1 2 #include "src/mat/matimpl.h" /*I "petscmat.h" I*/ 3 4 #undef __FUNCT__ 5 #define __FUNCT__ "MatAXPY" 6 /*@ 7 MatAXPY - Computes Y = a*X + Y. 8 9 Collective on Mat 10 11 Input Parameters: 12 + a - the scalar multiplier 13 . X - the first matrix 14 . Y - the second matrix 15 - str - either SAME_NONZERO_PATTERN, DIFFERENT_NONZERO_PATTERN or SUBSET_NONZERO_PATTERN 16 17 Contributed by: Matthew Knepley 18 19 Notes: 20 Will only be efficient if one has the SAME_NONZERO_PATTERN or SUBSET_NONZERO_PATTERN 21 22 Level: intermediate 23 24 .keywords: matrix, add 25 26 .seealso: MatAYPX() 27 @*/ 28 PetscErrorCode MatAXPY(const PetscScalar *a,Mat X,Mat Y,MatStructure str) 29 { 30 PetscErrorCode ierr; 31 int m1,m2,n1,n2; 32 33 PetscFunctionBegin; 34 PetscValidScalarPointer(a,1); 35 PetscValidHeaderSpecific(X,MAT_COOKIE,2); 36 PetscValidHeaderSpecific(Y,MAT_COOKIE,3); 37 38 ierr = MatGetSize(X,&m1,&n1);CHKERRQ(ierr); 39 ierr = MatGetSize(Y,&m2,&n2);CHKERRQ(ierr); 40 if (m1 != m2 || n1 != n2) SETERRQ4(PETSC_ERR_ARG_SIZ,"Non conforming matrix add: %d %d %d %d",m1,m2,n1,n2); 41 42 if (X->ops->axpy) { 43 ierr = (*X->ops->axpy)(a,X,Y,str);CHKERRQ(ierr); 44 } else { 45 ierr = MatAXPY_Basic(a,X,Y,str);CHKERRQ(ierr); 46 } 47 PetscFunctionReturn(0); 48 } 49 50 51 #undef __FUNCT__ 52 #define __FUNCT__ "MatAXPY_Basic" 53 PetscErrorCode MatAXPY_Basic(const PetscScalar *a,Mat X,Mat Y,MatStructure str) 54 { 55 int i,start,end,j,ncols,m,n; 56 PetscErrorCode ierr; 57 const int *row; 58 PetscScalar *val; 59 const PetscScalar *vals; 60 61 PetscFunctionBegin; 62 ierr = MatGetSize(X,&m,&n);CHKERRQ(ierr); 63 ierr = MatGetOwnershipRange(X,&start,&end);CHKERRQ(ierr); 64 if (*a == 1.0) { 65 for (i = start; i < end; i++) { 66 ierr = MatGetRow(X,i,&ncols,&row,&vals);CHKERRQ(ierr); 67 ierr = MatSetValues(Y,1,&i,ncols,row,vals,ADD_VALUES);CHKERRQ(ierr); 68 ierr = MatRestoreRow(X,i,&ncols,&row,&vals);CHKERRQ(ierr); 69 } 70 } else { 71 ierr = PetscMalloc((n+1)*sizeof(PetscScalar),&val);CHKERRQ(ierr); 72 for (i=start; i<end; i++) { 73 ierr = MatGetRow(X,i,&ncols,&row,&vals);CHKERRQ(ierr); 74 for (j=0; j<ncols; j++) { 75 val[j] = (*a)*vals[j]; 76 } 77 ierr = MatSetValues(Y,1,&i,ncols,row,val,ADD_VALUES);CHKERRQ(ierr); 78 ierr = MatRestoreRow(X,i,&ncols,&row,&vals);CHKERRQ(ierr); 79 } 80 ierr = PetscFree(val);CHKERRQ(ierr); 81 } 82 ierr = MatAssemblyBegin(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 83 ierr = MatAssemblyEnd(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 84 PetscFunctionReturn(0); 85 } 86 87 #undef __FUNCT__ 88 #define __FUNCT__ "MatShift" 89 /*@ 90 MatShift - Computes Y = Y + a I, where a is a PetscScalar and I is the identity matrix. 91 92 Collective on Mat 93 94 Input Parameters: 95 + Y - the matrices 96 - a - the PetscScalar 97 98 Level: intermediate 99 100 .keywords: matrix, add, shift 101 102 .seealso: MatDiagonalSet() 103 @*/ 104 PetscErrorCode MatShift(const PetscScalar *a,Mat Y) 105 { 106 PetscErrorCode ierr; 107 int i,start,end; 108 109 PetscFunctionBegin; 110 PetscValidScalarPointer(a,1); 111 PetscValidHeaderSpecific(Y,MAT_COOKIE,2); 112 if (Y->ops->shift) { 113 ierr = (*Y->ops->shift)(a,Y);CHKERRQ(ierr); 114 } else { 115 ierr = MatGetOwnershipRange(Y,&start,&end);CHKERRQ(ierr); 116 for (i=start; i<end; i++) { 117 ierr = MatSetValues(Y,1,&i,1,&i,a,ADD_VALUES);CHKERRQ(ierr); 118 } 119 ierr = MatAssemblyBegin(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 120 ierr = MatAssemblyEnd(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 121 } 122 PetscFunctionReturn(0); 123 } 124 125 #undef __FUNCT__ 126 #define __FUNCT__ "MatDiagonalSet" 127 /*@ 128 MatDiagonalSet - Computes Y = Y + D, where D is a diagonal matrix 129 that is represented as a vector. Or Y[i,i] = D[i] if InsertMode is 130 INSERT_VALUES. 