1 /*$Id: mpiaij.c,v 1.344 2001/08/10 03:30:48 bsmith Exp $*/ 2 3 #include "src/mat/impls/aij/mpi/mpiaij.h" 4 #include "src/vec/vecimpl.h" 5 #include "src/inline/spops.h" 6 7 EXTERN int MatSetUpMultiply_MPIAIJ(Mat); 8 EXTERN int DisAssemble_MPIAIJ(Mat); 9 EXTERN int MatSetValues_SeqAIJ(Mat,int,int*,int,int*,PetscScalar*,InsertMode); 10 EXTERN int MatGetRow_SeqAIJ(Mat,int,int*,int**,PetscScalar**); 11 EXTERN int MatRestoreRow_SeqAIJ(Mat,int,int*,int**,PetscScalar**); 12 EXTERN int MatPrintHelp_SeqAIJ(Mat); 13 EXTERN int MatUseSuperLU_DIST_MPIAIJ(Mat); 14 15 /* 16 Local utility routine that creates a mapping from the global column 17 number to the local number in the off-diagonal part of the local 18 storage of the matrix. When PETSC_USE_CTABLE is used this is scalable at 19 a slightly higher hash table cost; without it it is not scalable (each processor 20 has an order N integer array but is fast to acess. 21 */ 22 #undef __FUNCT__ 23 #define __FUNCT__ "CreateColmap_MPIAIJ_Private" 24 int CreateColmap_MPIAIJ_Private(Mat mat) 25 { 26 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 27 int n = aij->B->n,i,ierr; 28 29 PetscFunctionBegin; 30 #if defined (PETSC_USE_CTABLE) 31 ierr = PetscTableCreate(n,&aij->colmap);CHKERRQ(ierr); 32 for (i=0; i<n; i++){ 33 ierr = PetscTableAdd(aij->colmap,aij->garray[i]+1,i+1);CHKERRQ(ierr); 34 } 35 #else 36 ierr = PetscMalloc((mat->N+1)*sizeof(int),&aij->colmap);CHKERRQ(ierr); 37 PetscLogObjectMemory(mat,mat->N*sizeof(int)); 38 ierr = PetscMemzero(aij->colmap,mat->N*sizeof(int));CHKERRQ(ierr); 39 for (i=0; i<n; i++) aij->colmap[aij->garray[i]] = i+1; 40 #endif 41 PetscFunctionReturn(0); 42 } 43 44 #define CHUNKSIZE 15 45 #define MatSetValues_SeqAIJ_A_Private(row,col,value,addv) \ 46 { \ 47 \ 48 rp = aj + ai[row] + shift; ap = aa + ai[row] + shift; \ 49 rmax = aimax[row]; nrow = ailen[row]; \ 50 col1 = col - shift; \ 51 \ 52 low = 0; high = nrow; \ 53 while (high-low > 5) { \ 54 t = (low+high)/2; \ 55 if (rp[t] > col) high = t; \ 56 else low = t; \ 57 } \ 58 for (_i=low; _i<high; _i++) { \ 59 if (rp[_i] > col1) break; \ 60 if (rp[_i] == col1) { \ 61 if (addv == ADD_VALUES) ap[_i] += value; \ 62 else ap[_i] = value; \ 63 goto a_noinsert; \ 64 } \ 65 } \ 66 if (nonew == 1) goto a_noinsert; \ 67 else if (nonew == -1) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero into matrix"); \ 68 if (nrow >= rmax) { \ 69 /* there is no extra room in row, therefore enlarge */ \ 70 int new_nz = ai[am] + CHUNKSIZE,len,*new_i,*new_j; \ 71 PetscScalar *new_a; \ 72 \ 73 if (nonew == -2) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero in the matrix"); \ 74 \ 75 /* malloc new storage space */ \ 76 len = new_nz*(sizeof(int)+sizeof(PetscScalar))+(am+1)*sizeof(int); \ 77 ierr = PetscMalloc(len,&new_a);CHKERRQ(ierr); \ 78 new_j = (int*)(new_a + new_nz); \ 79 new_i = new_j + new_nz; \ 80 \ 81 /* copy over old data into new slots */ \ 82 for (ii=0; ii<row+1; ii++) {new_i[ii] = ai[ii];} \ 83 for (ii=row+1; ii<am+1; ii++) {new_i[ii] = ai[ii]+CHUNKSIZE;} \ 84 ierr = PetscMemcpy(new_j,aj,(ai[row]+nrow+shift)*sizeof(int));CHKERRQ(ierr); \ 85 len = (new_nz - CHUNKSIZE - ai[row] - nrow - shift); \ 86 ierr = PetscMemcpy(new_j+ai[row]+shift+nrow+CHUNKSIZE,aj+ai[row]+shift+nrow, \ 87 len*sizeof(int));CHKERRQ(ierr); \ 88 ierr = PetscMemcpy(new_a,aa,(ai[row]+nrow+shift)*sizeof(PetscScalar));CHKERRQ(ierr); \ 89 ierr = PetscMemcpy(new_a+ai[row]+shift+nrow+CHUNKSIZE,aa+ai[row]+shift+nrow, \ 90 len*sizeof(PetscScalar));CHKERRQ(ierr); \ 91 /* free up old matrix storage */ \ 92 \ 93 ierr = PetscFree(a->a);CHKERRQ(ierr); \ 94 if (!a->singlemalloc) { \ 95 ierr = PetscFree(a->i);CHKERRQ(ierr); \ 96 ierr = PetscFree(a->j);CHKERRQ(ierr); \ 97 } \ 98 aa = a->a = new_a; ai = a->i = new_i; aj = a->j = new_j; \ 99 a->singlemalloc = PETSC_TRUE; \ 100 \ 101 rp = aj + ai[row] + shift; ap = aa + ai[row] + shift; \ 102 rmax = aimax[row] = aimax[row] + CHUNKSIZE; \ 103 PetscLogObjectMemory(A,CHUNKSIZE*(sizeof(int) + sizeof(PetscScalar))); \ 104 a->maxnz += CHUNKSIZE; \ 105 a->reallocs++; \ 106 } \ 107 N = nrow++ - 1; a->nz++; \ 108 /* shift up all the later entries in this row */ \ 109 for (ii=N; ii>=_i; ii--) { \ 110 rp[ii+1] = rp[ii]; \ 111 ap[ii+1] = ap[ii]; \ 112 } \ 113 rp[_i] = col1; \ 114 ap[_i] = value; \ 115 a_noinsert: ; \ 116 ailen[row] = nrow; \ 117 } 118 119 #define MatSetValues_SeqAIJ_B_Private(row,col,value,addv) \ 120 { \ 121 \ 122 rp = bj + bi[row] + shift; ap = ba + bi[row] + shift; \ 123 rmax = bimax[row]; nrow = bilen[row]; \ 124 col1 = col - shift; \ 125 \ 126 low = 0; high = nrow; \ 127 while (high-low > 5) { \ 128 t = (low+high)/2; \ 129 if (rp[t] > col) high = t; \ 130 else low = t; \ 131 } \ 132 for (_i=low; _i<high; _i++) { \ 133 if (rp[_i] > col1) break; \ 134 if (rp[_i] == col1) { \ 135 if (addv == ADD_VALUES) ap[_i] += value; \ 136 else ap[_i] = value; \ 137 goto b_noinsert; \ 138 } \ 139 } \ 140 if (nonew == 1) goto b_noinsert; \ 141 else if (nonew == -1) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero into matrix"); \ 142 if (nrow >= rmax) { \ 143 /* there is no extra room in row, therefore enlarge */ \ 144 int new_nz = bi[bm] + CHUNKSIZE,len,*new_i,*new_j; \ 145 PetscScalar *new_a; \ 146 \ 147 if (nonew == -2) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero in the matrix"); \ 148 \ 149 /* malloc new storage space */ \ 150 len = new_nz*(sizeof(int)+sizeof(PetscScalar))+(bm+1)*sizeof(int); \ 151 ierr = PetscMalloc(len,&new_a);CHKERRQ(ierr); \ 152 new_j = (int*)(new_a + new_nz); \ 153 new_i = new_j + new_nz; \ 154 \ 155 /* copy over old data into new slots */ \ 156 for (ii=0; ii<row+1; ii++) {new_i[ii] = bi[ii];} \ 157 for (ii=row+1; ii<bm+1; ii++) {new_i[ii] = bi[ii]+CHUNKSIZE;} \ 158 ierr = PetscMemcpy(new_j,bj,(bi[row]+nrow+shift)*sizeof(int));CHKERRQ(ierr); \ 159 len = (new_nz - CHUNKSIZE - bi[row] - nrow - shift); \ 160 ierr = PetscMemcpy(new_j+bi[row]+shift+nrow+CHUNKSIZE,bj+bi[row]+shift+nrow, \ 161 len*sizeof(int));CHKERRQ(ierr); \ 162 ierr = PetscMemcpy(new_a,ba,(bi[row]+nrow+shift)*sizeof(PetscScalar));CHKERRQ(ierr); \ 163 ierr = PetscMemcpy(new_a+bi[row]+shift+nrow+CHUNKSIZE,ba+bi[row]+shift+nrow, \ 164 len*sizeof(PetscScalar));CHKERRQ(ierr); \ 165 /* free up old matrix storage */ \ 166 \ 167 ierr = PetscFree(b->a);CHKERRQ(ierr); \ 168 if (!b->singlemalloc) { \ 169 ierr = PetscFree(b->i);CHKERRQ(ierr); \ 170 ierr = PetscFree(b->j);CHKERRQ(ierr); \ 171 } \ 172 ba = b->a = new_a; bi = b->i = new_i; bj = b->j = new_j; \ 173 b->singlemalloc = PETSC_TRUE; \ 174 \ 175 rp = bj + bi[row] + shift; ap = ba + bi[row] + shift; \ 176 rmax = bimax[row] = bimax[row] + CHUNKSIZE; \ 177 PetscLogObjectMemory(B,CHUNKSIZE*(sizeof(int) + sizeof(PetscScalar))); \ 178 b->maxnz += CHUNKSIZE; \ 179 b->reallocs++; \ 180 } \ 181 N = nrow++ - 1; b->nz++; \ 182 /* shift up all the later entries in this row */ \ 183 for (ii=N; ii>=_i; ii--) { \ 184 rp[ii+1] = rp[ii]; \ 185 ap[ii+1] = ap[ii]; \ 186 } \ 187 rp[_i] = col1; \ 188 ap[_i] = value; \ 189 b_noinsert: ; \ 190 bilen[row] = nrow; \ 191 } 192 193 #undef __FUNCT__ 194 #define __FUNCT__ "MatSetValues_MPIAIJ" 195 int MatSetValues_MPIAIJ(Mat mat,int m,int *im,int n,int *in,PetscScalar *v,InsertMode addv) 196 { 197 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 198 PetscScalar value; 199 int ierr,i,j,rstart = aij->rstart,rend = aij->rend; 200 int cstart = aij->cstart,cend = aij->cend,row,col; 201 PetscTruth roworiented = aij->roworiented; 202 203 /* Some Variables required in the macro */ 204 Mat A = aij->A; 205 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 206 int *aimax = a->imax,*ai = a->i,*ailen = a->ilen,*aj = a->j; 207 PetscScalar *aa = a->a; 208 PetscTruth ignorezeroentries = (((a->ignorezeroentries)&&(addv==ADD_VALUES))?PETSC_TRUE:PETSC_FALSE); 209 Mat B = aij->B; 210 Mat_SeqAIJ *b = (Mat_SeqAIJ*)B->data; 211 int *bimax = b->imax,*bi = b->i,*bilen = b->ilen,*bj = b->j,bm = aij->B->m,am = aij->A->m; 212 PetscScalar *ba = b->a; 213 214 int *rp,ii,nrow,_i,rmax,N,col1,low,high,t; 215 int nonew = a->nonew,shift = a->indexshift; 216 PetscScalar *ap; 217 218 PetscFunctionBegin; 219 for (i=0; i<m; i++) { 220 if (im[i] < 0) continue; 221 #if defined(PETSC_USE_BOPT_g) 222 if (im[i] >= mat->M) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Row too large"); 223 #endif 224 if (im[i] >= rstart && im[i] < rend) { 225 row = im[i] - rstart; 226 for (j=0; j<n; j++) { 227 if (in[j] >= cstart && in[j] < cend){ 228 col = in[j] - cstart; 229 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 230 if (ignorezeroentries && value == 0.0) continue; 231 MatSetValues_SeqAIJ_A_Private(row,col,value,addv); 232 /* ierr = MatSetValues_SeqAIJ(aij->A,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 233 } else if (in[j] < 0) continue; 234 #if defined(PETSC_USE_BOPT_g) 235 else if (in[j] >= mat->N) {SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Column too large");} 236 #endif 237 else { 238 if (mat->was_assembled) { 239 if (!aij->colmap) { 240 ierr = CreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 241 } 242 #if defined (PETSC_USE_CTABLE) 243 ierr = PetscTableFind(aij->colmap,in[j]+1,&col);CHKERRQ(ierr); 244 col--; 245 #else 246 col = aij->colmap[in[j]] - 1; 247 #endif 248 if (col < 0 && !((Mat_SeqAIJ*)(aij->A->data))->nonew) { 249 ierr = DisAssemble_MPIAIJ(mat);CHKERRQ(ierr); 250 col = in[j]; 251 /* Reinitialize the variables required by MatSetValues_SeqAIJ_B_Private() */ 252 B = aij->B; 253 b = (Mat_SeqAIJ*)B->data; 254 bimax = b->imax; bi = b->i; bilen = b->ilen; bj = b->j; 255 ba = b->a; 256 } 257 } else col = in[j]; 258 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 259 if (ignorezeroentries && value == 0.0) continue; 260 MatSetValues_SeqAIJ_B_Private(row,col,value,addv); 261 /* ierr = MatSetValues_SeqAIJ(aij->B,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 262 } 263 } 264 } else { 265 if (!aij->donotstash) { 266 if (roworiented) { 267 if (ignorezeroentries && v[i*n] == 0.0) continue; 268 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n,in,v+i*n);CHKERRQ(ierr); 269 } else { 270 if (ignorezeroentries && v[i] == 0.0) continue; 271 ierr = MatStashValuesCol_Private(&mat->stash,im[i],n,in,v+i,m);CHKERRQ(ierr); 272 } 273 } 274 } 275 } 276 PetscFunctionReturn(0); 277 } 278 279 #undef __FUNCT__ 280 #define __FUNCT__ "MatGetValues_MPIAIJ" 281 int MatGetValues_MPIAIJ(Mat mat,int m,int *idxm,int n,int *idxn,PetscScalar *v) 282 { 283 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 284 int ierr,i,j,rstart = aij->rstart,rend = aij->rend; 285 int cstart = aij->cstart,cend = aij->cend,row,col; 286 287 PetscFunctionBegin; 288 for (i=0; i<m; i++) { 289 if (idxm[i] < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative row"); 290 if (idxm[i] >= mat->M) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Row too large"); 291 if (idxm[i] >= rstart && idxm[i] < rend) { 292 row = idxm[i] - rstart; 293 for (j=0; j<n; j++) { 294 if (idxn[j] < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative column"); 295 if (idxn[j] >= mat->N) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Column too large"); 296 if (idxn[j] >= cstart && idxn[j] < cend){ 297 