1 2 #include "src/mat/impls/baij/mpi/mpibaij.h" /*I "petscmat.h" I*/ 3 #include "mpisbaij.h" 4 #include "src/mat/impls/sbaij/seq/sbaij.h" 5 6 EXTERN PetscErrorCode MatSetUpMultiply_MPISBAIJ(Mat); 7 EXTERN PetscErrorCode MatSetUpMultiply_MPISBAIJ_2comm(Mat); 8 EXTERN PetscErrorCode DisAssemble_MPISBAIJ(Mat); 9 EXTERN PetscErrorCode MatIncreaseOverlap_MPISBAIJ(Mat,int,IS[],int); 10 EXTERN PetscErrorCode MatGetValues_SeqSBAIJ(Mat,int,const int[],int,const int[],PetscScalar []); 11 EXTERN PetscErrorCode MatGetValues_SeqBAIJ(Mat,int,const int[],int,const int[],PetscScalar []); 12 EXTERN PetscErrorCode MatSetValues_SeqSBAIJ(Mat,int,const int [],int,const int [],const PetscScalar [],InsertMode); 13 EXTERN PetscErrorCode MatSetValuesBlocked_SeqSBAIJ(Mat,int,const int[],int,const int[],const PetscScalar[],InsertMode); 14 EXTERN PetscErrorCode MatSetValuesBlocked_SeqBAIJ(Mat,int,const int[],int,const int[],const PetscScalar[],InsertMode); 15 EXTERN PetscErrorCode MatGetRow_SeqSBAIJ(Mat,int,int*,int**,PetscScalar**); 16 EXTERN PetscErrorCode MatRestoreRow_SeqSBAIJ(Mat,int,int*,int**,PetscScalar**); 17 EXTERN PetscErrorCode MatPrintHelp_SeqSBAIJ(Mat); 18 EXTERN PetscErrorCode MatZeroRows_SeqSBAIJ(Mat,IS,PetscScalar*); 19 EXTERN PetscErrorCode MatZeroRows_SeqBAIJ(Mat,IS,PetscScalar *); 20 EXTERN PetscErrorCode MatGetRowMax_MPISBAIJ(Mat,Vec); 21 EXTERN PetscErrorCode MatRelax_MPISBAIJ(Mat,Vec,PetscReal,MatSORType,PetscReal,int,int,Vec); 22 23 /* UGLY, ugly, ugly 24 When MatScalar == PetscScalar the function MatSetValuesBlocked_MPIBAIJ_MatScalar() does 25 not exist. Otherwise ..._MatScalar() takes matrix elements in single precision and 26 inserts them into the single precision data structure. The function MatSetValuesBlocked_MPIBAIJ() 27 converts the entries into single precision and then calls ..._MatScalar() to put them 28 into the single precision data structures. 29 */ 30 #if defined(PETSC_USE_MAT_SINGLE) 31 EXTERN PetscErrorCode MatSetValuesBlocked_SeqSBAIJ_MatScalar(Mat,int,const int[],int,const int[],const MatScalar[],InsertMode); 32 EXTERN PetscErrorCode MatSetValues_MPISBAIJ_MatScalar(Mat,int,const int[],int,const int[],const MatScalar[],InsertMode); 33 EXTERN PetscErrorCode MatSetValuesBlocked_MPISBAIJ_MatScalar(Mat,int,const int[],int,const int[],const MatScalar[],InsertMode); 34 EXTERN PetscErrorCode MatSetValues_MPISBAIJ_HT_MatScalar(Mat,int,const int[],int,const int[],const MatScalar[],InsertMode); 35 EXTERN PetscErrorCode MatSetValuesBlocked_MPISBAIJ_HT_MatScalar(Mat,int,const int[],int,const int[],const MatScalar[],InsertMode); 36 #else 37 #define MatSetValuesBlocked_SeqSBAIJ_MatScalar MatSetValuesBlocked_SeqSBAIJ 38 #define MatSetValues_MPISBAIJ_MatScalar MatSetValues_MPISBAIJ 39 #define MatSetValuesBlocked_MPISBAIJ_MatScalar MatSetValuesBlocked_MPISBAIJ 40 #define MatSetValues_MPISBAIJ_HT_MatScalar MatSetValues_MPISBAIJ_HT 41 #define MatSetValuesBlocked_MPISBAIJ_HT_MatScalar MatSetValuesBlocked_MPISBAIJ_HT 42 #endif 43 44 EXTERN_C_BEGIN 45 #undef __FUNCT__ 46 #define __FUNCT__ "MatStoreValues_MPISBAIJ" 47 PetscErrorCode MatStoreValues_MPISBAIJ(Mat mat) 48 { 49 Mat_MPISBAIJ *aij = (Mat_MPISBAIJ *)mat->data; 50 PetscErrorCode ierr; 51 52 PetscFunctionBegin; 53 ierr = MatStoreValues(aij->A);CHKERRQ(ierr); 54 ierr = MatStoreValues(aij->B);CHKERRQ(ierr); 55 PetscFunctionReturn(0); 56 } 57 EXTERN_C_END 58 59 EXTERN_C_BEGIN 60 #undef __FUNCT__ 61 #define __FUNCT__ "MatRetrieveValues_MPISBAIJ" 62 PetscErrorCode MatRetrieveValues_MPISBAIJ(Mat mat) 63 { 64 Mat_MPISBAIJ *aij = (Mat_MPISBAIJ *)mat->data; 65 PetscErrorCode ierr; 66 67 PetscFunctionBegin; 68 ierr = MatRetrieveValues(aij->A);CHKERRQ(ierr); 69 ierr = MatRetrieveValues(aij->B);CHKERRQ(ierr); 70 PetscFunctionReturn(0); 71 } 72 EXTERN_C_END 73 74 75 #define CHUNKSIZE 10 76 77 #define MatSetValues_SeqSBAIJ_A_Private(row,col,value,addv) \ 78 { \ 79 \ 80 brow = row/bs; \ 81 rp = aj + ai[brow]; ap = aa + bs2*ai[brow]; \ 82 rmax = aimax[brow]; nrow = ailen[brow]; \ 83 bcol = col/bs; \ 84 ridx = row % bs; cidx = col % bs; \ 85 low = 0; high = nrow; \ 86 while (high-low > 3) { \ 87 t = (low+high)/2; \ 88 if (rp[t] > bcol) high = t; \ 89 else low = t; \ 90 } \ 91 for (_i=low; _i<high; _i++) { \ 92 if (rp[_i] > bcol) break; \ 93 if (rp[_i] == bcol) { \ 94 bap = ap + bs2*_i + bs*cidx + ridx; \ 95 if (addv == ADD_VALUES) *bap += value; \ 96 else *bap = value; \ 97 goto a_noinsert; \ 98 } \ 99 } \ 100 if (a->nonew == 1) goto a_noinsert; \ 101 else if (a->nonew == -1) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%d, %d) into matrix", row, col); \ 102 if (nrow >= rmax) { \ 103 /* there is no extra room in row, therefore enlarge */ \ 104 int new_nz = ai[a->mbs] + CHUNKSIZE,len,*new_i,*new_j; \ 105 MatScalar *new_a; \ 106 \ 107 if (a->nonew == -2) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%d, %d) in the matrix", row, col); \ 108 \ 109 /* malloc new storage space */ \ 110 len = new_nz*(sizeof(int)+bs2*sizeof(MatScalar))+(a->mbs+1)*sizeof(int); \ 111 ierr = PetscMalloc(len,&new_a);CHKERRQ(ierr); \ 112 new_j = (int*)(new_a + bs2*new_nz); \ 113 new_i = new_j + new_nz; \ 114 \ 115 /* copy over old data into new slots */ \ 116 for (ii=0; ii<brow+1; ii++) {new_i[ii] = ai[ii];} \ 117 for (ii=brow+1; ii<a->mbs+1; ii++) {new_i[ii] = ai[ii]+CHUNKSIZE;} \ 118 ierr = PetscMemcpy(new_j,aj,(ai[brow]+nrow)*sizeof(int));CHKERRQ(ierr); \ 119 len = (new_nz - CHUNKSIZE - ai[brow] - nrow); \ 120 ierr = PetscMemcpy(new_j+ai[brow]+nrow+CHUNKSIZE,aj+ai[brow]+nrow,len*sizeof(int));CHKERRQ(ierr); \ 121 ierr = PetscMemcpy(new_a,aa,(ai[brow]+nrow)*bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 122 ierr = PetscMemzero(new_a+bs2*(ai[brow]+nrow),bs2*CHUNKSIZE*sizeof(PetscScalar));CHKERRQ(ierr); \ 123 ierr = PetscMemcpy(new_a+bs2*(ai[brow]+nrow+CHUNKSIZE), \ 124 aa+bs2*(ai[brow]+nrow),bs2*len*sizeof(MatScalar));CHKERRQ(ierr); \ 125 /* free up old matrix storage */ \ 126 ierr = PetscFree(a->a);CHKERRQ(ierr); \ 127 if (!a->singlemalloc) { \ 128 ierr = PetscFree(a->i);CHKERRQ(ierr); \ 129 ierr = PetscFree(a->j);CHKERRQ(ierr);\ 130 } \ 131 aa = a->a = new_a; ai = a->i = new_i; aj = a->j = new_j; \ 132 a->singlemalloc = PETSC_TRUE; \ 133 \ 134 rp = aj + ai[brow]; ap = aa + bs2*ai[brow]; \ 135 rmax = aimax[brow] = aimax[brow] + CHUNKSIZE; \ 136 PetscLogObjectMemory(A,CHUNKSIZE*(sizeof(int) + bs2*sizeof(MatScalar))); \ 137 a->maxnz += bs2*CHUNKSIZE; \ 138 a->reallocs++; \ 139 a->nz++; \ 140 } \ 141 N = nrow++ - 1; \ 142 /* shift up all the later entries in this row */ \ 143 for (ii=N; ii>=_i; ii--) { \ 144 rp[ii+1] = rp[ii]; \ 145 ierr = PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 146 } \ 147 if (N>=_i) { ierr = PetscMemzero(ap+bs2*_i,bs2*sizeof(MatScalar));CHKERRQ(ierr); } \ 148 rp[_i] = bcol; \ 149 ap[bs2*_i + bs*cidx + ridx] = value; \ 150 a_noinsert:; \ 151 ailen[brow] = nrow; \ 152 } 153 #ifndef MatSetValues_SeqBAIJ_B_Private 154 #define MatSetValues_SeqSBAIJ_B_Private(row,col,value,addv) \ 155 { \ 156 brow = row/bs; \ 157 rp = bj + bi[brow]; ap = ba + bs2*bi[brow]; \ 158 rmax = bimax[brow]; nrow = bilen[brow]; \ 159 bcol = col/bs; \ 160 ridx = row % bs; cidx = col % bs; \ 161 low = 0; high = nrow; \ 162 while (high-low > 3) { \ 163 t = (low+high)/2; \ 164 if (rp[t] > bcol) high = t; \ 165 else low = t; \ 166 } \ 167 for (_i=low; _i<high; _i++) { \ 168 if (rp[_i] > bcol) break; \ 169 if (rp[_i] == bcol) { \ 170 bap = ap + bs2*_i + bs*cidx + ridx; \ 171 if (addv == ADD_VALUES) *bap += value; \ 172 else *bap = value; \ 173 goto b_noinsert; \ 174 } \ 175 } \ 176 if (b->nonew == 1) goto b_noinsert; \ 177 else if (b->nonew == -1) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%d, %d) into matrix", row, col); \ 178 if (nrow >= rmax) { \ 179 /* there is no extra room in row, therefore enlarge */ \ 180 int new_nz = bi[b->mbs] + CHUNKSIZE,len,*new_i,*new_j; \ 181 MatScalar *new_a; \ 182 \ 183 if (b->nonew == -2) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%d, %d) in the matrix", row, col); \ 184 \ 185 /* malloc new storage space */ \ 186 len = new_nz*(sizeof(int)+bs2*sizeof(MatScalar))+(b->mbs+1)*sizeof(int); \ 187 ierr = PetscMalloc(len,&new_a);CHKERRQ(ierr); \ 188 new_j = (int*)(new_a + bs2*new_nz); \ 189 new_i = new_j + new_nz; \ 190 \ 191 /* copy over old data into new slots */ \ 192 for (ii=0; ii<brow+1; ii++) {new_i[ii] = bi[ii];} \ 193 for (ii=brow+1; ii<b->mbs+1; ii++) {new_i[ii] = bi[ii]+CHUNKSIZE;} \ 194 ierr = PetscMemcpy(new_j,bj,(bi[brow]+nrow)*sizeof(int));CHKERRQ(ierr); \ 195 len = (new_nz - CHUNKSIZE - bi[brow] - nrow); \ 196 ierr = PetscMemcpy(new_j+bi[brow]+nrow+CHUNKSIZE,bj+bi[brow]+nrow,len*sizeof(int));CHKERRQ(ierr); \ 197 ierr = PetscMemcpy(new_a,ba,(bi[brow]+nrow)*bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 198 ierr = PetscMemzero(new_a+bs2*(bi[brow]+nrow),bs2*CHUNKSIZE*sizeof(MatScalar));CHKERRQ(ierr); \ 199 ierr = PetscMemcpy(new_a+bs2*(bi[brow]+nrow+CHUNKSIZE), \ 200 ba+bs2*(bi[brow]+nrow),bs2*len*sizeof(MatScalar));CHKERRQ(ierr); \ 201 /* free up old matrix storage */ \ 202 ierr = PetscFree(b->a);CHKERRQ(ierr); \ 203 if (!b->singlemalloc) { \ 204 ierr = PetscFree(b->i);CHKERRQ(ierr); \ 205 ierr = PetscFree(b->j);CHKERRQ(ierr); \ 206 } \ 207 ba = b->a = new_a; bi = b->i = new_i; bj = b->j = new_j; \ 208 b->singlemalloc = PETSC_TRUE; \ 209 \ 210 rp = bj + bi[brow]; ap = ba + bs2*bi[brow]; \ 211 rmax = bimax[brow] = bimax[brow] + CHUNKSIZE; \ 212 PetscLogObjectMemory(B,CHUNKSIZE*(sizeof(int) + bs2*sizeof(MatScalar))); \ 213 b->maxnz += bs2*CHUNKSIZE; \ 214 b->reallocs++; \ 215 b->nz++; \ 216 } \ 217 N = nrow++ - 1; \ 218 /* shift up all the later entries in this row */ \ 219 for (ii=N; ii>=_i; ii--) { \ 220 rp[ii+1] = rp[ii]; \ 221 ierr = PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 222 } \ 223 if (N>=_i) { ierr = PetscMemzero(ap+bs2*_i,bs2*sizeof(MatScalar));CHKERRQ(ierr);} \ 224 rp[_i] = bcol; \ 225 ap[bs2*_i + bs*cidx + ridx] = value; \ 226 b_noinsert:; \ 227 bilen[brow] = nrow; \ 228 } 229 #endif 230 231 #if defined(PETSC_USE_MAT_SINGLE) 232 #undef __FUNCT__ 233 #define __FUNCT__ "MatSetValues_MPISBAIJ" 234 PetscErrorCode MatSetValues_MPISBAIJ(Mat mat,int m,const int im[],int n,const int in[],const PetscScalar v[],InsertMode addv) 235 { 236 Mat_MPISBAIJ *b = (Mat_MPISBAIJ*)mat->data; 237 PetscErrorCode ierr; 238 int i,N = m*n; 239 MatScalar *vsingle; 240 241 PetscFunctionBegin; 242 if (N > b->setvalueslen) { 243 if (b->setvaluescopy) {ierr = PetscFree(b->setvaluescopy);CHKERRQ(ierr);} 244 ierr = PetscMalloc(N*sizeof(MatScalar),&b->setvaluescopy);CHKERRQ(ierr); 245 b->setvalueslen = N; 246 } 247 vsingle = b->setvaluescopy; 248 249 for (i=0; i<N; i++) { 250 vsingle[i] = v[i]; 251 } 252 ierr = MatSetValues_MPISBAIJ_MatScalar(mat,m,im,n,in,vsingle,addv);CHKERRQ(ierr); 253 PetscFunctionReturn(0); 254 } 255 256 #undef __FUNCT__ 257 #define __FUNCT__ "MatSetValuesBlocked_MPISBAIJ" 258 PetscErrorCode MatSetValuesBlocked_MPISBAIJ(Mat mat,int m,const int im[],int n,const int in[],const PetscScalar v[],InsertMode addv) 259 { 260 Mat_MPIBAIJ *b = (Mat_MPIBAIJ*)mat->data; 261 PetscErrorCode ierr; 262 int i,N = m*n*b->bs2; 263 MatScalar *vsingle; 264 265 PetscFunctionBegin; 266 if (N > b->setvalueslen) { 267 if (b->setvaluescopy) {ierr = PetscFree(b->setvaluescopy);CHKERRQ(ierr);} 268 ierr = PetscMalloc(N*sizeof(MatScalar),&b->setvaluescopy);CHKERRQ(ierr); 269 b->setvalueslen = N; 270 } 271 vsingle = b->setvaluescopy; 272 for (i=0; i<N; i++) { 273 vsingle[i] = v[i]; 274 } 275 ierr = MatSetValuesBlocked_MPISBAIJ_MatScalar(mat,m,im,n,in,vsingle,addv);CHKERRQ(ierr); 276 PetscFunctionReturn(0); 277 } 278 279 #undef __FUNCT__ 280 #define __FUNCT__ "MatSetValues_MPISBAIJ_HT" 281 PetscErrorCode MatSetValues_MPISBAIJ_HT(Mat mat,int m,const int im[],int n,const int in[],const PetscScalar v[],InsertMode addv) 282 { 283 Mat_MPIBAIJ *b = (Mat_MPIBAIJ*)mat->data; 284 PetscErrorCode ierr; 285 int i,N = m*n; 286 MatScalar *vsingle; 287 288 PetscFunctionBegin; 289 SETERRQ(1,"Function not yet written for SBAIJ format"); 290 /* PetscFunctionReturn(0); */ 291 } 292 293 #undef __FUNCT__ 294 #define __FUNCT__ "MatSetValuesBlocked_MPISBAIJ_HT" 295 PetscErrorCode MatSetValuesBlocked_MPISBAIJ_HT(Mat mat,int m,const int im[],int n,const int in[],const PetscScalar v[],InsertMode addv) 296 { 297 Mat_MPIBAIJ *b = (Mat_MPIBAIJ*)mat->data; 298 PetscErrorCode ierr; 299 int i,N = m*n*b->bs2; 300 MatScalar *vsingle; 301 302 PetscFunctionBegin; 303 SETERRQ(1,"Function not yet written for SBAIJ format"); 304 /* PetscFunctionReturn(0); */ 305 } 306 #endif 307 308 /* Only add/insert a(i,j) with i<=j (blocks). 