131 132 Input Parameters: 133 + Y - the input matrix 134 . D - the diagonal matrix, represented as a vector 135 - i - INSERT_VALUES or ADD_VALUES 136 137 Collective on Mat and Vec 138 139 Level: intermediate 140 141 .keywords: matrix, add, shift, diagonal 142 143 .seealso: MatShift() 144 @*/ 145 PetscErrorCode MatDiagonalSet(Mat Y,Vec D,InsertMode is) 146 { 147 PetscErrorCode ierr; 148 int i,start,end; 149 150 PetscFunctionBegin; 151 PetscValidHeaderSpecific(Y,MAT_COOKIE,1); 152 PetscValidHeaderSpecific(D,VEC_COOKIE,2); 153 if (Y->ops->diagonalset) { 154 ierr = (*Y->ops->diagonalset)(Y,D,is);CHKERRQ(ierr); 155 } else { 156 int vstart,vend; 157 PetscScalar *v; 158 ierr = VecGetOwnershipRange(D,&vstart,&vend);CHKERRQ(ierr); 159 ierr = MatGetOwnershipRange(Y,&start,&end);CHKERRQ(ierr); 160 if (vstart != start || vend != end) { 161 SETERRQ4(PETSC_ERR_ARG_SIZ,"Vector ownership range not compatible with matrix: %d %d vec %d %d mat",vstart,vend,start,end); 162 } 163 ierr = VecGetArray(D,&v);CHKERRQ(ierr); 164 for (i=start; i<end; i++) { 165 ierr = MatSetValues(Y,1,&i,1,&i,v+i-start,is);CHKERRQ(ierr); 166 } 167 ierr = VecRestoreArray(D,&v);CHKERRQ(ierr); 168 ierr = MatAssemblyBegin(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 169 ierr = MatAssemblyEnd(Y,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 170 } 171 PetscFunctionReturn(0); 172 } 173 174 #undef __FUNCT__ 175 #define __FUNCT__ "MatAYPX" 176 /*@ 177 MatAYPX - Computes Y = X + a*Y. 178 179 Collective on Mat 180 181 Input Parameters: 182 + X,Y - the matrices 183 - a - the PetscScalar multiplier 184 185 Contributed by: Matthew Knepley 186 187 Notes: 188 This routine currently uses the MatAXPY() implementation. 189 190 This is slow, if you need it fast send email to petsc-maint@mcs.anl.gov 191 192 Level: intermediate 193 194 .keywords: matrix, add 195 196 .seealso: MatAXPY() 197 @*/ 198 PetscErrorCode MatAYPX(const PetscScalar *a,Mat X,Mat Y) 199 { 200 PetscScalar one = 1.0; 201 PetscErrorCode ierr; 202 int mX,mY,nX,nY; 203 204 PetscFunctionBegin; 205 PetscValidScalarPointer(a,1); 206 PetscValidHeaderSpecific(X,MAT_COOKIE,2); 207 PetscValidHeaderSpecific(Y,MAT_COOKIE,3); 208 209 ierr = MatGetSize(X,&mX,&nX);CHKERRQ(ierr); 210 ierr = MatGetSize(X,&mY,&nY);CHKERRQ(ierr); 211 if (mX != mY || nX != nY) SETERRQ4(PETSC_ERR_ARG_SIZ,"Non conforming matrices: %d %d first %d %d second",mX,mY,nX,nY); 212 213 ierr = MatScale(a,Y);CHKERRQ(ierr); 214 ierr = MatAXPY(&one,X,Y,DIFFERENT_NONZERO_PATTERN);CHKERRQ(ierr); 215 PetscFunctionReturn(0); 216 } 217 218 #undef __FUNCT__ 219 #define __FUNCT__ "MatComputeExplicitOperator" 220 /*@ 221 MatComputeExplicitOperator - Computes the explicit matrix 222 223 Collective on Mat 224 225 Input Parameter: 226 . inmat - the matrix 227 228 Output Parameter: 229 . mat - the explict preconditioned operator 230 231 Notes: 232 This computation is done by applying the operators to columns of the 233 identity matrix. 