1 2 #include <../src/mat/impls/aij/mpi/mpiaij.h> /*I "petscmat.h" I*/ 3 #include <petscblaslapack.h> 4 5 #undef __FUNCT__ 6 #define __FUNCT__ "MatFindNonZeroRows_MPIAIJ" 7 PetscErrorCode MatFindNonZeroRows_MPIAIJ(Mat M,IS *keptrows) 8 { 9 PetscErrorCode ierr; 10 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)M->data; 11 Mat_SeqAIJ *a = (Mat_SeqAIJ*)mat->A->data; 12 Mat_SeqAIJ *b = (Mat_SeqAIJ*)mat->B->data; 13 const PetscInt *ia,*ib; 14 const MatScalar *aa,*bb; 15 PetscInt na,nb,i,j,*rows,cnt=0,n0rows; 16 PetscInt m = M->rmap->n,rstart = M->rmap->rstart; 17 18 PetscFunctionBegin; 19 *keptrows = 0; 20 ia = a->i; 21 ib = b->i; 22 for (i=0; i<m; i++) { 23 na = ia[i+1] - ia[i]; 24 nb = ib[i+1] - ib[i]; 25 if (!na && !nb) { 26 cnt++; 27 goto ok1; 28 } 29 aa = a->a + ia[i]; 30 for (j=0; j<na; j++) { 31 if (aa[j] != 0.0) goto ok1; 32 } 33 bb = b->a + ib[i]; 34 for (j=0; j <nb; j++) { 35 if (bb[j] != 0.0) goto ok1; 36 } 37 cnt++; 38 ok1:; 39 } 40 ierr = MPI_Allreduce(&cnt,&n0rows,1,MPIU_INT,MPI_SUM,((PetscObject)M)->comm);CHKERRQ(ierr); 41 if (!n0rows) PetscFunctionReturn(0); 42 ierr = PetscMalloc((M->rmap->n-cnt)*sizeof(PetscInt),&rows);CHKERRQ(ierr); 43 cnt = 0; 44 for (i=0; i<m; i++) { 45 na = ia[i+1] - ia[i]; 46 nb = ib[i+1] - ib[i]; 47 if (!na && !nb) continue; 48 aa = a->a + ia[i]; 49 for(j=0; j<na;j++) { 50 if (aa[j] != 0.0) { 51 rows[cnt++] = rstart + i; 52 goto ok2; 53 } 54 } 55 bb = b->a + ib[i]; 56 for (j=0; j<nb; j++) { 57 if (bb[j] != 0.0) { 58 rows[cnt++] = rstart + i; 59 goto ok2; 60 } 61 } 62 ok2:; 63 } 64 ierr = ISCreateGeneral(PETSC_COMM_WORLD,cnt,rows,PETSC_OWN_POINTER,keptrows);CHKERRQ(ierr); 65 PetscFunctionReturn(0); 66 } 67 68 #undef __FUNCT__ 69 #define __FUNCT__ "MatDistribute_MPIAIJ" 70 /* 71 Distributes a SeqAIJ matrix across a set of processes. Code stolen from 72 MatLoad_MPIAIJ(). Horrible lack of reuse. Should be a routine for each matrix type. 73 74 Only for square matrices 75 */ 76 PetscErrorCode MatDistribute_MPIAIJ(MPI_Comm comm,Mat gmat,PetscInt m,MatReuse reuse,Mat *inmat) 77 { 78 PetscMPIInt rank,size; 79 PetscInt *rowners,*dlens,*olens,i,rstart,rend,j,jj,nz,*gmataj,cnt,row,*ld; 80 PetscErrorCode ierr; 81 Mat mat; 82 Mat_SeqAIJ *gmata; 83 PetscMPIInt tag; 84 MPI_Status status; 85 PetscBool aij; 86 MatScalar *gmataa,*ao,*ad,*gmataarestore=0; 87 88 PetscFunctionBegin; 89 CHKMEMQ; 90 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 91 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 92 if (!rank) { 93 ierr = PetscTypeCompare((PetscObject)gmat,MATSEQAIJ,&aij);CHKERRQ(ierr); 94 if (!aij) SETERRQ1(((PetscObject)gmat)->comm,PETSC_ERR_SUP,"Currently no support for input matrix of type %s\n",((PetscObject)gmat)->type_name); 95 } 96 if (reuse == MAT_INITIAL_MATRIX) { 97 ierr = MatCreate(comm,&mat);CHKERRQ(ierr); 98 ierr = MatSetSizes(mat,m,m,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 99 ierr = MatSetType(mat,MATAIJ);CHKERRQ(ierr); 100 ierr = PetscMalloc((size+1)*sizeof(PetscInt),&rowners);CHKERRQ(ierr); 101 ierr = PetscMalloc2(m,PetscInt,&dlens,m,PetscInt,&olens);CHKERRQ(ierr); 102 ierr = MPI_Allgather(&m,1,MPIU_INT,rowners+1,1,MPIU_INT,comm);CHKERRQ(ierr); 103 rowners[0] = 0; 104 for (i=2; i<=size; i++) { 105 rowners[i] += rowners[i-1]; 106 } 107 rstart = rowners[rank]; 108 rend = rowners[rank+1]; 109 ierr = PetscObjectGetNewTag((PetscObject)mat,&tag);CHKERRQ(ierr); 110 if (!rank) { 111 gmata = (Mat_SeqAIJ*) gmat->data; 112 /* send row lengths to all processors */ 113 for (i=0; i<m; i++) dlens[i] = gmata->ilen[i]; 114 for (i=1; i<size; i++) { 115 ierr = MPI_Send(gmata->ilen + rowners[i],rowners[i+1]-rowners[i],MPIU_INT,i,tag,comm);CHKERRQ(ierr); 116 } 117 /* determine number diagonal and off-diagonal counts */ 118 ierr = PetscMemzero(olens,m*sizeof(PetscInt));CHKERRQ(ierr); 119 ierr = PetscMalloc(m*sizeof(PetscInt),&ld);CHKERRQ(ierr); 120 ierr = PetscMemzero(ld,m*sizeof(PetscInt));CHKERRQ(ierr); 121 jj = 0; 122 for (i=0; i<m; i++) { 123 for (j=0; j<dlens[i]; j++) { 124 if (gmata->j[jj] < rstart) ld[i]++; 125 if (gmata->j[jj] < rstart || gmata->j[jj] >= rend) olens[i]++; 126 jj++; 127 } 128 } 129 /* send column indices to other processes */ 130 for (i=1; i<size; i++) { 131 nz = gmata->i[rowners[i+1]]-gmata->i[rowners[i]]; 132 ierr = MPI_Send(&nz,1,MPIU_INT,i,tag,comm);CHKERRQ(ierr); 133 ierr = MPI_Send(gmata->j + gmata->i[rowners[i]],nz,MPIU_INT,i,tag,comm);CHKERRQ(ierr); 134 } 135 136 /* send numerical values to other processes */ 137 for (i=1; i<size; i++) { 138 nz = gmata->i[rowners[i+1]]-gmata->i[rowners[i]]; 139 ierr = MPI_Send(gmata->a + gmata->i[rowners[i]],nz,MPIU_SCALAR,i,tag,comm);CHKERRQ(ierr); 140 } 141 gmataa = gmata->a; 142 gmataj = gmata->j; 143 144 } else { 145 /* receive row lengths */ 146 ierr = MPI_Recv(dlens,m,MPIU_INT,0,tag,comm,&status);CHKERRQ(ierr); 147 /* receive column indices */ 148 ierr = MPI_Recv(&nz,1,MPIU_INT,0,tag,comm,&status);CHKERRQ(ierr); 149 ierr = PetscMalloc2(nz,PetscScalar,&gmataa,nz,PetscInt,&gmataj);CHKERRQ(ierr); 150 ierr = MPI_Recv(gmataj,nz,MPIU_INT,0,tag,comm,&status);CHKERRQ(ierr); 151 /* determine number diagonal and off-diagonal counts */ 152 ierr = PetscMemzero(olens,m*sizeof(PetscInt));CHKERRQ(ierr); 153 ierr = PetscMalloc(m*sizeof(PetscInt),&ld);CHKERRQ(ierr); 154 ierr = PetscMemzero(ld,m*sizeof(PetscInt));CHKERRQ(ierr); 155 jj = 0; 156 for (i=0; i<m; i++) { 157 for (j=0; j<dlens[i]; j++) { 158 if (gmataj[jj] < rstart) ld[i]++; 159 if (gmataj[jj] < rstart || gmataj[jj] >= rend) olens[i]++; 160 jj++; 161 } 162 } 163 /* receive numerical values */ 164 ierr = PetscMemzero(gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); 165 ierr = MPI_Recv(gmataa,nz,MPIU_SCALAR,0,tag,comm,&status);CHKERRQ(ierr); 166 } 167 /* set preallocation */ 168 for (i=0; i<m; i++) { 169 dlens[i] -= olens[i]; 170 } 171 ierr = MatSeqAIJSetPreallocation(mat,0,dlens);CHKERRQ(ierr); 172 ierr = MatMPIAIJSetPreallocation(mat,0,dlens,0,olens);CHKERRQ(ierr); 173 174 for (i=0; i<m; i++) { 175 dlens[i] += olens[i]; 176 } 177 cnt = 0; 178 for (i=0; i<m; i++) { 179 row = rstart + i; 180 ierr = MatSetValues(mat,1,&row,dlens[i],gmataj+cnt,gmataa+cnt,INSERT_VALUES);CHKERRQ(ierr); 181 cnt += dlens[i]; 182 } 183 if (rank) { 184 ierr = PetscFree2(gmataa,gmataj);CHKERRQ(ierr); 185 } 186 ierr = PetscFree2(dlens,olens);CHKERRQ(ierr); 187 ierr = PetscFree(rowners);CHKERRQ(ierr); 188 ((Mat_MPIAIJ*)(mat->data))->ld = ld; 189 *inmat = mat; 190 } else { /* column indices are already set; only need to move over numerical values from process 0 */ 191 Mat_SeqAIJ *Ad = (Mat_SeqAIJ*)((Mat_MPIAIJ*)((*inmat)->data))->A->data; 192 Mat_SeqAIJ *Ao = (Mat_SeqAIJ*)((Mat_MPIAIJ*)((*inmat)->data))->B->data; 193 mat = *inmat; 194 ierr = PetscObjectGetNewTag((PetscObject)mat,&tag);CHKERRQ(ierr); 195 if (!rank) { 196 /* send numerical values to other processes */ 197 gmata = (Mat_SeqAIJ*) gmat->data; 198 ierr = MatGetOwnershipRanges(mat,(const PetscInt**)&rowners);CHKERRQ(ierr); 199 gmataa = gmata->a; 200 for (i=1; i<size; i++) { 201 nz = gmata->i[rowners[i+1]]-gmata->i[rowners[i]]; 202 ierr = MPI_Send(gmataa + gmata->i[rowners[i]],nz,MPIU_SCALAR,i,tag,comm);CHKERRQ(ierr); 203 } 204 nz = gmata->i[rowners[1]]-gmata->i[rowners[0]]; 205 } else { 206 /* receive numerical values from process 0*/ 207 nz = Ad->nz + Ao->nz; 208 ierr = PetscMalloc(nz*sizeof(PetscScalar),&gmataa);CHKERRQ(ierr); gmataarestore = gmataa; 209 ierr = MPI_Recv(gmataa,nz,MPIU_SCALAR,0,tag,comm,&status);CHKERRQ(ierr); 210 } 211 /* transfer numerical values into the diagonal A and off diagonal B parts of mat */ 212 ld = ((Mat_MPIAIJ*)(mat->data))->ld; 213 ad = Ad->a; 214 ao = Ao->a; 215 if (mat->rmap->n) { 216 i = 0; 217 nz = ld[i]; ierr = PetscMemcpy(ao,gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); ao += nz; gmataa += nz; 218 nz = Ad->i[i+1] - Ad->i[i]; ierr = PetscMemcpy(ad,gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); ad += nz; gmataa += nz; 219 } 220 for (i=1; i<mat->rmap->n; i++) { 221 nz = Ao->i[i] - Ao->i[i-1] - ld[i-1] + ld[i]; ierr = PetscMemcpy(ao,gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); ao += nz; gmataa += nz; 222 nz = Ad->i[i+1] - Ad->i[i]; ierr = PetscMemcpy(ad,gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); ad += nz; gmataa += nz; 223 } 224 i--; 225 if (mat->rmap->n) { 226 nz = Ao->i[i+1] - Ao->i[i] - ld[i]; ierr = PetscMemcpy(ao,gmataa,nz*sizeof(PetscScalar));CHKERRQ(ierr); ao += nz; gmataa += nz; 227 } 228 if (rank) { 229 ierr = PetscFree(gmataarestore);CHKERRQ(ierr); 230 } 231 } 232 ierr = MatAssemblyBegin(mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 233 ierr = MatAssemblyEnd(mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 234 CHKMEMQ; 235 PetscFunctionReturn(0); 236 } 237 238 /* 239 Local utility routine that creates a mapping from the global column 240 number to the local number in the off-diagonal part of the local 241 storage of the matrix. When PETSC_USE_CTABLE is used this is scalable at 242 a slightly higher hash table cost; without it it is not scalable (each processor 243 has an order N integer array but is fast to acess. 244 */ 245 #undef __FUNCT__ 246 #define __FUNCT__ "CreateColmap_MPIAIJ_Private" 247 PetscErrorCode CreateColmap_MPIAIJ_Private(Mat mat) 248 { 249 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 250 PetscErrorCode ierr; 251 PetscInt n = aij->B->cmap->n,i; 252 253 PetscFunctionBegin; 254 #if defined (PETSC_USE_CTABLE) 255 ierr = PetscTableCreate(n,&aij->colmap);CHKERRQ(ierr); 256 for (i=0; i<n; i++){ 257 ierr = PetscTableAdd(aij->colmap,aij->garray[i]+1,i+1);CHKERRQ(ierr); 258 } 259 #else 260 ierr = PetscMalloc((mat->cmap->N+1)*sizeof(PetscInt),&aij->colmap);CHKERRQ(ierr); 261 ierr = PetscLogObjectMemory(mat,mat->cmap->N*sizeof(PetscInt));CHKERRQ(ierr); 262 ierr = PetscMemzero(aij->colmap,mat->cmap->N*sizeof(PetscInt));CHKERRQ(ierr); 263 for (i=0; i<n; i++) aij->colmap[aij->garray[i]] = i+1; 264 #endif 265 PetscFunctionReturn(0); 266 } 267 268 #define MatSetValues_SeqAIJ_A_Private(row,col,value,addv) \ 269 { \ 270 if (col <= lastcol1) low1 = 0; else high1 = nrow1; \ 271 lastcol1 = col;\ 272 while (high1-low1 > 5) { \ 273 t = (low1+high1)/2; \ 274 if (rp1[t] > col) high1 = t; \ 275 else low1 = t; \ 276 } \ 277 for (_i=low1; _i<high1; _i++) { \ 278 if (rp1[_i] > col) break; \ 279 if (rp1[_i] == col) { \ 280 if (addv == ADD_VALUES) ap1[_i] += value; \ 281 else ap1[_i] = value; \ 282 goto a_noinsert; \ 283 } \ 284 } \ 285 if (value == 0.0 && ignorezeroentries) {low1 = 0; high1 = nrow1;goto a_noinsert;} \ 286 if (nonew == 1) {low1 = 0; high1 = nrow1; goto a_noinsert;} \ 287 if (nonew == -1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) into matrix", row, col); \ 288 MatSeqXAIJReallocateAIJ(A,am,1,nrow1,row,col,rmax1,aa,ai,aj,rp1,ap1,aimax,nonew,MatScalar); \ 289 N = nrow1++ - 1; a->nz++; high1++; \ 290 /* shift up all the later entries in this row */ \ 291 for (ii=N; ii>=_i; ii--) { \ 292 rp1[ii+1] = rp1[ii]; \ 293 ap1[ii+1] = ap1[ii]; \ 294 } \ 295 rp1[_i] = col; \ 296 ap1[_i] = value; \ 297 a_noinsert: ; \ 298 ailen[row] = nrow1; \ 299 } 300 301 302 #define MatSetValues_SeqAIJ_B_Private(row,col,value,addv) \ 303 { \ 304 if (col <= lastcol2) low2 = 0; else high2 = nrow2; \ 305 lastcol2 = col;\ 306 while (high2-low2 > 5) { \ 307 t = (low2+high2)/2; \ 308 if (rp2[t] > col) high2 = t; \ 309 else low2 = t; \ 310 } \ 311 for (_i=low2; _i<high2; _i++) { \ 312 if (rp2[_i] > col) break; \ 313 if (rp2[_i] == col) { \ 314 if (addv == ADD_VALUES) ap2[_i] += value; \ 315 else ap2[_i] = value; \ 316 goto b_noinsert; \ 317 } \ 318 } \ 319 if (value == 0.0 && ignorezeroentries) {low2 = 0; high2 = nrow2; goto b_noinsert;} \ 320 if (nonew == 1) {low2 = 0; high2 = nrow2; goto b_noinsert;} \ 321 if (nonew == -1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) into matrix", row, col); \ 322 MatSeqXAIJReallocateAIJ(B,bm,1,nrow2,row,col,rmax2,ba,bi,bj,rp2,ap2,bimax,nonew,MatScalar); \ 323 N = nrow2++ - 1; b->nz++; high2++; \ 324 /* shift up all the later entries in this row */ \ 325 for (ii=N; ii>=_i; ii--) { \ 326 rp2[ii+1] = rp2[ii]; \ 327 ap2[ii+1] = ap2[ii]; \ 328 } \ 329 rp2[_i] = col; \ 330 ap2[_i] = value; \ 331 b_noinsert: ; \ 332 bilen[row] = nrow2; \ 333 } 334 335 #undef __FUNCT__ 336 #define __FUNCT__ "MatSetValuesRow_MPIAIJ" 337 PetscErrorCode MatSetValuesRow_MPIAIJ(Mat A,PetscInt row,const PetscScalar v[]) 338 { 339 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)A->data; 340 Mat_SeqAIJ *a = (Mat_SeqAIJ*)mat->A->data,*b = (Mat_SeqAIJ*)mat->B->data; 341 PetscErrorCode ierr; 342 PetscInt l,*garray = mat->garray,diag; 343 344 PetscFunctionBegin; 345 /* code only works for square matrices A */ 346 347 /* find size of row to the left of the diagonal part */ 348 ierr = MatGetOwnershipRange(A,&diag,0);CHKERRQ(ierr); 349 row = row - diag; 350 for (l=0; l<b->i[row+1]-b->i[row]; l++) { 351 if (garray[b->j[b->i[row]+l]] > diag) break; 352 } 353 ierr = PetscMemcpy(b->a+b->i[row],v,l*sizeof(PetscScalar));CHKERRQ(ierr); 354 355 /* diagonal part */ 356 ierr = PetscMemcpy(a->a+a->i[row],v+l,(a->i[row+1]-a->i[row])*sizeof(PetscScalar));CHKERRQ(ierr); 357 358 /* right of diagonal part */ 359 ierr = PetscMemcpy(b->a+b->i[row]+l,v+l+a->i[row+1]-a->i[row],(b->i[row+1]-b->i[row]-l)*sizeof(PetscScalar));CHKERRQ(ierr); 360 PetscFunctionReturn(0); 361 } 362 363 #undef __FUNCT__ 364 #define __FUNCT__ "MatSetValues_MPIAIJ" 365 PetscErrorCode MatSetValues_MPIAIJ(Mat mat,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode addv) 366 { 367 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 368 PetscScalar value; 369 PetscErrorCode ierr; 370 PetscInt i,j,rstart = mat->rmap->rstart,rend = mat->rmap->rend; 371 PetscInt cstart = mat->cmap->rstart,cend = mat->cmap->rend,row,col; 372 PetscBool roworiented = aij->roworiented; 373 374 /* Some Variables required in the macro */ 375 Mat A = aij->A; 376 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 377 PetscInt *aimax = a->imax,*ai = a->i,*ailen = a->ilen,*aj = a->j; 378 MatScalar *aa = a->a; 379 PetscBool ignorezeroentries = a->ignorezeroentries; 380 Mat B = aij->B; 381 Mat_SeqAIJ *b = (Mat_SeqAIJ*)B->data; 382 PetscInt *bimax = b->imax,*bi = b->i,*bilen = b->ilen,*bj = b->j,bm = aij->B->rmap->n,am = aij->A->rmap->n; 383 MatScalar *ba = b->a; 384 385 PetscInt *rp1,*rp2,ii,nrow1,nrow2,_i,rmax1,rmax2,N,low1,high1,low2,high2,t,lastcol1,lastcol2; 386 PetscInt nonew = a->nonew; 387 MatScalar *ap1,*ap2; 388 389 PetscFunctionBegin; 390 if (v) PetscValidScalarPointer(v,6); 391 for (i=0; i<m; i++) { 392 if (im[i] < 0) continue; 393 #if defined(PETSC_USE_DEBUG) 394 if (im[i] >= mat->rmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",im[i],mat->rmap->N-1); 395 #endif 396 if (im[i] >= rstart && im[i] < rend) { 397 row = im[i] - rstart; 398 lastcol1 = -1; 399 rp1 = aj + ai[row]; 400 ap1 = aa + ai[row]; 401 rmax1 = aimax[row]; 402 nrow1 = ailen[row]; 403 low1 = 0; 404 high1 = nrow1; 405 lastcol2 = -1; 406 rp2 = bj + bi[row]; 407 ap2 = ba + bi[row]; 408 rmax2 = bimax[row]; 409 nrow2 = bilen[row]; 410 low2 = 0; 411 high2 = nrow2; 412 413 for (j=0; j<n; j++) { 414 if (v) {if (roworiented) value = v[i*n+j]; else value = v[i+j*m];} else value = 0.0; 415 if (ignorezeroentries && value == 0.0 && (addv == ADD_VALUES)) continue; 416 if (in[j] >= cstart && in[j] < cend){ 417 col = in[j] - cstart; 418 MatSetValues_SeqAIJ_A_Private(row,col,value,addv); 419 } else if (in[j] < 0) continue; 420 #if defined(PETSC_USE_DEBUG) 421 else if (in[j] >= mat->cmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[j],mat->cmap->N-1); 422 #endif 423 else { 424 if (mat->was_assembled) { 425 if (!aij->colmap) { 426 ierr = CreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 427 } 428 #if defined (PETSC_USE_CTABLE) 429 ierr = PetscTableFind(aij->colmap,in[j]+1,&col);CHKERRQ(ierr); 430 col--; 431 #else 432 col = aij->colmap[in[j]] - 1; 433 #endif 434 if (col < 0 && !((Mat_SeqAIJ*)(aij->A->data))->nonew) { 435 ierr = DisAssemble_MPIAIJ(mat);CHKERRQ(ierr); 436 col = in[j]; 437 /* Reinitialize the variables required by MatSetValues_SeqAIJ_B_Private() */ 438 B = aij->B; 439 b = (Mat_SeqAIJ*)B->data; 440 bimax = b->imax; bi = b->i; bilen = b->ilen; bj = b->j; ba = b->a; 441 rp2 = bj + bi[row]; 442 ap2 = ba + bi[row]; 443 rmax2 = bimax[row]; 444 nrow2 = bilen[row]; 445 low2 = 0; 446 high2 = nrow2; 447 bm = aij->B->rmap->n; 448 ba = b->a; 449 } 450 } else col = in[j]; 451 MatSetValues_SeqAIJ_B_Private(row,col,value,addv); 452 } 453 } 454 } else { 455 if (mat->nooffprocentries) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Setting off process row %D even though MatSetOption(,MAT_NO_OFF_PROC_ENTRIES,PETSC_TRUE) was set",im[i]); 456 if (!aij->donotstash) { 457 if (roworiented) { 458 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n,in,v+i*n,(PetscBool)(ignorezeroentries && (addv == ADD_VALUES)));CHKERRQ(ierr); 459 } else { 460 ierr = MatStashValuesCol_Private(&mat->stash,im[i],n,in,v+i,m,(PetscBool)(ignorezeroentries && (addv == ADD_VALUES)));CHKERRQ(ierr); 461 } 462 } 463 } 464 } 465 PetscFunctionReturn(0); 466 } 467 468 #undef __FUNCT__ 469 #define __FUNCT__ "MatGetValues_MPIAIJ" 470 PetscErrorCode MatGetValues_MPIAIJ(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[]) 471 { 472 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 473 PetscErrorCode ierr; 474 PetscInt i,j,rstart = mat->rmap->rstart,rend = mat->rmap->rend; 475 PetscInt cstart = mat->cmap->rstart,cend = mat->cmap->rend,row,col; 476 477 PetscFunctionBegin; 478 for (i=0; i<m; i++) { 479 if (idxm[i] < 0) continue; /* SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative row: %D",idxm[i]);*/ 480 if (idxm[i] >= mat->rmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",idxm[i],mat->rmap->N-1); 481 if (idxm[i] >= rstart && idxm[i] < rend) { 482 row = idxm[i] - rstart; 483 for (j=0; j<n; j++) { 484 if (idxn[j] < 0) continue; /* SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative column: %D",idxn[j]); */ 485 if (idxn[j] >= mat->cmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",idxn[j],mat->cmap->N-1); 486 if (idxn[j] >= cstart && idxn[j] < cend){ 487 col = idxn[j] - cstart; 488 ierr = MatGetValues(aij->A,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 489 } else { 490 if (!aij->colmap) { 491 ierr = CreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 492 } 493 #if defined (PETSC_USE_CTABLE) 494 ierr = PetscTableFind(aij->colmap,idxn[j]+1,&col);CHKERRQ(ierr); 495 col --; 496 #else 497 col = aij->colmap[idxn[j]] - 1; 498 #endif 499 if ((col < 0) || (aij->garray[col] != idxn[j])) *(v+i*n+j) = 0.0; 500 else { 501 ierr = MatGetValues(aij->B,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 502 } 503 } 504 } 505 } else { 506 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only local values currently supported"); 507 } 508 } 509 PetscFunctionReturn(0); 510 } 511 512 extern PetscErrorCode MatMultDiagonalBlock_MPIAIJ(Mat,Vec,Vec); 513 514 #undef __FUNCT__ 515 #define __FUNCT__ "MatAssemblyBegin_MPIAIJ" 516 PetscErrorCode MatAssemblyBegin_MPIAIJ(Mat mat,MatAssemblyType mode) 517 { 518 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 519 PetscErrorCode ierr; 520 PetscInt nstash,reallocs; 521 InsertMode addv; 522 523 PetscFunctionBegin; 524 if (aij->donotstash || mat->nooffprocentries) { 525 PetscFunctionReturn(0); 526 } 527 528 /* make sure all processors are either in INSERTMODE or ADDMODE */ 529 ierr = MPI_Allreduce(&mat->insertmode,&addv,1,MPI_INT,MPI_BOR,((PetscObject)mat)->comm);CHKERRQ(ierr); 530 if (addv == (ADD_VALUES|INSERT_VALUES)) SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_WRONGSTATE,"Some processors inserted others added"); 531 mat->insertmode = addv; /* in case this processor had no cache */ 532 533 ierr = MatStashScatterBegin_Private(mat,&mat->stash,mat->rmap->range);CHKERRQ(ierr); 534 ierr = MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);CHKERRQ(ierr); 535 ierr = PetscInfo2(aij->A,"Stash has %D entries, uses %D mallocs.\n",nstash,reallocs);CHKERRQ(ierr); 536 PetscFunctionReturn(0); 537 } 538 539 #undef __FUNCT__ 540 #define __FUNCT__ "MatAssemblyEnd_MPIAIJ" 541 PetscErrorCode MatAssemblyEnd_MPIAIJ(Mat mat,MatAssemblyType mode) 542 { 543 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 544 Mat_SeqAIJ *a=(Mat_SeqAIJ *)aij->A->data; 545 PetscErrorCode ierr; 546 PetscMPIInt n; 547 PetscInt i,j,rstart,ncols,flg; 548 PetscInt *row,*col; 549 PetscBool other_disassembled; 550 PetscScalar *val; 551 InsertMode addv = mat->insertmode; 552 553 /* do not use 'b = (Mat_SeqAIJ *)aij->B->data' as B can be reset in disassembly */ 554 PetscFunctionBegin; 555 if (!aij->donotstash && !mat->nooffprocentries) { 556 while (1) { 557 ierr = MatStashScatterGetMesg_Private(&mat->stash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 558 if (!flg) break; 559 560 for (i=0; i<n;) { 561 /* Now identify the consecutive vals belonging to the same row */ 562 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 563 if (j < n) ncols = j-i; 564 else ncols = n-i; 565 /* Now assemble all these values with a single function call */ 566 ierr = MatSetValues_MPIAIJ(mat,1,row+i,ncols,col+i,val+i,addv);CHKERRQ(ierr); 567 i = j; 568 } 569 } 570 ierr = MatStashScatterEnd_Private(&mat->stash);CHKERRQ(ierr); 571 } 572 ierr = MatAssemblyBegin(aij->A,mode);CHKERRQ(ierr); 573 ierr = MatAssemblyEnd(aij->A,mode);CHKERRQ(ierr); 574 575 /* determine if any processor has disassembled, if so we must 576 also disassemble ourselfs, in order that we may reassemble. */ 577 /* 578 if nonzero structure of submatrix B cannot change then we know that 579 no processor disassembled thus we can skip this stuff 580 */ 581 if (!((Mat_SeqAIJ*)aij->B->data)->nonew) { 582 ierr = MPI_Allreduce(&mat->was_assembled,&other_disassembled,1,MPI_INT,MPI_PROD,((PetscObject)mat)->comm);CHKERRQ(ierr); 583 if (mat->was_assembled && !other_disassembled) { 584 ierr = DisAssemble_MPIAIJ(mat);CHKERRQ(ierr); 585 } 586 } 587 if (!mat->was_assembled && mode == MAT_FINAL_ASSEMBLY) { 588 ierr = MatSetUpMultiply_MPIAIJ(mat);CHKERRQ(ierr); 589 } 590 ierr = MatSetOption(aij->B,MAT_USE_INODES,PETSC_FALSE);CHKERRQ(ierr); 591 ierr = MatSetOption(aij->B,MAT_CHECK_COMPRESSED_ROW,PETSC_FALSE);CHKERRQ(ierr); 592 ierr = MatAssemblyBegin(aij->B,mode);CHKERRQ(ierr); 593 ierr = MatAssemblyEnd(aij->B,mode);CHKERRQ(ierr); 594 595 ierr = PetscFree2(aij->rowvalues,aij->rowindices);CHKERRQ(ierr); 596 aij->rowvalues = 0; 597 598 /* used by MatAXPY() */ 599 a->xtoy = 0; ((Mat_SeqAIJ *)aij->B->data)->xtoy = 0; /* b->xtoy = 0 */ 600 a->XtoY = 0; ((Mat_SeqAIJ *)aij->B->data)->XtoY = 0; /* b->XtoY = 0 */ 601 602 ierr = VecDestroy(&aij->diag);CHKERRQ(ierr); 603 if (a->inode.size) mat->ops->multdiagonalblock = MatMultDiagonalBlock_MPIAIJ; 604 PetscFunctionReturn(0); 605 } 606 607 #undef __FUNCT__ 608 #define __FUNCT__ "MatZeroEntries_MPIAIJ" 609 PetscErrorCode MatZeroEntries_MPIAIJ(Mat A) 610 { 611 Mat_MPIAIJ *l = (Mat_MPIAIJ*)A->data; 612 PetscErrorCode ierr; 613 614 PetscFunctionBegin; 615 ierr = MatZeroEntries(l->A);CHKERRQ(ierr); 616 ierr = MatZeroEntries(l->B);CHKERRQ(ierr); 617 PetscFunctionReturn(0); 618 } 619 620 #undef __FUNCT__ 621 #define __FUNCT__ "MatZeroRows_MPIAIJ" 622 PetscErrorCode MatZeroRows_MPIAIJ(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 623 { 624 Mat_MPIAIJ *l = (Mat_MPIAIJ*)A->data; 625 PetscErrorCode ierr; 626 PetscMPIInt size = l->size,imdex,n,rank = l->rank,tag = ((PetscObject)A)->tag,lastidx = -1; 627 PetscInt i,*owners = A->rmap->range; 628 PetscInt *nprocs,j,idx,nsends,row; 629 PetscInt nmax,*svalues,*starts,*owner,nrecvs; 630 PetscInt *rvalues,count,base,slen,*source; 631 PetscInt *lens,*lrows,*values,rstart=A->rmap->rstart; 632 MPI_Comm comm = ((PetscObject)A)->comm; 633 MPI_Request *send_waits,*recv_waits; 634 MPI_Status recv_status,*send_status; 635 const PetscScalar *xx; 636 PetscScalar *bb; 637 #if defined(PETSC_DEBUG) 638 PetscBool found = PETSC_FALSE; 639 #endif 640 641 PetscFunctionBegin; 642 /* first count number of contributors to each processor */ 643 ierr = PetscMalloc(2*size*sizeof(PetscInt),&nprocs);CHKERRQ(ierr); 644 ierr = PetscMemzero(nprocs,2*size*sizeof(PetscInt));CHKERRQ(ierr); 645 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&owner);CHKERRQ(ierr); /* see note*/ 646 j = 0; 647 for (i=0; i<N; i++) { 648 if (lastidx > (idx = rows[i])) j = 0; 649 lastidx = idx; 650 for (; j<size; j++) { 651 if (idx >= owners[j] && idx < owners[j+1]) { 652 nprocs[2*j]++; 653 nprocs[2*j+1] = 1; 654 owner[i] = j; 655 #if defined(PETSC_DEBUG) 656 found = PETSC_TRUE; 657 #endif 658 break; 659 } 660 } 661 #if defined(PETSC_DEBUG) 662 if (!found) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Index out of range"); 663 found = PETSC_FALSE; 664 #endif 665 } 666 nsends = 0; for (i=0; i<size; i++) { nsends += nprocs[2*i+1];} 667 668 if (A->nooffproczerorows) { 669 if (nsends > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"You called MatSetOption(,MAT_NO_OFF_PROC_ZERO_ROWS,PETSC_TRUE) but set an off process zero row"); 670 nrecvs = nsends; 671 nmax = N; 672 } else { 673 /* inform other processors of number of messages and max length*/ 674 ierr = PetscMaxSum(comm,nprocs,&nmax,&nrecvs);CHKERRQ(ierr); 675 } 676 677 /* post receives: */ 678 ierr = PetscMalloc((nrecvs+1)*(nmax+1)*sizeof(PetscInt),&rvalues);CHKERRQ(ierr); 679 ierr = PetscMalloc((nrecvs+1)*sizeof(MPI_Request),&recv_waits);CHKERRQ(ierr); 680 for (i=0; i<nrecvs; i++) { 681 ierr = MPI_Irecv(rvalues+nmax*i,nmax,MPIU_INT,MPI_ANY_SOURCE,tag,comm,recv_waits+i);CHKERRQ(ierr); 682 } 683 684 /* do sends: 685 1) starts[i] gives the starting index in svalues for stuff going to 686 the ith processor 687 */ 688 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&svalues);CHKERRQ(ierr); 689 ierr = PetscMalloc((nsends+1)*sizeof(MPI_Request),&send_waits);CHKERRQ(ierr); 690 ierr = PetscMalloc((size+1)*sizeof(PetscInt),&starts);CHKERRQ(ierr); 691 starts[0] = 0; 692 for (i=1; i<size; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 693 for (i=0; i<N; i++) { 694 svalues[starts[owner[i]]++] = rows[i]; 695 } 696 697 starts[0] = 0; 698 for (i=1; i<size+1; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 699 count = 0; 700 for (i=0; i<size; i++) { 701 if (nprocs[2*i+1]) { 702 ierr = MPI_Isend(svalues+starts[i],nprocs[2*i],MPIU_INT,i,tag,comm,send_waits+count++);CHKERRQ(ierr); 703 } 704 } 705 ierr = PetscFree(starts);CHKERRQ(ierr); 706 707 base = owners[rank]; 708 709 /* wait on receives */ 710 ierr = PetscMalloc2(nrecvs,PetscInt,&lens,nrecvs,PetscInt,&source);CHKERRQ(ierr); 711 count = nrecvs; slen = 0; 712 while (count) { 713 ierr = MPI_Waitany(nrecvs,recv_waits,&imdex,&recv_status);CHKERRQ(ierr); 714 /* unpack receives into our local space */ 715 ierr = MPI_Get_count(&recv_status,MPIU_INT,&n);CHKERRQ(ierr); 716 source[imdex] = recv_status.MPI_SOURCE; 717 lens[imdex] = n; 718 slen += n; 719 count--; 720 } 721 ierr = PetscFree(recv_waits);CHKERRQ(ierr); 722 723 /* move the data into the send scatter */ 724 ierr = PetscMalloc((slen+1)*sizeof(PetscInt),&lrows);CHKERRQ(ierr); 725 count = 0; 726 for (i=0; i<nrecvs; i++) { 727 values = rvalues + i*nmax; 728 for (j=0; j<lens[i]; j++) { 729 lrows[count++] = values[j] - base; 730 } 731 } 732 ierr = PetscFree(rvalues);CHKERRQ(ierr); 733 ierr = PetscFree2(lens,source);CHKERRQ(ierr); 734 ierr = PetscFree(owner);CHKERRQ(ierr); 735 ierr = PetscFree(nprocs);CHKERRQ(ierr); 736 737 /* fix right hand side if needed */ 738 if (x && b) { 739 ierr = VecGetArrayRead(x,&xx);CHKERRQ(ierr); 740 ierr = VecGetArray(b,&bb);CHKERRQ(ierr); 741 for (i=0; i<slen; i++) { 742 bb[lrows[i]] = diag*xx[lrows[i]]; 743 } 744 ierr = VecRestoreArrayRead(x,&xx);CHKERRQ(ierr); 745 ierr = VecRestoreArray(b,&bb);CHKERRQ(ierr); 746 } 747 /* 748 Zero the required rows. If the "diagonal block" of the matrix 749 