309 Any a(i,j) with i>j input by user is ingored. 310 */ 311 #undef __FUNCT__ 312 #define __FUNCT__ "MatSetValues_MPIBAIJ_MatScalar" 313 PetscErrorCode MatSetValues_MPISBAIJ_MatScalar(Mat mat,int m,const int im[],int n,const int in[],const MatScalar v[],InsertMode addv) 314 { 315 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 316 MatScalar value; 317 PetscTruth roworiented = baij->roworiented; 318 PetscErrorCode ierr; 319 int i,j,row,col; 320 int rstart_orig=baij->rstart_bs; 321 int rend_orig=baij->rend_bs,cstart_orig=baij->cstart_bs; 322 int cend_orig=baij->cend_bs,bs=baij->bs; 323 324 /* Some Variables required in the macro */ 325 Mat A = baij->A; 326 Mat_SeqSBAIJ *a = (Mat_SeqSBAIJ*)(A)->data; 327 int *aimax=a->imax,*ai=a->i,*ailen=a->ilen,*aj=a->j; 328 MatScalar *aa=a->a; 329 330 Mat B = baij->B; 331 Mat_SeqBAIJ *b = (Mat_SeqBAIJ*)(B)->data; 332 int *bimax=b->imax,*bi=b->i,*bilen=b->ilen,*bj=b->j; 333 MatScalar *ba=b->a; 334 335 int *rp,ii,nrow,_i,rmax,N,brow,bcol; 336 int low,high,t,ridx,cidx,bs2=a->bs2; 337 MatScalar *ap,*bap; 338 339 /* for stash */ 340 int n_loc, *in_loc=0; 341 MatScalar *v_loc=0; 342 343 PetscFunctionBegin; 344 345 if(!baij->donotstash){ 346 ierr = PetscMalloc(n*sizeof(int),&in_loc);CHKERRQ(ierr); 347 ierr = PetscMalloc(n*sizeof(MatScalar),&v_loc);CHKERRQ(ierr); 348 } 349 350 for (i=0; i<m; i++) { 351 if (im[i] < 0) continue; 352 #if defined(PETSC_USE_BOPT_g) 353 if (im[i] >= mat->M) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %d max %d",im[i],mat->M-1); 354 #endif 355 if (im[i] >= rstart_orig && im[i] < rend_orig) { /* this processor entry */ 356 row = im[i] - rstart_orig; /* local row index */ 357 for (j=0; j<n; j++) { 358 if (im[i]/bs > in[j]/bs) continue; /* ignore lower triangular blocks */ 359 if (in[j] >= cstart_orig && in[j] < cend_orig){ /* diag entry (A) */ 360 col = in[j] - cstart_orig; /* local col index */ 361 brow = row/bs; bcol = col/bs; 362 if (brow > bcol) continue; /* ignore lower triangular blocks of A */ 363 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 364 MatSetValues_SeqSBAIJ_A_Private(row,col,value,addv); 365 /* ierr = MatSetValues_SeqBAIJ(baij->A,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 366 } else if (in[j] < 0) continue; 367 #if defined(PETSC_USE_BOPT_g) 368 else if (in[j] >= mat->N) {SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %d max %d",in[j],mat->N-1);} 369 #endif 370 else { /* off-diag entry (B) */ 371 if (mat->was_assembled) { 372 if (!baij->colmap) { 373 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 374 } 375 #if defined (PETSC_USE_CTABLE) 376 ierr = PetscTableFind(baij->colmap,in[j]/bs + 1,&col);CHKERRQ(ierr); 377 col = col - 1; 378 #else 379 col = baij->colmap[in[j]/bs] - 1; 380 #endif 381 if (col < 0 && !((Mat_SeqSBAIJ*)(baij->A->data))->nonew) { 382 ierr = DisAssemble_MPISBAIJ(mat);CHKERRQ(ierr); 383 col = in[j]; 384 /* Reinitialize the variables required by MatSetValues_SeqBAIJ_B_Private() */ 385 B = baij->B; 386 b = (Mat_SeqBAIJ*)(B)->data; 387 bimax=b->imax;bi=b->i;bilen=b->ilen;bj=b->j; 388 ba=b->a; 389 } else col += in[j]%bs; 390 } else col = in[j]; 391 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 392 MatSetValues_SeqSBAIJ_B_Private(row,col,value,addv); 393 /* ierr = MatSetValues_SeqBAIJ(baij->B,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 394 } 395 } 396 } else { /* off processor entry */ 397 if (!baij->donotstash) { 398 n_loc = 0; 399 for (j=0; j<n; j++){ 400 if (im[i]/bs > in[j]/bs) continue; /* ignore lower triangular blocks */ 401 in_loc[n_loc] = in[j]; 402 if (roworiented) { 403 v_loc[n_loc] = v[i*n+j]; 404 } else { 405 v_loc[n_loc] = v[j*m+i]; 406 } 407 n_loc++; 408 } 409 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n_loc,in_loc,v_loc);CHKERRQ(ierr); 410 } 411 } 412 } 413 414 if(!baij->donotstash){ 415 ierr = PetscFree(in_loc);CHKERRQ(ierr); 416 ierr = PetscFree(v_loc);CHKERRQ(ierr); 417 } 418 PetscFunctionReturn(0); 419 } 420 421 #undef __FUNCT__ 422 #define __FUNCT__ "MatSetValuesBlocked_MPISBAIJ_MatScalar" 423 PetscErrorCode MatSetValuesBlocked_MPISBAIJ_MatScalar(Mat mat,int m,const int im[],int n,const int in[],const MatScalar v[],InsertMode addv) 424 { 425 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 426 const MatScalar *value; 427 MatScalar *barray=baij->barray; 428 PetscTruth roworiented = baij->roworiented; 429 PetscErrorCode ierr; 430 int i,j,ii,jj,row,col,rstart=baij->rstart; 431 int rend=baij->rend,cstart=baij->cstart,stepval; 432 int cend=baij->cend,bs=baij->bs,bs2=baij->bs2; 433 434 PetscFunctionBegin; 435 if(!barray) { 436 ierr = PetscMalloc(bs2*sizeof(MatScalar),&barray);CHKERRQ(ierr); 437 baij->barray = barray; 438 } 439 440 if (roworiented) { 441 stepval = (n-1)*bs; 442 } else { 443 stepval = (m-1)*bs; 444 } 445 for (i=0; i<m; i++) { 446 if (im[i] < 0) continue; 447 #if defined(PETSC_USE_BOPT_g) 448 if (im[i] >= baij->Mbs) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large, row %d max %d",im[i],baij->Mbs-1); 449 #endif 450 if (im[i] >= rstart && im[i] < rend) { 451 row = im[i] - rstart; 452 for (j=0; j<n; j++) { 453 /* If NumCol = 1 then a copy is not required */ 454 if ((roworiented) && (n == 1)) { 455 barray = (MatScalar*) v + i*bs2; 456 } else if((!roworiented) && (m == 1)) { 457 barray = (MatScalar*) v + j*bs2; 458 } else { /* Here a copy is required */ 459 if (roworiented) { 460 value = v + i*(stepval+bs)*bs + j*bs; 461 } else { 462 value = v + j*(stepval+bs)*bs + i*bs; 463 } 464 for (ii=0; ii<bs; ii++,value+=stepval) { 465 for (jj=0; jj<bs; jj++) { 466 *barray++ = *value++; 467 } 468 } 469 barray -=bs2; 470 } 471 472 if (in[j] >= cstart && in[j] < cend){ 473 col = in[j] - cstart; 474 ierr = MatSetValuesBlocked_SeqSBAIJ(baij->A,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 475 } 476 else if (in[j] < 0) continue; 477 #if defined(PETSC_USE_BOPT_g) 478 else if (in[j] >= baij->Nbs) {SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large, col %d max %d",in[j],baij->Nbs-1);} 479 #endif 480 else { 481 if (mat->was_assembled) { 482 if (!baij->colmap) { 483 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 484 } 485 486 #if defined(PETSC_USE_BOPT_g) 487 #if defined (PETSC_USE_CTABLE) 488 { int data; 489 ierr = PetscTableFind(baij->colmap,in[j]+1,&data);CHKERRQ(ierr); 490 if ((data - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 491 } 492 #else 493 if ((baij->colmap[in[j]] - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 494 #endif 495 #endif 496 #if defined (PETSC_USE_CTABLE) 497 ierr = PetscTableFind(baij->colmap,in[j]+1,&col);CHKERRQ(ierr); 498 col = (col - 1)/bs; 499 #else 500 col = (baij->colmap[in[j]] - 1)/bs; 501 #endif 502 if (col < 0 && !((Mat_SeqBAIJ*)(baij->A->data))->nonew) { 503 ierr = DisAssemble_MPISBAIJ(mat);CHKERRQ(ierr); 504 col = in[j]; 505 } 506 } 507 else col = in[j]; 508 ierr = MatSetValuesBlocked_SeqBAIJ(baij->B,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 509 } 510 } 511 } else { 512 if (!baij->donotstash) { 513 if (roworiented) { 514 ierr = MatStashValuesRowBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 515 } else { 516 ierr = MatStashValuesColBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 517 } 518 } 519 } 520 } 521 PetscFunctionReturn(0); 522 } 523 524 #define HASH_KEY 0.6180339887 525 #define HASH(size,key,tmp) (tmp = (key)*HASH_KEY,(int)((size)*(tmp-(int)tmp))) 526 /* #define HASH(size,key) ((int)((size)*fmod(((key)*HASH_KEY),1))) */ 527 /* #define HASH(size,key,tmp) ((int)((size)*fmod(((key)*HASH_KEY),1))) */ 528 #undef __FUNCT__ 529 #define __FUNCT__ "MatSetValues_MPISBAIJ_HT_MatScalar" 530 PetscErrorCode MatSetValues_MPISBAIJ_HT_MatScalar(Mat mat,int m,const int im[],int n,const int in[],const MatScalar v[],InsertMode addv) 531 { 532 PetscFunctionBegin; 533 SETERRQ(1,"Function not yet written for SBAIJ format"); 534 /* PetscFunctionReturn(0); */ 535 } 536 537 #undef __FUNCT__ 538 #define __FUNCT__ "MatSetValuesBlocked_MPISBAIJ_HT_MatScalar" 539 PetscErrorCode MatSetValuesBlocked_MPISBAIJ_HT_MatScalar(Mat mat,int m,const int im[],int n,const int in[],const MatScalar v[],InsertMode addv) 540 { 541 PetscFunctionBegin; 542 SETERRQ(1,"Function not yet written for SBAIJ format"); 543 /* PetscFunctionReturn(0); */ 544 } 545 546 #undef __FUNCT__ 547 #define __FUNCT__ "MatGetValues_MPISBAIJ" 548 PetscErrorCode MatGetValues_MPISBAIJ(Mat mat,int m,const int idxm[],int n,const int idxn[],PetscScalar v[]) 549 { 550 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 551 PetscErrorCode ierr; 552 int bs=baij->bs,i,j,bsrstart = baij->rstart*bs,bsrend = baij->rend*bs; 553 int bscstart = baij->cstart*bs,bscend = baij->cend*bs,row,col,data; 554 555 PetscFunctionBegin; 556 for (i=0; i<m; i++) { 557 if (idxm[i] < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Negative row: %d",idxm[i]); 558 if (idxm[i] >= mat->M) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %d max %d",idxm[i],mat->M-1); 559 if (idxm[i] >= bsrstart && idxm[i] < bsrend) { 560 row = idxm[i] - bsrstart; 561 for (j=0; j<n; j++) { 562 if (idxn[j] < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Negative column %d",idxn[j]); 563 if (idxn[j] >= mat->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %d max %d",idxn[j],mat->N-1); 564 if (idxn[j] >= bscstart && idxn[j] < bscend){ 565 col = idxn[j] - bscstart; 566 ierr = MatGetValues_SeqSBAIJ(baij->A,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 567 } else { 568 if (!baij->colmap) { 569 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 570 } 571 #if defined (PETSC_USE_CTABLE) 572 ierr = PetscTableFind(baij->colmap,idxn[j]/bs+1,&data);CHKERRQ(ierr); 573 data --; 574 #else 575 data = baij->colmap[idxn[j]/bs]-1; 576 #endif 577 if((data < 0) || (baij->garray[data/bs] != idxn[j]/bs)) *(v+i*n+j) = 0.0; 578 else { 579 col = data + idxn[j]%bs; 580 ierr = MatGetValues_SeqBAIJ(baij->B,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 581 } 582 } 583 } 584 } else { 585 SETERRQ(PETSC_ERR_SUP,"Only local values currently supported"); 586 } 587 } 588 PetscFunctionReturn(0); 589 } 590 591 #undef __FUNCT__ 592 #define __FUNCT__ "MatNorm_MPISBAIJ" 593 PetscErrorCode MatNorm_MPISBAIJ(Mat mat,NormType type,PetscReal *norm) 594 { 595 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 596 /* Mat_SeqSBAIJ *amat = (Mat_SeqSBAIJ*)baij->A->data; */ 597 /* Mat_SeqBAIJ *bmat = (Mat_SeqBAIJ*)baij->B->data; */ 598 PetscErrorCode ierr; 599 PetscReal sum[2],*lnorm2; 600 601 PetscFunctionBegin; 602 if (baij->size == 1) { 603 ierr = MatNorm(baij->A,type,norm);CHKERRQ(ierr); 604 } else { 605 if (type == NORM_FROBENIUS) { 606 ierr = PetscMalloc(2*sizeof(PetscReal),&lnorm2);CHKERRQ(ierr); 607 ierr = MatNorm(baij->A,type,lnorm2);CHKERRQ(ierr); 608 *lnorm2 = (*lnorm2)*(*lnorm2); lnorm2++; /* squar power of norm(A) */ 609 ierr = MatNorm(baij->B,type,lnorm2);CHKERRQ(ierr); 610 *lnorm2 = (*lnorm2)*(*lnorm2); lnorm2--; /* squar power of norm(B) */ 611 /* 612 ierr = MPI_Comm_rank(mat->comm,&rank);CHKERRQ(ierr); 613 PetscSynchronizedPrintf(mat->comm,"[%d], lnorm2=%g, %g\n",rank,lnorm2[0],lnorm2[1]); 614 */ 615 ierr = MPI_Allreduce(lnorm2,&sum,2,MPIU_REAL,MPI_SUM,mat->comm);CHKERRQ(ierr); 616 /* 617 PetscSynchronizedPrintf(mat->comm,"[%d], sum=%g, %g\n",rank,sum[0],sum[1]); 618 PetscSynchronizedFlush(mat->comm); */ 619 620 *norm = sqrt(sum[0] + 2*sum[1]); 621 ierr = PetscFree(lnorm2);CHKERRQ(ierr); 622 } else { 623 SETERRQ(PETSC_ERR_SUP,"No support for this norm yet"); 624 } 625 } 626 PetscFunctionReturn(0); 627 } 628 629 /* 630 Creates the hash table, and sets the table 631 This table is created only once. 632 If new entried need to be added to the matrix 633 then the hash table has to be destroyed and 634 recreated. 