Actual source code: epsimpl.h

  1: /*
  2:    - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
  3:    SLEPc - Scalable Library for Eigenvalue Problem Computations
  4:    Copyright (c) 2002-, Universitat Politecnica de Valencia, Spain

  6:    This file is part of SLEPc.
  7:    SLEPc is distributed under a 2-clause BSD license (see LICENSE).
  8:    - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
  9: */

 11: #pragma once

 13: #include <slepceps.h>
 14: #include <slepc/private/bvimpl.h>

 16: /* SUBMANSEC = EPS */

 18: SLEPC_EXTERN PetscBool EPSRegisterAllCalled;
 19: SLEPC_EXTERN PetscBool EPSMonitorRegisterAllCalled;
 20: SLEPC_EXTERN PetscErrorCode EPSRegisterAll(void);
 21: SLEPC_EXTERN PetscErrorCode EPSMonitorRegisterAll(void);
 22: SLEPC_EXTERN PetscLogEvent EPS_SetUp,EPS_Solve,EPS_CISS_SVD;

 24: typedef struct _EPSOps *EPSOps;

 26: struct _EPSOps {
 27:   PetscErrorCode (*solve)(EPS);
 28:   PetscErrorCode (*setup)(EPS);
 29:   PetscErrorCode (*setupsort)(EPS);
 30:   PetscErrorCode (*setfromoptions)(EPS,PetscOptionItems);
 31:   PetscErrorCode (*publishoptions)(EPS);
 32:   PetscErrorCode (*destroy)(EPS);
 33:   PetscErrorCode (*reset)(EPS);
 34:   PetscErrorCode (*view)(EPS,PetscViewer);
 35:   PetscErrorCode (*backtransform)(EPS);
 36:   PetscErrorCode (*computevectors)(EPS);
 37:   PetscErrorCode (*setdefaultst)(EPS);
 38:   PetscErrorCode (*setdstype)(EPS);
 39: };

 41: /*
 42:    Maximum number of monitors you can run with a single EPS
 43: */
 44: #define MAXEPSMONITORS 5

 46: /*
 47:    The solution process goes through several states
 48: */
 49: typedef enum { EPS_STATE_INITIAL,
 50:                EPS_STATE_SETUP,
 51:                EPS_STATE_SOLVED,
 52:                EPS_STATE_EIGENVECTORS } EPSStateType;

 54: /*
 55:    To classify the different solvers into categories
 56: */
 57: typedef enum { EPS_CATEGORY_KRYLOV,      /* Krylov solver: relies on STApply and STBackTransform (same as OTHER) */
 58:                EPS_CATEGORY_PRECOND,     /* Preconditioned solver: uses ST only to manage preconditioner */
 59:                EPS_CATEGORY_CONTOUR,     /* Contour integral: ST used to solve linear systems at integration points */
 60:                EPS_CATEGORY_OTHER } EPSSolverType;

 62: /*
 63:    To check for unsupported features at EPSSetUp_XXX()
 64: */
 65: typedef enum { EPS_FEATURE_BALANCE=1,       /* balancing */
 66:                EPS_FEATURE_ARBITRARY=2,     /* arbitrary selection of eigepairs */
 67:                EPS_FEATURE_REGION=4,        /* nontrivial region for filtering */
 68:                EPS_FEATURE_EXTRACTION=8,    /* extraction technique different from Ritz */
 69:                EPS_FEATURE_CONVERGENCE=16,  /* convergence test selected by user */
 70:                EPS_FEATURE_STOPPING=32,     /* stopping test */
 71:                EPS_FEATURE_THRESHOLD=64,    /* threshold stopping test */
 72:                EPS_FEATURE_TWOSIDED=128     /* two-sided variant */
 73:              } EPSFeatureType;

 75: /*
 76:    Defines the EPS data structure
 77: */
 78: struct _p_EPS {
 79:   PETSCHEADER(struct _EPSOps);
 80:   /*------------------------- User parameters ---------------------------*/
 81:   PetscInt       max_it;           /* maximum number of iterations */
 82:   PetscInt       nev;              /* number of eigenvalues to compute */
 83:   PetscInt       ncv;              /* number of basis vectors */
 84:   PetscInt       mpd;              /* maximum dimension of projected problem */
 85:   PetscInt       nini,ninil;       /* number of initial vectors (negative means not copied yet) */
 86:   PetscInt       nds;              /* number of basis vectors of deflation space */
 87:   PetscScalar    target;           /* target value */
 88:   PetscReal      tol;              /* tolerance */
 89:   PetscReal      thres;            /* threshold */
 90:   PetscBool      threlative;       /* threshold is relative */
 91:   EPSConv        conv;             /* convergence test */
 92:   EPSStop        stop;             /* stopping test */
 93:   EPSWhich       which;            /* which part of the spectrum to be sought */
 94:   PetscReal      inta,intb;        /* interval [a,b] for spectrum slicing */
 95:   EPSProblemType problem_type;     /* which kind of problem to be solved */
 96:   EPSExtraction  extraction;       /* which kind of extraction to be applied */
 97:   EPSBalance     balance;          /* the balancing method */
 98:   PetscInt       balance_its;      /* number of iterations of the balancing method */
 99:   PetscReal      balance_cutoff;   /* cutoff value for balancing */
100:   PetscBool      trueres;          /* whether the true residual norm must be computed */
101:   PetscBool      trackall;         /* whether all the residuals must be computed */
102:   PetscBool      purify;           /* whether eigenvectors need to be purified */
103:   PetscBool      twosided;         /* whether to compute left eigenvectors (two-sided solver) */