is square and the user wishes to set the diagonal we use separate 750 code so that MatSetValues() is not called for each diagonal allocating 751 new memory, thus calling lots of mallocs and slowing things down. 752 753 */ 754 /* must zero l->B before l->A because the (diag) case below may put values into l->B*/ 755 ierr = MatZeroRows(l->B,slen,lrows,0.0,0,0);CHKERRQ(ierr); 756 if ((diag != 0.0) && (l->A->rmap->N == l->A->cmap->N)) { 757 ierr = MatZeroRows(l->A,slen,lrows,diag,0,0);CHKERRQ(ierr); 758 } else if (diag != 0.0) { 759 ierr = MatZeroRows(l->A,slen,lrows,0.0,0,0);CHKERRQ(ierr); 760 if (((Mat_SeqAIJ*)l->A->data)->nonew) { 761 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"MatZeroRows() on rectangular matrices cannot be used with the Mat options\n\ 762 MAT_NEW_NONZERO_LOCATIONS,MAT_NEW_NONZERO_LOCATION_ERR,MAT_NEW_NONZERO_ALLOCATION_ERR"); 763 } 764 for (i = 0; i < slen; i++) { 765 row = lrows[i] + rstart; 766 ierr = MatSetValues(A,1,&row,1,&row,&diag,INSERT_VALUES);CHKERRQ(ierr); 767 } 768 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 769 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 770 } else { 771 ierr = MatZeroRows(l->A,slen,lrows,0.0,0,0);CHKERRQ(ierr); 772 } 773 ierr = PetscFree(lrows);CHKERRQ(ierr); 774 775 /* wait on sends */ 776 if (nsends) { 777 ierr = PetscMalloc(nsends*sizeof(MPI_Status),&send_status);CHKERRQ(ierr); 778 ierr = MPI_Waitall(nsends,send_waits,send_status);CHKERRQ(ierr); 779 ierr = PetscFree(send_status);CHKERRQ(ierr); 780 } 781 ierr = PetscFree(send_waits);CHKERRQ(ierr); 782 ierr = PetscFree(svalues);CHKERRQ(ierr); 783 784 PetscFunctionReturn(0); 785 } 786 787 #undef __FUNCT__ 788 #define __FUNCT__ "MatZeroRowsColumns_MPIAIJ" 789 PetscErrorCode MatZeroRowsColumns_MPIAIJ(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 790 { 791 Mat_MPIAIJ *l = (Mat_MPIAIJ*)A->data; 792 PetscErrorCode ierr; 793 PetscMPIInt size = l->size,imdex,n,rank = l->rank,tag = ((PetscObject)A)->tag,lastidx = -1; 794 PetscInt i,*owners = A->rmap->range; 795 PetscInt *nprocs,j,idx,nsends; 796 PetscInt nmax,*svalues,*starts,*owner,nrecvs; 797 PetscInt *rvalues,count,base,slen,*source; 798 PetscInt *lens,*lrows,*values,m; 799 MPI_Comm comm = ((PetscObject)A)->comm; 800 MPI_Request *send_waits,*recv_waits; 801 MPI_Status recv_status,*send_status; 802 const PetscScalar *xx; 803 PetscScalar *bb,*mask; 804 Vec xmask,lmask; 805 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)l->B->data; 806 const PetscInt *aj, *ii,*ridx; 807 PetscScalar *aa; 808 #if defined(PETSC_DEBUG) 809 PetscBool found = PETSC_FALSE; 810 #endif 811 812 PetscFunctionBegin; 813 /* first count number of contributors to each processor */ 814 ierr = PetscMalloc(2*size*sizeof(PetscInt),&nprocs);CHKERRQ(ierr); 815 ierr = PetscMemzero(nprocs,2*size*sizeof(PetscInt));CHKERRQ(ierr); 816 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&owner);CHKERRQ(ierr); /* see note*/ 817 j = 0; 818 for (i=0; i<N; i++) { 819 if (lastidx > (idx = rows[i])) j = 0; 820 lastidx = idx; 821 for (; j<size; j++) { 822 if (idx >= owners[j] && idx < owners[j+1]) { 823 nprocs[2*j]++; 824 nprocs[2*j+1] = 1; 825 owner[i] = j; 826 #if defined(PETSC_DEBUG) 827 found = PETSC_TRUE; 828 #endif 829 break; 830 } 831 } 832 #if defined(PETSC_DEBUG) 833 if (!found) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Index out of range"); 834 found = PETSC_FALSE; 835 #endif 836 } 837 nsends = 0; for (i=0; i<size; i++) { nsends += nprocs[2*i+1];} 838 839 /* inform other processors of number of messages and max length*/ 840 ierr = PetscMaxSum(comm,nprocs,&nmax,&nrecvs);CHKERRQ(ierr); 841 842 /* post receives: */ 843 ierr = PetscMalloc((nrecvs+1)*(nmax+1)*sizeof(PetscInt),&rvalues);CHKERRQ(ierr); 844 ierr = PetscMalloc((nrecvs+1)*sizeof(MPI_Request),&recv_waits);CHKERRQ(ierr); 845 for (i=0; i<nrecvs; i++) { 846 ierr = MPI_Irecv(rvalues+nmax*i,nmax,MPIU_INT,MPI_ANY_SOURCE,tag,comm,recv_waits+i);CHKERRQ(ierr); 847 } 848 849 /* do sends: 850 1) starts[i] gives the starting index in svalues for stuff going to 851 the ith processor 852 */ 853 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&svalues);CHKERRQ(ierr); 854 ierr = PetscMalloc((nsends+1)*sizeof(MPI_Request),&send_waits);CHKERRQ(ierr); 855 ierr = PetscMalloc((size+1)*sizeof(PetscInt),&starts);CHKERRQ(ierr); 856 starts[0] = 0; 857 for (i=1; i<size; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 858 for (i=0; i<N; i++) { 859 svalues[starts[owner[i]]++] = rows[i]; 860 } 861 862 starts[0] = 0; 863 for (i=1; i<size+1; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 864 count = 0; 865 for (i=0; i<size; i++) { 866 if (nprocs[2*i+1]) { 867 ierr = MPI_Isend(svalues+starts[i],nprocs[2*i],MPIU_INT,i,tag,comm,send_waits+count++);CHKERRQ(ierr); 868 } 869 } 870 ierr = PetscFree(starts);CHKERRQ(ierr); 871 872 base = owners[rank]; 873 874 /* wait on receives */ 875 ierr = PetscMalloc2(nrecvs,PetscInt,&lens,nrecvs,PetscInt,&source);CHKERRQ(ierr); 876 count = nrecvs; slen = 0; 877 while (count) { 878 ierr = MPI_Waitany(nrecvs,recv_waits,&imdex,&recv_status);CHKERRQ(ierr); 879 /* unpack receives into our local space */ 880 ierr = MPI_Get_count(&recv_status,MPIU_INT,&n);CHKERRQ(ierr); 881 source[imdex] = recv_status.MPI_SOURCE; 882 lens[imdex] = n; 883 slen += n; 884 count--; 885 } 886 ierr = PetscFree(recv_waits);CHKERRQ(ierr); 887 888 /* move the data into the send scatter */ 889 ierr = PetscMalloc((slen+1)*sizeof(PetscInt),&lrows);CHKERRQ(ierr); 890 count = 0; 891 for (i=0; i<nrecvs; i++) { 892 values = rvalues + i*nmax; 893 for (j=0; j<lens[i]; j++) { 894 lrows[count++] = values[j] - base; 895 } 896 } 897 ierr = PetscFree(rvalues);CHKERRQ(ierr); 898 ierr = PetscFree2(lens,source);CHKERRQ(ierr); 899 ierr = PetscFree(owner);CHKERRQ(ierr); 900 ierr = PetscFree(nprocs);CHKERRQ(ierr); 901 /* lrows are the local rows to be zeroed, slen is the number of local rows */ 902 903 /* zero diagonal part of matrix */ 904 ierr = MatZeroRowsColumns(l->A,slen,lrows,diag,x,b);CHKERRQ(ierr); 905 906 /* handle off diagonal part of matrix */ 907 ierr = MatGetVecs(A,&xmask,PETSC_NULL);CHKERRQ(ierr); 908 ierr = VecDuplicate(l->lvec,&lmask);CHKERRQ(ierr); 909 ierr = VecGetArray(xmask,&bb);CHKERRQ(ierr); 910 for (i=0; i<slen; i++) { 911 bb[lrows[i]] = 1; 912 } 913 ierr = VecRestoreArray(xmask,&bb);CHKERRQ(ierr); 914 ierr = VecScatterBegin(l->Mvctx,xmask,lmask,ADD_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 915 ierr = VecScatterEnd(l->Mvctx,xmask,lmask,ADD_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 916 ierr = VecDestroy(&xmask);CHKERRQ(ierr); 917 ierr = VecScatterBegin(l->Mvctx,x,l->lvec,ADD_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 918 ierr = VecScatterEnd(l->Mvctx,x,l->lvec,ADD_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 919 ierr = VecGetArrayRead(l->lvec,&xx);CHKERRQ(ierr); 920 ierr = VecGetArray(lmask,&mask);CHKERRQ(ierr); 921 ierr = VecGetArray(b,&bb);CHKERRQ(ierr); 922 923 /* remove zeroed rows of off diagonal matrix */ 924 ii = aij->i; 925 for (i=0; i<slen; i++) { 926 ierr = PetscMemzero(aij->a + ii[lrows[i]],(ii[lrows[i]+1] - ii[lrows[i]])*sizeof(PetscScalar));CHKERRQ(ierr); 927 } 928 929 /* loop over all elements of off process part of matrix zeroing removed columns*/ 930 if (aij->compressedrow.use){ 931 m = aij->compressedrow.nrows; 932 ii = aij->compressedrow.i; 933 ridx = aij->compressedrow.rindex; 934 for (i=0; i<m; i++){ 935 n = ii[i+1] - ii[i]; 936 aj = aij->j + ii[i]; 937 aa = aij->a + ii[i]; 938 939 for (j=0; j<n; j++) { 940 if (PetscAbsScalar(mask[*aj])) { 941 bb[*ridx] -= *aa*xx[*aj]; 942 *aa = 0.0; 943 } 944 aa++; 945 aj++; 946 } 947 ridx++; 948 } 949 } else { /* do not use compressed row format */ 950 m = l->B->rmap->n; 951 for (i=0; i<m; i++) { 952 n = ii[i+1] - ii[i]; 953 aj = aij->j + ii[i]; 954 aa = aij->a + ii[i]; 955 for (j=0; j<n; j++) { 956 if (PetscAbsScalar(mask[*aj])) { 957 bb[i] -= *aa*xx[*aj]; 958 *aa = 0.0; 959 } 960 aa++; 961 aj++; 962 } 963 } 964 } 965 966 ierr = VecRestoreArray(b,&bb);CHKERRQ(ierr); 967 ierr = VecRestoreArray(lmask,&mask);CHKERRQ(ierr); 968 ierr = VecRestoreArrayRead(l->lvec,&xx);CHKERRQ(ierr); 969 ierr = VecDestroy(&lmask);CHKERRQ(ierr); 970 ierr = PetscFree(lrows);CHKERRQ(ierr); 971 972 /* wait on sends */ 973 if (nsends) { 974 ierr = PetscMalloc(nsends*sizeof(MPI_Status),&send_status);CHKERRQ(ierr); 975 ierr = MPI_Waitall(nsends,send_waits,send_status);CHKERRQ(ierr); 976 ierr = PetscFree(send_status);CHKERRQ(ierr); 977 } 978 ierr = PetscFree(send_waits);CHKERRQ(ierr); 979 ierr = PetscFree(svalues);CHKERRQ(ierr); 980 981 PetscFunctionReturn(0); 982 } 983 984 #undef __FUNCT__ 985 #define __FUNCT__ "MatMult_MPIAIJ" 986 PetscErrorCode MatMult_MPIAIJ(Mat A,Vec xx,Vec yy) 987 { 988 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 989 PetscErrorCode ierr; 990 PetscInt nt; 991 992 PetscFunctionBegin; 993 ierr = VecGetLocalSize(xx,&nt);CHKERRQ(ierr); 994 if (nt != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Incompatible partition of A (%D) and xx (%D)",A->cmap->n,nt); 995 ierr = VecScatterBegin(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 996 ierr = (*a->A->ops->mult)(a->A,xx,yy);CHKERRQ(ierr); 997 ierr = VecScatterEnd(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 998 ierr = (*a->B->ops->multadd)(a->B,a->lvec,yy,yy);CHKERRQ(ierr); 999 PetscFunctionReturn(0); 1000 } 1001 1002 #undef __FUNCT__ 1003 #define __FUNCT__ "MatMultDiagonalBlock_MPIAIJ" 1004 PetscErrorCode MatMultDiagonalBlock_MPIAIJ(Mat A,Vec bb,Vec xx) 1005 { 1006 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1007 PetscErrorCode ierr; 1008 1009 PetscFunctionBegin; 1010 ierr = MatMultDiagonalBlock(a->A,bb,xx);CHKERRQ(ierr); 1011 PetscFunctionReturn(0); 1012 } 1013 1014 #undef __FUNCT__ 1015 #define __FUNCT__ "MatMultAdd_MPIAIJ" 1016 PetscErrorCode MatMultAdd_MPIAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1017 { 1018 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1019 PetscErrorCode ierr; 1020 1021 PetscFunctionBegin; 1022 ierr = VecScatterBegin(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1023 ierr = (*a->A->ops->multadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 1024 ierr = VecScatterEnd(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1025 ierr = (*a->B->ops->multadd)(a->B,a->lvec,zz,zz);CHKERRQ(ierr); 1026 PetscFunctionReturn(0); 1027 } 1028 1029 #undef __FUNCT__ 1030 #define __FUNCT__ "MatMultTranspose_MPIAIJ" 1031 PetscErrorCode MatMultTranspose_MPIAIJ(Mat A,Vec xx,Vec yy) 1032 { 1033 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1034 PetscErrorCode ierr; 1035 PetscBool merged; 1036 1037 PetscFunctionBegin; 1038 ierr = VecScatterGetMerged(a->Mvctx,&merged);CHKERRQ(ierr); 1039 /* do nondiagonal part */ 1040 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1041 if (!merged) { 1042 /* send it on its way */ 1043 ierr = VecScatterBegin(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1044 /* do local part */ 1045 ierr = (*a->A->ops->multtranspose)(a->A,xx,yy);CHKERRQ(ierr); 1046 /* receive remote parts: note this assumes the values are not actually */ 1047 /* added in yy until the next line, */ 1048 ierr = VecScatterEnd(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1049 } else { 1050 /* do local part */ 1051 ierr = (*a->A->ops->multtranspose)(a->A,xx,yy);CHKERRQ(ierr); 1052 /* send it on its way */ 1053 ierr = VecScatterBegin(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1054 /* values actually were received in the Begin() but we need to call this nop */ 1055 ierr = VecScatterEnd(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1056 } 1057 PetscFunctionReturn(0); 1058 } 1059 1060 EXTERN_C_BEGIN 1061 #undef __FUNCT__ 1062 #define __FUNCT__ "MatIsTranspose_MPIAIJ" 1063 PetscErrorCode MatIsTranspose_MPIAIJ(Mat Amat,Mat Bmat,PetscReal tol,PetscBool *f) 1064 { 1065 MPI_Comm comm; 1066 Mat_MPIAIJ *Aij = (Mat_MPIAIJ *) Amat->data, *Bij; 1067 Mat Adia = Aij->A, Bdia, Aoff,Boff,*Aoffs,*Boffs; 1068 IS Me,Notme; 1069 PetscErrorCode ierr; 1070 PetscInt M,N,first,last,*notme,i; 1071 PetscMPIInt size; 1072 1073 PetscFunctionBegin; 1074 1075 /* Easy test: symmetric diagonal block */ 1076 Bij = (Mat_MPIAIJ *) Bmat->data; Bdia = Bij->A; 1077 ierr = MatIsTranspose(Adia,Bdia,tol,f);CHKERRQ(ierr); 1078 if (!*f) PetscFunctionReturn(0); 1079 ierr = PetscObjectGetComm((PetscObject)Amat,&comm);CHKERRQ(ierr); 1080 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 1081 if (size == 1) PetscFunctionReturn(0); 1082 1083 /* Hard test: off-diagonal block. This takes a MatGetSubMatrix. */ 1084 ierr = MatGetSize(Amat,&M,&N);CHKERRQ(ierr); 1085 ierr = MatGetOwnershipRange(Amat,&first,&last);CHKERRQ(ierr); 1086 ierr = PetscMalloc((N-last+first)*sizeof(PetscInt),¬me);CHKERRQ(ierr); 1087 for (i=0; i<first; i++) notme[i] = i; 1088 for (i=last; i<M; i++) notme[i-last+first] = i; 1089 ierr = ISCreateGeneral(MPI_COMM_SELF,N-last+first,notme,PETSC_COPY_VALUES,&Notme);CHKERRQ(ierr); 1090 ierr = ISCreateStride(MPI_COMM_SELF,last-first,first,1,&Me);CHKERRQ(ierr); 1091 ierr = MatGetSubMatrices(Amat,1,&Me,&Notme,MAT_INITIAL_MATRIX,&Aoffs);CHKERRQ(ierr); 1092 Aoff = Aoffs[0]; 1093 ierr = MatGetSubMatrices(Bmat,1,&Notme,&Me,MAT_INITIAL_MATRIX,&Boffs);CHKERRQ(ierr); 1094 Boff = Boffs[0]; 1095 ierr = MatIsTranspose(Aoff,Boff,tol,f);CHKERRQ(ierr); 1096 ierr = MatDestroyMatrices(1,&Aoffs);CHKERRQ(ierr); 1097 ierr = MatDestroyMatrices(1,&Boffs);CHKERRQ(ierr); 1098 ierr = ISDestroy(&Me);CHKERRQ(ierr); 1099 ierr = ISDestroy(&Notme);CHKERRQ(ierr); 1100 PetscFunctionReturn(0); 1101 } 1102 EXTERN_C_END 1103 1104 #undef __FUNCT__ 1105 #define __FUNCT__ "MatMultTransposeAdd_MPIAIJ" 1106 PetscErrorCode MatMultTransposeAdd_MPIAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1107 { 1108 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1109 PetscErrorCode ierr; 1110 1111 PetscFunctionBegin; 1112 /* do nondiagonal part */ 1113 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1114 /* send it on its way */ 1115 ierr = VecScatterBegin(a->Mvctx,a->lvec,zz,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1116 /* do local part */ 1117 ierr = (*a->A->ops->multtransposeadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 1118 /* receive remote parts */ 1119 ierr = VecScatterEnd(a->Mvctx,a->lvec,zz,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1120 PetscFunctionReturn(0); 1121 } 1122 1123 /* 1124 This only works correctly for square matrices where the subblock A->A is the 1125 diagonal block 1126 */ 1127 #undef __FUNCT__ 1128 #define __FUNCT__ "MatGetDiagonal_MPIAIJ" 1129 PetscErrorCode MatGetDiagonal_MPIAIJ(Mat A,Vec v) 1130 { 1131 PetscErrorCode ierr; 1132 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1133 1134 PetscFunctionBegin; 1135 if (A->rmap->N != A->cmap->N) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_SUP,"Supports only square matrix where A->A is diag block"); 1136 if (A->rmap->rstart != A->cmap->rstart || A->rmap->rend != A->cmap->rend) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"row partition must equal col partition"); 1137 ierr = MatGetDiagonal(a->A,v);CHKERRQ(ierr); 1138 PetscFunctionReturn(0); 1139 } 1140 1141 #undef __FUNCT__ 1142 #define __FUNCT__ "MatScale_MPIAIJ" 1143 PetscErrorCode MatScale_MPIAIJ(Mat A,PetscScalar aa) 1144 { 1145 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1146 PetscErrorCode ierr; 1147 1148 PetscFunctionBegin; 1149 ierr = MatScale(a->A,aa);CHKERRQ(ierr); 1150 ierr = MatScale(a->B,aa);CHKERRQ(ierr); 1151 PetscFunctionReturn(0); 1152 } 1153 1154 #undef __FUNCT__ 1155 #define __FUNCT__ "MatDestroy_MPIAIJ" 1156 PetscErrorCode MatDestroy_MPIAIJ(Mat mat) 1157 { 1158 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1159 PetscErrorCode ierr; 1160 1161 PetscFunctionBegin; 1162 #if defined(PETSC_USE_LOG) 1163 PetscLogObjectState((PetscObject)mat,"Rows=%D, Cols=%D",mat->rmap->N,mat->cmap->N); 1164 #endif 1165 ierr = MatStashDestroy_Private(&mat->stash);CHKERRQ(ierr); 1166 ierr = VecDestroy(&aij->diag);CHKERRQ(ierr); 1167 ierr = MatDestroy(&aij->A);CHKERRQ(ierr); 1168 ierr = MatDestroy(&aij->B);CHKERRQ(ierr); 1169 #if defined (PETSC_USE_CTABLE) 1170 if (aij->colmap) {ierr = PetscTableDestroy(aij->colmap);CHKERRQ(ierr);} 1171 #else 1172 ierr = PetscFree(aij->colmap);CHKERRQ(ierr); 1173 #endif 1174 ierr = PetscFree(aij->garray);CHKERRQ(ierr); 1175 ierr = VecDestroy(&aij->lvec);CHKERRQ(ierr); 1176 ierr = VecScatterDestroy(&aij->Mvctx);CHKERRQ(ierr); 1177 ierr = PetscFree2(aij->rowvalues,aij->rowindices);CHKERRQ(ierr); 1178 ierr = PetscFree(aij->ld);CHKERRQ(ierr); 1179 ierr = PetscFree(aij);CHKERRQ(ierr); 1180 1181 ierr = PetscObjectChangeTypeName((PetscObject)mat,0);CHKERRQ(ierr); 1182 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatStoreValues_C","",PETSC_NULL);CHKERRQ(ierr); 1183 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatRetrieveValues_C","",PETSC_NULL);CHKERRQ(ierr); 1184 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatGetDiagonalBlock_C","",PETSC_NULL);CHKERRQ(ierr); 1185 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatIsTranspose_C","",PETSC_NULL);CHKERRQ(ierr); 1186 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatMPIAIJSetPreallocation_C","",PETSC_NULL);CHKERRQ(ierr); 1187 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatMPIAIJSetPreallocationCSR_C","",PETSC_NULL);CHKERRQ(ierr); 1188 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatDiagonalScaleLocal_C","",PETSC_NULL);CHKERRQ(ierr); 1189 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatConvert_mpiaij_mpisbaij_C","",PETSC_NULL);CHKERRQ(ierr); 1190 PetscFunctionReturn(0); 1191 } 1192 1193 #undef __FUNCT__ 1194 #define __FUNCT__ "MatView_MPIAIJ_Binary" 1195 PetscErrorCode MatView_MPIAIJ_Binary(Mat mat,PetscViewer viewer) 1196 { 1197 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1198 Mat_SeqAIJ* A = (Mat_SeqAIJ*)aij->A->data; 1199 Mat_SeqAIJ* B = (Mat_SeqAIJ*)aij->B->data; 1200 PetscErrorCode ierr; 1201 PetscMPIInt rank,size,tag = ((PetscObject)viewer)->tag; 1202 int fd; 1203 PetscInt nz,header[4],*row_lengths,*range=0,rlen,i; 1204 PetscInt nzmax,*column_indices,j,k,col,*garray = aij->garray,cnt,cstart = mat->cmap->rstart,rnz; 1205 PetscScalar *column_values; 1206 PetscInt message_count,flowcontrolcount; 1207 1208 PetscFunctionBegin; 1209 ierr = MPI_Comm_rank(((PetscObject)mat)->comm,&rank);CHKERRQ(ierr); 1210 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 1211 nz = A->nz + B->nz; 1212 if (!rank) { 1213 header[0] = MAT_FILE_CLASSID; 1214 header[1] = mat->rmap->N; 1215 header[2] = mat->cmap->N; 1216 ierr = MPI_Reduce(&nz,&header[3],1,MPIU_INT,MPI_SUM,0,((PetscObject)mat)->comm);CHKERRQ(ierr); 1217 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 1218 ierr = PetscBinaryWrite(fd,header,4,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 1219 /* get largest number of rows any processor has */ 1220 rlen = mat->rmap->n; 1221 range = mat->rmap->range; 1222 for (i=1; i<size; i++) { 1223 rlen = PetscMax(rlen,range[i+1] - range[i]); 1224 } 1225 } else { 1226 ierr = MPI_Reduce(&nz,0,1,MPIU_INT,MPI_SUM,0,((PetscObject)mat)->comm);CHKERRQ(ierr); 1227 rlen = mat->rmap->n; 1228 } 1229 1230 /* load up the local row counts */ 1231 ierr = PetscMalloc((rlen+1)*sizeof(PetscInt),&row_lengths);CHKERRQ(ierr); 1232 for (i=0; i<mat->rmap->n; i++) { 1233 row_lengths[i] = A->i[i+1] - A->i[i] + B->i[i+1] - B->i[i]; 1234 } 1235 1236 /* store the row lengths to the file */ 1237 ierr = PetscViewerFlowControlStart(viewer,&message_count,&flowcontrolcount);CHKERRQ(ierr); 1238 if (!rank) { 1239 MPI_Status status; 1240 ierr = PetscBinaryWrite(fd,row_lengths,mat->rmap->n,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 1241 for (i=1; i<size; i++) { 1242 ierr = PetscViewerFlowControlStepMaster(viewer,i,message_count,flowcontrolcount);CHKERRQ(ierr); 1243 rlen = range[i+1] - range[i]; 1244 ierr = MPI_Recv(row_lengths,rlen,MPIU_INT,i,tag,((PetscObject)mat)->comm,&status);CHKERRQ(ierr); 1245 ierr = PetscBinaryWrite(fd,row_lengths,rlen,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 1246 } 1247 ierr = PetscViewerFlowControlEndMaster(viewer,message_count);CHKERRQ(ierr); 1248 } else { 1249 ierr = PetscViewerFlowControlStepWorker(viewer,rank,message_count);CHKERRQ(ierr); 1250 ierr = MPI_Send(row_lengths,mat->rmap->n,MPIU_INT,0,tag,((PetscObject)mat)->comm);CHKERRQ(ierr); 1251 ierr = PetscViewerFlowControlEndWorker(viewer,message_count);CHKERRQ(ierr); 1252 } 1253 ierr = PetscFree(row_lengths);CHKERRQ(ierr); 1254 1255 /* load up the local column indices */ 1256 nzmax = nz; /* )th processor needs space a largest processor needs */ 1257 ierr = MPI_Reduce(&nz,&nzmax,1,MPIU_INT,MPI_MAX,0,((PetscObject)mat)->comm);CHKERRQ(ierr); 1258 ierr = PetscMalloc((nzmax+1)*sizeof(PetscInt),&column_indices);CHKERRQ(ierr); 1259 cnt = 0; 1260 for (i=0; i<mat->rmap->n; i++) { 1261 for (j=B->i[i]; j<B->i[i+1]; j++) { 1262 if ( (col = garray[B->j[j]]) > cstart) break; 1263 column_indices[cnt++] = col; 1264 } 1265 for (k=A->i[i]; k<A->i[i+1]; k++) { 1266 column_indices[cnt++] = A->j[k] + cstart; 1267 } 1268 for (; j<B->i[i+1]; j++) { 1269 column_indices[cnt++] = garray[B->j[j]]; 1270 } 1271 } 1272 if (cnt != A->nz + B->nz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_LIB,"Internal PETSc error: cnt = %D nz = %D",cnt,A->nz+B->nz); 1273 1274 /* store the column indices to the file */ 1275 ierr = PetscViewerFlowControlStart(viewer,&message_count,&flowcontrolcount);CHKERRQ(ierr); 1276 if (!rank) { 1277 MPI_Status status; 1278 ierr = PetscBinaryWrite(fd,column_indices,nz,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 1279 for (i=1; i<size; i++) { 1280 ierr = PetscViewerFlowControlStepMaster(viewer,i,message_count,flowcontrolcount);CHKERRQ(ierr); 1281 ierr = MPI_Recv(&rnz,1,MPIU_INT,i,tag,((PetscObject)mat)->comm,&status);CHKERRQ(ierr); 1282 if (rnz > nzmax) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_LIB,"Internal PETSc error: nz = %D nzmax = %D",nz,nzmax); 1283 ierr = MPI_Recv(column_indices,rnz,MPIU_INT,i,tag,((PetscObject)mat)->comm,&status);CHKERRQ(ierr); 1284 ierr = PetscBinaryWrite(fd,column_indices,rnz,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 1285 } 1286 ierr = PetscViewerFlowControlEndMaster(viewer,message_count);CHKERRQ(ierr); 1287 } else { 1288 ierr = PetscViewerFlowControlStepWorker(viewer,rank,message_count);CHKERRQ(ierr); 1289 ierr = MPI_Send(&nz,1,MPIU_INT,0,tag,((PetscObject)mat)->comm);CHKERRQ(ierr); 1290 ierr = MPI_Send(column_indices,nz,MPIU_INT,0,tag,((PetscObject)mat)->comm);CHKERRQ(ierr); 1291 ierr = PetscViewerFlowControlEndWorker(viewer,message_count);CHKERRQ(ierr); 1292 } 1293 ierr = PetscFree(column_indices);CHKERRQ(ierr); 1294 1295 /* load up the local column values */ 1296 ierr = PetscMalloc((nzmax+1)*sizeof(PetscScalar),&column_values);CHKERRQ(ierr); 1297 cnt = 0; 1298 for (i=0; i<mat->rmap->n; i++) { 1299 for (j=B->i[i]; j<B->i[i+1]; j++) { 1300 if ( garray[B->j[j]] > cstart) break; 1301 column_values[cnt++] = B->a[j]; 1302 } 1303 for (k=A->i[i]; k<A->i[i+1]; k++) { 1304 column_values[cnt++] = A->a[k]; 1305 } 1306 for (; j<B->i[i+1]; j++) { 1307 column_values[cnt++] = B->a[j]; 1308 } 1309 } 1310 if (cnt != A->nz + B->nz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Internal PETSc error: cnt = %D nz = %D",cnt,A->nz+B->nz); 1311 1312 /* store the column values to the file */ 1313 ierr = PetscViewerFlowControlStart(viewer,&message_count,&flowcontrolcount);CHKERRQ(ierr); 1314 if (!rank) { 1315 MPI_Status status; 1316 ierr = PetscBinaryWrite(fd,column_values,nz,PETSC_SCALAR,PETSC_TRUE);CHKERRQ(ierr); 1317 for (i=1; i<size; i++) { 1318 ierr = PetscViewerFlowControlStepMaster(viewer,i,message_count,flowcontrolcount);CHKERRQ(ierr); 1319 ierr = MPI_Recv(&rnz,1,MPIU_INT,i,tag,((PetscObject)mat)->comm,&status);CHKERRQ(ierr); 1320 if (rnz > nzmax) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_LIB,"Internal PETSc error: nz = %D nzmax = %D",nz,nzmax); 1321 ierr = MPI_Recv(column_values,rnz,MPIU_SCALAR,i,tag,((PetscObject)mat)->comm,&status);CHKERRQ(ierr); 1322 ierr = PetscBinaryWrite(fd,column_values,rnz,PETSC_SCALAR,PETSC_TRUE);CHKERRQ(ierr); 1323 } 1324 ierr = PetscViewerFlowControlEndMaster(viewer,message_count);CHKERRQ(ierr); 1325 } else { 1326 ierr = PetscViewerFlowControlStepWorker(viewer,rank,message_count);CHKERRQ(ierr); 1327 ierr = MPI_Send(&nz,1,MPIU_INT,0,tag,((PetscObject)mat)->comm);CHKERRQ(ierr); 1328 ierr = MPI_Send(column_values,nz,MPIU_SCALAR,0,tag,((PetscObject)mat)->comm);CHKERRQ(ierr); 1329 ierr = PetscViewerFlowControlEndWorker(viewer,message_count);CHKERRQ(ierr); 1330 } 1331 ierr = PetscFree(column_values);CHKERRQ(ierr); 1332 PetscFunctionReturn(0); 1333 } 1334 1335 #undef __FUNCT__ 1336 #define __FUNCT__ "MatView_MPIAIJ_ASCIIorDraworSocket" 1337 PetscErrorCode MatView_MPIAIJ_ASCIIorDraworSocket(Mat mat,PetscViewer viewer) 1338 { 1339 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1340 PetscErrorCode ierr; 1341 PetscMPIInt rank = aij->rank,size = aij->size; 1342 PetscBool isdraw,iascii,isbinary; 1343 PetscViewer sviewer; 1344 PetscViewerFormat format; 1345 1346 PetscFunctionBegin; 1347 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);CHKERRQ(ierr); 1348 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);CHKERRQ(ierr); 1349 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 1350 if (iascii) { 1351 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 1352 if (format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 1353 MatInfo info; 1354 PetscBool inodes; 1355 1356 ierr = MPI_Comm_rank(((PetscObject)mat)->comm,&rank);CHKERRQ(ierr); 1357 ierr = MatGetInfo(mat,MAT_LOCAL,&info);CHKERRQ(ierr); 1358 ierr = MatInodeGetInodeSizes(aij->A,PETSC_NULL,(PetscInt **)&inodes,PETSC_NULL);CHKERRQ(ierr); 1359 if (!inodes) { 1360 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %D nz %D nz alloced %D mem %D, not using I-node routines\n", 1361 rank,mat->rmap->n,(PetscInt)info.nz_used,(PetscInt)info.nz_allocated,(PetscInt)info.memory);CHKERRQ(ierr); 1362 } else { 1363 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %D nz %D nz alloced %D mem %D, using I-node routines\n", 1364 rank,mat->rmap->n,(PetscInt)info.nz_used,(PetscInt)info.nz_allocated,(PetscInt)info.memory);CHKERRQ(ierr); 1365 } 1366 ierr = MatGetInfo(aij->A,MAT_LOCAL,&info);CHKERRQ(ierr); 1367 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] on-diagonal part: nz %D \n",rank,(PetscInt)info.nz_used);CHKERRQ(ierr); 1368 ierr = MatGetInfo(aij->B,MAT_LOCAL,&info);CHKERRQ(ierr); 1369 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] off-diagonal part: nz %D \n",rank,(PetscInt)info.nz_used);CHKERRQ(ierr); 1370 ierr = PetscViewerFlush(viewer);CHKERRQ(ierr); 1371 ierr = PetscViewerASCIIPrintf(viewer,"Information on VecScatter used in matrix-vector product: \n");CHKERRQ(ierr); 1372 ierr = VecScatterView(aij->Mvctx,viewer);CHKERRQ(ierr); 1373 PetscFunctionReturn(0); 1374 } else if (format == PETSC_VIEWER_ASCII_INFO) { 1375 PetscInt inodecount,inodelimit,*inodes; 1376 ierr = MatInodeGetInodeSizes(aij->A,&inodecount,&inodes,&inodelimit);CHKERRQ(ierr); 1377 if (inodes) { 1378 ierr = PetscViewerASCIIPrintf(viewer,"using I-node (on process 0) routines: found %D nodes, limit used is %D\n",inodecount,inodelimit);CHKERRQ(ierr); 1379 } else { 1380 ierr = PetscViewerASCIIPrintf(viewer,"not using I-node (on process 0) routines\n");CHKERRQ(ierr); 1381 } 1382 PetscFunctionReturn(0); 1383 } else if (format == PETSC_VIEWER_ASCII_FACTOR_INFO) { 1384 PetscFunctionReturn(0); 1385 } 1386 } else if (isbinary) { 1387 if (size == 1) { 1388 ierr = PetscObjectSetName((PetscObject)aij->A,((PetscObject)mat)->name);CHKERRQ(ierr); 1389 ierr = MatView(aij->A,viewer);CHKERRQ(ierr); 1390 } else { 1391 ierr = MatView_MPIAIJ_Binary(mat,viewer);CHKERRQ(ierr); 1392 } 1393 PetscFunctionReturn(0); 1394 } else if (isdraw) { 1395 PetscDraw draw; 1396 PetscBool isnull; 1397 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 1398 ierr = PetscDrawIsNull(draw,&isnull);CHKERRQ(ierr); if (isnull) PetscFunctionReturn(0); 1399 } 1400 1401 if (size == 1) { 1402 ierr = PetscObjectSetName((PetscObject)aij->A,((PetscObject)mat)->name);CHKERRQ(ierr); 1403 ierr = MatView(aij->A,viewer);CHKERRQ(ierr); 1404 } else { 1405 /* assemble the entire matrix onto first processor. */ 1406 Mat A; 1407 Mat_SeqAIJ *Aloc; 1408 PetscInt M = mat->rmap->N,N = mat->cmap->N,m,*ai,*aj,row,*cols,i,*ct; 1409 MatScalar *a; 1410 1411 if (mat->rmap->N > 1024) { 1412 PetscBool flg = PETSC_FALSE; 1413 1414 ierr = PetscOptionsGetBool(((PetscObject) mat)->prefix, "-mat_ascii_output_large", &flg,PETSC_NULL);CHKERRQ(ierr); 1415 if (!flg) { 1416 SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_ARG_OUTOFRANGE,"ASCII matrix output not allowed for matrices with more than 1024 rows, use binary format instead.