635 */ 636 #undef __FUNCT__ 637 #define __FUNCT__ "MatCreateHashTable_MPISBAIJ_Private" 638 PetscErrorCode MatCreateHashTable_MPISBAIJ_Private(Mat mat,PetscReal factor) 639 { 640 PetscFunctionBegin; 641 SETERRQ(1,"Function not yet written for SBAIJ format"); 642 /* PetscFunctionReturn(0); */ 643 } 644 645 #undef __FUNCT__ 646 #define __FUNCT__ "MatAssemblyBegin_MPISBAIJ" 647 PetscErrorCode MatAssemblyBegin_MPISBAIJ(Mat mat,MatAssemblyType mode) 648 { 649 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 650 PetscErrorCode ierr; 651 int nstash,reallocs; 652 InsertMode addv; 653 654 PetscFunctionBegin; 655 if (baij->donotstash) { 656 PetscFunctionReturn(0); 657 } 658 659 /* make sure all processors are either in INSERTMODE or ADDMODE */ 660 ierr = MPI_Allreduce(&mat->insertmode,&addv,1,MPI_INT,MPI_BOR,mat->comm);CHKERRQ(ierr); 661 if (addv == (ADD_VALUES|INSERT_VALUES)) { 662 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Some processors inserted others added"); 663 } 664 mat->insertmode = addv; /* in case this processor had no cache */ 665 666 ierr = MatStashScatterBegin_Private(&mat->stash,baij->rowners_bs);CHKERRQ(ierr); 667 ierr = MatStashScatterBegin_Private(&mat->bstash,baij->rowners);CHKERRQ(ierr); 668 ierr = MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);CHKERRQ(ierr); 669 PetscLogInfo(0,"MatAssemblyBegin_MPISBAIJ:Stash has %d entries,uses %d mallocs.\n",nstash,reallocs); 670 ierr = MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);CHKERRQ(ierr); 671 PetscLogInfo(0,"MatAssemblyBegin_MPISBAIJ:Block-Stash has %d entries, uses %d mallocs.\n",nstash,reallocs); 672 PetscFunctionReturn(0); 673 } 674 675 #undef __FUNCT__ 676 #define __FUNCT__ "MatAssemblyEnd_MPISBAIJ" 677 PetscErrorCode MatAssemblyEnd_MPISBAIJ(Mat mat,MatAssemblyType mode) 678 { 679 Mat_MPISBAIJ *baij=(Mat_MPISBAIJ*)mat->data; 680 Mat_SeqSBAIJ *a=(Mat_SeqSBAIJ*)baij->A->data; 681 Mat_SeqBAIJ *b=(Mat_SeqBAIJ*)baij->B->data; 682 PetscErrorCode ierr; 683 int i,j,rstart,ncols,n,flg,bs2=baij->bs2; 684 int *row,*col,other_disassembled; 685 PetscTruth r1,r2,r3; 686 MatScalar *val; 687 InsertMode addv = mat->insertmode; 688 689 PetscFunctionBegin; 690 691 if (!baij->donotstash) { 692 while (1) { 693 ierr = MatStashScatterGetMesg_Private(&mat->stash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 694 /* 695 PetscSynchronizedPrintf(mat->comm,"[%d]: in AssemblyEnd, stash, flg=%d\n",rank,flg); 696 PetscSynchronizedFlush(mat->comm); 697 */ 698 if (!flg) break; 699 700 for (i=0; i<n;) { 701 /* Now identify the consecutive vals belonging to the same row */ 702 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 703 if (j < n) ncols = j-i; 704 else ncols = n-i; 705 /* Now assemble all these values with a single function call */ 706 ierr = MatSetValues_MPISBAIJ_MatScalar(mat,1,row+i,ncols,col+i,val+i,addv);CHKERRQ(ierr); 707 i = j; 708 } 709 } 710 ierr = MatStashScatterEnd_Private(&mat->stash);CHKERRQ(ierr); 711 /* Now process the block-stash. Since the values are stashed column-oriented, 712 set the roworiented flag to column oriented, and after MatSetValues() 713 restore the original flags */ 714 r1 = baij->roworiented; 715 r2 = a->roworiented; 716 r3 = b->roworiented; 717 baij->roworiented = PETSC_FALSE; 718 a->roworiented = PETSC_FALSE; 719 b->roworiented = PETSC_FALSE; 720 while (1) { 721 ierr = MatStashScatterGetMesg_Private(&mat->bstash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 722 if (!flg) break; 723 724 for (i=0; i<n;) { 725 /* Now identify the consecutive vals belonging to the same row */ 726 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 727 if (j < n) ncols = j-i; 728 else ncols = n-i; 729 ierr = MatSetValuesBlocked_MPISBAIJ_MatScalar(mat,1,row+i,ncols,col+i,val+i*bs2,addv);CHKERRQ(ierr); 730 i = j; 731 } 732 } 733 ierr = MatStashScatterEnd_Private(&mat->bstash);CHKERRQ(ierr); 734 baij->roworiented = r1; 735 a->roworiented = r2; 736 b->roworiented = r3; 737 } 738 739 ierr = MatAssemblyBegin(baij->A,mode);CHKERRQ(ierr); 740 ierr = MatAssemblyEnd(baij->A,mode);CHKERRQ(ierr); 741 742 /* determine if any processor has disassembled, if so we must 743 also disassemble ourselfs, in order that we may reassemble. */ 744 /* 745 if nonzero structure of submatrix B cannot change then we know that 746 no processor disassembled thus we can skip this stuff 747 */ 748 if (!((Mat_SeqBAIJ*)baij->B->data)->nonew) { 749 ierr = MPI_Allreduce(&mat->was_assembled,&other_disassembled,1,MPI_INT,MPI_PROD,mat->comm);CHKERRQ(ierr); 750 if (mat->was_assembled && !other_disassembled) { 751 ierr = DisAssemble_MPISBAIJ(mat);CHKERRQ(ierr); 752 } 753 } 754 755 if (!mat->was_assembled && mode == MAT_FINAL_ASSEMBLY) { 756 ierr = MatSetUpMultiply_MPISBAIJ(mat);CHKERRQ(ierr); /* setup Mvctx and sMvctx */ 757 } 758 ierr = MatAssemblyBegin(baij->B,mode);CHKERRQ(ierr); 759 ierr = MatAssemblyEnd(baij->B,mode);CHKERRQ(ierr); 760 761 #if defined(PETSC_USE_BOPT_g) 762 if (baij->ht && mode== MAT_FINAL_ASSEMBLY) { 763 PetscLogInfo(0,"MatAssemblyEnd_MPISBAIJ:Average Hash Table Search in MatSetValues = %5.2f\n",((PetscReal)baij->ht_total_ct)/baij->ht_insert_ct); 764 baij->ht_total_ct = 0; 765 baij->ht_insert_ct = 0; 766 } 767 #endif 768 if (baij->ht_flag && !baij->ht && mode == MAT_FINAL_ASSEMBLY) { 769 ierr = MatCreateHashTable_MPISBAIJ_Private(mat,baij->ht_fact);CHKERRQ(ierr); 770 mat->ops->setvalues = MatSetValues_MPISBAIJ_HT; 771 mat->ops->setvaluesblocked = MatSetValuesBlocked_MPISBAIJ_HT; 772 } 773 774 if (baij->rowvalues) { 775 ierr = PetscFree(baij->rowvalues);CHKERRQ(ierr); 776 baij->rowvalues = 0; 777 } 778 779 PetscFunctionReturn(0); 780 } 781 782 #undef __FUNCT__ 783 #define __FUNCT__ "MatView_MPISBAIJ_ASCIIorDraworSocket" 784 static PetscErrorCode MatView_MPISBAIJ_ASCIIorDraworSocket(Mat mat,PetscViewer viewer) 785 { 786 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 787 PetscErrorCode ierr; 788 int bs = baij->bs,size = baij->size,rank = baij->rank; 789 PetscTruth iascii,isdraw; 790 PetscViewer sviewer; 791 PetscViewerFormat format; 792 793 PetscFunctionBegin; 794 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr); 795 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 796 if (iascii) { 797 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 798 if (format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 799 MatInfo info; 800 ierr = MPI_Comm_rank(mat->comm,&rank);CHKERRQ(ierr); 801 ierr = MatGetInfo(mat,MAT_LOCAL,&info);CHKERRQ(ierr); 802 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %d nz %d nz alloced %d bs %d mem %d\n", 803 rank,mat->m,(int)info.nz_used*bs,(int)info.nz_allocated*bs, 804 baij->bs,(int)info.memory);CHKERRQ(ierr); 805 ierr = MatGetInfo(baij->A,MAT_LOCAL,&info);CHKERRQ(ierr); 806 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] on-diagonal part: nz %d \n",rank,(int)info.nz_used*bs);CHKERRQ(ierr); 807 ierr = MatGetInfo(baij->B,MAT_LOCAL,&info);CHKERRQ(ierr); 808 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] off-diagonal part: nz %d \n",rank,(int)info.nz_used*bs);CHKERRQ(ierr); 809 ierr = PetscViewerFlush(viewer);CHKERRQ(ierr); 810 ierr = VecScatterView(baij->Mvctx,viewer);CHKERRQ(ierr); 811 PetscFunctionReturn(0); 812 } else if (format == PETSC_VIEWER_ASCII_INFO) { 813 ierr = PetscViewerASCIIPrintf(viewer," block size is %d\n",bs);CHKERRQ(ierr); 814 PetscFunctionReturn(0); 815 } 816 } 817 818 if (isdraw) { 819 PetscDraw draw; 820 PetscTruth isnull; 821 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 822 ierr = PetscDrawIsNull(draw,&isnull);CHKERRQ(ierr); if (isnull) PetscFunctionReturn(0); 823 } 824 825 if (size == 1) { 826 ierr = PetscObjectSetName((PetscObject)baij->A,mat->name);CHKERRQ(ierr); 827 ierr = MatView(baij->A,viewer);CHKERRQ(ierr); 828 } else { 829 /* assemble the entire matrix onto first processor. */ 830 Mat A; 831 Mat_SeqSBAIJ *Aloc; 832 Mat_SeqBAIJ *Bloc; 833 int M = mat->M,N = mat->N,*ai,*aj,col,i,j,k,*rvals,mbs = baij->mbs; 834 MatScalar *a; 835 836 /* Should this be the same type as mat? */ 837 if (!rank) { 838 ierr = MatCreate(mat->comm,M,N,M,N,&A);CHKERRQ(ierr); 839 } else { 840 ierr = MatCreate(mat->comm,0,0,M,N,&A);CHKERRQ(ierr); 841 } 842 ierr = MatSetType(A,MATMPISBAIJ);CHKERRQ(ierr); 843 ierr = MatMPISBAIJSetPreallocation(A,baij->bs,0,PETSC_NULL,0,PETSC_NULL);CHKERRQ(ierr); 844 PetscLogObjectParent(mat,A); 845 846 /* copy over the A part */ 847 Aloc = (Mat_SeqSBAIJ*)baij->A->data; 848 ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 849 ierr = PetscMalloc(bs*sizeof(int),&rvals);CHKERRQ(ierr); 850 851 for (i=0; i<mbs; i++) { 852 rvals[0] = bs*(baij->rstart + i); 853 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 854 for (j=ai[i]; j<ai[i+1]; j++) { 855 col = (baij->cstart+aj[j])*bs; 856 for (k=0; k<bs; k++) { 857 ierr = MatSetValues_MPISBAIJ_MatScalar(A,bs,rvals,1,&col,a,INSERT_VALUES);CHKERRQ(ierr); 858 col++; a += bs; 859 } 860 } 861 } 862 /* copy over the B part */ 863 Bloc = (Mat_SeqBAIJ*)baij->B->data; 864 ai = Bloc->i; aj = Bloc->j; a = Bloc->a; 865 for (i=0; i<mbs; i++) { 866 rvals[0] = bs*(baij->rstart + i); 867 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 868 for (j=ai[i]; j<ai[i+1]; j++) { 869 col = baij->garray[aj[j]]*bs; 870 for (k=0; k<bs; k++) { 871 ierr = MatSetValues_MPISBAIJ_MatScalar(A,bs,rvals,1,&col,a,INSERT_VALUES);CHKERRQ(ierr); 872 col++; a += bs; 873 } 874 } 875 } 876 ierr = PetscFree(rvals);CHKERRQ(ierr); 877 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 878 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 879 /* 880 Everyone has to call to draw the matrix since the graphics waits are 881 synchronized across all processors that share the PetscDraw object 882 */ 883 ierr = PetscViewerGetSingleton(viewer,&sviewer);CHKERRQ(ierr); 884 if (!rank) { 885 ierr = PetscObjectSetName((PetscObject)((Mat_MPISBAIJ*)(A->data))->A,mat->name);CHKERRQ(ierr); 886 ierr = MatView(((Mat_MPISBAIJ*)(A->data))->A,sviewer);CHKERRQ(ierr); 887 } 888 ierr = PetscViewerRestoreSingleton(viewer,&sviewer);CHKERRQ(ierr); 889 ierr = MatDestroy(A);CHKERRQ(ierr); 890 } 891 PetscFunctionReturn(0); 892 } 893 894 #undef __FUNCT__ 895 #define __FUNCT__ "MatView_MPISBAIJ" 896 PetscErrorCode MatView_MPISBAIJ(Mat mat,PetscViewer viewer) 897 { 898 PetscErrorCode ierr; 899 PetscTruth iascii,isdraw,issocket,isbinary; 900 901 PetscFunctionBegin; 902 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr); 903 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 904 