Actual source code: klu.c

petsc-3.7.7 2017-09-25
Report Typos and Errors
  2: /*
  3:    Provides an interface to the KLUv1.2 sparse solver

  5:    When build with PETSC_USE_64BIT_INDICES this will use SuiteSparse_long as the
  6:    integer type in KLU, otherwise it will use int. This means
  7:    all integers in this file are simply declared as PetscInt. Also it means
  8:    that KLU SuiteSparse_long version MUST be built with 64 bit integers when used.

 10: */
 11: #include <../src/mat/impls/aij/seq/aij.h>

 13: #if defined(PETSC_USE_64BIT_INDICES)
 14: #define klu_K_defaults                   klu_l_defaults
 15: #define klu_K_analyze(a,b,c,d)           klu_l_analyze((SuiteSparse_long)a,(SuiteSparse_long*)b,(SuiteSparse_long*)c,d)
 16: #define klu_K_analyze_given(a,b,c,d,e,f) klu_l_analyze_given((SuiteSparse_long)a,(SuiteSparse_long*)b,(SuiteSparse_long*)c,(SuiteSparse_long*)d,(SuiteSparse_long*)e,f)
 17: #define klu_K_free_symbolic              klu_l_free_symbolic
 18: #define klu_K_free_numeric               klu_l_free_numeric
 19: #define klu_K_common                     klu_l_common
 20: #define klu_K_symbolic                   klu_l_symbolic
 21: #define klu_K_numeric                    klu_l_numeric
 22: #if defined(PETSC_USE_COMPLEX)
 23: #define klu_K_factor(a,b,c,d,e)       klu_zl_factor((SuiteSparse_long*)a,(SuiteSparse_long*)b,c,d,e);
 24: #define klu_K_solve                   klu_zl_solve
 25: #define klu_K_tsolve                  klu_zl_tsolve
 26: #define klu_K_refactor                klu_zl_refactor
 27: #define klu_K_sort                    klu_zl_sort
 28: #define klu_K_flops                   klu_zl_flops
 29: #define klu_K_rgrowth                 klu_zl_rgrowth
 30: #define klu_K_condest                 klu_zl_condest
 31: #define klu_K_rcond                   klu_zl_rcond
 32: #define klu_K_scale                   klu_zl_scale
 33: #else
 34: #define klu_K_factor(a,b,c,d,e)       klu_l_factor((SuiteSparse_long*)a,(SuiteSparse_long*)b,c,d,e);
 35: #define klu_K_solve                   klu_l_solve
 36: #define klu_K_tsolve                  klu_l_tsolve
 37: #define klu_K_refactor                klu_l_refactor
 38: #define klu_K_sort                    klu_l_sort
 39: #define klu_K_flops                   klu_l_flops
 40: #define klu_K_rgrowth                 klu_l_rgrowth
 41: #define klu_K_condest                 klu_l_condest
 42: #define klu_K_rcond                   klu_l_rcond
 43: #define klu_K_scale                   klu_l_scale
 44: #endif
 45: #else
 46: #define klu_K_defaults                klu_defaults
 47: #define klu_K_analyze                 klu_analyze
 48: #define klu_K_analyze_given           klu_analyze_given
 49: #define klu_K_free_symbolic           klu_free_symbolic
 50: #define klu_K_free_numeric            klu_free_numeric
 51: #define klu_K_common                  klu_common
 52: #define klu_K_symbolic                klu_symbolic
 53: #define klu_K_numeric                 klu_numeric
 54: #if defined(PETSC_USE_COMPLEX)
 55: #define klu_K_factor                  klu_z_factor
 56: #define klu_K_solve                   klu_z_solve
 57: #define klu_K_tsolve                  klu_z_tsolve
 58: #define klu_K_refactor                klu_z_refactor
 59: #define klu_K_sort                    klu_z_sort
 60: #define klu_K_flops                   klu_z_flops
 61: #define klu_K_rgrowth                 klu_z_rgrowth
 62: #define klu_K_condest                 klu_z_condest
 63: #define klu_K_rcond                   klu_z_rcond
 64: #define klu_K_scale                   klu_z_scale
 65: #else
 66: #define klu_K_factor                  klu_factor
 67: #define klu_K_solve                   klu_solve
 68: #define klu_K_tsolve                  klu_tsolve
 69: #define klu_K_refactor                klu_refactor
 70: #define klu_K_sort                    klu_sort
 71: #define klu_K_flops                   klu_flops
 72: #define klu_K_rgrowth                 klu_rgrowth
 73: #define klu_K_condest                 klu_condest
 74: #define klu_K_rcond                   klu_rcond
 75: #define klu_K_scale                   klu_scale
 76: #endif
 77: #endif