col = idxn[j] - cstart; 298 ierr = MatGetValues(aij->A,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 299 } else { 300 if (!aij->colmap) { 301 ierr = CreateColmap_MPIAIJ_Private(mat);CHKERRQ(ierr); 302 } 303 #if defined (PETSC_USE_CTABLE) 304 ierr = PetscTableFind(aij->colmap,idxn[j]+1,&col);CHKERRQ(ierr); 305 col --; 306 #else 307 col = aij->colmap[idxn[j]] - 1; 308 #endif 309 if ((col < 0) || (aij->garray[col] != idxn[j])) *(v+i*n+j) = 0.0; 310 else { 311 ierr = MatGetValues(aij->B,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 312 } 313 } 314 } 315 } else { 316 SETERRQ(PETSC_ERR_SUP,"Only local values currently supported"); 317 } 318 } 319 PetscFunctionReturn(0); 320 } 321 322 #undef __FUNCT__ 323 #define __FUNCT__ "MatAssemblyBegin_MPIAIJ" 324 int MatAssemblyBegin_MPIAIJ(Mat mat,MatAssemblyType mode) 325 { 326 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 327 int ierr,nstash,reallocs; 328 InsertMode addv; 329 330 PetscFunctionBegin; 331 if (aij->donotstash) { 332 PetscFunctionReturn(0); 333 } 334 335 /* make sure all processors are either in INSERTMODE or ADDMODE */ 336 ierr = MPI_Allreduce(&mat->insertmode,&addv,1,MPI_INT,MPI_BOR,mat->comm);CHKERRQ(ierr); 337 if (addv == (ADD_VALUES|INSERT_VALUES)) { 338 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Some processors inserted others added"); 339 } 340 mat->insertmode = addv; /* in case this processor had no cache */ 341 342 ierr = MatStashScatterBegin_Private(&mat->stash,aij->rowners);CHKERRQ(ierr); 343 ierr = MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);CHKERRQ(ierr); 344 PetscLogInfo(aij->A,"MatAssemblyBegin_MPIAIJ:Stash has %d entries, uses %d mallocs.\n",nstash,reallocs); 345 PetscFunctionReturn(0); 346 } 347 348 349 #undef __FUNCT__ 350 #define __FUNCT__ "MatAssemblyEnd_MPIAIJ" 351 int MatAssemblyEnd_MPIAIJ(Mat mat,MatAssemblyType mode) 352 { 353 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 354 int i,j,rstart,ncols,n,ierr,flg; 355 int *row,*col,other_disassembled; 356 PetscScalar *val; 357 InsertMode addv = mat->insertmode; 358 #if defined(PETSC_HAVE_SUPERLUDIST) 359 PetscTruth flag; 360 #endif 361 362 PetscFunctionBegin; 363 if (!aij->donotstash) { 364 while (1) { 365 ierr = MatStashScatterGetMesg_Private(&mat->stash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 366 if (!flg) break; 367 368 for (i=0; i<n;) { 369 /* Now identify the consecutive vals belonging to the same row */ 370 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 371 if (j < n) ncols = j-i; 372 else ncols = n-i; 373 /* Now assemble all these values with a single function call */ 374 ierr = MatSetValues_MPIAIJ(mat,1,row+i,ncols,col+i,val+i,addv);CHKERRQ(ierr); 375 i = j; 376 } 377 } 378 ierr = MatStashScatterEnd_Private(&mat->stash);CHKERRQ(ierr); 379 } 380 381 ierr = MatAssemblyBegin(aij->A,mode);CHKERRQ(ierr); 382 ierr = MatAssemblyEnd(aij->A,mode);CHKERRQ(ierr); 383 384 /* determine if any processor has disassembled, if so we must 385 also disassemble ourselfs, in order that we may reassemble. */ 386 /* 387 if nonzero structure of submatrix B cannot change then we know that 388 no processor disassembled thus we can skip this stuff 389 */ 390 if (!((Mat_SeqAIJ*)aij->B->data)->nonew) { 391 ierr = MPI_Allreduce(&mat->was_assembled,&other_disassembled,1,MPI_INT,MPI_PROD,mat->comm);CHKERRQ(ierr); 392 if (mat->was_assembled && !other_disassembled) { 393 ierr = DisAssemble_MPIAIJ(mat);CHKERRQ(ierr); 394 } 395 } 396 397 if (!mat->was_assembled && mode == MAT_FINAL_ASSEMBLY) { 398 ierr = MatSetUpMultiply_MPIAIJ(mat);CHKERRQ(ierr); 399 } 400 ierr = MatAssemblyBegin(aij->B,mode);CHKERRQ(ierr); 401 ierr = MatAssemblyEnd(aij->B,mode);CHKERRQ(ierr); 402 403 if (aij->rowvalues) { 404 ierr = PetscFree(aij->rowvalues);CHKERRQ(ierr); 405 aij->rowvalues = 0; 406 } 407 #if defined(PETSC_HAVE_SUPERLUDIST) 408 ierr = PetscOptionsHasName(PETSC_NULL,"-mat_aij_superlu_dist",&flag);CHKERRQ(ierr); 409 if (flag) { ierr = MatUseSuperLU_DIST_MPIAIJ(mat);CHKERRQ(ierr); } 410 #endif 411 PetscFunctionReturn(0); 412 } 413 414 #undef __FUNCT__ 415 #define __FUNCT__ "MatZeroEntries_MPIAIJ" 416 int MatZeroEntries_MPIAIJ(Mat A) 417 { 418 Mat_MPIAIJ *l = (Mat_MPIAIJ*)A->data; 419 int ierr; 420 421 PetscFunctionBegin; 422 ierr = MatZeroEntries(l->A);CHKERRQ(ierr); 423 ierr = MatZeroEntries(l->B);CHKERRQ(ierr); 424 PetscFunctionReturn(0); 425 } 426 427 #undef __FUNCT__ 428 #define __FUNCT__ "MatZeroRows_MPIAIJ" 429 int MatZeroRows_MPIAIJ(Mat A,IS is,PetscScalar *diag) 430 { 431 Mat_MPIAIJ *l = (Mat_MPIAIJ*)A->data; 432 int i,ierr,N,*rows,*owners = l->rowners,size = l->size; 433 int *procs,*nprocs,j,idx,nsends,*work,row; 434 int nmax,*svalues,*starts,*owner,nrecvs,rank = l->rank; 435 int *rvalues,tag = A->tag,count,base,slen,n,*source; 436 int *lens,imdex,*lrows,*values,rstart=l->rstart; 437 MPI_Comm comm = A->comm; 438 MPI_Request *send_waits,*recv_waits; 439 MPI_Status recv_status,*send_status; 440 IS istmp; 441 PetscTruth found; 442 443 PetscFunctionBegin; 444 ierr = ISGetLocalSize(is,&N);CHKERRQ(ierr); 445 ierr = ISGetIndices(is,&rows);CHKERRQ(ierr); 446 447 /* first count number of contributors to each processor */ 448 ierr = PetscMalloc(2*size*sizeof(int),&nprocs);CHKERRQ(ierr); 449 ierr = PetscMemzero(nprocs,2*size*sizeof(int));CHKERRQ(ierr); 450 procs = nprocs + size; 451 ierr = PetscMalloc((N+1)*sizeof(int),&owner);CHKERRQ(ierr); /* see note*/ 452 for (i=0; i<N; i++) { 453 idx = rows[i]; 454 found = PETSC_FALSE; 455 for (j=0; j<size; j++) { 456 if (idx >= owners[j] && idx < owners[j+1]) { 457 nprocs[j]++; procs[j] = 1; owner[i] = j; found = PETSC_TRUE; break; 458 } 459 } 460 if (!found) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Index out of range"); 461 } 462 nsends = 0; for (i=0; i<size; i++) { nsends += procs[i];} 463 464 /* inform other processors of number of messages and max length*/ 465 ierr = PetscMalloc(2*size*sizeof(int),&work);CHKERRQ(ierr); 466 ierr = MPI_Allreduce(nprocs,work,2*size,MPI_INT,PetscMaxSum_Op,comm);CHKERRQ(ierr); 467 nrecvs = work[size+rank]; 468 nmax = work[rank]; 469 ierr = PetscFree(work);CHKERRQ(ierr); 470 471 /* post receives: */ 472 ierr = PetscMalloc((nrecvs+1)*(nmax+1)*sizeof(int),&rvalues);CHKERRQ(ierr); 473 ierr = PetscMalloc((nrecvs+1)*sizeof(MPI_Request),&recv_waits);CHKERRQ(ierr); 474 for (i=0; i<nrecvs; i++) { 475 ierr = MPI_Irecv(rvalues+nmax*i,nmax,MPI_INT,MPI_ANY_SOURCE,tag,comm,recv_waits+i);CHKERRQ(ierr); 476 } 477 478 /* do sends: 479 1) starts[i] gives the starting index in svalues for stuff going to 480 the ith processor 481 */ 482 ierr = PetscMalloc((N+1)*sizeof(int),&svalues);CHKERRQ(ierr); 483 ierr = PetscMalloc((nsends+1)*sizeof(MPI_Request),&send_waits);CHKERRQ(ierr); 484 ierr = PetscMalloc((size+1)*sizeof(int),&starts);CHKERRQ(ierr); 485 starts[0] = 0; 486 for (i=1; i<size; i++) { starts[i] = starts[i-1] + nprocs[i-1];} 487 for (i=0; i<N; i++) { 488 svalues[starts[owner[i]]++] = rows[i]; 489 } 490 ierr = ISRestoreIndices(is,&rows);CHKERRQ(ierr); 491 492 starts[0] = 0; 493 for (i=1; i<size+1; i++) { starts[i] = starts[i-1] + nprocs[i-1];} 494 count = 0; 495 for (i=0; i<size; i++) { 496 if (procs[i]) { 497 ierr = MPI_Isend(svalues+starts[i],nprocs[i],MPI_INT,i,tag,comm,send_waits+count++);CHKERRQ(ierr); 498 } 499 } 500 ierr = PetscFree(starts);CHKERRQ(ierr); 501 502 base = owners[rank]; 503 504 /* wait on receives */ 505 ierr = PetscMalloc(2*(nrecvs+1)*sizeof(int),&lens);CHKERRQ(ierr); 506 source = lens + nrecvs; 507 count = nrecvs; slen = 0; 508 while (count) { 509 ierr = MPI_Waitany(nrecvs,recv_waits,&imdex,&recv_status);CHKERRQ(ierr); 510 /* unpack receives into our local space */ 511 ierr = MPI_Get_count(&recv_status,MPI_INT,&n);CHKERRQ(ierr); 512 source[imdex] = recv_status.MPI_SOURCE; 513 lens[imdex] = n; 514 slen += n; 515 count--; 516 } 517 ierr = PetscFree(recv_waits);CHKERRQ(ierr); 518 519 /* move the data into the send scatter */ 520 ierr = PetscMalloc((slen+1)*sizeof(int),&lrows);CHKERRQ(ierr); 521 count = 0; 522 for (i=0; i<nrecvs; i++) { 523 values = rvalues + i*nmax; 524 for (j=0; j<lens[i]; j++) { 525 lrows[count++] = values[j] - base; 526 } 527 } 528 ierr = PetscFree(rvalues);CHKERRQ(ierr); 529 ierr = PetscFree(lens);CHKERRQ(ierr); 530 ierr = PetscFree(owner);CHKERRQ(ierr); 531 ierr = PetscFree(nprocs);CHKERRQ(ierr); 532 533 /* actually zap the local rows */ 534 ierr = ISCreateGeneral(PETSC_COMM_SELF,slen,lrows,&istmp);CHKERRQ(ierr); 535 PetscLogObjectParent(A,istmp); 536 537 /* 538 Zero the required rows. If the "diagonal block" of the matrix 539 is square and the user wishes to set the diagonal we use seperate 540 code so that MatSetValues() is not called for each diagonal allocating 541 new memory, thus calling lots of mallocs and slowing things down. 542 543 Contributed by: Mathew Knepley 544 */ 545 /* must zero l->B before l->A because the (diag) case below may put values into l->B*/ 546 ierr = MatZeroRows(l->B,istmp,0);CHKERRQ(ierr); 547 if (diag && (l->A->M == l->A->N)) { 548 ierr = MatZeroRows(l->A,istmp,diag);CHKERRQ(ierr); 549 } else if (diag) { 550 ierr = MatZeroRows(l->A,istmp,0);CHKERRQ(ierr); 551 if (((Mat_SeqAIJ*)l->A->data)->nonew) { 552 SETERRQ(PETSC_ERR_SUP,"MatZeroRows() on rectangular matrices cannot be used with the Mat options\n\ 553 MAT_NO_NEW_NONZERO_LOCATIONS,MAT_NEW_NONZERO_LOCATION_ERR,MAT_NEW_NONZERO_ALLOCATION_ERR"); 554 } 555 for (i = 0; i < slen; i++) { 556 row = lrows[i] + rstart; 557 ierr = MatSetValues(A,1,&row,1,&row,diag,INSERT_VALUES);CHKERRQ(ierr); 558 } 559 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 560 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 561 } else { 562 ierr = MatZeroRows(l->A,istmp,0);CHKERRQ(ierr); 563 } 564 ierr = ISDestroy(istmp);CHKERRQ(ierr); 565 ierr = PetscFree(lrows);CHKERRQ(ierr); 566 567 /* wait on sends */ 568 if (nsends) { 569 ierr = PetscMalloc(nsends*sizeof(MPI_Status),&send_status);CHKERRQ(ierr); 570 ierr = MPI_Waitall(nsends,send_waits,send_status);CHKERRQ(ierr); 571 ierr = PetscFree(send_status);CHKERRQ(ierr); 572 } 573 ierr = PetscFree(send_waits);CHKERRQ(ierr); 574 ierr = PetscFree(svalues);CHKERRQ(ierr); 575 576 PetscFunctionReturn(0); 577 } 578 579 #undef __FUNCT__ 580 #define __FUNCT__ "MatMult_MPIAIJ" 581 int MatMult_MPIAIJ(Mat A,Vec xx,Vec yy) 582 { 583 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 584 int ierr,nt; 585 586 PetscFunctionBegin; 587 ierr = VecGetLocalSize(xx,&nt);CHKERRQ(ierr); 588 if (nt != A->n) { 589 SETERRQ2(PETSC_ERR_ARG_SIZ,"Incompatible partition of A (%d) and xx (%d)",A->n,nt); 590 } 591 ierr = VecScatterBegin(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 592 ierr = (*a->A->ops->mult)(a->A,xx,yy);CHKERRQ(ierr); 593 ierr = VecScatterEnd(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 594 ierr = (*a->B->ops->multadd)(a->B,a->lvec,yy,yy);CHKERRQ(ierr); 595 PetscFunctionReturn(0); 596 } 597 598 #undef __FUNCT__ 599 #define __FUNCT__ "MatMultAdd_MPIAIJ" 600 int MatMultAdd_MPIAIJ(Mat A,Vec xx,Vec yy,Vec zz) 601 { 602 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 603 int ierr; 604 605 PetscFunctionBegin; 606 ierr = VecScatterBegin(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 