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_SOCKET,&issocket);CHKERRQ(ierr); 905 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_BINARY,&isbinary);CHKERRQ(ierr); 906 if (iascii || isdraw || issocket || isbinary) { 907 ierr = MatView_MPISBAIJ_ASCIIorDraworSocket(mat,viewer);CHKERRQ(ierr); 908 } else { 909 SETERRQ1(1,"Viewer type %s not supported by MPISBAIJ matrices",((PetscObject)viewer)->type_name); 910 } 911 PetscFunctionReturn(0); 912 } 913 914 #undef __FUNCT__ 915 #define __FUNCT__ "MatDestroy_MPISBAIJ" 916 PetscErrorCode MatDestroy_MPISBAIJ(Mat mat) 917 { 918 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 919 PetscErrorCode ierr; 920 921 PetscFunctionBegin; 922 #if defined(PETSC_USE_LOG) 923 PetscLogObjectState((PetscObject)mat,"Rows=%d,Cols=%d",mat->M,mat->N); 924 #endif 925 ierr = MatStashDestroy_Private(&mat->stash);CHKERRQ(ierr); 926 ierr = MatStashDestroy_Private(&mat->bstash);CHKERRQ(ierr); 927 ierr = PetscFree(baij->rowners);CHKERRQ(ierr); 928 ierr = MatDestroy(baij->A);CHKERRQ(ierr); 929 ierr = MatDestroy(baij->B);CHKERRQ(ierr); 930 #if defined (PETSC_USE_CTABLE) 931 if (baij->colmap) {ierr = PetscTableDelete(baij->colmap);CHKERRQ(ierr);} 932 #else 933 if (baij->colmap) {ierr = PetscFree(baij->colmap);CHKERRQ(ierr);} 934 #endif 935 if (baij->garray) {ierr = PetscFree(baij->garray);CHKERRQ(ierr);} 936 if (baij->lvec) {ierr = VecDestroy(baij->lvec);CHKERRQ(ierr);} 937 if (baij->Mvctx) {ierr = VecScatterDestroy(baij->Mvctx);CHKERRQ(ierr);} 938 if (baij->slvec0) { 939 ierr = VecDestroy(baij->slvec0);CHKERRQ(ierr); 940 ierr = VecDestroy(baij->slvec0b);CHKERRQ(ierr); 941 } 942 if (baij->slvec1) { 943 ierr = VecDestroy(baij->slvec1);CHKERRQ(ierr); 944 ierr = VecDestroy(baij->slvec1a);CHKERRQ(ierr); 945 ierr = VecDestroy(baij->slvec1b);CHKERRQ(ierr); 946 } 947 if (baij->sMvctx) {ierr = VecScatterDestroy(baij->sMvctx);CHKERRQ(ierr);} 948 if (baij->rowvalues) {ierr = PetscFree(baij->rowvalues);CHKERRQ(ierr);} 949 if (baij->barray) {ierr = PetscFree(baij->barray);CHKERRQ(ierr);} 950 if (baij->hd) {ierr = PetscFree(baij->hd);CHKERRQ(ierr);} 951 #if defined(PETSC_USE_MAT_SINGLE) 952 if (baij->setvaluescopy) {ierr = PetscFree(baij->setvaluescopy);CHKERRQ(ierr);} 953 #endif 954 ierr = PetscFree(baij);CHKERRQ(ierr); 955 PetscFunctionReturn(0); 956 } 957 958 #undef __FUNCT__ 959 #define __FUNCT__ "MatMult_MPISBAIJ" 960 PetscErrorCode MatMult_MPISBAIJ(Mat A,Vec xx,Vec yy) 961 { 962 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 963 PetscErrorCode ierr; 964 int nt,mbs=a->mbs,bs=a->bs; 965 PetscScalar *x,*from,zero=0.0; 966 967 PetscFunctionBegin; 968 ierr = VecGetLocalSize(xx,&nt);CHKERRQ(ierr); 969 if (nt != A->n) { 970 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible partition of A and xx"); 971 } 972 ierr = VecGetLocalSize(yy,&nt);CHKERRQ(ierr); 973 if (nt != A->m) { 974 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible parition of A and yy"); 975 } 976 977 /* diagonal part */ 978 ierr = (*a->A->ops->mult)(a->A,xx,a->slvec1a);CHKERRQ(ierr); 979 ierr = VecSet(&zero,a->slvec1b);CHKERRQ(ierr); 980 981 /* subdiagonal part */ 982 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->slvec0b);CHKERRQ(ierr); 983 984 /* copy x into the vec slvec0 */ 985 ierr = VecGetArray(a->slvec0,&from);CHKERRQ(ierr); 986 ierr = VecGetArray(xx,&x);CHKERRQ(ierr); 987 ierr = PetscMemcpy(from,x,bs*mbs*sizeof(MatScalar));CHKERRQ(ierr); 988 ierr = VecRestoreArray(a->slvec0,&from);CHKERRQ(ierr); 989 990 ierr = VecScatterBegin(a->slvec0,a->slvec1,ADD_VALUES,SCATTER_FORWARD,a->sMvctx);CHKERRQ(ierr); 991 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 992 ierr = VecScatterEnd(a->slvec0,a->slvec1,ADD_VALUES,SCATTER_FORWARD,a->sMvctx);CHKERRQ(ierr); 993 994 /* supperdiagonal part */ 995 ierr = (*a->B->ops->multadd)(a->B,a->slvec1b,a->slvec1a,yy);CHKERRQ(ierr); 996 997 PetscFunctionReturn(0); 998 } 999 1000 #undef __FUNCT__ 1001 #define __FUNCT__ "MatMult_MPISBAIJ_2comm" 1002 PetscErrorCode MatMult_MPISBAIJ_2comm(Mat A,Vec xx,Vec yy) 1003 { 1004 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1005 PetscErrorCode ierr; 1006 int nt; 1007 1008 PetscFunctionBegin; 1009 ierr = VecGetLocalSize(xx,&nt);CHKERRQ(ierr); 1010 if (nt != A->n) { 1011 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible partition of A and xx"); 1012 } 1013 ierr = VecGetLocalSize(yy,&nt);CHKERRQ(ierr); 1014 if (nt != A->m) { 1015 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible parition of A and yy"); 1016 } 1017 1018 ierr = VecScatterBegin(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 1019 /* do diagonal part */ 1020 ierr = (*a->A->ops->mult)(a->A,xx,yy);CHKERRQ(ierr); 1021 /* do supperdiagonal part */ 1022 ierr = VecScatterEnd(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 1023 ierr = (*a->B->ops->multadd)(a->B,a->lvec,yy,yy);CHKERRQ(ierr); 1024 /* do subdiagonal part */ 1025 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1026 ierr = VecScatterBegin(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 1027 ierr = VecScatterEnd(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 1028 1029 PetscFunctionReturn(0); 1030 } 1031 1032 #undef __FUNCT__ 1033 #define __FUNCT__ "MatMultAdd_MPISBAIJ" 1034 PetscErrorCode MatMultAdd_MPISBAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1035 { 1036 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1037 PetscErrorCode ierr; 1038 int mbs=a->mbs,bs=a->bs; 1039 PetscScalar *x,*from,zero=0.0; 1040 1041 PetscFunctionBegin; 1042 /* 1043 PetscSynchronizedPrintf(A->comm," MatMultAdd is called ...\n"); 1044 PetscSynchronizedFlush(A->comm); 1045 */ 1046 /* diagonal part */ 1047 ierr = (*a->A->ops->multadd)(a->A,xx,yy,a->slvec1a);CHKERRQ(ierr); 1048 ierr = VecSet(&zero,a->slvec1b);CHKERRQ(ierr); 1049 1050 /* subdiagonal part */ 1051 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->slvec0b);CHKERRQ(ierr); 1052 1053 /* copy x into the vec slvec0 */ 1054 ierr = VecGetArray(a->slvec0,&from);CHKERRQ(ierr); 1055 ierr = VecGetArray(xx,&x);CHKERRQ(ierr); 1056 ierr = PetscMemcpy(from,x,bs*mbs*sizeof(MatScalar));CHKERRQ(ierr); 1057 ierr = VecRestoreArray(a->slvec0,&from);CHKERRQ(ierr); 1058 1059 ierr = VecScatterBegin(a->slvec0,a->slvec1,ADD_VALUES,SCATTER_FORWARD,a->sMvctx);CHKERRQ(ierr); 1060 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 1061 ierr = VecScatterEnd(a->slvec0,a->slvec1,ADD_VALUES,SCATTER_FORWARD,a->sMvctx);CHKERRQ(ierr); 1062 1063 /* supperdiagonal part */ 1064 ierr = (*a->B->ops->multadd)(a->B,a->slvec1b,a->slvec1a,zz);CHKERRQ(ierr); 1065 1066 PetscFunctionReturn(0); 1067 } 1068 1069 #undef __FUNCT__ 1070 #define __FUNCT__ "MatMultAdd_MPISBAIJ_2comm" 1071 PetscErrorCode MatMultAdd_MPISBAIJ_2comm(Mat A,Vec xx,Vec yy,Vec zz) 1072 { 1073 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1074 PetscErrorCode ierr; 1075 1076 PetscFunctionBegin; 1077 ierr = VecScatterBegin(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 1078 /* do diagonal part */ 1079 ierr = (*a->A->ops->multadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 1080 /* do supperdiagonal part */ 1081 ierr = VecScatterEnd(xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD,a->Mvctx);CHKERRQ(ierr); 1082 ierr = (*a->B->ops->multadd)(a->B,a->lvec,zz,zz);CHKERRQ(ierr); 1083 1084 /* do subdiagonal part */ 1085 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1086 ierr = VecScatterBegin(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 1087 ierr = VecScatterEnd(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);CHKERRQ(ierr); 1088 1089 PetscFunctionReturn(0); 1090 } 1091 1092 #undef __FUNCT__ 1093 #define __FUNCT__ "MatMultTranspose_MPISBAIJ" 1094 PetscErrorCode MatMultTranspose_MPISBAIJ(Mat A,Vec xx,Vec yy) 1095 { 1096 PetscErrorCode ierr; 1097 1098 PetscFunctionBegin; 1099 ierr = MatMult(A,xx,yy);CHKERRQ(ierr); 1100 PetscFunctionReturn(0); 1101 } 1102 1103 #undef __FUNCT__ 1104 #define __FUNCT__ "MatMultTransposeAdd_MPISBAIJ" 1105 PetscErrorCode MatMultTransposeAdd_MPISBAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1106 { 1107 PetscErrorCode ierr; 1108 1109 PetscFunctionBegin; 1110 ierr = MatMultAdd(A,xx,yy,zz);CHKERRQ(ierr); 1111 PetscFunctionReturn(0); 1112 } 1113 1114 /* 1115 This only works correctly for square matrices where the subblock A->A is the 1116 diagonal block 1117 */ 1118 #undef __FUNCT__ 1119 #define __FUNCT__ "MatGetDiagonal_MPISBAIJ" 1120 PetscErrorCode MatGetDiagonal_MPISBAIJ(Mat A,Vec v) 1121 { 1122 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1123 PetscErrorCode ierr; 1124 1125 PetscFunctionBegin; 1126 /* if (a->M != a->N) SETERRQ(PETSC_ERR_SUP,"Supports only square matrix where A->A is diag block"); */ 1127 ierr = MatGetDiagonal(a->A,v);CHKERRQ(ierr); 1128 PetscFunctionReturn(0); 1129 } 1130 1131 #undef __FUNCT__ 1132 #define __FUNCT__ "MatScale_MPISBAIJ" 1133 PetscErrorCode MatScale_MPISBAIJ(const PetscScalar *aa,Mat A) 1134 { 1135 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1136 PetscErrorCode ierr; 1137 1138 PetscFunctionBegin; 1139 ierr = MatScale(aa,a->A);CHKERRQ(ierr); 1140 ierr = MatScale(aa,a->B);CHKERRQ(ierr); 1141 PetscFunctionReturn(0); 1142 } 1143 1144 #undef __FUNCT__ 1145 #define __FUNCT__ "MatGetRow_MPISBAIJ" 1146 PetscErrorCode MatGetRow_MPISBAIJ(Mat matin,int row,int *nz,int **idx,PetscScalar **v) 1147 { 1148 Mat_MPISBAIJ *mat = (Mat_MPISBAIJ*)matin->data; 1149 PetscScalar *vworkA,*vworkB,**pvA,**pvB,*v_p; 1150 PetscErrorCode ierr; 1151 int bs = mat->bs,bs2 = mat->bs2,i,*cworkA,*cworkB,**pcA,**pcB; 1152 int nztot,nzA,nzB,lrow,brstart = mat->rstart*bs,brend = mat->rend*bs; 1153 int *cmap,*idx_p,cstart = mat->cstart; 1154 1155 PetscFunctionBegin; 1156 if (mat->getrowactive == PETSC_TRUE) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Already active"); 1157 mat->getrowactive = PETSC_TRUE; 1158 1159 if (!mat->rowvalues && (idx || v)) { 1160 /* 1161 allocate enough space to hold information from the longest row. 1162 */ 1163 Mat_SeqSBAIJ *Aa = (Mat_SeqSBAIJ*)mat->A->data; 1164 Mat_SeqBAIJ *Ba = (Mat_SeqBAIJ*)mat->B->data; 1165 int max = 1,mbs = mat->mbs,tmp; 1166 for (i=0; i<mbs; i++) { 1167 tmp = Aa->i[i+1] - Aa->i[i] + Ba->i[i+1] - Ba->i[i]; /* row length */ 1168 if (max < tmp) { max = tmp; } 1169 } 1170 ierr = PetscMalloc(max*bs2*(sizeof(int)+sizeof(PetscScalar)),&mat->rowvalues);CHKERRQ(ierr); 1171 mat->rowindices = (int*)(mat->rowvalues + max*bs2); 1172 } 1173 1174 if (row < brstart || row >= brend) SETERRQ(PETSC_ERR_SUP,"Only local rows") 1175 lrow = row - brstart; /* local row index */ 1176 1177 pvA = &vworkA; pcA = &cworkA; pvB = &vworkB; pcB = &cworkB; 1178 if (!v) {pvA = 0; pvB = 0;} 1179 if (!idx) {pcA = 0; if (!v) pcB = 0;} 1180 ierr = (*mat->A->ops->getrow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1181 ierr = (*mat->B->ops->getrow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1182 nztot = nzA + nzB; 1183 1184 cmap = mat->garray; 1185 if (v || idx) { 1186 if (nztot) { 1187 /* Sort by increasing column numbers, assuming A and B already sorted */ 1188 int imark = -1; 1189 if (v) { 1190 *v = v_p = mat->rowvalues; 1191 for (i=0; i<nzB; i++) { 1192 if (cmap[cworkB[i]/bs] < cstart) v_p[i] = vworkB[i]; 1193 else break; 1194 } 1195 imark = i; 1196 for (i=0; i<nzA; i++) v_p[imark+i] = vworkA[i]; 1197 for (i=imark; i<nzB; i++) v_p[nzA+i] = vworkB[i]; 1198 } 1199 if (idx) { 1200 *idx = idx_p = mat->rowindices; 1201 if (imark > -1) { 1202 for (i=0; i<imark; i++) { 1203 idx_p[i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs; 1204 } 1205 } else { 1206 for (i=0; i<nzB; i++) { 1207 if (cmap[cworkB[i]/bs] < cstart) 