1 2 #include <petsc/private/matimpl.h> /*I "petscmat.h" I*/ 3 4 typedef struct { 5 Mat A; /* sparse matrix */ 6 Mat U,V; /* dense tall-skinny matrices */ 7 Vec c; /* sequential vector containing the diagonal of C */ 8 Vec work1,work2; /* sequential vectors that hold partial products */ 9 PetscMPIInt nwork; /* length of work vectors */ 10 Vec xl,yl; /* auxiliary sequential vectors for matmult operation */ 11 } Mat_LRC; 12 13 14 #undef __FUNCT__ 15 #define __FUNCT__ "MatMult_LRC" 16 PetscErrorCode MatMult_LRC(Mat N,Vec x,Vec y) 17 { 18 Mat_LRC *Na = (Mat_LRC*)N->data; 19 PetscErrorCode ierr; 20 PetscScalar *w1,*w2; 21 const PetscScalar *a; 22 23 PetscFunctionBegin; 24 ierr = VecGetArrayRead(x,&a);CHKERRQ(ierr); 25 ierr = VecPlaceArray(Na->xl,a);CHKERRQ(ierr); 26 ierr = VecGetLocalVector(y,Na->yl);CHKERRQ(ierr); 27 28 /* multiply the local part of V with the local part of x */ 29 #if defined(PETSC_USE_COMPLEX) 30 ierr = MatMultHermitianTranspose(Na->V,Na->xl,Na->work1);CHKERRQ(ierr); 31 #else 32 ierr = MatMultTranspose(Na->V,Na->xl,Na->work1);CHKERRQ(ierr); 33 #endif 34 35 /* form the sum of all the local multiplies: this is work2 = V'*x = 36 sum_{all processors} work1 */ 37 ierr = VecGetArray(Na->work1,&w1);CHKERRQ(ierr); 38 ierr = VecGetArray(Na->work2,&w2);CHKERRQ(ierr); 39 ierr = MPIU_Allreduce(w1,w2,Na->nwork,MPIU_SCALAR,MPIU_SUM,PetscObjectComm((PetscObject)N));CHKERRQ(ierr); 40 ierr = VecRestoreArray(Na->work1,&w1);CHKERRQ(ierr); 41 ierr = VecRestoreArray(Na->work2,&w2);CHKERRQ(ierr); 42 43 if (Na->c) { /* work2 = C*work2 */ 44 ierr = VecPointwiseMult(Na->work2,Na->c,Na->work2);CHKERRQ(ierr); 45 } 46 47 if (Na->A) { 48 /* form y = A*x */ 49 ierr = MatMult(Na->A,x,y);CHKERRQ(ierr); 50 /* multiply-add y = y + U*work2 */ 51 ierr = MatMultAdd(Na->U,Na->work2,Na->yl,Na->yl);CHKERRQ(ierr); 52 } else { 53 /* multiply y = U*work2 */ 54 ierr = MatMult(Na->U,Na->work2,Na->yl);CHKERRQ(ierr); 55 } 56 57 ierr = VecRestoreArrayRead(x,&a);CHKERRQ(ierr); 58 ierr = VecResetArray(Na->xl);CHKERRQ(ierr); 59 ierr = VecRestoreLocalVector(y,Na->yl);CHKERRQ(ierr); 60 PetscFunctionReturn(0); 61 } 62 63 #undef __FUNCT__ 64 #define __FUNCT__ "MatDestroy_LRC" 65 PetscErrorCode MatDestroy_LRC(Mat N) 66 { 67 Mat_LRC *Na = (Mat_LRC*)N->data; 68 PetscErrorCode ierr; 69 70 PetscFunctionBegin; 71 ierr = MatDestroy(&Na->A);CHKERRQ(ierr); 72 ierr = MatDestroy(&Na->U);CHKERRQ(ierr); 73 ierr = MatDestroy(&Na->V);CHKERRQ(ierr); 74 ierr = VecDestroy(&Na->c);CHKERRQ(ierr); 75 ierr = VecDestroy(&Na->work1);CHKERRQ(ierr); 76 ierr = VecDestroy(&Na->work2);CHKERRQ(ierr); 77 ierr = VecDestroy(&Na->xl);CHKERRQ(ierr); 78 ierr = VecDestroy(&Na->yl);CHKERRQ(ierr); 79 ierr = PetscFree(N->data);CHKERRQ(ierr); 80 ierr = PetscObjectComposeFunction((PetscObject)N,"MatLRCGetMats_C",0);CHKERRQ(ierr); 81 PetscFunctionReturn(0); 82 } 83 84 #undef __FUNCT__ 