\nYou can override this restriction using -mat_ascii_output_large."); 1417 } 1418 } 1419 1420 ierr = MatCreate(((PetscObject)mat)->comm,&A);CHKERRQ(ierr); 1421 if (!rank) { 1422 ierr = MatSetSizes(A,M,N,M,N);CHKERRQ(ierr); 1423 } else { 1424 ierr = MatSetSizes(A,0,0,M,N);CHKERRQ(ierr); 1425 } 1426 /* This is just a temporary matrix, so explicitly using MATMPIAIJ is probably best */ 1427 ierr = MatSetType(A,MATMPIAIJ);CHKERRQ(ierr); 1428 ierr = MatMPIAIJSetPreallocation(A,0,PETSC_NULL,0,PETSC_NULL);CHKERRQ(ierr); 1429 ierr = PetscLogObjectParent(mat,A);CHKERRQ(ierr); 1430 1431 /* copy over the A part */ 1432 Aloc = (Mat_SeqAIJ*)aij->A->data; 1433 m = aij->A->rmap->n; ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1434 row = mat->rmap->rstart; 1435 for (i=0; i<ai[m]; i++) {aj[i] += mat->cmap->rstart ;} 1436 for (i=0; i<m; i++) { 1437 ierr = MatSetValues(A,1,&row,ai[i+1]-ai[i],aj,a,INSERT_VALUES);CHKERRQ(ierr); 1438 row++; a += ai[i+1]-ai[i]; aj += ai[i+1]-ai[i]; 1439 } 1440 aj = Aloc->j; 1441 for (i=0; i<ai[m]; i++) {aj[i] -= mat->cmap->rstart;} 1442 1443 /* copy over the B part */ 1444 Aloc = (Mat_SeqAIJ*)aij->B->data; 1445 m = aij->B->rmap->n; ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1446 row = mat->rmap->rstart; 1447 ierr = PetscMalloc((ai[m]+1)*sizeof(PetscInt),&cols);CHKERRQ(ierr); 1448 ct = cols; 1449 for (i=0; i<ai[m]; i++) {cols[i] = aij->garray[aj[i]];} 1450 for (i=0; i<m; i++) { 1451 ierr = MatSetValues(A,1,&row,ai[i+1]-ai[i],cols,a,INSERT_VALUES);CHKERRQ(ierr); 1452 row++; a += ai[i+1]-ai[i]; cols += ai[i+1]-ai[i]; 1453 } 1454 ierr = PetscFree(ct);CHKERRQ(ierr); 1455 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1456 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1457 /* 1458 Everyone has to call to draw the matrix since the graphics waits are 1459 synchronized across all processors that share the PetscDraw object 1460 */ 1461 ierr = PetscViewerGetSingleton(viewer,&sviewer);CHKERRQ(ierr); 1462 if (!rank) { 1463 ierr = PetscObjectSetName((PetscObject)((Mat_MPIAIJ*)(A->data))->A,((PetscObject)mat)->name);CHKERRQ(ierr); 1464 /* Set the type name to MATMPIAIJ so that the correct type can be printed out by PetscObjectPrintClassNamePrefixType() in MatView_SeqAIJ_ASCII()*/ 1465 PetscStrcpy(((PetscObject)((Mat_MPIAIJ*)(A->data))->A)->type_name,MATMPIAIJ); 1466 ierr = MatView(((Mat_MPIAIJ*)(A->data))->A,sviewer);CHKERRQ(ierr); 1467 } 1468 ierr = PetscViewerRestoreSingleton(viewer,&sviewer);CHKERRQ(ierr); 1469 ierr = MatDestroy(&A);CHKERRQ(ierr); 1470 } 1471 PetscFunctionReturn(0); 1472 } 1473 1474 #undef __FUNCT__ 1475 #define __FUNCT__ "MatView_MPIAIJ" 1476 PetscErrorCode MatView_MPIAIJ(Mat mat,PetscViewer viewer) 1477 { 1478 PetscErrorCode ierr; 1479 PetscBool iascii,isdraw,issocket,isbinary; 1480 1481 PetscFunctionBegin; 1482 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);CHKERRQ(ierr); 1483 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);CHKERRQ(ierr); 1484 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 1485 ierr = PetscTypeCompare((PetscObject)viewer,PETSCVIEWERSOCKET,&issocket);CHKERRQ(ierr); 1486 if (iascii || isdraw || isbinary || issocket) { 1487 ierr = MatView_MPIAIJ_ASCIIorDraworSocket(mat,viewer);CHKERRQ(ierr); 1488 } else { 1489 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"Viewer type %s not supported by MPIAIJ matrices",((PetscObject)viewer)->type_name); 1490 } 1491 PetscFunctionReturn(0); 1492 } 1493 1494 #undef __FUNCT__ 1495 #define __FUNCT__ "MatSOR_MPIAIJ" 1496 PetscErrorCode MatSOR_MPIAIJ(Mat matin,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx) 1497 { 1498 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 1499 PetscErrorCode ierr; 1500 Vec bb1 = 0; 1501 PetscBool hasop; 1502 1503 PetscFunctionBegin; 1504 if (its > 1 || ~flag & SOR_ZERO_INITIAL_GUESS || flag & SOR_EISENSTAT) { 1505 ierr = VecDuplicate(bb,&bb1);CHKERRQ(ierr); 1506 } 1507 1508 if (flag == SOR_APPLY_UPPER) { 1509 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 1510 PetscFunctionReturn(0); 1511 } 1512 1513 if ((flag & SOR_LOCAL_SYMMETRIC_SWEEP) == SOR_LOCAL_SYMMETRIC_SWEEP){ 1514 if (flag & SOR_ZERO_INITIAL_GUESS) { 1515 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 1516 its--; 1517 } 1518 1519 while (its--) { 1520 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1521 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1522 1523 /* update rhs: bb1 = bb - B*x */ 1524 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 1525 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 1526 1527 /* local sweep */ 1528 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_SYMMETRIC_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 1529 } 1530 } else if (flag & SOR_LOCAL_FORWARD_SWEEP){ 1531 if (flag & SOR_ZERO_INITIAL_GUESS) { 1532 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 1533 its--; 1534 } 1535 while (its--) { 1536 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1537 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1538 1539 /* update rhs: bb1 = bb - B*x */ 1540 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 1541 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 1542 1543 /* local sweep */ 1544 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_FORWARD_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 1545 } 1546 } else if (flag & SOR_LOCAL_BACKWARD_SWEEP){ 1547 if (flag & SOR_ZERO_INITIAL_GUESS) { 1548 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 1549 its--; 1550 } 1551 while (its--) { 1552 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1553 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1554 1555 /* update rhs: bb1 = bb - B*x */ 1556 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 1557 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 1558 1559 /* local sweep */ 1560 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_BACKWARD_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 1561 } 1562 } else if (flag & SOR_EISENSTAT) { 1563 Vec xx1; 1564 1565 ierr = VecDuplicate(bb,&xx1);CHKERRQ(ierr); 1566 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,(MatSORType)(SOR_ZERO_INITIAL_GUESS | SOR_LOCAL_BACKWARD_SWEEP),fshift,lits,1,xx);CHKERRQ(ierr); 1567 1568 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1569 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1570 if (!mat->diag) { 1571 ierr = MatGetVecs(matin,&mat->diag,PETSC_NULL);CHKERRQ(ierr); 1572 ierr = MatGetDiagonal(matin,mat->diag);CHKERRQ(ierr); 1573 } 1574 ierr = MatHasOperation(matin,MATOP_MULT_DIAGONAL_BLOCK,&hasop);CHKERRQ(ierr); 1575 if (hasop) { 1576 ierr = MatMultDiagonalBlock(matin,xx,bb1);CHKERRQ(ierr); 1577 } else { 1578 ierr = VecPointwiseMult(bb1,mat->diag,xx);CHKERRQ(ierr); 1579 } 1580 ierr = VecAYPX(bb1,(omega-2.0)/omega,bb);CHKERRQ(ierr); 1581 1582 ierr = MatMultAdd(mat->B,mat->lvec,bb1,bb1);CHKERRQ(ierr); 1583 1584 /* local sweep */ 1585 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,(MatSORType)(SOR_ZERO_INITIAL_GUESS | SOR_LOCAL_FORWARD_SWEEP),fshift,lits,1,xx1);CHKERRQ(ierr); 1586 ierr = VecAXPY(xx,1.0,xx1);CHKERRQ(ierr); 1587 ierr = VecDestroy(&xx1);CHKERRQ(ierr); 1588 } else SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"Parallel SOR not supported"); 1589 1590 ierr = VecDestroy(&bb1);CHKERRQ(ierr); 1591 PetscFunctionReturn(0); 1592 } 1593 1594 #undef __FUNCT__ 1595 #define __FUNCT__ "MatPermute_MPIAIJ" 1596 PetscErrorCode MatPermute_MPIAIJ(Mat A,IS rowp,IS colp,Mat *B) 1597 { 1598 MPI_Comm comm,pcomm; 1599 PetscInt first,local_size,nrows; 1600 const PetscInt *rows; 1601 PetscMPIInt size; 1602 IS crowp,growp,irowp,lrowp,lcolp,icolp; 1603 PetscErrorCode ierr; 1604 1605 PetscFunctionBegin; 1606 ierr = PetscObjectGetComm((PetscObject)A,&comm);CHKERRQ(ierr); 1607 /* make a collective version of 'rowp' */ 1608 ierr = PetscObjectGetComm((PetscObject)rowp,&pcomm);CHKERRQ(ierr); 1609 if (pcomm==comm) { 1610 crowp = rowp; 1611 } else { 1612 ierr = ISGetSize(rowp,&nrows);CHKERRQ(ierr); 1613 ierr = ISGetIndices(rowp,&rows);CHKERRQ(ierr); 1614 ierr = ISCreateGeneral(comm,nrows,rows,PETSC_COPY_VALUES,&crowp);CHKERRQ(ierr); 1615 ierr = ISRestoreIndices(rowp,&rows);CHKERRQ(ierr); 1616 } 1617 /* collect the global row permutation and invert it */ 1618 ierr = ISAllGather(crowp,&growp);CHKERRQ(ierr); 1619 ierr = ISSetPermutation(growp);CHKERRQ(ierr); 1620 if (pcomm!=comm) { 1621 ierr = ISDestroy(&crowp);CHKERRQ(ierr); 1622 } 1623 ierr = ISInvertPermutation(growp,PETSC_DECIDE,&irowp);CHKERRQ(ierr); 1624 /* get the local target indices */ 1625 ierr = MatGetOwnershipRange(A,&first,PETSC_NULL);CHKERRQ(ierr); 1626 ierr = MatGetLocalSize(A,&local_size,PETSC_NULL);CHKERRQ(ierr); 1627 ierr = ISGetIndices(irowp,&rows);CHKERRQ(ierr); 1628 ierr = ISCreateGeneral(MPI_COMM_SELF,local_size,rows+first,PETSC_COPY_VALUES,&lrowp);CHKERRQ(ierr); 1629 ierr = ISRestoreIndices(irowp,&rows);CHKERRQ(ierr); 1630 ierr = ISDestroy(&irowp);CHKERRQ(ierr); 1631 /* the column permutation is so much easier; 1632 make a local version of 'colp' and invert it */ 1633 ierr = PetscObjectGetComm((PetscObject)colp,&pcomm);CHKERRQ(ierr); 1634 ierr = MPI_Comm_size(pcomm,&size);CHKERRQ(ierr); 1635 if (size==1) { 1636 lcolp = colp; 1637 } else { 1638 ierr = ISGetSize(colp,&nrows);CHKERRQ(ierr); 1639 ierr = ISGetIndices(colp,&rows);CHKERRQ(ierr); 1640 ierr = ISCreateGeneral(MPI_COMM_SELF,nrows,rows,PETSC_COPY_VALUES,&lcolp);CHKERRQ(ierr); 1641 } 1642 ierr = ISSetPermutation(lcolp);CHKERRQ(ierr); 1643 ierr = ISInvertPermutation(lcolp,PETSC_DECIDE,&icolp);CHKERRQ(ierr); 1644 ierr = ISSetPermutation(icolp);CHKERRQ(ierr); 1645 if (size>1) { 1646 ierr = ISRestoreIndices(colp,&rows);CHKERRQ(ierr); 1647 ierr = ISDestroy(&lcolp);CHKERRQ(ierr); 1648 } 1649 /* now we just get the submatrix */ 1650 ierr = MatGetSubMatrix_MPIAIJ_Private(A,lrowp,icolp,local_size,MAT_INITIAL_MATRIX,B);CHKERRQ(ierr); 1651 /* clean up */ 1652 ierr = ISDestroy(&lrowp);CHKERRQ(ierr); 1653 ierr = ISDestroy(&icolp);CHKERRQ(ierr); 1654 PetscFunctionReturn(0); 1655 } 1656 1657 #undef __FUNCT__ 1658 #define __FUNCT__ "MatGetInfo_MPIAIJ" 1659 PetscErrorCode MatGetInfo_MPIAIJ(Mat matin,MatInfoType flag,MatInfo *info) 1660 { 1661 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 1662 Mat A = mat->A,B = mat->B; 1663 PetscErrorCode ierr; 1664 PetscReal isend[5],irecv[5]; 1665 1666 PetscFunctionBegin; 1667 info->block_size = 1.0; 1668 ierr = MatGetInfo(A,MAT_LOCAL,info);CHKERRQ(ierr); 1669 isend[0] = info->nz_used; isend[1] = info->nz_allocated; isend[2] = info->nz_unneeded; 1670 isend[3] = info->memory; isend[4] = info->mallocs; 1671 ierr = MatGetInfo(B,MAT_LOCAL,info);CHKERRQ(ierr); 1672 isend[0] += info->nz_used; isend[1] += info->nz_allocated; isend[2] += info->nz_unneeded; 1673 isend[3] += info->memory; isend[4] += info->mallocs; 1674 if (flag == MAT_LOCAL) { 1675 info->nz_used = isend[0]; 1676 info->nz_allocated = isend[1]; 1677 info->nz_unneeded = isend[2]; 1678 info->memory = isend[3]; 1679 info->mallocs = isend[4]; 1680 } else if (flag == MAT_GLOBAL_MAX) { 1681 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPIU_MAX,((PetscObject)matin)->comm);CHKERRQ(ierr); 1682 info->nz_used = irecv[0]; 1683 info->nz_allocated = irecv[1]; 1684 info->nz_unneeded = irecv[2]; 1685 info->memory = irecv[3]; 1686 info->mallocs = irecv[4]; 1687 } else if (flag == MAT_GLOBAL_SUM) { 1688 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPIU_SUM,((PetscObject)matin)->comm);CHKERRQ(ierr); 1689 info->nz_used = irecv[0]; 1690 info->nz_allocated = irecv[1]; 1691 info->nz_unneeded = irecv[2]; 1692 info->memory = irecv[3]; 1693 info->mallocs = irecv[4]; 1694 } 1695 info->fill_ratio_given = 0; /* no parallel LU/ILU/Cholesky */ 1696 info->fill_ratio_needed = 0; 1697 info->factor_mallocs = 0; 1698 1699 PetscFunctionReturn(0); 1700 } 1701 1702 #undef __FUNCT__ 1703 #define __FUNCT__ "MatSetOption_MPIAIJ" 1704 PetscErrorCode MatSetOption_MPIAIJ(Mat A,MatOption op,PetscBool flg) 1705 { 1706 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1707 PetscErrorCode ierr; 1708 1709 PetscFunctionBegin; 1710 switch (op) { 1711 case MAT_NEW_NONZERO_LOCATIONS: 1712 case MAT_NEW_NONZERO_ALLOCATION_ERR: 1713 case MAT_UNUSED_NONZERO_LOCATION_ERR: 1714 case MAT_KEEP_NONZERO_PATTERN: 1715 case MAT_NEW_NONZERO_LOCATION_ERR: 1716 case MAT_USE_INODES: 1717 case MAT_IGNORE_ZERO_ENTRIES: 1718 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1719 ierr = MatSetOption(a->B,op,flg);CHKERRQ(ierr); 1720 break; 1721 case MAT_ROW_ORIENTED: 1722 a->roworiented = flg; 1723 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1724 ierr = MatSetOption(a->B,op,flg);CHKERRQ(ierr); 1725 break; 1726 case MAT_NEW_DIAGONALS: 1727 ierr = PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);CHKERRQ(ierr); 1728 break; 1729 case MAT_IGNORE_OFF_PROC_ENTRIES: 1730 a->donotstash = PETSC_TRUE; 1731 break; 1732 case MAT_SPD: 1733 A->spd_set = PETSC_TRUE; 1734 A->spd = flg; 1735 if (flg) { 1736 A->symmetric = PETSC_TRUE; 1737 A->structurally_symmetric = PETSC_TRUE; 1738 A->symmetric_set = PETSC_TRUE; 1739 A->structurally_symmetric_set = PETSC_TRUE; 1740 } 1741 break; 1742 case MAT_SYMMETRIC: 1743 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1744 break; 1745 case MAT_STRUCTURALLY_SYMMETRIC: 1746 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1747 break; 1748 case MAT_HERMITIAN: 1749 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1750 break; 1751 case MAT_SYMMETRY_ETERNAL: 1752 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1753 break; 1754 default: 1755 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"unknown option %d",op); 1756 } 1757 PetscFunctionReturn(0); 1758 } 1759 1760 #undef __FUNCT__ 1761 #define __FUNCT__ "MatGetRow_MPIAIJ" 1762 PetscErrorCode MatGetRow_MPIAIJ(Mat matin,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 1763 { 1764 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 1765 PetscScalar *vworkA,*vworkB,**pvA,**pvB,*v_p; 1766 PetscErrorCode ierr; 1767 PetscInt i,*cworkA,*cworkB,**pcA,**pcB,cstart = matin->cmap->rstart; 1768 PetscInt nztot,nzA,nzB,lrow,rstart = matin->rmap->rstart,rend = matin->rmap->rend; 1769 PetscInt *cmap,*idx_p; 1770 1771 PetscFunctionBegin; 1772 if (mat->getrowactive) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Already active"); 1773 mat->getrowactive = PETSC_TRUE; 1774 1775 if (!mat->rowvalues && (idx || v)) { 1776 /* 1777 allocate enough space to hold information from the longest row. 1778 */ 1779 Mat_SeqAIJ *Aa = (Mat_SeqAIJ*)mat->A->data,*Ba = (Mat_SeqAIJ*)mat->B->data; 1780 PetscInt max = 1,tmp; 1781 for (i=0; i<matin->rmap->n; i++) { 1782 tmp = Aa->i[i+1] - Aa->i[i] + Ba->i[i+1] - Ba->i[i]; 1783 if (max < tmp) { max = tmp; } 1784 } 1785 ierr = PetscMalloc2(max,PetscScalar,&mat->rowvalues,max,PetscInt,&mat->rowindices);CHKERRQ(ierr); 1786 } 1787 1788 if (row < rstart || row >= rend) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Only local rows"); 1789 lrow = row - rstart; 1790 1791 pvA = &vworkA; pcA = &cworkA; pvB = &vworkB; pcB = &cworkB; 1792 if (!v) {pvA = 0; pvB = 0;} 1793 if (!idx) {pcA = 0; if (!v) pcB = 0;} 1794 ierr = (*mat->A->ops->getrow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1795 ierr = (*mat->B->ops->getrow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1796 nztot = nzA + nzB; 1797 1798 cmap = mat->garray; 1799 if (v || idx) { 1800 if (nztot) { 1801 /* Sort by increasing column numbers, assuming A and B already sorted */ 1802 PetscInt imark = -1; 1803 if (v) { 1804 *v = v_p = mat->rowvalues; 1805 for (i=0; i<nzB; i++) { 1806 if (cmap[cworkB[i]] < cstart) v_p[i] = vworkB[i]; 1807 else break; 1808 } 1809 imark = i; 1810 for (i=0; i<nzA; i++) v_p[imark+i] = vworkA[i]; 1811 for (i=imark; i<nzB; i++) v_p[nzA+i] = vworkB[i]; 1812 } 1813 if (idx) { 1814 *idx = idx_p = mat->rowindices; 1815 if (imark > -1) { 1816 for (i=0; i<imark; i++) { 1817 idx_p[i] = cmap[cworkB[i]]; 1818 } 1819 } else { 1820 for (i=0; i<nzB; i++) { 1821 if (cmap[cworkB[i]] < cstart) idx_p[i] = cmap[cworkB[i]]; 1822 else break; 1823 } 1824 imark = i; 1825 } 1826 for (i=0; i<nzA; i++) idx_p[imark+i] = cstart + cworkA[i]; 1827 for (i=imark; i<nzB; i++) idx_p[nzA+i] = cmap[cworkB[i]]; 1828 } 1829 } else { 1830 if (idx) *idx = 0; 1831 if (v) *v = 0; 1832 } 1833 } 1834 *nz = nztot; 1835 ierr = (*mat->A->ops->restorerow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1836 ierr = (*mat->B->ops->restorerow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1837 PetscFunctionReturn(0); 1838 } 1839 1840 #undef __FUNCT__ 1841 #define __FUNCT__ "MatRestoreRow_MPIAIJ" 1842 PetscErrorCode MatRestoreRow_MPIAIJ(Mat mat,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 1843 { 1844 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1845 1846 PetscFunctionBegin; 1847 if (!aij->getrowactive) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"MatGetRow() must be called first"); 1848 aij->getrowactive = PETSC_FALSE; 1849 PetscFunctionReturn(0); 1850 } 1851 1852 #undef __FUNCT__ 1853 #define __FUNCT__ "MatNorm_MPIAIJ" 1854 PetscErrorCode MatNorm_MPIAIJ(Mat mat,NormType type,PetscReal *norm) 1855 { 1856 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1857 Mat_SeqAIJ *amat = (Mat_SeqAIJ*)aij->A->data,*bmat = (Mat_SeqAIJ*)aij->B->data; 1858 PetscErrorCode ierr; 1859 PetscInt i,j,cstart = mat->cmap->rstart; 1860 PetscReal sum = 0.0; 1861 MatScalar *v; 1862 1863 PetscFunctionBegin; 1864 if (aij->size == 1) { 1865 ierr = MatNorm(aij->A,type,norm);CHKERRQ(ierr); 1866 } else { 1867 if (type == NORM_FROBENIUS) { 1868 v = amat->a; 1869 for (i=0; i<amat->nz; i++) { 1870 #if defined(PETSC_USE_COMPLEX) 1871 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 1872 #else 1873 sum += (*v)*(*v); v++; 1874 #endif 1875 } 1876 v = bmat->a; 1877 for (i=0; i<bmat->nz; i++) { 1878 #if defined(PETSC_USE_COMPLEX) 1879 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 1880 #else 1881 sum += (*v)*(*v); v++; 1882 #endif 1883 } 1884 ierr = MPI_Allreduce(&sum,norm,1,MPIU_REAL,MPIU_SUM,((PetscObject)mat)->comm);CHKERRQ(ierr); 1885 *norm = sqrt(*norm); 1886 } else if (type == NORM_1) { /* max column norm */ 1887 PetscReal *tmp,*tmp2; 1888 PetscInt *jj,*garray = aij->garray; 1889 ierr = PetscMalloc((mat->cmap->N+1)*sizeof(PetscReal),&tmp);CHKERRQ(ierr); 1890 ierr = PetscMalloc((mat->cmap->N+1)*sizeof(PetscReal),&tmp2);CHKERRQ(ierr); 1891 ierr = PetscMemzero(tmp,mat->cmap->N*sizeof(PetscReal));CHKERRQ(ierr); 1892 *norm = 0.0; 1893 v = amat->a; jj = amat->j; 1894 for (j=0; j<amat->nz; j++) { 1895 tmp[cstart + *jj++ ] += PetscAbsScalar(*v); v++; 1896 } 1897 v = bmat->a; jj = bmat->j; 1898 for (j=0; j<bmat->nz; j++) { 1899 tmp[garray[*jj++]] += PetscAbsScalar(*v); v++; 1900 } 1901 ierr = MPI_Allreduce(tmp,tmp2,mat->cmap->N,MPIU_REAL,MPIU_SUM,((PetscObject)mat)->comm);CHKERRQ(ierr); 1902 for (j=0; j<mat->cmap->N; j++) { 1903 if (tmp2[j] > *norm) *norm = tmp2[j]; 1904 } 1905 ierr = PetscFree(tmp);CHKERRQ(ierr); 1906 ierr = PetscFree(tmp2);CHKERRQ(ierr); 1907 } else if (type == NORM_INFINITY) { /* max row norm */ 1908 PetscReal ntemp = 0.0; 1909 for (j=0; j<aij->A->rmap->n; j++) { 1910 v = amat->a + amat->i[j]; 1911 sum = 0.0; 1912 for (i=0; i<amat->i[j+1]-amat->i[j]; i++) { 1913 sum += PetscAbsScalar(*v); v++; 1914 } 1915 v = bmat->a + bmat->i[j]; 1916 for (i=0; i<bmat->i[j+1]-bmat->i[j]; i++) { 1917 sum += PetscAbsScalar(*v); v++; 1918 } 1919 if (sum > ntemp) ntemp = sum; 1920 } 1921 ierr = MPI_Allreduce(&ntemp,norm,1,MPIU_REAL,MPIU_MAX,((PetscObject)mat)->comm);CHKERRQ(ierr); 1922 } else { 1923 SETERRQ(((PetscObject)mat)->comm,PETSC_ERR_SUP,"No support for two norm"); 1924 } 1925 } 1926 PetscFunctionReturn(0); 1927 } 1928 1929 #undef __FUNCT__ 1930 #define __FUNCT__ "MatTranspose_MPIAIJ" 1931 PetscErrorCode MatTranspose_MPIAIJ(Mat A,MatReuse reuse,Mat *matout) 1932 { 1933 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1934 Mat_SeqAIJ *Aloc=(Mat_SeqAIJ*)a->A->data,*Bloc=(Mat_SeqAIJ*)a->B->data; 1935 PetscErrorCode ierr; 1936 PetscInt M = A->rmap->N,N = A->cmap->N,ma,na,mb,*ai,*aj,*bi,*bj,row,*cols,*cols_tmp,i,*d_nnz; 1937 PetscInt cstart=A->cmap->rstart,ncol; 1938 Mat B; 1939 MatScalar *array; 1940 1941 PetscFunctionBegin; 1942 if (reuse == MAT_REUSE_MATRIX && A == *matout && M != N) SETERRQ(((PetscObject)A)->comm,PETSC_ERR_ARG_SIZ,"Square matrix only for in-place"); 1943 1944 ma = A->rmap->n; na = A->cmap->n; mb = a->B->rmap->n; 1945 ai = Aloc->i; aj = Aloc->j; 1946 bi = Bloc->i; bj = Bloc->j; 1947 if (reuse == MAT_INITIAL_MATRIX || *matout == A) { 1948 /* compute d_nnz for preallocation; o_nnz is approximated by d_nnz to avoid communication */ 1949 ierr = PetscMalloc((1+na)*sizeof(PetscInt),&d_nnz);CHKERRQ(ierr); 1950 ierr = PetscMemzero(d_nnz,(1+na)*sizeof(PetscInt));CHKERRQ(ierr); 1951 for (i=0; i<ai[ma]; i++){ 1952 d_nnz[aj[i]] ++; 1953 aj[i] += cstart; /* global col index to be used by MatSetValues() */ 1954 } 1955 1956 ierr = MatCreate(((PetscObject)A)->comm,&B);CHKERRQ(ierr); 1957 ierr = MatSetSizes(B,A->cmap->n,A->rmap->n,N,M);CHKERRQ(ierr); 1958 ierr = MatSetType(B,((PetscObject)A)->type_name);CHKERRQ(ierr); 1959 ierr = MatMPIAIJSetPreallocation(B,0,d_nnz,0,d_nnz);CHKERRQ(ierr); 1960 ierr = PetscFree(d_nnz);CHKERRQ(ierr); 1961 } else { 1962 B = *matout; 1963 } 1964 1965 /* copy over the A part */ 1966 array = Aloc->a; 1967 row = A->rmap->rstart; 1968 for (i=0; i<ma; i++) { 1969 ncol = ai[i+1]-ai[i]; 1970 ierr = MatSetValues(B,ncol,aj,1,&row,array,INSERT_VALUES);CHKERRQ(ierr); 1971 row++; array += ncol; aj += ncol; 1972 } 1973 aj = Aloc->j; 1974 for (i=0; i<ai[ma]; i++) aj[i] -= cstart; /* resume local col index */ 1975 1976 /* copy over the B part */ 1977 ierr = PetscMalloc(bi[mb]*sizeof(PetscInt),&cols);CHKERRQ(ierr); 1978 ierr = PetscMemzero(cols,bi[mb]*sizeof(PetscInt));CHKERRQ(ierr); 1979 array = Bloc->a; 1980 row = A->rmap->rstart; 1981 for (i=0; i<bi[mb]; i++) {cols[i] = a->garray[bj[i]];} 1982 cols_tmp = cols; 1983 for (i=0; i<mb; i++) { 1984 ncol = bi[i+1]-bi[i]; 1985 ierr = MatSetValues(B,ncol,cols_tmp,1,&row,array,INSERT_VALUES);CHKERRQ(ierr); 1986 row++; array += ncol; cols_tmp += ncol; 1987 } 1988 ierr = PetscFree(cols);CHKERRQ(ierr); 1989 1990 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1991 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1992 if (reuse == MAT_INITIAL_MATRIX || *matout != A) { 1993 *matout = B; 1994 } else { 1995 ierr = MatHeaderMerge(A,B);CHKERRQ(ierr); 1996 } 1997 PetscFunctionReturn(0); 1998 } 1999 2000 #undef __FUNCT__ 2001 #define __FUNCT__ "MatDiagonalScale_MPIAIJ" 2002 PetscErrorCode MatDiagonalScale_MPIAIJ(Mat mat,Vec ll,Vec rr) 2003 { 2004 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 2005 Mat a = aij->A,b = aij->B; 2006 PetscErrorCode ierr; 2007 PetscInt s1,s2,s3; 2008 2009 PetscFunctionBegin; 2010 ierr = MatGetLocalSize(mat,&s2,&s3);CHKERRQ(ierr); 2011 if (rr) { 2012 ierr = VecGetLocalSize(rr,&s1);CHKERRQ(ierr); 2013 if (s1!=s3) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"right vector non-conforming local size"); 2014 /* Overlap communication with computation. */ 2015 ierr = VecScatterBegin(aij->Mvctx,rr,aij->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2016 } 2017 if (ll) { 2018 ierr = VecGetLocalSize(ll,&s1);CHKERRQ(ierr); 2019 if (s1!