1208 idx_p[i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs ; 1209 else break; 1210 } 1211 imark = i; 1212 } 1213 for (i=0; i<nzA; i++) idx_p[imark+i] = cstart*bs + cworkA[i]; 1214 for (i=imark; i<nzB; i++) idx_p[nzA+i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs ; 1215 } 1216 } else { 1217 if (idx) *idx = 0; 1218 if (v) *v = 0; 1219 } 1220 } 1221 *nz = nztot; 1222 ierr = (*mat->A->ops->restorerow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1223 ierr = (*mat->B->ops->restorerow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1224 PetscFunctionReturn(0); 1225 } 1226 1227 #undef __FUNCT__ 1228 #define __FUNCT__ "MatRestoreRow_MPISBAIJ" 1229 PetscErrorCode MatRestoreRow_MPISBAIJ(Mat mat,int row,int *nz,int **idx,PetscScalar **v) 1230 { 1231 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 1232 1233 PetscFunctionBegin; 1234 if (baij->getrowactive == PETSC_FALSE) { 1235 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"MatGetRow not called"); 1236 } 1237 baij->getrowactive = PETSC_FALSE; 1238 PetscFunctionReturn(0); 1239 } 1240 1241 #undef __FUNCT__ 1242 #define __FUNCT__ "MatGetBlockSize_MPISBAIJ" 1243 PetscErrorCode MatGetBlockSize_MPISBAIJ(Mat mat,int *bs) 1244 { 1245 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 1246 1247 PetscFunctionBegin; 1248 *bs = baij->bs; 1249 PetscFunctionReturn(0); 1250 } 1251 1252 #undef __FUNCT__ 1253 #define __FUNCT__ "MatZeroEntries_MPISBAIJ" 1254 PetscErrorCode MatZeroEntries_MPISBAIJ(Mat A) 1255 { 1256 Mat_MPISBAIJ *l = (Mat_MPISBAIJ*)A->data; 1257 PetscErrorCode ierr; 1258 1259 PetscFunctionBegin; 1260 ierr = MatZeroEntries(l->A);CHKERRQ(ierr); 1261 ierr = MatZeroEntries(l->B);CHKERRQ(ierr); 1262 PetscFunctionReturn(0); 1263 } 1264 1265 #undef __FUNCT__ 1266 #define __FUNCT__ "MatGetInfo_MPISBAIJ" 1267 PetscErrorCode MatGetInfo_MPISBAIJ(Mat matin,MatInfoType flag,MatInfo *info) 1268 { 1269 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)matin->data; 1270 Mat A = a->A,B = a->B; 1271 PetscErrorCode ierr; 1272 PetscReal isend[5],irecv[5]; 1273 1274 PetscFunctionBegin; 1275 info->block_size = (PetscReal)a->bs; 1276 ierr = MatGetInfo(A,MAT_LOCAL,info);CHKERRQ(ierr); 1277 isend[0] = info->nz_used; isend[1] = info->nz_allocated; isend[2] = info->nz_unneeded; 1278 isend[3] = info->memory; isend[4] = info->mallocs; 1279 ierr = MatGetInfo(B,MAT_LOCAL,info);CHKERRQ(ierr); 1280 isend[0] += info->nz_used; isend[1] += info->nz_allocated; isend[2] += info->nz_unneeded; 1281 isend[3] += info->memory; isend[4] += info->mallocs; 1282 if (flag == MAT_LOCAL) { 1283 info->nz_used = isend[0]; 1284 info->nz_allocated = isend[1]; 1285 info->nz_unneeded = isend[2]; 1286 info->memory = isend[3]; 1287 info->mallocs = isend[4]; 1288 } else if (flag == MAT_GLOBAL_MAX) { 1289 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_MAX,matin->comm);CHKERRQ(ierr); 1290 info->nz_used = irecv[0]; 1291 info->nz_allocated = irecv[1]; 1292 info->nz_unneeded = irecv[2]; 1293 info->memory = irecv[3]; 1294 info->mallocs = irecv[4]; 1295 } else if (flag == MAT_GLOBAL_SUM) { 1296 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_SUM,matin->comm);CHKERRQ(ierr); 1297 info->nz_used = irecv[0]; 1298 info->nz_allocated = irecv[1]; 1299 info->nz_unneeded = irecv[2]; 1300 info->memory = irecv[3]; 1301 info->mallocs = irecv[4]; 1302 } else { 1303 SETERRQ1(1,"Unknown MatInfoType argument %d",flag); 1304 } 1305 info->rows_global = (PetscReal)A->M; 1306 info->columns_global = (PetscReal)A->N; 1307 info->rows_local = (PetscReal)A->m; 1308 info->columns_local = (PetscReal)A->N; 1309 info->fill_ratio_given = 0; /* no parallel LU/ILU/Cholesky */ 1310 info->fill_ratio_needed = 0; 1311 info->factor_mallocs = 0; 1312 PetscFunctionReturn(0); 1313 } 1314 1315 #undef __FUNCT__ 1316 #define __FUNCT__ "MatSetOption_MPISBAIJ" 1317 PetscErrorCode MatSetOption_MPISBAIJ(Mat A,MatOption op) 1318 { 1319 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1320 PetscErrorCode ierr; 1321 1322 PetscFunctionBegin; 1323 switch (op) { 1324 case MAT_NO_NEW_NONZERO_LOCATIONS: 1325 case MAT_YES_NEW_NONZERO_LOCATIONS: 1326 case MAT_COLUMNS_UNSORTED: 1327 case MAT_COLUMNS_SORTED: 1328 case MAT_NEW_NONZERO_ALLOCATION_ERR: 1329 case MAT_KEEP_ZEROED_ROWS: 1330 case MAT_NEW_NONZERO_LOCATION_ERR: 1331 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 1332 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 1333 break; 1334 case MAT_ROW_ORIENTED: 1335 a->roworiented = PETSC_TRUE; 1336 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 1337 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 1338 break; 1339 case MAT_ROWS_SORTED: 1340 case MAT_ROWS_UNSORTED: 1341 case MAT_YES_NEW_DIAGONALS: 1342 PetscLogInfo(A,"Info:MatSetOption_MPIBAIJ:Option ignored\n"); 1343 break; 1344 case MAT_COLUMN_ORIENTED: 1345 a->roworiented = PETSC_FALSE; 1346 ierr = MatSetOption(a->A,op);CHKERRQ(ierr); 1347 ierr = MatSetOption(a->B,op);CHKERRQ(ierr); 1348 break; 1349 case MAT_IGNORE_OFF_PROC_ENTRIES: 1350 a->donotstash = PETSC_TRUE; 1351 break; 1352 case MAT_NO_NEW_DIAGONALS: 1353 SETERRQ(PETSC_ERR_SUP,"MAT_NO_NEW_DIAGONALS"); 1354 case MAT_USE_HASH_TABLE: 1355 a->ht_flag = PETSC_TRUE; 1356 break; 1357 case MAT_NOT_SYMMETRIC: 1358 case MAT_NOT_STRUCTURALLY_SYMMETRIC: 1359 case MAT_HERMITIAN: 1360 SETERRQ(PETSC_ERR_SUP,"Matrix must be symmetric"); 1361 case MAT_SYMMETRIC: 1362 case MAT_STRUCTURALLY_SYMMETRIC: 1363 case MAT_NOT_HERMITIAN: 1364 case MAT_SYMMETRY_ETERNAL: 1365 case MAT_NOT_SYMMETRY_ETERNAL: 1366 break; 1367 default: 1368 SETERRQ(PETSC_ERR_SUP,"unknown option"); 1369 } 1370 PetscFunctionReturn(0); 1371 } 1372 1373 #undef __FUNCT__ 1374 #define __FUNCT__ "MatTranspose_MPISBAIJ" 1375 PetscErrorCode MatTranspose_MPISBAIJ(Mat A,Mat *B) 1376 { 1377 PetscErrorCode ierr; 1378 PetscFunctionBegin; 1379 ierr = MatDuplicate(A,MAT_COPY_VALUES,B);CHKERRQ(ierr); 1380 PetscFunctionReturn(0); 1381 } 1382 1383 #undef __FUNCT__ 1384 #define __FUNCT__ "MatDiagonalScale_MPISBAIJ" 1385 PetscErrorCode MatDiagonalScale_MPISBAIJ(Mat mat,Vec ll,Vec rr) 1386 { 1387 Mat_MPISBAIJ *baij = (Mat_MPISBAIJ*)mat->data; 1388 Mat a = baij->A,b = baij->B; 1389 PetscErrorCode ierr; 1390 int s1,s2,s3; 1391 1392 PetscFunctionBegin; 1393 if (ll != rr) { 1394 SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"For symmetric format, left and right scaling vectors must be same\n"); 1395 } 1396 ierr = MatGetLocalSize(mat,&s2,&s3);CHKERRQ(ierr); 1397 if (rr) { 1398 ierr = VecGetLocalSize(rr,&s1);CHKERRQ(ierr); 1399 if (s1!=s3) SETERRQ(PETSC_ERR_ARG_SIZ,"right vector non-conforming local size"); 1400 /* Overlap communication with computation. */ 1401 ierr = VecScatterBegin(rr,baij->lvec,INSERT_VALUES,SCATTER_FORWARD,baij->Mvctx);CHKERRQ(ierr); 1402 /*} if (ll) { */ 1403 ierr = VecGetLocalSize(ll,&s1);CHKERRQ(ierr); 1404 if (s1!=s2) SETERRQ(PETSC_ERR_ARG_SIZ,"left vector non-conforming local size"); 1405 ierr = (*b->ops->diagonalscale)(b,ll,PETSC_NULL);CHKERRQ(ierr); 1406 /* } */ 1407 /* scale the diagonal block */ 1408 ierr = (*a->ops->diagonalscale)(a,ll,rr);CHKERRQ(ierr); 1409 1410 /* if (rr) { */ 1411 /* Do a scatter end and then right scale the off-diagonal block */ 1412 ierr = VecScatterEnd(rr,baij->lvec,INSERT_VALUES,SCATTER_FORWARD,baij->Mvctx);CHKERRQ(ierr); 1413 ierr = (*b->ops->diagonalscale)(b,PETSC_NULL,baij->lvec);CHKERRQ(ierr); 1414 } 1415 1416 PetscFunctionReturn(0); 1417 } 1418 1419 #undef __FUNCT__ 1420 #define __FUNCT__ "MatZeroRows_MPISBAIJ" 1421 PetscErrorCode MatZeroRows_MPISBAIJ(Mat A,IS is,const PetscScalar *diag) 1422 { 1423 PetscFunctionBegin; 1424 SETERRQ(PETSC_ERR_SUP,"No support for this function yet"); 1425 } 1426 1427 #undef __FUNCT__ 1428 #define __FUNCT__ "MatPrintHelp_MPISBAIJ" 1429 PetscErrorCode MatPrintHelp_MPISBAIJ(Mat A) 1430 { 1431 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1432 MPI_Comm comm = A->comm; 1433 static int called = 0; 1434 PetscErrorCode ierr; 1435 1436 PetscFunctionBegin; 1437 if (!a->rank) { 1438 ierr = MatPrintHelp_SeqSBAIJ(a->A);CHKERRQ(ierr); 1439 } 1440 if (called) {PetscFunctionReturn(0);} else called = 1; 1441 ierr = (*PetscHelpPrintf)(comm," Options for MATMPISBAIJ matrix format (the defaults):\n");CHKERRQ(ierr); 1442 ierr = (*PetscHelpPrintf)(comm," -mat_use_hash_table <factor>: Use hashtable for efficient matrix assembly\n");CHKERRQ(ierr); 1443 PetscFunctionReturn(0); 1444 } 1445 1446 #undef __FUNCT__ 1447 #define __FUNCT__ "MatSetUnfactored_MPISBAIJ" 1448 PetscErrorCode MatSetUnfactored_MPISBAIJ(Mat A) 1449 { 1450 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 1451 PetscErrorCode ierr; 1452 1453 PetscFunctionBegin; 1454 ierr = MatSetUnfactored(a->A);CHKERRQ(ierr); 1455 PetscFunctionReturn(0); 1456 } 1457 1458 static PetscErrorCode MatDuplicate_MPISBAIJ(Mat,MatDuplicateOption,Mat *); 1459 1460 #undef __FUNCT__ 1461 #define __FUNCT__ "MatEqual_MPISBAIJ" 1462 PetscErrorCode MatEqual_MPISBAIJ(Mat A,Mat B,PetscTruth *flag) 1463 { 1464 Mat_MPISBAIJ *matB = (Mat_MPISBAIJ*)B->data,*matA = (Mat_MPISBAIJ*)A->data; 1465 Mat a,b,c,d; 1466 PetscTruth flg; 1467 PetscErrorCode ierr; 1468 1469 PetscFunctionBegin; 1470 a = matA->A; b = matA->B; 1471 c = matB->A; d = matB->B; 1472 1473 ierr = MatEqual(a,c,&flg);CHKERRQ(ierr); 1474 if (flg == PETSC_TRUE) { 1475 ierr = MatEqual(b,d,&flg);CHKERRQ(ierr); 1476 } 1477 ierr = MPI_Allreduce(&flg,flag,1,MPI_INT,MPI_LAND,A->comm);CHKERRQ(ierr); 1478 PetscFunctionReturn(0); 1479 } 1480 1481 #undef __FUNCT__ 1482 #define __FUNCT__ "MatSetUpPreallocation_MPISBAIJ" 1483 PetscErrorCode MatSetUpPreallocation_MPISBAIJ(Mat A) 1484 { 1485 PetscErrorCode ierr; 1486 1487 PetscFunctionBegin; 1488 ierr = MatMPISBAIJSetPreallocation(A,1,PETSC_DEFAULT,0,PETSC_DEFAULT,0);CHKERRQ(ierr); 1489 PetscFunctionReturn(0); 1490 } 1491 1492 #undef __FUNCT__ 1493 #define __FUNCT__ "MatGetSubMatrices_MPISBAIJ" 1494 PetscErrorCode MatGetSubMatrices_MPISBAIJ(Mat A,int n,const IS irow[],const IS icol[],MatReuse scall,Mat *B[]) 1495 { 1496 PetscErrorCode ierr; 1497 int i; 1498 PetscTruth flg; 1499 1500 PetscFunctionBegin; 1501 for (i=0; i<n; i++) { 1502 ierr = ISEqual(irow[i],icol[i],&flg);CHKERRQ(ierr); 1503 if (!flg) { 1504 SETERRQ(1,"Can only get symmetric submatrix for MPISBAIJ matrices"); 1505 } 1506 } 1507 ierr = MatGetSubMatrices_MPIBAIJ(A,n,irow,icol,scall,B);CHKERRQ(ierr); 1508 PetscFunctionReturn(0); 1509 } 1510 1511 1512 /* -------------------------------------------------------------------*/ 1513 static struct _MatOps MatOps_Values = { 1514 MatSetValues_MPISBAIJ, 1515 MatGetRow_MPISBAIJ, 1516 MatRestoreRow_MPISBAIJ, 1517 MatMult_MPISBAIJ, 1518 /* 4*/ MatMultAdd_MPISBAIJ, 1519 MatMultTranspose_MPISBAIJ, 1520 MatMultTransposeAdd_MPISBAIJ, 1521 0, 1522 0, 1523 0, 1524 /*10*/ 0, 1525 0, 1526 0, 1527 MatRelax_MPISBAIJ, 1528 MatTranspose_MPISBAIJ, 1529 /*15*/ MatGetInfo_MPISBAIJ, 1530 MatEqual_MPISBAIJ, 1531 MatGetDiagonal_MPISBAIJ, 1532 MatDiagonalScale_MPISBAIJ, 1533 MatNorm_MPISBAIJ, 1534 /*20*/ MatAssemblyBegin_MPISBAIJ, 1535 MatAssemblyEnd_MPISBAIJ, 1536 0, 1537 MatSetOption_MPISBAIJ, 1538 MatZeroEntries_MPISBAIJ, 1539 /*25*/ MatZeroRows_MPISBAIJ, 1540 0, 1541 0, 1542 0, 1543 0, 1544 /*30*/ MatSetUpPreallocation_MPISBAIJ, 1545 0, 1546 0, 1547 0, 1548 0, 1549 /*35*/ MatDuplicate_MPISBAIJ, 1550 0, 1551 0, 1552 0, 1553 0, 1554 /*40*/ 0, 1555 MatGetSubMatrices_MPISBAIJ, 1556 MatIncreaseOverlap_MPISBAIJ, 1557 MatGetValues_MPISBAIJ, 1558 0, 1559 /*45*/ MatPrintHelp_MPISBAIJ, 1560 MatScale_MPISBAIJ, 1561 0, 1562 0, 1563 0, 1564 /*50*/ MatGetBlockSize_MPISBAIJ, 1565 0, 1566 0, 1567 0, 1568 0, 1569 /*55*/ 0, 1570 0, 1571 