
if (!eps->setupcalled){ ierr = EPSSetUp(eps);CHKERRQ(ierr); }
|
|
|
50 |
ierr = STResetNumberLinearIterations(eps->OP);
|
|
|
51 |
eps->evecsavailable = PETSC_FALSE;
|
|
|
52 |
ierr = PetscLogEventBegin(EPS_Solve,eps,eps->V[0],eps->V[0],0);CHKERRQ(ierr);
|
|
|
53 |
ierr = STPreSolve(eps->OP,eps);CHKERRQ(ierr);
|
|
|
54 |
ierr = (*eps->ops->solve)(eps);CHKERRQ(ierr);
|
|
|
55 |
ierr = STPostSolve(eps->OP,eps);CHKERRQ(ierr);
|
|
|
56 |
ierr = PetscLogEventEnd(EPS_Solve,eps,eps->V[0],eps->V[0],0);CHKERRQ(ierr);
|
|
|
57 |
if (!eps->reason) {
|
|
|
58 |
SETERRQ(1,"Internal error, solver returned without setting converged reason");
|
|
|
59 |
}
|
|
|
60 |
|
|
|
61 |
/* Map eigenvalues back to the original problem, necessary in some
|
|
|
62 |
* spectral transformations */
|
|
|
63 |
ierr = (*eps->ops->backtransform)(eps);CHKERRQ(ierr);
|
|
|
64 |
|
|
|
65 |
#ifndef PETSC_USE_COMPLEX
|
|
|
66 |
/* reorder conjugate eigenvalues (positive imaginary first) */
|
|
|
67 |
for (i=0; i<eps->nconv-1; i++) {
|
|
|
68 |
PetscScalar minus = -1.0;
|
|
|
69 |
if (eps->eigi[i] != 0) {
|
|
|
70 |
if (eps->eigi[i] < 0) {
|
|
|
71 |
eps->eigi[i] = -eps->eigi[i];
|
|
|
72 |
eps->eigi[i+1] = -eps->eigi[i+1];
|
|
|
73 |
ierr = VecScale(&minus, eps->V[i+1]); CHKERRQ(ierr);
|
|
|
74 |
}
|
|
|
75 |
i++;
|
|
|
76 |
}
|
|
|
77 |
}
|
|
|
78 |
#endif
|
|
|
79 |
|
|
|
80 |
/* sort eigenvalues according to eps->which parameter */
|
|
|
81 |
if (eps->perm) {
|
|
|
82 |
ierr = PetscFree(eps->perm); CHKERRQ(ierr);
|
|
|
83 |
eps->perm = PETSC_NULL;
|
|
|
84 |
}
|
|
|
85 |
if (eps->nconv > 0) {
|
|
|
86 |
ierr = PetscMalloc(sizeof(int)*eps->nconv, &eps->perm); CHKERRQ(ierr);
|
|
|
87 |
ierr = EPSSortEigenvalues(eps->nconv, eps->eigr, eps->eigi, eps->which, eps->nconv, eps->perm); CHKERRQ(ierr);
|
|
|
88 |
}
|
|
|
89 |
|
|
|
90 |
ierr = PetscOptionsHasName(eps->prefix,"-eps_view",&flg);CHKERRQ(ierr);
|
|
|
91 |
if (flg && !PetscPreLoadingOn) { ierr = EPSView(eps,PETSC_VIEWER_STDOUT_WORLD);CHKERRQ(ierr); }
|
|
|
92 |
|
|
|
93 |
ierr = PetscOptionsHasName(eps->prefix,"-eps_plot_eigs",&flg);CHKERRQ(ierr);
|
|
|
94 |
if (flg) {
|
|
|
95 |
ierr = PetscViewerDrawOpen(PETSC_COMM_SELF,0,"Computed Eigenvalues",
|
|
|
96 |
PETSC_DECIDE,PETSC_DECIDE,300,300,&viewer);CHKERRQ(ierr);
|
|
|
97 |
ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr);
|
|
|
98 |
ierr = PetscDrawSPCreate(draw,1,&drawsp);CHKERRQ(ierr);
|
|
|
99 |
for( i=0; i<eps->nconv; i++ ) {
|
|
|
100 |
#if defined(PETSC_USE_COMPLEX)
|
|
|
101 |
re = PetscRealPart(eps->eigr[i]);
|
|
|
102 |
im = PetscImaginaryPart(eps->eigi[i]);
|
|
|
103 |
#else
|
|
|
104 |
re = eps->eigr[i];
|
|
|
105 |
im = eps->eigi[i];
|
|
|
106 |
#endif
|
|
|
107 |
ierr = PetscDrawSPAddPoint(drawsp,&re,&im);CHKERRQ(ierr);
|
|
|
108 |
}
|
|
|
109 |
ierr = PetscDrawSPDraw(drawsp);CHKERRQ(ierr);
|
|
|
110 |
ierr = PetscDrawSPDestroy(drawsp);CHKERRQ(ierr);
|
|
|
111 |
ierr = PetscViewerDestroy(viewer);CHKERRQ(ierr);
|
|
|
112 |
}
|
|
|
113 |
|
|
|
114 |
PetscFunctionReturn(0);
|
|
|
115 |
}
|
|
|
116 |
|
|
|
117 |
#undef __FUNCT__
|
|
|
118 |
#define __FUNCT__ "EPSGetIterationNumber"
|
|
|
119 |
/*@
|
|
|
120 |
EPSGetIterationNumber - Gets the current iteration number. If the
|
|
|
121 |
call to EPSSolve() is complete, then it returns the number of iterations
|
|
|
122 |
carried out by the solution method.
|
|
|
123 |
|
|
|
124 |
Not Collective
|
|
|
125 |
|
|
|
126 |
Input Parameter:
|
|
|
127 |
. eps - the eigensolver context
|
|
|
128 |
|
|
|
129 |
Output Parameter:
|
|
|
130 |
. its - number of iterations
|
|
|
131 |
|
|
|
132 |
Level: intermediate
|
|
|
133 |
|
|
|
134 |
Notes:
|
|
|
135 |
During the i-th iteration this call returns i-1. If EPSSolve() is
|
|
|
136 |
complete, then parameter "its" contains either the iteration number at
|
|
|
137 |
which convergence was successfully reached, or failure was detected.