105:   /*-------------- User-provided functions and contexts -----------------*/
106:   EPSConvergenceTestFn      *converged;
107:   EPSConvergenceTestFn      *convergeduser;
108:   PetscCtxDestroyFn         *convergeddestroy;
109:   EPSStoppingTestFn         *stopping;
110:   EPSStoppingTestFn         *stoppinguser;
111:   PetscCtxDestroyFn         *stoppingdestroy;
112:   SlepcArbitrarySelectionFn *arbitrary;
113:   PetscCtxDestroyFn         *arbitrarydestroy;
114:   void                      *convergedctx;
115:   void                      *stoppingctx;
116:   void                      *arbitraryctx;
117:   EPSMonitorFn              *monitor[MAXEPSMONITORS];
118:   PetscCtxDestroyFn         *monitordestroy[MAXEPSMONITORS];
119:   void                      *monitorcontext[MAXEPSMONITORS];
120:   PetscInt                  numbermonitors;

122:   /*----------------- Child objects and working data -------------------*/
123:   ST             st;               /* spectral transformation object */
124:   DS             ds;               /* direct solver object */
125:   BV             V;                /* set of basis vectors and computed eigenvectors */
126:   BV             W;                /* left basis vectors (if left eigenvectors requested) */
127:   RG             rg;               /* optional region for filtering */
128:   SlepcSC        sc;               /* sorting criterion data */
129:   Vec            D;                /* diagonal matrix for balancing */
130:   Vec            *IS,*ISL;         /* references to user-provided initial spaces */
131:   Vec            *defl;            /* references to user-provided deflation space */
132:   PetscScalar    *eigr,*eigi;      /* real and imaginary parts of eigenvalues */
133:   PetscReal      *errest;          /* error estimates */
134:   PetscScalar    *rr,*ri;          /* values computed by user's arbitrary selection function */
135:   PetscInt       *perm;            /* permutation for eigenvalue ordering */
136:   PetscInt       nwork;            /* number of work vectors */
137:   Vec            *work;            /* work vectors */
138:   void           *data;            /* placeholder for solver-specific stuff */

140:   /* ----------------------- Status variables --------------------------*/
141:   EPSStateType   state;            /* initial -> setup -> solved -> eigenvectors */
142:   EPSSolverType  categ;            /* solver category */
143:   PetscInt       nconv;            /* number of converged eigenvalues */
144:   PetscInt       its;              /* number of iterations so far computed */
145:   PetscInt       n,nloc;           /* problem dimensions (global, local) */
146:   PetscReal      nrma,nrmb;        /* computed matrix norms */
147:   PetscBool      useds;            /* whether the solver uses the DS object or not */
148:   PetscBool      isgeneralized;
149:   PetscBool      ispositive;
150:   PetscBool      ishermitian;
151:   PetscBool      isstructured;
152:   EPSConvergedReason reason;
153: };

155: /*
156:     Macros to test valid EPS arguments
157: */
158: #if !defined(PETSC_USE_DEBUG)

160: #define EPSCheckSolved(h,arg) do {(void)(h);} while (0)

162: #else

164: #define EPSCheckSolved(h,arg) \
165:   do { \
166:     PetscCheck((h)->state>=EPS_STATE_SOLVED,PetscObjectComm((PetscObject)(h)),PETSC_ERR_ARG_WRONGSTATE,"Must call EPSSolve() first: Parameter #%d",arg); \
167:   } while (0)

169: #endif

171: /*
172:     Macros to check settings at EPSSetUp()
173: */

175: /* EPSCheckHermitianDefinite: the problem is HEP or GHEP */
176: #define EPSCheckHermitianDefiniteCondition(eps,condition,msg) \
177:   do { \
178:     if (condition) { \
179:       PetscCheck((eps)->ishermitian,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s cannot be used for non-%s problems",((PetscObject)(eps))->type_name,(msg),SLEPC_STRING_HERMITIAN); \
180:       PetscCheck(!(eps)->isgeneralized || (eps)->ispositive,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s requires that the problem is %s-definite",((PetscObject)(eps))->type_name,(msg),SLEPC_STRING_HERMITIAN); \
181:     } \
182:   } while (0)
183: #define EPSCheckHermitianDefinite(eps) EPSCheckHermitianDefiniteCondition(eps,PETSC_TRUE,"")

