1 2 /* 3 Provides an interface to the KLUv1.2 sparse solver 4 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. 9 10 */ 11 #include <../src/mat/impls/aij/seq/aij.h> 12 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 78 79 EXTERN_C_BEGIN 80 #include <klu.h> 81 EXTERN_C_END 82 83 static const char *KluOrderingTypes[] = {"AMD","COLAMD"}; 84 static const char *scale[] ={"NONE","SUM","MAX"}; 85 86 typedef struct { 87 klu_K_common Common; 88 klu_K_symbolic *Symbolic; 89 klu_K_numeric *Numeric; 90 PetscInt *perm_c,*perm_r; 91 MatStructure flg; 92 PetscBool PetscMatOrdering; 93 PetscBool CleanUpKLU; 94 } Mat_KLU; 95 96 static PetscErrorCode MatDestroy_KLU(Mat A) 97 { 98 Mat_KLU *lu=(Mat_KLU*)A->data; 99 100 PetscFunctionBegin; 101 if (lu->CleanUpKLU) { 102 klu_K_free_symbolic(&lu->Symbolic,&lu->Common); 103 klu_K_free_numeric(&lu->Numeric,&lu->Common); 104 PetscCall(PetscFree2(lu->perm_r,lu->perm_c)); 105 } 106 PetscCall(PetscFree(A->data)); 107 PetscFunctionReturn(0); 108 } 109 110 static PetscErrorCode MatSolveTranspose_KLU(Mat A,Vec b,Vec x) 111 { 112 Mat_KLU *lu = (Mat_KLU*)A->data; 113 PetscScalar *xa; 114 PetscInt status; 115 116 PetscFunctionBegin; 117 /* KLU uses a column major format, solve Ax = b by klu_*_solve */ 118 /* ----------------------------------*/ 119 PetscCall(VecCopy(b,x)); /* klu_solve stores the solution in rhs */ 120 PetscCall(VecGetArray(x,&xa)); 121 status = klu_K_solve(lu->Symbolic,lu->Numeric,A->rmap->n,1,(PetscReal*)xa,&lu->Common); 122 PetscCheck(status == 1,PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Solve failed"); 123 PetscCall(VecRestoreArray(x,&xa)); 124 PetscFunctionReturn(0); 125 } 126 127 static PetscErrorCode MatSolve_KLU(Mat A,Vec b,Vec x) 128 { 129 Mat_KLU *lu = (Mat_KLU*)A->data; 130 PetscScalar *xa; 131 PetscInt status; 132 133 PetscFunctionBegin; 134 /* KLU uses a column major format, solve A^Tx = b by klu_*_tsolve */ 135 /* ----------------------------------*/ 136 PetscCall(VecCopy(b,x)); /* klu_solve stores the solution in rhs */ 137 PetscCall(VecGetArray(x,&xa)); 138 #if defined(PETSC_USE_COMPLEX) 139 PetscInt conj_solve=1; 140 status = klu_K_tsolve(lu->Symbolic,lu->Numeric,A->rmap->n,1,(PetscReal*)xa,conj_solve,&lu->Common); /* conjugate solve */ 141 #else 142 status = klu_K_tsolve(lu->Symbolic,lu->Numeric,A->rmap->n,1,xa,&lu->Common); 143 #endif 144 PetscCheck(status == 1,PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Solve failed"); 145 PetscCall(VecRestoreArray(x,&xa)); 146 PetscFunctionReturn(0); 147 } 148 149 static PetscErrorCode MatLUFactorNumeric_KLU(Mat F,Mat A,const MatFactorInfo *info) 150 { 151 Mat_KLU *lu = (Mat_KLU*)(F)->data; 152 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 153 PetscInt *ai = a->i,*aj=a->j; 154 PetscScalar *av = a->a; 155 156 PetscFunctionBegin; 157 /* numeric factorization of A' */ 158 /* ----------------------------*/ 159 160 if (lu->flg == SAME_NONZERO_PATTERN && lu->Numeric) { 161 klu_K_free_numeric(&lu->Numeric,&lu->Common); 162 } 163 lu->Numeric = klu_K_factor(ai,aj,(PetscReal*)av,lu->Symbolic,&lu->Common); 164 PetscCheck(lu->Numeric,PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Numeric factorization failed"); 165 166 lu->flg = SAME_NONZERO_PATTERN; 167 lu->CleanUpKLU = PETSC_TRUE; 168 F->ops->solve = MatSolve_KLU; 169 F->ops->solvetranspose = MatSolveTranspose_KLU; 170 