1 2 #include <../src/mat/impls/aij/mpi/mpiaij.h> 3 #include <../src/mat/impls/baij/mpi/mpibaij.h> 4 #include <petsc/private/isimpl.h> 5 6 #undef __FUNCT__ 7 #define __FUNCT__ "MatFDColoringApply_BAIJ" 8 PetscErrorCode MatFDColoringApply_BAIJ(Mat J,MatFDColoring coloring,Vec x1,void *sctx) 9 { 10 PetscErrorCode (*f)(void*,Vec,Vec,void*) = (PetscErrorCode (*)(void*,Vec,Vec,void*))coloring->f; 11 PetscErrorCode ierr; 12 PetscInt k,cstart,cend,l,row,col,nz,spidx,i,j; 13 PetscScalar dx=0.0,*w3_array,*dy_i,*dy=coloring->dy; 14 PetscScalar *vscale_array; 15 const PetscScalar *xx; 16 PetscReal epsilon=coloring->error_rel,umin=coloring->umin,unorm; 17 Vec w1=coloring->w1,w2=coloring->w2,w3,vscale=coloring->vscale; 18 void *fctx=coloring->fctx; 19 PetscInt ctype=coloring->ctype,nxloc,nrows_k; 20 PetscScalar *valaddr; 21 MatEntry *Jentry=coloring->matentry; 22 MatEntry2 *Jentry2=coloring->matentry2; 23 const PetscInt ncolors=coloring->ncolors,*ncolumns=coloring->ncolumns,*nrows=coloring->nrows; 24 PetscInt bs=J->rmap->bs; 25 26 PetscFunctionBegin; 27 /* (1) Set w1 = F(x1) */ 28 if (!coloring->fset) { 29 ierr = PetscLogEventBegin(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 30 ierr = (*f)(sctx,x1,w1,fctx);CHKERRQ(ierr); 31 ierr = PetscLogEventEnd(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 32 } else { 33 coloring->fset = PETSC_FALSE; 34 } 35 36 /* (2) Compute vscale = 1./dx - the local scale factors, including ghost points */ 37 ierr = VecGetLocalSize(x1,&nxloc);CHKERRQ(ierr); 38 if (coloring->htype[0] == 'w') { 39 /* vscale = dx is a constant scalar */ 40 ierr = VecNorm(x1,NORM_2,&unorm);CHKERRQ(ierr); 41 dx = 1.0/(PetscSqrtReal(1.0 + unorm)*epsilon); 42 } else { 43 ierr = VecGetArrayRead(x1,&xx);CHKERRQ(ierr); 44 ierr = VecGetArray(vscale,&vscale_array);CHKERRQ(ierr); 45 for (col=0; col<nxloc; col++) { 46 dx = xx[col]; 47 if (PetscAbsScalar(dx) < umin) { 48 if (PetscRealPart(dx) >= 0.0) dx = umin; 49 else if (PetscRealPart(dx) < 0.0 ) dx = -umin; 50 } 51 dx *= epsilon; 52 vscale_array[col] = 1.0/dx; 53 } 54 ierr = VecRestoreArrayRead(x1,&xx);CHKERRQ(ierr); 55 ierr = VecRestoreArray(vscale,&vscale_array);CHKERRQ(ierr); 56 } 57 if (ctype == IS_COLORING_GLOBAL && coloring->htype[0] == 'd') { 58 ierr = VecGhostUpdateBegin(vscale,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 59 ierr = VecGhostUpdateEnd(vscale,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 60 } 61 62 /* (3) Loop over each color */ 63 if (!coloring->w3) { 64 ierr = VecDuplicate(x1,&coloring->w3);CHKERRQ(ierr); 65 ierr = PetscLogObjectParent((PetscObject)coloring,(PetscObject)coloring->w3);CHKERRQ(ierr); 66 } 67 w3 = coloring->w3; 68 69 ierr = VecGetOwnershipRange(x1,&cstart,&cend);CHKERRQ(ierr); /* used by ghosted vscale */ 70 if (vscale) { 71 ierr = VecGetArray(vscale,&vscale_array);CHKERRQ(ierr); 72 } 73 nz = 0; 74 for (k=0; k<ncolors; k++) { 75 coloring->currentcolor = k; 76 77 /* 78 (3-1) Loop over each column associated with color 79 adding the perturbation to the vector w3 = x1 + dx. 80 */ 81 ierr = VecCopy(x1,w3);CHKERRQ(ierr); 82 dy_i = dy; 83 for (i=0; i<bs; i++) { /* Loop over a block of columns */ 84 ierr = VecGetArray(w3,&w3_array);CHKERRQ(ierr); 85 