 80: EXTERN_C_BEGIN
 81: #include <klu.h>
 82: EXTERN_C_END

 84: static const char *KluOrderingTypes[] = {"AMD","COLAMD","PETSC"};
 85: static const char *scale[] ={"NONE","SUM","MAX"};

 87: typedef struct {
 88:   klu_K_common   Common;
 89:   klu_K_symbolic *Symbolic;
 90:   klu_K_numeric  *Numeric;
 91:   PetscInt     *perm_c,*perm_r;
 92:   MatStructure flg;
 93:   PetscBool    PetscMatOrdering;

 95:   /* Flag to clean up KLU objects during Destroy */
 96:   PetscBool CleanUpKLU;
 97: } Mat_KLU;

101: static PetscErrorCode MatDestroy_KLU(Mat A)
102: {
104:   Mat_KLU    *lu=(Mat_KLU*)A->spptr;

107:   if (lu && lu->CleanUpKLU) {
108:     klu_K_free_symbolic(&lu->Symbolic,&lu->Common);
109:     klu_K_free_numeric(&lu->Numeric,&lu->Common);
110:     PetscFree2(lu->perm_r,lu->perm_c);
111:   }
112:   PetscFree(A->spptr);
113:   MatDestroy_SeqAIJ(A);
114:   return(0);
115: }

119: static PetscErrorCode MatSolveTranspose_KLU(Mat A,Vec b,Vec x)
120: {
121:   Mat_KLU       *lu = (Mat_KLU*)A->spptr;
122:   PetscScalar    *xa;
124:   PetscInt       status;

127:   /* KLU uses a column major format, solve Ax = b by klu_*_solve */
128:   /* ----------------------------------*/
129:   VecCopy(b,x); /* klu_solve stores the solution in rhs */
130:   VecGetArray(x,&xa);
131:   status = klu_K_solve(lu->Symbolic,lu->Numeric,A->rmap->n,1,(PetscReal*)xa,&lu->Common);
132:   if (status != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Solve failed");
133:   VecRestoreArray(x,&xa);
134:   return(0);
135: }

139: static PetscErrorCode MatSolve_KLU(Mat A,Vec b,Vec x)
140: {
141:   Mat_KLU       *lu = (Mat_KLU*)A->spptr;
142:   PetscScalar    *xa;
144:   PetscInt       status;

147:   /* KLU uses a column major format, solve A^Tx = b by klu_*_tsolve */
148:   /* ----------------------------------*/
149:   VecCopy(b,x); /* klu_solve stores the solution in rhs */
150:   VecGetArray(x,&xa);
151: #if defined(PETSC_USE_COMPLEX)
152:   PetscInt conj_solve=1;
153:   status = klu_K_tsolve(lu->Symbolic,lu->Numeric,A->rmap->n,1,(PetscReal*)xa,conj_solve,&lu->Common); /* conjugate solve */
154: #else
155:   status = klu_K_tsolve(lu->Symbolic,lu->Numeric,A->rmap->n,1,xa,&lu->Common);
156: #endif  
157:   if (status != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Solve failed");
158:   VecRestoreArray(x,&xa);
159:   return(0);
160: }

164: static PetscErrorCode MatLUFactorNumeric_KLU(Mat F,Mat A,const MatFactorInfo *info)
165: {
166:   Mat_KLU        *lu = (Mat_KLU*)(F)->spptr;
167:   Mat_SeqAIJ     *a  = (Mat_SeqAIJ*)A->data;
168:   PetscInt       *ai = a->i,*aj=a->j;
169:   PetscScalar    *av = a->a;

172:   /* numeric factorization of A' */
173:   /* ----------------------------*/

175:   if (lu->flg == SAME_NONZERO_PATTERN && lu->Numeric) {
176:     klu_K_free_numeric(&lu->Numeric,&lu->Common);
177:   }
178:   lu->Numeric = klu_K_factor(ai,aj,(PetscReal*)av,lu->Symbolic,&lu->Common);
179:   if(!lu->Numeric) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Numeric factorization failed");