607 ierr = (*a->A->ops->multadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 608 ierr = VecScatterEnd(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 609 ierr = (*a->B->ops->multadd)(a->B,a->lvec,zz,zz);CHKERRQ(ierr); 610 PetscFunctionReturn(0); 611 } 612 613 #undef __FUNCT__ 614 #define __FUNCT__ "MatMultTranspose_MPIAIJ" 615 int MatMultTranspose_MPIAIJ(Mat A,Vec xx,Vec yy) 616 { 617 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 618 int ierr; 619 620 PetscFunctionBegin; 621 /* do nondiagonal part */ 622 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 623 /* send it on its way */ 624 ierr = VecScatterBegin(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 625 /* do local part */ 626 ierr = (*a->A->ops->multtranspose)(a->A,xx,yy);CHKERRQ(ierr); 627 /* receive remote parts: note this assumes the values are not actually */ 628 /* inserted in yy until the next line, which is true for my implementation*/ 629 /* but is not perhaps always true. */ 630 ierr = VecScatterEnd(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 631 PetscFunctionReturn(0); 632 } 633 634 #undef __FUNCT__ 635 #define __FUNCT__ "MatMultTransposeAdd_MPIAIJ" 636 int MatMultTransposeAdd_MPIAIJ(Mat A,Vec xx,Vec yy,Vec zz) 637 { 638 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 639 int ierr; 640 641 PetscFunctionBegin; 642 /* do nondiagonal part */ 643 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 644 /* send it on its way */ 645 ierr = VecScatterBegin(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 646 /* do local part */ 647 ierr = (*a->A->ops->multtransposeadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 648 /* receive remote parts: note this assumes the values are not actually */ 649 /* inserted in yy until the next line, which is true for my implementation*/ 650 /* but is not perhaps always true. */ 651 ierr = VecScatterEnd(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 652 PetscFunctionReturn(0); 653 } 654 655 /* 656 This only works correctly for square matrices where the subblock A->A is the 657 diagonal block 658 */ 659 #undef __FUNCT__ 660 #define __FUNCT__ "MatGetDiagonal_MPIAIJ" 661 int MatGetDiagonal_MPIAIJ(Mat A,Vec v) 662 { 663 int ierr; 664 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 665 666 PetscFunctionBegin; 667 if (A->M != A->N) SETERRQ(PETSC_ERR_SUP,"Supports only square matrix where A->A is diag block"); 668 if (a->rstart != a->cstart || a->rend != a->cend) { 669 SETERRQ(PETSC_ERR_ARG_SIZ,"row partition must equal col partition"); 670 } 671 ierr = MatGetDiagonal(a->A,v);CHKERRQ(ierr); 672 PetscFunctionReturn(0); 673 } 674 675 #undef __FUNCT__ 676 #define __FUNCT__ "MatScale_MPIAIJ" 677 int MatScale_MPIAIJ(PetscScalar *aa,Mat A) 678 { 679 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 680 int ierr; 681 682 PetscFunctionBegin; 683 ierr = MatScale(aa,a->A);CHKERRQ(ierr); 684 ierr = MatScale(aa,a->B);CHKERRQ(ierr); 685 PetscFunctionReturn(0); 686 } 687 688 #undef __FUNCT__ 689 #define __FUNCT__ "MatDestroy_MPIAIJ" 690 int MatDestroy_MPIAIJ(Mat mat) 691 { 692 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 693 int ierr; 694 695 PetscFunctionBegin; 696 #if defined(PETSC_USE_LOG) 697 PetscLogObjectState((PetscObject)mat,"Rows=%d, Cols=%d",mat->M,mat->N); 698 #endif 699 ierr = MatStashDestroy_Private(&mat->stash);CHKERRQ(ierr); 700 ierr = PetscFree(aij->rowners);CHKERRQ(ierr); 701 ierr = MatDestroy(aij->A);CHKERRQ(ierr); 702 ierr = MatDestroy(aij->B);CHKERRQ(ierr); 703 #if defined (PETSC_USE_CTABLE) 704 if (aij->colmap) {ierr = PetscTableDelete(aij->colmap);CHKERRQ(ierr);} 705 #else 706 if (aij->colmap) {ierr = PetscFree(aij->colmap);CHKERRQ(ierr);} 707 #endif 708 if (aij->garray) {ierr = PetscFree(aij->garray);CHKERRQ(ierr);} 709 if (aij->lvec) {ierr = VecDestroy(aij->lvec);CHKERRQ(ierr);} 710 if (aij->Mvctx) {ierr = VecScatterDestroy(aij->Mvctx);CHKERRQ(ierr);} 711 if (aij->rowvalues) {ierr = PetscFree(aij->rowvalues);CHKERRQ(ierr);} 712 ierr = PetscFree(aij);CHKERRQ(ierr); 713 PetscFunctionReturn(0); 714 } 715 716 #undef __FUNCT__ 717 #define __FUNCT__ "MatView_MPIAIJ_ASCIIorDraworSocket" 718 int MatView_MPIAIJ_ASCIIorDraworSocket(Mat mat,PetscViewer viewer) 719 { 720 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 721 Mat_SeqAIJ* C = (Mat_SeqAIJ*)aij->A->data; 722 int ierr,shift = C->indexshift,rank = aij->rank,size = aij->size; 723 PetscTruth isdraw,isascii,flg; 724 PetscViewer sviewer; 725 PetscViewerFormat format; 726 727 PetscFunctionBegin; 728 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 729 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&isascii);CHKERRQ(ierr); 730 if (isascii) { 731 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 732 if (format == PETSC_VIEWER_ASCII_INFO_LONG) { 733 MatInfo info; 734 ierr = MPI_Comm_rank(mat->comm,&rank);CHKERRQ(ierr); 735 ierr = MatGetInfo(mat,MAT_LOCAL,&info);CHKERRQ(ierr); 736 ierr = PetscOptionsHasName(PETSC_NULL,"-mat_aij_no_inode",&flg);CHKERRQ(ierr); 737 if (flg) { 738 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %d nz %d nz alloced %d mem %d, not using I-node routines\n", 739 rank,mat->m,(int)info.nz_used,(int)info.nz_allocated,(int)info.memory);CHKERRQ(ierr); 740 } else { 741 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %d nz %d nz alloced %d mem %d, using I-node routines\n", 742 rank,mat->m,(int)info.nz_used,(int)info.nz_allocated,(int)info.memory);CHKERRQ(ierr); 743 } 744 ierr = MatGetInfo(aij->A,MAT_LOCAL,&info);CHKERRQ(ierr); 745 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] on-diagonal part: nz %d \n",rank,(int)info.nz_used);CHKERRQ(ierr); 746 ierr = MatGetInfo(aij->B,MAT_LOCAL,&info);CHKERRQ(ierr); 747 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] off-diagonal part: nz %d \n",rank,(int)info.nz_used);CHKERRQ(ierr); 748 ierr = PetscViewerFlush(viewer);CHKERRQ(ierr); 749 ierr = VecScatterView(aij->Mvctx,viewer);CHKERRQ(ierr); 750 PetscFunctionReturn(0); 751 } else if (format == PETSC_VIEWER_ASCII_INFO) { 752 PetscFunctionReturn(0); 753 } 754 } else if (isdraw) { 755 PetscDraw draw; 756 PetscTruth isnull; 757 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 758 ierr = PetscDrawIsNull(draw,&isnull);CHKERRQ(ierr); if (isnull) PetscFunctionReturn(0); 759 } 760 761 if (size == 1) { 762 ierr = PetscObjectSetName((PetscObject)aij->A,mat->name);CHKERRQ(ierr); 763 ierr = MatView(aij->A,viewer);CHKERRQ(ierr); 764 } else { 765 /* assemble the entire matrix onto first processor. */ 766 Mat A; 767 Mat_SeqAIJ *Aloc; 768 int M = mat->M,N = mat->N,m,*ai,*aj,row,*cols,i,*ct; 769 PetscScalar *a; 770 771 if (!rank) { 772 ierr = MatCreateMPIAIJ(mat->comm,M,N,M,N,0,PETSC_NULL,0,PETSC_NULL,&A);CHKERRQ(ierr); 773 } else { 774 ierr = MatCreateMPIAIJ(mat->comm,0,0,M,N,0,PETSC_NULL,0,PETSC_NULL,&A);CHKERRQ(ierr); 775 } 776 PetscLogObjectParent(mat,A); 777 778 /* copy over the A part */ 779 Aloc = (Mat_SeqAIJ*)aij->A->data; 780 m = aij->A->m; ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 781 row = aij->rstart; 782 for (i=0; i<ai[m]+shift; i++) {aj[i] += aij->cstart + shift;} 783 for (i=0; i<m; i++) { 784 ierr = MatSetValues(A,1,&row,ai[i+1]-ai[i],aj,a,INSERT_VALUES);CHKERRQ(ierr); 785 row++; a += ai[i+1]-ai[i]; aj += ai[i+1]-ai[i]; 786 } 787 aj = Aloc->j; 788 for (i=0; i<ai[m]+shift; i++) {aj[i] -= aij->cstart + shift;} 789 790 /* copy over the B part */ 791 Aloc = (Mat_SeqAIJ*)aij->B->data; 792 m = aij->B->m; ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 793 row = aij->rstart; 794 ierr = PetscMalloc((ai[m]+1)*sizeof(int),&cols);CHKERRQ(ierr); 795 ct = cols; 796 for (i=0; i<ai[m]+shift; i++) {cols[i] = aij->garray[aj[i]+shift];} 797 for (i=0; i<m; i++) { 798 ierr = MatSetValues(A,1,&row,ai[i+1]-ai[i],cols,a,INSERT_VALUES);CHKERRQ(ierr); 799 row++; a += ai[i+1]-ai[i]; cols += ai[i+1]-ai[i]; 800 } 801 ierr = PetscFree(ct);CHKERRQ(ierr); 802 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 803 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 804 /* 805 Everyone has to call to draw the matrix since the graphics waits are 806 synchronized across all processors that share the PetscDraw object 807 */ 808 ierr = PetscViewerGetSingleton(viewer,&sviewer);CHKERRQ(ierr); 809 if (!rank) { 810 ierr = PetscObjectSetName((PetscObject)((Mat_MPIAIJ*)(A->data))->A,mat->name);CHKERRQ(ierr); 811 ierr = MatView(((Mat_MPIAIJ*)(A->data))->A,sviewer);CHKERRQ(ierr); 812 } 813 ierr = PetscViewerRestoreSingleton(viewer,&sviewer);CHKERRQ(ierr); 814 ierr = MatDestroy(A);CHKERRQ(ierr); 815 } 816 PetscFunctionReturn(0); 817 } 818 819 #undef __FUNCT__ 820 #define __FUNCT__ "MatView_MPIAIJ" 821 int MatView_MPIAIJ(Mat mat,PetscViewer viewer) 822 { 823 int ierr; 824 PetscTruth isascii,isdraw,issocket,isbinary; 825 826 PetscFunctionBegin; 827 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&isascii);CHKERRQ(ierr); 828 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 829 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_BINARY,&isbinary);CHKERRQ(ierr); 830 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_SOCKET,&issocket);CHKERRQ(ierr); 831 if (isascii || isdraw || isbinary || issocket) { 832 ierr = MatView_MPIAIJ_ASCIIorDraworSocket(mat,viewer);CHKERRQ(ierr); 833 } else { 834 SETERRQ1(1,"Viewer type %s not supported by MPIAIJ matrices",((PetscObject)viewer)->type_name); 835 } 836 PetscFunctionReturn(0); 837 } 838 839 840 841 #undef __FUNCT__ 842 #define __FUNCT__ "MatRelax_MPIAIJ" 843 int MatRelax_MPIAIJ(Mat matin,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,int its,int lits,Vec xx) 844 { 845 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 846 int ierr; 847 Vec bb1; 848 PetscScalar mone=-1.0; 849 850 PetscFunctionBegin; 851 if (its <= 0 || lits <= 0) SETERRQ2(PETSC_ERR_ARG_WRONG,"Relaxation requires global its %d and local its %d both positive",its,lits); 852 853 ierr = VecDuplicate(bb,&bb1);CHKERRQ(ierr); 854 855 if ((flag & SOR_LOCAL_SYMMETRIC_SWEEP) == SOR_LOCAL_SYMMETRIC_SWEEP){ 856 if (flag & SOR_ZERO_INITIAL_GUESS) { 857 ierr = (*mat->A->ops->relax)(mat->A,bb,omega,flag,fshift,lits,PETSC_NULL,xx);CHKERRQ(ierr); 858 its--; 859 } 860 861 while (its--) { 862 ierr = VecScatterBegin(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 863 ierr = VecScatterEnd(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 864 865 /* update rhs: bb1 = bb - B*x */ 866 ierr = VecScale(&mone,mat->lvec);CHKERRQ(ierr); 867 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 868 869 /* local sweep */ 870 ierr = (*mat->A->ops->relax)(mat->A,bb1,omega,SOR_SYMMETRIC_SWEEP,fshift,lits,PETSC_NULL,xx); 871 CHKERRQ(ierr); 872 } 873 } else if (flag & SOR_LOCAL_FORWARD_SWEEP){ 874 if (flag & SOR_ZERO_INITIAL_GUESS) { 875 ierr = (*mat->A->ops->relax)(mat->A,bb,omega,flag,fshift,lits,PETSC_NULL,xx);CHKERRQ(ierr); 876 its--; 877 } 878 while (its--) { 879 ierr = VecScatterBegin(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 880 ierr = VecScatterEnd(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 881 882 /* update rhs: bb1 = bb - B*x */ 883 ierr = VecScale(&mone,mat->lvec);CHKERRQ(ierr); 884 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 885 886 /* local