PetscFunctionReturn(0); 171 } 172 173 static PetscErrorCode MatLUFactorSymbolic_KLU(Mat F,Mat A,IS r,IS c,const MatFactorInfo *info) 174 { 175 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 176 Mat_KLU *lu = (Mat_KLU*)(F->data); 177 PetscInt i,*ai = a->i,*aj = a->j,m=A->rmap->n,n=A->cmap->n; 178 const PetscInt *ra,*ca; 179 180 PetscFunctionBegin; 181 if (lu->PetscMatOrdering) { 182 PetscCall(ISGetIndices(r,&ra)); 183 PetscCall(ISGetIndices(c,&ca)); 184 PetscCall(PetscMalloc2(m,&lu->perm_r,n,&lu->perm_c)); 185 /* we cannot simply memcpy on 64 bit archs */ 186 for (i = 0; i < m; i++) lu->perm_r[i] = ra[i]; 187 for (i = 0; i < n; i++) lu->perm_c[i] = ca[i]; 188 PetscCall(ISRestoreIndices(r,&ra)); 189 PetscCall(ISRestoreIndices(c,&ca)); 190 } 191 192 /* symbolic factorization of A' */ 193 /* ---------------------------------------------------------------------- */ 194 if (r) { 195 lu->PetscMatOrdering = PETSC_TRUE; 196 lu->Symbolic = klu_K_analyze_given(n,ai,aj,lu->perm_c,lu->perm_r,&lu->Common); 197 } else { /* use klu internal ordering */ 198 lu->Symbolic = klu_K_analyze(n,ai,aj,&lu->Common); 199 } 200 PetscCheck(lu->Symbolic,PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Symbolic Factorization failed"); 201 202 lu->flg = DIFFERENT_NONZERO_PATTERN; 203 lu->CleanUpKLU = PETSC_TRUE; 204 (F)->ops->lufactornumeric = MatLUFactorNumeric_KLU; 205 PetscFunctionReturn(0); 206 } 207 208 static PetscErrorCode MatView_Info_KLU(Mat A,PetscViewer viewer) 209 { 210 Mat_KLU *lu= (Mat_KLU*)A->data; 211 klu_K_numeric *Numeric=(klu_K_numeric*)lu->Numeric; 212 213 PetscFunctionBegin; 214 PetscCall(PetscViewerASCIIPrintf(viewer,"KLU stats:\n")); 215 PetscCall(PetscViewerASCIIPrintf(viewer," Number of diagonal blocks: %" PetscInt_FMT "\n",(PetscInt)(Numeric->nblocks))); 216 PetscCall(PetscViewerASCIIPrintf(viewer," Total nonzeros=%" PetscInt_FMT "\n",(PetscInt)(Numeric->lnz+Numeric->unz))); 217 PetscCall(PetscViewerASCIIPrintf(viewer,"KLU runtime parameters:\n")); 218 /* Control parameters used by numeric factorization */ 219 PetscCall(PetscViewerASCIIPrintf(viewer," Partial pivoting tolerance: %g\n",lu->Common.tol)); 220 /* BTF preordering */ 221 PetscCall(PetscViewerASCIIPrintf(viewer," BTF preordering enabled: %" PetscInt_FMT "\n",(PetscInt)(lu->Common.btf))); 222 /* mat ordering */ 223 if (!lu->PetscMatOrdering) { 224 PetscCall(PetscViewerASCIIPrintf(viewer," Ordering: %s (not using the PETSc ordering)\n",KluOrderingTypes[(int)lu->Common.ordering])); 225 } 226 /* matrix row scaling */ 227 PetscCall(PetscViewerASCIIPrintf(viewer, " Matrix row scaling: %s\n",scale[(int)lu->Common.scale])); 228 PetscFunctionReturn(0); 229 } 230 231 static PetscErrorCode MatView_KLU(Mat A,PetscViewer viewer) 232 { 233 PetscBool iascii; 234 PetscViewerFormat format; 235 236 PetscFunctionBegin; 237 PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii)); 238 if (iascii) { 239 PetscCall(PetscViewerGetFormat(viewer,&format)); 240 if (format == PETSC_VIEWER_ASCII_INFO) { 241 PetscCall(MatView_Info_KLU(A,viewer)); 242 } 243 } 244 PetscFunctionReturn(0); 245 } 246 247 PetscErrorCode MatFactorGetSolverType_seqaij_klu(Mat A,MatSolverType *type) 248 { 249 PetscFunctionBegin; 250 *type = MATSOLVERKLU; 251 PetscFunctionReturn(0); 252 } 253 254 /*MC 255 MATSOLVERKLU = "klu" - A matrix type providing direct solvers (LU) for sequential matrices 256 via the external package KLU. 257 258 ./configure --download-suitesparse to install PETSc to use KLU 259 260 Use -pc_type lu -pc_factor_mat_solver_type klu to use this direct solver 261 262 Consult KLU documentation for more information on the options database keys below. 