181:   lu->flg                = SAME_NONZERO_PATTERN;
182:   lu->CleanUpKLU         = PETSC_TRUE;
183:   F->ops->solve          = MatSolve_KLU;
184:   F->ops->solvetranspose = MatSolveTranspose_KLU;
185:   return(0);
186: }

190: static PetscErrorCode MatLUFactorSymbolic_KLU(Mat F,Mat A,IS r,IS c,const MatFactorInfo *info)
191: {
192:   Mat_SeqAIJ     *a  = (Mat_SeqAIJ*)A->data;
193:   Mat_KLU       *lu = (Mat_KLU*)(F->spptr);
195:   PetscInt       i,*ai = a->i,*aj = a->j,m=A->rmap->n,n=A->cmap->n;
196:   const PetscInt *ra,*ca;

199:   if (lu->PetscMatOrdering) {
200:     ISGetIndices(r,&ra);
201:     ISGetIndices(c,&ca);
202:     PetscMalloc2(m,&lu->perm_r,n,&lu->perm_c);
203:     /* we cannot simply memcpy on 64 bit archs */
204:     for (i = 0; i < m; i++) lu->perm_r[i] = ra[i];
205:     for (i = 0; i < n; i++) lu->perm_c[i] = ca[i];
206:     ISRestoreIndices(r,&ra);
207:     ISRestoreIndices(c,&ca);
208:   }

210:   /* symbolic factorization of A' */
211:   /* ---------------------------------------------------------------------- */
212:   if (lu->PetscMatOrdering) { /* use Petsc ordering */
213:     lu->Symbolic = klu_K_analyze_given(n,ai,aj,lu->perm_c,lu->perm_r,&lu->Common);
214:   } else { /* use klu internal ordering */
215:     lu->Symbolic = klu_K_analyze(n,ai,aj,&lu->Common);
216:   }
217:   if (!lu->Symbolic) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Symbolic Factorization failed");

219:   lu->flg                   = DIFFERENT_NONZERO_PATTERN;
220:   lu->CleanUpKLU            = PETSC_TRUE;
221:   (F)->ops->lufactornumeric = MatLUFactorNumeric_KLU;
222:   return(0);
223: }

227: static PetscErrorCode MatFactorInfo_KLU(Mat A,PetscViewer viewer)
228: {
229:   Mat_KLU       *lu= (Mat_KLU*)A->spptr;
230:   klu_K_numeric *Numeric=(klu_K_numeric*)lu->Numeric;

234:   /* check if matrix is KLU type */
235:   if (A->ops->solve != MatSolve_KLU) return(0);

237:   PetscViewerASCIIPrintf(viewer,"KLU stats:\n");
238:   PetscViewerASCIIPrintf(viewer,"  Number of diagonal blocks: %d\n",Numeric->nblocks);
239:   PetscViewerASCIIPrintf(viewer,"  Total nonzeros=%d\n",Numeric->lnz+Numeric->unz);

241:   PetscViewerASCIIPrintf(viewer,"KLU runtime parameters:\n");

243:   /* Control parameters used by numeric factorization */
244:   PetscViewerASCIIPrintf(viewer,"  Partial pivoting tolerance: %g\n",lu->Common.tol);
245:   /* BTF preordering */
246:   PetscViewerASCIIPrintf(viewer,"  BTF preordering enabled: %d\n",lu->Common.btf);
247:   /* mat ordering */
248:   if (!lu->PetscMatOrdering) {
249:     PetscViewerASCIIPrintf(viewer,"  Ordering: %s (not using the PETSc ordering)\n",KluOrderingTypes[(int)lu->Common.ordering]);
250:   } else {
251:     PetscViewerASCIIPrintf(viewer,"  Using PETSc ordering\n");
252:   }
253:   /* matrix row scaling */
254:   PetscViewerASCIIPrintf(viewer, "  Matrix row scaling: %s\n",scale[(int)lu->Common.scale]);
255:   return(0);
256: }

260: static PetscErrorCode MatView_KLU(Mat A,PetscViewer viewer)
261: {
262:   PetscErrorCode    ierr;
263:   PetscBool         iascii;
264:   PetscViewerFormat format;

267:   MatView_SeqAIJ(A,viewer);

269:   PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);
270:   if (iascii) {
271:     PetscViewerGetFormat(viewer,&format);
272:     if (format == PETSC_VIEWER_ASCII_INFO) {
273:       MatFactorInfo_KLU(A,viewer);
274:     }
275:   }
276:   return(0);
277: }

281: PetscErrorCode MatFactorGetSolverPackage_seqaij_klu(Mat A,const MatSolverPackage *type)
282: {
284:   *type = MATSOLVERKLU;
285:   return(0);
286: }


289: /*MC
290:   MATSOLVERKLU = "klu" - A matrix type providing direct solvers (LU) for sequential matrices
291:   via the external package KLU.