85 #define __FUNCT__ "MatLRCGetMats_LRC" 86 PetscErrorCode MatLRCGetMats_LRC(Mat N,Mat *A,Mat *U,Vec *c,Mat *V) 87 { 88 Mat_LRC *Na = (Mat_LRC*)N->data; 89 90 PetscFunctionBegin; 91 if (A) *A = Na->A; 92 if (U) *U = Na->U; 93 if (c) *c = Na->c; 94 if (V) *V = Na->V; 95 PetscFunctionReturn(0); 96 } 97 98 #undef __FUNCT__ 99 #define __FUNCT__ "MatLRCGetMats" 100 /*@ 101 MatLRCGetMats - Returns the constituents of an LRC matrix 102 103 Collective on Mat 104 105 Input Parameter: 106 . N - matrix of type LRC 107 108 Output Parameters: 109 + A - the (sparse) matrix 110 . U, V - two dense rectangular (tall and skinny) matrices 111 - c - a sequential vector containing the diagonal of C 112 113 Note: 114 The returned matrices need not be destroyed by the caller. 115 116 Level: intermediate 117 118 .seealso: MatCreateLRC() 119 @*/ 120 PetscErrorCode MatLRCGetMats(Mat N,Mat *A,Mat *U,Vec *c,Mat *V) 121 { 122 PetscErrorCode ierr; 123 124 PetscFunctionBegin; 125 ierr = PetscUseMethod(N,"MatLRCGetMats_C",(Mat,Mat*,Mat*,Vec*,Mat*),(N,A,U,c,V));CHKERRQ(ierr); 126 PetscFunctionReturn(0); 127 } 128 129 #undef __FUNCT__ 130 #define __FUNCT__ "MatCreateLRC" 131 /*@ 132 MatCreateLRC - Creates a new matrix object that behaves like A + U*C*V' 133 134 Collective on Mat 135 136 Input Parameters: 137 + A - the (sparse) matrix (can be NULL) 138 . U, V - two dense rectangular (tall and skinny) matrices 139 - c - a sequential vector containing the diagonal of C (can be NULL) 140 141 Output Parameter: 142 . N - the matrix that represents A + U*C*V' 143 144 Notes: 145 The matrix A + U*C*V' is not formed! Rather the new matrix 146 object performs the matrix-vector product by first multiplying by 147 A and then adding the other term. 148 149 C is a diagonal matrix (represented as a vector) of order k, 150 where k is the number of columns of both U and V. 151 152 If A is NULL then the new object behaves like a low-rank matrix U*C*V'. 153 154 Use V=U (or V=NULL) for a symmetric low-rank correction, A + U*C*U'. 155 156 If c is NULL then the low-rank correction is just U*V'. 157 158 Level: intermediate 159 160 .seealso: MatLRCGetMats() 161 @*/ 162 PetscErrorCode MatCreateLRC(Mat A,Mat U,Vec c,Mat V,Mat *N) 163 { 164 PetscErrorCode ierr; 165 PetscBool match; 166 PetscInt m,n,k,m1,n1,k1; 167 Mat_LRC *Na; 168 169 PetscFunctionBegin; 170 if (A) PetscValidHeaderSpecific(A,MAT_CLASSID,1); 171 PetscValidHeaderSpecific(U,MAT_CLASSID,2); 172 if (c) PetscValidHeaderSpecific(c,VEC_CLASSID,3); 173 if (V) PetscValidHeaderSpecific(V,MAT_CLASSID,4); 174 else V=U; 175 if (A) PetscCheckSameComm(A,1,U,2); 176 PetscCheckSameComm(U,2,V,4); 177 178 ierr = PetscObjectTypeCompareAny((PetscObject)U,&match,MATSEQDENSE,MATMPIDENSE,"");CHKERRQ(ierr); 179 if (!match) SETERRQ(PetscObjectComm((PetscObject)U),PETSC_ERR_SUP,"Matrix U must be of type dense"); 180 if (V) { 181 ierr = PetscObjectTypeCompareAny((PetscObject)V,&match,MATSEQDENSE,MATMPIDENSE,"");CHKERRQ(ierr); 182 if (!match) SETERRQ(PetscObjectComm((PetscObject)V),PETSC_ERR_SUP,"Matrix V must be of type dense"); 183 } 184 185 ierr = MatGetSize(U,NULL,&k);CHKERRQ(ierr); 186 ierr = MatGetSize(V,NULL,&k1);CHKERRQ(ierr); 187 if (k!