if (ctype == IS_COLORING_GLOBAL) w3_array -= cstart; /* shift pointer so global index can be used */ 86 if (coloring->htype[0] == 'w') { 87 for (l=0; l<ncolumns[k]; l++) { 88 col = i + bs*coloring->columns[k][l]; /* local column (in global index!) of the matrix we are probing for */ 89 w3_array[col] += 1.0/dx; 90 if (i) w3_array[col-1] -= 1.0/dx; /* resume original w3[col-1] */ 91 } 92 } else { /* htype == 'ds' */ 93 vscale_array -= cstart; /* shift pointer so global index can be used */ 94 for (l=0; l<ncolumns[k]; l++) { 95 col = i + bs*coloring->columns[k][l]; /* local column (in global index!) of the matrix we are probing for */ 96 w3_array[col] += 1.0/vscale_array[col]; 97 if (i) w3_array[col-1] -= 1.0/vscale_array[col-1]; /* resume original w3[col-1] */ 98 } 99 vscale_array += cstart; 100 } 101 if (ctype == IS_COLORING_GLOBAL) w3_array += cstart; 102 ierr = VecRestoreArray(w3,&w3_array);CHKERRQ(ierr); 103 104 /* 105 (3-2) Evaluate function at w3 = x1 + dx (here dx is a vector of perturbations) 106 w2 = F(x1 + dx) - F(x1) 107 */ 108 ierr = PetscLogEventBegin(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 109 ierr = VecPlaceArray(w2,dy_i);CHKERRQ(ierr); /* place w2 to the array dy_i */ 110 ierr = (*f)(sctx,w3,w2,fctx);CHKERRQ(ierr); 111 ierr = PetscLogEventEnd(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 112 ierr = VecAXPY(w2,-1.0,w1);CHKERRQ(ierr); 113 ierr = VecResetArray(w2);CHKERRQ(ierr); 114 dy_i += nxloc; /* points to dy+i*nxloc */ 115 } 116 117 /* 118 (3-3) Loop over rows of vector, putting results into Jacobian matrix 119 */ 120 nrows_k = nrows[k]; 121 if (coloring->htype[0] == 'w') { 122 for (l=0; l<nrows_k; l++) { 123 row = bs*Jentry2[nz].row; /* local row index */ 124 valaddr = Jentry2[nz++].valaddr; 125 spidx = 0; 126 dy_i = dy; 127 for (i=0; i<bs; i++) { /* column of the block */ 128 for (j=0; j<bs; j++) { /* row of the block */ 129 valaddr[spidx++] = dy_i[row+j]*dx; 130 } 131 dy_i += nxloc; /* points to dy+i*nxloc */ 132 } 133 } 134 } else { /* htype == 'ds' */ 135 for (l=0; l<nrows_k; l++) { 136 row = bs*Jentry[nz].row; /* local row index */ 137 col = bs*Jentry[nz].col; /* local column index */ 138 valaddr = Jentry[nz++].valaddr; 139 spidx = 0; 140 dy_i = dy; 141 for (i=0; i<bs; i++) { /* column of the block */ 142 for (j=0; j<bs; j++) { /* row of the block */ 143 valaddr[spidx++] = dy_i[row+j]*vscale_array[col+i]; 144 } 145 dy_i += nxloc; /* points to dy+i*nxloc */ 146 } 147 } 148 } 149 } 150 ierr = MatAssemblyBegin(J,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 151 ierr = MatAssemblyEnd(J,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 152 if (vscale) { 153 ierr = VecRestoreArray(vscale,&vscale_array);CHKERRQ(ierr); 154 } 155 156 coloring->currentcolor = -1; 157 PetscFunctionReturn(0); 158 } 159 160 #include <petscdm.h> 161 162 #undef __FUNCT__ 163 #define __FUNCT__ "MatFDColoringApply_AIJ" 164 PetscErrorCode MatFDColoringApply_AIJ(Mat J,MatFDColoring coloring,Vec x1,void *sctx) 165 { 166 PetscErrorCode (*f)(void*,Vec,Vec,void*) = (PetscErrorCode (*)(void*,Vec,Vec,void*))coloring->f; 167 PetscErrorCode ierr; 168 PetscInt k,cstart,cend,l,row,col,nz; 169 PetscScalar dx=0.0,*y,*w3_array; 170 const