sweep */ 887 ierr = (*mat->A->ops->relax)(mat->A,bb1,omega,SOR_FORWARD_SWEEP,fshift,lits,PETSC_NULL,xx); 888 CHKERRQ(ierr); 889 } 890 } else if (flag & SOR_LOCAL_BACKWARD_SWEEP){ 891 if (flag & SOR_ZERO_INITIAL_GUESS) { 892 ierr = (*mat->A->ops->relax)(mat->A,bb,omega,flag,fshift,lits,PETSC_NULL,xx);CHKERRQ(ierr); 893 its--; 894 } 895 while (its--) { 896 ierr = VecScatterBegin(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 897 ierr = VecScatterEnd(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 898 899 /* update rhs: bb1 = bb - B*x */ 900 ierr = VecScale(&mone,mat->lvec);CHKERRQ(ierr); 901 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 902 903 /* local sweep */ 904 ierr = (*mat->A->ops->relax)(mat->A,bb1,omega,SOR_BACKWARD_SWEEP,fshift,lits,PETSC_NULL,xx); 905 CHKERRQ(ierr); 906 } 907 } else { 908 SETERRQ(PETSC_ERR_SUP,"Parallel SOR not supported"); 909 } 910 911 ierr = VecDestroy(bb1);CHKERRQ(ierr); 912 PetscFunctionReturn(0); 913 } 914 915 #undef __FUNCT__ 916 #define __FUNCT__ "MatGetInfo_MPIAIJ" 917 int MatGetInfo_MPIAIJ(Mat matin,MatInfoType flag,MatInfo *info) 918 { 919 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 920 Mat A = mat->A,B = mat->B; 921 int ierr; 922 PetscReal isend[5],irecv[5]; 923 924 PetscFunctionBegin; 925 info->block_size = 1.0; 926 ierr = MatGetInfo(A,MAT_LOCAL,info);CHKERRQ(ierr); 927 isend[0] = info->nz_used; isend[1] = info->nz_allocated; isend[2] = info->nz_unneeded; 928 isend[3] = info->memory; isend[4] = info->mallocs; 929 ierr = MatGetInfo(B,MAT_LOCAL,info);CHKERRQ(ierr); 930 isend[0] += info->nz_used; isend[1] += info->nz_allocated; isend[2] += info->nz_unneeded; 931 isend[3] += info->memory; isend[4] += info->mallocs; 932 if (flag == MAT_LOCAL) { 933 info->nz_used = isend[0]; 934 info->nz_allocated = isend[1]; 935 info->nz_unneeded = isend[2]; 936 info->memory = isend[3]; 937 info->mallocs = isend[4]; 938 } else if (flag == MAT_GLOBAL_MAX) { 939 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_MAX,matin->comm);CHKERRQ(ierr); 940 info->nz_used = irecv[0]; 941 info->nz_allocated = irecv[1]; 942 info->nz_unneeded = irecv[2]; 943 info->memory = irecv[3]; 944 info->mallocs = irecv[4]; 945 } else if (flag == MAT_GLOBAL_SUM) { 946 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_SUM,matin->comm);CHKERRQ(ierr); 947 info->nz_used = irecv[0]; 948 info->nz_allocated = irecv[1]; 949 info->nz_unneeded = irecv[2]; 950 info->memory = irecv[3]; 951 info->mallocs = irecv[4]; 952 } 953 info->fill_ratio_given = 0; /* no parallel LU/ILU/Cholesky */ 954 info->fill_ratio_needed = 0; 955 info->factor_mallocs = 0; 956 info->rows_global = (double)matin->M; 957 info->columns_global = (double)matin->N; 958 info->rows_local = (double)matin->m; 959 info->columns_local = (double)matin->N; 960 961 PetscFunctionReturn(0); 962 } 963 964 #undef __FUNCT__ 965 #define __FUNCT__ "MatSetOption_MPIAIJ" 966 int MatSetOption_MPIAIJ(Mat A,MatOption op) 967 { 968 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 969 int ierr; 970 971 PetscFunctionBegin; 972 switch (op) { 973 case MAT_NO_NEW_NONZERO_LOCATIONS: 974 case MAT_YES_NEW_NONZERO_LOCATIONS: 975 case MAT_COLUMNS_UNSORTED: 976 case MAT_COLUMNS_SORTED: 977 case MAT_NEW_NONZERO_ALLOCATION_ERR: 978 case MAT_KEEP_ZEROED_ROWS: 979 case MAT_NEW_NONZERO_LOCATION_ERR: 980 case MAT_USE_INODES: 981 case MAT_DO_NOT_USE_INODES: 982 case MAT_IGNORE_ZERO_ENTRIES: 983 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 984 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 985 break; 986 case MAT_ROW_ORIENTED: 987 a->roworiented = PETSC_TRUE; 988 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 989 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 990 break; 991 case MAT_ROWS_SORTED: 992 case MAT_ROWS_UNSORTED: 993 case MAT_YES_NEW_DIAGONALS: 994 case MAT_USE_SINGLE_PRECISION_SOLVES: 995 PetscLogInfo(A,"MatSetOption_MPIAIJ:Option ignored\n"); 996 break; 997 case MAT_COLUMN_ORIENTED: 998 a->roworiented = PETSC_FALSE; 999 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 1000 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 1001 break; 1002 case MAT_IGNORE_OFF_PROC_ENTRIES: 1003 a->donotstash = PETSC_TRUE; 1004 break; 1005 case MAT_NO_NEW_DIAGONALS: 1006 SETERRQ(PETSC_ERR_SUP,"MAT_NO_NEW_DIAGONALS"); 1007 break; 1008 default: 1009 SETERRQ(PETSC_ERR_SUP,"unknown option"); 1010 break; 1011 } 1012 PetscFunctionReturn(0); 1013 } 1014 1015 #undef __FUNCT__ 1016 #define __FUNCT__ "MatGetRow_MPIAIJ" 1017 int MatGetRow_MPIAIJ(Mat matin,int row,int *nz,int **idx,PetscScalar **v) 1018 { 1019 Mat_MPIAIJ *mat = (Mat_MPIAIJ*)matin->data; 1020 PetscScalar *vworkA,*vworkB,**pvA,**pvB,*v_p; 1021 int i,ierr,*cworkA,*cworkB,**pcA,**pcB,cstart = mat->cstart; 1022 int nztot,nzA,nzB,lrow,rstart = mat->rstart,rend = mat->rend; 1023 int *cmap,*idx_p; 1024 1025 PetscFunctionBegin; 1026 if (mat->getrowactive == PETSC_TRUE) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Already active"); 1027 mat->getrowactive = PETSC_TRUE; 1028 1029 if (!mat->rowvalues && (idx || v)) { 1030 /* 1031 allocate enough space to hold information from the longest row. 1032 */ 1033 Mat_SeqAIJ *Aa = (Mat_SeqAIJ*)mat->A->data,*Ba = (Mat_SeqAIJ*)mat->B->data; 1034 int max = 1,tmp; 1035 for (i=0; i<matin->m; i++) { 1036 tmp = Aa->i[i+1] - Aa->i[i] + Ba->i[i+1] - Ba->i[i]; 1037 if (max < tmp) { max = tmp; } 1038 } 1039 ierr = PetscMalloc(max*(sizeof(int)+sizeof(PetscScalar)),&mat->rowvalues);CHKERRQ(ierr); 1040 mat->rowindices = (int*)(mat->rowvalues + max); 1041 } 1042 1043 if (row < rstart || row >= rend) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Only local rows") 1044 lrow = row - rstart; 1045 1046 pvA = &vworkA; pcA = &cworkA; pvB = &vworkB; pcB = &cworkB; 1047 if (!v) {pvA = 0; pvB = 0;} 1048 if (!idx) {pcA = 0; if (!v) pcB = 0;} 1049 ierr = (*mat->A->ops->getrow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1050 ierr = (*mat->B->ops->getrow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1051 nztot = nzA + nzB; 1052 1053 cmap = mat->garray; 1054 if (v || idx) { 1055 if (nztot) { 1056 /* Sort by increasing column numbers, assuming A and B already sorted */ 1057 int imark = -1; 1058 if (v) { 1059 *v = v_p = mat->rowvalues; 1060 for (i=0; i<nzB; i++) { 1061 if (cmap[cworkB[i]] < cstart) v_p[i] = vworkB[i]; 1062 else break; 1063 } 1064 imark = i; 1065 for (i=0; i<nzA; i++) v_p[imark+i] = vworkA[i]; 1066 for (i=imark; i<nzB; i++) v_p[nzA+i] = vworkB[i]; 1067 } 1068 if (idx) { 1069 *idx = idx_p = mat->rowindices; 1070 if (imark > -1) { 1071 for (i=0; i<imark; i++) { 1072 idx_p[i] = cmap[cworkB[i]]; 1073 } 1074 } else { 1075 for (i=0; i<nzB; i++) { 1076 if (cmap[cworkB[i]] < cstart) idx_p[i] = cmap[cworkB[i]]; 1077 else break; 1078 } 1079 imark = i; 1080 } 1081 for (i=0; i<nzA; i++) idx_p[imark+i] = cstart + cworkA[i]; 1082 for (i=imark; i<nzB; i++) idx_p[nzA+i] = cmap[cworkB[i]]; 1083 } 1084 } else { 1085 if (idx) *idx = 0; 1086 if (v) *v = 0; 1087 } 1088 } 1089 *nz = nztot; 1090 ierr = (*mat->A->ops->restorerow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1091 ierr = (*mat->B->ops->restorerow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1092 PetscFunctionReturn(0); 1093 } 1094 1095 #undef __FUNCT__ 1096 #define __FUNCT__ "MatRestoreRow_MPIAIJ" 1097 int MatRestoreRow_MPIAIJ(Mat mat,int row,int *nz,int **idx,PetscScalar **v) 1098 { 1099 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1100 1101 PetscFunctionBegin; 1102 if (aij->getrowactive == PETSC_FALSE) { 1103 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"MatGetRow not called"); 1104 } 1105 aij->getrowactive = PETSC_FALSE; 1106 PetscFunctionReturn(0); 1107 } 1108 1109 #undef __FUNCT__ 1110 #define __FUNCT__ "MatNorm_MPIAIJ" 1111 int MatNorm_MPIAIJ(Mat mat,NormType type,PetscReal *norm) 1112 { 1113 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1114 Mat_SeqAIJ *amat = (Mat_SeqAIJ*)aij->A->data,*bmat = (Mat_SeqAIJ*)aij->B->data; 1115 int ierr,i,j,cstart = aij->cstart,shift = amat->indexshift; 1116 PetscReal sum = 0.0; 1117 PetscScalar *v; 1118 1119 PetscFunctionBegin; 1120 if (aij->size == 1) { 1121 ierr = MatNorm(aij->A,type,norm);CHKERRQ(ierr); 1122 } else { 1123 if (type == NORM_FROBENIUS) { 1124 v = amat->a; 1125 for (i=0; i<amat->nz; i++) { 1126 #if defined(PETSC_USE_COMPLEX) 1127 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 1128 #else 1129 sum += (*v)*(*v); v++; 1130 #endif 1131 } 1132 v = bmat->a; 1133 for (i=0; i<bmat->nz; i++) { 1134 #if defined(PETSC_USE_COMPLEX) 1135 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 1136 #else 1137 sum += (*v)*(*v); v++; 1138 #endif 1139 } 1140 ierr = MPI_Allreduce(&sum,norm,1,MPIU_REAL,MPI_SUM,mat->comm);CHKERRQ(ierr); 1141 *norm = sqrt(*norm); 1142 } else if (type == NORM_1) { /* max column norm */ 1143 PetscReal *tmp,*tmp2; 1144 int *jj,*garray = aij->garray; 1145 ierr = PetscMalloc((mat->N+1)*sizeof(PetscReal),&tmp);CHKERRQ(ierr); 1146 ierr = PetscMalloc((mat->N+1)*sizeof(PetscReal),&tmp2);CHKERRQ(ierr); 1147 ierr = PetscMemzero(tmp,mat->N*sizeof(PetscReal));CHKERRQ(ierr); 1148 *norm = 0.0; 1149 v = amat->a; jj = amat->j; 1150 for (j=0; j<amat->nz; j++) { 1151 tmp[cstart + *jj++ + shift] += PetscAbsScalar(*v); v++; 1152 } 1153 v = bmat->a; jj = bmat->j; 1154 for (j=0; j<bmat->nz; j++) { 1155 tmp[garray[*jj++ + shift]] += PetscAbsScalar(*v); v++; 1156 } 1157 ierr = MPI_Allreduce(tmp,tmp2,mat->N,MPIU_REAL,MPI_SUM,mat->comm);CHKERRQ(ierr); 1158 for (j=0; j<mat->N; j++) { 1159 if (tmp2[j] > *norm) *norm = tmp2[j]; 1160 } 1161 ierr = PetscFree(tmp);CHKERRQ(ierr); 1162 ierr = PetscFree(tmp2);CHKERRQ(ierr); 1163 } else if (type == NORM_INFINITY) { /* max row norm */ 1164 PetscReal ntemp = 0.0; 1165 for (j=0; j<aij->A->m; j++) { 1166 v = amat->a + amat->i[j] + shift; 1167 sum = 0.0; 1168 for (i=0; i<amat->i[j+1]-amat->i[j]; i++) { 1169 sum += PetscAbsScalar(*v); v++; 1170 } 1171 v = bmat->a + bmat->i[j] + shift; 1172 for (i=0; i<bmat->i[j+1]-bmat->i[j]; i++) { 1173 sum += PetscAbsScalar(*v); v++; 1174 } 1175 if (sum > ntemp) ntemp = sum; 1176 } 1177 ierr = MPI_Allreduce(&ntemp,norm,1,MPIU_REAL,MPI_MAX,mat->comm);CHKERRQ(ierr); 1178 } else { 1179 SETERRQ(PETSC_ERR_SUP,"No support for two norm"); 1180 } 1181 } 1182 PetscFunctionReturn(0); 1183 } 1184 1185 #undef __FUNCT__ 1186 #define __FUNCT__ "MatTranspose_MPIAIJ" 1187 int MatTranspose_MPIAIJ(Mat A,Mat *matout) 1188 { 1189 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1190 Mat_SeqAIJ *Aloc = (Mat_SeqAIJ*)a->A->data; 1191 int ierr,shift = Aloc->indexshift; 1192 int M = A->M,N = A->N,m,*ai,*aj,row,*cols,i,*ct; 1193 Mat B; 1194 PetscScalar *array; 1195 1196 PetscFunctionBegin; 1197 if (!matout && M != N) { 1198 SETERRQ(PETSC_ERR_ARG_SIZ,"Square matrix only for in-place"); 1199 } 1200 1201 ierr = MatCreateMPIAIJ(A->comm,A->n,A->m,N,M,0,PETSC_NULL,0,PETSC_NULL,&B);CHKERRQ(ierr); 1202 1203 /* copy over the A part */ 1204 Aloc = (Mat_SeqAIJ*)a->A->data; 1205 m = a->A->m; ai = Aloc->i; aj = Aloc->j; array = Aloc->a; 1206 row = a->rstart; 1207 for (i=0; i<ai[m]+shift; i++) {aj[i] += a->cstart + shift;} 1208 for (i=0; i<m; i++) { 1209 ierr = MatSetValues(B,ai[i+1]-ai[i],aj,1,&row,array,INSERT_VALUES);CHKERRQ(ierr); 1210 row++; array += ai[i+1]-ai[i]; aj += ai[i+1]-ai[i]; 1211 } 1212 aj = Aloc->j; 1213 for (i=0; i<ai[m]+shift; i++) {aj[i] -= a->cstart + shift;} 1214 1215 /* copy over the B part */ 1216 Aloc = (Mat_SeqAIJ*)a->B->data; 1217 m = a->B->m; ai = Aloc->i; aj = Aloc->j; array = Aloc->a; 1218 row = a->rstart; 1219 ierr = PetscMalloc((1+ai[m]-shift)*sizeof(int),&cols);CHKERRQ(ierr); 1220 ct = cols; 1221 for (i=0; i<ai[m]+shift; i++) {cols[i] = a->garray[aj[i]+shift];} 1222 for (i=0; i<m; i++) { 1223 ierr = MatSetValues(B,ai[i+1]-ai[i],cols,1,&row,array,INSERT_VALUES);CHKERRQ(ierr); 1224 row++; array += ai[i+1]-ai[i]; cols += ai[i+1]-ai[i]; 1225 } 1226 ierr = PetscFree(ct);CHKERRQ(ierr); 1227 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1228 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1229 if (matout) { 1230 *matout = B; 1231 } else { 1232 ierr = MatHeaderCopy(A,B);CHKERRQ(ierr); 1233 } 1234 PetscFunctionReturn(0); 1235 } 1236 1237 #undef __FUNCT__ 1238 #define __FUNCT__ "MatDiagonalScale_MPIAIJ" 1239 int MatDiagonalScale_MPIAIJ(Mat mat,Vec ll,Vec rr) 1240 { 1241 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)mat->data; 1242 Mat a = aij->A,b = aij->B; 1243 int ierr,s1,s2,s3; 1244 1245 PetscFunctionBegin; 1246 ierr = MatGetLocalSize(mat,&s2,&s3);CHKERRQ(ierr); 1247 if (rr) { 1248 ierr = VecGetLocalSize(rr,&s1);CHKERRQ(ierr); 1249 if (s1!