MatSetUnfactored_MPISBAIJ, 1572 0, 1573 MatSetValuesBlocked_MPISBAIJ, 1574 /*60*/ 0, 1575 0, 1576 0, 1577 MatGetPetscMaps_Petsc, 1578 0, 1579 /*65*/ 0, 1580 0, 1581 0, 1582 0, 1583 0, 1584 /*70*/ MatGetRowMax_MPISBAIJ, 1585 0, 1586 0, 1587 0, 1588 0, 1589 /*75*/ 0, 1590 0, 1591 0, 1592 0, 1593 0, 1594 /*80*/ 0, 1595 0, 1596 0, 1597 0, 1598 /*85*/ MatLoad_MPISBAIJ 1599 }; 1600 1601 1602 EXTERN_C_BEGIN 1603 #undef __FUNCT__ 1604 #define __FUNCT__ "MatGetDiagonalBlock_MPISBAIJ" 1605 PetscErrorCode MatGetDiagonalBlock_MPISBAIJ(Mat A,PetscTruth *iscopy,MatReuse reuse,Mat *a) 1606 { 1607 PetscFunctionBegin; 1608 *a = ((Mat_MPISBAIJ *)A->data)->A; 1609 *iscopy = PETSC_FALSE; 1610 PetscFunctionReturn(0); 1611 } 1612 EXTERN_C_END 1613 1614 EXTERN_C_BEGIN 1615 #undef __FUNCT__ 1616 #define __FUNCT__ "MatMPISBAIJSetPreallocation_MPISBAIJ" 1617 PetscErrorCode MatMPISBAIJSetPreallocation_MPISBAIJ(Mat B,int bs,int d_nz,int *d_nnz,int o_nz,int *o_nnz) 1618 { 1619 Mat_MPISBAIJ *b; 1620 PetscErrorCode ierr; 1621 int i,mbs,Mbs; 1622 1623 PetscFunctionBegin; 1624 ierr = PetscOptionsGetInt(B->prefix,"-mat_block_size",&bs,PETSC_NULL);CHKERRQ(ierr); 1625 1626 if (bs < 1) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Invalid block size specified, must be positive"); 1627 if (d_nz == PETSC_DECIDE || d_nz == PETSC_DEFAULT) d_nz = 3; 1628 if (o_nz == PETSC_DECIDE || o_nz == PETSC_DEFAULT) o_nz = 1; 1629 if (d_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"d_nz cannot be less than 0: value %d",d_nz); 1630 if (o_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"o_nz cannot be less than 0: value %d",o_nz); 1631 if (d_nnz) { 1632 for (i=0; i<B->m/bs; i++) { 1633 if (d_nnz[i] < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"d_nnz cannot be less than -1: local row %d value %d",i,d_nnz[i]); 1634 } 1635 } 1636 if (o_nnz) { 1637 for (i=0; i<B->m/bs; i++) { 1638 if (o_nnz[i] < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"o_nnz cannot be less than -1: local row %d value %d",i,o_nnz[i]); 1639 } 1640 } 1641 B->preallocated = PETSC_TRUE; 1642 ierr = PetscSplitOwnershipBlock(B->comm,bs,&B->m,&B->M);CHKERRQ(ierr); 1643 ierr = PetscSplitOwnershipBlock(B->comm,bs,&B->n,&B->N);CHKERRQ(ierr); 1644 ierr = PetscMapCreateMPI(B->comm,B->m,B->M,&B->rmap);CHKERRQ(ierr); 1645 ierr = PetscMapCreateMPI(B->comm,B->m,B->M,&B->cmap);CHKERRQ(ierr); 1646 1647 b = (Mat_MPISBAIJ*)B->data; 1648 mbs = B->m/bs; 1649 Mbs = B->M/bs; 1650 if (mbs*bs != B->m) { 1651 SETERRQ2(PETSC_ERR_ARG_SIZ,"No of local rows %d must be divisible by blocksize %d",B->m,bs); 1652 } 1653 1654 b->bs = bs; 1655 b->bs2 = bs*bs; 1656 b->mbs = mbs; 1657 b->nbs = mbs; 1658 b->Mbs = Mbs; 1659 b->Nbs = Mbs; 1660 1661 ierr = MPI_Allgather(&b->mbs,1,MPI_INT,b->rowners+1,1,MPI_INT,B->comm);CHKERRQ(ierr); 1662 b->rowners[0] = 0; 1663 for (i=2; i<=b->size; i++) { 1664 b->rowners[i] += b->rowners[i-1]; 1665 } 1666 b->rstart = b->rowners[b->rank]; 1667 b->rend = b->rowners[b->rank+1]; 1668 b->cstart = b->rstart; 1669 b->cend = b->rend; 1670 for (i=0; i<=b->size; i++) { 1671 b->rowners_bs[i] = b->rowners[i]*bs; 1672 } 1673 b->rstart_bs = b-> rstart*bs; 1674 b->rend_bs = b->rend*bs; 1675 1676 b->cstart_bs = b->cstart*bs; 1677 b->cend_bs = b->cend*bs; 1678 1679 ierr = MatCreate(PETSC_COMM_SELF,B->m,B->m,B->m,B->m,&b->A);CHKERRQ(ierr); 1680 ierr = MatSetType(b->A,MATSEQSBAIJ);CHKERRQ(ierr); 1681 ierr = MatSeqSBAIJSetPreallocation(b->A,bs,d_nz,d_nnz);CHKERRQ(ierr); 1682 PetscLogObjectParent(B,b->A); 1683 1684 ierr = MatCreate(PETSC_COMM_SELF,B->m,B->M,B->m,B->M,&b->B);CHKERRQ(ierr); 1685 ierr = MatSetType(b->B,MATSEQBAIJ);CHKERRQ(ierr); 1686 ierr = MatSeqBAIJSetPreallocation(b->B,bs,o_nz,o_nnz);CHKERRQ(ierr); 1687 PetscLogObjectParent(B,b->B); 1688 1689 /* build cache for off array entries formed */ 1690 ierr = MatStashCreate_Private(B->comm,bs,&B->bstash);CHKERRQ(ierr); 1691 1692 PetscFunctionReturn(0); 1693 } 1694 EXTERN_C_END 1695 1696 /*MC 1697 MATMPISBAIJ - MATMPISBAIJ = "mpisbaij" - A matrix type to be used for distributed symmetric sparse block matrices, 1698 based on block compressed sparse row format. Only the upper triangular portion of the matrix is stored. 1699 1700 Options Database Keys: 1701 . -mat_type mpisbaij - sets the matrix type to "mpisbaij" during a call to MatSetFromOptions() 1702 1703 Level: beginner 1704 1705 .seealso: MatCreateMPISBAIJ 1706 M*/ 1707 1708 EXTERN_C_BEGIN 1709 #undef __FUNCT__ 1710 #define __FUNCT__ "MatCreate_MPISBAIJ" 1711 PetscErrorCode MatCreate_MPISBAIJ(Mat B) 1712 { 1713 Mat_MPISBAIJ *b; 1714 PetscErrorCode ierr; 1715 PetscTruth flg; 1716 1717 PetscFunctionBegin; 1718 1719 ierr = PetscNew(Mat_MPISBAIJ,&b);CHKERRQ(ierr); 1720 B->data = (void*)b; 1721 ierr = PetscMemzero(b,sizeof(Mat_MPISBAIJ));CHKERRQ(ierr); 1722 ierr = PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 1723 1724 B->ops->destroy = MatDestroy_MPISBAIJ; 1725 B->ops->view = MatView_MPISBAIJ; 1726 B->mapping = 0; 1727 B->factor = 0; 1728 B->assembled = PETSC_FALSE; 1729 1730 B->insertmode = NOT_SET_VALUES; 1731 ierr = MPI_Comm_rank(B->comm,&b->rank);CHKERRQ(ierr); 1732 ierr = MPI_Comm_size(B->comm,&b->size);CHKERRQ(ierr); 1733 1734 /* build local table of row and column ownerships */ 1735 ierr = PetscMalloc(3*(b->size+2)*sizeof(int),&b->rowners);CHKERRQ(ierr); 1736 b->cowners = b->rowners + b->size + 2; 1737 b->rowners_bs = b->cowners + b->size + 2; 1738 PetscLogObjectMemory(B,3*(b->size+2)*sizeof(int)+sizeof(struct _p_Mat)+sizeof(Mat_MPISBAIJ)); 1739 1740 /* build cache for off array entries formed */ 1741 ierr = MatStashCreate_Private(B->comm,1,&B->stash);CHKERRQ(ierr); 1742 b->donotstash = PETSC_FALSE; 1743 b->colmap = PETSC_NULL; 1744 b->garray = PETSC_NULL; 1745 b->roworiented = PETSC_TRUE; 1746 1747 #if defined(PETSC_USE_MAT_SINGLE) 1748 /* stuff for MatSetValues_XXX in single precision */ 1749 b->setvalueslen = 0; 1750 b->setvaluescopy = PETSC_NULL; 1751 #endif 1752 1753 /* stuff used in block assembly */ 1754 b->barray = 0; 1755 1756 /* stuff used for matrix vector multiply */ 1757 b->lvec = 0; 1758 b->Mvctx = 0; 1759 b->slvec0 = 0; 1760 b->slvec0b = 0; 1761 b->slvec1 = 0; 1762 b->slvec1a = 0; 1763 b->slvec1b = 0; 1764 b->sMvctx = 0; 1765 1766 /* stuff for MatGetRow() */ 1767 b->rowindices = 0; 1768 b->rowvalues = 0; 1769 b->getrowactive = PETSC_FALSE; 1770 1771 /* hash table stuff */ 1772 b->ht = 0; 1773 b->hd = 0; 1774 b->ht_size = 0; 1775 b->ht_flag = PETSC_FALSE; 1776 b->ht_fact = 0; 1777 b->ht_total_ct = 0; 1778 b->ht_insert_ct = 0; 1779 1780 ierr = PetscOptionsHasName(B->prefix,"-mat_use_hash_table",&flg);CHKERRQ(ierr); 1781 if (flg) { 1782 PetscReal fact = 1.39; 1783 ierr = MatSetOption(B,MAT_USE_HASH_TABLE);CHKERRQ(ierr); 1784 ierr = PetscOptionsGetReal(B->prefix,"-mat_use_hash_table",&fact,PETSC_NULL);CHKERRQ(ierr); 1785 if (fact <= 1.0) fact = 1.39; 1786 ierr = MatMPIBAIJSetHashTableFactor(B,fact);CHKERRQ(ierr); 1787 PetscLogInfo(0,"MatCreateMPISBAIJ:Hash table Factor used %5.2f\n",fact); 1788 } 1789 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatStoreValues_C", 1790 "MatStoreValues_MPISBAIJ", 1791 MatStoreValues_MPISBAIJ);CHKERRQ(ierr); 1792 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatRetrieveValues_C", 1793 "MatRetrieveValues_MPISBAIJ", 1794 MatRetrieveValues_MPISBAIJ);CHKERRQ(ierr); 1795 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetDiagonalBlock_C", 1796 "MatGetDiagonalBlock_MPISBAIJ", 1797 MatGetDiagonalBlock_MPISBAIJ);CHKERRQ(ierr); 1798 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMPISBAIJSetPreallocation_C", 1799 "MatMPISBAIJSetPreallocation_MPISBAIJ", 1800 MatMPISBAIJSetPreallocation_MPISBAIJ);CHKERRQ(ierr); 1801 B->symmetric = PETSC_TRUE; 1802 B->structurally_symmetric = PETSC_TRUE; 1803 B->symmetric_set = PETSC_TRUE; 1804 B->structurally_symmetric_set = PETSC_TRUE; 1805 PetscFunctionReturn(0); 1806 } 1807 EXTERN_C_END 1808 1809 /*MC 1810 MATSBAIJ - MATSBAIJ = "sbaij" - A matrix type to be used for symmetric block sparse matrices. 1811 1812 This matrix type is identical to MATSEQSBAIJ when constructed with a single process communicator, 1813 and MATMPISBAIJ otherwise. 1814 1815 Options Database Keys: 1816 . -mat_type sbaij - sets the matrix type to "sbaij" during a call to MatSetFromOptions() 1817 1818 Level: beginner 1819 1820 .seealso: MatCreateMPISBAIJ,MATSEQSBAIJ,MATMPISBAIJ 1821 M*/ 1822 1823 EXTERN_C_BEGIN 1824 #undef __FUNCT__ 1825 #define __FUNCT__ "MatCreate_SBAIJ" 1826 PetscErrorCode MatCreate_SBAIJ(Mat A) 1827 { 1828 PetscErrorCode ierr; 1829 int size; 1830 1831 PetscFunctionBegin; 1832 ierr = PetscObjectChangeTypeName((PetscObject)A,MATSBAIJ);CHKERRQ(ierr); 1833 ierr = MPI_Comm_size(A->comm,&size);CHKERRQ(ierr); 1834 if (size == 1) { 1835 ierr = MatSetType(A,MATSEQSBAIJ);CHKERRQ(ierr); 1836 } else { 1837 ierr = MatSetType(A,MATMPISBAIJ);CHKERRQ(ierr); 1838 } 1839 PetscFunctionReturn(0); 1840 } 1841 EXTERN_C_END 1842 1843 #undef __FUNCT__ 1844 #define __FUNCT__ "MatMPISBAIJSetPreallocation" 1845 /*@C 1846 MatMPISBAIJSetPreallocation - For good matrix assembly performance 1847 the user should preallocate the matrix storage by setting the parameters 1848 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 1849 performance can be increased by more than a factor of 50. 1850 1851 Collective on Mat 1852 1853 Input Parameters: 1854 + A - the matrix 1855 . bs - size of blockk 1856 . d_nz - number of block nonzeros per block row in diagonal portion of local 1857 submatrix (same for all local rows) 1858 . d_nnz - array containing the number of block nonzeros in the various block rows 1859 in the upper triangular and diagonal part of the in diagonal portion of the local 1860 (possibly different for each block row) or PETSC_NULL. You must leave room 1861 for the diagonal entry even if it is zero. 1862 . o_nz - number of block nonzeros per block row in the off-diagonal portion of local 1863 submatrix (same for all local rows). 1864 - o_nnz - array containing the number of nonzeros in the various block rows of the 1865 off-diagonal portion of the local submatrix (possibly different for 1866 each block row) or PETSC_NULL. 1867 1868 1869 Options Database Keys: 1870 . -mat_no_unroll - uses code that does not unroll the loops in the 1871 block calculations (much slower) 1872 . -mat_block_size - size of the blocks to use 1873 1874 Notes: 1875 1876 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one processor 1877 than it must be used on all processors that share the object for that argument. 1878 1879 Storage Information: 1880 For a square global matrix we define each processor's diagonal portion 1881 to be its local rows and the corresponding columns (a square submatrix); 1882 each processor's off-diagonal portion encompasses the remainder of the 1883 local matrix (a rectangular submatrix). 1884 1885 The user can specify preallocated storage for the diagonal part of 1886 the local submatrix with either d_nz or d_nnz (not both). Set 1887 d_nz=PETSC_DEFAULT and d_nnz=PETSC_NULL for PETSc to control dynamic 1888 memory allocation. Likewise, specify preallocated storage for the 1889 off-diagonal part of the local submatrix with o_nz or o_nnz (not both). 1890 1891 Consider a processor that owns rows 3, 4 and 5 of a parallel matrix. In 1892 the figure below we depict these three local rows and all columns (0-11). 1893 1894 .vb 1895 0 1 2 3 4 5 6 7 8 9 10 11 1896 ------------------- 1897 row 3 | o o o d d d o o o o o o 1898 row 4 | o o o d d d o o o o o o 1899 row 5 | o o o d d d o o o o o o 1900 ------------------- 1901 .ve 1902 1903 Thus, any entries in the d locations are stored in the d (diagonal) 1904 submatrix, and any entries in the o locations are stored in the 1905 o (off-diagonal) submatrix. Note that the d matrix is stored in 1906 MatSeqSBAIJ format and the o submatrix in MATSEQBAIJ format. 