PetscScalar *xx; 171 PetscScalar *vscale_array; 172 PetscReal epsilon=coloring->error_rel,umin=coloring->umin,unorm; 173 Vec w1=coloring->w1,w2=coloring->w2,w3,vscale=coloring->vscale; 174 void *fctx=coloring->fctx; 175 PetscInt ctype=coloring->ctype,nxloc,nrows_k; 176 MatEntry *Jentry=coloring->matentry; 177 MatEntry2 *Jentry2=coloring->matentry2; 178 const PetscInt ncolors=coloring->ncolors,*ncolumns=coloring->ncolumns,*nrows=coloring->nrows; 179 180 PetscFunctionBegin; 181 if (ctype == IS_COLORING_LOCAL) { 182 Vec x1local; 183 DM dm; 184 ierr = MatGetDM(J,&dm);CHKERRQ(ierr); 185 ierr = DMGetLocalVector(dm,&x1local);CHKERRQ(ierr); 186 ierr = DMGlobalToLocalBegin(dm,x1,INSERT_VALUES,x1local);CHKERRQ(ierr); 187 ierr = DMGlobalToLocalEnd(dm,x1,INSERT_VALUES,x1local);CHKERRQ(ierr); 188 x1 = x1local; 189 } 190 191 /* (1) Set w1 = F(x1) */ 192 if (!coloring->fset) { 193 ierr = PetscLogEventBegin(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 194 ierr = (*f)(sctx,x1,w1,fctx);CHKERRQ(ierr); 195 ierr = PetscLogEventEnd(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 196 } else { 197 coloring->fset = PETSC_FALSE; 198 } 199 200 /* (2) Compute vscale = 1./dx - the local scale factors, including ghost points */ 201 if (coloring->htype[0] == 'w') { 202 /* vscale = 1./dx is a constant scalar */ 203 ierr = VecNorm(x1,NORM_2,&unorm);CHKERRQ(ierr); 204 dx = 1.0/(PetscSqrtReal(1.0 + unorm)*epsilon); 205 } else { 206 ierr = VecGetLocalSize(x1,&nxloc);CHKERRQ(ierr); 207 ierr = VecGetArrayRead(x1,&xx);CHKERRQ(ierr); 208 ierr = VecGetArray(vscale,&vscale_array);CHKERRQ(ierr); 209 for (col=0; col<nxloc; col++) { 210 dx = xx[col]; 211 if (PetscAbsScalar(dx) < umin) { 212 if (PetscRealPart(dx) >= 0.0) dx = umin; 213 else if (PetscRealPart(dx) < 0.0 ) dx = -umin; 214 } 215 dx *= epsilon; 216 vscale_array[col] = 1.0/dx; 217 } 218 ierr = VecRestoreArrayRead(x1,&xx);CHKERRQ(ierr); 219 ierr = VecRestoreArray(vscale,&vscale_array);CHKERRQ(ierr); 220 } 221 if (ctype == IS_COLORING_GLOBAL && coloring->htype[0] == 'd') { 222 ierr = VecGhostUpdateBegin(vscale,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 223 ierr = VecGhostUpdateEnd(vscale,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 224 } 225 226 /* (3) Loop over each color */ 227 if (!coloring->w3) { 228 ierr = VecDuplicate(x1,&coloring->w3);CHKERRQ(ierr); 229 ierr = PetscLogObjectParent((PetscObject)coloring,(PetscObject)coloring->w3);CHKERRQ(ierr); 230 } 231 w3 = coloring->w3; 232 233 ierr = VecGetOwnershipRange(x1,&cstart,&cend);CHKERRQ(ierr); /* used by ghosted vscale */ 234 if (vscale) { 235 ierr = VecGetArray(vscale,&vscale_array);CHKERRQ(ierr); 236 } 237 nz = 0; 238 239 if (coloring->bcols > 1) { /* use blocked insertion of Jentry */ 240 PetscInt i,m=J->rmap->n,nbcols,bcols=coloring->bcols; 241 PetscScalar *dy=coloring->dy,*dy_k; 242 243 nbcols = 0; 244 for (k=0; k<ncolors; k+=bcols) { 245 coloring->currentcolor = k; 246 247 /* 248 (3-1) Loop over each column associated with color 249 adding the perturbation to the vector w3 = x1 + dx. 250 */ 251 252 dy_k = dy; 253 if (k + bcols > ncolors) bcols = ncolors - k; 254 for (i=0; i<bcols; i++) { 255 256 ierr = VecCopy(x1,w3);CHKERRQ(ierr); 257 ierr = VecGetArray(w3,&w3_array);CHKERRQ(ierr); 258 if (ctype == IS_COLORING_GLOBAL) w3_array -= cstart; /* shift pointer so global index can be used */ 259 if (coloring->htype[0] == 'w') { 260 for (l=0; l<ncolumns[k+i]; l++) { 261 col = coloring->columns[k+i][l]; /* local column (in global index!) of the matrix we are probing for */ 262 w3_array[col] += 1.0/dx; 263 } 264 } else { /* htype == 'ds' */ 265 vscale_array -= cstart; /* shift pointer so global index can be used */ 266 for (l=0; l<ncolumns[k+i]; l++) { 267 col = coloring->columns[k+i][l]; /* local column (in global index!) of the matrix we are probing for */ 268 w3_array[col] += 1.0/vscale_array[col]; 269 } 270 vscale_array += cstart; 271 } 272 if (ctype == IS_COLORING_GLOBAL) w3_array += cstart; 273 ierr = VecRestoreArray(w3,&w3_array);CHKERRQ(ierr); 274 275 /* 276 (3-2) Evaluate function at w3 = x1 + dx (here dx is a vector of perturbations) 277 w2 = F(x1 + dx) - F(x1) 278 */ 279 ierr = PetscLogEventBegin(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 280 ierr = VecPlaceArray(w2,dy_k);CHKERRQ(ierr); /* place w2 to the array dy_i */ 281 ierr = (*f)(sctx,w3,w2,fctx);CHKERRQ(ierr); 282 ierr = PetscLogEventEnd(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 283 ierr = VecAXPY(w2,-1.0,w1);CHKERRQ(ierr); 284 ierr = VecResetArray(w2);CHKERRQ(ierr); 285 dy_k += m; /* points to dy+i*nxloc */ 286 } 287 288 /* 289 (3-3) Loop over block rows of vector, putting results into Jacobian matrix 290 */ 291 nrows_k = nrows[nbcols++]; 292 ierr = VecGetArray(w2,&y);CHKERRQ(ierr); 293 294 if (coloring->htype[0] == 'w') { 295 for (l=0; l<nrows_k; l++) { 296 row = Jentry2[nz].row; /* local row index */ 297 *(Jentry2[nz++].valaddr) = dy[row]*dx; 298 } 299 } else { /* htype == 'ds' */ 300 for (l=0; l<nrows_k; l++) { 301 row = Jentry[nz].row; /* local row index */ 302 *(Jentry[nz].valaddr) = dy[row]*vscale_array[Jentry[nz].col]; 303 nz++; 304 } 305 } 306 ierr = VecRestoreArray(w2,&y);CHKERRQ(ierr); 307 } 308 } else { /* bcols == 1 */ 309 for (k=0; k<ncolors; k++) { 310 coloring->currentcolor = k; 311 312 /* 313 (3-1) Loop over each column associated with color 314 adding the perturbation to the vector w3 = x1 + dx. 315 */ 316 ierr = VecCopy(x1,w3);CHKERRQ(ierr); 317 ierr = VecGetArray(w3,&w3_array);CHKERRQ(ierr); 318 if (ctype == IS_COLORING_GLOBAL) w3_array -= cstart; /* shift pointer so global index can be used */ 319 if (coloring->htype[0] == 'w') { 320 for (l=0; l<ncolumns[k]; l++) { 321 col = coloring->columns[k][l]; /* local column (in global index!) of the matrix we are probing for */ 322 w3_array[col] += 1.0/dx; 323 } 324 } else { /* htype == 'ds' */ 325 vscale_array -= cstart; /* shift pointer so global index can be used */ 326 for (l=0; l<ncolumns[k]; l++) { 327 col = coloring->columns[k][l]; /* local column (in global index!) of the matrix we are probing for */ 328 w3_array[col] += 1.0/vscale_array[col]; 329 } 330 vscale_array += cstart; 331 } 332 if (ctype == IS_COLORING_GLOBAL) w3_array += cstart; 333 ierr = VecRestoreArray(w3,&w3_array);CHKERRQ(ierr); 334 335 /* 336 (3-2) Evaluate function at w3 = x1 + dx (here dx is a vector of perturbations) 337 w2 = F(x1 + dx) - F(x1) 338 */ 339 ierr = PetscLogEventBegin(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 340 ierr = (*f)(sctx,w3,w2,fctx);CHKERRQ(ierr); 341 ierr = PetscLogEventEnd(MAT_FDColoringFunction,0,0,0,0);CHKERRQ(ierr); 342 ierr = VecAXPY(w2,-1.0,w1);CHKERRQ(ierr); 343 344 /* 345 (3-3) Loop over rows of vector, putting results into Jacobian matrix 346 */ 347 nrows_k = nrows[k]; 348 ierr = VecGetArray(w2,&y);CHKERRQ(ierr); 349 if (coloring->htype[0] == 'w') { 350 for (l=0; l<nrows_k; l++) { 351 row = Jentry2[nz].row; /* local row index */ 352 *(Jentry2[nz++].valaddr) = y[row]*dx; 353 } 354 } else { /* htype == 'ds' */ 355 for (l=0; l<nrows_k; l++) { 356 row = Jentry[nz].row; /* local row index */ 357 *(Jentry[nz].valaddr) = y[row]*vscale_array[Jentry[nz].col]; 358 nz++; 359 } 360 } 361 ierr = VecRestoreArray(w2,&y);CHKERRQ(ierr); 362 } 363 } 364 365 ierr = MatAssemblyBegin(J,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 366 ierr = MatAssemblyEnd(J,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 367 if (vscale) { 368 ierr = VecRestoreArray(vscale,&vscale_array);CHKERRQ(ierr); 369 } 370 coloring->currentcolor = -1; 371 if (ctype == IS_COLORING_LOCAL) { 372 DM dm; 373 ierr = MatGetDM(J,&dm);CHKERRQ(ierr); 374 ierr = DMRestoreLocalVector(dm,&x1);CHKERRQ(ierr); 375 } 376 PetscFunctionReturn(0); 377 } 378 379 #undef __FUNCT__ 380 #define __FUNCT__ "MatFDColoringSetUp_MPIXAIJ" 381 PetscErrorCode MatFDColoringSetUp_MPIXAIJ(Mat mat,ISColoring iscoloring,MatFDColoring c) 382 { 383 PetscErrorCode ierr; 384 PetscMPIInt size,*ncolsonproc,*disp,nn; 385 PetscInt i,n,nrows,nrows_i,j,k,m,ncols,col,*rowhit,cstart,cend,colb; 386 const PetscInt *is,*A_ci,*A_cj,*B_ci,*B_cj,*row=NULL,*ltog=NULL; 387 PetscInt nis=iscoloring->n,nctot,*cols; 388 IS *isa; 389 ISLocalToGlobalMapping map=mat->cmap->mapping; 390 PetscInt ctype=c->ctype,*spidxA,*spidxB,nz,bs,bs2,spidx; 391 Mat A,B; 392 PetscScalar *A_val,*B_val,**valaddrhit; 393 MatEntry *Jentry; 394 MatEntry2 *Jentry2; 395 PetscBool isBAIJ; 396 PetscInt bcols=c->bcols; 397 #if defined(PETSC_USE_CTABLE) 398 PetscTable colmap=NULL; 399 #else 400 PetscInt *colmap=NULL; /* local col number of off-diag col */ 401 #endif 402 403 PetscFunctionBegin; 404 if (ctype == IS_COLORING_LOCAL) { 405 if (!map) SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_ARG_INCOMP,"When using ghosted differencing matrix must have local to global mapping provided with MatSetLocalToGlobalMapping"); 406 ierr = ISLocalToGlobalMappingGetIndices(map,<og);CHKERRQ(ierr); 407 } 408 409 ierr = MatGetBlockSize(mat,&bs);CHKERRQ(ierr); 410 ierr = PetscObjectTypeCompare((PetscObject)mat,MATMPIBAIJ,&isBAIJ);CHKERRQ(ierr); 411 if (isBAIJ) { 412 Mat_MPIBAIJ *baij=(Mat_MPIBAIJ*)mat->data; 413 Mat_SeqBAIJ *spA,*spB; 414 A = baij->A; spA = (Mat_SeqBAIJ*)A->data; A_val = spA->a; 415 B = baij->B; spB = (Mat_SeqBAIJ*)B->data; B_val = spB->a; 416 nz = spA->nz + spB->nz; /* total nonzero entries of mat */ 417 if (!baij->colmap) { 418 ierr = MatCreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 419 } 420 colmap = baij->colmap; 421 ierr = MatGetColumnIJ_SeqBAIJ_Color(A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&spidxA,NULL);CHKERRQ(ierr); 422 ierr = MatGetColumnIJ_SeqBAIJ_Color(B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&spidxB,NULL);CHKERRQ(ierr); 423 424 if (ctype == IS_COLORING_GLOBAL && c->htype[0] == 'd') { /* create vscale for storing dx */ 425 PetscInt *garray; 426 ierr = PetscMalloc1(B->cmap->n,&garray);CHKERRQ(ierr); 427 for (i=0; i<baij->B->cmap->n/bs; i++) { 428 for (j=0; j<bs; j++) { 429 garray[i*bs+j] = bs*baij->garray[i]+j; 430 } 431 } 432 ierr = VecCreateGhost(PetscObjectComm((PetscObject)mat),mat->cmap->n,PETSC_DETERMINE,B->cmap->n,garray,&c->vscale);CHKERRQ(ierr); 433 ierr = PetscFree(garray);CHKERRQ(ierr); 434 } 435 } else { 436 Mat_MPIAIJ *aij=(Mat_MPIAIJ*)mat->data; 437 Mat_SeqAIJ *spA,*spB; 438 A = aij->A; spA = (Mat_SeqAIJ*)A->data; A_val = spA->a; 439 B = aij->B; spB = (Mat_SeqAIJ*)B->data; B_val = spB->a; 440 nz = spA->nz + spB->nz; /* total nonzero entries of mat */ 441 if (!aij->colmap) { 442 /* Allow access to data structures of local part of matrix 443 - creates aij->colmap which maps global column number to local number in part B */ 444 ierr = MatCreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 445 } 446 colmap = aij->colmap; 447 ierr = MatGetColumnIJ_SeqAIJ_Color(A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&spidxA,NULL);CHKERRQ(ierr); 448 ierr = MatGetColumnIJ_SeqAIJ_Color(B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&spidxB,NULL);CHKERRQ(ierr); 449 450 bs = 1; /* only bs=1 is supported for non MPIBAIJ matrix */ 451 452 if (ctype == IS_COLORING_GLOBAL && c->htype[0] == 'd') { /* create vscale for storing dx */ 453 ierr = VecCreateGhost(PetscObjectComm((PetscObject)mat),mat->cmap->n,PETSC_DETERMINE,B->cmap->n,aij->garray,&c->vscale);CHKERRQ(ierr); 454 } 455 } 456 457 m = mat->rmap->n/bs; 458 cstart = mat->cmap->rstart/bs; 459 cend = mat->cmap->rend/bs; 460 461 ierr = PetscMalloc1(nis,&c->ncolumns);CHKERRQ(ierr); 462 ierr = PetscMalloc1(nis,&c->columns);CHKERRQ(ierr); 463 ierr = PetscCalloc1(nis,&c->nrows);CHKERRQ(ierr); 464 ierr = PetscLogObjectMemory((PetscObject)c,3*nis*sizeof(PetscInt));CHKERRQ(ierr); 465 466 if (c->htype[0] == 'd') { 467 ierr = PetscMalloc1(nz,&Jentry);CHKERRQ(ierr); 468 ierr = PetscLogObjectMemory((PetscObject)c,nz*sizeof(MatEntry));CHKERRQ(ierr); 469 c->matentry = Jentry; 470 } else if (c->htype[0] == 'w') { 471 ierr = PetscMalloc1(nz,&Jentry2);CHKERRQ(ierr); 472 ierr = PetscLogObjectMemory((PetscObject)c,nz*sizeof(MatEntry2));CHKERRQ(ierr); 473 c->matentry2 = Jentry2; 474 } else SETERRQ(PetscObjectComm((PetscObject)mat),PETSC_ERR_SUP,"htype is not supported"); 475 476 ierr = PetscMalloc2(m+1,&rowhit,m+1,&valaddrhit);CHKERRQ(ierr); 477 nz = 0; 478 ierr = ISColoringGetIS(iscoloring,PETSC_IGNORE,&isa);CHKERRQ(ierr); 479 for (i=0; i<nis; i++) { /* for each local color */ 480 ierr = ISGetLocalSize(isa[i],&n);CHKERRQ(ierr); 481 ierr = ISGetIndices(isa[i],&is);CHKERRQ(ierr); 482 483 c->ncolumns[i] = n; /* local number of columns of this color on this process */ 484 if (n) { 485 ierr = PetscMalloc1(n,&c->columns[i]);CHKERRQ(ierr); 486 ierr = PetscLogObjectMemory((PetscObject)c,n*sizeof(PetscInt));CHKERRQ(ierr); 487 ierr = PetscMemcpy(c->columns[i],is,n*sizeof(PetscInt));CHKERRQ(ierr); 488 } else { 489 c->columns[i] = 0; 490 } 