=s3) SETERRQ(PETSC_ERR_ARG_SIZ,"right vector non-conforming local size"); 1250 /* Overlap communication with computation. */ 1251 ierr = VecScatterBegin(rr,aij->lvec,INSERT_VALUES,SCATTER_FORWARD,aij->Mvctx);CHKERRQ(ierr); 1252 } 1253 if (ll) { 1254 ierr = VecGetLocalSize(ll,&s1);CHKERRQ(ierr); 1255 if (s1!=s2) SETERRQ(PETSC_ERR_ARG_SIZ,"left vector non-conforming local size"); 1256 ierr = (*b->ops->diagonalscale)(b,ll,0);CHKERRQ(ierr); 1257 } 1258 /* scale the diagonal block */ 1259 ierr = (*a->ops->diagonalscale)(a,ll,rr);CHKERRQ(ierr); 1260 1261 if (rr) { 1262 /* Do a scatter end and then right scale the off-diagonal block */ 1263 ierr = VecScatterEnd(rr,aij->lvec,INSERT_VALUES,SCATTER_FORWARD,aij->Mvctx);CHKERRQ(ierr); 1264 ierr = (*b->ops->diagonalscale)(b,0,aij->lvec);CHKERRQ(ierr); 1265 } 1266 1267 PetscFunctionReturn(0); 1268 } 1269 1270 1271 #undef __FUNCT__ 1272 #define __FUNCT__ "MatPrintHelp_MPIAIJ" 1273 int MatPrintHelp_MPIAIJ(Mat A) 1274 { 1275 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1276 int ierr; 1277 1278 PetscFunctionBegin; 1279 if (!a->rank) { 1280 ierr = MatPrintHelp_SeqAIJ(a->A);CHKERRQ(ierr); 1281 } 1282 PetscFunctionReturn(0); 1283 } 1284 1285 #undef __FUNCT__ 1286 #define __FUNCT__ "MatGetBlockSize_MPIAIJ" 1287 int MatGetBlockSize_MPIAIJ(Mat A,int *bs) 1288 { 1289 PetscFunctionBegin; 1290 *bs = 1; 1291 PetscFunctionReturn(0); 1292 } 1293 #undef __FUNCT__ 1294 #define __FUNCT__ "MatSetUnfactored_MPIAIJ" 1295 int MatSetUnfactored_MPIAIJ(Mat A) 1296 { 1297 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1298 int ierr; 1299 1300 PetscFunctionBegin; 1301 ierr = MatSetUnfactored(a->A);CHKERRQ(ierr); 1302 PetscFunctionReturn(0); 1303 } 1304 1305 #undef __FUNCT__ 1306 #define __FUNCT__ "MatEqual_MPIAIJ" 1307 int MatEqual_MPIAIJ(Mat A,Mat B,PetscTruth *flag) 1308 { 1309 Mat_MPIAIJ *matB = (Mat_MPIAIJ*)B->data,*matA = (Mat_MPIAIJ*)A->data; 1310 Mat a,b,c,d; 1311 PetscTruth flg; 1312 int ierr; 1313 1314 PetscFunctionBegin; 1315 ierr = PetscTypeCompare((PetscObject)B,MATMPIAIJ,&flg);CHKERRQ(ierr); 1316 if (!flg) SETERRQ(PETSC_ERR_ARG_INCOMP,"Matrices must be same type"); 1317 a = matA->A; b = matA->B; 1318 c = matB->A; d = matB->B; 1319 1320 ierr = MatEqual(a,c,&flg);CHKERRQ(ierr); 1321 if (flg == PETSC_TRUE) { 1322 ierr = MatEqual(b,d,&flg);CHKERRQ(ierr); 1323 } 1324 ierr = MPI_Allreduce(&flg,flag,1,MPI_INT,MPI_LAND,A->comm);CHKERRQ(ierr); 1325 PetscFunctionReturn(0); 1326 } 1327 1328 #undef __FUNCT__ 1329 #define __FUNCT__ "MatCopy_MPIAIJ" 1330 int MatCopy_MPIAIJ(Mat A,Mat B,MatStructure str) 1331 { 1332 int ierr; 1333 Mat_MPIAIJ *a = (Mat_MPIAIJ *)A->data; 1334 Mat_MPIAIJ *b = (Mat_MPIAIJ *)B->data; 1335 PetscTruth flg; 1336 1337 PetscFunctionBegin; 1338 ierr = PetscTypeCompare((PetscObject)B,MATMPIAIJ,&flg);CHKERRQ(ierr); 1339 if (str != SAME_NONZERO_PATTERN || !flg) { 1340 /* because of the column compression in the off-processor part of the matrix a->B, 1341 the number of columns in a->B and b->B may be different, hence we cannot call 1342 the MatCopy() directly on the two parts. If need be, we can provide a more 1343 efficient copy than the MatCopy_Basic() by first uncompressing the a->B matrices 1344 then copying the submatrices */ 1345 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 1346 } else { 1347 ierr = MatCopy(a->A,b->A,str);CHKERRQ(ierr); 1348 ierr = MatCopy(a->B,b->B,str);CHKERRQ(ierr); 1349 } 1350 PetscFunctionReturn(0); 1351 } 1352 1353 #undef __FUNCT__ 1354 #define __FUNCT__ "MatSetUpPreallocation_MPIAIJ" 1355 int MatSetUpPreallocation_MPIAIJ(Mat A) 1356 { 1357 int ierr; 1358 1359 PetscFunctionBegin; 1360 ierr = MatMPIAIJSetPreallocation(A,PETSC_DEFAULT,0,PETSC_DEFAULT,0);CHKERRQ(ierr); 1361 PetscFunctionReturn(0); 1362 } 1363 1364 EXTERN int MatDuplicate_MPIAIJ(Mat,MatDuplicateOption,Mat *); 1365 EXTERN int MatIncreaseOverlap_MPIAIJ(Mat,int,IS *,int); 1366 EXTERN int MatFDColoringCreate_MPIAIJ(Mat,ISColoring,MatFDColoring); 1367 EXTERN int MatGetSubMatrices_MPIAIJ (Mat,int,IS *,IS *,MatReuse,Mat **); 1368 EXTERN int MatGetSubMatrix_MPIAIJ (Mat,IS,IS,int,MatReuse,Mat *); 1369 #if !defined(PETSC_USE_COMPLEX) && !defined(PETSC_USE_SINGLE) 1370 EXTERN int MatLUFactorSymbolic_MPIAIJ_TFS(Mat,IS,IS,MatLUInfo*,Mat*); 1371 #endif 1372 1373 #include "petscblaslapack.h" 1374 1375 #undef __FUNCT__ 1376 #define __FUNCT__ "MatAXPY_MPIAIJ" 1377 int MatAXPY_MPIAIJ(PetscScalar *a,Mat X,Mat Y,MatStructure str) 1378 { 1379 int ierr,one; 1380 Mat_MPIAIJ *xx = (Mat_MPIAIJ *)X->data,*yy = (Mat_MPIAIJ *)Y->data; 1381 Mat_SeqAIJ *x,*y; 1382 1383 PetscFunctionBegin; 1384 if (str == SAME_NONZERO_PATTERN) { 1385 x = (Mat_SeqAIJ *)xx->A->data; 1386 y = (Mat_SeqAIJ *)yy->A->data; 1387 BLaxpy_(&x->nz,a,x->a,&one,y->a,&one); 1388 x = (Mat_SeqAIJ *)xx->B->data; 1389 y = (Mat_SeqAIJ *)yy->B->data; 1390 BLaxpy_(&x->nz,a,x->a,&one,y->a,&one); 1391 } else { 1392 ierr = MatAXPY_Basic(a,X,Y,str);CHKERRQ(ierr); 1393 } 1394 PetscFunctionReturn(0); 1395 } 1396 1397 /* -------------------------------------------------------------------*/ 1398 static struct _MatOps MatOps_Values = {MatSetValues_MPIAIJ, 1399 MatGetRow_MPIAIJ, 1400 MatRestoreRow_MPIAIJ, 1401 MatMult_MPIAIJ, 1402 MatMultAdd_MPIAIJ, 1403 MatMultTranspose_MPIAIJ, 1404 MatMultTransposeAdd_MPIAIJ, 1405 0, 1406 0, 1407 0, 1408 0, 1409 0, 1410 0, 1411 MatRelax_MPIAIJ, 1412 MatTranspose_MPIAIJ, 1413 MatGetInfo_MPIAIJ, 1414 MatEqual_MPIAIJ, 1415 MatGetDiagonal_MPIAIJ, 1416 MatDiagonalScale_MPIAIJ, 1417 MatNorm_MPIAIJ, 1418 MatAssemblyBegin_MPIAIJ, 1419 MatAssemblyEnd_MPIAIJ, 1420 0, 1421 MatSetOption_MPIAIJ, 1422 MatZeroEntries_MPIAIJ, 1423 MatZeroRows_MPIAIJ, 1424 #if !defined(PETSC_USE_COMPLEX) && !defined(PETSC_USE_SINGLE) 1425 MatLUFactorSymbolic_MPIAIJ_TFS, 1426 #else 1427 0, 1428 #endif 1429 0, 1430 0, 1431 0, 1432 MatSetUpPreallocation_MPIAIJ, 1433 0, 1434 0, 1435 0, 1436 0, 1437 MatDuplicate_MPIAIJ, 1438 0, 1439 0, 1440 0, 1441 0, 1442 MatAXPY_MPIAIJ, 1443 MatGetSubMatrices_MPIAIJ, 1444 MatIncreaseOverlap_MPIAIJ, 1445 MatGetValues_MPIAIJ, 1446 MatCopy_MPIAIJ, 1447 MatPrintHelp_MPIAIJ, 1448 MatScale_MPIAIJ, 1449 0, 1450 0, 1451 0, 1452 MatGetBlockSize_MPIAIJ, 1453 0, 1454 0, 1455 0, 1456 0, 1457 MatFDColoringCreate_MPIAIJ, 1458 0, 1459 MatSetUnfactored_MPIAIJ, 1460 0, 1461 0, 1462 MatGetSubMatrix_MPIAIJ, 1463 MatDestroy_MPIAIJ, 1464 MatView_MPIAIJ, 1465 MatGetPetscMaps_Petsc, 1466 0, 1467 0, 1468 0, 1469 0, 1470 0, 1471 0, 1472 0, 1473 0, 1474 MatSetColoring_MPIAIJ, 1475 MatSetValuesAdic_MPIAIJ, 1476 MatSetValuesAdifor_MPIAIJ 1477 }; 1478 1479 /* ----------------------------------------------------------------------------------------*/ 1480 1481 EXTERN_C_BEGIN 1482 #undef __FUNCT__ 1483 #define __FUNCT__ "MatStoreValues_MPIAIJ" 1484 int MatStoreValues_MPIAIJ(Mat mat) 1485 { 1486 Mat_MPIAIJ *aij = (Mat_MPIAIJ *)mat->data; 1487 int ierr; 1488 1489 PetscFunctionBegin; 1490 ierr = MatStoreValues(aij->A);CHKERRQ(ierr); 1491 ierr = MatStoreValues(aij->B);CHKERRQ(ierr); 1492 PetscFunctionReturn(0); 1493 } 1494 EXTERN_C_END 1495 1496 EXTERN_C_BEGIN 1497 #undef __FUNCT__ 1498 #define __FUNCT__ "MatRetrieveValues_MPIAIJ" 1499 int MatRetrieveValues_MPIAIJ(Mat mat) 1500 { 1501 Mat_MPIAIJ *aij = (Mat_MPIAIJ *)mat->data; 1502 int ierr; 1503 1504 PetscFunctionBegin; 1505 ierr = MatRetrieveValues(aij->A);CHKERRQ(ierr); 1506 ierr = MatRetrieveValues(aij->B);CHKERRQ(ierr); 1507 PetscFunctionReturn(0); 1508 } 1509 EXTERN_C_END 1510 1511 #include "petscpc.h" 1512 EXTERN_C_BEGIN 1513 EXTERN int MatGetDiagonalBlock_MPIAIJ(Mat,PetscTruth *,MatReuse,Mat *); 1514 EXTERN_C_END 1515 1516 EXTERN_C_BEGIN 1517 #undef __FUNCT__ 1518 #define __FUNCT__ "MatCreate_MPIAIJ" 1519 int MatCreate_MPIAIJ(Mat B) 1520 { 1521 Mat_MPIAIJ *b; 1522 int ierr,i,size; 1523 1524 PetscFunctionBegin; 1525 ierr = MPI_Comm_size(B->comm,&size);CHKERRQ(ierr); 1526 1527 ierr = PetscNew(Mat_MPIAIJ,&b);CHKERRQ(ierr); 1528 B->data = (void*)b; 1529 ierr = PetscMemzero(b,sizeof(Mat_MPIAIJ));CHKERRQ(ierr); 1530 ierr = PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 1531 B->factor = 0; 1532 B->assembled = PETSC_FALSE; 1533 B->mapping = 0; 1534 1535 B->insertmode = NOT_SET_VALUES; 1536 b->size = size; 1537 ierr = MPI_Comm_rank(B->comm,&b->rank);CHKERRQ(ierr); 1538 1539 ierr = PetscSplitOwnership(B->comm,&B->m,&B->M);CHKERRQ(ierr); 1540 ierr = PetscSplitOwnership(B->comm,&B->n,&B->N);CHKERRQ(ierr); 1541 1542 /* the information in the maps duplicates the information computed below, eventually 1543 we should remove the duplicate information that is not contained in the maps */ 1544 ierr = PetscMapCreateMPI(B->comm,B->m,B->M,&B->rmap);CHKERRQ(ierr); 1545 ierr = PetscMapCreateMPI(B->comm,B->n,B->N,&B->cmap);CHKERRQ(ierr); 1546 1547 /* build local table of row and column ownerships */ 1548 ierr = PetscMalloc(2*(b->size+2)*sizeof(int),&b->rowners);CHKERRQ(ierr); 1549 PetscLogObjectMemory(B,2*(b->size+2)*sizeof(int)+sizeof(struct _p_Mat)+sizeof(Mat_MPIAIJ)); 1550 b->cowners = b->rowners + b->size + 2; 1551 ierr = MPI_Allgather(&B->m,1,MPI_INT,b->rowners+1,1,MPI_INT,B->comm);CHKERRQ(ierr); 1552 b->rowners[0] = 0; 1553 for (i=2; i<=b->size; i++) { 1554 b->rowners[i] += b->rowners[i-1]; 1555 } 1556 b->rstart = b->rowners[b->rank]; 1557 b->rend = b->rowners[b->rank+1]; 1558 ierr = MPI_Allgather(&B->n,1,MPI_INT,b->cowners+1,1,MPI_INT,B->comm);CHKERRQ(ierr); 1559 b->cowners[0] = 0; 1560 for (i=2; i<=b->size; i++) { 1561 b->cowners[i] += b->cowners[i-1]; 1562 } 1563 b->cstart = b->cowners[b->rank]; 1564 b->cend = b->cowners[b->rank+1]; 1565 1566 /* build cache for off array entries formed */ 1567 ierr = MatStashCreate_Private(B->comm,1,&B->stash);CHKERRQ(ierr); 1568 b->donotstash = PETSC_FALSE; 1569 b->colmap = 0; 1570 b->garray = 0; 1571 b->roworiented = PETSC_TRUE; 