|
|
|
138 |
Call EPSGetConvergedReason() to determine if the solver converged or
|
|
|
139 |
failed and why.
|
|
|
140 |
|
|
|
141 |
@*/
|
|
|
142 |
PetscErrorCode EPSGetIterationNumber(EPS eps,int *its)
|
|
|
143 |
{
|
|
|
144 |
PetscFunctionBegin;
|
|
|
145 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
146 |
PetscValidIntPointer(its,2);
|
|
|
147 |
*its = eps->its;
|
|
|
148 |
PetscFunctionReturn(0);
|
|
|
149 |
}
|
|
|
150 |
|
|
|
151 |
#undef __FUNCT__
|
|
|
152 |
#define __FUNCT__ "EPSGetNumberLinearIterations"
|
|
|
153 |
/*@
|
|
|
154 |
EPSGetNumberLinearIterations - Gets the total number of iterations
|
|
|
155 |
required by the linear solves associated to the ST object during the
|
|
|
156 |
last EPSSolve() call.
|
|
|
157 |
|
|
|
158 |
Not Collective
|
|
|
159 |
|
|
|
160 |
Input Parameter:
|
|
|
161 |
. eps - EPS context
|
|
|
162 |
|
|
|
163 |
Output Parameter:
|
|
|
164 |
. lits - number of linear iterations
|
|
|
165 |
|
|
|
166 |
Notes:
|
|
|
167 |
When the eigensolver algorithm invokes STApply() then a linear system
|
|
|
168 |
must be solved (except in the case of standard eigenproblems and shift
|
|
|
169 |
transformation). The number of iterations required in this solve is
|
|
|
170 |
accumulated into a counter whose value is returned by this function.
|
|
|
171 |
|
|
|
172 |
The iteration counter is reset to zero at each successive call to EPSSolve().
|
|
|
173 |
|
|
|
174 |
Level: intermediate
|
|
|
175 |
|
|
|
176 |
@*/
|
|
|
177 |
PetscErrorCode EPSGetNumberLinearIterations(EPS eps,int* lits)
|
|
|
178 |
{
|
|
|
179 |
PetscFunctionBegin;
|
|
|
180 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
181 |
PetscValidIntPointer(lits,2);
|
|
|
182 |
STGetNumberLinearIterations(eps->OP, lits);
|
|
|
183 |
PetscFunctionReturn(0);
|
|
|
184 |
}
|
|
|
185 |
|
|
|
186 |
#undef __FUNCT__
|
|
|
187 |
#define __FUNCT__ "EPSGetConverged"
|
|
|
188 |
/*@
|
|
|
189 |
EPSGetConverged - Gets the number of converged eigenpairs.
|
|
|
190 |
|
|
|
191 |
Not Collective
|
|
|
192 |
|
|
|
193 |
Input Parameter:
|
|
|
194 |
. eps - the eigensolver context
|
|
|
195 |
|
|
|
196 |
Output Parameter:
|
|
|
197 |
. nconv - number of converged eigenpairs
|
|
|
198 |
|
|
|
199 |
Note:
|
|
|
200 |
This function should be called after EPSSolve() has finished.
|
|
|
201 |
|
|
|
202 |
Level: beginner
|
|
|
203 |
|
|
|
204 |
.seealso: EPSSetDimensions()
|
|
|
205 |
@*/
|
|
|
206 |
PetscErrorCode EPSGetConverged(EPS eps,int *nconv)
|
|
|
207 |
{
|
|
|
208 |
PetscFunctionBegin;
|
|
|
209 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
210 |
if (nconv) *nconv = eps->nconv;
|
|
|
211 |
PetscFunctionReturn(0);
|
|
|
212 |
}
|
|
|
213 |
|
|
|
214 |
|
|
|
215 |
#undef __FUNCT__
|
|
|
216 |
#define __FUNCT__ "EPSGetConvergedReason"
|
|
|
217 |
/*@C
|
|
|
218 |
EPSGetConvergedReason - Gets the reason why the EPSSolve() iteration was
|
|
|
219 |
stopped.
|
|
|
220 |
|
|
|
221 |
Not Collective
|
|
|
222 |
|
|
|
223 |
Input Parameter:
|
|
|
224 |
. eps - the eigensolver context
|
|
|
225 |
|
|
|
226 |
Output Parameter:
|
|
|
227 |
. reason - negative value indicates diverged, positive value converged
|
|
|
228 |
(see EPSConvergedReason)
|
|
|
229 |
|
|
|
230 |
Possible values for reason:
|
|
|
231 |
+ EPS_CONVERGED_TOL - converged up to tolerance
|
|
|
232 |
. EPS_DIVERGED_ITS - required more than its to reach convergence
|
|
|
233 |
. EPS_DIVERGED_BREAKDOWN - generic breakdown in method
|
|
|
234 |
- EPS_DIVERGED_NONSYMMETRIC - The operator is nonsymmetric
|
|
|
235 |
|
|
|
236 |
Level: intermediate
|
|
|
237 |
|
|
|
238 |
Notes: Can only be called after the call to EPSSolve() is complete.
|
|
|
239 |
|
|
|
240 |
.seealso: EPSSetTolerances(), EPSSolve(), EPSConvergedReason
|
|
|
241 |
@*/
|
|
|
242 |
PetscErrorCode EPSGetConvergedReason(EPS eps,EPSConvergedReason *reason)
|
|
|
243 |
{
|
|
|
244 |
PetscFunctionBegin;
|
|
|
245 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
246 |
*reason = eps->reason;
|
|
|
247 |
PetscFunctionReturn(0);
|
|
|
248 |
}
|
|
|
249 |
|
|
|
250 |
#undef __FUNCT__
|
|
|
251 |
#define __FUNCT__ "EPSGetInvariantSubspace"
|
|
|
252 |
/*@
|
| 761 |
dsic.upv.es!jroman |
253 |
EPSGetInvariantSubspace - Gets an orthonormal basis of the computed invariant
|
|
|
254 |
subspace.