491 492 if (ctype == IS_COLORING_GLOBAL) { 493 /* Determine nctot, the total (parallel) number of columns of this color */ 494 ierr = MPI_Comm_size(PetscObjectComm((PetscObject)mat),&size);CHKERRQ(ierr); 495 ierr = PetscMalloc2(size,&ncolsonproc,size,&disp);CHKERRQ(ierr); 496 497 /* ncolsonproc[j]: local ncolumns on proc[j] of this color */ 498 ierr = PetscMPIIntCast(n,&nn);CHKERRQ(ierr); 499 ierr = MPI_Allgather(&nn,1,MPI_INT,ncolsonproc,1,MPI_INT,PetscObjectComm((PetscObject)mat));CHKERRQ(ierr); 500 nctot = 0; for (j=0; j<size; j++) nctot += ncolsonproc[j]; 501 if (!nctot) { 502 ierr = PetscInfo(mat,"Coloring of matrix has some unneeded colors with no corresponding rows\n");CHKERRQ(ierr); 503 } 504 505 disp[0] = 0; 506 for (j=1; j<size; j++) { 507 disp[j] = disp[j-1] + ncolsonproc[j-1]; 508 } 509 510 /* Get cols, the complete list of columns for this color on each process */ 511 ierr = PetscMalloc1(nctot+1,&cols);CHKERRQ(ierr); 512 ierr = MPI_Allgatherv((void*)is,n,MPIU_INT,cols,ncolsonproc,disp,MPIU_INT,PetscObjectComm((PetscObject)mat));CHKERRQ(ierr); 513 ierr = PetscFree2(ncolsonproc,disp);CHKERRQ(ierr); 514 } else if (ctype == IS_COLORING_LOCAL) { 515 /* Determine local number of columns of this color on this process, including ghost points */ 516 nctot = n; 517 ierr = PetscMalloc1(nctot+1,&cols);CHKERRQ(ierr); 518 ierr = PetscMemcpy(cols,is,n*sizeof(PetscInt));CHKERRQ(ierr); 519 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Not provided for this MatFDColoring type"); 520 521 /* Mark all rows affect by these columns */ 522 ierr = PetscMemzero(rowhit,m*sizeof(PetscInt));CHKERRQ(ierr); 523 bs2 = bs*bs; 524 nrows_i = 0; 525 for (j=0; j<nctot; j++) { /* loop over columns*/ 526 if (ctype == IS_COLORING_LOCAL) { 527 col = ltog[cols[j]]; 528 } else { 529 col = cols[j]; 530 } 531 if (col >= cstart && col < cend) { /* column is in A, diagonal block of mat */ 532 row = A_cj + A_ci[col-cstart]; 533 nrows = A_ci[col-cstart+1] - A_ci[col-cstart]; 534 nrows_i += nrows; 535 /* loop over columns of A marking them in rowhit */ 536 for (k=0; k<nrows; k++) { 537 /* set valaddrhit for part A */ 538 spidx = bs2*spidxA[A_ci[col-cstart] + k]; 539 valaddrhit[*row] = &A_val[spidx]; 540 rowhit[*row++] = col - cstart + 1; /* local column index */ 541 } 542 } else { /* column is in B, off-diagonal block of mat */ 543 #if defined(PETSC_USE_CTABLE) 544 ierr = PetscTableFind(colmap,col+1,&colb);CHKERRQ(ierr); 545 colb--; 546 #else 547 colb = colmap[col] - 1; /* local column index */ 548 #endif 549 if (colb == -1) { 550 nrows = 0; 551 } else { 552 colb = colb/bs; 553 row = B_cj + B_ci[colb]; 554 nrows = B_ci[colb+1] - B_ci[colb]; 555 } 556 nrows_i += nrows; 557 /* loop over columns of B marking them in rowhit */ 558 for (k=0; k<nrows; k++) { 559 /* set valaddrhit for part B */ 560 spidx = bs2*spidxB[B_ci[colb] + k]; 561 valaddrhit[*row] = &B_val[spidx]; 562 rowhit[*row++] = colb + 1 + cend - cstart; /* local column index */ 563 } 564 } 565 } 566 c->nrows[i] = nrows_i; 567 568 if (c->htype[0] == 'd') { 569 for (j=0; j<m; j++) { 570 if (rowhit[j]) { 571 Jentry[nz].row = j; /* local row index */ 572 Jentry[nz].col = rowhit[j] - 1; /* local column index */ 573 