263 264 Options Database Keys: 265 + -mat_klu_pivot_tol <0.001> - Partial pivoting tolerance 266 . -mat_klu_use_btf <1> - Use BTF preordering 267 . -mat_klu_ordering <AMD> - KLU reordering scheme to reduce fill-in (choose one of) AMD COLAMD PETSC 268 - -mat_klu_row_scale <NONE> - Matrix row scaling (choose one of) NONE SUM MAX 269 270 Note: KLU is part of SuiteSparse http://faculty.cse.tamu.edu/davis/suitesparse.html 271 272 Level: beginner 273 274 .seealso: PCLU, MATSOLVERUMFPACK, MATSOLVERCHOLMOD, PCFactorSetMatSolverType(), MatSolverType 275 M*/ 276 277 PETSC_INTERN PetscErrorCode MatGetFactor_seqaij_klu(Mat A,MatFactorType ftype,Mat *F) 278 { 279 Mat B; 280 Mat_KLU *lu; 281 PetscErrorCode ierr; 282 PetscInt m=A->rmap->n,n=A->cmap->n,idx = 0,status; 283 PetscBool flg; 284 285 PetscFunctionBegin; 286 /* Create the factorization matrix F */ 287 PetscCall(MatCreate(PetscObjectComm((PetscObject)A),&B)); 288 PetscCall(MatSetSizes(B,PETSC_DECIDE,PETSC_DECIDE,m,n)); 289 PetscCall(PetscStrallocpy("klu",&((PetscObject)B)->type_name)); 290 PetscCall(MatSetUp(B)); 291 292 PetscCall(PetscNewLog(B,&lu)); 293 294 B->data = lu; 295 B->ops->getinfo = MatGetInfo_External; 296 B->ops->lufactorsymbolic = MatLUFactorSymbolic_KLU; 297 B->ops->destroy = MatDestroy_KLU; 298 B->ops->view = MatView_KLU; 299 300 PetscCall(PetscObjectComposeFunction((PetscObject)B,"MatFactorGetSolverType_C",MatFactorGetSolverType_seqaij_klu)); 301 302 B->factortype = MAT_FACTOR_LU; 303 B->assembled = PETSC_TRUE; /* required by -ksp_view */ 304 B->preallocated = PETSC_TRUE; 305 306 PetscCall(PetscFree(B->solvertype)); 307 PetscCall(PetscStrallocpy(MATSOLVERKLU,&B->solvertype)); 308 B->canuseordering = PETSC_TRUE; 309 PetscCall(PetscStrallocpy(MATORDERINGEXTERNAL,(char**)&B->preferredordering[MAT_FACTOR_LU])); 310 311 /* initializations */ 312 /* ------------------------------------------------*/ 313 /* get the default control parameters */ 314 status = klu_K_defaults(&lu->Common); 315 PetscCheck(status > 0,PETSC_COMM_SELF,PETSC_ERR_LIB,"KLU Initialization failed"); 316 317 lu->Common.scale = 0; /* No row scaling */ 318 319 ierr = PetscOptionsBegin(PetscObjectComm((PetscObject)A),((PetscObject)A)->prefix,"KLU Options","Mat");PetscCall(ierr); 320 /* Partial pivoting tolerance */ 321 PetscCall(PetscOptionsReal("-mat_klu_pivot_tol","Partial pivoting tolerance","None",lu->Common.tol,&lu->Common.tol,NULL)); 322 /* BTF pre-ordering */ 323 PetscCall(PetscOptionsInt("-mat_klu_use_btf","Enable BTF preordering","None",(PetscInt)lu->Common.btf,(PetscInt*)&lu->Common.btf,NULL)); 324 /* Matrix reordering */ 325 PetscCall(PetscOptionsEList("-mat_klu_ordering","Internal ordering method","None",KluOrderingTypes,sizeof(KluOrderingTypes)/sizeof(KluOrderingTypes[0]),KluOrderingTypes[0],&idx,&flg)); 326 lu->Common.ordering = (int)idx; 327 /* Matrix row scaling */ 328 PetscCall(PetscOptionsEList("-mat_klu_row_scale","Matrix row scaling","None",scale,3,scale[0],&idx,&flg)); 329 ierr = PetscOptionsEnd();PetscCall(ierr); 330 *F = B; 331 PetscFunctionReturn(0); 332 } 333