|
| 528 |
dsic.upv.es!antodo |
255 |
|
|
|
256 |
Not Collective
|
|
|
257 |
|
|
|
258 |
Input Parameter:
|
|
|
259 |
. eps - the eigensolver context
|
|
|
260 |
|
|
|
261 |
Output Parameter:
|
|
|
262 |
. v - an array of vectors
|
|
|
263 |
|
|
|
264 |
Notes:
|
|
|
265 |
This function should be called after EPSSolve() has finished.
|
|
|
266 |
|
|
|
267 |
The user should provide in v an array of nconv vectors, where nconv is
|
|
|
268 |
the value returned by EPSGetConverged().
|
|
|
269 |
|
| 761 |
dsic.upv.es!jroman |
270 |
The first k vectors returned in v span an invariant subspace associated
|
|
|
271 |
with the first k computed eigenvalues (note that this is not true if the
|
|
|
272 |
k-th eigenvalue is complex and matrix A is real; in this case the first
|
|
|
273 |
k+1 vectors should be used). An invariant subspace X of A satisfies Ax
|
| 528 |
dsic.upv.es!antodo |
274 |
in X for all x in X (a similar definition applies for generalized
|
|
|
275 |
eigenproblems).
|
|
|
276 |
|
|
|
277 |
Level: intermediate
|
|
|
278 |
|
|
|
279 |
.seealso: EPSGetEigenpair(), EPSGetConverged(), EPSSolve()
|
|
|
280 |
@*/
|
|
|
281 |
PetscErrorCode EPSGetInvariantSubspace(EPS eps, Vec *v)
|
|
|
282 |
{
|
|
|
283 |
PetscErrorCode ierr;
|
|
|
284 |
int i;
|
|
|
285 |
|
|
|
286 |
PetscFunctionBegin;
|
|
|
287 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
288 |
PetscValidHeaderSpecific(v,VEC_COOKIE,3);
|
|
|
289 |
if (!eps->V) {
|
|
|
290 |
SETERRQ(PETSC_ERR_ARG_WRONGSTATE, "EPSSolve must be called first");
|
|
|
291 |
}
|
|
|
292 |
for (i=0;i<eps->nconv;i++) {
|
|
|
293 |
ierr = VecCopy(eps->V[i],v[i]);CHKERRQ(ierr);
|
|
|
294 |
}
|
|
|
295 |
PetscFunctionReturn(0);
|
|
|
296 |
}
|
|
|
297 |
|
|
|
298 |
#undef __FUNCT__
|
|
|
299 |
#define __FUNCT__ "EPSGetEigenpair"
|
|
|
300 |
/*@
|
|
|
301 |
EPSGetEigenpair - Gets the i-th solution of the eigenproblem
|
|
|
302 |
as computed by EPSSolve(). The solution consists in both the eigenvalue
|
| 761 |
dsic.upv.es!jroman |
303 |
and the eigenvector.
|
| 528 |
dsic.upv.es!antodo |
304 |
|
|
|
305 |
Not Collective
|
|
|
306 |
|
|
|
307 |
Input Parameters:
|
|
|
308 |
+ eps - eigensolver context
|
|
|
309 |
- i - index of the solution
|
|
|
310 |
|
|
|
311 |
Output Parameters:
|
|
|
312 |
+ eigr - real part of eigenvalue
|
|
|
313 |
. eigi - imaginary part of eigenvalue
|
|
|
314 |
. Vr - real part of eigenvector
|
|
|
315 |
- Vi - imaginary part of eigenvector
|
|
|
316 |
|
|
|
317 |
Notes:
|
|
|
318 |
If the eigenvalue is real, then eigi and Vi are set to zero. In the
|
|
|
319 |
complex case (e.g. with BOPT=O_complex) the eigenvalue is stored
|
| 761 |
dsic.upv.es!jroman |
320 |
directly in eigr (eigi is set to zero) and the eigenvector in Vr (Vi is
|
| 528 |
dsic.upv.es!antodo |
321 |
set to zero).
|
|
|
322 |
|
|
|
323 |
The index i should be a value between 0 and nconv (see EPSGetConverged()).
|
|
|
324 |
Eigenpairs are indexed according to the ordering criterion established
|
|
|
325 |
with EPSSetWhichEigenpairs().