Jentry[nz].valaddr = valaddrhit[j]; /* address of mat value for this entry */ 574 nz++; 575 } 576 } 577 } else { /* c->htype == 'wp' */ 578 for (j=0; j<m; j++) { 579 if (rowhit[j]) { 580 Jentry2[nz].row = j; /* local row index */ 581 Jentry2[nz].valaddr = valaddrhit[j]; /* address of mat value for this entry */ 582 nz++; 583 } 584 } 585 } 586 ierr = PetscFree(cols);CHKERRQ(ierr); 587 } 588 589 if (bcols > 1) { /* reorder Jentry for faster MatFDColoringApply() */ 590 ierr = MatFDColoringSetUpBlocked_AIJ_Private(mat,c,nz);CHKERRQ(ierr); 591 } 592 593 if (isBAIJ) { 594 ierr = MatRestoreColumnIJ_SeqBAIJ_Color(A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&spidxA,NULL);CHKERRQ(ierr); 595 ierr = MatRestoreColumnIJ_SeqBAIJ_Color(B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&spidxB,NULL);CHKERRQ(ierr); 596 ierr = PetscMalloc1(bs*mat->rmap->n,&c->dy);CHKERRQ(ierr); 597 } else { 598 ierr = MatRestoreColumnIJ_SeqAIJ_Color(A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&spidxA,NULL);CHKERRQ(ierr); 599 ierr = MatRestoreColumnIJ_SeqAIJ_Color(B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&spidxB,NULL);CHKERRQ(ierr); 600 } 601 602 ierr = ISColoringRestoreIS(iscoloring,&isa);CHKERRQ(ierr); 603 ierr = PetscFree2(rowhit,valaddrhit);CHKERRQ(ierr); 604 605 if (ctype == IS_COLORING_LOCAL) { 606 ierr = ISLocalToGlobalMappingRestoreIndices(map,<og);CHKERRQ(ierr); 607 } 608 ierr = PetscInfo3(c,"ncolors %D, brows %D and bcols %D are used.\n",c->ncolors,c->brows,c->bcols);CHKERRQ(ierr); 609 PetscFunctionReturn(0); 610 } 611 612 #undef __FUNCT__ 613 #define __FUNCT__ "MatFDColoringCreate_MPIXAIJ" 614 PetscErrorCode MatFDColoringCreate_MPIXAIJ(Mat mat,ISColoring iscoloring,MatFDColoring c) 615 { 616 PetscErrorCode ierr; 617 PetscInt bs,nis=iscoloring->n,m=mat->rmap->n; 618 PetscBool isBAIJ; 619 620 PetscFunctionBegin; 621 /* set default brows and bcols for speedup inserting the dense matrix into sparse Jacobian; 622 bcols is chosen s.t. dy-array takes 50% of memory space as mat */ 623 ierr = MatGetBlockSize(mat,&bs);CHKERRQ(ierr); 624 ierr = PetscObjectTypeCompare((PetscObject)mat,MATMPIBAIJ,&isBAIJ);CHKERRQ(ierr); 625 if (isBAIJ || m == 0) { 626 c->brows = m; 627 c->bcols = 1; 628 } else { /* mpiaij matrix */ 629 /* bcols is chosen s.t. dy-array takes 50% of local memory space as mat */ 630 Mat_MPIAIJ *aij=(Mat_MPIAIJ*)mat->data; 631 Mat_SeqAIJ *spA,*spB; 632 Mat A,B; 633 PetscInt nz,brows,bcols; 634 PetscReal mem; 635 636 bs = 1; /* only bs=1 is supported for MPIAIJ matrix */ 637 638 A = aij->A; spA = (Mat_SeqAIJ*)A->data; 639 B = aij->B; spB = (Mat_SeqAIJ*)B->data; 640 nz = spA->nz + spB->nz; /* total local nonzero entries of mat */ 641 mem = nz*(sizeof(PetscScalar) + sizeof(PetscInt)) + 3*m*sizeof(PetscInt); 642 bcols = (PetscInt)(0.5*mem /(m*sizeof(PetscScalar))); 643 brows = 1000/bcols; 644 if (bcols > nis) bcols = nis; 645 if (brows == 0 || brows > m) brows = m; 646 c->brows = brows; 647 c->bcols = bcols; 648 } 649 650 c->M = mat->rmap->N/bs; /* set the global rows and columns and local rows */ 651 c->N = mat->cmap->N/bs; 652 c->m = mat->rmap->n/bs; 653 c->rstart = mat->rmap->rstart/bs; 654 c->ncolors = nis; 655 PetscFunctionReturn(0); 656 } 657