185: /* EPSCheckHermitian: the problem is HEP, GHEP, or GHIEP */
186: #define EPSCheckHermitianCondition(eps,condition,msg) \
187:   do { \
188:     if (condition) { \
189:       PetscCheck((eps)->ishermitian,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s cannot be used for non-%s problems",((PetscObject)(eps))->type_name,(msg),SLEPC_STRING_HERMITIAN); \
190:     } \
191:   } while (0)
192: #define EPSCheckHermitian(eps) EPSCheckHermitianCondition(eps,PETSC_TRUE,"")

194: /* EPSCheckDefinite: the problem is not GHIEP */
195: #define EPSCheckDefiniteCondition(eps,condition,msg) \
196:   do { \
197:     if (condition) { \
198:       PetscCheck(!(eps)->isgeneralized || !(eps)->ishermitian || (eps)->ispositive,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s cannot be used for %s-indefinite problems",((PetscObject)(eps))->type_name,(msg),SLEPC_STRING_HERMITIAN); \
199:     } \
200:   } while (0)
201: #define EPSCheckDefinite(eps) EPSCheckDefiniteCondition(eps,PETSC_TRUE,"")

203: /* EPSCheckStandard: the problem is HEP or NHEP */
204: #define EPSCheckStandardCondition(eps,condition,msg) \
205:   do { \
206:     if (condition) { \
207:       PetscCheck(!(eps)->isgeneralized,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s cannot be used for generalized problems",((PetscObject)(eps))->type_name,(msg)); \
208:     } \
209:   } while (0)
210: #define EPSCheckStandard(eps) EPSCheckStandardCondition(eps,PETSC_TRUE,"")

212: /* EPSCheckNotStructured: the problem is not structured */
213: #define EPSCheckNotStructuredCondition(eps,condition,msg) \
214:   do { \
215:     if (condition) { \
216:       PetscCheck(!(eps)->isstructured,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s does not provide support for structured eigenproblems",((PetscObject)(eps))->type_name,(msg)); \
217:     } \
218:   } while (0)
219: #define EPSCheckNotStructured(eps) EPSCheckNotStructuredCondition(eps,PETSC_TRUE,"")

221: /* EPSCheckSinvert: shift-and-invert ST */
222: #define EPSCheckSinvertCondition(eps,condition,msg) \
223:   do { \
224:     if (condition) { \
225:       PetscBool __flg; \
226:       PetscCall(PetscObjectTypeCompare((PetscObject)(eps)->st,STSINVERT,&__flg)); \
227:       PetscCheck(__flg,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s requires a shift-and-invert spectral transform",((PetscObject)(eps))->type_name,(msg)); \
228:     } \
229:   } while (0)
230: #define EPSCheckSinvert(eps) EPSCheckSinvertCondition(eps,PETSC_TRUE,"")

232: /* EPSCheckSinvertCayley: shift-and-invert or Cayley ST */
233: #define EPSCheckSinvertCayleyCondition(eps,condition,msg) \
234:   do { \
235:     if (condition) { \
236:       PetscBool __flg; \
237:       PetscCall(PetscObjectTypeCompareAny((PetscObject)(eps)->st,&__flg,STSINVERT,STCAYLEY,"")); \
238:       PetscCheck(__flg,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s requires shift-and-invert or Cayley transform",((PetscObject)(eps))->type_name,(msg)); \
239:     } \
240:   } while (0)
241: #define EPSCheckSinvertCayley(eps) EPSCheckSinvertCayleyCondition(eps,PETSC_TRUE,"")