|
|
|
326 |
|
|
|
327 |
Level: beginner
|
|
|
328 |
|
|
|
329 |
.seealso: EPSSolve(), EPSGetConverged(), EPSSetWhichEigenpairs(),
|
|
|
330 |
EPSGetInvariantSubspace()
|
|
|
331 |
@*/
|
|
|
332 |
PetscErrorCode EPSGetEigenpair(EPS eps, int i, PetscScalar *eigr, PetscScalar *eigi, Vec Vr, Vec Vi)
|
|
|
333 |
{
|
|
|
334 |
PetscErrorCode ierr;
|
|
|
335 |
int k;
|
|
|
336 |
PetscScalar zero = 0.0;
|
|
|
337 |
#ifndef PETSC_USE_COMPLEX
|
|
|
338 |
PetscScalar minus = -1.0;
|
|
|
339 |
#endif
|
|
|
340 |
|
|
|
341 |
PetscFunctionBegin;
|
|
|
342 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
343 |
if (!eps->eigr || !eps->eigi || !eps->V) {
|
|
|
344 |
SETERRQ(PETSC_ERR_ARG_WRONGSTATE, "EPSSolve must be called first");
|
|
|
345 |
}
|
|
|
346 |
if (i<0 || i>=eps->nconv) {
|
|
|
347 |
SETERRQ(PETSC_ERR_ARG_OUTOFRANGE, "Argument 2 out of range");
|
|
|
348 |
}
|
|
|
349 |
if (!eps->evecsavailable && (Vr || Vi) ) {
|
|
|
350 |
ierr = (*eps->ops->computevectors)(eps);CHKERRQ(ierr);
|
|
|
351 |
}
|
|
|
352 |
|
|
|
353 |
if (!eps->perm) k = i;
|
|
|
354 |
else k = eps->perm[i];
|
|
|
355 |
#ifdef PETSC_USE_COMPLEX
|
|
|
356 |
if (eigr) *eigr = eps->eigr[k];
|
|
|
357 |
if (eigi) *eigi = 0;
|
|
|
358 |
if (Vr) { ierr = VecCopy(eps->AV[k], Vr); CHKERRQ(ierr); }
|
|
|
359 |
if (Vi) { ierr = VecSet(&zero, Vi); CHKERRQ(ierr); }
|
|
|
360 |
#else
|
|
|
361 |
if (eigr) *eigr = eps->eigr[k];
|
|
|
362 |
if (eigi) *eigi = eps->eigi[k];
|
|
|
363 |
if (eps->eigi[k] > 0) { /* first value of conjugate pair */
|
|
|
364 |
if (Vr) { ierr = VecCopy(eps->AV[k], Vr); CHKERRQ(ierr); }
|
|
|
365 |
if (Vi) { ierr = VecCopy(eps->AV[k+1], Vi); CHKERRQ(ierr); }
|
|
|
366 |
} else if (eps->eigi[k] < 0) { /* second value of conjugate pair */
|
|
|
367 |
if (Vr) { ierr = VecCopy(eps->AV[k-1], Vr); CHKERRQ(ierr); }
|
|
|
368 |
if (Vi) {
|
|
|
369 |
ierr = VecCopy(eps->AV[k], Vi); CHKERRQ(ierr);
|
|
|
370 |
ierr = VecScale(&minus, Vi); CHKERRQ(ierr);
|
|
|
371 |
}
|
|
|
372 |
} else { /* real eigenvalue */
|
|
|
373 |
if (Vr) { ierr = VecCopy(eps->AV[k], Vr); CHKERRQ(ierr); }
|
|
|
374 |
if (Vi) { ierr = VecSet(&zero, Vi); CHKERRQ(ierr); }
|
|
|
375 |
}
|
|
|
376 |
#endif
|
|
|
377 |
|
|
|
378 |
PetscFunctionReturn(0);
|
|
|
379 |
}
|
|
|
380 |
|
|
|
381 |
#undef __FUNCT__
|
|
|
382 |
#define __FUNCT__ "EPSGetErrorEstimate"
|
|
|
383 |
/*@
|
| 761 |
dsic.upv.es!jroman |
384 |
EPSGetErrorEstimate - Returns the error estimate associated to the i-th
|
|
|
385 |
computed eigenpair.
|
| 528 |
dsic.upv.es!antodo |
386 |
|
|
|
387 |
Not Collective
|
|
|
388 |
|
|
|
389 |
Input Parameter:
|
|
|
390 |
+ eps - eigensolver context
|
|
|
391 |
- i - index of eigenpair
|
|
|
392 |
|
|
|
393 |
Output Parameter:
|
|
|
394 |
. errest - the error estimate
|
|
|
395 |
|
| 761 |
dsic.upv.es!jroman |
396 |
Notes:
|
|
|
397 |
This is the error estimate used internally by the eigensolver. The actual
|
|
|
398 |
error bound can be computed with EPSComputeRelativeError(). See also the user's
|
|
|
399 |
manual for details.
|
|
|
400 |
|
| 528 |
dsic.upv.es!antodo |
401 |
Level: advanced
|
|
|
402 |
|
|
|
403 |
.seealso: EPSComputeRelativeError()
|
|
|
404 |
@*/
|
|
|
405 |
PetscErrorCode EPSGetErrorEstimate(EPS eps, int i, PetscReal *errest)
|
|
|
406 |
{
|
|
|
407 |
PetscFunctionBegin;
|
|
|
408 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
409 |
if (!eps->eigr || !eps->eigi) {
|
|
|
410 |
SETERRQ(PETSC_ERR_ARG_WRONGSTATE, "EPSSolve must be called first");
|
|
|
411 |
}
|
|
|
412 |
if (i<0 || i>=eps->nconv) {
|
|
|
413 |
SETERRQ(PETSC_ERR_ARG_OUTOFRANGE, "Argument 2 out of range");
|
|
|
414 |
}
|
|
|
415 |
if (eps->perm) i = eps->perm[i];
|
|
|
416 |
if (errest) *errest = eps->errest[i];
|
|
|
417 |
PetscFunctionReturn(0);
|
|
|
418 |
}
|
|
|
419 |
|
|
|
420 |
|
|
|
421 |
#undef __FUNCT__
|
|
|
422 |
#define __FUNCT__ "EPSComputeResidualNorm"
|
|
|
423 |
/*@
|
| 761 |
dsic.upv.es!jroman |
424 |
EPSComputeResidualNorm - Computes the norm of the residual vector associated with
|
|
|
425 |
the i-th computed eigenpair.
|
| 528 |
dsic.upv.es!antodo |
426 |
|
|
|
427 |
Collective on EPS
|
|
|
428 |
|
|
|
429 |
Input Parameter:
|
|
|
430 |
. eps - the eigensolver context
|
|
|
431 |
. i - the solution index
|
|
|
432 |
|
|
|
433 |
Output Parameter:
|
| 761 |
dsic.upv.es!jroman |
434 |
. norm - the residual norm, computed as ||Ax-kBx||_2 where k is the
|
| 528 |
dsic.upv.es!antodo |
435 |
eigenvalue and x is the eigenvector.
|
| 761 |
dsic.upv.es!jroman |
436 |
If k=0 then the residual norm is computed as ||Ax||_2.
|
| 528 |
dsic.upv.es!antodo |
437 |
|
|
|
438 |
Notes:
|
|
|
439 |
The index i should be a value between 0 and nconv (see EPSGetConverged()).
|
|
|
440 |
Eigenpairs are indexed according to the ordering criterion established
|
|
|
441 |
with EPSSetWhichEigenpairs().