1907 1908 Now d_nz should indicate the number of block nonzeros per row in the upper triangular 1909 plus the diagonal part of the d matrix, 1910 and o_nz should indicate the number of block nonzeros per row in the o matrix. 1911 In general, for PDE problems in which most nonzeros are near the diagonal, 1912 one expects d_nz >> o_nz. For large problems you MUST preallocate memory 1913 or you will get TERRIBLE performance; see the users' manual chapter on 1914 matrices. 1915 1916 Level: intermediate 1917 1918 .keywords: matrix, block, aij, compressed row, sparse, parallel 1919 1920 .seealso: MatCreate(), MatCreateSeqSBAIJ(), MatSetValues(), MatCreateMPIBAIJ() 1921 @*/ 1922 PetscErrorCode MatMPISBAIJSetPreallocation(Mat B,int bs,int d_nz,const int d_nnz[],int o_nz,const int o_nnz[]) 1923 { 1924 PetscErrorCode ierr,(*f)(Mat,int,int,const int[],int,const int[]); 1925 1926 PetscFunctionBegin; 1927 ierr = PetscObjectQueryFunction((PetscObject)B,"MatMPISBAIJSetPreallocation_C",(void (**)(void))&f);CHKERRQ(ierr); 1928 if (f) { 1929 ierr = (*f)(B,bs,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 1930 } 1931 PetscFunctionReturn(0); 1932 } 1933 1934 #undef __FUNCT__ 1935 #define __FUNCT__ "MatCreateMPISBAIJ" 1936 /*@C 1937 MatCreateMPISBAIJ - Creates a sparse parallel matrix in symmetric block AIJ format 1938 (block compressed row). For good matrix assembly performance 1939 the user should preallocate the matrix storage by setting the parameters 1940 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 1941 performance can be increased by more than a factor of 50. 1942 1943 Collective on MPI_Comm 1944 1945 Input Parameters: 1946 + comm - MPI communicator 1947 . bs - size of blockk 1948 . m - number of local rows (or PETSC_DECIDE to have calculated if M is given) 1949 This value should be the same as the local size used in creating the 1950 y vector for the matrix-vector product y = Ax. 1951 . n - number of local columns (or PETSC_DECIDE to have calculated if N is given) 1952 This value should be the same as the local size used in creating the 1953 x vector for the matrix-vector product y = Ax. 1954 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 1955 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 1956 . d_nz - number of block nonzeros per block row in diagonal portion of local 1957 submatrix (same for all local rows) 1958 . d_nnz - array containing the number of block nonzeros in the various block rows 1959 in the upper triangular portion of the in diagonal portion of the local 1960 (possibly different for each block block row) or PETSC_NULL. 1961 You must leave room for the diagonal entry even if it is zero. 1962 . o_nz - number of block nonzeros per block row in the off-diagonal portion of local 1963 submatrix (same for all local rows). 1964 - o_nnz - array containing the number of nonzeros in the various block rows of the 1965 off-diagonal portion of the local submatrix (possibly different for 1966 each block row) or PETSC_NULL. 1967 1968 Output Parameter: 1969 . A - the matrix 1970 1971 Options Database Keys: 1972 . -mat_no_unroll - uses code that does not unroll the loops in the 1973 block calculations (much slower) 1974 . -mat_block_size - size of the blocks to use 1975 . -mat_mpi - use the parallel matrix data structures even on one processor 1976 (defaults to using SeqBAIJ format on one processor) 1977 1978 Notes: 1979 The user MUST specify either the local or global matrix dimensions 1980 (possibly both). 1981 1982 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one processor 1983 than it must be used on all processors that share the object for that argument. 1984 1985 Storage Information: 1986 For a square global matrix we define each processor's diagonal portion 1987 to be its local rows and the corresponding columns (a square submatrix); 1988 each processor's off-diagonal portion encompasses the remainder of the 1989 local matrix (a rectangular submatrix). 1990 1991 The user can specify preallocated storage for the diagonal part of 1992 the local submatrix with either d_nz or d_nnz (not both). Set 1993 d_nz=PETSC_DEFAULT and d_nnz=PETSC_NULL for PETSc to control dynamic 1994 memory allocation. Likewise, specify preallocated storage for the 1995 off-diagonal part of the local submatrix with o_nz or o_nnz (not both). 1996 1997 Consider a processor that owns rows 3, 4 and 5 of a parallel matrix. In 1998 the figure below we depict these three local rows and all columns (0-11). 1999 2000 .vb 2001 0 1 2 3 4 5 6 7 8 9 10 11 2002 ------------------- 2003 row 3 | o o o d d d o o o o o o 2004 row 4 | o o o d d d o o o o o o 2005 row 5 | o o o d d d o o o o o o 2006 ------------------- 2007 .ve 2008 2009 Thus, any entries in the d locations are stored in the d (diagonal) 2010 submatrix, and any entries in the o locations are stored in the 2011 o (off-diagonal) submatrix. Note that the d matrix is stored in 2012 MatSeqSBAIJ format and the o submatrix in MATSEQBAIJ format. 2013 2014 Now d_nz should indicate the number of block nonzeros per row in the upper triangular 2015 plus the diagonal part of the d matrix, 2016 and o_nz should indicate the number of block nonzeros per row in the o matrix. 2017 In general, for PDE problems in which most nonzeros are near the diagonal, 2018 one expects d_nz >> o_nz. For large problems you MUST preallocate memory 2019 or you will get TERRIBLE performance; see the users' manual chapter on 2020 matrices. 2021 2022 Level: intermediate 2023 2024 .keywords: matrix, block, aij, compressed row, sparse, parallel 2025 2026 .seealso: MatCreate(), MatCreateSeqSBAIJ(), MatSetValues(), MatCreateMPIBAIJ() 2027 @*/ 2028 2029 PetscErrorCode MatCreateMPISBAIJ(MPI_Comm comm,int bs,int m,int n,int M,int N,int d_nz,const int d_nnz[],int o_nz,const int o_nnz[],Mat *A) 2030 { 2031 PetscErrorCode ierr; 2032 int size; 2033 2034 PetscFunctionBegin; 2035 ierr = MatCreate(comm,m,n,M,N,A);CHKERRQ(ierr); 2036 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 2037 if (size > 1) { 2038 ierr = MatSetType(*A,MATMPISBAIJ);CHKERRQ(ierr); 2039 ierr = MatMPISBAIJSetPreallocation(*A,bs,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 2040 } else { 2041 ierr = MatSetType(*A,MATSEQSBAIJ);CHKERRQ(ierr); 2042 ierr = MatSeqSBAIJSetPreallocation(*A,bs,d_nz,d_nnz);CHKERRQ(ierr); 2043 } 2044 PetscFunctionReturn(0); 2045 } 2046 2047 2048 #undef __FUNCT__ 2049 #define __FUNCT__ "MatDuplicate_MPISBAIJ" 2050 static PetscErrorCode MatDuplicate_MPISBAIJ(Mat matin,MatDuplicateOption cpvalues,Mat *newmat) 2051 { 2052 Mat mat; 2053 Mat_MPISBAIJ *a,*oldmat = (Mat_MPISBAIJ*)matin->data; 2054 PetscErrorCode ierr; 2055 int len=0,nt,bs=oldmat->bs,mbs=oldmat->mbs; 2056 PetscScalar *array; 2057 2058 PetscFunctionBegin; 2059 *newmat = 0; 2060 ierr = MatCreate(matin->comm,matin->m,matin->n,matin->M,matin->N,&mat);CHKERRQ(ierr); 2061 ierr = MatSetType(mat,matin->type_name);CHKERRQ(ierr); 2062 ierr = PetscMemcpy(mat->ops,matin->ops,sizeof(struct _MatOps));CHKERRQ(ierr); 2063 2064 mat->factor = matin->factor; 2065 mat->preallocated = PETSC_TRUE; 2066 mat->assembled = PETSC_TRUE; 2067 mat->insertmode = NOT_SET_VALUES; 2068 2069 a = (Mat_MPISBAIJ*)mat->data; 2070 a->bs = oldmat->bs; 2071 a->bs2 = oldmat->bs2; 2072 a->mbs = oldmat->mbs; 2073 a->nbs = oldmat->nbs; 2074 a->Mbs = oldmat->Mbs; 2075 a->Nbs = oldmat->Nbs; 2076 2077 a->rstart = oldmat->rstart; 2078 a->rend = oldmat->rend; 2079 a->cstart = oldmat->cstart; 2080 a->cend = oldmat->cend; 2081 a->size = oldmat->size; 2082 a->rank = oldmat->rank; 2083 a->donotstash = oldmat->donotstash; 2084 a->roworiented = oldmat->roworiented; 2085 a->rowindices = 0; 2086 a->rowvalues = 0; 2087 a->getrowactive = PETSC_FALSE; 2088 a->barray = 0; 2089 a->rstart_bs = oldmat->rstart_bs; 2090 a->rend_bs = oldmat->rend_bs; 2091 a->cstart_bs = oldmat->cstart_bs; 2092 a->cend_bs = oldmat->cend_bs; 2093 2094 /* hash table stuff */ 2095 a->ht = 0; 2096 a->hd = 0; 2097 a->ht_size = 0; 2098 a->ht_flag = oldmat->ht_flag; 2099 a->ht_fact = oldmat->ht_fact; 2100 a->ht_total_ct = 0; 2101 a->ht_insert_ct = 0; 2102 2103 ierr = PetscMemcpy(a->rowners,oldmat->rowners,3*(a->size+2)*sizeof(int));CHKERRQ(ierr); 2104 ierr = MatStashCreate_Private(matin->comm,1,&mat->stash);CHKERRQ(ierr); 2105 ierr = MatStashCreate_Private(matin->comm,oldmat->bs,&mat->bstash);CHKERRQ(ierr); 2106 if (oldmat->colmap) { 2107 #if defined (PETSC_USE_CTABLE) 2108 ierr = PetscTableCreateCopy(oldmat->colmap,&a->colmap);CHKERRQ(ierr); 2109 #else 2110 ierr = PetscMalloc((a->Nbs)*sizeof(int),&a->colmap);CHKERRQ(ierr); 2111 PetscLogObjectMemory(mat,(a->Nbs)*sizeof(int)); 2112 ierr = PetscMemcpy(a->colmap,oldmat->colmap,(a->Nbs)*sizeof(int));CHKERRQ(ierr); 2113 #endif 2114 } else a->colmap = 0; 2115 2116 if (oldmat->garray && (len = ((Mat_SeqBAIJ*)(oldmat->B->data))->nbs)) { 2117 ierr = PetscMalloc(len*sizeof(int),&a->garray);CHKERRQ(ierr); 2118 PetscLogObjectMemory(mat,len*sizeof(int)); 2119 ierr = PetscMemcpy(a->garray,oldmat->garray,len*sizeof(int));CHKERRQ(ierr); 2120 } else a->garray = 0; 2121 2122 ierr = VecDuplicate(oldmat->lvec,&a->lvec);CHKERRQ(ierr); 2123 PetscLogObjectParent(mat,a->lvec); 2124 ierr = VecScatterCopy(oldmat->Mvctx,&a->Mvctx);CHKERRQ(ierr); 2125 PetscLogObjectParent(mat,a->Mvctx); 2126 2127 ierr = VecDuplicate(oldmat->slvec0,&a->slvec0);CHKERRQ(ierr); 2128 PetscLogObjectParent(mat,a->slvec0); 2129 ierr = VecDuplicate(oldmat->slvec1,&a->slvec1);CHKERRQ(ierr); 2130 PetscLogObjectParent(mat,a->slvec1); 2131 2132 ierr = VecGetLocalSize(a->slvec1,&nt);CHKERRQ(ierr); 2133 ierr = VecGetArray(a->slvec1,&array);CHKERRQ(ierr); 2134 ierr = VecCreateSeqWithArray(PETSC_COMM_SELF,bs*mbs,array,&a->slvec1a);CHKERRQ(ierr); 2135 ierr = VecCreateSeqWithArray(PETSC_COMM_SELF,nt-bs*mbs,array+bs*mbs,&a->slvec1b);CHKERRQ(ierr); 2136 ierr = VecRestoreArray(a->slvec1,&array);CHKERRQ(ierr); 2137 ierr = VecGetArray(a->slvec0,&array);CHKERRQ(ierr); 2138 ierr = VecCreateSeqWithArray(PETSC_COMM_SELF,nt-bs*mbs,array+bs*mbs,&a->slvec0b);CHKERRQ(ierr); 2139 ierr = VecRestoreArray(a->slvec0,&array);CHKERRQ(ierr); 2140 PetscLogObjectParent(mat,a->slvec0); 2141 PetscLogObjectParent(mat,a->slvec1); 2142 PetscLogObjectParent(mat,a->slvec0b); 2143 PetscLogObjectParent(mat,a->slvec1a); 2144 PetscLogObjectParent(mat,a->slvec1b); 2145 2146 /* ierr = VecScatterCopy(oldmat->sMvctx,&a->sMvctx); - not written yet, replaced by the lazy trick: */ 2147 ierr = PetscObjectReference((PetscObject)oldmat->sMvctx);CHKERRQ(ierr); 2148 a->sMvctx = oldmat->sMvctx; 2149 PetscLogObjectParent(mat,a->sMvctx); 2150 2151 ierr = MatDuplicate(oldmat->A,cpvalues,&a->A);CHKERRQ(ierr); 2152 PetscLogObjectParent(mat,a->A); 2153 ierr = MatDuplicate(oldmat->B,cpvalues,&a->B);CHKERRQ(ierr); 2154 PetscLogObjectParent(mat,a->B); 2155 ierr = PetscFListDuplicate(mat->qlist,&matin->qlist);CHKERRQ(ierr); 2156 *newmat = mat; 2157 PetscFunctionReturn(0); 2158 } 2159 2160 #include "petscsys.h" 2161 2162 #undef __FUNCT__ 2163 #define __FUNCT__ "MatLoad_MPISBAIJ" 2164 PetscErrorCode MatLoad_MPISBAIJ(PetscViewer viewer,const MatType type,Mat *newmat) 2165 { 2166 Mat A; 2167 PetscErrorCode ierr; 2168 int i,nz,j,rstart,rend,fd; 2169 PetscScalar *vals,*buf; 2170 MPI_Comm comm = ((PetscObject)viewer)->comm; 2171 MPI_Status status; 2172 int header[4],rank,size,*rowlengths = 0,M,N,m,*rowners,*browners,maxnz,*cols; 