243: /* Check for unsupported features */
244: #define EPSCheckUnsupportedCondition(eps,mask,condition,msg) \
245:   do { \
246:     if (condition) { \
247:       PetscCheck(!((mask) & EPS_FEATURE_BALANCE) || (eps)->balance==EPS_BALANCE_NONE,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s does not support balancing",((PetscObject)(eps))->type_name,(msg)); \
248:       PetscCheck(!((mask) & EPS_FEATURE_ARBITRARY) || !(eps)->arbitrary,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s does not support arbitrary selection of eigenpairs",((PetscObject)(eps))->type_name,(msg)); \
249:       if ((mask) & EPS_FEATURE_REGION) { \
250:         PetscBool      __istrivial; \
251:         PetscCall(RGIsTrivial((eps)->rg,&__istrivial)); \
252:         PetscCheck(__istrivial,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s does not support region filtering",((PetscObject)(eps))->type_name,(msg)); \
253:       } \
254:       PetscCheck(!((mask) & EPS_FEATURE_EXTRACTION) || (eps)->extraction==EPS_RITZ,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s only supports Ritz extraction",((PetscObject)(eps))->type_name,(msg)); \
255:       PetscCheck(!((mask) & EPS_FEATURE_CONVERGENCE) || (eps)->converged==EPSConvergedRelative,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s only supports the default convergence test",((PetscObject)(eps))->type_name,(msg)); \
256:       PetscCheck(!((mask) & EPS_FEATURE_STOPPING) || (eps)->stopping==EPSStoppingBasic,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s only supports the default stopping test",((PetscObject)(eps))->type_name,(msg)); \
257:       PetscCheck(!((mask) & EPS_FEATURE_THRESHOLD) || (eps)->stopping!=EPSStoppingThreshold,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s does not support the threshold stopping test",((PetscObject)(eps))->type_name,(msg)); \
258:       PetscCheck(!((mask) & EPS_FEATURE_TWOSIDED) || !(eps)->twosided,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_SUP,"The solver '%s'%s cannot compute left eigenvectors (no two-sided variant)",((PetscObject)(eps))->type_name,(msg)); \
259:     } \
260:   } while (0)
261: #define EPSCheckUnsupported(eps,mask) EPSCheckUnsupportedCondition(eps,mask,PETSC_TRUE,"")

263: /* Check for ignored features */
264: #define EPSCheckIgnoredCondition(eps,mask,condition,msg) \
265:   do { \
266:     if (condition) { \
267:       if (((mask) & EPS_FEATURE_BALANCE) && (eps)->balance!=EPS_BALANCE_NONE) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the balancing settings\n",((PetscObject)(eps))->type_name,(msg))); \
268:       if (((mask) & EPS_FEATURE_ARBITRARY) && (eps)->arbitrary) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the settings for arbitrary selection of eigenpairs\n",((PetscObject)(eps))->type_name,(msg))); \
269:       if ((mask) & EPS_FEATURE_REGION) { \
270:         PetscBool __istrivial; \
271:         PetscCall(RGIsTrivial((eps)->rg,&__istrivial)); \
272:         if (!__istrivial) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the specified region\n",((PetscObject)(eps))->type_name,(msg))); \
273:       } \
274:       if (((mask) & EPS_FEATURE_EXTRACTION) && (eps)->extraction!=EPS_RITZ) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the extraction settings\n",((PetscObject)(eps))->type_name,(msg))); \
275:       if (((mask) & EPS_FEATURE_CONVERGENCE) && (eps)->converged!=EPSConvergedRelative) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the convergence test settings\n",((PetscObject)(eps))->type_name,(msg))); \
276:       if (((mask) & EPS_FEATURE_STOPPING) && (eps)->stopping!=EPSStoppingBasic) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the stopping test settings\n",((PetscObject)(eps))->type_name,(msg))); \
277:       if (((mask) & EPS_FEATURE_TWOSIDED) && (eps)->twosided) PetscCall(PetscInfo((eps),"The solver '%s'%s ignores the two-sided flag\n",((PetscObject)(eps))->type_name,(msg))); \
278:     } \
279:   } while (0)
280: #define EPSCheckIgnored(eps,mask) EPSCheckIgnoredCondition(eps,mask,PETSC_TRUE,"")

282: /*
283:     EPSSetCtxThreshold - Fills EPSStoppingCtx with data needed for the threshold stopping test

285:     k = number of converged approximations, n = total number of available approximations
286: */
287: #define EPSSetCtxThreshold(eps,eigr,eigi,err_est,k,n) \
288:   do { \
289:     if ((eps)->stop==EPS_STOP_THRESHOLD && k) { \
290:       PetscScalar __kr=(eigr)[k-1],__ki=(eigi)[k-1],__krn=0.0,__kin=0.0,__kr0=(eigr)[0],__ki0=(eigi)[0]; \
291:       PetscCall(STBackTransform((eps)->st,1,&__kr,&__ki)); \
292:       PetscCall(STBackTransform((eps)->st,1,&__kr0,&__ki0)); \
293:       if (n>k) { \
294:         __krn=(eigr)[k];__kin=(eigi)[k]; \
295:         PetscCall(STBackTransform((eps)->st,1,&__krn,&__kin)); \
296:       } \
297:       if ((eps)->which==EPS_LARGEST_MAGNITUDE || (eps)->which==EPS_SMALLEST_MAGNITUDE) { \
298:         ((EPSStoppingCtx)(eps)->stoppingctx)->firstev = SlepcAbsEigenvalue(__kr0,__ki0); \
299:         ((EPSStoppingCtx)(eps)->stoppingctx)->lastev  = SlepcAbsEigenvalue(__kr,__ki); \
300:         ((EPSStoppingCtx)(eps)->stoppingctx)->firstnc = SlepcAbsEigenvalue(__krn,__kin); \
301:       } else { \
302:         ((EPSStoppingCtx)(eps)->stoppingctx)->firstev = PetscRealPart(__kr0); \
303:         ((EPSStoppingCtx)(eps)->stoppingctx)->lastev  = PetscRealPart(__kr); \
304:         ((EPSStoppingCtx)(eps)->stoppingctx)->firstnc = PetscRealPart(__krn); \
305:       } \
306:       ((EPSStoppingCtx)(eps)->stoppingctx)->errest  = (err_est)[k]; \
307:       ((EPSStoppingCtx)(eps)->stoppingctx)->napprox = n; \
308:     } \
309:   } while (0)