1572 1573 /* stuff used for matrix vector multiply */ 1574 b->lvec = PETSC_NULL; 1575 b->Mvctx = PETSC_NULL; 1576 1577 /* stuff for MatGetRow() */ 1578 b->rowindices = 0; 1579 b->rowvalues = 0; 1580 b->getrowactive = PETSC_FALSE; 1581 1582 /* #if defined(PETSC_HAVE_SUPERLUDIST) 1583 1584 ierr = PetscOptionsHasName(PETSC_NULL,"-mat_aij_superlu_dist",&flg);CHKERRQ(ierr); 1585 if (flg) { ierr = MatUseSuperLU_DIST_MPIAIJ(B);CHKERRQ(ierr); } 1586 #endif */ 1587 1588 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatStoreValues_C", 1589 "MatStoreValues_MPIAIJ", 1590 MatStoreValues_MPIAIJ);CHKERRQ(ierr); 1591 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatRetrieveValues_C", 1592 "MatRetrieveValues_MPIAIJ", 1593 MatRetrieveValues_MPIAIJ);CHKERRQ(ierr); 1594 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetDiagonalBlock_C", 1595 "MatGetDiagonalBlock_MPIAIJ", 1596 MatGetDiagonalBlock_MPIAIJ);CHKERRQ(ierr); 1597 1598 PetscFunctionReturn(0); 1599 } 1600 EXTERN_C_END 1601 1602 #undef __FUNCT__ 1603 #define __FUNCT__ "MatDuplicate_MPIAIJ" 1604 int MatDuplicate_MPIAIJ(Mat matin,MatDuplicateOption cpvalues,Mat *newmat) 1605 { 1606 Mat mat; 1607 Mat_MPIAIJ *a,*oldmat = (Mat_MPIAIJ*)matin->data; 1608 int ierr; 1609 1610 PetscFunctionBegin; 1611 *newmat = 0; 1612 ierr = MatCreate(matin->comm,matin->m,matin->n,matin->M,matin->N,&mat);CHKERRQ(ierr); 1613 ierr = MatSetType(mat,MATMPIAIJ);CHKERRQ(ierr); 1614 a = (Mat_MPIAIJ*)mat->data; 1615 ierr = PetscMemcpy(mat->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 1616 mat->factor = matin->factor; 1617 mat->assembled = PETSC_TRUE; 1618 mat->insertmode = NOT_SET_VALUES; 1619 mat->preallocated = PETSC_TRUE; 1620 1621 a->rstart = oldmat->rstart; 1622 a->rend = oldmat->rend; 1623 a->cstart = oldmat->cstart; 1624 a->cend = oldmat->cend; 1625 a->size = oldmat->size; 1626 a->rank = oldmat->rank; 1627 a->donotstash = oldmat->donotstash; 1628 a->roworiented = oldmat->roworiented; 1629 a->rowindices = 0; 1630 a->rowvalues = 0; 1631 a->getrowactive = PETSC_FALSE; 1632 1633 ierr = PetscMemcpy(a->rowners,oldmat->rowners,2*(a->size+2)*sizeof(int));CHKERRQ(ierr); 1634 ierr = MatStashCreate_Private(matin->comm,1,&mat->stash);CHKERRQ(ierr); 1635 if (oldmat->colmap) { 1636 #if defined (PETSC_USE_CTABLE) 1637 ierr = PetscTableCreateCopy(oldmat->colmap,&a->colmap);CHKERRQ(ierr); 1638 #else 1639 ierr = PetscMalloc((mat->N)*sizeof(int),&a->colmap);CHKERRQ(ierr); 1640 PetscLogObjectMemory(mat,(mat->N)*sizeof(int)); 1641 ierr = PetscMemcpy(a->colmap,oldmat->colmap,(mat->N)*sizeof(int));CHKERRQ(ierr); 1642 #endif 1643 } else a->colmap = 0; 1644 if (oldmat->garray) { 1645 int len; 1646 len = oldmat->B->n; 1647 ierr = PetscMalloc((len+1)*sizeof(int),&a->garray);CHKERRQ(ierr); 1648 PetscLogObjectMemory(mat,len*sizeof(int)); 1649 if (len) { ierr = PetscMemcpy(a->garray,oldmat->garray,len*sizeof(int));CHKERRQ(ierr); } 1650 } else a->garray = 0; 1651 1652 ierr = VecDuplicate(oldmat->lvec,&a->lvec);CHKERRQ(ierr); 1653 PetscLogObjectParent(mat,a->lvec); 1654 ierr = VecScatterCopy(oldmat->Mvctx,&a->Mvctx);CHKERRQ(ierr); 1655 PetscLogObjectParent(mat,a->Mvctx); 1656 ierr = MatDuplicate(oldmat->A,cpvalues,&a->A);CHKERRQ(ierr); 1657 PetscLogObjectParent(mat,a->A); 1658 ierr = MatDuplicate(oldmat->B,cpvalues,&a->B);CHKERRQ(ierr); 1659 PetscLogObjectParent(mat,a->B); 1660 ierr = PetscFListDuplicate(matin->qlist,&mat->qlist);CHKERRQ(ierr); 1661 *newmat = mat; 1662 PetscFunctionReturn(0); 1663 } 1664 1665 #include "petscsys.h" 1666 1667 EXTERN_C_BEGIN 1668 #undef __FUNCT__ 1669 #define __FUNCT__ "MatLoad_MPIAIJ" 1670 int MatLoad_MPIAIJ(PetscViewer viewer,MatType type,Mat *newmat) 1671 { 1672 Mat A; 1673 PetscScalar *vals,*svals; 1674 MPI_Comm comm = ((PetscObject)viewer)->comm; 1675 MPI_Status status; 1676 int i,nz,ierr,j,rstart,rend,fd; 1677 int header[4],rank,size,*rowlengths = 0,M,N,m,*rowners,maxnz,*cols; 1678 int *ourlens,*sndcounts = 0,*procsnz = 0,*offlens,jj,*mycols,*smycols; 1679 int tag = ((PetscObject)viewer)->tag,cend,cstart,n; 1680 1681 PetscFunctionBegin; 1682 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 1683 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 1684 if (!rank) { 1685 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 1686 ierr = PetscBinaryRead(fd,(char *)header,4,PETSC_INT);CHKERRQ(ierr); 1687 if (header[0] != MAT_FILE_COOKIE) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"not matrix object"); 1688 if (header[3] < 0) { 1689 SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"Matrix in special format on disk, cannot load as MPIAIJ"); 1690 } 1691 } 1692 1693 ierr = MPI_Bcast(header+1,3,MPI_INT,0,comm);CHKERRQ(ierr); 1694 M = header[1]; N = header[2]; 1695 /* determine ownership of all rows */ 1696 m = M/size + ((M % size) > rank); 1697 ierr = PetscMalloc((size+2)*sizeof(int),&rowners);CHKERRQ(ierr); 1698 ierr = MPI_Allgather(&m,1,MPI_INT,rowners+1,1,MPI_INT,comm);CHKERRQ(ierr); 1699 rowners[0] = 0; 1700 for (i=2; i<=size; i++) { 1701 rowners[i] += rowners[i-1]; 1702 } 1703 rstart = rowners[rank]; 1704 rend = rowners[rank+1]; 1705 1706 /* distribute row lengths to all processors */ 1707 ierr = PetscMalloc(2*(rend-rstart+1)*sizeof(int),&ourlens);CHKERRQ(ierr); 1708 offlens = ourlens + (rend-rstart); 1709 if (!rank) { 1710 ierr = PetscMalloc(M*sizeof(int),&rowlengths);CHKERRQ(ierr); 1711 ierr = PetscBinaryRead(fd,rowlengths,M,PETSC_INT);CHKERRQ(ierr); 1712 ierr = PetscMalloc(size*sizeof(int),&sndcounts);CHKERRQ(ierr); 1713 for (i=0; i<size; i++) sndcounts[i] = rowners[i+1] - rowners[i]; 1714 ierr = MPI_Scatterv(rowlengths,sndcounts,rowners,MPI_INT,ourlens,rend-rstart,MPI_INT,0,comm);CHKERRQ(ierr); 1715 ierr = PetscFree(sndcounts);CHKERRQ(ierr); 1716 } else { 1717 ierr = MPI_Scatterv(0,0,0,MPI_INT,ourlens,rend-rstart,MPI_INT,0,comm);CHKERRQ(ierr); 1718 } 1719 1720 if (!rank) { 1721 /* calculate the number of nonzeros on each processor */ 1722 ierr = PetscMalloc(size*sizeof(int),&procsnz);CHKERRQ(ierr); 1723 ierr = PetscMemzero(procsnz,size*sizeof(int));CHKERRQ(ierr); 1724 for (i=0; i<size; i++) { 1725 for (j=rowners[i]; j< rowners[i+1]; j++) { 1726 procsnz[i] += rowlengths[j]; 1727 } 1728 } 1729 ierr = PetscFree(rowlengths);CHKERRQ(ierr); 1730 1731 /* determine max buffer needed and allocate it */ 1732 maxnz = 0; 1733 for (i=0; i<size; i++) { 1734 maxnz = PetscMax(maxnz,procsnz[i]); 1735 } 1736 ierr = PetscMalloc(maxnz*sizeof(int),&cols);CHKERRQ(ierr); 1737 1738 /* read in my part of the matrix column indices */ 1739 nz = procsnz[0]; 1740 ierr = PetscMalloc(nz*sizeof(int),&mycols);CHKERRQ(ierr); 1741 ierr = PetscBinaryRead(fd,mycols,nz,PETSC_INT);CHKERRQ(ierr); 1742 1743 /* read in every one elses and ship off */ 1744 for (i=1; i<size; i++) { 1745 nz = procsnz[i]; 1746 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 1747 ierr = MPI_Send(cols,nz,MPI_INT,i,tag,comm);CHKERRQ(ierr); 1748 } 1749 ierr = PetscFree(cols);CHKERRQ(ierr); 1750 } else { 1751 /* determine buffer space needed for message */ 1752 nz = 0; 1753 for (i=0; i<m; i++) { 1754 nz += ourlens[i]; 1755 } 1756 ierr = PetscMalloc((nz+1)*sizeof(int),&mycols);CHKERRQ(ierr); 1757 1758 /* receive message of column indices*/ 1759 ierr = MPI_Recv(mycols,nz,MPI_INT,0,tag,comm,&status);CHKERRQ(ierr); 1760 ierr = MPI_Get_count(&status,MPI_INT,&maxnz);CHKERRQ(ierr); 1761 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 1762 } 1763 1764 /* determine column ownership if matrix is not square */ 1765 if (N != M) { 1766 n = N/size + ((N % size) > rank); 1767 ierr = MPI_Scan(&n,&cend,1,MPI_INT,MPI_SUM,comm);CHKERRQ(ierr); 1768 cstart = cend - n; 1769 } else { 1770 cstart = rstart; 1771 cend = rend; 1772 n = cend - cstart; 1773 } 1774 1775 /* loop over local rows, determining number of off diagonal entries */ 1776 ierr = PetscMemzero(offlens,m*sizeof(int));CHKERRQ(ierr); 1777 jj = 0; 1778 for (i=0; i<m; i++) { 1779 for (j=0; j<ourlens[i]; j++) { 1780 if (mycols[jj] < cstart || mycols[jj] >= cend) offlens[i]++; 1781 jj++; 1782 } 1783 } 1784 1785 /* create our matrix */ 1786 for (i=0; i<m; i++) { 1787 ourlens[i] -= offlens[i]; 1788 } 1789 ierr = MatCreateMPIAIJ(comm,m,n,M,N,0,ourlens,0,offlens,newmat);CHKERRQ(ierr); 1790 A = *newmat; 1791 ierr = MatSetOption(A,MAT_COLUMNS_SORTED);CHKERRQ(ierr); 1792 for (i=0; i<m; i++) { 1793 ourlens[i] += offlens[i]; 1794 } 1795 1796 if (!rank) { 1797 ierr = PetscMalloc(maxnz*sizeof(PetscScalar),&vals);CHKERRQ(ierr); 1798 1799 /* read in my part of the matrix numerical values */ 1800 nz = procsnz[0]; 1801 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 1802 1803 /* insert into matrix */ 1804 jj = rstart; 1805 smycols = mycols; 1806 svals = vals; 1807 for (i=0; i<m; i++) { 1808 ierr = MatSetValues(A,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);CHKERRQ(ierr); 1809 smycols += ourlens[i]; 1810 svals += ourlens[i]; 1811 jj++; 1812 } 1813 1814 /* read in other processors and ship out */ 1815 for (i=1; i<size; i++) { 1816 nz = procsnz[i]; 1817 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 1818 ierr = MPI_Send(vals,nz,MPIU_SCALAR,i,A->tag,comm);CHKERRQ(ierr); 1819 } 1820 ierr = PetscFree(procsnz);CHKERRQ(ierr); 1821 } else { 1822 /* receive numeric values */ 1823 ierr = PetscMalloc((nz+1)*sizeof(PetscScalar),&vals);CHKERRQ(ierr); 1824 1825 /* receive message of values*/ 1826 ierr = MPI_Recv(vals,nz,MPIU_SCALAR,0,A->tag,comm,&status);CHKERRQ(ierr); 1827 ierr = MPI_Get_count(&status,MPIU_SCALAR,&maxnz);CHKERRQ(ierr); 1828 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 1829 1830 /* insert into matrix */ 1831 jj = rstart; 1832 smycols = mycols; 1833 svals = vals; 1834 for (i=0; i<m; i++) { 1835 ierr = MatSetValues(A,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);CHKERRQ(ierr); 1836 smycols += ourlens[i]; 1837 svals += ourlens[i]; 1838 jj++; 1839 } 1840 } 1841 ierr = PetscFree(ourlens);CHKERRQ(ierr); 1842 ierr = PetscFree(vals);CHKERRQ(ierr); 1843 ierr = PetscFree(mycols);CHKERRQ(ierr); 1844 ierr = PetscFree(rowners);CHKERRQ(ierr); 1845 1846 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1847 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1848 PetscFunctionReturn(0); 1849 } 1850 EXTERN_C_END 1851 1852 #undef __FUNCT__ 1853 #define __FUNCT__ "MatGetSubMatrix_MPIAIJ" 1854 /* 1855 Not great since it makes two copies of the submatrix, first an SeqAIJ 1856 in local and then by concatenating the local matrices the end result. 