293:   ./configure --download-suitesparse to install PETSc to use KLU

295:   Use -pc_type lu -pc_factor_mat_solver_package klu to us this direct solver

297:   Consult KLU documentation for more information on the options database keys below.

299:   Options Database Keys:
300: + -mat_klu_pivot_tol <0.001>                  - Partial pivoting tolerance
301: . -mat_klu_use_btf <1>                        - Use BTF preordering
302: . -mat_klu_ordering <AMD>                     - KLU reordering scheme to reduce fill-in (choose one of) AMD COLAMD PETSC
303: - -mat_klu_row_scale <NONE>                   - Matrix row scaling (choose one of) NONE SUM MAX 

305:    Note: KLU is part of SuiteSparse http://faculty.cse.tamu.edu/davis/suitesparse.html

307:    Level: beginner

309: .seealso: PCLU, MATSOLVERUMFPACK, MATSOLVERCHOLMOD, PCFactorSetMatSolverPackage(), MatSolverPackage
310: M*/

314: PETSC_INTERN PetscErrorCode MatGetFactor_seqaij_klu(Mat A,MatFactorType ftype,Mat *F)
315: {
316:   Mat            B;
317:   Mat_KLU       *lu;
319:   PetscInt       m=A->rmap->n,n=A->cmap->n,idx,status;
320:   PetscBool      flg;

323:   /* Create the factorization matrix F */
324:   MatCreate(PetscObjectComm((PetscObject)A),&B);
325:   MatSetSizes(B,PETSC_DECIDE,PETSC_DECIDE,m,n);
326:   MatSetType(B,((PetscObject)A)->type_name);
327:   MatSeqAIJSetPreallocation(B,0,NULL);
328:   PetscNewLog(B,&lu);

330:   B->spptr                 = lu;
331:   B->ops->lufactorsymbolic = MatLUFactorSymbolic_KLU;
332:   B->ops->destroy          = MatDestroy_KLU;
333:   B->ops->view             = MatView_KLU;

335:   PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverPackage_C",MatFactorGetSolverPackage_seqaij_klu);

337:   B->factortype   = MAT_FACTOR_LU;
338:   B->assembled    = PETSC_TRUE;           /* required by -ksp_view */
339:   B->preallocated = PETSC_TRUE;

341:   PetscFree(B->solvertype);
342:   PetscStrallocpy(MATSOLVERKLU,&B->solvertype);

344:   /* initializations */
345:   /* ------------------------------------------------*/
346:   /* get the default control parameters */
347:   status = klu_K_defaults(&lu->Common);
348:   if(status <= 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Initialization failed");

350:   lu->Common.scale = 0; /* No row scaling */

352:   PetscOptionsBegin(PetscObjectComm((PetscObject)A),((PetscObject)A)->prefix,"KLU Options","Mat");
353:   /* Partial pivoting tolerance */
354:   PetscOptionsReal("-mat_klu_pivot_tol","Partial pivoting tolerance","None",lu->Common.tol,&lu->Common.tol,NULL);
355:   /* BTF pre-ordering */
356:   PetscOptionsInt("-mat_klu_use_btf","Enable BTF preordering","None",(PetscInt)lu->Common.btf,(PetscInt*)&lu->Common.btf,NULL);
357:   /* Matrix reordering */
358:   PetscOptionsEList("-mat_klu_ordering","Internal ordering method","None",KluOrderingTypes,sizeof(KluOrderingTypes)/sizeof(KluOrderingTypes[0]),KluOrderingTypes[0],&idx,&flg);
359:   if (flg) {
360:     if ((int)idx == 2) lu->PetscMatOrdering = PETSC_TRUE;   /* use Petsc mat ordering (note: size is for the transpose, and PETSc r = Klu perm_c) */
361:     else lu->Common.ordering = (int)idx;
362:   }
363:   /* Matrix row scaling */
364:   PetscOptionsEList("-mat_klu_row_scale","Matrix row scaling","None",scale,3,scale[0],&idx,&flg);
365:   PetscOptionsEnd();
366:   *F = B;
367:   return(0);
368: }