311: /*
312:   EPS_SetInnerProduct - set B matrix for inner product if appropriate.
313: */
314: static inline PetscErrorCode EPS_SetInnerProduct(EPS eps)
315: {
316:   Mat            B;

318:   PetscFunctionBegin;
319:   if (!eps->V) PetscCall(EPSGetBV(eps,&eps->V));
320:   if (eps->ispositive || (eps->isgeneralized && eps->ishermitian)) {
321:     PetscCall(STGetBilinearForm(eps->st,&B));
322:     PetscCall(BVSetMatrix(eps->V,B,PetscNot(eps->ispositive)));
323:     if (eps->twosided) PetscCall(BVSetMatrix(eps->W,B,PetscNot(eps->ispositive)));
324:     PetscCall(MatDestroy(&B));
325:   } else PetscCall(BVSetMatrix(eps->V,NULL,PETSC_FALSE));
326:   PetscFunctionReturn(PETSC_SUCCESS);
327: }

329: /*
330:   EPS_Purify - purify the first k vectors in the V basis
331: */
332: static inline PetscErrorCode EPS_Purify(EPS eps,PetscInt k)
333: {
334:   PetscInt       i;
335:   Vec            v,z;

337:   PetscFunctionBegin;
338:   PetscCall(BVCreateVec(eps->V,&v));
339:   for (i=0;i<k;i++) {
340:     PetscCall(BVCopyVec(eps->V,i,v));
341:     PetscCall(BVGetColumn(eps->V,i,&z));
342:     PetscCall(STApply(eps->st,v,z));
343:     PetscCall(BVRestoreColumn(eps->V,i,&z));
344:   }
345:   PetscCall(VecDestroy(&v));
346:   PetscFunctionReturn(PETSC_SUCCESS);
347: }

349: /*
350:   EPS_KSPSetOperators - Sets the KSP matrices, see also ST_KSPSetOperators()
351: */
352: static inline PetscErrorCode EPS_KSPSetOperators(KSP ksp,Mat A,Mat B)
353: {
354:   const char     *prefix;

356:   PetscFunctionBegin;
357:   PetscCall(KSPSetOperators(ksp,A,B));
358:   PetscCall(MatGetOptionsPrefix(B,&prefix));
359:   if (!prefix) {
360:     /* set Mat prefix to be the same as KSP to enable setting command-line options (e.g. MUMPS)
361:        only applies if the Mat has no user-defined prefix */
362:     PetscCall(KSPGetOptionsPrefix(ksp,&prefix));
363:     PetscCall(MatSetOptionsPrefix(B,prefix));
364:   }
365:   PetscFunctionReturn(PETSC_SUCCESS);
366: }

368: /*
369:   EPS_GetActualConverged - Gets the actual value of nconv; in special cases the
370:   number of available eigenvalues is larger than the computed ones
371: */
372: static inline PetscErrorCode EPS_GetActualConverged(EPS eps,PetscInt *nconv)
373: {
374:   PetscFunctionBegin;
375:   *nconv = eps->nconv;
376:   if (eps->isstructured) {
377:     if (eps->problem_type == EPS_BSE && (eps->which == EPS_SMALLEST_MAGNITUDE || eps->which == EPS_LARGEST_MAGNITUDE || eps->which == EPS_TARGET_MAGNITUDE)) *nconv *= 2;
378:     else if (eps->problem_type == EPS_HAMILT && (eps->which == EPS_SMALLEST_MAGNITUDE || eps->which == EPS_LARGEST_MAGNITUDE || eps->which == EPS_TARGET_MAGNITUDE)) *nconv *= 2;
379:     else if (eps->problem_type == EPS_LREP && (eps->which == EPS_SMALLEST_MAGNITUDE || eps->which == EPS_LARGEST_MAGNITUDE || eps->which == EPS_TARGET_MAGNITUDE)) *nconv *= 2;
380:   }
381:   PetscFunctionReturn(PETSC_SUCCESS);
382: }

384: static inline PetscErrorCode EPS_GetEigenvector_BSE(EPS eps,BV V,PetscInt i,Vec Vr,Vec Vi)
385: {
386:   PetscInt  k;
387:   Vec       v0,v1,w,w0,w1;
388:   Mat       H;
389:   IS        is[2];