=s2) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"left vector non-conforming local size"); 2020 ierr = (*b->ops->diagonalscale)(b,ll,0);CHKERRQ(ierr); 2021 } 2022 /* scale the diagonal block */ 2023 ierr = (*a->ops->diagonalscale)(a,ll,rr);CHKERRQ(ierr); 2024 2025 if (rr) { 2026 /* Do a scatter end and then right scale the off-diagonal block */ 2027 ierr = VecScatterEnd(aij->Mvctx,rr,aij->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2028 ierr = (*b->ops->diagonalscale)(b,0,aij->lvec);CHKERRQ(ierr); 2029 } 2030 2031 PetscFunctionReturn(0); 2032 } 2033 2034 #undef __FUNCT__ 2035 #define __FUNCT__ "MatSetBlockSize_MPIAIJ" 2036 PetscErrorCode MatSetBlockSize_MPIAIJ(Mat A,PetscInt bs) 2037 { 2038 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2039 PetscErrorCode ierr; 2040 2041 PetscFunctionBegin; 2042 ierr = MatSetBlockSize(a->A,bs);CHKERRQ(ierr); 2043 ierr = MatSetBlockSize(a->B,bs);CHKERRQ(ierr); 2044 ierr = PetscLayoutSetBlockSize(A->rmap,bs);CHKERRQ(ierr); 2045 ierr = PetscLayoutSetBlockSize(A->cmap,bs);CHKERRQ(ierr); 2046 PetscFunctionReturn(0); 2047 } 2048 #undef __FUNCT__ 2049 #define __FUNCT__ "MatSetUnfactored_MPIAIJ" 2050 PetscErrorCode MatSetUnfactored_MPIAIJ(Mat A) 2051 { 2052 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2053 PetscErrorCode ierr; 2054 2055 PetscFunctionBegin; 2056 ierr = MatSetUnfactored(a->A);CHKERRQ(ierr); 2057 PetscFunctionReturn(0); 2058 } 2059 2060 #undef __FUNCT__ 2061 #define __FUNCT__ "MatEqual_MPIAIJ" 2062 PetscErrorCode MatEqual_MPIAIJ(Mat A,Mat B,PetscBool *flag) 2063 { 2064 Mat_MPIAIJ *matB = (Mat_MPIAIJ*)B->data,*matA = (Mat_MPIAIJ*)A->data; 2065 Mat a,b,c,d; 2066 PetscBool flg; 2067 PetscErrorCode ierr; 2068 2069 PetscFunctionBegin; 2070 a = matA->A; b = matA->B; 2071 c = matB->A; d = matB->B; 2072 2073 ierr = MatEqual(a,c,&flg);CHKERRQ(ierr); 2074 if (flg) { 2075 ierr = MatEqual(b,d,&flg);CHKERRQ(ierr); 2076 } 2077 ierr = MPI_Allreduce(&flg,flag,1,MPI_INT,MPI_LAND,((PetscObject)A)->comm);CHKERRQ(ierr); 2078 PetscFunctionReturn(0); 2079 } 2080 2081 #undef __FUNCT__ 2082 #define __FUNCT__ "MatCopy_MPIAIJ" 2083 PetscErrorCode MatCopy_MPIAIJ(Mat A,Mat B,MatStructure str) 2084 { 2085 PetscErrorCode ierr; 2086 Mat_MPIAIJ *a = (Mat_MPIAIJ *)A->data; 2087 Mat_MPIAIJ *b = (Mat_MPIAIJ *)B->data; 2088 2089 PetscFunctionBegin; 2090 /* If the two matrices don't have the same copy implementation, they aren't compatible for fast copy. */ 2091 if ((str != SAME_NONZERO_PATTERN) || (A->ops->copy != B->ops->copy)) { 2092 /* because of the column compression in the off-processor part of the matrix a->B, 2093 the number of columns in a->B and b->B may be different, hence we cannot call 2094 the MatCopy() directly on the two parts. If need be, we can provide a more 2095 efficient copy than the MatCopy_Basic() by first uncompressing the a->B matrices 2096 then copying the submatrices */ 2097 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 2098 } else { 2099 ierr = MatCopy(a->A,b->A,str);CHKERRQ(ierr); 2100 ierr = MatCopy(a->B,b->B,str);CHKERRQ(ierr); 2101 } 2102 PetscFunctionReturn(0); 2103 } 2104 2105 #undef __FUNCT__ 2106 #define __FUNCT__ "MatSetUpPreallocation_MPIAIJ" 2107 PetscErrorCode MatSetUpPreallocation_MPIAIJ(Mat A) 2108 { 2109 PetscErrorCode ierr; 2110 2111 PetscFunctionBegin; 2112 ierr = MatMPIAIJSetPreallocation(A,PETSC_DEFAULT,0,PETSC_DEFAULT,0);CHKERRQ(ierr); 2113 PetscFunctionReturn(0); 2114 } 2115 2116 #undef __FUNCT__ 2117 #define __FUNCT__ "MatAXPY_MPIAIJ" 2118 PetscErrorCode MatAXPY_MPIAIJ(Mat Y,PetscScalar a,Mat X,MatStructure str) 2119 { 2120 PetscErrorCode ierr; 2121 PetscInt i; 2122 Mat_MPIAIJ *xx = (Mat_MPIAIJ *)X->data,*yy = (Mat_MPIAIJ *)Y->data; 2123 PetscBLASInt bnz,one=1; 2124 Mat_SeqAIJ *x,*y; 2125 2126 PetscFunctionBegin; 2127 if (str == SAME_NONZERO_PATTERN) { 2128 PetscScalar alpha = a; 2129 x = (Mat_SeqAIJ *)xx->A->data; 2130 y = (Mat_SeqAIJ *)yy->A->data; 2131 bnz = PetscBLASIntCast(x->nz); 2132 BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one); 2133 x = (Mat_SeqAIJ *)xx->B->data; 2134 y = (Mat_SeqAIJ *)yy->B->data; 2135 bnz = PetscBLASIntCast(x->nz); 2136 BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one); 2137 } else if (str == SUBSET_NONZERO_PATTERN) { 2138 ierr = MatAXPY_SeqAIJ(yy->A,a,xx->A,str);CHKERRQ(ierr); 2139 2140 x = (Mat_SeqAIJ *)xx->B->data; 2141 y = (Mat_SeqAIJ *)yy->B->data; 2142 if (y->xtoy && y->XtoY != xx->B) { 2143 ierr = PetscFree(y->xtoy);CHKERRQ(ierr); 2144 ierr = MatDestroy(&y->XtoY);CHKERRQ(ierr); 2145 } 2146 if (!y->xtoy) { /* get xtoy */ 2147 ierr = MatAXPYGetxtoy_Private(xx->B->rmap->n,x->i,x->j,xx->garray,y->i,y->j,yy->garray,&y->xtoy);CHKERRQ(ierr); 2148 y->XtoY = xx->B; 2149 ierr = PetscObjectReference((PetscObject)xx->B);CHKERRQ(ierr); 2150 } 2151 for (i=0; i<x->nz; i++) y->a[y->xtoy[i]] += a*(x->a[i]); 2152 } else { 2153 Mat B; 2154 PetscInt *nnz_d,*nnz_o; 2155 ierr = PetscMalloc(yy->A->rmap->N*sizeof(PetscInt),&nnz_d);CHKERRQ(ierr); 2156 ierr = PetscMalloc(yy->B->rmap->N*sizeof(PetscInt),&nnz_o);CHKERRQ(ierr); 2157 ierr = MatCreate(((PetscObject)Y)->comm,&B);CHKERRQ(ierr); 2158 ierr = PetscObjectSetName((PetscObject)B,((PetscObject)Y)->name);CHKERRQ(ierr); 2159 ierr = MatSetSizes(B,Y->rmap->n,Y->cmap->n,Y->rmap->N,Y->cmap->N);CHKERRQ(ierr); 2160 ierr = MatSetType(B,MATMPIAIJ);CHKERRQ(ierr); 2161 ierr = MatAXPYGetPreallocation_SeqAIJ(yy->A,xx->A,nnz_d);CHKERRQ(ierr); 2162 ierr = MatAXPYGetPreallocation_SeqAIJ(yy->B,xx->B,nnz_o);CHKERRQ(ierr); 2163 ierr = MatMPIAIJSetPreallocation(B,PETSC_NULL,nnz_d,PETSC_NULL,nnz_o);CHKERRQ(ierr); 2164 ierr = MatAXPY_BasicWithPreallocation(B,Y,a,X,str);CHKERRQ(ierr); 2165 ierr = MatHeaderReplace(Y,B); 2166 ierr = PetscFree(nnz_d);CHKERRQ(ierr); 2167 ierr = PetscFree(nnz_o);CHKERRQ(ierr); 2168 } 2169 PetscFunctionReturn(0); 2170 } 2171 2172 extern PetscErrorCode MatConjugate_SeqAIJ(Mat); 2173 2174 #undef __FUNCT__ 2175 #define __FUNCT__ "MatConjugate_MPIAIJ" 2176 PetscErrorCode MatConjugate_MPIAIJ(Mat mat) 2177 { 2178 #if defined(PETSC_USE_COMPLEX) 2179 PetscErrorCode ierr; 2180 Mat_MPIAIJ *aij = (Mat_MPIAIJ *)mat->data; 2181 2182 PetscFunctionBegin; 2183 ierr = MatConjugate_SeqAIJ(aij->A);CHKERRQ(ierr); 2184 ierr = MatConjugate_SeqAIJ(aij->B);CHKERRQ(ierr); 2185 #else 2186 PetscFunctionBegin; 2187 #endif 2188 PetscFunctionReturn(0); 2189 } 2190 2191 #undef __FUNCT__ 2192 #define __FUNCT__ "MatRealPart_MPIAIJ" 2193 PetscErrorCode MatRealPart_MPIAIJ(Mat A) 2194 { 2195 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2196 PetscErrorCode ierr; 2197 2198 PetscFunctionBegin; 2199 ierr = MatRealPart(a->A);CHKERRQ(ierr); 2200 ierr = MatRealPart(a->B);CHKERRQ(ierr); 2201 PetscFunctionReturn(0); 2202 } 2203 2204 #undef __FUNCT__ 2205 #define __FUNCT__ "MatImaginaryPart_MPIAIJ" 2206 PetscErrorCode MatImaginaryPart_MPIAIJ(Mat A) 2207 { 2208 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2209 PetscErrorCode ierr; 2210 2211 PetscFunctionBegin; 2212 ierr = MatImaginaryPart(a->A);CHKERRQ(ierr); 2213 ierr = MatImaginaryPart(a->B);CHKERRQ(ierr); 2214 PetscFunctionReturn(0); 2215 } 2216 2217 #ifdef PETSC_HAVE_PBGL 2218 2219 #include <boost/parallel/mpi/bsp_process_group.hpp> 2220 #include <boost/graph/distributed/ilu_default_graph.hpp> 2221 #include <boost/graph/distributed/ilu_0_block.hpp> 2222 #include <boost/graph/distributed/ilu_preconditioner.hpp> 2223 #include <boost/graph/distributed/petsc/interface.hpp> 2224 #include <boost/multi_array.hpp> 2225 #include <boost/parallel/distributed_property_map->hpp> 2226 2227 #undef __FUNCT__ 2228 #define __FUNCT__ "MatILUFactorSymbolic_MPIAIJ" 2229 /* 2230 This uses the parallel ILU factorization of Peter Gottschling <pgottsch@osl.iu.edu> 2231 */ 2232 PetscErrorCode MatILUFactorSymbolic_MPIAIJ(Mat fact,Mat A, IS isrow, IS iscol, const MatFactorInfo *info) 2233 { 2234 namespace petsc = boost::distributed::petsc; 2235 2236 namespace graph_dist = boost::graph::distributed; 2237 using boost::graph::distributed::ilu_default::process_group_type; 2238 using boost::graph::ilu_permuted; 2239 2240 PetscBool row_identity, col_identity; 2241 PetscContainer c; 2242 PetscInt m, n, M, N; 2243 PetscErrorCode ierr; 2244 2245 PetscFunctionBegin; 2246 if (info->levels != 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only levels = 0 supported for parallel ilu"); 2247 ierr = ISIdentity(isrow, &row_identity);CHKERRQ(ierr); 2248 ierr = ISIdentity(iscol, &col_identity);CHKERRQ(ierr); 2249 if (!row_identity || !col_identity) { 2250 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Row and column permutations must be identity for parallel ILU"); 2251 } 2252 2253 process_group_type pg; 2254 typedef graph_dist::ilu_default::ilu_level_graph_type lgraph_type; 2255 lgraph_type* lgraph_p = new lgraph_type(petsc::num_global_vertices(A), pg, petsc::matrix_distribution(A, pg)); 2256 lgraph_type& level_graph = *lgraph_p; 2257 graph_dist::ilu_default::graph_type& graph(level_graph.graph); 2258 2259 petsc::read_matrix(A, graph, get(boost::edge_weight, graph)); 2260 ilu_permuted(level_graph); 2261 2262 /* put together the new matrix */ 2263 ierr = MatCreate(((PetscObject)A)->comm, fact);CHKERRQ(ierr); 2264 ierr = MatGetLocalSize(A, &m, &n);CHKERRQ(ierr); 2265 ierr = MatGetSize(A, &M, &N);CHKERRQ(ierr); 2266 ierr = MatSetSizes(fact, m, n, M, N);CHKERRQ(ierr); 2267 ierr = MatSetType(fact, ((PetscObject)A)->type_name);CHKERRQ(ierr); 2268 ierr = MatAssemblyBegin(fact, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2269 ierr = MatAssemblyEnd(fact, MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2270 2271 ierr = PetscContainerCreate(((PetscObject)A)->comm, &c); 2272 ierr = PetscContainerSetPointer(c, lgraph_p); 2273 ierr = PetscObjectCompose((PetscObject) (fact), "graph", (PetscObject) c); 2274 PetscFunctionReturn(0); 2275 } 2276 2277 #undef __FUNCT__ 2278 #define __FUNCT__ "MatLUFactorNumeric_MPIAIJ" 2279 PetscErrorCode MatLUFactorNumeric_MPIAIJ(Mat B,Mat A, const MatFactorInfo *info) 2280 { 2281 PetscFunctionBegin; 2282 PetscFunctionReturn(0); 2283 } 2284 2285 #undef __FUNCT__ 2286 #define __FUNCT__ "MatSolve_MPIAIJ" 2287 /* 2288 This uses the parallel ILU factorization of Peter Gottschling <pgottsch@osl.iu.edu> 2289 */ 2290 PetscErrorCode MatSolve_MPIAIJ(Mat A, Vec b, Vec x) 2291 { 2292 namespace graph_dist = boost::graph::distributed; 2293 2294 typedef graph_dist::ilu_default::ilu_level_graph_type lgraph_type; 2295 lgraph_type* lgraph_p; 2296 PetscContainer c; 2297 PetscErrorCode ierr; 2298 2299 PetscFunctionBegin; 2300 ierr = PetscObjectQuery((PetscObject) A, "graph", (PetscObject *) &c);CHKERRQ(ierr); 2301 ierr = PetscContainerGetPointer(c, (void **) &lgraph_p);CHKERRQ(ierr); 2302 ierr = VecCopy(b, x);CHKERRQ(ierr); 2303 2304 PetscScalar* array_x; 2305 ierr = VecGetArray(x, &array_x);CHKERRQ(ierr); 2306 PetscInt sx; 2307 ierr = VecGetSize(x, &sx);CHKERRQ(ierr); 2308 2309 PetscScalar* array_b; 2310 ierr = VecGetArray(b, &array_b);CHKERRQ(ierr); 2311 PetscInt sb; 2312 ierr = VecGetSize(b, &sb);CHKERRQ(ierr); 2313 2314 lgraph_type& level_graph = *lgraph_p; 2315 graph_dist::ilu_default::graph_type& graph(level_graph.graph); 2316 2317 typedef boost::multi_array_ref<PetscScalar, 1> array_ref_type; 2318 array_ref_type ref_b(array_b, boost::extents[num_vertices(graph)]), 2319 ref_x(array_x, boost::extents[num_vertices(graph)]); 2320 2321 typedef boost::iterator_property_map<array_ref_type::iterator, 2322 boost::property_map<graph_dist::ilu_default::graph_type, boost::vertex_index_t>::type> gvector_type; 2323 gvector_type vector_b(ref_b.begin(), get(boost::vertex_index, graph)), 2324 vector_x(ref_x.begin(), get(boost::vertex_index, graph)); 2325 2326 ilu_set_solve(*lgraph_p, vector_b, vector_x); 2327 2328 PetscFunctionReturn(0); 2329 } 2330 #endif 2331 2332 typedef struct { /* used by MatGetRedundantMatrix() for reusing matredundant */ 2333 PetscInt nzlocal,nsends,nrecvs; 2334 PetscMPIInt *send_rank,*recv_rank; 2335 PetscInt *sbuf_nz,*rbuf_nz,*sbuf_j,**rbuf_j; 2336 PetscScalar *sbuf_a,**rbuf_a; 2337 PetscErrorCode (*MatDestroy)(Mat); 2338 } Mat_Redundant; 2339 2340 #undef __FUNCT__ 2341 #define __FUNCT__ "PetscContainerDestroy_MatRedundant" 2342 PetscErrorCode PetscContainerDestroy_MatRedundant(void *ptr) 2343 { 2344 PetscErrorCode ierr; 2345 Mat_Redundant *redund=(Mat_Redundant*)ptr; 2346 PetscInt i; 2347 2348 PetscFunctionBegin; 2349 ierr = PetscFree2(redund->send_rank,redund->recv_rank);CHKERRQ(ierr); 2350 ierr = PetscFree(redund->sbuf_j);CHKERRQ(ierr); 2351 ierr = PetscFree(redund->sbuf_a);CHKERRQ(ierr); 2352 for (i=0; i<redund->nrecvs; i++){ 2353 ierr = PetscFree(redund->rbuf_j[i]);CHKERRQ(ierr); 2354 ierr = PetscFree(redund->rbuf_a[i]);CHKERRQ(ierr); 2355 } 2356 ierr = PetscFree4(redund->sbuf_nz,redund->rbuf_nz,redund->rbuf_j,redund->rbuf_a);CHKERRQ(ierr); 2357 ierr = PetscFree(redund);CHKERRQ(ierr); 2358 PetscFunctionReturn(0); 2359 } 2360 2361 #undef __FUNCT__ 2362 #define __FUNCT__ "MatDestroy_MatRedundant" 2363 PetscErrorCode MatDestroy_MatRedundant(Mat A) 2364 { 2365 PetscErrorCode ierr; 2366 PetscContainer container; 2367 Mat_Redundant *redund=PETSC_NULL; 2368 2369 PetscFunctionBegin; 2370 ierr = PetscObjectQuery((PetscObject)A,"Mat_Redundant",(PetscObject *)&container);CHKERRQ(ierr); 2371 if (container) { 2372 ierr = PetscContainerGetPointer(container,(void **)&redund);CHKERRQ(ierr); 2373 } else { 2374 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Container does not exit"); 2375 } 2376 A->ops->destroy = redund->MatDestroy; 2377 ierr = PetscObjectCompose((PetscObject)A,"Mat_Redundant",0);CHKERRQ(ierr); 2378 ierr = (*A->ops->destroy)(A);CHKERRQ(ierr); 2379 ierr = PetscContainerDestroy(&container);CHKERRQ(ierr); 2380 PetscFunctionReturn(0); 2381 } 2382 2383 #undef __FUNCT__ 2384 #define __FUNCT__ "MatGetRedundantMatrix_MPIAIJ" 2385 PetscErrorCode MatGetRedundantMatrix_MPIAIJ(Mat mat,PetscInt nsubcomm,MPI_Comm subcomm,PetscInt mlocal_sub,MatReuse reuse,Mat *matredundant) 2386 { 2387 PetscMPIInt rank,size; 2388 MPI_Comm comm=((PetscObject)mat)->comm; 2389 PetscErrorCode ierr; 2390 PetscInt nsends=0,nrecvs=0,i,rownz_max=0; 2391 PetscMPIInt *send_rank=PETSC_NULL,*recv_rank=PETSC_NULL; 2392 PetscInt *rowrange=mat->rmap->range; 2393 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 2394 Mat A=aij->A,B=aij->B,C=*matredundant; 2395 Mat_SeqAIJ *a=(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)B->data; 2396 PetscScalar *sbuf_a; 2397 PetscInt nzlocal=a->nz+b->nz; 2398 PetscInt j,cstart=mat->cmap->rstart,cend=mat->cmap->rend,row,nzA,nzB,ncols,*cworkA,*cworkB; 2399 PetscInt rstart=mat->rmap->rstart,rend=mat->rmap->rend,*bmap=aij->garray,M,N; 2400 PetscInt *cols,ctmp,lwrite,*rptr,l,*sbuf_j; 2401 MatScalar *aworkA,*aworkB; 2402 PetscScalar *vals; 2403 PetscMPIInt tag1,tag2,tag3,imdex; 2404 MPI_Request *s_waits1=PETSC_NULL,*s_waits2=PETSC_NULL,*s_waits3=PETSC_NULL, 2405 *r_waits1=PETSC_NULL,*r_waits2=PETSC_NULL,*r_waits3=PETSC_NULL; 2406 MPI_Status recv_status,*send_status; 2407 PetscInt *sbuf_nz=PETSC_NULL,*rbuf_nz=PETSC_NULL,count; 2408 PetscInt **rbuf_j=PETSC_NULL; 2409 PetscScalar **rbuf_a=PETSC_NULL; 2410 Mat_Redundant *redund=PETSC_NULL; 2411 PetscContainer container; 2412 2413 PetscFunctionBegin; 2414 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 2415 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 2416 2417 if (reuse == MAT_REUSE_MATRIX) { 2418 ierr = MatGetSize(C,&M,&N);CHKERRQ(ierr); 2419 if (M != N || M != mat->rmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. Wrong global size"); 2420 ierr = MatGetLocalSize(C,&M,&N);CHKERRQ(ierr); 2421 if (M != N || M != mlocal_sub) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. Wrong local size"); 2422 ierr = PetscObjectQuery((PetscObject)C,"Mat_Redundant",(PetscObject *)&container);CHKERRQ(ierr); 2423 if (container) { 2424 ierr = PetscContainerGetPointer(container,(void **)&redund);CHKERRQ(ierr); 2425 } else { 2426 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Container does not exit"); 2427 } 2428 if (nzlocal != redund->nzlocal) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. Wrong nzlocal"); 2429 2430 nsends = redund->nsends; 2431 nrecvs = redund->nrecvs; 2432 send_rank = redund->send_rank; 2433 recv_rank = redund->recv_rank; 2434 sbuf_nz = redund->sbuf_nz; 2435 rbuf_nz = redund->rbuf_nz; 2436 sbuf_j = redund->sbuf_j; 2437 sbuf_a = redund->sbuf_a; 2438 rbuf_j = redund->rbuf_j; 2439 rbuf_a = redund->rbuf_a; 2440 } 2441 2442 if (reuse == MAT_INITIAL_MATRIX){ 2443 PetscMPIInt subrank,subsize; 2444 PetscInt nleftover,np_subcomm; 2445 /* get the destination processors' id send_rank, nsends and nrecvs */ 2446 ierr = MPI_Comm_rank(subcomm,&subrank);CHKERRQ(ierr); 2447 ierr = MPI_Comm_size(subcomm,&subsize);CHKERRQ(ierr); 2448 ierr = PetscMalloc2(size,PetscMPIInt,&send_rank,size,PetscMPIInt,&recv_rank); 2449 np_subcomm = size/nsubcomm; 2450 nleftover = size - nsubcomm*np_subcomm; 2451 nsends = 0; nrecvs = 0; 2452 for (i=0; i<size; i++){ /* i=rank*/ 2453 if (subrank == i/nsubcomm && rank != i){ /* my_subrank == other's subrank */ 2454 send_rank[nsends] = i; nsends++; 2455 recv_rank[nrecvs++] = i; 2456 } 2457 } 2458 if (rank >= size - nleftover){/* this proc is a leftover processor */ 2459 i = size-nleftover-1; 2460 j = 0; 2461 while (j < nsubcomm - nleftover){ 2462 send_rank[nsends++] = i; 2463 i--; j++; 2464 } 2465 } 2466 2467 if (nleftover && subsize == size/nsubcomm && subrank==subsize-1){ /* this proc recvs from leftover processors */ 2468 for (i=0; i<nleftover; i++){ 2469 recv_rank[nrecvs++] = size-nleftover+i; 2470 } 2471 } 2472 2473 /* allocate sbuf_j, sbuf_a */ 2474 i = nzlocal + rowrange[rank+1] - rowrange[rank] + 2; 2475 ierr = PetscMalloc(i*sizeof(PetscInt),&sbuf_j);CHKERRQ(ierr); 2476 ierr = PetscMalloc((nzlocal+1)*sizeof(PetscScalar),&sbuf_a);CHKERRQ(ierr); 2477 } /* endof if (reuse == MAT_INITIAL_MATRIX) */ 2478 2479 /* copy mat's local entries into the buffers */ 2480 if (reuse == MAT_INITIAL_MATRIX){ 2481 rownz_max = 0; 2482 rptr = sbuf_j; 2483 cols = sbuf_j + rend-rstart + 1; 2484 vals = sbuf_a; 2485 rptr[0] = 0; 2486 for (i=0; i<rend-rstart; i++){ 2487 row = i + rstart; 2488 nzA = a->i[i+1] - a->i[i]; nzB = b->i[i+1] - b->i[i]; 2489 ncols = nzA + nzB; 2490 cworkA = a->j + a->i[i]; cworkB = b->j + b->i[i]; 2491 aworkA = a->a + a->i[i]; aworkB = b->a + b->i[i]; 2492 /* load the column indices for this row into cols */ 2493 lwrite = 0; 2494 for (l=0; l<nzB; l++) { 2495 if ((ctmp = bmap[cworkB[l]]) < cstart){ 2496 vals[lwrite] = aworkB[l]; 2497 cols[lwrite++] = ctmp; 2498 } 2499 } 2500 for (l=0; l<nzA; l++){ 2501 vals[lwrite] = aworkA[l]; 2502 cols[lwrite++] = cstart + cworkA[l]; 2503 } 2504 for (l=0; l<nzB; l++) { 2505 if ((ctmp = bmap[cworkB[l]]) >= cend){ 2506 vals[lwrite] = aworkB[l]; 2507 cols[lwrite++] = ctmp; 2508 } 2509 } 2510 vals += ncols; 2511 cols += ncols; 2512 rptr[i+1] = rptr[i] + ncols; 2513 if (rownz_max < ncols) rownz_max = ncols; 2514 } 2515 if (rptr[rend-rstart] != a->nz + b->nz) SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_PLIB, "rptr[%d] %d != %d + %d",rend-rstart,rptr[rend-rstart+1],a->nz,b->nz); 2516 } else { /* only copy matrix values into sbuf_a */ 2517 rptr = sbuf_j; 2518 vals = sbuf_a; 2519 rptr[0] = 0; 2520 for (i=0; i<rend-rstart; i++){ 2521 row = i + rstart; 2522 nzA = a->i[i+1] - a->i[i]; nzB = b->i[i+1] - b->i[i]; 2523 ncols = nzA + nzB; 2524 cworkA = a->j + a->i[i]; cworkB = b->j + b->i[i]; 2525 aworkA = a->a + a->i[i]; aworkB = b->a + b->i[i]; 2526 lwrite = 0; 2527 for (l=0; l<nzB; l++) { 2528 if ((ctmp = bmap[cworkB[l]]) < cstart) vals[lwrite++] = aworkB[l]; 2529 } 2530 for (l=0; l<nzA; l++) vals[lwrite++] = aworkA[l]; 2531 for (l=0; l<nzB; l++) { 2532 if ((ctmp = bmap[cworkB[l]]) >= cend) vals[lwrite++] = aworkB[l]; 2533 } 2534 vals += ncols; 2535 rptr[i+1] = rptr[i] + ncols; 2536 } 2537 } /* endof if (reuse == MAT_INITIAL_MATRIX) */ 2538 2539 /* send nzlocal to others, and recv other's nzlocal */ 2540 /*--------------------------------------------------*/ 2541 if (reuse == MAT_INITIAL_MATRIX){ 2542 ierr = PetscMalloc2(3*(nsends + nrecvs)+1,MPI_Request,&s_waits3,nsends+1,MPI_Status,&send_status);CHKERRQ(ierr); 2543 s_waits2 = s_waits3 + nsends; 2544 s_waits1 = s_waits2 + nsends; 2545 r_waits1 = s_waits1 + nsends; 2546 r_waits2 = r_waits1 + nrecvs; 2547 r_waits3 = r_waits2 + nrecvs; 2548 } else { 2549 ierr = PetscMalloc2(nsends + nrecvs +1,MPI_Request,&s_waits3,nsends+1,MPI_Status,&send_status);CHKERRQ(ierr); 2550 r_waits3 = s_waits3 + nsends; 2551 } 2552 2553 ierr = PetscObjectGetNewTag((PetscObject)mat,&tag3);CHKERRQ(ierr); 2554 if (reuse == MAT_INITIAL_MATRIX){ 2555 /* get new tags to keep the communication clean */ 2556 ierr = PetscObjectGetNewTag((PetscObject)mat,&tag1);CHKERRQ(ierr); 2557 ierr = PetscObjectGetNewTag((PetscObject)mat,&tag2);CHKERRQ(ierr); 2558 ierr = PetscMalloc4(nsends,PetscInt,&sbuf_nz,nrecvs,PetscInt,&rbuf_nz,nrecvs,PetscInt*,&rbuf_j,nrecvs,PetscScalar*,&rbuf_a);CHKERRQ(ierr); 2559 2560 /* post receives of other's nzlocal */ 2561 for (i=0; i<nrecvs; i++){ 2562 ierr = MPI_Irecv(rbuf_nz+i,1,MPIU_INT,MPI_ANY_SOURCE,tag1,comm,r_waits1+i);CHKERRQ(ierr); 2563 } 2564 /* send nzlocal to others */ 2565 for (i=0; i<nsends; i++){ 2566 sbuf_nz[i] = nzlocal; 2567 ierr = MPI_Isend(sbuf_nz+i,1,MPIU_INT,send_rank[i],tag1,comm,s_waits1+i);CHKERRQ(ierr); 2568 } 2569 /* wait on receives of nzlocal; allocate space for rbuf_j, rbuf_a */ 2570 count = nrecvs; 2571 while (count) { 2572 ierr = MPI_Waitany(nrecvs,r_waits1,&imdex,&recv_status);CHKERRQ(ierr); 2573 recv_rank[imdex] = recv_status.MPI_SOURCE; 2574 /* allocate rbuf_a and rbuf_j; then post receives of rbuf_j */ 2575 ierr = PetscMalloc((rbuf_nz[imdex]+1)*sizeof(PetscScalar),&rbuf_a[imdex]);CHKERRQ(ierr); 2576 2577 i = rowrange[recv_status.MPI_SOURCE+1] - rowrange[recv_status.MPI_SOURCE]; /* number of expected mat->i */ 2578 rbuf_nz[imdex] += i + 2; 2579 ierr = PetscMalloc(rbuf_nz[imdex]*sizeof(PetscInt),&rbuf_j[imdex]);CHKERRQ(ierr); 2580 ierr = MPI_Irecv(rbuf_j[imdex],rbuf_nz[imdex],MPIU_INT,recv_status.MPI_SOURCE,tag2,comm,r_waits2+imdex);CHKERRQ(ierr); 2581 count--; 2582 } 2583 /* wait on sends of nzlocal */ 2584 if (nsends) {ierr = MPI_Waitall(nsends,s_waits1,send_status);CHKERRQ(ierr);} 2585 /* send mat->i,j to others, and recv from other's */ 2586 /*------------------------------------------------*/ 2587 for (i=0; i<nsends; i++){ 2588 j = nzlocal + rowrange[rank+1] - rowrange[rank] + 1; 2589 ierr = MPI_Isend(sbuf_j,j,MPIU_INT,send_rank[i],tag2,comm,s_waits2+i);CHKERRQ(ierr); 2590 } 2591 /* wait on receives of mat->i,j */ 2592 /*------------------------------*/ 2593 count = nrecvs; 2594 while (count) { 2595 ierr = MPI_Waitany(nrecvs,r_waits2,&imdex,&recv_status);CHKERRQ(ierr); 2596 if (recv_rank[imdex] != recv_status.MPI_SOURCE) SETERRQ2(PETSC_COMM_SELF,1, "recv_rank %d != MPI_SOURCE %d",recv_rank[imdex],recv_status.MPI_SOURCE); 2597 count--; 2598 } 2599 /* wait on sends of mat->i,j */ 2600 /*---------------------------*/ 2601 if (nsends) { 2602 ierr = MPI_Waitall(nsends,s_waits2,send_status);CHKERRQ(ierr); 2603 } 2604 } /* endof if (reuse == MAT_INITIAL_MATRIX) */ 2605 2606 /* post receives, send and receive mat->a */ 2607 /*----------------------------------------*/ 2608 for (imdex=0; imdex<nrecvs; imdex++) { 2609 ierr = MPI_Irecv(rbuf_a[imdex],rbuf_nz[imdex],MPIU_SCALAR,recv_rank[imdex],tag3,comm,r_waits3+imdex);CHKERRQ(ierr); 2610 } 2611 for (i=0; i<nsends; i++){ 2612 ierr = MPI_Isend(sbuf_a,nzlocal,MPIU_SCALAR,send_rank[i],tag3,comm,s_waits3+i);CHKERRQ(ierr); 2613 } 2614 count = nrecvs; 2615 while (count) { 2616 ierr = MPI_Waitany(nrecvs,r_waits3,&imdex,&recv_status);CHKERRQ(ierr); 2617 if (recv_rank[imdex] != recv_status.MPI_SOURCE) SETERRQ2(PETSC_COMM_SELF,1, "recv_rank %d != MPI_SOURCE %d",recv_rank[imdex],recv_status.MPI_SOURCE); 2618 count--; 2619 } 2620 if (nsends) { 2621 ierr = MPI_Waitall(nsends,s_waits3,send_status);CHKERRQ(ierr); 2622 } 2623 2624 ierr = PetscFree2(s_waits3,send_status);CHKERRQ(ierr); 2625 2626 /* create redundant matrix */ 2627 /*-------------------------*/ 2628 if (reuse == MAT_INITIAL_MATRIX){ 2629 /* compute rownz_max for preallocation */ 2630 for (imdex=0; imdex<nrecvs; imdex++){ 2631 j = rowrange[recv_rank[imdex]+1] - rowrange[recv_rank[imdex]]; 2632 rptr = rbuf_j[imdex]; 2633 for (i=0; i<j; i++){ 2634 ncols = rptr[i+1] - rptr[i]; 2635 if (rownz_max < ncols) rownz_max = ncols; 2636 } 2637 } 2638 2639 ierr = MatCreate(subcomm,&C);CHKERRQ(ierr); 2640 ierr = MatSetSizes(C,mlocal_sub,mlocal_sub,PETSC_DECIDE,PETSC_DECIDE);CHKERRQ(ierr); 2641 ierr = MatSetFromOptions(C);CHKERRQ(ierr); 2642 ierr = MatSeqAIJSetPreallocation(C,rownz_max,PETSC_NULL);CHKERRQ(ierr); 2643 ierr = MatMPIAIJSetPreallocation(C,rownz_max,PETSC_NULL,rownz_max,PETSC_NULL);CHKERRQ(ierr); 2644 } else { 2645 C = *matredundant; 2646 } 2647 2648 /* insert local matrix entries */ 2649 rptr = sbuf_j; 2650 cols = sbuf_j + rend-rstart + 1; 2651 vals = sbuf_a; 2652 for (i=0; i<rend-rstart; i++){ 2653 row = i + rstart; 2654 ncols = rptr[i+1] - rptr[i]; 2655 ierr = MatSetValues(C,1,&row,ncols,cols,vals,INSERT_VALUES);CHKERRQ(ierr); 2656 vals += ncols; 2657 cols += ncols; 2658 } 2659 /* insert received matrix entries */ 2660 for (imdex=0; imdex<nrecvs; imdex++){ 2661 rstart = rowrange[recv_rank[imdex]]; 2662 rend = rowrange[recv_rank[imdex]+1]; 2663 rptr = rbuf_j[imdex]; 2664 cols = rbuf_j[imdex] + rend-rstart + 1; 2665 vals = rbuf_a[imdex]; 2666 for (i=0; i<rend-rstart; i++){ 2667 row = i + rstart; 2668 ncols = rptr[i+1] - rptr[i]; 2669 ierr = MatSetValues(C,1,&row,ncols,cols,vals,INSERT_VALUES);CHKERRQ(ierr); 2670 vals += ncols; 2671 cols += ncols; 2672 } 2673 } 2674 ierr = MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2675 ierr = MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2676 ierr = MatGetSize(C,&M,&N);CHKERRQ(ierr); 2677 if (M != mat->rmap->N || N != mat->cmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"redundant mat size %d != input mat size %d",M,mat->rmap->N); 2678 if (reuse == MAT_INITIAL_MATRIX){ 2679 PetscContainer container; 2680 *matredundant = C; 2681 /* create a supporting struct and attach it to C for reuse */ 2682 ierr = PetscNewLog(C,Mat_Redundant,&redund);CHKERRQ(ierr); 2683 ierr = PetscContainerCreate(PETSC_COMM_SELF,&container);CHKERRQ(ierr); 2684 ierr = PetscContainerSetPointer(container,redund);CHKERRQ(ierr); 2685 ierr = PetscObjectCompose((PetscObject)C,"Mat_Redundant",(PetscObject)container);CHKERRQ(ierr); 2686 ierr = PetscContainerSetUserDestroy(container,PetscContainerDestroy_MatRedundant);CHKERRQ(ierr); 2687 2688 redund->nzlocal = nzlocal; 2689 redund->nsends = nsends; 2690 redund->nrecvs = nrecvs; 2691 redund->send_rank = send_rank; 2692 redund->recv_rank = recv_rank; 2693 redund->sbuf_nz = sbuf_nz; 2694 redund->rbuf_nz = rbuf_nz; 2695 redund->sbuf_j = sbuf_j; 2696 redund->sbuf_a = sbuf_a; 2697 redund->rbuf_j = rbuf_j; 2698 redund->rbuf_a = rbuf_a; 2699 2700 redund->MatDestroy = C->ops->destroy; 2701 C->ops->destroy = MatDestroy_MatRedundant; 2702 } 2703 PetscFunctionReturn(0); 2704 } 2705 2706 #undef __FUNCT__ 2707 #define __FUNCT__ "MatGetRowMaxAbs_MPIAIJ" 2708 PetscErrorCode MatGetRowMaxAbs_MPIAIJ(Mat A, Vec v, PetscInt idx[]) 2709 { 2710 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2711 PetscErrorCode ierr; 2712 PetscInt i,*idxb = 0; 2713 PetscScalar *va,*vb; 2714 Vec vtmp; 2715 2716 PetscFunctionBegin; 2717 ierr = MatGetRowMaxAbs(a->A,v,idx);CHKERRQ(ierr); 2718 ierr = VecGetArray(v,&va);CHKERRQ(ierr); 2719 if (idx) { 2720 for (i=0; i<A->rmap->n; i++) { 2721 if (PetscAbsScalar(va[i])) idx[i] += A->cmap->rstart; 2722 } 2723 } 2724 2725 ierr = VecCreateSeq(PETSC_COMM_SELF,A->rmap->n,&vtmp);CHKERRQ(ierr); 2726 if (idx) { 2727 ierr = PetscMalloc(A->rmap->n*sizeof(PetscInt),&idxb);CHKERRQ(ierr); 2728 } 2729 ierr = MatGetRowMaxAbs(a->B,vtmp,idxb);CHKERRQ(ierr); 2730 ierr = VecGetArray(vtmp,&vb);CHKERRQ(ierr); 2731 2732 for (i=0; i<A->rmap->n; i++){ 2733 if (PetscAbsScalar(va[i]) < PetscAbsScalar(vb[i])) { 2734 va[i] = vb[i]; 2735 if (idx) idx[i] = a->garray[idxb[i]]; 2736 } 2737 } 2738 2739 ierr = VecRestoreArray(v,&va);CHKERRQ(ierr); 2740 ierr = VecRestoreArray(vtmp,&vb);CHKERRQ(ierr); 2741 ierr = PetscFree(idxb);CHKERRQ(ierr); 2742 ierr = VecDestroy(&vtmp);CHKERRQ(ierr); 2743 PetscFunctionReturn(0); 2744 } 2745 2746 #undef __FUNCT__ 2747 #define __FUNCT__ "MatGetRowMinAbs_MPIAIJ" 2748 PetscErrorCode MatGetRowMinAbs_MPIAIJ(Mat A, Vec v, PetscInt idx[]) 2749 { 2750 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2751 PetscErrorCode ierr; 2752 PetscInt i,*idxb = 0; 2753 PetscScalar *va,*vb; 2754 Vec vtmp; 2755 2756 PetscFunctionBegin; 2757 ierr = MatGetRowMinAbs(a->A,v,idx);CHKERRQ(ierr); 2758 ierr = VecGetArray(v,&va);CHKERRQ(ierr); 2759 if (idx) { 2760 for (i=0; i<A->cmap->n; i++) { 2761 if (PetscAbsScalar(va[i])) idx[i] += A->cmap->rstart; 2762 } 2763 } 2764 2765 ierr = VecCreateSeq(PETSC_COMM_SELF,A->rmap->n,&vtmp);CHKERRQ(ierr); 2766 if (idx) { 2767 ierr = PetscMalloc(A->rmap->n*sizeof(PetscInt),&idxb);CHKERRQ(ierr); 2768 } 2769 ierr = MatGetRowMinAbs(a->B,vtmp,idxb);CHKERRQ(ierr); 2770 ierr = VecGetArray(vtmp,&vb);CHKERRQ(ierr); 2771 2772 for (i=0; i<A->rmap->n; i++){ 2773 if (PetscAbsScalar(va[i]) > PetscAbsScalar(vb[i])) { 2774 va[i] = vb[i]; 2775 if (idx) idx[i] = a->garray[idxb[i]]; 2776 } 2777 } 2778 2779 ierr = VecRestoreArray(v,&va);CHKERRQ(ierr); 2780 ierr = VecRestoreArray(vtmp,&vb);CHKERRQ(ierr); 2781 ierr = PetscFree(idxb);CHKERRQ(ierr); 2782 ierr = VecDestroy(&vtmp);CHKERRQ(ierr); 2783 PetscFunctionReturn(0); 2784 } 2785 2786 #undef __FUNCT__ 2787 #define __FUNCT__ "MatGetRowMin_MPIAIJ" 2788 PetscErrorCode MatGetRowMin_MPIAIJ(Mat A, Vec v, PetscInt idx[]) 2789 { 2790 Mat_MPIAIJ *mat = (Mat_MPIAIJ *) A->data; 2791 PetscInt n = A->rmap->n; 2792 PetscInt cstart = A->cmap->rstart; 2793 PetscInt *cmap = mat->garray; 2794 PetscInt *diagIdx, *offdiagIdx; 2795 Vec diagV, offdiagV; 2796 PetscScalar *a, *diagA, *offdiagA; 2797 PetscInt r; 2798 PetscErrorCode ierr; 2799 2800 PetscFunctionBegin; 2801 ierr = PetscMalloc2(n,PetscInt,&diagIdx,n,PetscInt,&offdiagIdx);CHKERRQ(ierr); 2802 ierr = VecCreateSeq(((PetscObject)A)->comm, n, &diagV);CHKERRQ(ierr); 2803 ierr = VecCreateSeq(((PetscObject)A)->comm, n, &offdiagV);CHKERRQ(ierr); 2804 ierr = MatGetRowMin(mat->A, diagV, diagIdx);CHKERRQ(ierr); 2805 ierr = MatGetRowMin(mat->B, offdiagV, offdiagIdx);CHKERRQ(ierr); 2806 ierr = VecGetArray(v, &a);CHKERRQ(ierr); 2807 ierr = VecGetArray(diagV, &diagA);CHKERRQ(ierr); 2808 ierr = VecGetArray(offdiagV, &offdiagA);CHKERRQ(ierr); 2809 for(r = 0; r < n; ++r) { 2810 if (PetscAbsScalar(diagA[r]) <= PetscAbsScalar(offdiagA[r])) { 2811 a[r] = diagA[r]; 2812 idx[r] = cstart + diagIdx[r]; 2813 } else { 2814 a[r] = offdiagA[r]; 2815 idx[r] = cmap[offdiagIdx[r]]; 2816 } 2817 } 2818 ierr = VecRestoreArray(v, &a);CHKERRQ(ierr); 2819 ierr = VecRestoreArray(diagV, &diagA);CHKERRQ(ierr); 2820 ierr = VecRestoreArray(offdiagV, &offdiagA);CHKERRQ(ierr); 2821 ierr = VecDestroy(&diagV);CHKERRQ(ierr); 2822 ierr = VecDestroy(&offdiagV);CHKERRQ(ierr); 2823 ierr = PetscFree2(diagIdx, offdiagIdx);CHKERRQ(ierr); 2824 PetscFunctionReturn(0); 2825 } 2826 2827 #undef __FUNCT__ 2828 #define __FUNCT__ "MatGetRowMax_MPIAIJ" 2829 PetscErrorCode MatGetRowMax_MPIAIJ(Mat A, Vec v, PetscInt idx[]) 2830 { 2831 Mat_MPIAIJ *mat = (Mat_MPIAIJ *) A->data; 