234 235 Currently, this routine uses a dense matrix format when 1 processor 236 is used and a sparse format otherwise. This routine is costly in general, 237 and is recommended for use only with relatively small systems. 238 239 Level: advanced 240 241 .keywords: Mat, compute, explicit, operator 242 243 @*/ 244 PetscErrorCode MatComputeExplicitOperator(Mat inmat,Mat *mat) 245 { 246 Vec in,out; 247 PetscErrorCode ierr; 248 int i,M,m,size,*rows,start,end; 249 MPI_Comm comm; 250 PetscScalar *array,zero = 0.0,one = 1.0; 251 252 PetscFunctionBegin; 253 PetscValidHeaderSpecific(inmat,MAT_COOKIE,1); 254 PetscValidPointer(mat,2); 255 256 comm = inmat->comm; 257 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 258 259 ierr = MatGetLocalSize(inmat,&m,0);CHKERRQ(ierr); 260 ierr = MatGetSize(inmat,&M,0);CHKERRQ(ierr); 261 ierr = VecCreateMPI(comm,m,M,&in);CHKERRQ(ierr); 262 ierr = VecDuplicate(in,&out);CHKERRQ(ierr); 263 ierr = VecGetOwnershipRange(in,&start,&end);CHKERRQ(ierr); 264 ierr = PetscMalloc((m+1)*sizeof(int),&rows);CHKERRQ(ierr); 265 for (i=0; i<m; i++) {rows[i] = start + i;} 266 267 ierr = MatCreate(comm,m,m,M,M,mat);CHKERRQ(ierr); 268 if (size == 1) { 269 ierr = MatSetType(*mat,MATSEQDENSE);CHKERRQ(ierr); 270 ierr = MatSeqDenseSetPreallocation(*mat,PETSC_NULL);CHKERRQ(ierr); 271 } else { 272 ierr = MatSetType(*mat,MATMPIAIJ);CHKERRQ(ierr); 273 ierr = MatMPIAIJSetPreallocation(*mat,0,PETSC_NULL,0,PETSC_NULL);CHKERRQ(ierr); 274 } 275 276 for (i=0; i<M; i++) { 277 278 ierr = VecSet(&zero,in);CHKERRQ(ierr); 279 ierr = VecSetValues(in,1,&i,&one,INSERT_VALUES);CHKERRQ(ierr); 280 ierr = VecAssemblyBegin(in);CHKERRQ(ierr); 281 ierr = VecAssemblyEnd(in);CHKERRQ(ierr); 282 283 ierr = MatMult(inmat,in,out);CHKERRQ(ierr); 284 285 ierr = VecGetArray(out,&array);CHKERRQ(ierr); 286 ierr = MatSetValues(*mat,m,rows,1,&i,array,INSERT_VALUES);CHKERRQ(ierr); 287 ierr = VecRestoreArray(out,&array);CHKERRQ(ierr); 288 289 } 290 ierr = PetscFree(rows);CHKERRQ(ierr); 291 ierr = VecDestroy(out);CHKERRQ(ierr); 292 ierr = VecDestroy(in);CHKERRQ(ierr); 293 ierr = MatAssemblyBegin(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 294 ierr = MatAssemblyEnd(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 295 PetscFunctionReturn(0); 296 } 297 298 /* Get the map xtoy which is used by MatAXPY() in the case of SUBSET_NONZERO_PATTERN */ 299 #undef __FUNCT__ 300 #define __FUNCT__ "MatAXPYGetxtoy_Private" 301 PetscErrorCode MatAXPYGetxtoy_Private(int m,int *xi,int *xj,int *xgarray, int *yi,int *yj,int *ygarray, int **xtoy) 302 { 303 PetscErrorCode ierr; 304 int row,i,nz,xcol,ycol,jx,jy,*x2y; 305 306 PetscFunctionBegin; 307 ierr = PetscMalloc(xi[m]*sizeof(int),&x2y);CHKERRQ(ierr); 308 i = 0; 309 for (row=0; row<m; row++){ 310 nz = xi[1] - xi[0]; 311 jy = 0; 312 for (jx=0; jx<nz; jx++,jy++){ 313 if (xgarray && ygarray){ 314 xcol = xgarray[xj[*xi + jx]]; 315 ycol = ygarray[yj[*yi + jy]]; 316 } else { 317 xcol = xj[*xi + jx]; 318 ycol = yj[*yi + jy]; /* col index for y */ 319 } 320 while ( ycol < xcol ) { 321 jy++; 322 if (ygarray){ 323 ycol = ygarray[yj[*yi + jy]]; 324 } else { 325 ycol = yj[*yi + jy]; 326 } 327 } 328 if (xcol != ycol) SETERRQ2(PETSC_ERR_ARG_WRONG,"X matrix entry (%d,%d) is not in Y matrix",row,ycol); 329 x2y[i++] = *yi + jy; 330 } 331 xi++; yi++; 332 } 333 *xtoy = x2y; 334 PetscFunctionReturn(0); 335 } 336