1857 Writing it directly would be much like MatGetSubMatrices_MPIAIJ() 1858 */ 1859 int MatGetSubMatrix_MPIAIJ(Mat mat,IS isrow,IS iscol,int csize,MatReuse call,Mat *newmat) 1860 { 1861 int ierr,i,m,n,rstart,row,rend,nz,*cwork,size,rank,j; 1862 int *ii,*jj,nlocal,*dlens,*olens,dlen,olen,jend; 1863 Mat *local,M,Mreuse; 1864 PetscScalar *vwork,*aa; 1865 MPI_Comm comm = mat->comm; 1866 Mat_SeqAIJ *aij; 1867 1868 1869 PetscFunctionBegin; 1870 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 1871 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 1872 1873 if (call == MAT_REUSE_MATRIX) { 1874 ierr = PetscObjectQuery((PetscObject)*newmat,"SubMatrix",(PetscObject *)&Mreuse);CHKERRQ(ierr); 1875 if (!Mreuse) SETERRQ(1,"Submatrix passed in was not used before, cannot reuse"); 1876 local = &Mreuse; 1877 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_REUSE_MATRIX,&local);CHKERRQ(ierr); 1878 } else { 1879 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_INITIAL_MATRIX,&local);CHKERRQ(ierr); 1880 Mreuse = *local; 1881 ierr = PetscFree(local);CHKERRQ(ierr); 1882 } 1883 1884 /* 1885 m - number of local rows 1886 n - number of columns (same on all processors) 1887 rstart - first row in new global matrix generated 1888 */ 1889 ierr = MatGetSize(Mreuse,&m,&n);CHKERRQ(ierr); 1890 if (call == MAT_INITIAL_MATRIX) { 1891 aij = (Mat_SeqAIJ*)(Mreuse)->data; 1892 if (aij->indexshift) SETERRQ(PETSC_ERR_SUP,"No support for index shifted matrix"); 1893 ii = aij->i; 1894 jj = aij->j; 1895 1896 /* 1897 Determine the number of non-zeros in the diagonal and off-diagonal 1898 portions of the matrix in order to do correct preallocation 1899 */ 1900 1901 /* first get start and end of "diagonal" columns */ 1902 if (csize == PETSC_DECIDE) { 1903 nlocal = n/size + ((n % size) > rank); 1904 } else { 1905 nlocal = csize; 1906 } 1907 ierr = MPI_Scan(&nlocal,&rend,1,MPI_INT,MPI_SUM,comm);CHKERRQ(ierr); 1908 rstart = rend - nlocal; 1909 if (rank == size - 1 && rend != n) { 1910 SETERRQ(1,"Local column sizes do not add up to total number of columns"); 1911 } 1912 1913 /* next, compute all the lengths */ 1914 ierr = PetscMalloc((2*m+1)*sizeof(int),&dlens);CHKERRQ(ierr); 1915 olens = dlens + m; 1916 for (i=0; i<m; i++) { 1917 jend = ii[i+1] - ii[i]; 1918 olen = 0; 1919 dlen = 0; 1920 for (j=0; j<jend; j++) { 1921 if (*jj < rstart || *jj >= rend) olen++; 1922 else dlen++; 1923 jj++; 1924 } 1925 olens[i] = olen; 1926 dlens[i] = dlen; 1927 } 1928 ierr = MatCreateMPIAIJ(comm,m,nlocal,PETSC_DECIDE,n,0,dlens,0,olens,&M);CHKERRQ(ierr); 1929 ierr = PetscFree(dlens);CHKERRQ(ierr); 1930 } else { 1931 int ml,nl; 1932 1933 M = *newmat; 1934 ierr = MatGetLocalSize(M,&ml,&nl);CHKERRQ(ierr); 1935 if (ml != m) SETERRQ(PETSC_ERR_ARG_SIZ,"Previous matrix must be same size/layout as request"); 1936 ierr = MatZeroEntries(M);CHKERRQ(ierr); 1937 /* 1938 The next two lines are needed so we may call MatSetValues_MPIAIJ() below directly, 1939 rather than the slower MatSetValues(). 1940 */ 1941 M->was_assembled = PETSC_TRUE; 1942 M->assembled = PETSC_FALSE; 1943 } 1944 ierr = MatGetOwnershipRange(M,&rstart,&rend);CHKERRQ(ierr); 1945 aij = (Mat_SeqAIJ*)(Mreuse)->data; 1946 if (aij->indexshift) SETERRQ(PETSC_ERR_SUP,"No support for index shifted matrix"); 1947 ii = aij->i; 1948 jj = aij->j; 1949 aa = aij->a; 1950 for (i=0; i<m; i++) { 1951 row = rstart + i; 1952 nz = ii[i+1] - ii[i]; 1953 cwork = jj; jj += nz; 1954 vwork = aa; aa += nz; 1955 ierr = MatSetValues_MPIAIJ(M,1,&row,nz,cwork,vwork,INSERT_VALUES);CHKERRQ(ierr); 1956 } 1957 1958 ierr = MatAssemblyBegin(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1959 ierr = MatAssemblyEnd(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1960 *newmat = M; 1961 1962 /* save submatrix used in processor for next request */ 1963 if (call == MAT_INITIAL_MATRIX) { 1964 ierr = PetscObjectCompose((PetscObject)M,"SubMatrix",(PetscObject)Mreuse);CHKERRQ(ierr); 1965 ierr = PetscObjectDereference((PetscObject)Mreuse);CHKERRQ(ierr); 1966 } 1967 1968 PetscFunctionReturn(0); 1969 } 1970 1971 #undef __FUNCT__ 1972 #define __FUNCT__ "MatMPIAIJSetPreallocation" 1973 /*@C 1974 MatMPIAIJSetPreallocation - Creates a sparse parallel matrix in AIJ format 1975 (the default parallel PETSc format). For good matrix assembly performance 1976 the user should preallocate the matrix storage by setting the parameters 1977 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 1978 performance can be increased by more than a factor of 50. 1979 1980 Collective on MPI_Comm 1981 1982 Input Parameters: 1983 + A - the matrix 1984 . d_nz - number of nonzeros per row in DIAGONAL portion of local submatrix 1985 (same value is used for all local rows) 1986 . d_nnz - array containing the number of nonzeros in the various rows of the 1987 DIAGONAL portion of the local submatrix (possibly different for each row) 1988 or PETSC_NULL, if d_nz is used to specify the nonzero structure. 1989 The size of this array is equal to the number of local rows, i.e 'm'. 1990 You must leave room for the diagonal entry even if it is zero. 1991 . o_nz - number of nonzeros per row in the OFF-DIAGONAL portion of local 1992 submatrix (same value is used for all local rows). 1993 - o_nnz - array containing the number of nonzeros in the various rows of the 1994 OFF-DIAGONAL portion of the local submatrix (possibly different for 1995 each row) or PETSC_NULL, if o_nz is used to specify the nonzero 1996 structure. The size of this array is equal to the number 1997 of local rows, i.e 'm'. 1998 1999 The AIJ format (also called the Yale sparse matrix format or 2000 compressed row storage), is fully compatible with standard Fortran 77 2001 storage. That is, the stored row and column indices can begin at 2002 either one (as in Fortran) or zero. See the users manual for details. 2003 2004 The user MUST specify either the local or global matrix dimensions 2005 (possibly both). 2006 2007 The parallel matrix is partitioned such that the first m0 rows belong to 2008 process 0, the next m1 rows belong to process 1, the next m2 rows belong 2009 to process 2 etc.. where m0,m1,m2... are the input parameter 'm'. 2010 2011 The DIAGONAL portion of the local submatrix of a processor can be defined 2012 as the submatrix which is obtained by extraction the part corresponding 2013 to the rows r1-r2 and columns r1-r2 of the global matrix, where r1 is the 2014 first row that belongs to the processor, and r2 is the last row belonging 2015 to the this processor. This is a square mxm matrix. The remaining portion 2016 of the local submatrix (mxN) constitute the OFF-DIAGONAL portion. 2017 2018 If o_nnz, d_nnz are specified, then o_nz, and d_nz are ignored. 2019 2020 By default, this format uses inodes (identical nodes) when possible. 2021 We search for consecutive rows with the same nonzero structure, thereby 2022 reusing matrix information to achieve increased efficiency. 2023 2024 Options Database Keys: 2025 + -mat_aij_no_inode - Do not use inodes 2026 . -mat_aij_inode_limit <limit> - Sets inode limit (max limit=5) 2027 - -mat_aij_oneindex - Internally use indexing starting at 1 2028 rather than 0. Note that when calling MatSetValues(), 2029 the user still MUST index entries starting at 0! 2030 2031 Example usage: 2032 2033 Consider the following 8x8 matrix with 34 non-zero values, that is 2034 assembled across 3 processors. Lets assume that proc0 owns 3 rows, 2035 proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown 2036 as follows: 2037 2038 .vb 2039 1 2 0 | 0 3 0 | 0 4 2040 Proc0 0 5 6 | 7 0 0 | 8 0 2041 9 0 10 | 11 0 0 | 12 0 2042 ------------------------------------- 2043 13 0 14 | 15 16 17 | 0 0 2044 Proc1 0 18 0 | 19 20 21 | 0 0 2045 0 0 0 | 22 23 0 | 24 0 2046 ------------------------------------- 2047 Proc2 25 26 27 | 0 0 28 | 29 0 2048 30 0 0 | 31 32 33 | 0 34 2049 .ve 2050 2051 This can be represented as a collection of submatrices as: 2052 2053 .vb 2054 A B C 2055 D E F 2056 G H I 2057 .ve 2058 2059 Where the submatrices A,B,C are owned by proc0, D,E,F are 2060 owned by proc1, G,H,I are owned by proc2. 2061 2062 The 'm' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 2063 The 'n' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 2064 The 'M','N' parameters are 8,8, and have the same values on all procs. 2065 2066 The DIAGONAL submatrices corresponding to proc0,proc1,proc2 are 2067 submatrices [A], [E], [I] respectively. The OFF-DIAGONAL submatrices 2068 corresponding to proc0,proc1,proc2 are [BC], [DF], [GH] respectively. 2069 Internally, each processor stores the DIAGONAL part, and the OFF-DIAGONAL 2070 part as SeqAIJ matrices. for eg: proc1 will store [E] as a SeqAIJ 2071 matrix, ans [DF] as another SeqAIJ matrix. 2072 2073 When d_nz, o_nz parameters are specified, d_nz storage elements are 2074 allocated for every row of the local diagonal submatrix, and o_nz 2075 storage locations are allocated for every row of the OFF-DIAGONAL submat. 2076 One way to choose d_nz and o_nz is to use the max nonzerors per local 2077 rows for each of the local DIAGONAL, and the OFF-DIAGONAL submatrices. 2078 In this case, the values of d_nz,o_nz are: 2079 .vb 2080 proc0 : dnz = 2, o_nz = 2 2081 proc1 : dnz = 3, o_nz = 2 2082 proc2 : dnz = 1, o_nz = 4 2083 .ve 2084 We are allocating m*(d_nz+o_nz) storage locations for every proc. This 2085 translates to 3*(2+2)=12 for proc0, 3*(3+2)=15 for proc1, 2*(1+4)=10 2086 for proc3. i.e we are using 12+15+10=37 storage locations to store 2087 34 values. 2088 2089 When d_nnz, o_nnz parameters are specified, the storage is specified 2090 for every row, coresponding to both DIAGONAL and OFF-DIAGONAL submatrices. 2091 In the above case the values for d_nnz,o_nnz are: 2092 .vb 2093 proc0: d_nnz = [2,2,2] and o_nnz = [2,2,2] 2094 proc1: d_nnz = [3,3,2] and o_nnz = [2,1,1] 2095 proc2: d_nnz = [1,1] and o_nnz = [4,4] 2096 .ve 2097 Here the space allocated is sum of all the above values i.e 34, and 2098 hence pre-allocation is perfect. 2099 2100 Level: intermediate 2101 2102 .keywords: matrix, aij, compressed row, sparse, parallel 2103 2104 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues() 2105 @*/ 2106 int MatMPIAIJSetPreallocation(Mat B,int d_nz,int *d_nnz,int o_nz,int *o_nnz) 2107 { 2108 Mat_MPIAIJ *b; 2109 int ierr,i; 2110 PetscTruth flg2; 2111 2112 PetscFunctionBegin; 2113 ierr = PetscTypeCompare((PetscObject)B,MATMPIAIJ,&flg2);CHKERRQ(ierr); 2114 if (!flg2) PetscFunctionReturn(0); 2115 B->preallocated = PETSC_TRUE; 2116 if (d_nz == PETSC_DEFAULT || d_nz == PETSC_DECIDE) d_nz = 5; 2117 if (o_nz == PETSC_DEFAULT || o_nz == PETSC_DECIDE) o_nz = 2; 2118 if (d_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"d_nz cannot be less than 0: value %d",d_nz); 2119 if (o_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"o_nz cannot be less than 0: value %d",o_nz); 2120 if (d_nnz) { 2121 for (i=0; i<B->m; i++) { 2122 if (d_nnz[i] < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"d_nnz cannot be less than 0: local row %d value %d",i,d_nnz[i]); 2123 } 2124 } 2125 if (o_nnz) { 2126 for (i=0; i<B->m; i++) { 2127 if (o_nnz[i] < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"o_nnz cannot be less than 0: local row %d value %d",i,o_nnz[i]); 2128 } 2129 } 2130 b = (Mat_MPIAIJ*)B->data; 2131 2132 ierr = MatCreateSeqAIJ(PETSC_COMM_SELF,B->m,B->n,d_nz,d_nnz,&b->A);CHKERRQ(ierr); 2133 PetscLogObjectParent(B,b->A); 2134 ierr = MatCreateSeqAIJ(PETSC_COMM_SELF,B->m,B->N,o_nz,o_nnz,&b->B);CHKERRQ(ierr); 2135 PetscLogObjectParent(B,b->B); 2136 2137 PetscFunctionReturn(0); 2138 } 2139 2140 #undef __FUNCT__ 2141 #define __FUNCT__ "MatCreateMPIAIJ" 2142 /*@C 2143 MatCreateMPIAIJ - Creates a sparse parallel matrix in AIJ format 2144 (the default parallel PETSc format). For good matrix assembly performance 2145 the user should preallocate the matrix storage by setting the parameters 2146 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 2147 performance can be increased by more than a factor of 50. 