2832 PetscInt n = A->rmap->n; 2833 PetscInt cstart = A->cmap->rstart; 2834 PetscInt *cmap = mat->garray; 2835 PetscInt *diagIdx, *offdiagIdx; 2836 Vec diagV, offdiagV; 2837 PetscScalar *a, *diagA, *offdiagA; 2838 PetscInt r; 2839 PetscErrorCode ierr; 2840 2841 PetscFunctionBegin; 2842 ierr = PetscMalloc2(n,PetscInt,&diagIdx,n,PetscInt,&offdiagIdx);CHKERRQ(ierr); 2843 ierr = VecCreateSeq(((PetscObject)A)->comm, n, &diagV);CHKERRQ(ierr); 2844 ierr = VecCreateSeq(((PetscObject)A)->comm, n, &offdiagV);CHKERRQ(ierr); 2845 ierr = MatGetRowMax(mat->A, diagV, diagIdx);CHKERRQ(ierr); 2846 ierr = MatGetRowMax(mat->B, offdiagV, offdiagIdx);CHKERRQ(ierr); 2847 ierr = VecGetArray(v, &a);CHKERRQ(ierr); 2848 ierr = VecGetArray(diagV, &diagA);CHKERRQ(ierr); 2849 ierr = VecGetArray(offdiagV, &offdiagA);CHKERRQ(ierr); 2850 for(r = 0; r < n; ++r) { 2851 if (PetscAbsScalar(diagA[r]) >= PetscAbsScalar(offdiagA[r])) { 2852 a[r] = diagA[r]; 2853 idx[r] = cstart + diagIdx[r]; 2854 } else { 2855 a[r] = offdiagA[r]; 2856 idx[r] = cmap[offdiagIdx[r]]; 2857 } 2858 } 2859 ierr = VecRestoreArray(v, &a);CHKERRQ(ierr); 2860 ierr = VecRestoreArray(diagV, &diagA);CHKERRQ(ierr); 2861 ierr = VecRestoreArray(offdiagV, &offdiagA);CHKERRQ(ierr); 2862 ierr = VecDestroy(&diagV);CHKERRQ(ierr); 2863 ierr = VecDestroy(&offdiagV);CHKERRQ(ierr); 2864 ierr = PetscFree2(diagIdx, offdiagIdx);CHKERRQ(ierr); 2865 PetscFunctionReturn(0); 2866 } 2867 2868 #undef __FUNCT__ 2869 #define __FUNCT__ "MatGetSeqNonzerostructure_MPIAIJ" 2870 PetscErrorCode MatGetSeqNonzerostructure_MPIAIJ(Mat mat,Mat *newmat) 2871 { 2872 PetscErrorCode ierr; 2873 Mat *dummy; 2874 2875 PetscFunctionBegin; 2876 ierr = MatGetSubMatrix_MPIAIJ_All(mat,MAT_DO_NOT_GET_VALUES,MAT_INITIAL_MATRIX,&dummy);CHKERRQ(ierr); 2877 *newmat = *dummy; 2878 ierr = PetscFree(dummy);CHKERRQ(ierr); 2879 PetscFunctionReturn(0); 2880 } 2881 2882 extern PetscErrorCode MatFDColoringApply_AIJ(Mat,MatFDColoring,Vec,MatStructure*,void*); 2883 /* -------------------------------------------------------------------*/ 2884 static struct _MatOps MatOps_Values = {MatSetValues_MPIAIJ, 2885 MatGetRow_MPIAIJ, 2886 MatRestoreRow_MPIAIJ, 2887 MatMult_MPIAIJ, 2888 /* 4*/ MatMultAdd_MPIAIJ, 2889 MatMultTranspose_MPIAIJ, 2890 MatMultTransposeAdd_MPIAIJ, 2891 #ifdef PETSC_HAVE_PBGL 2892 MatSolve_MPIAIJ, 2893 #else 2894 0, 2895 #endif 2896 0, 2897 0, 2898 /*10*/ 0, 2899 0, 2900 0, 2901 MatSOR_MPIAIJ, 2902 MatTranspose_MPIAIJ, 2903 /*15*/ MatGetInfo_MPIAIJ, 2904 MatEqual_MPIAIJ, 2905 MatGetDiagonal_MPIAIJ, 2906 MatDiagonalScale_MPIAIJ, 2907 MatNorm_MPIAIJ, 2908 /*20*/ MatAssemblyBegin_MPIAIJ, 2909 MatAssemblyEnd_MPIAIJ, 2910 MatSetOption_MPIAIJ, 2911 MatZeroEntries_MPIAIJ, 2912 /*24*/ MatZeroRows_MPIAIJ, 2913 0, 2914 #ifdef PETSC_HAVE_PBGL 2915 0, 2916 #else 2917 0, 2918 #endif 2919 0, 2920 0, 2921 /*29*/ MatSetUpPreallocation_MPIAIJ, 2922 #ifdef PETSC_HAVE_PBGL 2923 0, 2924 #else 2925 0, 2926 #endif 2927 0, 2928 0, 2929 0, 2930 /*34*/ MatDuplicate_MPIAIJ, 2931 0, 2932 0, 2933 0, 2934 0, 2935 /*39*/ MatAXPY_MPIAIJ, 2936 MatGetSubMatrices_MPIAIJ, 2937 MatIncreaseOverlap_MPIAIJ, 2938 MatGetValues_MPIAIJ, 2939 MatCopy_MPIAIJ, 2940 /*44*/ MatGetRowMax_MPIAIJ, 2941 MatScale_MPIAIJ, 2942 0, 2943 0, 2944 MatZeroRowsColumns_MPIAIJ, 2945 /*49*/ MatSetBlockSize_MPIAIJ, 2946 0, 2947 0, 2948 0, 2949 0, 2950 /*54*/ MatFDColoringCreate_MPIAIJ, 2951 0, 2952 MatSetUnfactored_MPIAIJ, 2953 MatPermute_MPIAIJ, 2954 0, 2955 /*59*/ MatGetSubMatrix_MPIAIJ, 2956 MatDestroy_MPIAIJ, 2957 MatView_MPIAIJ, 2958 0, 2959 0, 2960 /*64*/ 0, 2961 0, 2962 0, 2963 0, 2964 0, 2965 /*69*/ MatGetRowMaxAbs_MPIAIJ, 2966 MatGetRowMinAbs_MPIAIJ, 2967 0, 2968 MatSetColoring_MPIAIJ, 2969 #if defined(PETSC_HAVE_ADIC) 2970 MatSetValuesAdic_MPIAIJ, 2971 #else 2972 0, 2973 #endif 2974 MatSetValuesAdifor_MPIAIJ, 2975 /*75*/ MatFDColoringApply_AIJ, 2976 0, 2977 0, 2978 0, 2979 0, 2980 /*80*/ 0, 2981 0, 2982 0, 2983 /*83*/ MatLoad_MPIAIJ, 2984 0, 2985 0, 2986 0, 2987 0, 2988 0, 2989 /*89*/ MatMatMult_MPIAIJ_MPIAIJ, 2990 MatMatMultSymbolic_MPIAIJ_MPIAIJ, 2991 MatMatMultNumeric_MPIAIJ_MPIAIJ, 2992 MatPtAP_Basic, 2993 MatPtAPSymbolic_MPIAIJ, 2994 /*94*/ MatPtAPNumeric_MPIAIJ, 2995 0, 2996 0, 2997 0, 2998 0, 2999 /*99*/ 0, 3000 MatPtAPSymbolic_MPIAIJ_MPIAIJ, 3001 MatPtAPNumeric_MPIAIJ_MPIAIJ, 3002 MatConjugate_MPIAIJ, 3003 0, 3004 /*104*/MatSetValuesRow_MPIAIJ, 3005 MatRealPart_MPIAIJ, 3006 MatImaginaryPart_MPIAIJ, 3007 0, 3008 0, 3009 /*109*/0, 3010 MatGetRedundantMatrix_MPIAIJ, 3011 MatGetRowMin_MPIAIJ, 3012 0, 3013 0, 3014 /*114*/MatGetSeqNonzerostructure_MPIAIJ, 3015 0, 3016 0, 3017 0, 3018 0, 3019 /*119*/0, 3020 0, 3021 0, 3022 0, 3023 MatGetMultiProcBlock_MPIAIJ, 3024 /*124*/MatFindNonZeroRows_MPIAIJ, 3025 0, 3026 0, 3027 0, 3028 MatGetSubMatricesParallel_MPIAIJ 3029 }; 3030 3031 /* ----------------------------------------------------------------------------------------*/ 3032 3033 EXTERN_C_BEGIN 3034 #undef __FUNCT__ 3035 #define __FUNCT__ "MatStoreValues_MPIAIJ" 3036 PetscErrorCode MatStoreValues_MPIAIJ(Mat mat) 3037 { 3038 Mat_MPIAIJ *aij = (Mat_MPIAIJ *)mat->data; 3039 PetscErrorCode ierr; 3040 3041 PetscFunctionBegin; 3042 ierr = MatStoreValues(aij->A);CHKERRQ(ierr); 3043 ierr = MatStoreValues(aij->B);CHKERRQ(ierr); 3044 PetscFunctionReturn(0); 3045 } 3046 EXTERN_C_END 3047 3048 EXTERN_C_BEGIN 3049 #undef __FUNCT__ 3050 #define __FUNCT__ "MatRetrieveValues_MPIAIJ" 3051 PetscErrorCode MatRetrieveValues_MPIAIJ(Mat mat) 3052 { 3053 Mat_MPIAIJ *aij = (Mat_MPIAIJ *)mat->data; 3054 PetscErrorCode ierr; 3055 3056 PetscFunctionBegin; 3057 ierr = MatRetrieveValues(aij->A);CHKERRQ(ierr); 3058 ierr = MatRetrieveValues(aij->B);CHKERRQ(ierr); 3059 PetscFunctionReturn(0); 3060 } 3061 EXTERN_C_END 3062 3063 EXTERN_C_BEGIN 3064 #undef __FUNCT__ 3065 #define __FUNCT__ "MatMPIAIJSetPreallocation_MPIAIJ" 3066 PetscErrorCode MatMPIAIJSetPreallocation_MPIAIJ(Mat B,PetscInt d_nz,const PetscInt d_nnz[],PetscInt o_nz,const PetscInt o_nnz[]) 3067 { 3068 Mat_MPIAIJ *b; 3069 PetscErrorCode ierr; 3070 PetscInt i; 3071 3072 PetscFunctionBegin; 3073 if (d_nz == PETSC_DEFAULT || d_nz == PETSC_DECIDE) d_nz = 5; 3074 if (o_nz == PETSC_DEFAULT || o_nz == PETSC_DECIDE) o_nz = 2; 3075 if (d_nz < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"d_nz cannot be less than 0: value %D",d_nz); 3076 if (o_nz < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"o_nz cannot be less than 0: value %D",o_nz); 3077 3078 ierr = PetscLayoutSetBlockSize(B->rmap,1);CHKERRQ(ierr); 3079 ierr = PetscLayoutSetBlockSize(B->cmap,1);CHKERRQ(ierr); 3080 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 3081 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 3082 if (d_nnz) { 3083 for (i=0; i<B->rmap->n; i++) { 3084 if (d_nnz[i] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"d_nnz cannot be less than 0: local row %D value %D",i,d_nnz[i]); 3085 } 3086 } 3087 if (o_nnz) { 3088 for (i=0; i<B->rmap->n; i++) { 3089 if (o_nnz[i] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"o_nnz cannot be less than 0: local row %D value %D",i,o_nnz[i]); 3090 } 3091 } 3092 b = (Mat_MPIAIJ*)B->data; 3093 3094 if (!B->preallocated) { 3095 /* Explicitly create 2 MATSEQAIJ matrices. */ 3096 ierr = MatCreate(PETSC_COMM_SELF,&b->A);CHKERRQ(ierr); 3097 ierr = MatSetSizes(b->A,B->rmap->n,B->cmap->n,B->rmap->n,B->cmap->n);CHKERRQ(ierr); 3098 ierr = MatSetType(b->A,MATSEQAIJ);CHKERRQ(ierr); 3099 ierr = PetscLogObjectParent(B,b->A);CHKERRQ(ierr); 3100 ierr = MatCreate(PETSC_COMM_SELF,&b->B);CHKERRQ(ierr); 3101 ierr = MatSetSizes(b->B,B->rmap->n,B->cmap->N,B->rmap->n,B->cmap->N);CHKERRQ(ierr); 3102 ierr = MatSetType(b->B,MATSEQAIJ);CHKERRQ(ierr); 3103 ierr = PetscLogObjectParent(B,b->B);CHKERRQ(ierr); 3104 } 3105 3106 ierr = MatSeqAIJSetPreallocation(b->A,d_nz,d_nnz);CHKERRQ(ierr); 3107 ierr = MatSeqAIJSetPreallocation(b->B,o_nz,o_nnz);CHKERRQ(ierr); 3108 B->preallocated = PETSC_TRUE; 3109 PetscFunctionReturn(0); 3110 } 3111 EXTERN_C_END 3112 3113 #undef __FUNCT__ 3114 #define __FUNCT__ "MatDuplicate_MPIAIJ" 3115 PetscErrorCode MatDuplicate_MPIAIJ(Mat matin,MatDuplicateOption cpvalues,Mat *newmat) 3116 { 3117 Mat mat; 3118 Mat_MPIAIJ *a,*oldmat = (Mat_MPIAIJ*)matin->data; 3119 PetscErrorCode ierr; 3120 3121 PetscFunctionBegin; 3122 *newmat = 0; 3123 ierr = MatCreate(((PetscObject)matin)->comm,&mat);CHKERRQ(ierr); 3124 ierr = MatSetSizes(mat,matin->rmap->n,matin->cmap->n,matin->rmap->N,matin->cmap->N);CHKERRQ(ierr); 3125 ierr = MatSetType(mat,((PetscObject)matin)->type_name);CHKERRQ(ierr); 3126 ierr = PetscMemcpy(mat->ops,matin->ops,sizeof(struct _MatOps));CHKERRQ(ierr); 3127 a = (Mat_MPIAIJ*)mat->data; 3128 3129 mat->factortype = matin->factortype; 3130 mat->rmap->bs = matin->rmap->bs; 3131 mat->assembled = PETSC_TRUE; 3132 mat->insertmode = NOT_SET_VALUES; 3133 mat->preallocated = PETSC_TRUE; 3134 3135 a->size = oldmat->size; 3136 a->rank = oldmat->rank; 3137 a->donotstash = oldmat->donotstash; 3138 a->roworiented = oldmat->roworiented; 3139 a->rowindices = 0; 3140 a->rowvalues = 0; 3141 a->getrowactive = PETSC_FALSE; 3142 3143 ierr = PetscLayoutCopy(matin->rmap,&mat->rmap);CHKERRQ(ierr); 3144 ierr = PetscLayoutCopy(matin->cmap,&mat->cmap);CHKERRQ(ierr); 3145 3146 if (oldmat->colmap) { 3147 #if defined (PETSC_USE_CTABLE) 3148 ierr = PetscTableCreateCopy(oldmat->colmap,&a->colmap);CHKERRQ(ierr); 3149 #else 3150 ierr = PetscMalloc((mat->cmap->N)*sizeof(PetscInt),&a->colmap);CHKERRQ(ierr); 3151 ierr = PetscLogObjectMemory(mat,(mat->cmap->N)*sizeof(PetscInt));CHKERRQ(ierr); 3152 ierr = PetscMemcpy(a->colmap,oldmat->colmap,(mat->cmap->N)*sizeof(PetscInt));CHKERRQ(ierr); 3153 #endif 3154 } else a->colmap = 0; 3155 if (oldmat->garray) { 3156 PetscInt len; 3157 len = oldmat->B->cmap->n; 3158 ierr = PetscMalloc((len+1)*sizeof(PetscInt),&a->garray);CHKERRQ(ierr); 3159 ierr = PetscLogObjectMemory(mat,len*sizeof(PetscInt));CHKERRQ(ierr); 3160 if (len) { ierr = PetscMemcpy(a->garray,oldmat->garray,len*sizeof(PetscInt));CHKERRQ(ierr); } 3161 } else a->garray = 0; 3162 3163 ierr = VecDuplicate(oldmat->lvec,&a->lvec);CHKERRQ(ierr); 3164 ierr = PetscLogObjectParent(mat,a->lvec);CHKERRQ(ierr); 3165 ierr = VecScatterCopy(oldmat->Mvctx,&a->Mvctx);CHKERRQ(ierr); 3166 ierr = PetscLogObjectParent(mat,a->Mvctx);CHKERRQ(ierr); 3167 ierr = MatDuplicate(oldmat->A,cpvalues,&a->A);CHKERRQ(ierr); 3168 ierr = PetscLogObjectParent(mat,a->A);CHKERRQ(ierr); 3169 ierr = MatDuplicate(oldmat->B,cpvalues,&a->B);CHKERRQ(ierr); 3170 ierr = PetscLogObjectParent(mat,a->B);CHKERRQ(ierr); 3171 ierr = PetscFListDuplicate(((PetscObject)matin)->qlist,&((PetscObject)mat)->qlist);CHKERRQ(ierr); 3172 *newmat = mat; 3173 PetscFunctionReturn(0); 3174 } 3175 3176 /* 3177 Allows sending/receiving larger messages then 2 gigabytes in a single call 3178 */ 3179 static int MPILong_Send(void *mess,PetscInt cnt, MPI_Datatype type,int to, int tag, MPI_Comm comm) 3180 { 3181 int ierr; 3182 static PetscInt CHUNKSIZE = 250000000; /* 250,000,000 */ 3183 PetscInt i,numchunks; 3184 PetscMPIInt icnt; 3185 3186 numchunks = cnt/CHUNKSIZE + 1; 3187 for (i=0; i<numchunks; i++) { 3188 icnt = PetscMPIIntCast((i < numchunks-1) ? CHUNKSIZE : cnt - (numchunks-1)*CHUNKSIZE); 3189 ierr = MPI_Send(mess,icnt,type,to,tag,comm);if (ierr) return ierr; 3190 if (type == MPIU_INT) { 3191 mess = (void*) (((PetscInt*)mess) + CHUNKSIZE); 3192 } else if (type == MPIU_SCALAR) { 3193 mess = (void*) (((PetscScalar*)mess) + CHUNKSIZE); 3194 } else SETERRQ(comm,PETSC_ERR_SUP,"No support for this datatype"); 3195 } 3196 return 0; 3197 } 3198 static int MPILong_Recv(void *mess,PetscInt cnt, MPI_Datatype type,int from, int tag, MPI_Comm comm) 3199 { 3200 int ierr; 3201 static PetscInt CHUNKSIZE = 250000000; /* 250,000,000 */ 3202 MPI_Status status; 3203 PetscInt i,numchunks; 3204 PetscMPIInt icnt; 3205 3206 numchunks = cnt/CHUNKSIZE + 1; 3207 for (i=0; i<numchunks; i++) { 3208 icnt = PetscMPIIntCast((i < numchunks-1) ? CHUNKSIZE : cnt - (numchunks-1)*CHUNKSIZE); 3209 ierr = MPI_Recv(mess,icnt,type,from,tag,comm,&status);if (ierr) return ierr; 3210 if (type == MPIU_INT) { 3211 mess = (void*) (((PetscInt*)mess) + CHUNKSIZE); 3212 } else if (type == MPIU_SCALAR) { 3213 mess = (void*) (((PetscScalar*)mess) + CHUNKSIZE); 3214 } else SETERRQ(comm,PETSC_ERR_SUP,"No support for this datatype"); 3215 } 3216 return 0; 3217 } 3218 3219 #undef __FUNCT__ 3220 #define __FUNCT__ "MatLoad_MPIAIJ" 3221 PetscErrorCode MatLoad_MPIAIJ(Mat newMat, PetscViewer viewer) 3222 { 3223 PetscScalar *vals,*svals; 3224 MPI_Comm comm = ((PetscObject)viewer)->comm; 3225 PetscErrorCode ierr; 3226 PetscMPIInt rank,size,tag = ((PetscObject)viewer)->tag; 3227 PetscInt i,nz,j,rstart,rend,mmax,maxnz = 0,grows,gcols; 3228 PetscInt header[4],*rowlengths = 0,M,N,m,*cols; 3229 PetscInt *ourlens = PETSC_NULL,*procsnz = PETSC_NULL,*offlens = PETSC_NULL,jj,*mycols,*smycols; 3230 PetscInt cend,cstart,n,*rowners,sizesset=1; 3231 int fd; 3232 3233 PetscFunctionBegin; 3234 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 3235 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 3236 if (!rank) { 3237 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 3238 ierr = PetscBinaryRead(fd,(char *)header,4,PETSC_INT);CHKERRQ(ierr); 3239 if (header[0] != MAT_FILE_CLASSID) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"not matrix object"); 3240 } 3241 3242 if (newMat->rmap->n < 0 && newMat->rmap->N < 0 && newMat->cmap->n < 0 && newMat->cmap->N < 0) sizesset = 0; 3243 3244 ierr = MPI_Bcast(header+1,3,MPIU_INT,0,comm);CHKERRQ(ierr); 3245 M = header[1]; N = header[2]; 3246 /* If global rows/cols are set to PETSC_DECIDE, set it to the sizes given in the file */ 3247 if (sizesset && newMat->rmap->N < 0) newMat->rmap->N = M; 3248 if (sizesset && newMat->cmap->N < 0) newMat->cmap->N = N; 3249 3250 /* If global sizes are set, check if they are consistent with that given in the file */ 3251 if (sizesset) { 3252 ierr = MatGetSize(newMat,&grows,&gcols);CHKERRQ(ierr); 3253 } 3254 if (sizesset && newMat->rmap->N != grows) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED, "Inconsistent # of rows:Matrix in file has (%d) and input matrix has (%d)",M,grows); 3255 if (sizesset && newMat->cmap->N != gcols) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED, "Inconsistent # of cols:Matrix in file has (%d) and input matrix has (%d)",N,gcols); 3256 3257 /* determine ownership of all rows */ 3258 if (newMat->rmap->n < 0 ) m = M/size + ((M % size) > rank); /* PETSC_DECIDE */ 3259 else m = newMat->rmap->n; /* Set by user */ 3260 3261 ierr = PetscMalloc((size+1)*sizeof(PetscInt),&rowners);CHKERRQ(ierr); 3262 ierr = MPI_Allgather(&m,1,MPIU_INT,rowners+1,1,MPIU_INT,comm);CHKERRQ(ierr); 3263 3264 /* First process needs enough room for process with most rows */ 3265 if (!rank) { 3266 mmax = rowners[1]; 3267 for (i=2; i<size; i++) { 3268 mmax = PetscMax(mmax,rowners[i]); 3269 } 3270 } else mmax = m; 3271 3272 rowners[0] = 0; 3273 for (i=2; i<=size; i++) { 3274 rowners[i] += rowners[i-1]; 3275 } 3276 rstart = rowners[rank]; 3277 rend = rowners[rank+1]; 3278 3279 /* distribute row lengths to all processors */ 3280 ierr = PetscMalloc2(mmax,PetscInt,&ourlens,mmax,PetscInt,&offlens);CHKERRQ(ierr); 3281 if (!rank) { 3282 ierr = PetscBinaryRead(fd,ourlens,m,PETSC_INT);CHKERRQ(ierr); 3283 ierr = PetscMalloc(m*sizeof(PetscInt),&rowlengths);CHKERRQ(ierr); 3284 ierr = PetscMalloc(size*sizeof(PetscInt),&procsnz);CHKERRQ(ierr); 3285 ierr = PetscMemzero(procsnz,size*sizeof(PetscInt));CHKERRQ(ierr); 3286 for (j=0; j<m; j++) { 3287 procsnz[0] += ourlens[j]; 3288 } 3289 for (i=1; i<size; i++) { 3290 ierr = PetscBinaryRead(fd,rowlengths,rowners[i+1]-rowners[i],PETSC_INT);CHKERRQ(ierr); 3291 /* calculate the number of nonzeros on each processor */ 3292 for (j=0; j<rowners[i+1]-rowners[i]; j++) { 3293 procsnz[i] += rowlengths[j]; 3294 } 3295 ierr = MPILong_Send(rowlengths,rowners[i+1]-rowners[i],MPIU_INT,i,tag,comm);CHKERRQ(ierr); 3296 } 3297 ierr = PetscFree(rowlengths);CHKERRQ(ierr); 3298 } else { 3299 ierr = MPILong_Recv(ourlens,m,MPIU_INT,0,tag,comm);CHKERRQ(ierr); 3300 } 3301 3302 if (!rank) { 3303 /* determine max buffer needed and allocate it */ 3304 maxnz = 0; 3305 for (i=0; i<size; i++) { 3306 maxnz = PetscMax(maxnz,procsnz[i]); 3307 } 3308 ierr = PetscMalloc(maxnz*sizeof(PetscInt),&cols);CHKERRQ(ierr); 3309 3310 /* read in my part of the matrix column indices */ 3311 nz = procsnz[0]; 3312 ierr = PetscMalloc(nz*sizeof(PetscInt),&mycols);CHKERRQ(ierr); 3313 ierr = PetscBinaryRead(fd,mycols,nz,PETSC_INT);CHKERRQ(ierr); 3314 3315 /* read in every one elses and ship off */ 3316 for (i=1; i<size; i++) { 3317 nz = procsnz[i]; 3318 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 3319 ierr = MPILong_Send(cols,nz,MPIU_INT,i,tag,comm);CHKERRQ(ierr); 3320 } 3321 ierr = PetscFree(cols);CHKERRQ(ierr); 3322 } else { 3323 /* determine buffer space needed for message */ 3324 nz = 0; 3325 for (i=0; i<m; i++) { 3326 nz += ourlens[i]; 3327 } 3328 ierr = PetscMalloc(nz*sizeof(PetscInt),&mycols);CHKERRQ(ierr); 3329 3330 /* receive message of column indices*/ 3331 ierr = MPILong_Recv(mycols,nz,MPIU_INT,0,tag,comm);CHKERRQ(ierr); 3332 } 3333 3334 /* determine column ownership if matrix is not square */ 3335 if (N != M) { 3336 if (newMat->cmap->n < 0) n = N/size + ((N % size) > rank); 3337 else n = newMat->cmap->n; 3338 ierr = MPI_Scan(&n,&cend,1,MPIU_INT,MPI_SUM,comm);CHKERRQ(ierr); 3339 cstart = cend - n; 3340 } else { 3341 cstart = rstart; 3342 cend = rend; 3343 n = cend - cstart; 3344 } 3345 3346 /* loop over local rows, determining number of off diagonal entries */ 3347 ierr = PetscMemzero(offlens,m*sizeof(PetscInt));CHKERRQ(ierr); 3348 jj = 0; 3349 for (i=0; i<m; i++) { 3350 for (j=0; j<ourlens[i]; j++) { 3351 if (mycols[jj] < cstart || mycols[jj] >= cend) offlens[i]++; 3352 jj++; 3353 } 3354 } 3355 3356 for (i=0; i<m; i++) { 3357 ourlens[i] -= offlens[i]; 3358 } 3359 if (!sizesset) { 3360 ierr = MatSetSizes(newMat,m,n,M,N);CHKERRQ(ierr); 3361 } 3362 ierr = MatMPIAIJSetPreallocation(newMat,0,ourlens,0,offlens);CHKERRQ(ierr); 3363 3364 for (i=0; i<m; i++) { 3365 ourlens[i] += offlens[i]; 3366 } 3367 3368 if (!rank) { 3369 ierr = PetscMalloc((maxnz+1)*sizeof(PetscScalar),&vals);CHKERRQ(ierr); 3370 3371 /* read in my part of the matrix numerical values */ 3372 nz = procsnz[0]; 3373 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 3374 3375 /* insert into matrix */ 3376 jj = rstart; 3377 smycols = mycols; 3378 svals = vals; 3379 for (i=0; i<m; i++) { 3380 ierr = MatSetValues_MPIAIJ(newMat,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);CHKERRQ(ierr); 3381 smycols += ourlens[i]; 3382 svals += ourlens[i]; 3383 jj++; 3384 } 3385 3386 /* read in other processors and ship out */ 3387 for (i=1; i<size; i++) { 3388 nz = procsnz[i]; 3389 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 3390 ierr = MPILong_Send(vals,nz,MPIU_SCALAR,i,((PetscObject)newMat)->tag,comm);CHKERRQ(ierr); 3391 } 3392 ierr = PetscFree(procsnz);CHKERRQ(ierr); 3393 } else { 3394 /* receive numeric values */ 3395 ierr = PetscMalloc((nz+1)*sizeof(PetscScalar),&vals);CHKERRQ(ierr); 3396 3397 /* receive message of values*/ 3398 ierr = MPILong_Recv(vals,nz,MPIU_SCALAR,0,((PetscObject)newMat)->tag,comm);CHKERRQ(ierr); 3399 3400 /* insert into matrix */ 3401 jj = rstart; 3402 smycols = mycols; 3403 svals = vals; 3404 for (i=0; i<m; i++) { 3405 ierr = MatSetValues_MPIAIJ(newMat,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);CHKERRQ(ierr); 3406 smycols += ourlens[i]; 3407 svals += ourlens[i]; 3408 jj++; 3409 } 3410 } 3411 ierr = PetscFree2(ourlens,offlens);CHKERRQ(ierr); 3412 ierr = PetscFree(vals);CHKERRQ(ierr); 3413 ierr = PetscFree(mycols);CHKERRQ(ierr); 3414 ierr = PetscFree(rowners);CHKERRQ(ierr); 3415 3416 ierr = MatAssemblyBegin(newMat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3417 ierr = MatAssemblyEnd(newMat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3418 PetscFunctionReturn(0); 3419 } 3420 3421 #undef __FUNCT__ 3422 #define __FUNCT__ "MatGetSubMatrix_MPIAIJ" 3423 PetscErrorCode MatGetSubMatrix_MPIAIJ(Mat mat,IS isrow,IS iscol,MatReuse call,Mat *newmat) 3424 { 3425 PetscErrorCode ierr; 3426 IS iscol_local; 3427 PetscInt csize; 3428 3429 PetscFunctionBegin; 3430 ierr = ISGetLocalSize(iscol,&csize);CHKERRQ(ierr); 3431 if (call == MAT_REUSE_MATRIX) { 3432 ierr = PetscObjectQuery((PetscObject)*newmat,"ISAllGather",(PetscObject*)&iscol_local);CHKERRQ(ierr); 3433 if (!iscol_local) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Submatrix passed in was not used before, cannot reuse"); 3434 } else { 3435 ierr = ISAllGather(iscol,&iscol_local);CHKERRQ(ierr); 3436 } 3437 ierr = MatGetSubMatrix_MPIAIJ_Private(mat,isrow,iscol_local,csize,call,newmat);CHKERRQ(ierr); 3438 if (call == MAT_INITIAL_MATRIX) { 3439 ierr = PetscObjectCompose((PetscObject)*newmat,"ISAllGather",(PetscObject)iscol_local);CHKERRQ(ierr); 3440 ierr = ISDestroy(&iscol_local);CHKERRQ(ierr); 3441 } 3442 PetscFunctionReturn(0); 3443 } 3444 3445 #undef __FUNCT__ 3446 #define __FUNCT__ "MatGetSubMatrix_MPIAIJ_Private" 3447 /* 3448 Not great since it makes two copies of the submatrix, first an SeqAIJ 3449 in local and then by concatenating the local matrices the end result. 3450 Writing it directly would be much like MatGetSubMatrices_MPIAIJ() 3451 3452 Note: This requires a sequential iscol with all indices. 3453 */ 3454 PetscErrorCode MatGetSubMatrix_MPIAIJ_Private(Mat mat,IS isrow,IS iscol,PetscInt csize,MatReuse call,Mat *newmat) 3455 { 3456 PetscErrorCode ierr; 3457 PetscMPIInt rank,size; 3458 PetscInt i,m,n,rstart,row,rend,nz,*cwork,j; 3459 PetscInt *ii,*jj,nlocal,*dlens,*olens,dlen,olen,jend,mglobal; 3460 Mat *local,M,Mreuse; 3461 MatScalar *vwork,*aa; 3462 MPI_Comm comm = ((PetscObject)mat)->comm; 3463 Mat_SeqAIJ *aij; 3464 3465 3466 PetscFunctionBegin; 3467 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 3468 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 3469 3470 if (call == MAT_REUSE_MATRIX) { 3471 ierr = PetscObjectQuery((PetscObject)*newmat,"SubMatrix",(PetscObject *)&Mreuse);CHKERRQ(ierr); 3472 if (!Mreuse) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Submatrix passed in was not used before, cannot reuse"); 3473 local = &Mreuse; 3474 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_REUSE_MATRIX,&local);CHKERRQ(ierr); 3475 } else { 3476 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_INITIAL_MATRIX,&local);CHKERRQ(ierr); 3477 Mreuse = *local; 3478 ierr = PetscFree(local);CHKERRQ(ierr); 3479 } 3480 3481 /* 3482 m - number of local rows 3483 n - number of columns (same on all processors) 3484 rstart - first row in new global matrix generated 3485 */ 3486 ierr = MatGetSize(Mreuse,&m,&n);CHKERRQ(ierr); 3487 if (call == MAT_INITIAL_MATRIX) { 3488 aij = (Mat_SeqAIJ*)(Mreuse)->data; 3489 ii = aij->i; 3490 jj = aij->j; 3491 3492 /* 3493 Determine the number of non-zeros in the diagonal and off-diagonal 3494 portions of the matrix in order to do correct preallocation 3495 */ 3496 3497 /* first get start and end of "diagonal" columns */ 3498 if (csize == PETSC_DECIDE) { 3499 ierr = ISGetSize(isrow,&mglobal);CHKERRQ(ierr); 3500 if (mglobal == n) { /* square matrix */ 3501 nlocal = m; 3502 } else { 3503 nlocal = n/size + ((n % size) > rank); 3504 } 3505 } else { 3506 nlocal = csize; 3507 } 3508 ierr = MPI_Scan(&nlocal,&rend,1,MPIU_INT,MPI_SUM,comm);CHKERRQ(ierr); 3509 rstart = rend - nlocal; 3510 if (rank == size - 1 && rend != n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Local column sizes %D do not add up to total number of columns %D",rend,n); 3511 3512 /* next, compute all the lengths */ 3513 ierr = PetscMalloc((2*m+1)*sizeof(PetscInt),&dlens);CHKERRQ(ierr); 3514 olens = dlens + m; 3515 for (i=0; i<m; i++) { 3516 jend = ii[i+1] - ii[i]; 3517 olen = 0; 3518 dlen = 0; 3519 for (j=0; j<jend; j++) { 3520 if (*jj < rstart || *jj >= rend) olen++; 3521 else dlen++; 3522 jj++; 3523 } 3524 olens[i] = olen; 3525 dlens[i] = dlen; 3526 } 3527 ierr = MatCreate(comm,&M);CHKERRQ(ierr); 3528 ierr = MatSetSizes(M,m,nlocal,PETSC_DECIDE,n);CHKERRQ(ierr); 3529 ierr = MatSetType(M,((PetscObject)mat)->type_name);CHKERRQ(ierr); 3530 ierr = MatMPIAIJSetPreallocation(M,0,dlens,0,olens);CHKERRQ(ierr); 3531 ierr = PetscFree(dlens);CHKERRQ(ierr); 3532 } else { 3533 PetscInt ml,nl; 3534 3535 M = *newmat; 3536 ierr = MatGetLocalSize(M,&ml,&nl);CHKERRQ(ierr); 3537 if (ml != m) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Previous matrix must be same size/layout as request"); 3538 ierr = MatZeroEntries(M);CHKERRQ(ierr); 3539 /* 3540 The next two lines are needed so we may call MatSetValues_MPIAIJ() below directly, 3541 rather than the slower MatSetValues(). 3542 */ 3543 M->was_assembled = PETSC_TRUE; 3544 M->assembled = PETSC_FALSE; 3545 } 3546 ierr = MatGetOwnershipRange(M,&rstart,&rend);CHKERRQ(ierr); 3547 aij = (Mat_SeqAIJ*)(Mreuse)->data; 3548 ii = aij->i; 3549 jj = aij->j; 3550 aa = aij->a; 3551 for (i=0; i<m; i++) { 3552 row = rstart + i; 3553 nz = ii[i+1] - ii[i]; 3554 cwork = jj; jj += nz; 3555 vwork = aa; aa += nz; 3556 ierr = MatSetValues_MPIAIJ(M,1,&row,nz,cwork,vwork,INSERT_VALUES);CHKERRQ(ierr); 3557 } 3558 3559 ierr = MatAssemblyBegin(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3560 ierr = MatAssemblyEnd(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3561 *newmat = M; 3562 3563 /* save submatrix used in processor for next request */ 3564 if (call == MAT_INITIAL_MATRIX) { 3565 ierr = PetscObjectCompose((PetscObject)M,"SubMatrix",(PetscObject)Mreuse);CHKERRQ(ierr); 3566 ierr = PetscObjectDereference((PetscObject)Mreuse);CHKERRQ(ierr); 3567 } 3568 3569 PetscFunctionReturn(0); 3570 } 3571 3572 EXTERN_C_BEGIN 3573 #undef __FUNCT__ 3574 #define __FUNCT__ "MatMPIAIJSetPreallocationCSR_MPIAIJ" 3575 PetscErrorCode MatMPIAIJSetPreallocationCSR_MPIAIJ(Mat B,const PetscInt Ii[],const PetscInt J[],const PetscScalar v[]) 3576 { 3577 PetscInt m,cstart, cend,j,nnz,i,d; 3578 PetscInt *d_nnz,*o_nnz,nnz_max = 0,rstart,ii; 3579 const PetscInt *JJ; 3580 PetscScalar *values; 3581 PetscErrorCode ierr; 3582 3583 PetscFunctionBegin; 3584 if (Ii[0]) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Ii[0] must be 0 it is %D",Ii[0]); 3585 3586 ierr = PetscLayoutSetBlockSize(B->rmap,1);CHKERRQ(ierr); 3587 ierr = PetscLayoutSetBlockSize(B->cmap,1);CHKERRQ(ierr); 3588 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 3589 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 3590 m = B->rmap->n; 3591 cstart = B->cmap->rstart; 3592 cend = B->cmap->rend; 3593 rstart = B->rmap->rstart; 3594 3595 ierr = PetscMalloc2(m,PetscInt,&d_nnz,m,PetscInt,&o_nnz);CHKERRQ(ierr); 3596 3597 #if defined(PETSC_USE_DEBUGGING) 3598 for (i=0; i<m; i++) { 3599 nnz = Ii[i+1]- Ii[i]; 3600 JJ = J + Ii[i]; 3601 if (nnz < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Local row %D has a negative %D number of columns",i,nnz); 3602 if (nnz && (JJ[0] < 0)) SETERRRQ1(PETSC_ERR_ARG_WRONGSTATE,"Row %D starts with negative column index",i,j); 3603 if (nnz && (JJ[nnz-1] >= B->cmap->N) SETERRRQ3(PETSC_ERR_ARG_WRONGSTATE,"Row %D ends with too large a column index %D (max allowed %D)",i,JJ[nnz-1],B->cmap->N); 3604 } 3605 #endif 3606 3607 for (i=0; i<m; i++) { 3608 nnz = Ii[i+1]- Ii[i]; 3609 JJ = J + Ii[i]; 3610 nnz_max = PetscMax(nnz_max,nnz); 3611 d = 0; 3612 for (j=0; j<nnz; j++) { 3613 if (cstart <= JJ[j] && JJ[j] < cend) d++; 3614 } 3615 d_nnz[i] = d; 3616 o_nnz[i] = nnz - d; 3617 } 3618 ierr = MatMPIAIJSetPreallocation(B,0,d_nnz,0,o_nnz);CHKERRQ(ierr); 3619 ierr = PetscFree2(d_nnz,o_nnz);CHKERRQ(ierr); 3620 3621 if (v) values = (PetscScalar*)v; 3622 else { 3623 ierr = PetscMalloc((nnz_max+1)*sizeof(PetscScalar),&values);CHKERRQ(ierr); 3624 ierr = PetscMemzero(values,nnz_max*sizeof(PetscScalar));CHKERRQ(ierr); 3625 } 3626 3627 for (i=0; i<m; i++) { 3628 ii = i + rstart; 3629 nnz = Ii[i+1]- Ii[i]; 3630 ierr = MatSetValues_MPIAIJ(B,1,&ii,nnz,J+Ii[i],values+(v ? Ii[i] : 0),INSERT_VALUES);CHKERRQ(ierr); 3631 } 3632 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3633 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3634 3635 if (!v) { 3636 ierr = PetscFree(values);CHKERRQ(ierr); 3637 } 3638 PetscFunctionReturn(0); 3639 } 3640 EXTERN_C_END 3641 3642 #undef __FUNCT__ 3643 #define __FUNCT__ "MatMPIAIJSetPreallocationCSR" 3644 /*@ 3645 MatMPIAIJSetPreallocationCSR - Allocates memory for a sparse parallel matrix in AIJ format 3646 (the default parallel PETSc format). 