|
|
|
442 |
|
|
|
443 |
Level: beginner
|
|
|
444 |
|
|
|
445 |
.seealso: EPSSolve(), EPSGetConverged(), EPSSetWhichEigenpairs()
|
|
|
446 |
@*/
|
|
|
447 |
PetscErrorCode EPSComputeResidualNorm(EPS eps, int i, PetscReal *norm)
|
|
|
448 |
{
|
|
|
449 |
PetscErrorCode ierr;
|
|
|
450 |
Vec u, v, w, xr, xi;
|
|
|
451 |
Mat A, B;
|
|
|
452 |
PetscScalar alpha, kr, ki;
|
|
|
453 |
#ifndef PETSC_USE_COMPLEX
|
|
|
454 |
PetscReal ni, nr;
|
|
|
455 |
#endif
|
|
|
456 |
|
|
|
457 |
PetscFunctionBegin;
|
|
|
458 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
459 |
ierr = STGetOperators(eps->OP,&A,&B);CHKERRQ(ierr);
|
|
|
460 |
ierr = VecDuplicate(eps->vec_initial,&u); CHKERRQ(ierr);
|
|
|
461 |
ierr = VecDuplicate(eps->vec_initial,&v); CHKERRQ(ierr);
|
|
|
462 |
ierr = VecDuplicate(eps->vec_initial,&w); CHKERRQ(ierr);
|
|
|
463 |
ierr = VecDuplicate(eps->vec_initial,&xr); CHKERRQ(ierr);
|
|
|
464 |
ierr = VecDuplicate(eps->vec_initial,&xi); CHKERRQ(ierr);
|
|
|
465 |
ierr = EPSGetEigenpair(eps,i,&kr,&ki,xr,xi); CHKERRQ(ierr);
|
|
|
466 |
|
|
|
467 |
#ifndef PETSC_USE_COMPLEX
|
|
|
468 |
if (ki == 0 ||
|
|
|
469 |
PetscAbsScalar(ki) < PetscAbsScalar(kr*PETSC_MACHINE_EPSILON)) {
|
|
|
470 |
#endif
|
|
|
471 |
ierr = MatMult( A, xr, u ); CHKERRQ(ierr); /* u=A*x */
|
|
|
472 |
if (PetscAbsScalar(kr) > PETSC_MACHINE_EPSILON) {
|
|
|
473 |
if (eps->isgeneralized) { ierr = MatMult( B, xr, w ); CHKERRQ(ierr); }
|
|
|
474 |
else { ierr = VecCopy( xr, w ); CHKERRQ(ierr); } /* w=B*x */
|
|
|
475 |
alpha = -kr;
|
|
|
476 |
ierr = VecAXPY( &alpha, w, u ); CHKERRQ(ierr); /* u=A*x-k*B*x */
|
|
|
477 |
}
|
|
|
478 |
ierr = VecNorm( u, NORM_2, norm); CHKERRQ(ierr);
|
|
|
479 |
#ifndef PETSC_USE_COMPLEX
|
|
|
480 |
} else {
|
|
|
481 |
ierr = MatMult( A, xr, u ); CHKERRQ(ierr); /* u=A*xr */
|
|
|
482 |
if (eps->isgeneralized) { ierr = MatMult( B, xr, v ); CHKERRQ(ierr); }
|
|
|
483 |
else { ierr = VecCopy( xr, v ); CHKERRQ(ierr); } /* v=B*xr */
|
|
|
484 |
alpha = -kr;
|
|
|
485 |
ierr = VecAXPY( &alpha, v, u ); CHKERRQ(ierr); /* u=A*xr-kr*B*xr */
|
|
|
486 |
if (eps->isgeneralized) { ierr = MatMult( B, xi, w ); CHKERRQ(ierr); }
|
|
|
487 |
else { ierr = VecCopy( xi, w ); CHKERRQ(ierr); } /* w=B*xi */
|
|
|
488 |
alpha = ki;
|
|
|
489 |
ierr = VecAXPY( &alpha, w, u ); CHKERRQ(ierr); /* u=A*xr-kr*B*xr+ki*B*xi */
|
|
|
490 |
ierr = VecNorm( u, NORM_2, &nr ); CHKERRQ(ierr);
|
|
|
491 |
ierr = MatMult( A, xi, u ); CHKERRQ(ierr); /* u=A*xi */
|
|
|
492 |
alpha = -kr;
|
|
|
493 |
ierr = VecAXPY( &alpha, w, u ); CHKERRQ(ierr); /* u=A*xi-kr*B*xi */
|
|
|
494 |
alpha = -ki;
|
|
|
495 |
ierr = VecAXPY( &alpha, v, u ); CHKERRQ(ierr); /* u=A*xi-kr*B*xi-ki*B*xr */
|
|
|
496 |
ierr = VecNorm( u, NORM_2, &ni ); CHKERRQ(ierr);
|
|
|
497 |
*norm = SlepcAbsEigenvalue( nr, ni );
|
|
|
498 |
}
|
|
|
499 |
#endif
|
|
|
500 |
|
|
|
501 |
ierr = VecDestroy(w); CHKERRQ(ierr);
|
|
|
502 |
ierr = VecDestroy(v); CHKERRQ(ierr);
|
|
|
503 |
ierr = VecDestroy(u); CHKERRQ(ierr);
|
|
|
504 |
ierr = VecDestroy(xr); CHKERRQ(ierr);
|
|
|
505 |
ierr = VecDestroy(xi); CHKERRQ(ierr);
|
|
|
506 |
PetscFunctionReturn(0);
|
|
|
507 |
}
|
|
|
508 |
|
|
|
509 |
#undef __FUNCT__
|
|
|
510 |
#define __FUNCT__ "EPSComputeRelativeError"
|
|
|
511 |
/*@
|
| 761 |
dsic.upv.es!jroman |
512 |
EPSComputeRelativeError - Computes the relative error bound associated
|
|
|
513 |
with the i-th computed eigenpair.