391:   PetscFunctionBegin;
392:   PetscCheck(eps->which == EPS_SMALLEST_MAGNITUDE || eps->which == EPS_LARGEST_MAGNITUDE || eps->which == EPS_TARGET_MAGNITUDE,PetscObjectComm((PetscObject)(eps)),PETSC_ERR_PLIB,"Inconsistent state");
393:   /* BSE problem, even index is +lambda, odd index is -lambda */
394:   k = eps->perm[i/2];
395:   if (i%2) {
396:     /* eigenvector of -lambda is J*conj(x) where J=[0 I; I 0] and x is eigenvector of lambda */
397:     PetscCall(VecDuplicate(Vr?Vr:Vi,&w));
398:     PetscCall(STGetMatrix(eps->st,0,&H));
399:     PetscCall(MatNestGetISs(H,is,NULL));
400:     if (Vr) {
401:       PetscCall(BV_GetEigenvector(V,k,eps->eigi[k],w,NULL));
402:       PetscCall(VecConjugate(w));
403:       PetscCall(VecGetSubVector(w,is[0],&w0));
404:       PetscCall(VecGetSubVector(w,is[1],&w1));
405:       PetscCall(VecGetSubVector(Vr,is[0],&v0));
406:       PetscCall(VecGetSubVector(Vr,is[1],&v1));
407:       PetscCall(VecCopy(w1,v0));
408:       PetscCall(VecCopy(w0,v1));
409:       PetscCall(VecRestoreSubVector(w,is[0],&w0));
410:       PetscCall(VecRestoreSubVector(w,is[1],&w1));
411:       PetscCall(VecRestoreSubVector(Vr,is[0],&v0));
412:       PetscCall(VecRestoreSubVector(Vr,is[1],&v1));
413:     }
414: #if !defined(PETSC_USE_COMPLEX)
415:     if (Vi) {
416:       PetscCall(BV_GetEigenvector(V,k,eps->eigi[k],NULL,w));
417:       PetscCall(VecScale(w,-1.0));
418:       PetscCall(VecGetSubVector(w,is[0],&w0));
419:       PetscCall(VecGetSubVector(w,is[1],&w1));
420:       PetscCall(VecGetSubVector(Vi,is[0],&v0));
421:       PetscCall(VecGetSubVector(Vi,is[1],&v1));
422:       PetscCall(VecCopy(w1,v0));
423:       PetscCall(VecCopy(w0,v1));
424:       PetscCall(VecRestoreSubVector(w,is[0],&w0));
425:       PetscCall(VecRestoreSubVector(w,is[1],&w1));
426:       PetscCall(VecRestoreSubVector(Vi,is[0],&v0));
427:       PetscCall(VecRestoreSubVector(Vi,is[1],&v1));
428:     }
429: #endif
430:     PetscCall(VecDestroy(&w));
431:   } else {
432:     PetscCall(BV_GetEigenvector(V,k,eps->eigi[k],Vr,Vi));
433:   }
434:   PetscFunctionReturn(PETSC_SUCCESS);
435: }

437: static inline PetscErrorCode EPS_GetEigenvector_HAMILT(EPS eps,BV V,PetscInt i,Vec Vr,Vec Vi)
438: {
439: #if !defined(PETSC_USE_COMPLEX)
440:   PetscInt  k;
441:   Vec       w;
442:   PetscInt  k0,k1,k2,iquad;
443:   PetscReal nrm,nrmr=0.0,nrmi=0.0,sgn;
444: #endif