=k1) SETERRQ2(PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_INCOMP,"U and V have different number of columns (%D vs %D)",k,k1); 188 ierr = MatGetLocalSize(U,&m,NULL);CHKERRQ(ierr); 189 ierr = MatGetLocalSize(V,&n,NULL);CHKERRQ(ierr); 190 if (A) { 191 ierr = MatGetLocalSize(A,&m1,&n1);CHKERRQ(ierr); 192 if (m!=m1) SETERRQ2(PetscObjectComm((PetscObject)U),PETSC_ERR_ARG_INCOMP,"Local dimensions of U %D and A %D do not match",m,m1); 193 if (n!=n1) SETERRQ2(PetscObjectComm((PetscObject)V),PETSC_ERR_ARG_INCOMP,"Local dimensions of V %D and A %D do not match",n,n1); 194 } 195 if (c) { 196 ierr = VecGetSize(c,&k1);CHKERRQ(ierr); 197 if (k!=k1) SETERRQ2(PetscObjectComm((PetscObject)c),PETSC_ERR_ARG_INCOMP,"The length of c %D does not match the number of columns of U and V (%D)",k1,k); 198 ierr = VecGetLocalSize(c,&k1);CHKERRQ(ierr); 199 if (k!=k1) SETERRQ(PetscObjectComm((PetscObject)c),PETSC_ERR_ARG_INCOMP,"c must be a sequential vector"); 200 } 201 202 ierr = MatCreate(PetscObjectComm((PetscObject)U),N);CHKERRQ(ierr); 203 ierr = MatSetSizes(*N,m,n,PETSC_DECIDE,PETSC_DECIDE);CHKERRQ(ierr); 204 ierr = PetscObjectChangeTypeName((PetscObject)*N,MATLRC);CHKERRQ(ierr); 205 206 ierr = PetscNewLog(*N,&Na);CHKERRQ(ierr); 207 (*N)->data = (void*)Na; 208 Na->A = A; 209 Na->c = c; 210 211 ierr = MatDenseGetLocalMatrix(U,&Na->U);CHKERRQ(ierr); 212 ierr = MatDenseGetLocalMatrix(V,&Na->V);CHKERRQ(ierr); 213 if (A) { ierr = PetscObjectReference((PetscObject)A);CHKERRQ(ierr); } 214 ierr = PetscObjectReference((PetscObject)Na->U);CHKERRQ(ierr); 215 ierr = PetscObjectReference((PetscObject)Na->V);CHKERRQ(ierr); 216 if (c) { ierr = PetscObjectReference((PetscObject)c);CHKERRQ(ierr); } 217 218 ierr = VecCreateSeq(PETSC_COMM_SELF,U->cmap->N,&Na->work1);CHKERRQ(ierr); 219 ierr = VecDuplicate(Na->work1,&Na->work2);CHKERRQ(ierr); 220 ierr = PetscMPIIntCast(U->cmap->N,&Na->nwork);CHKERRQ(ierr); 221 222 ierr = VecCreateSeqWithArray(PETSC_COMM_SELF,1,V->rmap->n,NULL,&Na->xl);CHKERRQ(ierr); 223 ierr = VecCreateSeqWithArray(PETSC_COMM_SELF,1,U->rmap->n,NULL,&Na->yl);CHKERRQ(ierr); 224 225 (*N)->ops->destroy = MatDestroy_LRC; 226 (*N)->ops->mult = MatMult_LRC; 227 (*N)->assembled = PETSC_TRUE; 228 (*N)->preallocated = PETSC_TRUE; 229 (*N)->cmap->N = V->rmap->N; 230 (*N)->rmap->N = U->rmap->N; 231 (*N)->cmap->n = V->rmap->n; 232 (*N)->rmap->n = U->rmap->n; 233 234 ierr = PetscObjectComposeFunction((PetscObject)(*N),"MatLRCGetMats_C",MatLRCGetMats_LRC);CHKERRQ(ierr); 235 PetscFunctionReturn(0); 236 } 237 238