2148 2149 Collective on MPI_Comm 2150 2151 Input Parameters: 2152 + comm - MPI communicator 2153 . m - number of local rows (or PETSC_DECIDE to have calculated if M is given) 2154 This value should be the same as the local size used in creating the 2155 y vector for the matrix-vector product y = Ax. 2156 . n - This value should be the same as the local size used in creating the 2157 x vector for the matrix-vector product y = Ax. (or PETSC_DECIDE to have 2158 calculated if N is given) For square matrices n is almost always m. 2159 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 2160 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 2161 . d_nz - number of nonzeros per row in DIAGONAL portion of local submatrix 2162 (same value is used for all local rows) 2163 . d_nnz - array containing the number of nonzeros in the various rows of the 2164 DIAGONAL portion of the local submatrix (possibly different for each row) 2165 or PETSC_NULL, if d_nz is used to specify the nonzero structure. 2166 The size of this array is equal to the number of local rows, i.e 'm'. 2167 You must leave room for the diagonal entry even if it is zero. 2168 . o_nz - number of nonzeros per row in the OFF-DIAGONAL portion of local 2169 submatrix (same value is used for all local rows). 2170 - o_nnz - array containing the number of nonzeros in the various rows of the 2171 OFF-DIAGONAL portion of the local submatrix (possibly different for 2172 each row) or PETSC_NULL, if o_nz is used to specify the nonzero 2173 structure. The size of this array is equal to the number 2174 of local rows, i.e 'm'. 2175 2176 Output Parameter: 2177 . A - the matrix 2178 2179 Notes: 2180 m,n,M,N parameters specify the size of the matrix, and its partitioning across 2181 processors, while d_nz,d_nnz,o_nz,o_nnz parameters specify the approximate 2182 storage requirements for this matrix. 2183 2184 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one 2185 processor than it must be used on all processors that share the object for 2186 that argument. 2187 2188 The AIJ format (also called the Yale sparse matrix format or 2189 compressed row storage), is fully compatible with standard Fortran 77 2190 storage. That is, the stored row and column indices can begin at 2191 either one (as in Fortran) or zero. See the users manual for details. 2192 2193 The user MUST specify either the local or global matrix dimensions 2194 (possibly both). 2195 2196 The parallel matrix is partitioned such that the first m0 rows belong to 2197 process 0, the next m1 rows belong to process 1, the next m2 rows belong 2198 to process 2 etc.. where m0,m1,m2... are the input parameter 'm'. 2199 2200 The DIAGONAL portion of the local submatrix of a processor can be defined 2201 as the submatrix which is obtained by extraction the part corresponding 2202 to the rows r1-r2 and columns r1-r2 of the global matrix, where r1 is the 2203 first row that belongs to the processor, and r2 is the last row belonging 2204 to the this processor. This is a square mxm matrix. The remaining portion 2205 of the local submatrix (mxN) constitute the OFF-DIAGONAL portion. 2206 2207 If o_nnz, d_nnz are specified, then o_nz, and d_nz are ignored. 2208 2209 By default, this format uses inodes (identical nodes) when possible. 2210 We search for consecutive rows with the same nonzero structure, thereby 2211 reusing matrix information to achieve increased efficiency. 2212 2213 Options Database Keys: 2214 + -mat_aij_no_inode - Do not use inodes 2215 . -mat_aij_inode_limit <limit> - Sets inode limit (max limit=5) 2216 - -mat_aij_oneindex - Internally use indexing starting at 1 2217 rather than 0. Note that when calling MatSetValues(), 2218 the user still MUST index entries starting at 0! 2219 2220 2221 Example usage: 2222 2223 Consider the following 8x8 matrix with 34 non-zero values, that is 2224 assembled across 3 processors. Lets assume that proc0 owns 3 rows, 2225 proc1 owns 3 rows, proc2 owns 2 rows. This division can be shown 2226 as follows: 2227 2228 .vb 2229 1 2 0 | 0 3 0 | 0 4 2230 Proc0 0 5 6 | 7 0 0 | 8 0 2231 9 0 10 | 11 0 0 | 12 0 2232 ------------------------------------- 2233 13 0 14 | 15 16 17 | 0 0 2234 Proc1 0 18 0 | 19 20 21 | 0 0 2235 0 0 0 | 22 23 0 | 24 0 2236 ------------------------------------- 2237 Proc2 25 26 27 | 0 0 28 | 29 0 2238 30 0 0 | 31 32 33 | 0 34 2239 .ve 2240 2241 This can be represented as a collection of submatrices as: 2242 2243 .vb 2244 A B C 2245 D E F 2246 G H I 2247 .ve 2248 2249 Where the submatrices A,B,C are owned by proc0, D,E,F are 2250 owned by proc1, G,H,I are owned by proc2. 2251 2252 The 'm' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 2253 The 'n' parameters for proc0,proc1,proc2 are 3,3,2 respectively. 2254 The 'M','N' parameters are 8,8, and have the same values on all procs. 2255 2256 The DIAGONAL submatrices corresponding to proc0,proc1,proc2 are 2257 submatrices [A], [E], [I] respectively. The OFF-DIAGONAL submatrices 2258 corresponding to proc0,proc1,proc2 are [BC], [DF], [GH] respectively. 2259 Internally, each processor stores the DIAGONAL part, and the OFF-DIAGONAL 2260 part as SeqAIJ matrices. for eg: proc1 will store [E] as a SeqAIJ 2261 matrix, ans [DF] as another SeqAIJ matrix. 2262 2263 When d_nz, o_nz parameters are specified, d_nz storage elements are 2264 allocated for every row of the local diagonal submatrix, and o_nz 2265 storage locations are allocated for every row of the OFF-DIAGONAL submat. 2266 One way to choose d_nz and o_nz is to use the max nonzerors per local 2267 rows for each of the local DIAGONAL, and the OFF-DIAGONAL submatrices. 2268 In this case, the values of d_nz,o_nz are: 2269 .vb 2270 proc0 : dnz = 2, o_nz = 2 2271 proc1 : dnz = 3, o_nz = 2 2272 proc2 : dnz = 1, o_nz = 4 2273 .ve 2274 We are allocating m*(d_nz+o_nz) storage locations for every proc. This 2275 translates to 3*(2+2)=12 for proc0, 3*(3+2)=15 for proc1, 2*(1+4)=10 2276 for proc3. i.e we are using 12+15+10=37 storage locations to store 2277 34 values. 2278 2279 When d_nnz, o_nnz parameters are specified, the storage is specified 2280 for every row, coresponding to both DIAGONAL and OFF-DIAGONAL submatrices. 2281 In the above case the values for d_nnz,o_nnz are: 2282 .vb 2283 proc0: d_nnz = [2,2,2] and o_nnz = [2,2,2] 2284 proc1: d_nnz = [3,3,2] and o_nnz = [2,1,1] 2285 proc2: d_nnz = [1,1] and o_nnz = [4,4] 2286 .ve 2287 Here the space allocated is sum of all the above values i.e 34, and 2288 hence pre-allocation is perfect. 2289 2290 Level: intermediate 2291 2292 .keywords: matrix, aij, compressed row, sparse, parallel 2293 2294 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues() 2295 @*/ 2296 int MatCreateMPIAIJ(MPI_Comm comm,int m,int n,int M,int N,int d_nz,int *d_nnz,int o_nz,int *o_nnz,Mat *A) 2297 { 2298 int ierr,size; 2299 2300 PetscFunctionBegin; 2301 ierr = MatCreate(comm,m,n,M,N,A);CHKERRQ(ierr); 2302 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 2303 if (size > 1) { 2304 ierr = MatSetType(*A,MATMPIAIJ);CHKERRQ(ierr); 2305 ierr = MatMPIAIJSetPreallocation(*A,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 2306 } else { 2307 ierr = MatSetType(*A,MATSEQAIJ);CHKERRQ(ierr); 2308 ierr = MatSeqAIJSetPreallocation(*A,d_nz,d_nnz);CHKERRQ(ierr); 2309 } 2310 PetscFunctionReturn(0); 2311 } 2312 2313 #undef __FUNCT__ 2314 #define __FUNCT__ "MatMPIAIJGetSeqAIJ" 2315 int MatMPIAIJGetSeqAIJ(Mat A,Mat *Ad,Mat *Ao,int **colmap) 2316 { 2317 Mat_MPIAIJ *a = (Mat_MPIAIJ *)A->data; 2318 PetscFunctionBegin; 2319 *Ad = a->A; 2320 *Ao = a->B; 2321 *colmap = a->garray; 2322 PetscFunctionReturn(0); 2323 } 2324 2325 #undef __FUNCT__ 2326 #define __FUNCT__ "MatSetColoring_MPIAIJ" 2327 int MatSetColoring_MPIAIJ(Mat A,ISColoring coloring) 2328 { 2329 int ierr; 2330 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2331 2332 PetscFunctionBegin; 2333 if (coloring->ctype == IS_COLORING_LOCAL) { 2334 int *allcolors,*colors,i; 2335 ISColoring ocoloring; 2336 2337 /* set coloring for diagonal portion */ 2338 ierr = MatSetColoring_SeqAIJ(a->A,coloring);CHKERRQ(ierr); 2339 2340 /* set coloring for off-diagonal portion */ 2341 ierr = ISAllGatherIndices(A->comm,coloring->n,coloring->colors,PETSC_NULL,&allcolors);CHKERRQ(ierr); 2342 ierr = PetscMalloc((a->B->n+1)*sizeof(int),&colors);CHKERRQ(ierr); 2343 for (i=0; i<a->B->n; i++) { 2344 colors[i] = allcolors[a->garray[i]]; 2345 } 2346 ierr = PetscFree(allcolors);CHKERRQ(ierr); 2347 ierr = ISColoringCreate(MPI_COMM_SELF,a->B->n,colors,&ocoloring);CHKERRQ(ierr); 2348 ierr = MatSetColoring_SeqAIJ(a->B,ocoloring);CHKERRQ(ierr); 2349 ierr = ISColoringDestroy(ocoloring);CHKERRQ(ierr); 2350 } else if (coloring->ctype == IS_COLORING_GHOSTED) { 2351 int *colors,i,*larray; 2352 ISColoring ocoloring; 2353 2354 /* set coloring for diagonal portion */ 2355 ierr = PetscMalloc((a->A->n+1)*sizeof(int),&larray);CHKERRQ(ierr); 2356 for (i=0; i<a->A->n; i++) { 2357 larray[i] = i + a->cstart; 2358 } 2359 ierr = ISGlobalToLocalMappingApply(A->mapping,IS_GTOLM_MASK,a->A->n,larray,PETSC_NULL,larray);CHKERRQ(ierr); 2360 ierr = PetscMalloc((a->A->n+1)*sizeof(int),&colors);CHKERRQ(ierr); 2361 for (i=0; i<a->A->n; i++) { 2362 colors[i] = coloring->colors[larray[i]]; 2363 } 2364 ierr = PetscFree(larray);CHKERRQ(ierr); 2365 ierr = ISColoringCreate(MPI_COMM_SELF,a->A->n,colors,&ocoloring);CHKERRQ(ierr); 2366 ierr = MatSetColoring_SeqAIJ(a->A,ocoloring);CHKERRQ(ierr); 2367 ierr = ISColoringDestroy(ocoloring);CHKERRQ(ierr); 2368 2369 /* set coloring for off-diagonal portion */ 2370 ierr = PetscMalloc((a->B->n+1)*sizeof(int),&larray);CHKERRQ(ierr); 2371 ierr = ISGlobalToLocalMappingApply(A->mapping,IS_GTOLM_MASK,a->B->n,a->garray,PETSC_NULL,larray);CHKERRQ(ierr); 2372 ierr = PetscMalloc((a->B->n+1)*sizeof(int),&colors);CHKERRQ(ierr); 2373 for (i=0; i<a->B->n; i++) { 2374 colors[i] = coloring->colors[larray[i]]; 2375 } 2376 ierr = PetscFree(larray);CHKERRQ(ierr); 2377 ierr = ISColoringCreate(MPI_COMM_SELF,a->B->n,colors,&ocoloring);CHKERRQ(ierr); 2378 ierr = MatSetColoring_SeqAIJ(a->B,ocoloring);CHKERRQ(ierr); 2379 ierr = ISColoringDestroy(ocoloring);CHKERRQ(ierr); 2380 } else { 2381 SETERRQ1(1,"No support ISColoringType %d",coloring->ctype); 2382 } 2383 2384 PetscFunctionReturn(0); 2385 } 2386 2387 #undef __FUNCT__ 2388 #define __FUNCT__ "MatSetValuesAdic_MPIAIJ" 2389 int MatSetValuesAdic_MPIAIJ(Mat A,void *advalues) 2390 { 2391 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2392 int ierr; 2393 2394 PetscFunctionBegin; 2395 ierr = MatSetValuesAdic_SeqAIJ(a->A,advalues);CHKERRQ(ierr); 2396 ierr = MatSetValuesAdic_SeqAIJ(a->B,advalues);CHKERRQ(ierr); 2397 PetscFunctionReturn(0); 2398 } 2399 2400 #undef __FUNCT__ 2401 #define __FUNCT__ "MatSetValuesAdifor_MPIAIJ" 2402 int MatSetValuesAdifor_MPIAIJ(Mat A,int nl,void *advalues) 2403 { 2404 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 2405 int ierr; 2406 2407 PetscFunctionBegin; 2408 ierr = MatSetValuesAdifor_SeqAIJ(a->A,nl,advalues);CHKERRQ(ierr); 2409 ierr = MatSetValuesAdifor_SeqAIJ(a->B,nl,advalues);CHKERRQ(ierr); 2410 PetscFunctionReturn(0); 2411 } 2412