446:   PetscFunctionBegin;
447: #if !defined(PETSC_USE_COMPLEX)
448:   k = eps->perm[i/2];
449:   if (eps->eigi[k]==0.0) { /* real eigenvalue */
450:     if (Vr) {
451:       PetscCall(BVCopyVec(V,k+eps->ncv/2+1,Vr));
452:       PetscCall(BVGetColumn(V,k,&w));
453:       PetscCall(VecAXPY(Vr,(i%2)?-eps->eigr[k]:eps->eigr[k],w));
454:       PetscCall(BVRestoreColumn(V,k,&w));
455:       PetscCall(VecNorm(Vr,NORM_2,&nrmr));
456:     }
457:     if (Vi) PetscCall(VecZeroEntries(Vi));
458:     nrm = nrmr;
459:   } else if (eps->eigr[k]==0.0 ) { /* purely imaginary eigenvalue */
460:     if (Vr) {
461:       PetscCall(BVCopyVec(V,k+eps->ncv/2+1,Vr));
462:       PetscCall(VecNorm(Vr,NORM_2,&nrmr));
463:     }
464:     if (Vi) {
465:       PetscCall(BVCopyVec(V,k,Vi));
466:       PetscCall(VecScale(Vi,(i%2)?-eps->eigi[k]:eps->eigi[k]));
467:       PetscCall(VecNorm(Vi,NORM_2,&nrmi));
468:     }
469:     nrm = SlepcAbs(nrmr,nrmi);
470:   } else { /* quadruple eigenvalue (-conj(lambda),-lambda,lambda,conj(lambda)) */
471:     iquad = i%2;  /* index within the 4 values */
472:     if (i>=2) {
473:       k2 = eps->perm[(i-2)/2];
474:       if (eps->eigr[k]==eps->eigr[k2] && eps->eigi[k]==-eps->eigi[k2]) iquad += 2;
475:     }
476:     k0 = (iquad<2)? k: k2;
477:     k1 = k0+1;
478:     /* Vr+Vi*i obtained as eig*u+v where u=ur+ui*i is stored in cols k0 (ur) and k1 (ui) and
479:        v=vr+vi*i is in cols shifted by ncv/2+1.
480:        For lambda=eigr+eigi*i:
481:         Vr+Vi*i = (eigr+eigi*i)(ur+ui*i) + vr+vi*i
482:         Vr+Vi*i = eigr*(ur+ui*i) - eigi*ui+eigi*ur*i + vr+vi*i
483:         Vr+Vi*i = eigr*ur-eigi*ui+vr + (eigi*ur+eigr*ui+vi)*i
484:        For -conj(lambda): eigr, ui and vi have the signs changed
485:        For       -lambda: eigr and eigi have the signs changed
486:        For  conj(lambda): eigi, ui and vi have the signs changed   */
487:     if (Vr) {
488:       sgn = (iquad<2)? -1.0: 1.0;
489:       PetscCall(BVCopyVec(V,k0,Vr));                  /* ur */
490:       PetscCall(VecScale(Vr,sgn*eps->eigr[k0]));
491:       PetscCall(BVGetColumn(V,k1,&w));                /* ui */
492:       PetscCall(VecAXPY(Vr,-sgn*eps->eigi[k0],w));
493:       PetscCall(BVRestoreColumn(V,k1,&w));
494:       PetscCall(BVGetColumn(V,k0+eps->ncv/2+1,&w));   /* vr */
495:       PetscCall(VecAXPY(Vr,1.0,w));
496:       PetscCall(BVRestoreColumn(V,k0+eps->ncv/2+1,&w));
497:       PetscCall(VecNorm(Vr,NORM_2,&nrmr));
498:     }
499:     if (Vi) {
500:       sgn = (iquad%2)? -1.0: 1.0;
501:       PetscCall(BVCopyVec(V,k0,Vi));                  /* ur */
502:       PetscCall(VecScale(Vi,sgn*eps->eigi[k0]));
503:       PetscCall(BVGetColumn(V,k1,&w));                /* ui */
504:       PetscCall(VecAXPY(Vi,sgn*eps->eigr[k0],w));
505:       PetscCall(BVRestoreColumn(V,k1,&w));
506:       PetscCall(BVGetColumn(V,k1+eps->ncv/2+1,&w));   /* vi */
507:       sgn = (iquad%3)? 1.0: -1.0;
508:       PetscCall(VecAXPY(Vi,sgn,w));
509:       PetscCall(BVRestoreColumn(V,k1+eps->ncv/2+1,&w));
510:       PetscCall(VecNorm(Vi,NORM_2,&nrmi));
511:     }
512:     nrm = SlepcAbs(nrmr,nrmi);
513:   }
514:   if (Vr) PetscCall(VecScale(Vr,1.0/nrm));
515:   if (Vi) PetscCall(VecScale(Vi,1.0/nrm));
516: #endif
517:   PetscFunctionReturn(PETSC_SUCCESS);
518: }

520: static inline PetscErrorCode EPS_GetEigenvector_LREP(EPS eps,BV V,PetscInt i,Vec Vr,Vec Vi)
521: {
522:   PetscInt  k;
523:   Vec       v;
524:   Mat       H;
525:   IS        is[2];

527:   PetscFunctionBegin;
528:   /* LREP problem, even index is +lambda, odd index is -lambda */
529:   k = eps->perm[i/2];
530:   PetscCall(BV_GetEigenvector(V,k,eps->eigi[k],Vr,Vi));
531:   if (i%2) {
532:     /* eigenvector of -lambda is S*x where S=[I 0; 0 -I] and x is eigenvector of lambda */
533:     PetscCall(STGetMatrix(eps->st,0,&H));
534:     PetscCall(MatNestGetISs(H,is,NULL));
535:     PetscCall(VecGetSubVector(Vr,is[1],&v));
536:     PetscCall(VecScale(v,-1.0));
537:     PetscCall(VecRestoreSubVector(Vr,is[1],&v));
538:   }
539:   PetscFunctionReturn(PETSC_SUCCESS);
540: }