|
| 528 |
dsic.upv.es!antodo |
514 |
|
|
|
515 |
Collective on EPS
|
|
|
516 |
|
|
|
517 |
Input Parameter:
|
|
|
518 |
. eps - the eigensolver context
|
|
|
519 |
. i - the solution index
|
|
|
520 |
|
|
|
521 |
Output Parameter:
|
| 761 |
dsic.upv.es!jroman |
522 |
. error - the relative error bound, computed as ||Ax-kBx||_2/||kx||_2 where
|
|
|
523 |
k is the eigenvalue and x is the eigenvector.
|
|
|
524 |
If k=0 the relative error is computed as ||Ax||_2/||x||_2.
|
| 528 |
dsic.upv.es!antodo |
525 |
|
|
|
526 |
Level: beginner
|
|
|
527 |
|
| 761 |
dsic.upv.es!jroman |
528 |
.seealso: EPSSolve(), EPSComputeResidualNorm(), EPSGetErrorEstimate()
|
| 528 |
dsic.upv.es!antodo |
529 |
@*/
|
|
|
530 |
PetscErrorCode EPSComputeRelativeError(EPS eps, int i, PetscReal *error)
|
|
|
531 |
{
|
|
|
532 |
PetscErrorCode ierr;
|
|
|
533 |
Vec xr, xi;
|
|
|
534 |
PetscScalar kr, ki;
|
|
|
535 |
PetscReal norm, er;
|
|
|
536 |
#ifndef PETSC_USE_COMPLEX
|
|
|
537 |
Vec u;
|
|
|
538 |
PetscScalar alpha;
|
|
|
539 |
PetscReal ei;
|
|
|
540 |
#endif
|
|
|
541 |
|
|
|
542 |
PetscFunctionBegin;
|
|
|
543 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
544 |
ierr = EPSComputeResidualNorm(eps,i,&norm); CHKERRQ(ierr);
|
|
|
545 |
ierr = VecDuplicate(eps->vec_initial,&xr); CHKERRQ(ierr);
|
|
|
546 |
ierr = VecDuplicate(eps->vec_initial,&xi); CHKERRQ(ierr);
|
|
|
547 |
ierr = EPSGetEigenpair(eps,i,&kr,&ki,xr,xi); CHKERRQ(ierr);
|
|
|
548 |
|
|
|
549 |
#ifndef PETSC_USE_COMPLEX
|
|
|
550 |
if (ki == 0 ||
|
|
|
551 |
PetscAbsScalar(ki) < PetscAbsScalar(kr*PETSC_MACHINE_EPSILON)) {
|
|
|
552 |
#endif
|
|
|
553 |
if (PetscAbsScalar(kr) > PETSC_MACHINE_EPSILON) {
|
|
|
554 |
ierr = VecScale(&kr, xr); CHKERRQ(ierr);
|
|
|
555 |
}
|
|
|
556 |
ierr = VecNorm(xr, NORM_2, &er); CHKERRQ(ierr);
|
|
|
557 |
*error = norm / er;
|
|
|
558 |
#ifndef PETSC_USE_COMPLEX
|
|
|
559 |
} else {
|
|
|
560 |
ierr = VecDuplicate(xi, &u); CHKERRQ(ierr);
|
|
|
561 |
ierr = VecCopy(xi, u); CHKERRQ(ierr);
|
|
|
562 |
alpha = -ki;
|
|
|
563 |
ierr = VecAXPBY(&kr, &alpha, xr, u); CHKERRQ(ierr);
|
|
|
564 |
ierr = VecNorm(u, NORM_2, &er); CHKERRQ(ierr);
|
|
|
565 |
ierr = VecAXPBY(&kr, &ki, xr, xi); CHKERRQ(ierr);
|
|
|
566 |
ierr = VecNorm(xi, NORM_2, &ei); CHKERRQ(ierr);
|
|
|
567 |
ierr = VecDestroy(u); CHKERRQ(ierr);
|
|
|
568 |
*error = norm / SlepcAbsEigenvalue(er, ei);
|
|
|
569 |
}
|
|
|
570 |
#endif
|
|
|
571 |
|
|
|
572 |
ierr = VecDestroy(xr); CHKERRQ(ierr);
|
|
|
573 |
ierr = VecDestroy(xi); CHKERRQ(ierr);
|
|
|
574 |
PetscFunctionReturn(0);
|
|
|
575 |
}
|
|
|
576 |
|
|
|
577 |
#undef __FUNCT__
|
|
|
578 |
#define __FUNCT__ "EPSReverseProjection"
|
|
|
579 |
/*@
|
|
|
580 |
EPSReverseProjection - Compute the operation V=V*S, where the columns of
|
|
|
581 |
V are m of the basis vectors of the EPS object and S is an mxm dense
|
|
|
582 |
matrix.
|
|
|
583 |
|
|
|
584 |
Collective on EPS
|
|
|
585 |
|
|
|
586 |
Input Parameter:
|
|
|
587 |
+ eps - the eigenproblem solver context
|
|
|
588 |
. V - basis vectors
|
|
|
589 |
. S - pointer to the values of matrix S
|
|
|
590 |
. k - starting column
|
|
|
591 |
. m - dimension of matrix S
|
|
|
592 |
- work - workarea of m vectors for intermediate results
|
|
|
593 |
|
|
|
594 |
Level: developer
|
|
|
595 |
|
|
|
596 |
@*/
|
|
|
597 |
PetscErrorCode EPSReverseProjection(EPS eps,Vec* V,PetscScalar *S,int k,int m,Vec* work)
|
|
|
598 |
{
|
|
|
599 |
PetscErrorCode ierr;
|
|
|
600 |
int i;
|
|
|
601 |
PetscScalar zero = 0.0;
|
|
|
602 |
|
|
|
603 |
PetscFunctionBegin;
|
| 756 |
dsic.upv.es!antodo |
604 |
ierr = PetscLogEventBegin(EPS_ReverseProjection,eps,0,0,0);CHKERRQ(ierr);
|
| 528 |
dsic.upv.es!antodo |
605 |
for (i=k;i<m;i++) {
|
|
|
606 |
ierr = VecSet(&zero,work[i]);CHKERRQ(ierr);
|
|
|
607 |
ierr = VecMAXPY(m,S+m*i,work[i],V);CHKERRQ(ierr);
|
|
|
608 |
}
|
|
|
609 |
for (i=k;i<m;i++) {
|
|
|
610 |
ierr = VecCopy(work[i],V[i]);CHKERRQ(ierr);
|
|
|
611 |
}
|
| 756 |
dsic.upv.es!antodo |
612 |
ierr = PetscLogEventEnd(EPS_ReverseProjection,eps,0,0,0);CHKERRQ(ierr);
|
| 528 |
dsic.upv.es!antodo |
613 |
PetscFunctionReturn(0);
|
|
|
614 |
}
|
|
|
615 |
|
| 677 |
dsic.upv.es!antodo |
616 |
#define SWAP(a,b,t) {t=a;a=b;b=t;}
|
|
|
617 |
|
| 528 |
dsic.upv.es!antodo |
618 |
#undef __FUNCT__
|
| 677 |
dsic.upv.es!antodo |
619 |
#define __FUNCT__ "EPSSortEigenvalues"
|
| 528 |
dsic.upv.es!antodo |
620 |
/*@
|
| 677 |
dsic.upv.es!antodo |
621 |
EPSSortEigenvalues - Sorts a list of eigenvalues according to a certain
|
|
|
622 |
criterion.