3647 3648 Collective on MPI_Comm 3649 3650 Input Parameters: 3651 + B - the matrix 3652 . i - the indices into j for the start of each local row (starts with zero) 3653 . j - the column indices for each local row (starts with zero) 3654 - v - optional values in the matrix 3655 3656 Level: developer 3657 3658 Notes: 3659 The i, j, and a arrays ARE copied by this routine into the internal format used by PETSc; 3660 thus you CANNOT change the matrix entries by changing the values of a[] after you have 3661 called this routine. Use MatCreateMPIAIJWithSplitArrays() to avoid needing to copy the arrays. 3662 3663 The i and j indices are 0 based, and i indices are indices corresponding to the local j array. 3664 3665 The format which is used for the sparse matrix input, is equivalent to a 3666 row-major ordering.. i.e for the following matrix, the input data expected is 3667 as shown: 3668 3669 1 0 0 3670 2 0 3 P0 3671 ------- 3672 4 5 6 P1 3673 3674 Process0 [P0]: rows_owned=[0,1] 3675 i = {0,1,3} [size = nrow+1 = 2+1] 3676 j = {0,0,2} [size = nz = 6] 3677 v = {1,2,3} [size = nz = 6] 3678 3679 Process1 [P1]: rows_owned=[2] 3680 i = {0,3} [size = nrow+1 = 1+1] 3681 j = {0,1,2} [size = nz = 6] 3682 v = {4,5,6} [size = nz = 6] 3683 3684 .keywords: matrix, aij, compressed row, sparse, parallel 3685 3686 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatMPIAIJSetPreallocation(), MatCreateMPIAIJ(), MPIAIJ, 3687 MatCreateSeqAIJWithArrays(), MatCreateMPIAIJWithSplitArrays() 3688 @*/ 3689 PetscErrorCode MatMPIAIJSetPreallocationCSR(Mat B,const PetscInt i[],const PetscInt j[], const PetscScalar v[]) 3690 { 3691 PetscErrorCode ierr; 3692 3693 PetscFunctionBegin; 3694 ierr = PetscTryMethod(B,"MatMPIAIJSetPreallocationCSR_C",(Mat,const PetscInt[],const PetscInt[],const PetscScalar[]),(B,i,j,v));CHKERRQ(ierr); 3695 PetscFunctionReturn(0); 3696 } 3697 3698 #undef __FUNCT__ 3699 #define __FUNCT__ "MatMPIAIJSetPreallocation" 3700 /*@C 3701 MatMPIAIJSetPreallocation - Preallocates memory for a sparse parallel matrix in AIJ format 3702 (the default parallel PETSc format). For good matrix assembly performance 3703 the user should preallocate the matrix storage by setting the parameters 3704 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 3705 performance can be increased by more than a factor of 50. 3706 3707 Collective on MPI_Comm 3708 3709 Input Parameters: 3710 + A - the matrix 3711 . d_nz - number of nonzeros per row in DIAGONAL portion of local submatrix 3712 (same value is used for all local rows) 3713 . d_nnz - array containing the number of nonzeros in the various rows of the 3714 DIAGONAL portion of the local submatrix (possibly different for each row) 3715 or PETSC_NULL, if d_nz is used to specify the nonzero structure. 3716 The size of this array is equal to the number of local rows, i.e 'm'. 3717 You must leave room for the diagonal entry even if it is zero. 3718 . o_nz - number of nonzeros per row in the OFF-DIAGONAL portion of local 3719 submatrix (same value is used for all local rows). 3720 - o_nnz - array containing the number of nonzeros in the various rows of the 3721 OFF-DIAGONAL portion of the local submatrix (possibly different for 3722 each row) or PETSC_NULL, if o_nz is used to specify the nonzero 3723 structure. The size of this array is equal to the number 3724 of local rows, i.e 'm'. 3725 3726 If the *_nnz parameter is given then the *_nz parameter is ignored 3727 3728 The AIJ format (also called the Yale sparse matrix format or 3729 compressed row storage (CSR)), is fully compatible with standard Fortran 77 3730 storage. The stored row and column indices begin with zero. 3731 See the <A href="../../docs/manual.pdf#nameddest=ch_mat">Mat chapter of the users manual</A> for details. 3732 3733 The parallel matrix is partitioned such that the first m0 rows belong to 3734 process 0, the next m1 rows belong to process 1, the next m2 rows belong 3735 to process 2 etc.. where m0,m1,m2... are the input parameter 'm'. 3736 3737 The DIAGONAL portion of the local submatrix of a processor can be defined 3738 as the submatrix which is obtained by extraction the part corresponding to 3739 the rows r1-r2 and columns c1-c2 of the global matrix, where r1 is the 3740 first row that belongs to the processor, r2 is the last row belonging to 3741 the this processor, and c1-c2 is range of indices of the local part of a 3742 vector suitable for applying the matrix to. This is an mxn matrix. In the 3743 common case of a square matrix, the row and column ranges are the same and 3744 the DIAGONAL part is also square. The remaining portion of the local 3745 submatrix (mxN) constitute the OFF-DIAGONAL portion. 3746 3747 If o_nnz, d_nnz are specified, then o_nz, and d_nz are ignored. 3748 3749 You can call MatGetInfo() to get information on how effective the preallocation was; 3750 for example the fields mallocs,nz_allocated,nz_used,nz_unneeded; 3751 You can also run with the option -info and look for messages with the string 3752 malloc in them to see if additional memory allocation was needed. 3753 3754 Example usage: 3755 3756 Consider the following 8x8 matrix with 34 non-zero values, that is 3757 assembled across 3 processors. Lets assume that proc0 owns 3 rows, 3758 proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown 3759 as follows: 3760 3761 .vb 3762 1 2 0 | 0 3 0 | 0 4 3763 Proc0 0 5 6 | 7 0 0 | 8 0 3764 9 0 10 | 11 0 0 | 12 0 3765 ------------------------------------- 3766 13 0 14 | 15 16 17 | 0 0 3767 Proc1 0 18 0 | 19 20 21 | 0 0 3768 0 0 0 | 22 23 0 | 24 0 3769 ------------------------------------- 3770 Proc2 25 26 27 | 0 0 28 | 29 0 3771 30 0 0 | 31 32 33 | 0 34 3772 .ve 3773 3774 This can be represented as a collection of submatrices as: 3775 3776 .vb 3777 A B C 3778 D E F 3779 G H I 3780 .ve 3781 3782 Where the submatrices A,B,C are owned by proc0, D,E,F are 3783 owned by proc1, G,H,I are owned by proc2. 3784 3785 The 'm' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 3786 The 'n' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 3787 The 'M','N' parameters are 8,8, and have the same values on all procs. 3788 3789 The DIAGONAL submatrices corresponding to proc0,proc1,proc2 are 3790 submatrices [A], [E], [I] respectively. The OFF-DIAGONAL submatrices 3791 corresponding to proc0,proc1,proc2 are [BC], [DF], [GH] respectively. 3792 Internally, each processor stores the DIAGONAL part, and the OFF-DIAGONAL 3793 part as SeqAIJ matrices. for eg: proc1 will store [E] as a SeqAIJ 3794 matrix, ans [DF] as another SeqAIJ matrix. 3795 3796 When d_nz, o_nz parameters are specified, d_nz storage elements are 3797 allocated for every row of the local diagonal submatrix, and o_nz 3798 storage locations are allocated for every row of the OFF-DIAGONAL submat. 3799 One way to choose d_nz and o_nz is to use the max nonzerors per local 3800 rows for each of the local DIAGONAL, and the OFF-DIAGONAL submatrices. 3801 In this case, the values of d_nz,o_nz are: 3802 .vb 3803 proc0 : dnz = 2, o_nz = 2 3804 proc1 : dnz = 3, o_nz = 2 3805 proc2 : dnz = 1, o_nz = 4 3806 .ve 3807 We are allocating m*(d_nz+o_nz) storage locations for every proc. This 3808 translates to 3*(2+2)=12 for proc0, 3*(3+2)=15 for proc1, 2*(1+4)=10 3809 for proc3. i.e we are using 12+15+10=37 storage locations to store 3810 34 values. 3811 3812 When d_nnz, o_nnz parameters are specified, the storage is specified 3813 for every row, coresponding to both DIAGONAL and OFF-DIAGONAL submatrices. 3814 In the above case the values for d_nnz,o_nnz are: 3815 .vb 3816 proc0: d_nnz = [2,2,2] and o_nnz = [2,2,2] 3817 proc1: d_nnz = [3,3,2] and o_nnz = [2,1,1] 3818 proc2: d_nnz = [1,1] and o_nnz = [4,4] 3819 .ve 3820 Here the space allocated is sum of all the above values i.e 34, and 3821 hence pre-allocation is perfect. 3822 3823 Level: intermediate 3824 3825 .keywords: matrix, aij, compressed row, sparse, parallel 3826 3827 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatCreateMPIAIJ(), MatMPIAIJSetPreallocationCSR(), 3828 MPIAIJ, MatGetInfo() 3829 @*/ 3830 PetscErrorCode MatMPIAIJSetPreallocation(Mat B,PetscInt d_nz,const PetscInt d_nnz[],PetscInt o_nz,const PetscInt o_nnz[]) 3831 { 3832 PetscErrorCode ierr; 3833 3834 PetscFunctionBegin; 3835 ierr = PetscTryMethod(B,"MatMPIAIJSetPreallocation_C",(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[]),(B,d_nz,d_nnz,o_nz,o_nnz));CHKERRQ(ierr); 3836 PetscFunctionReturn(0); 3837 } 3838 3839 #undef __FUNCT__ 3840 #define __FUNCT__ "MatCreateMPIAIJWithArrays" 3841 /*@ 3842 MatCreateMPIAIJWithArrays - creates a MPI AIJ matrix using arrays that contain in standard 3843 CSR format the local rows. 3844 3845 Collective on MPI_Comm 3846 3847 Input Parameters: 3848 + comm - MPI communicator 3849 . m - number of local rows (Cannot be PETSC_DECIDE) 3850 . n - This value should be the same as the local size used in creating the 3851 x vector for the matrix-vector product y = Ax. (or PETSC_DECIDE to have 3852 calculated if N is given) For square matrices n is almost always m. 3853 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 3854 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 3855 . i - row indices 3856 . j - column indices 3857 - a - matrix values 3858 3859 Output Parameter: 3860 . mat - the matrix 3861 3862 Level: intermediate 3863 3864 Notes: 3865 The i, j, and a arrays ARE copied by this routine into the internal format used by PETSc; 3866 thus you CANNOT change the matrix entries by changing the values of a[] after you have 3867 called this routine. Use MatCreateMPIAIJWithSplitArrays() to avoid needing to copy the arrays. 3868 3869 The i and j indices are 0 based, and i indices are indices corresponding to the local j array. 3870 3871 The format which is used for the sparse matrix input, is equivalent to a 3872 row-major ordering.. i.e for the following matrix, the input data expected is 3873 as shown: 3874 3875 1 0 0 3876 2 0 3 P0 3877 ------- 3878 4 5 6 P1 3879 3880 Process0 [P0]: rows_owned=[0,1] 3881 i = {0,1,3} [size = nrow+1 = 2+1] 3882 j = {0,0,2} [size = nz = 6] 3883 v = {1,2,3} [size = nz = 6] 3884 3885 Process1 [P1]: rows_owned=[2] 3886 i = {0,3} [size = nrow+1 = 1+1] 3887 j = {0,1,2} [size = nz = 6] 3888 v = {4,5,6} [size = nz = 6] 3889 3890 .keywords: matrix, aij, compressed row, sparse, parallel 3891 3892 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatMPIAIJSetPreallocation(), MatMPIAIJSetPreallocationCSR(), 3893 MPIAIJ, MatCreateMPIAIJ(), MatCreateMPIAIJWithSplitArrays() 3894 @*/ 3895 PetscErrorCode MatCreateMPIAIJWithArrays(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt M,PetscInt N,const PetscInt i[],const PetscInt j[],const PetscScalar a[],Mat *mat) 3896 { 3897 PetscErrorCode ierr; 3898 3899 PetscFunctionBegin; 3900 if (i[0]) { 3901 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"i (row indices) must start with 0"); 3902 } 3903 if (m < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"local number of rows (m) cannot be PETSC_DECIDE, or negative"); 3904 ierr = MatCreate(comm,mat);CHKERRQ(ierr); 3905 ierr = MatSetSizes(*mat,m,n,M,N);CHKERRQ(ierr); 3906 ierr = MatSetType(*mat,MATMPIAIJ);CHKERRQ(ierr); 3907 ierr = MatMPIAIJSetPreallocationCSR(*mat,i,j,a);CHKERRQ(ierr); 3908 PetscFunctionReturn(0); 3909 } 3910 3911 #undef __FUNCT__ 3912 #define __FUNCT__ "MatCreateMPIAIJ" 3913 /*@C 3914 MatCreateMPIAIJ - Creates a sparse parallel matrix in AIJ format 3915 (the default parallel PETSc format). For good matrix assembly performance 3916 the user should preallocate the matrix storage by setting the parameters 3917 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 3918 performance can be increased by more than a factor of 50. 3919 3920 Collective on MPI_Comm 3921 3922 Input Parameters: 3923 + comm - MPI communicator 3924 . m - number of local rows (or PETSC_DECIDE to have calculated if M is given) 3925 This value should be the same as the local size used in creating the 3926 y vector for the matrix-vector product y = Ax. 3927 . n - This value should be the same as the local size used in creating the 3928 x vector for the matrix-vector product y = Ax. (or PETSC_DECIDE to have 3929 calculated if N is given) For square matrices n is almost always m. 3930 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 3931 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 3932 . d_nz - number of nonzeros per row in DIAGONAL portion of local submatrix 3933 (same value is used for all local rows) 3934 . d_nnz - array containing the number of nonzeros in the various rows of the 3935 DIAGONAL portion of the local submatrix (possibly different for each row) 3936 or PETSC_NULL, if d_nz is used to specify the nonzero structure. 3937 The size of this array is equal to the number of local rows, i.e 'm'. 3938 You must leave room for the diagonal entry even if it is zero. 3939 . o_nz - number of nonzeros per row in the OFF-DIAGONAL portion of local 3940 submatrix (same value is used for all local rows). 3941 - o_nnz - array containing the number of nonzeros in the various rows of the 3942 OFF-DIAGONAL portion of the local submatrix (possibly different for 3943 each row) or PETSC_NULL, if o_nz is used to specify the nonzero 3944 structure. The size of this array is equal to the number 3945 of local rows, i.e 'm'. 3946 3947 Output Parameter: 3948 . A - the matrix 3949 3950 It is recommended that one use the MatCreate(), MatSetType() and/or MatSetFromOptions(), 3951 MatXXXXSetPreallocation() paradgm instead of this routine directly. 3952 [MatXXXXSetPreallocation() is, for example, MatSeqAIJSetPreallocation] 3953 3954 Notes: 3955 If the *_nnz parameter is given then the *_nz parameter is ignored 3956 3957 m,n,M,N parameters specify the size of the matrix, and its partitioning across 3958 processors, while d_nz,d_nnz,o_nz,o_nnz parameters specify the approximate 3959 storage requirements for this matrix. 3960 3961 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one 3962 processor than it must be used on all processors that share the object for 3963 that argument. 3964 3965 The user MUST specify either the local or global matrix dimensions 3966 (possibly both). 3967 3968 The parallel matrix is partitioned across processors such that the 3969 first m0 rows belong to process 0, the next m1 rows belong to 3970 process 1, the next m2 rows belong to process 2 etc.. where 3971 m0,m1,m2,.. are the input parameter 'm'. i.e each processor stores 3972 values corresponding to [m x N] submatrix. 3973 3974 The columns are logically partitioned with the n0 columns belonging 3975 to 0th partition, the next n1 columns belonging to the next 3976 partition etc.. where n0,n1,n2... are the the input parameter 'n'. 3977 3978 The DIAGONAL portion of the local submatrix on any given processor 3979 is the submatrix corresponding to the rows and columns m,n 3980 corresponding to the given processor. i.e diagonal matrix on 3981 process 0 is [m0 x n0], diagonal matrix on process 1 is [m1 x n1] 3982 etc. The remaining portion of the local submatrix [m x (N-n)] 3983 constitute the OFF-DIAGONAL portion. The example below better 3984 illustrates this concept. 3985 3986 For a square global matrix we define each processor's diagonal portion 3987 to be its local rows and the corresponding columns (a square submatrix); 3988 each processor's off-diagonal portion encompasses the remainder of the 3989 local matrix (a rectangular submatrix). 3990 3991 If o_nnz, d_nnz are specified, then o_nz, and d_nz are ignored. 3992 3993 When calling this routine with a single process communicator, a matrix of 3994 type SEQAIJ is returned. If a matrix of type MPIAIJ is desired for this 3995 type of communicator, use the construction mechanism: 3996 MatCreate(...,&A); MatSetType(A,MATMPIAIJ); MatSetSizes(A, m,n,M,N); MatMPIAIJSetPreallocation(A,...); 3997 3998 By default, this format uses inodes (identical nodes) when possible. 3999 We search for consecutive rows with the same nonzero structure, thereby 4000 reusing matrix information to achieve increased efficiency. 4001 4002 Options Database Keys: 4003 + -mat_no_inode - Do not use inodes 4004 . -mat_inode_limit <limit> - Sets inode limit (max limit=5) 4005 - -mat_aij_oneindex - Internally use indexing starting at 1 4006 rather than 0. Note that when calling MatSetValues(), 4007 the user still MUST index entries starting at 0! 4008 4009 4010 Example usage: 4011 4012 Consider the following 8x8 matrix with 34 non-zero values, that is 4013 assembled across 3 processors. Lets assume that proc0 owns 3 rows, 4014 proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown 4015 as follows: 4016 4017 .vb 4018 1 2 0 | 0 3 0 | 0 4 4019 Proc0 0 5 6 | 7 0 0 | 8 0 4020 9 0 10 | 11 0 0 | 12 0 4021 ------------------------------------- 4022 13 0 14 | 15 16 17 | 0 0 4023 Proc1 0 18 0 | 19 20 21 | 0 0 4024 0 0 0 | 22 23 0 | 24 0 4025 ------------------------------------- 4026 Proc2 25 26 27 | 0 0 28 | 29 0 4027 30 0 0 | 31 32 33 | 0 34 4028 .ve 4029 4030 This can be represented as a collection of submatrices as: 4031 4032 .vb 4033 A B C 4034 D E F 4035 G H I 4036 .ve 4037 4038 Where the submatrices A,B,C are owned by proc0, D,E,F are 4039 owned by proc1, G,H,I are owned by proc2. 4040 4041 The 'm' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 4042 The 'n' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 4043 The 'M','N' parameters are 8,8, and have the same values on all procs. 4044 4045 The DIAGONAL submatrices corresponding to proc0,proc1,proc2 are 4046 submatrices [A], [E], [I] respectively. The OFF-DIAGONAL submatrices 4047 corresponding to proc0,proc1,proc2 are [BC], [DF], [GH] respectively. 4048 Internally, each processor stores the DIAGONAL part, and the OFF-DIAGONAL 4049 part as SeqAIJ matrices. for eg: proc1 will store [E] as a SeqAIJ 4050 matrix, ans [DF] as another SeqAIJ matrix. 4051 4052 When d_nz, o_nz parameters are specified, d_nz storage elements are 4053 allocated for every row of the local diagonal submatrix, and o_nz 4054 storage locations are allocated for every row of the OFF-DIAGONAL submat. 4055 One way to choose d_nz and o_nz is to use the max nonzerors per local 4056 rows for each of the local DIAGONAL, and the OFF-DIAGONAL submatrices. 4057 In this case, the values of d_nz,o_nz are: 4058 .vb 4059 proc0 : dnz = 2, o_nz = 2 4060 proc1 : dnz = 3, o_nz = 2 4061 proc2 : dnz = 1, o_nz = 4 4062 .ve 4063 We are allocating m*(d_nz+o_nz) storage locations for every proc. This 4064 translates to 3*(2+2)=12 for proc0, 3*(3+2)=15 for proc1, 2*(1+4)=10 4065 for proc3. i.e we are using 12+15+10=37 storage locations to store 4066 34 values. 4067 4068 When d_nnz, o_nnz parameters are specified, the storage is specified 4069 for every row, coresponding to both DIAGONAL and OFF-DIAGONAL submatrices. 4070 In the above case the values for d_nnz,o_nnz are: 4071 .vb 4072 proc0: d_nnz = [2,2,2] and o_nnz = [2,2,2] 4073 proc1: d_nnz = [3,3,2] and o_nnz = [2,1,1] 4074 proc2: d_nnz = [1,1] and o_nnz = [4,4] 4075 .ve 4076 Here the space allocated is sum of all the above values i.e 34, and 4077 hence pre-allocation is perfect. 4078 4079 Level: intermediate 4080 4081 .keywords: matrix, aij, compressed row, sparse, parallel 4082 4083 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatMPIAIJSetPreallocation(), MatMPIAIJSetPreallocationCSR(), 4084 MPIAIJ, MatCreateMPIAIJWithArrays() 4085 @*/ 4086 PetscErrorCode MatCreateMPIAIJ(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt M,PetscInt N,PetscInt d_nz,const PetscInt d_nnz[],PetscInt o_nz,const PetscInt o_nnz[],Mat *A) 4087 { 4088 PetscErrorCode ierr; 4089 PetscMPIInt size; 4090 4091 PetscFunctionBegin; 4092 ierr = MatCreate(comm,A);CHKERRQ(ierr); 4093 ierr = MatSetSizes(*A,m,n,M,N);CHKERRQ(ierr); 4094 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 4095 if (size > 1) { 4096 ierr = MatSetType(*A,MATMPIAIJ);CHKERRQ(ierr); 4097 ierr = MatMPIAIJSetPreallocation(*A,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 4098 } else { 4099 ierr = MatSetType(*A,MATSEQAIJ);CHKERRQ(ierr); 4100 ierr = MatSeqAIJSetPreallocation(*A,d_nz,d_nnz);CHKERRQ(ierr); 4101 } 4102 PetscFunctionReturn(0); 4103 } 4104 4105 #undef __FUNCT__ 4106 #define __FUNCT__ "MatMPIAIJGetSeqAIJ" 4107 PetscErrorCode MatMPIAIJGetSeqAIJ(Mat A,Mat *Ad,Mat *Ao,PetscInt *colmap[]) 4108 { 4109 Mat_MPIAIJ *a = (Mat_MPIAIJ *)A->data; 4110 4111 PetscFunctionBegin; 4112 *Ad = a->A; 4113 *Ao = a->B; 4114 *colmap = a->garray; 4115 PetscFunctionReturn(0); 4116 } 4117 4118 #undef __FUNCT__ 4119 #define __FUNCT__ "MatSetColoring_MPIAIJ" 4120 PetscErrorCode MatSetColoring_MPIAIJ(Mat A,ISColoring coloring) 4121 { 4122 PetscErrorCode ierr; 4123 PetscInt i; 4124 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 4125 4126 PetscFunctionBegin; 4127 if (coloring->ctype == IS_COLORING_GLOBAL) { 4128 ISColoringValue *allcolors,*colors; 4129 ISColoring ocoloring; 4130 4131 /* set coloring for diagonal portion */ 4132 ierr = MatSetColoring_SeqAIJ(a->A,coloring);CHKERRQ(ierr); 4133 4134 /* set coloring for off-diagonal portion */ 4135 ierr = ISAllGatherColors(((PetscObject)A)->comm,coloring->n,coloring->colors,PETSC_NULL,&allcolors);CHKERRQ(ierr); 4136 ierr = PetscMalloc((a->B->cmap->n+1)*sizeof(ISColoringValue),&colors);CHKERRQ(ierr); 4137 for (i=0; i<a->B->cmap->n; i++) { 4138 colors[i] = allcolors[a->garray[i]]; 4139 } 4140 ierr = PetscFree(allcolors);CHKERRQ(ierr); 4141 ierr = ISColoringCreate(MPI_COMM_SELF,coloring->n,a->B->cmap->n,colors,&ocoloring);CHKERRQ(ierr); 4142 ierr = MatSetColoring_SeqAIJ(a->B,ocoloring);CHKERRQ(ierr); 4143 ierr = ISColoringDestroy(&ocoloring);CHKERRQ(ierr); 4144 } else if (coloring->ctype == IS_COLORING_GHOSTED) { 4145 ISColoringValue *colors; 4146 PetscInt *larray; 4147 ISColoring ocoloring; 4148 4149 /* set coloring for diagonal portion */ 4150 ierr = PetscMalloc((a->A->cmap->n+1)*sizeof(PetscInt),&larray);CHKERRQ(ierr); 4151 for (i=0; i<a->A->cmap->n; i++) { 4152 larray[i] = i + A->cmap->rstart; 4153 } 4154 ierr = ISGlobalToLocalMappingApply(A->cmapping,IS_GTOLM_MASK,a->A->cmap->n,larray,PETSC_NULL,larray);CHKERRQ(ierr); 4155 ierr = PetscMalloc((a->A->cmap->n+1)*sizeof(ISColoringValue),&colors);CHKERRQ(ierr); 4156 for (i=0; i<a->A->cmap->n; i++) { 4157 colors[i] = coloring->colors[larray[i]]; 4158 } 4159 ierr = PetscFree(larray);CHKERRQ(ierr); 4160 ierr = ISColoringCreate(PETSC_COMM_SELF,coloring->n,a->A->cmap->n,colors,&ocoloring);CHKERRQ(ierr); 4161 ierr = MatSetColoring_SeqAIJ(a->A,ocoloring);CHKERRQ(ierr); 4162 ierr = ISColoringDestroy(&ocoloring);CHKERRQ(ierr); 4163 4164 /* set coloring for off-diagonal portion */ 4165 ierr = PetscMalloc((a->B->cmap->n+1)*sizeof(PetscInt),&larray);CHKERRQ(ierr); 4166 ierr = ISGlobalToLocalMappingApply(A->cmapping,IS_GTOLM_MASK,a->B->cmap->n,a->garray,PETSC_NULL,larray);CHKERRQ(ierr); 4167 ierr = PetscMalloc((a->B->cmap->n+1)*sizeof(ISColoringValue),&colors);CHKERRQ(ierr); 4168 for (i=0; i<a->B->cmap->n; i++) { 4169 colors[i] = coloring->colors[larray[i]]; 4170 } 4171 ierr = PetscFree(larray);CHKERRQ(ierr); 4172 ierr = ISColoringCreate(MPI_COMM_SELF,coloring->n,a->B->cmap->n,colors,&ocoloring);CHKERRQ(ierr); 4173 ierr = MatSetColoring_SeqAIJ(a->B,ocoloring);CHKERRQ(ierr); 4174 ierr = ISColoringDestroy(&ocoloring);CHKERRQ(ierr); 4175 } else SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support ISColoringType %d",(int)coloring->ctype); 4176 4177 PetscFunctionReturn(0); 4178 } 4179 4180 #if defined(PETSC_HAVE_ADIC) 4181 #undef __FUNCT__ 4182 #define __FUNCT__ "MatSetValuesAdic_MPIAIJ" 4183 PetscErrorCode MatSetValuesAdic_MPIAIJ(Mat A,void *advalues) 4184 { 4185 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 4186 PetscErrorCode ierr; 4187 4188 PetscFunctionBegin; 4189 ierr = MatSetValuesAdic_SeqAIJ(a->A,advalues);CHKERRQ(ierr); 4190 ierr = MatSetValuesAdic_SeqAIJ(a->B,advalues);CHKERRQ(ierr); 4191 PetscFunctionReturn(0); 4192 } 4193 #endif 4194 4195 #undef __FUNCT__ 4196 #define __FUNCT__ "MatSetValuesAdifor_MPIAIJ" 4197 PetscErrorCode MatSetValuesAdifor_MPIAIJ(Mat A,PetscInt nl,void *advalues) 4198 { 4199 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 4200 PetscErrorCode ierr; 4201 4202 PetscFunctionBegin; 4203 ierr = MatSetValuesAdifor_SeqAIJ(a->A,nl,advalues);CHKERRQ(ierr); 4204 ierr = MatSetValuesAdifor_SeqAIJ(a->B,nl,advalues);CHKERRQ(ierr); 4205 PetscFunctionReturn(0); 4206 } 4207 4208 #undef __FUNCT__ 4209 #define __FUNCT__ "MatMerge" 4210 /*@ 4211 MatMerge - Creates a single large PETSc matrix by concatinating sequential 4212 matrices from each processor 4213 4214 Collective on MPI_Comm 4215 4216 Input Parameters: 4217 + comm - the communicators the parallel matrix will live on 4218 . inmat - the input sequential matrices 4219 . n - number of local columns (or PETSC_DECIDE) 4220 - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 4221 4222 Output Parameter: 4223 . outmat - the parallel matrix generated 4224 4225 Level: advanced 4226 4227 Notes: The number of columns of the matrix in EACH processor MUST be the same. 4228 4229 @*/ 4230 PetscErrorCode MatMerge(MPI_Comm comm,Mat inmat,PetscInt n,MatReuse scall,Mat *outmat) 4231 { 4232 PetscErrorCode ierr; 4233 PetscInt m,N,i,rstart,nnz,Ii,*dnz,*onz; 4234 PetscInt *indx; 4235 PetscScalar *values; 4236 4237 PetscFunctionBegin; 4238 ierr = MatGetSize(inmat,&m,&N);CHKERRQ(ierr); 4239 if (scall == MAT_INITIAL_MATRIX){ 4240 /* count nonzeros in each row, for diagonal and off diagonal portion of matrix */ 4241 if (n == PETSC_DECIDE){ 4242 ierr = PetscSplitOwnership(comm,&n,&N);CHKERRQ(ierr); 4243 } 4244 ierr = MPI_Scan(&m, &rstart,1,MPIU_INT,MPI_SUM,comm);CHKERRQ(ierr); 4245 rstart -= m; 4246 4247 ierr = MatPreallocateInitialize(comm,m,n,dnz,onz);CHKERRQ(ierr); 4248 for (i=0;i<m;i++) { 4249 ierr = MatGetRow_SeqAIJ(inmat,i,&nnz,&indx,PETSC_NULL);CHKERRQ(ierr); 4250 ierr = MatPreallocateSet(i+rstart,nnz,indx,dnz,onz);CHKERRQ(ierr); 4251 ierr = MatRestoreRow_SeqAIJ(inmat,i,&nnz,&indx,PETSC_NULL);CHKERRQ(ierr); 4252 } 4253 /* This routine will ONLY return MPIAIJ type matrix */ 4254 ierr = MatCreate(comm,outmat);CHKERRQ(ierr); 4255 ierr = MatSetSizes(*outmat,m,n,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 4256 ierr = MatSetType(*outmat,MATMPIAIJ);CHKERRQ(ierr); 4257 ierr = MatMPIAIJSetPreallocation(*outmat,0,dnz,0,onz);CHKERRQ(ierr); 4258 ierr = MatPreallocateFinalize(dnz,onz);CHKERRQ(ierr); 4259 4260 } else if (scall == MAT_REUSE_MATRIX){ 4261 ierr = MatGetOwnershipRange(*outmat,&rstart,PETSC_NULL);CHKERRQ(ierr); 4262 } else { 4263 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Invalid MatReuse %d",(int)scall); 4264 } 4265 4266 for (i=0;i<m;i++) { 4267 ierr = MatGetRow_SeqAIJ(inmat,i,&nnz,&indx,&values);CHKERRQ(ierr); 4268 Ii = i + rstart; 4269 ierr = MatSetValues(*outmat,1,&Ii,nnz,indx,values,INSERT_VALUES);CHKERRQ(ierr); 4270 ierr = MatRestoreRow_SeqAIJ(inmat,i,&nnz,&indx,&values);CHKERRQ(ierr); 4271 } 4272 ierr = MatDestroy(&inmat);CHKERRQ(ierr); 4273 ierr = MatAssemblyBegin(*outmat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4274 ierr = MatAssemblyEnd(*outmat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4275 4276 PetscFunctionReturn(0); 4277 } 4278 4279 #undef __FUNCT__ 4280 #define __FUNCT__ "MatFileSplit" 4281 PetscErrorCode MatFileSplit(Mat A,char *outfile) 4282 { 4283 PetscErrorCode ierr; 4284 PetscMPIInt rank; 4285 PetscInt m,N,i,rstart,nnz; 4286 size_t len; 4287 const PetscInt *indx; 4288 PetscViewer out; 4289 char *name; 4290 Mat B; 4291 const PetscScalar *values; 4292 4293 PetscFunctionBegin; 4294 ierr = MatGetLocalSize(A,&m,0);CHKERRQ(ierr); 4295 ierr = MatGetSize(A,0,&N);CHKERRQ(ierr); 4296 /* Should this be the type of the diagonal block of A? */ 4297 ierr = MatCreate(PETSC_COMM_SELF,&B);CHKERRQ(ierr); 4298 ierr = MatSetSizes(B,m,N,m,N);CHKERRQ(ierr); 4299 ierr = MatSetType(B,MATSEQAIJ);CHKERRQ(ierr); 4300 ierr = MatSeqAIJSetPreallocation(B,0,PETSC_NULL);CHKERRQ(ierr); 4301 ierr = MatGetOwnershipRange(A,&rstart,0);CHKERRQ(ierr); 4302 for (i=0;i<m;i++) { 4303 ierr = MatGetRow(A,i+rstart,&nnz,&indx,&values);CHKERRQ(ierr); 4304 ierr = MatSetValues(B,1,&i,nnz,indx,values,INSERT_VALUES);CHKERRQ(ierr); 4305 ierr = MatRestoreRow(A,i+rstart,&nnz,&indx,&values);CHKERRQ(ierr); 4306 } 4307 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4308 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4309 4310 ierr = MPI_Comm_rank(((PetscObject)A)->comm,&rank);CHKERRQ(ierr); 4311 ierr = PetscStrlen(outfile,&len);CHKERRQ(ierr); 4312 ierr = PetscMalloc((len+5)*sizeof(char),&name);CHKERRQ(ierr); 4313 sprintf(name,"%s.