542: /*
543:   EPS_GetEigenvector - Gets the i-th eigenvector taking into account the case
544:   where i exceeds the number of computed vectors (structure-preserving solver).
545:   The argument V should be eps->V for right eigenvectors, eps->W for left ones.
546: */
547: static inline PetscErrorCode EPS_GetEigenvector(EPS eps,BV V,PetscInt i,Vec Vr,Vec Vi)
548: {
549:   PetscInt  k;
550:   PetscBool reduced;
551:   Mat       H;

553:   PetscFunctionBegin;
554:   if (!eps->isstructured) {
555:     k = eps->perm[i];
556:     PetscCall(BV_GetEigenvector(V,k,eps->eigi[k],Vr,Vi));
557:   } else {
558:     switch (eps->problem_type) {
559:       case EPS_BSE:
560:         PetscCall(EPS_GetEigenvector_BSE(eps,V,i,Vr,Vi));
561:         break;
562:       case EPS_HAMILT:
563:         PetscCall(EPS_GetEigenvector_HAMILT(eps,V,i,Vr,Vi));
564:         break;
565:       case EPS_LREP:
566:         PetscCall(STGetMatrix(eps->st,0,&H));
567:         PetscCall(SlepcCheckMatLREPReduced(H,&reduced));
568:         if (reduced) PetscCall(EPS_GetEigenvector_LREP(eps,V,i,Vr,Vi));
569:         else PetscCall(EPS_GetEigenvector_BSE(eps,V,i,Vr,Vi));
570:         break;
571:       default:
572:         SETERRQ(PetscObjectComm((PetscObject)eps),PETSC_ERR_LIB,"Inconsistent state");
573:     }
574:   }
575:   PetscFunctionReturn(PETSC_SUCCESS);
576: }

578: SLEPC_INTERN PetscErrorCode EPSSetWhichEigenpairs_Default(EPS);
579: SLEPC_INTERN PetscErrorCode EPSSetDimensions_Default(EPS,PetscInt*,PetscInt*,PetscInt*);
580: SLEPC_INTERN PetscErrorCode EPSBackTransform_Default(EPS);
581: SLEPC_INTERN PetscErrorCode EPSComputeVectors(EPS);
582: SLEPC_INTERN PetscErrorCode EPSComputeVectors_Hermitian(EPS);
583: SLEPC_INTERN PetscErrorCode EPSComputeVectors_Schur(EPS);
584: SLEPC_INTERN PetscErrorCode EPSComputeVectors_Indefinite(EPS);
585: SLEPC_INTERN PetscErrorCode EPSComputeVectors_Twosided(EPS);
586: SLEPC_INTERN PetscErrorCode EPSComputeVectors_Slice(EPS);
587: SLEPC_INTERN PetscErrorCode EPSComputeResidualNorm_Private(EPS,PetscBool,PetscScalar,PetscScalar,Vec,Vec,Vec*,PetscReal*);
588: SLEPC_INTERN PetscErrorCode EPSComputeRitzVector(EPS,PetscScalar*,PetscScalar*,BV,Vec,Vec);
589: SLEPC_INTERN PetscErrorCode EPSGetStartVector(EPS,PetscInt,PetscBool*);
590: SLEPC_INTERN PetscErrorCode EPSGetLeftStartVector(EPS,PetscInt,PetscBool*);
591: SLEPC_INTERN PetscErrorCode MatEstimateSpectralRange_EPS(Mat,PetscReal*,PetscReal*);

593: /* Private functions of the solver implementations */

595: SLEPC_INTERN PetscErrorCode EPSDelayedArnoldi(EPS,PetscScalar*,PetscInt,PetscInt,PetscInt*,PetscReal*,PetscBool*);
596: SLEPC_INTERN PetscErrorCode EPSDelayedArnoldi1(EPS,PetscScalar*,PetscInt,PetscInt,PetscInt*,PetscReal*,PetscBool*);
597: SLEPC_INTERN PetscErrorCode EPSKrylovConvergence(EPS,PetscBool,PetscInt,PetscInt,PetscReal,PetscReal,PetscReal,PetscInt*);
598: SLEPC_INTERN PetscErrorCode EPSPseudoLanczos(EPS,PetscReal*,PetscReal*,PetscReal*,PetscInt,PetscInt*,PetscBool*,PetscBool*,PetscReal*,Vec);
599: SLEPC_INTERN PetscErrorCode EPSBuildBalance_Krylov(EPS);
600: SLEPC_INTERN PetscErrorCode EPSSetDefaultST(EPS);
601: SLEPC_INTERN PetscErrorCode EPSSetDefaultST_Precond(EPS);
602: SLEPC_INTERN PetscErrorCode EPSSetDefaultST_GMRES(EPS);
603: SLEPC_INTERN PetscErrorCode EPSSetDefaultST_NoFactor(EPS);
604: SLEPC_INTERN PetscErrorCode EPSSetUpSort_Basic(EPS);
605: SLEPC_INTERN PetscErrorCode EPSSetUpSort_Default(EPS);