|
| 528 |
dsic.upv.es!antodo |
623 |
|
| 677 |
dsic.upv.es!antodo |
624 |
Not Collective
|
| 528 |
dsic.upv.es!antodo |
625 |
|
| 677 |
dsic.upv.es!antodo |
626 |
Input Parameters:
|
|
|
627 |
+ n - number of eigenvalue in the list
|
|
|
628 |
. eig - pointer to the array containing the eigenvalues
|
|
|
629 |
. eigi - imaginary part of the eigenvalues (only when using real numbers)
|
|
|
630 |
. which - sorting criterion
|
|
|
631 |
- nev - number of wanted eigenvalues
|
| 528 |
dsic.upv.es!antodo |
632 |
|
| 677 |
dsic.upv.es!antodo |
633 |
Output Parameter:
|
|
|
634 |
. permout - resulting permutation
|
| 528 |
dsic.upv.es!antodo |
635 |
|
| 677 |
dsic.upv.es!antodo |
636 |
Notes:
|
|
|
637 |
The result is a list of indices in the original eigenvalue array
|
|
|
638 |
corresponding to the first nev eigenvalues sorted in the specified
|
|
|
639 |
criterion
|
| 528 |
dsic.upv.es!antodo |
640 |
|
| 677 |
dsic.upv.es!antodo |
641 |
Level: developer
|
| 528 |
dsic.upv.es!antodo |
642 |
|
| 677 |
dsic.upv.es!antodo |
643 |
.seealso: EPSDenseNHEPSorted(), EPSSetWhichEigenpairs()
|
| 528 |
dsic.upv.es!antodo |
644 |
@*/
|
| 677 |
dsic.upv.es!antodo |
645 |
PetscErrorCode EPSSortEigenvalues(int n,PetscScalar *eig,PetscScalar *eigi,EPSWhich which,int nev,int *permout)
|
| 528 |
dsic.upv.es!antodo |
646 |
{
|
|
|
647 |
PetscErrorCode ierr;
|
| 677 |
dsic.upv.es!antodo |
648 |
int i,*perm;
|
|
|
649 |
PetscReal *values;
|
| 528 |
dsic.upv.es!antodo |
650 |
|
|
|
651 |
PetscFunctionBegin;
|
| 677 |
dsic.upv.es!antodo |
652 |
ierr = PetscMalloc(n*sizeof(int),&perm);CHKERRQ(ierr);
|
|
|
653 |
ierr = PetscMalloc(n*sizeof(PetscReal),&values);CHKERRQ(ierr);
|
|
|
654 |
for (i=0; i<n; i++) { perm[i] = i;}
|
| 528 |
dsic.upv.es!antodo |
655 |
|
| 677 |
dsic.upv.es!antodo |
656 |
switch(which) {
|
|
|
657 |
case EPS_LARGEST_MAGNITUDE:
|
|
|
658 |
case EPS_SMALLEST_MAGNITUDE:
|
|
|
659 |
for (i=0; i<n; i++) { values[i] = SlepcAbsEigenvalue(eig[i],eigi[i]); }
|
|
|
660 |
break;
|
|
|
661 |
case EPS_LARGEST_REAL:
|
|
|
662 |
case EPS_SMALLEST_REAL:
|
|
|
663 |
for (i=0; i<n; i++) { values[i] = PetscRealPart(eig[i]); }
|
|
|
664 |
break;
|
|
|
665 |
case EPS_LARGEST_IMAGINARY:
|
|
|
666 |
case EPS_SMALLEST_IMAGINARY:
|
|
|
667 |
#if defined(PETSC_USE_COMPLEX)
|
|
|
668 |
for (i=0; i<n; i++) { values[i] = PetscImaginaryPart(eig[i]); }
|
|
|
669 |
#else
|
|
|
670 |
for (i=0; i<n; i++) { values[i] = PetscAbsReal(eigi[i]); }
|
|
|
671 |
#endif
|
|
|
672 |
break;
|
|
|
673 |
default: SETERRQ(1,"Wrong value of which");
|
|
|
674 |
}
|
| 528 |
dsic.upv.es!antodo |
675 |
|
| 677 |
dsic.upv.es!antodo |
676 |
ierr = PetscSortRealWithPermutation(n,values,perm);CHKERRQ(ierr);
|
| 528 |
dsic.upv.es!antodo |
677 |
|
| 677 |
dsic.upv.es!antodo |
678 |
switch(which) {
|
|
|
679 |
case EPS_LARGEST_MAGNITUDE:
|
|
|
680 |
case EPS_LARGEST_REAL:
|
|
|
681 |
case EPS_LARGEST_IMAGINARY:
|
|
|
682 |
for (i=0; i<nev; i++) { permout[i] = perm[n-1-i]; }
|
|
|
683 |
break;
|
|
|
684 |
case EPS_SMALLEST_MAGNITUDE:
|
|
|
685 |
case EPS_SMALLEST_REAL:
|
|
|
686 |
case EPS_SMALLEST_IMAGINARY:
|
|
|
687 |
for (i=0; i<nev; i++) { permout[i] = perm[i]; }
|
|
|
688 |
break;
|
|
|
689 |
default: SETERRQ(1,"Wrong value of which");
|
| 528 |
dsic.upv.es!antodo |
690 |
}
|
|
|
691 |
|
| 677 |
dsic.upv.es!antodo |
692 |
#if !defined(PETSC_USE_COMPLEX)
|
|
|
693 |
for (i=0; i<nev-1; i++) {