%d",outfile,rank); 4314 ierr = PetscViewerBinaryOpen(PETSC_COMM_SELF,name,FILE_MODE_APPEND,&out);CHKERRQ(ierr); 4315 ierr = PetscFree(name); 4316 ierr = MatView(B,out);CHKERRQ(ierr); 4317 ierr = PetscViewerDestroy(&out);CHKERRQ(ierr); 4318 ierr = MatDestroy(&B);CHKERRQ(ierr); 4319 PetscFunctionReturn(0); 4320 } 4321 4322 extern PetscErrorCode MatDestroy_MPIAIJ(Mat); 4323 #undef __FUNCT__ 4324 #define __FUNCT__ "MatDestroy_MPIAIJ_SeqsToMPI" 4325 PetscErrorCode MatDestroy_MPIAIJ_SeqsToMPI(Mat A) 4326 { 4327 PetscErrorCode ierr; 4328 Mat_Merge_SeqsToMPI *merge; 4329 PetscContainer container; 4330 4331 PetscFunctionBegin; 4332 ierr = PetscObjectQuery((PetscObject)A,"MatMergeSeqsToMPI",(PetscObject *)&container);CHKERRQ(ierr); 4333 if (container) { 4334 ierr = PetscContainerGetPointer(container,(void **)&merge);CHKERRQ(ierr); 4335 ierr = PetscFree(merge->id_r);CHKERRQ(ierr); 4336 ierr = PetscFree(merge->len_s);CHKERRQ(ierr); 4337 ierr = PetscFree(merge->len_r);CHKERRQ(ierr); 4338 ierr = PetscFree(merge->bi);CHKERRQ(ierr); 4339 ierr = PetscFree(merge->bj);CHKERRQ(ierr); 4340 ierr = PetscFree(merge->buf_ri[0]);CHKERRQ(ierr); 4341 ierr = PetscFree(merge->buf_ri);CHKERRQ(ierr); 4342 ierr = PetscFree(merge->buf_rj[0]);CHKERRQ(ierr); 4343 ierr = PetscFree(merge->buf_rj);CHKERRQ(ierr); 4344 ierr = PetscFree(merge->coi);CHKERRQ(ierr); 4345 ierr = PetscFree(merge->coj);CHKERRQ(ierr); 4346 ierr = PetscFree(merge->owners_co);CHKERRQ(ierr); 4347 ierr = PetscLayoutDestroy(&merge->rowmap);CHKERRQ(ierr); 4348 4349 ierr = PetscContainerDestroy(&container);CHKERRQ(ierr); 4350 ierr = PetscObjectCompose((PetscObject)A,"MatMergeSeqsToMPI",0);CHKERRQ(ierr); 4351 } 4352 ierr = PetscFree(merge);CHKERRQ(ierr); 4353 4354 ierr = MatDestroy_MPIAIJ(A);CHKERRQ(ierr); 4355 PetscFunctionReturn(0); 4356 } 4357 4358 #include <../src/mat/utils/freespace.h> 4359 #include <petscbt.h> 4360 4361 #undef __FUNCT__ 4362 #define __FUNCT__ "MatMerge_SeqsToMPINumeric" 4363 /*@C 4364 MatMerge_SeqsToMPI - Creates a MPIAIJ matrix by adding sequential 4365 matrices from each processor 4366 4367 Collective on MPI_Comm 4368 4369 Input Parameters: 4370 + comm - the communicators the parallel matrix will live on 4371 . seqmat - the input sequential matrices 4372 . m - number of local rows (or PETSC_DECIDE) 4373 . n - number of local columns (or PETSC_DECIDE) 4374 - scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 4375 4376 Output Parameter: 4377 . mpimat - the parallel matrix generated 4378 4379 Level: advanced 4380 4381 Notes: 4382 The dimensions of the sequential matrix in each processor MUST be the same. 4383 The input seqmat is included into the container "Mat_Merge_SeqsToMPI", and will be 4384 destroyed when mpimat is destroyed. Call PetscObjectQuery() to access seqmat. 4385 @*/ 4386 PetscErrorCode MatMerge_SeqsToMPINumeric(Mat seqmat,Mat mpimat) 4387 { 4388 PetscErrorCode ierr; 4389 MPI_Comm comm=((PetscObject)mpimat)->comm; 4390 Mat_SeqAIJ *a=(Mat_SeqAIJ*)seqmat->data; 4391 PetscMPIInt size,rank,taga,*len_s; 4392 PetscInt N=mpimat->cmap->N,i,j,*owners,*ai=a->i,*aj=a->j; 4393 PetscInt proc,m; 4394 PetscInt **buf_ri,**buf_rj; 4395 PetscInt k,anzi,*bj_i,*bi,*bj,arow,bnzi,nextaj; 4396 PetscInt nrows,**buf_ri_k,**nextrow,**nextai; 4397 MPI_Request *s_waits,*r_waits; 4398 MPI_Status *status; 4399 MatScalar *aa=a->a; 4400 MatScalar **abuf_r,*ba_i; 4401 Mat_Merge_SeqsToMPI *merge; 4402 PetscContainer container; 4403 4404 PetscFunctionBegin; 4405 ierr = PetscLogEventBegin(MAT_Seqstompinum,seqmat,0,0,0);CHKERRQ(ierr); 4406 4407 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 4408 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 4409 4410 ierr = PetscObjectQuery((PetscObject)mpimat,"MatMergeSeqsToMPI",(PetscObject *)&container);CHKERRQ(ierr); 4411 if (container) { 4412 ierr = PetscContainerGetPointer(container,(void **)&merge);CHKERRQ(ierr); 4413 } 4414 bi = merge->bi; 4415 bj = merge->bj; 4416 buf_ri = merge->buf_ri; 4417 buf_rj = merge->buf_rj; 4418 4419 ierr = PetscMalloc(size*sizeof(MPI_Status),&status);CHKERRQ(ierr); 4420 owners = merge->rowmap->range; 4421 len_s = merge->len_s; 4422 4423 /* send and recv matrix values */ 4424 /*-----------------------------*/ 4425 ierr = PetscObjectGetNewTag((PetscObject)mpimat,&taga);CHKERRQ(ierr); 4426 ierr = PetscPostIrecvScalar(comm,taga,merge->nrecv,merge->id_r,merge->len_r,&abuf_r,&r_waits);CHKERRQ(ierr); 4427 4428 ierr = PetscMalloc((merge->nsend+1)*sizeof(MPI_Request),&s_waits);CHKERRQ(ierr); 4429 for (proc=0,k=0; proc<size; proc++){ 4430 if (!len_s[proc]) continue; 4431 i = owners[proc]; 4432 ierr = MPI_Isend(aa+ai[i],len_s[proc],MPIU_MATSCALAR,proc,taga,comm,s_waits+k);CHKERRQ(ierr); 4433 k++; 4434 } 4435 4436 if (merge->nrecv) {ierr = MPI_Waitall(merge->nrecv,r_waits,status);CHKERRQ(ierr);} 4437 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,s_waits,status);CHKERRQ(ierr);} 4438 ierr = PetscFree(status);CHKERRQ(ierr); 4439 4440 ierr = PetscFree(s_waits);CHKERRQ(ierr); 4441 ierr = PetscFree(r_waits);CHKERRQ(ierr); 4442 4443 /* insert mat values of mpimat */ 4444 /*----------------------------*/ 4445 ierr = PetscMalloc(N*sizeof(PetscScalar),&ba_i);CHKERRQ(ierr); 4446 ierr = PetscMalloc3(merge->nrecv,PetscInt*,&buf_ri_k,merge->nrecv,PetscInt*,&nextrow,merge->nrecv,PetscInt*,&nextai);CHKERRQ(ierr); 4447 4448 for (k=0; k<merge->nrecv; k++){ 4449 buf_ri_k[k] = buf_ri[k]; /* beginning of k-th recved i-structure */ 4450 nrows = *(buf_ri_k[k]); 4451 nextrow[k] = buf_ri_k[k]+1; /* next row number of k-th recved i-structure */ 4452 nextai[k] = buf_ri_k[k] + (nrows + 1);/* poins to the next i-structure of k-th recved i-structure */ 4453 } 4454 4455 /* set values of ba */ 4456 m = merge->rowmap->n; 4457 for (i=0; i<m; i++) { 4458 arow = owners[rank] + i; 4459 bj_i = bj+bi[i]; /* col indices of the i-th row of mpimat */ 4460 bnzi = bi[i+1] - bi[i]; 4461 ierr = PetscMemzero(ba_i,bnzi*sizeof(PetscScalar));CHKERRQ(ierr); 4462 4463 /* add local non-zero vals of this proc's seqmat into ba */ 4464 anzi = ai[arow+1] - ai[arow]; 4465 aj = a->j + ai[arow]; 4466 aa = a->a + ai[arow]; 4467 nextaj = 0; 4468 for (j=0; nextaj<anzi; j++){ 4469 if (*(bj_i + j) == aj[nextaj]){ /* bcol == acol */ 4470 ba_i[j] += aa[nextaj++]; 4471 } 4472 } 4473 4474 /* add received vals into ba */ 4475 for (k=0; k<merge->nrecv; k++){ /* k-th received message */ 4476 /* i-th row */ 4477 if (i == *nextrow[k]) { 4478 anzi = *(nextai[k]+1) - *nextai[k]; 4479 aj = buf_rj[k] + *(nextai[k]); 4480 aa = abuf_r[k] + *(nextai[k]); 4481 nextaj = 0; 4482 for (j=0; nextaj<anzi; j++){ 4483 if (*(bj_i + j) == aj[nextaj]){ /* bcol == acol */ 4484 ba_i[j] += aa[nextaj++]; 4485 } 4486 } 4487 nextrow[k]++; nextai[k]++; 4488 } 4489 } 4490 ierr = MatSetValues(mpimat,1,&arow,bnzi,bj_i,ba_i,INSERT_VALUES);CHKERRQ(ierr); 4491 } 4492 ierr = MatAssemblyBegin(mpimat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4493 ierr = MatAssemblyEnd(mpimat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4494 4495 ierr = PetscFree(abuf_r[0]);CHKERRQ(ierr); 4496 ierr = PetscFree(abuf_r);CHKERRQ(ierr); 4497 ierr = PetscFree(ba_i);CHKERRQ(ierr); 4498 ierr = PetscFree3(buf_ri_k,nextrow,nextai);CHKERRQ(ierr); 4499 ierr = PetscLogEventEnd(MAT_Seqstompinum,seqmat,0,0,0);CHKERRQ(ierr); 4500 PetscFunctionReturn(0); 4501 } 4502 4503 #undef __FUNCT__ 4504 #define __FUNCT__ "MatMerge_SeqsToMPISymbolic" 4505 PetscErrorCode MatMerge_SeqsToMPISymbolic(MPI_Comm comm,Mat seqmat,PetscInt m,PetscInt n,Mat *mpimat) 4506 { 4507 PetscErrorCode ierr; 4508 Mat B_mpi; 4509 Mat_SeqAIJ *a=(Mat_SeqAIJ*)seqmat->data; 4510 PetscMPIInt size,rank,tagi,tagj,*len_s,*len_si,*len_ri; 4511 PetscInt **buf_rj,**buf_ri,**buf_ri_k; 4512 PetscInt M=seqmat->rmap->n,N=seqmat->cmap->n,i,*owners,*ai=a->i,*aj=a->j; 4513 PetscInt len,proc,*dnz,*onz; 4514 PetscInt k,anzi,*bi,*bj,*lnk,nlnk,arow,bnzi,nspacedouble=0; 4515 PetscInt nrows,*buf_s,*buf_si,*buf_si_i,**nextrow,**nextai; 4516 MPI_Request *si_waits,*sj_waits,*ri_waits,*rj_waits; 4517 MPI_Status *status; 4518 PetscFreeSpaceList free_space=PETSC_NULL,current_space=PETSC_NULL; 4519 PetscBT lnkbt; 4520 Mat_Merge_SeqsToMPI *merge; 4521 PetscContainer container; 4522 4523 PetscFunctionBegin; 4524 ierr = PetscLogEventBegin(MAT_Seqstompisym,seqmat,0,0,0);CHKERRQ(ierr); 4525 4526 /* make sure it is a PETSc comm */ 4527 ierr = PetscCommDuplicate(comm,&comm,PETSC_NULL);CHKERRQ(ierr); 4528 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 4529 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 4530 4531 ierr = PetscNew(Mat_Merge_SeqsToMPI,&merge);CHKERRQ(ierr); 4532 ierr = PetscMalloc(size*sizeof(MPI_Status),&status);CHKERRQ(ierr); 4533 4534 /* determine row ownership */ 4535 /*---------------------------------------------------------*/ 4536 ierr = PetscLayoutCreate(comm,&merge->rowmap);CHKERRQ(ierr); 4537 ierr = PetscLayoutSetLocalSize(merge->rowmap,m);CHKERRQ(ierr); 4538 ierr = PetscLayoutSetSize(merge->rowmap,M);CHKERRQ(ierr); 4539 ierr = PetscLayoutSetBlockSize(merge->rowmap,1);CHKERRQ(ierr); 4540 ierr = PetscLayoutSetUp(merge->rowmap);CHKERRQ(ierr); 4541 ierr = PetscMalloc(size*sizeof(PetscMPIInt),&len_si);CHKERRQ(ierr); 4542 ierr = PetscMalloc(size*sizeof(PetscMPIInt),&merge->len_s);CHKERRQ(ierr); 4543 4544 m = merge->rowmap->n; 4545 M = merge->rowmap->N; 4546 owners = merge->rowmap->range; 4547 4548 /* determine the number of messages to send, their lengths */ 4549 /*---------------------------------------------------------*/ 4550 len_s = merge->len_s; 4551 4552 len = 0; /* length of buf_si[] */ 4553 merge->nsend = 0; 4554 for (proc=0; proc<size; proc++){ 4555 len_si[proc] = 0; 4556 if (proc == rank){ 4557 len_s[proc] = 0; 4558 } else { 4559 len_si[proc] = owners[proc+1] - owners[proc] + 1; 4560 len_s[proc] = ai[owners[proc+1]] - ai[owners[proc]]; /* num of rows to be sent to [proc] */ 4561 } 4562 if (len_s[proc]) { 4563 merge->nsend++; 4564 nrows = 0; 4565 for (i=owners[proc]; i<owners[proc+1]; i++){ 4566 if (ai[i+1] > ai[i]) nrows++; 4567 } 4568 len_si[proc] = 2*(nrows+1); 4569 len += len_si[proc]; 4570 } 4571 } 4572 4573 /* determine the number and length of messages to receive for ij-structure */ 4574 /*-------------------------------------------------------------------------*/ 4575 ierr = PetscGatherNumberOfMessages(comm,PETSC_NULL,len_s,&merge->nrecv);CHKERRQ(ierr); 4576 ierr = PetscGatherMessageLengths2(comm,merge->nsend,merge->nrecv,len_s,len_si,&merge->id_r,&merge->len_r,&len_ri);CHKERRQ(ierr); 4577 4578 /* post the Irecv of j-structure */ 4579 /*-------------------------------*/ 4580 ierr = PetscCommGetNewTag(comm,&tagj);CHKERRQ(ierr); 4581 ierr = PetscPostIrecvInt(comm,tagj,merge->nrecv,merge->id_r,merge->len_r,&buf_rj,&rj_waits);CHKERRQ(ierr); 4582 4583 /* post the Isend of j-structure */ 4584 /*--------------------------------*/ 4585 ierr = PetscMalloc2(merge->nsend,MPI_Request,&si_waits,merge->nsend,MPI_Request,&sj_waits);CHKERRQ(ierr); 4586 4587 for (proc=0, k=0; proc<size; proc++){ 4588 if (!len_s[proc]) continue; 4589 i = owners[proc]; 4590 ierr = MPI_Isend(aj+ai[i],len_s[proc],MPIU_INT,proc,tagj,comm,sj_waits+k);CHKERRQ(ierr); 4591 k++; 4592 } 4593 4594 /* receives and sends of j-structure are complete */ 4595 /*------------------------------------------------*/ 4596 if (merge->nrecv) {ierr = MPI_Waitall(merge->nrecv,rj_waits,status);CHKERRQ(ierr);} 4597 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,sj_waits,status);CHKERRQ(ierr);} 4598 4599 /* send and recv i-structure */ 4600 /*---------------------------*/ 4601 ierr = PetscCommGetNewTag(comm,&tagi);CHKERRQ(ierr); 4602 ierr = PetscPostIrecvInt(comm,tagi,merge->nrecv,merge->id_r,len_ri,&buf_ri,&ri_waits);CHKERRQ(ierr); 4603 4604 ierr = PetscMalloc((len+1)*sizeof(PetscInt),&buf_s);CHKERRQ(ierr); 4605 buf_si = buf_s; /* points to the beginning of k-th msg to be sent */ 4606 for (proc=0,k=0; proc<size; proc++){ 4607 if (!len_s[proc]) continue; 4608 /* form outgoing message for i-structure: 4609 buf_si[0]: nrows to be sent 4610 [1:nrows]: row index (global) 4611 [nrows+1:2*nrows+1]: i-structure index 4612 */ 4613 /*-------------------------------------------*/ 4614 nrows = len_si[proc]/2 - 1; 4615 buf_si_i = buf_si + nrows+1; 4616 buf_si[0] = nrows; 4617 buf_si_i[0] = 0; 4618 nrows = 0; 4619 for (i=owners[proc]; i<owners[proc+1]; i++){ 4620 anzi = ai[i+1] - ai[i]; 4621 if (anzi) { 4622 buf_si_i[nrows+1] = buf_si_i[nrows] + anzi; /* i-structure */ 4623 buf_si[nrows+1] = i-owners[proc]; /* local row index */ 4624 nrows++; 4625 } 4626 } 4627 ierr = MPI_Isend(buf_si,len_si[proc],MPIU_INT,proc,tagi,comm,si_waits+k);CHKERRQ(ierr); 4628 k++; 4629 buf_si += len_si[proc]; 4630 } 4631 4632 if (merge->nrecv) {ierr = MPI_Waitall(merge->nrecv,ri_waits,status);CHKERRQ(ierr);} 4633 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,si_waits,status);CHKERRQ(ierr);} 4634 4635 ierr = PetscInfo2(seqmat,"nsend: %D, nrecv: %D\n",merge->nsend,merge->nrecv);CHKERRQ(ierr); 4636 for (i=0; i<merge->nrecv; i++){ 4637 ierr = PetscInfo3(seqmat,"recv len_ri=%D, len_rj=%D from [%D]\n",len_ri[i],merge->len_r[i],merge->id_r[i]);CHKERRQ(ierr); 4638 } 4639 4640 ierr = PetscFree(len_si);CHKERRQ(ierr); 4641 ierr = PetscFree(len_ri);CHKERRQ(ierr); 4642 ierr = PetscFree(rj_waits);CHKERRQ(ierr); 4643 ierr = PetscFree2(si_waits,sj_waits);CHKERRQ(ierr); 4644 ierr = PetscFree(ri_waits);CHKERRQ(ierr); 4645 ierr = PetscFree(buf_s);CHKERRQ(ierr); 4646 ierr = PetscFree(status);CHKERRQ(ierr); 4647 4648 /* compute a local seq matrix in each processor */ 4649 /*----------------------------------------------*/ 4650 /* allocate bi array and free space for accumulating nonzero column info */ 4651 ierr = PetscMalloc((m+1)*sizeof(PetscInt),&bi);CHKERRQ(ierr); 4652 bi[0] = 0; 4653 4654 /* create and initialize a linked list */ 4655 nlnk = N+1; 4656 ierr = PetscLLCreate(N,N,nlnk,lnk,lnkbt);CHKERRQ(ierr); 4657 4658 /* initial FreeSpace size is 2*(num of local nnz(seqmat)) */ 4659 len = 0; 4660 len = ai[owners[rank+1]] - ai[owners[rank]]; 4661 ierr = PetscFreeSpaceGet((PetscInt)(2*len+1),&free_space);CHKERRQ(ierr); 4662 current_space = free_space; 4663 4664 /* determine symbolic info for each local row */ 4665 ierr = PetscMalloc3(merge->nrecv,PetscInt*,&buf_ri_k,merge->nrecv,PetscInt*,&nextrow,merge->nrecv,PetscInt*,&nextai);CHKERRQ(ierr); 4666 4667 for (k=0; k<merge->nrecv; k++){ 4668 buf_ri_k[k] = buf_ri[k]; /* beginning of k-th recved i-structure */ 4669 nrows = *buf_ri_k[k]; 4670 nextrow[k] = buf_ri_k[k] + 1; /* next row number of k-th recved i-structure */ 4671 nextai[k] = buf_ri_k[k] + (nrows + 1);/* poins to the next i-structure of k-th recved i-structure */ 4672 } 4673 4674 ierr = MatPreallocateInitialize(comm,m,n,dnz,onz);CHKERRQ(ierr); 4675 len = 0; 4676 for (i=0;i<m;i++) { 4677 bnzi = 0; 4678 /* add local non-zero cols of this proc's seqmat into lnk */ 4679 arow = owners[rank] + i; 4680 anzi = ai[arow+1] - ai[arow]; 4681 aj = a->j + ai[arow]; 4682 ierr = PetscLLAdd(anzi,aj,N,nlnk,lnk,lnkbt);CHKERRQ(ierr); 4683 bnzi += nlnk; 4684 /* add received col data into lnk */ 4685 for (k=0; k<merge->nrecv; k++){ /* k-th received message */ 4686 if (i == *nextrow[k]) { /* i-th row */ 4687 anzi = *(nextai[k]+1) - *nextai[k]; 4688 aj = buf_rj[k] + *nextai[k]; 4689 ierr = PetscLLAdd(anzi,aj,N,nlnk,lnk,lnkbt);CHKERRQ(ierr); 4690 bnzi += nlnk; 4691 nextrow[k]++; nextai[k]++; 4692 } 4693 } 4694 if (len < bnzi) len = bnzi; /* =max(bnzi) */ 4695 4696 /* if free space is not available, make more free space */ 4697 if (current_space->local_remaining<bnzi) { 4698 ierr = PetscFreeSpaceGet(bnzi+current_space->total_array_size,¤t_space);CHKERRQ(ierr); 4699 nspacedouble++; 4700 } 4701 /* copy data into free space, then initialize lnk */ 4702 ierr = PetscLLClean(N,N,bnzi,lnk,current_space->array,lnkbt);CHKERRQ(ierr); 4703 ierr = MatPreallocateSet(i+owners[rank],bnzi,current_space->array,dnz,onz);CHKERRQ(ierr); 4704 4705 current_space->array += bnzi; 4706 current_space->local_used += bnzi; 4707 current_space->local_remaining -= bnzi; 4708 4709 bi[i+1] = bi[i] + bnzi; 4710 } 4711 4712 ierr = PetscFree3(buf_ri_k,nextrow,nextai);CHKERRQ(ierr); 4713 4714 ierr = PetscMalloc((bi[m]+1)*sizeof(PetscInt),&bj);CHKERRQ(ierr); 4715 ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr); 4716 ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr); 4717 4718 /* create symbolic parallel matrix B_mpi */ 4719 /*---------------------------------------*/ 4720 ierr = MatCreate(comm,&B_mpi);CHKERRQ(ierr); 4721 if (n==PETSC_DECIDE) { 4722 ierr = MatSetSizes(B_mpi,m,n,PETSC_DETERMINE,N);CHKERRQ(ierr); 4723 } else { 4724 ierr = MatSetSizes(B_mpi,m,n,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 4725 } 4726 ierr = MatSetType(B_mpi,MATMPIAIJ);CHKERRQ(ierr); 4727 ierr = MatMPIAIJSetPreallocation(B_mpi,0,dnz,0,onz);CHKERRQ(ierr); 4728 ierr = MatPreallocateFinalize(dnz,onz);CHKERRQ(ierr); 4729 4730 /* B_mpi is not ready for use - assembly will be done by MatMerge_SeqsToMPINumeric() */ 4731 B_mpi->assembled = PETSC_FALSE; 4732 B_mpi->ops->destroy = MatDestroy_MPIAIJ_SeqsToMPI; 4733 merge->bi = bi; 4734 merge->bj = bj; 4735 merge->buf_ri = buf_ri; 4736 merge->buf_rj = buf_rj; 4737 merge->coi = PETSC_NULL; 4738 merge->coj = PETSC_NULL; 4739 merge->owners_co = PETSC_NULL; 4740 4741 /* attach the supporting struct to B_mpi for reuse */ 4742 ierr = PetscContainerCreate(PETSC_COMM_SELF,&container);CHKERRQ(ierr); 4743 ierr = PetscContainerSetPointer(container,merge);CHKERRQ(ierr); 4744 ierr = PetscObjectCompose((PetscObject)B_mpi,"MatMergeSeqsToMPI",(PetscObject)container);CHKERRQ(ierr); 4745 *mpimat = B_mpi; 4746 4747 ierr = PetscCommDestroy(&comm);CHKERRQ(ierr); 4748 ierr = PetscLogEventEnd(MAT_Seqstompisym,seqmat,0,0,0);CHKERRQ(ierr); 4749 PetscFunctionReturn(0); 4750 } 4751 4752 #undef __FUNCT__ 4753 #define __FUNCT__ "MatMerge_SeqsToMPI" 4754 PetscErrorCode MatMerge_SeqsToMPI(MPI_Comm comm,Mat seqmat,PetscInt m,PetscInt n,MatReuse scall,Mat *mpimat) 4755 { 4756 PetscErrorCode ierr; 4757 4758 PetscFunctionBegin; 4759 ierr = PetscLogEventBegin(MAT_Seqstompi,seqmat,0,0,0);CHKERRQ(ierr); 4760 if (scall == MAT_INITIAL_MATRIX){ 4761 ierr = MatMerge_SeqsToMPISymbolic(comm,seqmat,m,n,mpimat);CHKERRQ(ierr); 4762 } 4763 ierr = MatMerge_SeqsToMPINumeric(seqmat,*mpimat);CHKERRQ(ierr); 4764 ierr = PetscLogEventEnd(MAT_Seqstompi,seqmat,0,0,0);CHKERRQ(ierr); 4765 PetscFunctionReturn(0); 4766 } 4767 4768 #undef __FUNCT__ 4769 #define __FUNCT__ "MatMPIAIJGetLocalMat" 4770 /*@ 4771 MatMPIAIJGetLocalMat - Creates a SeqAIJ from a MPIAIJ matrix by taking all its local rows and putting them into a sequential vector with 4772 mlocal rows and n columns. Where mlocal is the row count obtained with MatGetLocalSize() and n is the global column count obtained 4773 with MatGetSize() 4774 4775 Not Collective 4776 4777 Input Parameters: 4778 + A - the matrix 4779 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 4780 4781 Output Parameter: 4782 . A_loc - the local sequential matrix generated 4783 4784 Level: developer 4785 4786 .seealso: MatGetOwnerShipRange(), MatMPIAIJGetLocalMatCondensed() 4787 4788 @*/ 4789 PetscErrorCode MatMPIAIJGetLocalMat(Mat A,MatReuse scall,Mat *A_loc) 4790 { 4791 PetscErrorCode ierr; 4792 Mat_MPIAIJ *mpimat=(Mat_MPIAIJ*)A->data; 4793 Mat_SeqAIJ *mat,*a=(Mat_SeqAIJ*)(mpimat->A)->data,*b=(Mat_SeqAIJ*)(mpimat->B)->data; 4794 PetscInt *ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j,*cmap=mpimat->garray; 4795 MatScalar *aa=a->a,*ba=b->a,*cam; 4796 PetscScalar *ca; 4797 PetscInt am=A->rmap->n,i,j,k,cstart=A->cmap->rstart; 4798 PetscInt *ci,*cj,col,ncols_d,ncols_o,jo; 4799 PetscBool match; 4800 4801 PetscFunctionBegin; 4802 ierr = PetscTypeCompare((PetscObject)A,MATMPIAIJ,&match);CHKERRQ(ierr); 4803 if (!match) SETERRQ(((PetscObject)A)->comm, PETSC_ERR_SUP,"Requires MPIAIJ matrix as input"); 4804 ierr = PetscLogEventBegin(MAT_Getlocalmat,A,0,0,0);CHKERRQ(ierr); 4805 if (scall == MAT_INITIAL_MATRIX){ 4806 ierr = PetscMalloc((1+am)*sizeof(PetscInt),&ci);CHKERRQ(ierr); 4807 ci[0] = 0; 4808 for (i=0; i<am; i++){ 4809 ci[i+1] = ci[i] + (ai[i+1] - ai[i]) + (bi[i+1] - bi[i]); 4810 } 4811 ierr = PetscMalloc((1+ci[am])*sizeof(PetscInt),&cj);CHKERRQ(ierr); 4812 ierr = PetscMalloc((1+ci[am])*sizeof(PetscScalar),&ca);CHKERRQ(ierr); 4813 k = 0; 4814 for (i=0; i<am; i++) { 4815 ncols_o = bi[i+1] - bi[i]; 4816 ncols_d = ai[i+1] - ai[i]; 4817 /* off-diagonal portion of A */ 4818 for (jo=0; jo<ncols_o; jo++) { 4819 col = cmap[*bj]; 4820 if (col >= cstart) break; 4821 cj[k] = col; bj++; 4822 ca[k++] = *ba++; 4823 } 4824 /* diagonal portion of A */ 4825 for (j=0; j<ncols_d; j++) { 4826 cj[k] = cstart + *aj++; 4827 ca[k++] = *aa++; 4828 } 4829 /* off-diagonal portion of A */ 4830 for (j=jo; j<ncols_o; j++) { 4831 cj[k] = cmap[*bj++]; 4832 ca[k++] = *ba++; 4833 } 4834 } 4835 /* put together the new matrix */ 4836 ierr = MatCreateSeqAIJWithArrays(PETSC_COMM_SELF,am,A->cmap->N,ci,cj,ca,A_loc);CHKERRQ(ierr); 4837 /* MatCreateSeqAIJWithArrays flags matrix so PETSc doesn't free the user's arrays. */ 4838 /* Since these are PETSc arrays, change flags to free them as necessary. */ 4839 mat = (Mat_SeqAIJ*)(*A_loc)->data; 4840 mat->free_a = PETSC_TRUE; 4841 mat->free_ij = PETSC_TRUE; 4842 mat->nonew = 0; 4843 } else if (scall == MAT_REUSE_MATRIX){ 4844 mat=(Mat_SeqAIJ*)(*A_loc)->data; 4845 ci = mat->i; cj = mat->j; cam = mat->a; 4846 for (i=0; i<am; i++) { 4847 /* off-diagonal portion of A */ 4848 ncols_o = bi[i+1] - bi[i]; 4849 for (jo=0; jo<ncols_o; jo++) { 4850 col = cmap[*bj]; 4851 if (col >= cstart) break; 4852 *cam++ = *ba++; bj++; 4853 } 4854 /* diagonal portion of A */ 4855 ncols_d = ai[i+1] - ai[i]; 4856 for (j=0; j<ncols_d; j++) *cam++ = *aa++; 4857 /* off-diagonal portion of A */ 4858 for (j=jo; j<ncols_o; j++) { 4859 *cam++ = *ba++; bj++; 4860 } 4861 } 4862 } else SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Invalid MatReuse %d",(int)scall); 4863 ierr = PetscLogEventEnd(MAT_Getlocalmat,A,0,0,0);CHKERRQ(ierr); 4864 PetscFunctionReturn(0); 4865 } 4866 4867 #undef __FUNCT__ 4868 #define __FUNCT__ "MatMPIAIJGetLocalMatCondensed" 4869 /*@C 4870 MatMPIAIJGetLocalMatCondensed - Creates a SeqAIJ matrix from an MPIAIJ matrix by taking all its local rows and NON-ZERO columns 4871 4872 Not Collective 4873 4874 Input Parameters: 4875 + A - the matrix 4876 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 4877 - row, col - index sets of rows and columns to extract (or PETSC_NULL) 4878 4879 Output Parameter: 4880 . A_loc - the local sequential matrix generated 4881 4882 Level: developer 4883 4884 .seealso: MatGetOwnershipRange(), MatMPIAIJGetLocalMat() 4885 4886 @*/ 4887 PetscErrorCode MatMPIAIJGetLocalMatCondensed(Mat A,MatReuse scall,IS *row,IS *col,Mat *A_loc) 4888 { 4889 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data; 4890 PetscErrorCode ierr; 4891 PetscInt i,start,end,ncols,nzA,nzB,*cmap,imark,*idx; 4892 IS isrowa,iscola; 4893 Mat *aloc; 4894 PetscBool match; 4895 4896 PetscFunctionBegin; 4897 ierr = PetscTypeCompare((PetscObject)A,MATMPIAIJ,&match);CHKERRQ(ierr); 4898 if (!match) SETERRQ(((PetscObject)A)->comm, PETSC_ERR_SUP,"Requires MPIAIJ matrix as input"); 4899 ierr = PetscLogEventBegin(MAT_Getlocalmatcondensed,A,0,0,0);CHKERRQ(ierr); 4900 if (!row){ 4901 start = A->rmap->rstart; end = A->rmap->rend; 4902 ierr = ISCreateStride(PETSC_COMM_SELF,end-start,start,1,&isrowa);CHKERRQ(ierr); 4903 } else { 4904 isrowa = *row; 4905 } 4906 if (!col){ 4907 start = A->cmap->rstart; 4908 cmap = a->garray; 4909 nzA = a->A->cmap->n; 4910 nzB = a->B->cmap->n; 4911 ierr = PetscMalloc((nzA+nzB)*sizeof(PetscInt), &idx);CHKERRQ(ierr); 4912 ncols = 0; 4913 for (i=0; i<nzB; i++) { 4914 if (cmap[i] < start) idx[ncols++] = cmap[i]; 4915 else break; 4916 } 4917 imark = i; 4918 for (i=0; i<nzA; i++) idx[ncols++] = start + i; 4919 for (i=imark; i<nzB; i++) idx[ncols++] = cmap[i]; 4920 ierr = ISCreateGeneral(PETSC_COMM_SELF,ncols,idx,PETSC_OWN_POINTER,&iscola);CHKERRQ(ierr); 4921 } else { 4922 iscola = *col; 4923 } 4924 if (scall != MAT_INITIAL_MATRIX){ 4925 ierr = PetscMalloc(sizeof(Mat),&aloc);CHKERRQ(ierr); 4926 aloc[0] = *A_loc; 4927 } 4928 ierr = MatGetSubMatrices(A,1,&isrowa,&iscola,scall,&aloc);CHKERRQ(ierr); 4929 *A_loc = aloc[0]; 4930 ierr = PetscFree(aloc);CHKERRQ(ierr); 4931 if (!row){ 4932 ierr = ISDestroy(&isrowa);CHKERRQ(ierr); 4933 } 4934 if (!col){ 4935 ierr = ISDestroy(&iscola);CHKERRQ(ierr); 4936 } 4937 ierr = PetscLogEventEnd(MAT_Getlocalmatcondensed,A,0,0,0);CHKERRQ(ierr); 4938 PetscFunctionReturn(0); 4939 } 4940 4941 #undef __FUNCT__ 4942 #define __FUNCT__ "MatGetBrowsOfAcols" 4943 /*@C 4944 MatGetBrowsOfAcols - Creates a SeqAIJ matrix by taking rows of B that equal to nonzero columns of local A 4945 4946 Collective on Mat 4947 4948 Input Parameters: 4949 + A,B - the matrices in mpiaij format 4950 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 4951 - rowb, colb - index sets of rows and columns of B to extract (or PETSC_NULL) 4952 4953 Output Parameter: 4954 + rowb, colb - index sets of rows and columns of B to extract 4955 . brstart - row index of B_seq from which next B->rmap->n rows are taken from B's local rows 4956 - B_seq - the sequential matrix generated 4957 4958 Level: developer 4959 4960 @*/ 4961 PetscErrorCode MatGetBrowsOfAcols(Mat A,Mat B,MatReuse scall,IS *rowb,IS *colb,PetscInt *brstart,Mat *B_seq) 4962 { 4963 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data; 4964 PetscErrorCode ierr; 4965 PetscInt *idx,i,start,ncols,nzA,nzB,*cmap,imark; 4966 IS isrowb,iscolb; 4967 Mat *bseq; 4968 4969 PetscFunctionBegin; 4970 if (A->cmap->rstart != B->rmap->rstart || A->cmap->rend != B->rmap->rend){ 4971 SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Matrix local dimensions are incompatible, (%D, %D) != (%D,%D)",A->cmap->rstart,A->cmap->rend,B->rmap->rstart,B->rmap->rend); 4972 } 4973 ierr = PetscLogEventBegin(MAT_GetBrowsOfAcols,A,B,0,0);CHKERRQ(ierr); 4974 4975 if (scall == MAT_INITIAL_MATRIX){ 4976 start = A->cmap->rstart; 4977 cmap = a->garray; 4978 nzA = a->A->cmap->n; 4979 nzB = a->B->cmap->n; 4980 ierr = PetscMalloc((nzA+nzB)*sizeof(PetscInt), &idx);CHKERRQ(ierr); 4981 ncols = 0; 4982 for (i=0; i<nzB; i++) { /* row < local row index */ 4983 if (cmap[i] < start) idx[ncols++] = cmap[i]; 4984 else break; 4985 } 4986 imark = i; 4987 for (i=0; i<nzA; i++) idx[ncols++] = start + i; /* local rows */ 4988 for (i=imark; i<nzB; i++) idx[ncols++] = cmap[i]; /* row > local row index */ 4989 ierr = ISCreateGeneral(PETSC_COMM_SELF,ncols,idx,PETSC_OWN_POINTER,&isrowb);CHKERRQ(ierr); 4990 *brstart = imark; 4991 ierr = ISCreateStride(PETSC_COMM_SELF,B->cmap->N,0,1,&iscolb);CHKERRQ(ierr); 4992 } else { 4993 if (!rowb || !colb) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"IS rowb and colb must be provided for MAT_REUSE_MATRIX"); 4994 isrowb = *rowb; iscolb = *colb; 4995 ierr = PetscMalloc(sizeof(Mat),&bseq);CHKERRQ(ierr); 4996 bseq[0] = *B_seq; 4997 } 4998 ierr = MatGetSubMatrices(B,1,&isrowb,&iscolb,scall,&bseq);CHKERRQ(ierr); 4999 *B_seq = bseq[0]; 5000 ierr = PetscFree(bseq);CHKERRQ(ierr); 5001 if (!rowb){ 5002 ierr = ISDestroy(&isrowb);CHKERRQ(ierr); 5003 } else { 5004 *rowb = isrowb; 5005 } 5006 if (!colb){ 5007 ierr = ISDestroy(&iscolb);CHKERRQ(ierr); 5008 } else { 5009 *colb = iscolb; 5010 } 5011 ierr = PetscLogEventEnd(MAT_GetBrowsOfAcols,A,B,0,0);CHKERRQ(ierr); 5012 PetscFunctionReturn(0); 5013 } 5014 5015 #undef __FUNCT__ 5016 #define __FUNCT__ "MatGetBrowsOfAoCols" 5017 /*@C 5018 MatGetBrowsOfAoCols - Creates a SeqAIJ matrix by taking rows of B that equal to nonzero columns 5019 of the OFF-DIAGONAL portion of local A 5020 5021 Collective on Mat 5022 5023 Input Parameters: 5024 + A,B - the matrices in mpiaij format 5025 . scall - either MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX 5026 . startsj - starting point in B's sending and receiving j-arrays, saved for MAT_REUSE (or PETSC_NULL) 5027 . startsj_r - similar to startsj for receives 5028 - bufa_ptr - array for sending matrix values, saved for MAT_REUSE (or PETSC_NULL) 5029 5030 Output Parameter: 5031 + B_oth - the sequential matrix generated 5032 5033 Level: developer 5034 5035 @*/ 5036 PetscErrorCode MatGetBrowsOfAoCols(Mat A,Mat B,MatReuse scall,PetscInt **startsj,PetscInt **startsj_r,MatScalar **bufa_ptr,Mat *B_oth) 5037 { 5038 VecScatter_MPI_General *gen_to,*gen_from; 5039 PetscErrorCode ierr; 5040 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data; 5041 Mat_SeqAIJ *b_oth; 5042 VecScatter ctx=a->Mvctx; 5043 MPI_Comm comm=((PetscObject)ctx)->comm; 5044 PetscMPIInt *rprocs,*sprocs,tag=((PetscObject)ctx)->tag,rank; 5045 PetscInt *rowlen,*bufj,*bufJ,ncols,aBn=a->B->cmap->n,row,*b_othi,*b_othj; 5046 PetscScalar *rvalues,*svalues; 5047 MatScalar *b_otha,*bufa,*bufA; 5048 PetscInt i,j,k,l,ll,nrecvs,nsends,nrows,*srow,*rstarts,*rstartsj = 0,*sstarts,*sstartsj,len; 5049 MPI_Request *rwaits = PETSC_NULL,*swaits = PETSC_NULL; 5050 MPI_Status *sstatus,rstatus; 5051 PetscMPIInt jj; 5052 PetscInt *cols,sbs,rbs; 5053 PetscScalar *vals; 5054 5055 PetscFunctionBegin; 5056 if (A->cmap->rstart != B->rmap->rstart || A->cmap->rend != B->rmap->rend){ 5057 SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Matrix local dimensions are incompatible, (%d, %d) != (%d,%d)",A->cmap->rstart,A->cmap->rend,B->rmap->rstart,B->rmap->rend); 5058 } 5059 ierr = PetscLogEventBegin(MAT_GetBrowsOfAocols,A,B,0,0);CHKERRQ(ierr); 5060 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 5061 5062 gen_to = (VecScatter_MPI_General*)ctx->todata; 5063 gen_from = (VecScatter_MPI_General*)ctx->fromdata; 5064 rvalues = gen_from->values; /* holds the length of receiving row */ 5065 svalues = gen_to->values; /* holds the length of sending row */ 5066 nrecvs = gen_from->n; 5067 nsends = gen_to->n; 5068 5069 ierr = PetscMalloc2(nrecvs,MPI_Request,&rwaits,nsends,MPI_Request,&swaits);CHKERRQ(ierr); 5070 srow = gen_to->indices; /* local row index to be sent */ 5071 sstarts = gen_to->starts; 5072 sprocs = gen_to->procs; 5073 sstatus = gen_to->sstatus; 5074 sbs = gen_to->bs; 5075 rstarts = gen_from->starts; 5076 rprocs = gen_from->procs; 5077 rbs = gen_from->bs; 5078 5079 if (!startsj || !bufa_ptr) scall = MAT_INITIAL_MATRIX; 5080 if (scall == MAT_INITIAL_MATRIX){ 5081 /* i-array */ 5082 /*---------*/ 5083 /* post receives */ 5084 for (i=0; i<nrecvs; i++){ 5085 rowlen = (PetscInt*)rvalues + rstarts[i]*rbs; 5086 nrows = (rstarts[i+1]-rstarts[i])*rbs; /* num of indices to be received */ 5087 ierr = MPI_Irecv(rowlen,nrows,MPIU_INT,rprocs[i],tag,comm,rwaits+i);CHKERRQ(ierr); 5088 } 5089 5090 /* pack the outgoing message */ 5091 ierr = PetscMalloc2(nsends+1,PetscInt,&sstartsj,nrecvs+1,PetscInt,&rstartsj);CHKERRQ(ierr); 5092 sstartsj[0] = 0; rstartsj[0] = 0; 5093 len = 0; /* total length of j or a array to be sent */ 5094 k = 0; 5095 for (i=0; i<nsends; i++){ 5096 rowlen = (PetscInt*)svalues + sstarts[i]*sbs; 5097 nrows = sstarts[i+1]-sstarts[i]; /* num of block rows */ 5098 for (j=0; j<nrows; j++) { 5099 row = srow[k] + B->rmap->range[rank]; /* global row idx */ 5100 for (l=0; l<sbs; l++){ 5101 ierr = MatGetRow_MPIAIJ(B,row+l,&ncols,PETSC_NULL,PETSC_NULL);CHKERRQ(ierr); /* rowlength */ 5102 rowlen[j*sbs+l] = ncols; 5103 len += ncols; 5104 ierr = MatRestoreRow_MPIAIJ(B,row+l,&ncols,PETSC_NULL,PETSC_NULL);CHKERRQ(ierr); 5105 } 5106 k++; 5107 } 5108 ierr = MPI_Isend(rowlen,nrows*sbs,MPIU_INT,sprocs[i],tag,comm,swaits+i);CHKERRQ(ierr); 5109 sstartsj[i+1] = len; /* starting point of (i+1)-th outgoing msg in bufj and bufa */ 5110 } 5111 /* recvs and sends of i-array are completed */ 5112 i = nrecvs; 5113 while (i--) { 5114 ierr = MPI_Waitany(nrecvs,rwaits,&jj,&rstatus);CHKERRQ(ierr); 5115 } 5116 if (nsends) {ierr = MPI_Waitall(nsends,swaits,sstatus);CHKERRQ(ierr);} 5117 5118 /* allocate buffers for sending j and a arrays */ 5119 ierr = PetscMalloc((len+1)*sizeof(PetscInt),&bufj);CHKERRQ(ierr); 5120 ierr = PetscMalloc((len+1)*sizeof(PetscScalar),&bufa);CHKERRQ(ierr); 5121 5122 /* create i-array of B_oth */ 5123 ierr = PetscMalloc((aBn+2)*sizeof(PetscInt),&b_othi);CHKERRQ(ierr); 5124 b_othi[0] = 0; 5125 len = 0; /* total length of j or a array to be received */ 5126 k = 0; 5127 for (i=0; i<nrecvs; i++){ 5128 rowlen = (PetscInt*)rvalues + rstarts[i]*rbs; 5129 nrows = rbs*(rstarts[i+1]-rstarts[i]); /* num of rows to be recieved */ 5130 for (j=0; j<nrows; j++) { 5131 b_othi[k+1] = b_othi[k] + rowlen[j]; 5132 len += rowlen[j]; k++; 5133 } 5134 rstartsj[i+1] = len; /* starting point of (i+1)-th incoming msg in bufj and bufa */ 5135 } 5136 5137 /* allocate space for j and a arrrays of B_oth */ 5138 ierr = PetscMalloc((b_othi[aBn]+1)*sizeof(PetscInt),&b_othj);CHKERRQ(ierr); 5139 ierr = PetscMalloc((b_othi[aBn]+1)*sizeof(MatScalar),&b_otha);CHKERRQ(ierr); 5140 5141 /* j-array */ 5142 /*---------*/ 5143 /* post receives of j-array */ 5144 for (i=0; i<nrecvs; i++){ 5145 nrows = rstartsj[i+1]-rstartsj[i]; /* length of the msg received */ 5146 ierr = MPI_Irecv(b_othj+rstartsj[i],nrows,MPIU_INT,rprocs[i],tag,comm,rwaits+i);CHKERRQ(ierr); 5147 } 5148 5149 /* pack the outgoing message j-array */ 5150 k = 0; 5151 for (i=0; i<nsends; i++){ 5152 nrows = sstarts[i+1]-sstarts[i]; /* num of block rows */ 5153 bufJ = bufj+sstartsj[i]; 5154 for (j=0; j<nrows; j++) { 5155 row = srow[k++] + B->rmap->range[rank]; /* global row idx */ 5156 for (ll=0; ll<sbs; ll++){ 5157 ierr = MatGetRow_MPIAIJ(B,row+ll,&ncols,&cols,PETSC_NULL);CHKERRQ(ierr); 5158 for (l=0; l<ncols; l++){ 5159 *bufJ++ = cols[l]; 5160 } 5161 ierr = MatRestoreRow_MPIAIJ(B,row+ll,&ncols,&cols,PETSC_NULL);CHKERRQ(ierr); 5162 } 5163 } 5164 ierr = MPI_Isend(bufj+sstartsj[i],sstartsj[i+1]-sstartsj[i],MPIU_INT,sprocs[i],tag,comm,swaits+i);CHKERRQ(ierr); 5165 } 5166 5167 /* recvs and sends of j-array are completed */ 5168 i = nrecvs; 5169 while (i--) { 5170 ierr = MPI_Waitany(nrecvs,rwaits,&jj,&rstatus);CHKERRQ(ierr); 5171 } 5172 if (nsends) {ierr = MPI_Waitall(nsends,swaits,sstatus);CHKERRQ(ierr);} 5173 } else if (scall == MAT_REUSE_MATRIX){ 5174 sstartsj = *startsj; 5175 rstartsj = *startsj_r; 5176 bufa = *bufa_ptr; 5177 b_oth = (Mat_SeqAIJ*)(*B_oth)->data; 5178 b_otha = b_oth->a; 5179 } else { 5180 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE, "Matrix P does not posses an object container"); 5181 } 5182 5183 /* a-array */ 5184 /*---------*/ 5185 /* post receives of a-array */ 5186 for (i=0; i<nrecvs; i++){ 5187 nrows = rstartsj[i+1]-rstartsj[i]; /* length of the msg received */ 5188 ierr = MPI_Irecv(b_otha+rstartsj[i],nrows,MPIU_SCALAR,rprocs[i],tag,comm,rwaits+i);CHKERRQ(ierr); 5189 } 5190 5191 /* pack the outgoing message a-array */ 5192 k = 0; 5193 for (i=0; i<nsends; i++){ 5194 nrows = sstarts[i+1]-sstarts[i]; /* num of block rows */ 5195 bufA = bufa+sstartsj[i]; 5196 for (j=0; j<nrows; j++) { 5197 row = srow[k++] + B->rmap->range[rank]; /* global row idx */ 5198 for (ll=0; ll<sbs; ll++){ 5199 ierr = MatGetRow_MPIAIJ(B,row+ll,&ncols,PETSC_NULL,&vals);CHKERRQ(ierr); 5200 for (l=0; l<ncols; l++){ 5201 *bufA++ = vals[l]; 5202 } 5203 ierr = MatRestoreRow_MPIAIJ(B,row+ll,&ncols,PETSC_NULL,&vals);CHKERRQ(ierr); 5204 } 5205 } 5206 ierr = MPI_Isend(bufa+sstartsj[i],sstartsj[i+1]-sstartsj[i],MPIU_SCALAR,sprocs[i],tag,comm,swaits+i);CHKERRQ(ierr); 5207 } 5208 /* recvs and sends of a-array are completed */ 5209 i = nrecvs; 5210 while (i--) { 5211 ierr = MPI_Waitany(nrecvs,rwaits,&jj,&rstatus);CHKERRQ(ierr); 5212 } 5213 if (nsends) {ierr = MPI_Waitall(nsends,swaits,sstatus);CHKERRQ(ierr);} 5214 ierr = PetscFree2(rwaits,swaits);CHKERRQ(ierr); 5215 5216 if (scall == MAT_INITIAL_MATRIX){ 5217 /* put together the new matrix */ 5218 ierr = MatCreateSeqAIJWithArrays(PETSC_COMM_SELF,aBn,B->cmap->N,b_othi,b_othj,b_otha,B_oth);CHKERRQ(ierr); 5219 5220 /* MatCreateSeqAIJWithArrays flags matrix so PETSc doesn't free the user's arrays. */ 5221 /* Since these are PETSc arrays, change flags to free them as necessary. */ 5222 b_oth = (Mat_SeqAIJ *)(*B_oth)->data; 5223 b_oth->free_a = PETSC_TRUE; 5224 b_oth->free_ij = PETSC_TRUE; 5225 b_oth->nonew = 0; 5226 5227 ierr = PetscFree(bufj);CHKERRQ(ierr); 5228 if (!startsj || !bufa_ptr){ 5229 ierr = PetscFree2(sstartsj,rstartsj);CHKERRQ(ierr); 5230 ierr = PetscFree(bufa_ptr);CHKERRQ(ierr); 5231 } else { 5232 *startsj = sstartsj; 5233 *startsj_r = rstartsj; 5234 *bufa_ptr = bufa; 5235 } 5236 } 5237 ierr = PetscLogEventEnd(MAT_GetBrowsOfAocols,A,B,0,0);CHKERRQ(ierr); 5238 PetscFunctionReturn(0); 5239 } 5240 5241 #undef __FUNCT__ 5242 #define __FUNCT__ "MatGetCommunicationStructs" 5243 /*@C 5244 MatGetCommunicationStructs - Provides access to the communication structures used in matrix-vector multiplication. 5245 5246 Not Collective 5247 5248 Input Parameters: 5249 . A - The matrix in mpiaij format 5250 5251 Output Parameter: 5252 + lvec - The local vector holding off-process values from the argument to a matrix-vector product 5253 . colmap - A map from global column index to local index into lvec 5254 - multScatter - A scatter from the argument of a matrix-vector product to lvec 5255 5256 Level: developer 5257 5258 @*/ 5259 #if defined (PETSC_USE_CTABLE) 5260 PetscErrorCode MatGetCommunicationStructs(Mat A, Vec *lvec, PetscTable *colmap, VecScatter *multScatter) 5261 #else 5262 PetscErrorCode MatGetCommunicationStructs(Mat A, Vec *lvec, PetscInt *colmap[], VecScatter *multScatter) 5263 #endif 5264 { 5265 Mat_MPIAIJ *a; 5266 5267 PetscFunctionBegin; 5268 PetscValidHeaderSpecific(A, MAT_CLASSID, 1); 5269 PetscValidPointer(lvec, 2); 5270 PetscValidPointer(colmap, 3); 5271 PetscValidPointer(multScatter, 4); 5272 a = (Mat_MPIAIJ *) A->data; 5273 if (lvec) *lvec = a->lvec; 5274 if (colmap) *colmap = a->colmap; 5275 if (multScatter) *multScatter = a->Mvctx; 5276 PetscFunctionReturn(0); 5277 } 5278 5279 EXTERN_C_BEGIN 5280 extern PetscErrorCode MatConvert_MPIAIJ_MPIAIJCRL(Mat,const MatType,MatReuse,Mat*); 5281 extern PetscErrorCode MatConvert_MPIAIJ_MPIAIJPERM(Mat,const MatType,MatReuse,Mat*); 5282 extern PetscErrorCode MatConvert_MPIAIJ_MPISBAIJ(Mat,const MatType,MatReuse,Mat*); 5283 EXTERN_C_END 5284 5285 #undef __FUNCT__ 5286 #define __FUNCT__ "MatMatMultNumeric_MPIDense_MPIAIJ" 5287 /* 5288 Computes (B'*A')' since computing B*A directly is untenable 5289 5290 n p p 5291 ( ) ( ) ( ) 5292 m ( A ) * n ( B ) = m ( C ) 5293 ( ) ( ) ( ) 5294 5295 */ 5296 PetscErrorCode MatMatMultNumeric_MPIDense_MPIAIJ(Mat A,Mat B,Mat C) 5297 { 5298 PetscErrorCode ierr; 5299 Mat At,Bt,Ct; 5300 5301 PetscFunctionBegin; 5302 ierr = MatTranspose(A,MAT_INITIAL_MATRIX,&At);CHKERRQ(ierr); 5303 ierr = MatTranspose(B,MAT_INITIAL_MATRIX,&Bt);CHKERRQ(ierr); 5304 ierr = MatMatMult(Bt,At,MAT_INITIAL_MATRIX,1.0,&Ct);CHKERRQ(ierr); 5305 ierr = MatDestroy(&At);CHKERRQ(ierr); 5306 ierr = MatDestroy(&Bt);CHKERRQ(ierr); 5307 ierr = MatTranspose(Ct,MAT_REUSE_MATRIX,&C);CHKERRQ(ierr); 5308 ierr = MatDestroy(&Ct);CHKERRQ(ierr); 5309 PetscFunctionReturn(0); 5310 } 5311 5312 #undef __FUNCT__ 5313 #define __FUNCT__ "MatMatMultSymbolic_MPIDense_MPIAIJ" 5314 PetscErrorCode MatMatMultSymbolic_MPIDense_MPIAIJ(Mat A,Mat B,PetscReal fill,Mat *C) 5315 { 5316 PetscErrorCode ierr; 5317 PetscInt m=A->rmap->n,n=B->cmap->n; 5318 Mat Cmat; 5319 5320 PetscFunctionBegin; 5321 if (A->cmap->n != B->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"A->cmap->n %d != B->rmap->n %d\n",A->cmap->n,B->rmap->n); 5322 ierr = MatCreate(((PetscObject)A)->comm,&Cmat);CHKERRQ(ierr); 5323 ierr = MatSetSizes(Cmat,m,n,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 5324 ierr = MatSetType(Cmat,MATMPIDENSE);CHKERRQ(ierr); 5325 ierr = MatMPIDenseSetPreallocation(Cmat,PETSC_NULL);CHKERRQ(ierr); 5326 ierr = MatAssemblyBegin(Cmat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5327 ierr = MatAssemblyEnd(Cmat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5328 *C = Cmat; 5329 PetscFunctionReturn(0); 5330 } 5331 5332 /* ----------------------------------------------------------------*/ 5333 #undef __FUNCT__ 5334 #define __FUNCT__ "MatMatMult_MPIDense_MPIAIJ" 5335 PetscErrorCode MatMatMult_MPIDense_MPIAIJ(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 5336 { 5337 PetscErrorCode ierr; 5338 5339 PetscFunctionBegin; 5340 if (scall == MAT_INITIAL_MATRIX){ 5341 ierr = MatMatMultSymbolic_MPIDense_MPIAIJ(A,B,fill,C);CHKERRQ(ierr); 5342 } 5343 ierr = MatMatMultNumeric_MPIDense_MPIAIJ(A,B,*C);CHKERRQ(ierr); 5344 PetscFunctionReturn(0); 5345 } 5346 5347 EXTERN_C_BEGIN 5348 #if defined(PETSC_HAVE_MUMPS) 5349 extern PetscErrorCode MatGetFactor_aij_mumps(Mat,MatFactorType,Mat*); 5350 #endif 5351 #if defined(PETSC_HAVE_PASTIX) 5352 extern PetscErrorCode MatGetFactor_mpiaij_pastix(Mat,MatFactorType,Mat*); 5353 #endif 5354 #if defined(PETSC_HAVE_SUPERLU_DIST) 5355 extern PetscErrorCode MatGetFactor_mpiaij_superlu_dist(Mat,MatFactorType,Mat*); 5356 #endif 5357 #if defined(PETSC_HAVE_SPOOLES) 5358 extern PetscErrorCode MatGetFactor_mpiaij_spooles(Mat,MatFactorType,Mat*); 5359 #endif 5360 EXTERN_C_END 5361 5362 /*MC 5363 MATMPIAIJ - MATMPIAIJ = "mpiaij" - A matrix type to be used for parallel sparse matrices. 5364 5365 Options Database Keys: 5366 . -mat_type mpiaij - sets the matrix type to "mpiaij" during a call to MatSetFromOptions() 5367 5368 Level: beginner 5369 5370 .seealso: MatCreateMPIAIJ() 5371 M*/ 5372 5373 EXTERN_C_BEGIN 5374 #undef __FUNCT__ 5375 #define __FUNCT__ "MatCreate_MPIAIJ" 5376 PetscErrorCode MatCreate_MPIAIJ(Mat B) 5377 { 5378 Mat_MPIAIJ *b; 5379 PetscErrorCode ierr; 5380 PetscMPIInt size; 5381 5382 PetscFunctionBegin; 5383 ierr = MPI_Comm_size(((PetscObject)B)->comm,&size);CHKERRQ(ierr); 5384 5385 ierr = PetscNewLog(B,Mat_MPIAIJ,&b);CHKERRQ(ierr); 5386 B->data = (void*)b; 5387 ierr = PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 5388 B->rmap->bs = 1; 5389 B->assembled = PETSC_FALSE; 5390 5391 B->insertmode = NOT_SET_VALUES; 5392 b->size = size; 5393 ierr = MPI_Comm_rank(((PetscObject)B)->comm,&b->rank);CHKERRQ(ierr); 5394 5395 /* build cache for off array entries formed */ 5396 ierr = MatStashCreate_Private(((PetscObject)B)->comm,1,&B->stash);CHKERRQ(ierr); 5397 b->donotstash = PETSC_FALSE; 5398 b->colmap = 0; 5399 b->garray = 0; 5400 b->roworiented = PETSC_TRUE; 5401 5402 /* stuff used for matrix vector multiply */ 5403 b->lvec = PETSC_NULL; 5404 b->Mvctx = PETSC_NULL; 5405 5406 /* stuff for MatGetRow() */ 5407 b->rowindices = 0; 5408 b->rowvalues = 0; 5409 b->getrowactive = PETSC_FALSE; 5410 5411 #if defined(PETSC_HAVE_SPOOLES) 5412 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetFactor_spooles_C", 5413 "MatGetFactor_mpiaij_spooles", 5414 MatGetFactor_mpiaij_spooles);CHKERRQ(ierr); 5415 #endif 5416 #if defined(PETSC_HAVE_MUMPS) 5417 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetFactor_mumps_C", 5418 "MatGetFactor_aij_mumps", 5419 MatGetFactor_aij_mumps);CHKERRQ(ierr); 5420 #endif 5421 #if defined(PETSC_HAVE_PASTIX) 5422 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetFactor_pastix_C", 5423 "MatGetFactor_mpiaij_pastix", 5424 MatGetFactor_mpiaij_pastix);CHKERRQ(ierr); 5425 #endif 5426 #if defined(PETSC_HAVE_SUPERLU_DIST) 5427 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetFactor_superlu_dist_C", 5428 "MatGetFactor_mpiaij_superlu_dist", 5429 MatGetFactor_mpiaij_superlu_dist);CHKERRQ(ierr); 5430 #endif 5431 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatStoreValues_C", 5432 "MatStoreValues_MPIAIJ", 5433 MatStoreValues_MPIAIJ);CHKERRQ(ierr); 5434 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatRetrieveValues_C", 5435 "MatRetrieveValues_MPIAIJ", 5436 MatRetrieveValues_MPIAIJ);CHKERRQ(ierr); 5437 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetDiagonalBlock_C", 5438 "MatGetDiagonalBlock_MPIAIJ", 5439 MatGetDiagonalBlock_MPIAIJ);CHKERRQ(ierr); 5440 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatIsTranspose_C", 5441 "MatIsTranspose_MPIAIJ", 5442 MatIsTranspose_MPIAIJ);CHKERRQ(ierr); 5443 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMPIAIJSetPreallocation_C", 5444 "MatMPIAIJSetPreallocation_MPIAIJ", 5445 MatMPIAIJSetPreallocation_MPIAIJ);CHKERRQ(ierr); 5446 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMPIAIJSetPreallocationCSR_C", 5447 "MatMPIAIJSetPreallocationCSR_MPIAIJ", 5448 MatMPIAIJSetPreallocationCSR_MPIAIJ);CHKERRQ(ierr); 5449 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatDiagonalScaleLocal_C", 5450 "MatDiagonalScaleLocal_MPIAIJ", 5451 MatDiagonalScaleLocal_MPIAIJ);CHKERRQ(ierr); 5452 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatConvert_mpiaij_mpiaijperm_C", 5453 "MatConvert_MPIAIJ_MPIAIJPERM", 5454 MatConvert_MPIAIJ_MPIAIJPERM);CHKERRQ(ierr); 5455 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatConvert_mpiaij_mpiaijcrl_C", 5456 "MatConvert_MPIAIJ_MPIAIJCRL", 5457 MatConvert_MPIAIJ_MPIAIJCRL);CHKERRQ(ierr); 5458 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatConvert_mpiaij_mpisbaij_C", 5459 "MatConvert_MPIAIJ_MPISBAIJ", 5460 MatConvert_MPIAIJ_MPISBAIJ);CHKERRQ(ierr); 5461 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMatMult_mpidense_mpiaij_C", 5462 "MatMatMult_MPIDense_MPIAIJ", 5463 MatMatMult_MPIDense_MPIAIJ);CHKERRQ(ierr); 5464 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMatMultSymbolic_mpidense_mpiaij_C", 5465 "MatMatMultSymbolic_MPIDense_MPIAIJ", 5466 MatMatMultSymbolic_MPIDense_MPIAIJ);CHKERRQ(ierr); 5467 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMatMultNumeric_mpidense_mpiaij_C", 5468 "MatMatMultNumeric_MPIDense_MPIAIJ", 5469 MatMatMultNumeric_MPIDense_MPIAIJ);CHKERRQ(ierr); 5470 ierr = PetscObjectChangeTypeName((PetscObject)B,MATMPIAIJ);CHKERRQ(ierr); 5471 PetscFunctionReturn(0); 5472 } 5473 EXTERN_C_END 5474 5475 #undef __FUNCT__ 5476 #define __FUNCT__ "MatCreateMPIAIJWithSplitArrays" 5477 /*@ 5478 MatCreateMPIAIJWithSplitArrays - creates a MPI AIJ matrix using arrays that contain the "diagonal" 5479 and "off-diagonal" part of the matrix in CSR format. 5480 5481 Collective on MPI_Comm 5482 5483 Input Parameters: 5484 + comm - MPI communicator 5485 . m - number of local rows (Cannot be PETSC_DECIDE) 5486 . n - This value should be the same as the local size used in creating the 5487 x vector for the matrix-vector product y = Ax. (or PETSC_DECIDE to have 5488 calculated if N is given) For square matrices n is almost always m. 5489 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 5490 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 5491 . i - row indices for "diagonal" portion of matrix 5492 . j - column indices 5493 . a - matrix values 5494 . oi - row indices for "off-diagonal" portion of matrix 5495 . oj - column indices 5496 - oa - matrix values 5497 5498 Output Parameter: 5499 . mat - the matrix 5500 5501 Level: advanced 5502 5503 Notes: 5504 The i, j, and a arrays ARE NOT copied by this routine into the internal format used by PETSc. The user 5505 must free the arrays once the matrix has been destroyed and not before. 5506 5507 The i and j indices are 0 based 5508 5509 See MatCreateMPIAIJ() for the definition of "diagonal" and "off-diagonal" portion of the matrix 5510 5511 This sets local rows and cannot be used to set off-processor values. 5512 5513 You cannot later use MatSetValues() to change values in this matrix. 5514 5515 .keywords: matrix, aij, compressed row, sparse, parallel 5516 5517 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatMPIAIJSetPreallocation(), MatMPIAIJSetPreallocationCSR(), 5518 MPIAIJ, MatCreateMPIAIJ(), MatCreateMPIAIJWithArrays() 5519 @*/ 5520 PetscErrorCode MatCreateMPIAIJWithSplitArrays(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt M,PetscInt N,PetscInt i[],PetscInt j[],PetscScalar a[], 5521 PetscInt oi[], PetscInt oj[],PetscScalar oa[],Mat *mat) 5522 { 5523 PetscErrorCode ierr; 5524 Mat_MPIAIJ *maij; 5525 5526 PetscFunctionBegin; 5527 if (m < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"local number of rows (m) cannot be PETSC_DECIDE, or negative"); 5528 if (i[0]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"i (row indices) must start with 0"); 5529 if (oi[0]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"oi (row indices) must start with 0"); 5530 ierr = MatCreate(comm,mat);CHKERRQ(ierr); 5531 ierr = MatSetSizes(*mat,m,n,M,N);CHKERRQ(ierr); 5532 ierr = MatSetType(*mat,MATMPIAIJ);CHKERRQ(ierr); 5533 maij = (Mat_MPIAIJ*) (*mat)->data; 5534 maij->donotstash = PETSC_TRUE; 5535 (*mat)->preallocated = PETSC_TRUE; 5536 5537 ierr = PetscLayoutSetBlockSize((*mat)->rmap,1);CHKERRQ(ierr); 5538 ierr = PetscLayoutSetBlockSize((*mat)->cmap,1);CHKERRQ(ierr); 5539 ierr = PetscLayoutSetUp((*mat)->rmap);CHKERRQ(ierr); 5540 ierr = PetscLayoutSetUp((*mat)->cmap);CHKERRQ(ierr); 5541 5542 ierr = MatCreateSeqAIJWithArrays(PETSC_COMM_SELF,m,n,i,j,a,&maij->A);CHKERRQ(ierr); 5543 ierr = MatCreateSeqAIJWithArrays(PETSC_COMM_SELF,m,(*mat)->cmap->N,oi,oj,oa,&maij->B);CHKERRQ(ierr); 5544 5545 ierr = MatAssemblyBegin(maij->A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5546 ierr = MatAssemblyEnd(maij->A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5547 ierr = MatAssemblyBegin(maij->B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5548 ierr = MatAssemblyEnd(maij->B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5549 5550 ierr = MatAssemblyBegin(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5551 ierr = MatAssemblyEnd(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 5552 PetscFunctionReturn(0); 5553 } 5554 5555 /* 5556 Special version for direct calls from Fortran 5557 */ 5558 #include <private/fortranimpl.h> 5559 5560 #if defined(PETSC_HAVE_FORTRAN_CAPS) 5561 #define matsetvaluesmpiaij_ MATSETVALUESMPIAIJ 5562 #elif !defined(PETSC_HAVE_FORTRAN_UNDERSCORE) 5563 #define matsetvaluesmpiaij_ matsetvaluesmpiaij 5564 #endif 5565 5566 /* Change these macros so can be used in void function */ 5567 #undef CHKERRQ 5568 #define CHKERRQ(ierr) CHKERRABORT(PETSC_COMM_WORLD,ierr) 5569 #undef SETERRQ2 5570 #define SETERRQ2(comm,ierr,b,c,d) CHKERRABORT(comm,ierr) 5571 #undef SETERRQ 5572 #define SETERRQ(c,ierr,b) CHKERRABORT(c,ierr) 5573 5574 EXTERN_C_BEGIN 5575 #undef __FUNCT__ 5576 #define __FUNCT__ "matsetvaluesmpiaij_" 5577 void PETSC_STDCALL matsetvaluesmpiaij_(Mat *mmat,PetscInt *mm,const PetscInt im[],PetscInt *mn,const PetscInt in[],const PetscScalar v[],InsertMode *maddv,PetscErrorCode *_ierr) 5578 { 5579 Mat mat = *mmat; 5580 PetscInt m = *mm, n = *mn; 5581 InsertMode addv = *maddv; 5582 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 5583 PetscScalar value; 5584 PetscErrorCode ierr; 5585 5586 ierr = MatPreallocated(mat);CHKERRQ(ierr); 5587 if (mat->insertmode == NOT_SET_VALUES) { 5588 mat->insertmode = addv; 5589 } 5590 #if defined(PETSC_USE_DEBUG) 5591 else if (mat->insertmode != addv) { 5592 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 5593 } 5594 #endif 5595 { 5596 PetscInt i,j,rstart = mat->rmap->rstart,rend = mat->rmap->rend; 5597 PetscInt cstart = mat->cmap->rstart,cend = mat->cmap->rend,row,col; 5598 PetscBool roworiented = aij->roworiented; 5599 5600 /* Some Variables required in the macro */ 5601 Mat A = aij->A; 5602 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 5603 PetscInt *aimax = a->imax,*ai = a->i,*ailen = a->ilen,*aj = a->j; 5604 MatScalar *aa = a->a; 5605 PetscBool ignorezeroentries = (((a->ignorezeroentries)&&(addv==ADD_VALUES))?PETSC_TRUE:PETSC_FALSE); 5606 Mat B = aij->B; 5607 Mat_SeqAIJ *b = (Mat_SeqAIJ*)B->data; 5608 PetscInt *bimax = b->imax,*bi = b->i,*bilen = b->ilen,*bj = b->j,bm = aij->B->rmap->n,am = aij->A->rmap->n; 5609 MatScalar *ba = b->a; 5610 5611 PetscInt *rp1,*rp2,ii,nrow1,nrow2,_i,rmax1,rmax2,N,low1,high1,low2,high2,t,lastcol1,lastcol2; 5612 PetscInt nonew = a->nonew; 5613 MatScalar *ap1,*ap2; 5614 5615 PetscFunctionBegin; 5616 for (i=0; i<m; i++) { 5617 if (im[i] < 0) continue; 5618 #if defined(PETSC_USE_DEBUG) 5619 if (im[i] >= mat->rmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",im[i],mat->rmap->N-1); 5620 #endif 5621 if (im[i] >= rstart && im[i] < rend) { 5622 row = im[i] - rstart; 5623 lastcol1 = -1; 5624 rp1 = aj + ai[row]; 5625 ap1 = aa + ai[row]; 5626 rmax1 = aimax[row]; 5627 nrow1 = ailen[row]; 5628 low1 = 0; 5629 high1 = nrow1; 5630 lastcol2 = -1; 5631 rp2 = bj + bi[row]; 5632 ap2 = ba + bi[row]; 5633 rmax2 = bimax[row]; 5634 nrow2 = bilen[row]; 5635 low2 = 0; 5636 high2 = nrow2; 5637 5638 for (j=0; j<n; j++) { 5639 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 5640 if (ignorezeroentries && value == 0.0 && (addv == ADD_VALUES)) continue; 5641 if (in[j] >= cstart && in[j] < cend){ 5642 col = in[j] - cstart; 5643 MatSetValues_SeqAIJ_A_Private(row,col,value,addv); 5644 } else if (in[j] < 0) continue; 5645 #if defined(PETSC_USE_DEBUG) 5646 else if (in[j] >= mat->cmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[j],mat->cmap->N-1); 5647 #endif 5648 else { 5649 if (mat->was_assembled) { 5650 if (!aij->colmap) { 5651 ierr = CreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 5652 } 5653 #if defined (PETSC_USE_CTABLE) 5654 ierr = PetscTableFind(aij->colmap,in[j]+1,&col);CHKERRQ(ierr); 5655 col--; 5656 #else 5657 col = aij->colmap[in[j]] - 1; 5658 #endif 5659 if (col < 0 && !((Mat_SeqAIJ*)(aij->A->data))->nonew) { 5660 ierr = DisAssemble_MPIAIJ(mat);CHKERRQ(ierr); 5661 col = in[j]; 5662 /* Reinitialize the variables required by MatSetValues_SeqAIJ_B_Private() */ 5663 B = aij->B; 5664 b = (Mat_SeqAIJ*)B->data; 5665 bimax = b->imax; bi = b->i; bilen = b->ilen; bj = b->j; 5666 rp2 = bj + bi[row]; 5667 ap2 = ba + bi[row]; 5668 rmax2 = bimax[row]; 5669 nrow2 = bilen[row]; 5670 low2 = 0; 5671 high2 = nrow2; 5672 bm = aij->B->rmap->n; 5673 ba = b->a; 5674 } 5675 } else col = in[j]; 5676 MatSetValues_SeqAIJ_B_Private(row,col,value,addv); 5677 } 5678 } 5679 } else { 5680 if (!aij->donotstash) { 5681 if (roworiented) { 5682 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n,in,v+i*n,(PetscBool)(ignorezeroentries && (addv == ADD_VALUES)));CHKERRQ(ierr); 5683 } else { 5684 ierr = MatStashValuesCol_Private(&mat->stash,im[i],n,in,v+i,m,(PetscBool)(ignorezeroentries && (addv == ADD_VALUES)));CHKERRQ(ierr); 5685 } 5686 } 5687 } 5688 }} 5689 PetscFunctionReturnVoid(); 5690 } 5691 EXTERN_C_END 5692 5693