|
|
|
694 |
if (eigi[permout[i]] != 0.0) {
|
|
|
695 |
if (eig[permout[i]] == eig[permout[i+1]] &&
|
|
|
696 |
eigi[permout[i]] == -eigi[permout[i+1]] &&
|
|
|
697 |
eigi[permout[i]] < 0.0) {
|
|
|
698 |
int tmp;
|
|
|
699 |
SWAP(permout[i], permout[i+1], tmp);
|
|
|
700 |
}
|
|
|
701 |
i++;
|
|
|
702 |
}
|
|
|
703 |
}
|
|
|
704 |
#endif
|
| 528 |
dsic.upv.es!antodo |
705 |
|
| 677 |
dsic.upv.es!antodo |
706 |
ierr = PetscFree(values);CHKERRQ(ierr);
|
|
|
707 |
ierr = PetscFree(perm);CHKERRQ(ierr);
|
| 528 |
dsic.upv.es!antodo |
708 |
PetscFunctionReturn(0);
|
|
|
709 |
}
|
| 689 |
dsic.upv.es!jroman |
710 |
|
|
|
711 |
#undef __FUNCT__
|
|
|
712 |
#define __FUNCT__ "EPSGetStartVector"
|
|
|
713 |
/*@
|
|
|
714 |
EPSGetStartVector - Gets a vector to be used as the starting vector
|
|
|
715 |
in an Arnoldi or Lanczos reduction.
|
|
|
716 |
|
|
|
717 |
Collective on EPS and Vec
|
|
|
718 |
|
|
|
719 |
Input Parameters:
|
|
|
720 |
+ eps - the eigensolver context
|
|
|
721 |
- i - index of the Arnoldi/Lanczos step
|
|
|
722 |
|
|
|
723 |
Input/Output Parameter:
|
|
|
724 |
. vec - the start vector
|
|
|
725 |
|
|
|
726 |
Input/Output Parameter:
|
|
|
727 |
. norm - the start vector
|
|
|
728 |
|
|
|
729 |
Notes:
|
|
|
730 |
The start vector is computed from another vector: for the first step (i=0),
|
|
|
731 |
the initial vector is used (see EPSGetInitialVector()); otherwise a random
|
|
|
732 |
vector is created. Then this vector is forced to be in the range of OP and
|
|
|
733 |
orthonormalized with respect to all V-vectors up to i-1.
|
|
|
734 |
|
|
|
735 |
The caller must pass a vector already allocated with dimensions conforming
|
|
|
736 |
to the initial vector. This vector is overwritten.
|
|
|
737 |
|
|
|
738 |
Level: developer
|
|
|
739 |
|
|
|
740 |
.seealso: EPSGetInitialVector()
|
|
|
741 |
|
|
|
742 |
@*/
|
|
|
743 |
PetscErrorCode EPSGetStartVector(EPS eps,int i,Vec vec)
|
|
|
744 |
{
|
|
|
745 |
PetscErrorCode ierr;
|
|
|
746 |
PetscTruth breakdown;
|
|
|
747 |
PetscScalar t;
|
|
|
748 |
PetscReal norm;
|
|
|
749 |
Vec w;
|
|
|
750 |
|
|
|
751 |
PetscFunctionBegin;
|
|
|
752 |
PetscValidHeaderSpecific(eps,EPS_COOKIE,1);
|
|
|
753 |
PetscValidHeaderSpecific(vec,VEC_COOKIE,3);
|
|
|
754 |
|
|
|
755 |
/* For the first step, use the initial vector, otherwise a random one */
|
|
|
756 |
if (i==0) {
|
|
|
757 |
w = eps->vec_initial;
|
|
|
758 |
}
|
|
|
759 |
else {
|
|
|
760 |
ierr = VecDuplicate(eps->vec_initial,&w);CHKERRQ(ierr);
|
|
|
761 |
ierr = SlepcVecSetRandom(w);CHKERRQ(ierr);
|
|
|
762 |
}
|
|
|
763 |
|
|
|
764 |
/* Force the vector to be in the range of OP */
|
|
|
765 |
ierr = STApply(eps->OP,w,vec);CHKERRQ(ierr);
|
|
|
766 |
|
|
|
767 |
/* Orthonormalize the vector with respect to previous vectors */
|
|
|
768 |
ierr = EPSOrthogonalize(eps,i+eps->nds,eps->DSV,vec,PETSC_NULL,&norm,&breakdown);CHKERRQ(ierr);
|
| 750 |
dsic.upv.es!antodo |
769 |
if (breakdown) {
|
|
|
770 |
if (i==0) { SETERRQ(1,"Initial vector is zero or belongs to the deflation space"); }
|
|
|
771 |
else { SETERRQ(1,"Unable to generate more start vectors"); }
|
|
|
772 |
}
|
| 689 |
dsic.upv.es!jroman |
773 |
t = 1 / norm;
|
|
|
774 |
ierr = VecScale(&t,vec);CHKERRQ(ierr);
|
|
|
775 |
|
|
|
776 |
if (i!=0) {
|
|
|
777 |
ierr = VecDestroy(w);CHKERRQ(ierr);
|
|
|
778 |
}
|
|
|
779 |
|
|
|
780 |
PetscFunctionReturn(0);
|
|
|
781 |
}
|
|
|
782 |
|