2173 int *locrowlens,*sndcounts = 0,*procsnz = 0,jj,*mycols,*ibuf; 2174 int tag = ((PetscObject)viewer)->tag,bs=1,Mbs,mbs,extra_rows; 2175 int *dlens,*odlens,*mask,*masked1,*masked2,rowcount,odcount; 2176 int dcount,kmax,k,nzcount,tmp; 2177 2178 PetscFunctionBegin; 2179 ierr = PetscOptionsGetInt(PETSC_NULL,"-matload_block_size",&bs,PETSC_NULL);CHKERRQ(ierr); 2180 2181 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 2182 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 2183 if (!rank) { 2184 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 2185 ierr = PetscBinaryRead(fd,(char *)header,4,PETSC_INT);CHKERRQ(ierr); 2186 if (header[0] != MAT_FILE_COOKIE) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"not matrix object"); 2187 if (header[3] < 0) { 2188 SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"Matrix stored in special format, cannot load as MPISBAIJ"); 2189 } 2190 } 2191 2192 ierr = MPI_Bcast(header+1,3,MPI_INT,0,comm);CHKERRQ(ierr); 2193 M = header[1]; N = header[2]; 2194 2195 if (M != N) SETERRQ(PETSC_ERR_SUP,"Can only do square matrices"); 2196 2197 /* 2198 This code adds extra rows to make sure the number of rows is 2199 divisible by the blocksize 2200 */ 2201 Mbs = M/bs; 2202 extra_rows = bs - M + bs*(Mbs); 2203 if (extra_rows == bs) extra_rows = 0; 2204 else Mbs++; 2205 if (extra_rows &&!rank) { 2206 PetscLogInfo(0,"MatLoad_MPISBAIJ:Padding loaded matrix to match blocksize\n"); 2207 } 2208 2209 /* determine ownership of all rows */ 2210 mbs = Mbs/size + ((Mbs % size) > rank); 2211 m = mbs*bs; 2212 ierr = PetscMalloc(2*(size+2)*sizeof(int),&rowners);CHKERRQ(ierr); 2213 browners = rowners + size + 1; 2214 ierr = MPI_Allgather(&mbs,1,MPI_INT,rowners+1,1,MPI_INT,comm);CHKERRQ(ierr); 2215 rowners[0] = 0; 2216 for (i=2; i<=size; i++) rowners[i] += rowners[i-1]; 2217 for (i=0; i<=size; i++) browners[i] = rowners[i]*bs; 2218 rstart = rowners[rank]; 2219 rend = rowners[rank+1]; 2220 2221 /* distribute row lengths to all processors */ 2222 ierr = PetscMalloc((rend-rstart)*bs*sizeof(int),&locrowlens);CHKERRQ(ierr); 2223 if (!rank) { 2224 ierr = PetscMalloc((M+extra_rows)*sizeof(int),&rowlengths);CHKERRQ(ierr); 2225 ierr = PetscBinaryRead(fd,rowlengths,M,PETSC_INT);CHKERRQ(ierr); 2226 for (i=0; i<extra_rows; i++) rowlengths[M+i] = 1; 2227 ierr = PetscMalloc(size*sizeof(int),&sndcounts);CHKERRQ(ierr); 2228 for (i=0; i<size; i++) sndcounts[i] = browners[i+1] - browners[i]; 2229 ierr = MPI_Scatterv(rowlengths,sndcounts,browners,MPI_INT,locrowlens,(rend-rstart)*bs,MPI_INT,0,comm);CHKERRQ(ierr); 2230 ierr = PetscFree(sndcounts);CHKERRQ(ierr); 2231 } else { 2232 ierr = MPI_Scatterv(0,0,0,MPI_INT,locrowlens,(rend-rstart)*bs,MPI_INT,0,comm);CHKERRQ(ierr); 2233 } 2234 2235 if (!rank) { /* procs[0] */ 2236 /* calculate the number of nonzeros on each processor */ 2237 ierr = PetscMalloc(size*sizeof(int),&procsnz);CHKERRQ(ierr); 2238 ierr = PetscMemzero(procsnz,size*sizeof(int));CHKERRQ(ierr); 2239 for (i=0; i<size; i++) { 2240 for (j=rowners[i]*bs; j< rowners[i+1]*bs; j++) { 2241 procsnz[i] += rowlengths[j]; 2242 } 2243 } 2244 ierr = PetscFree(rowlengths);CHKERRQ(ierr); 2245 2246 /* determine max buffer needed and allocate it */ 2247 maxnz = 0; 2248 for (i=0; i<size; i++) { 2249 maxnz = PetscMax(maxnz,procsnz[i]); 2250 } 2251 ierr = PetscMalloc(maxnz*sizeof(int),&cols);CHKERRQ(ierr); 2252 2253 /* read in my part of the matrix column indices */ 2254 nz = procsnz[0]; 2255 ierr = PetscMalloc(nz*sizeof(int),&ibuf);CHKERRQ(ierr); 2256 mycols = ibuf; 2257 if (size == 1) nz -= extra_rows; 2258 ierr = PetscBinaryRead(fd,mycols,nz,PETSC_INT);CHKERRQ(ierr); 2259 if (size == 1) for (i=0; i< extra_rows; i++) { mycols[nz+i] = M+i; } 2260 2261 /* read in every ones (except the last) and ship off */ 2262 for (i=1; i<size-1; i++) { 2263 nz = procsnz[i]; 2264 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 2265 ierr = MPI_Send(cols,nz,MPI_INT,i,tag,comm);CHKERRQ(ierr); 2266 } 2267 /* read in the stuff for the last proc */ 2268 if (size != 1) { 2269 nz = procsnz[size-1] - extra_rows; /* the extra rows are not on the disk */ 2270 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 2271 for (i=0; i<extra_rows; i++) cols[nz+i] = M+i; 2272 ierr = MPI_Send(cols,nz+extra_rows,MPI_INT,size-1,tag,comm);CHKERRQ(ierr); 2273 } 2274 ierr = PetscFree(cols);CHKERRQ(ierr); 2275 } else { /* procs[i], i>0 */ 2276 /* determine buffer space needed for message */ 2277 nz = 0; 2278 for (i=0; i<m; i++) { 2279 nz += locrowlens[i]; 2280 } 2281 ierr = PetscMalloc(nz*sizeof(int),&ibuf);CHKERRQ(ierr); 2282 mycols = ibuf; 2283 /* receive message of column indices*/ 2284 ierr = MPI_Recv(mycols,nz,MPI_INT,0,tag,comm,&status);CHKERRQ(ierr); 2285 ierr = MPI_Get_count(&status,MPI_INT,&maxnz);CHKERRQ(ierr); 2286 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 2287 } 2288 2289 /* loop over local rows, determining number of off diagonal entries */ 2290 ierr = PetscMalloc(2*(rend-rstart+1)*sizeof(int),&dlens);CHKERRQ(ierr); 2291 odlens = dlens + (rend-rstart); 2292 ierr = PetscMalloc(3*Mbs*sizeof(int),&mask);CHKERRQ(ierr); 2293 ierr = PetscMemzero(mask,3*Mbs*sizeof(int));CHKERRQ(ierr); 2294 masked1 = mask + Mbs; 2295 masked2 = masked1 + Mbs; 2296 rowcount = 0; nzcount = 0; 2297 for (i=0; i<mbs; i++) { 2298 dcount = 0; 2299 odcount = 0; 2300 for (j=0; j<bs; j++) { 2301 kmax = locrowlens[rowcount]; 2302 for (k=0; k<kmax; k++) { 2303 tmp = mycols[nzcount++]/bs; /* block col. index */ 2304 if (!mask[tmp]) { 2305 mask[tmp] = 1; 2306 if (tmp < rstart || tmp >= rend) masked2[odcount++] = tmp; /* entry in off-diag portion */ 2307 else masked1[dcount++] = tmp; /* entry in diag portion */ 2308 } 2309 } 2310 rowcount++; 2311 } 2312 2313 dlens[i] = dcount; /* d_nzz[i] */ 2314 odlens[i] = odcount; /* o_nzz[i] */ 2315 2316 /* zero out the mask elements we set */ 2317 for (j=0; j<dcount; j++) mask[masked1[j]] = 0; 2318 for (j=0; j<odcount; j++) mask[masked2[j]] = 0; 2319 } 2320 2321 /* create our matrix */ 2322 ierr = MatCreate(comm,m,m,PETSC_DETERMINE,PETSC_DETERMINE,&A);CHKERRQ(ierr); 2323 ierr = MatSetType(A,type);CHKERRQ(ierr); 2324 ierr = MatMPISBAIJSetPreallocation(A,bs,0,dlens,0,odlens);CHKERRQ(ierr); 2325 ierr = MatSetOption(A,MAT_COLUMNS_SORTED);CHKERRQ(ierr); 2326 2327 if (!rank) { 2328 ierr = PetscMalloc(maxnz*sizeof(PetscScalar),&buf);CHKERRQ(ierr); 2329 /* read in my part of the matrix numerical values */ 2330 nz = procsnz[0]; 2331 vals = buf; 2332 mycols = ibuf; 2333 if (size == 1) nz -= extra_rows; 2334 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 2335 if (size == 1) for (i=0; i< extra_rows; i++) { vals[nz+i] = 1.0; } 2336 2337 /* insert into matrix */ 2338 jj = rstart*bs; 2339 for (i=0; i<m; i++) { 2340 ierr = MatSetValues(A,1,&jj,locrowlens[i],mycols,vals,INSERT_VALUES);CHKERRQ(ierr); 2341 mycols += locrowlens[i]; 2342 vals += locrowlens[i]; 2343 jj++; 2344 } 2345 2346 /* read in other processors (except the last one) and ship out */ 2347 for (i=1; i<size-1; i++) { 2348 nz = procsnz[i]; 2349 vals = buf; 2350 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 2351 ierr = MPI_Send(vals,nz,MPIU_SCALAR,i,A->tag,comm);CHKERRQ(ierr); 2352 } 2353 /* the last proc */ 2354 if (size != 1){ 2355 nz = procsnz[i] - extra_rows; 2356 vals = buf; 2357 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 2358 for (i=0; i<extra_rows; i++) vals[nz+i] = 1.0; 2359 ierr = MPI_Send(vals,nz+extra_rows,MPIU_SCALAR,size-1,A->tag,comm);CHKERRQ(ierr); 2360 } 2361 ierr = PetscFree(procsnz);CHKERRQ(ierr); 2362 2363 } else { 2364 /* receive numeric values */ 2365 ierr = PetscMalloc(nz*sizeof(PetscScalar),&buf);CHKERRQ(ierr); 2366 2367 /* receive message of values*/ 2368 vals = buf; 2369 mycols = ibuf; 2370 ierr = MPI_Recv(vals,nz,MPIU_SCALAR,0,A->tag,comm,&status);CHKERRQ(ierr); 2371 ierr = MPI_Get_count(&status,MPIU_SCALAR,&maxnz);CHKERRQ(ierr); 2372 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 2373 2374 /* insert into matrix */ 2375 jj = rstart*bs; 2376 for (i=0; i<m; i++) { 2377 ierr = MatSetValues_MPISBAIJ(A,1,&jj,locrowlens[i],mycols,vals,INSERT_VALUES);CHKERRQ(ierr); 2378 mycols += locrowlens[i]; 2379 vals += locrowlens[i]; 2380 jj++; 2381 } 2382 } 2383 2384 ierr = PetscFree(locrowlens);CHKERRQ(ierr); 2385 ierr = PetscFree(buf);CHKERRQ(ierr); 2386 ierr = PetscFree(ibuf);CHKERRQ(ierr); 2387 ierr = PetscFree(rowners);CHKERRQ(ierr); 2388 ierr = PetscFree(dlens);CHKERRQ(ierr); 2389 ierr = PetscFree(mask);CHKERRQ(ierr); 2390 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2391 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2392 *newmat = A; 2393 PetscFunctionReturn(0); 2394 } 2395 2396 #undef __FUNCT__ 2397 #define __FUNCT__ "MatMPISBAIJSetHashTableFactor" 2398 /*@ 2399 MatMPISBAIJSetHashTableFactor - Sets the factor required to compute the size of the HashTable. 2400 2401 Input Parameters: 2402 . mat - the matrix 2403 . fact - factor 2404 2405 Collective on Mat 2406 2407 Level: advanced 2408 2409 Notes: 2410 This can also be set by the command line option: -mat_use_hash_table fact 2411 2412 .keywords: matrix, hashtable, factor, HT 2413 2414 .seealso: MatSetOption() 2415 @*/ 2416 PetscErrorCode MatMPISBAIJSetHashTableFactor(Mat mat,PetscReal fact) 2417 { 2418 PetscFunctionBegin; 2419 SETERRQ(1,"Function not yet written for SBAIJ format"); 2420 /* PetscFunctionReturn(0); */ 2421 } 2422 2423 #undef __FUNCT__ 2424 #define __FUNCT__ "MatGetRowMax_MPISBAIJ" 2425 PetscErrorCode MatGetRowMax_MPISBAIJ(Mat A,Vec v) 2426 { 2427 Mat_MPISBAIJ *a = (Mat_MPISBAIJ*)A->data; 2428 Mat_SeqBAIJ *b = (Mat_SeqBAIJ*)(a->B)->data; 2429 PetscReal atmp; 2430 PetscReal *work,*svalues,*rvalues; 2431 PetscErrorCode ierr; 2432 int i,bs,mbs,*bi,*bj,brow,j,ncols,krow,kcol,col,row,Mbs,bcol; 2433 int rank,size,*rowners_bs,dest,count,source; 2434 PetscScalar *va; 2435 MatScalar *ba; 2436 MPI_Status stat; 2437 2438 PetscFunctionBegin; 2439 ierr = MatGetRowMax(a->A,v);CHKERRQ(ierr); 2440 ierr = VecGetArray(v,&va);CHKERRQ(ierr); 2441 2442 ierr = MPI_Comm_size(A->comm,&size);CHKERRQ(ierr); 2443 ierr = MPI_Comm_rank(A->comm,&rank);CHKERRQ(ierr); 2444 2445 bs = a->bs; 2446 mbs = a->mbs; 2447 Mbs = a->Mbs; 2448 ba = b->a; 2449 bi = b->i; 2450 bj = b->j; 2451 /* 2452 PetscSynchronizedPrintf(A->comm,"[%d] M: %d, bs: %d, mbs: %d \n",rank,bs*Mbs,bs,mbs); 2453 PetscSynchronizedFlush(A->comm); 2454 */ 2455 2456 /* find ownerships */ 2457 rowners_bs = a->rowners_bs; 2458 /* 2459 if (!rank){ 2460 for (i=0; i<size+1; i++) PetscPrintf(PETSC_COMM_SELF," rowners_bs[%d]: %d\n",i,rowners_bs[i]); 2461 } 2462 */ 2463 2464 /* each proc creates an array to be distributed */ 2465 ierr = PetscMalloc(bs*Mbs*sizeof(PetscReal),&work);CHKERRQ(ierr); 2466 ierr = PetscMemzero(work,bs*Mbs*sizeof(PetscReal));CHKERRQ(ierr); 2467 2468 /* row_max for B */ 2469 if (rank != size-1){ 2470 for (i=0; i<mbs; i++) { 2471 ncols = bi[1] - bi[0]; bi++; 2472 brow = bs*i; 2473 for (j=0; j<ncols; j++){ 2474 bcol = bs*(*bj); 2475 for (kcol=0; kcol<bs; kcol++){ 2476 col = bcol + kcol; /* local col index */ 2477 col += rowners_bs[rank+1]; /* global col index */ 2478 /* PetscPrintf(PETSC_COMM_SELF,"[%d], col: %d\n",rank,col); */ 2479 for (krow=0; krow<bs; krow++){ 2480 atmp = PetscAbsScalar(*ba); ba++; 2481 row = brow + krow; /* local row index */ 2482 /* printf("val[%d,%d]: %g\n",row,col,atmp); */ 2483 if (PetscRealPart(va[row]) < atmp) va[row] = atmp; 2484 if (work[col] < atmp) work[col] = atmp; 2485 } 2486 } 2487 bj++; 2488 } 2489 } 2490 /* 2491 PetscPrintf(PETSC_COMM_SELF,"[%d], work: ",rank); 2492 for (i=0; i<bs*Mbs; i++) PetscPrintf(PETSC_COMM_SELF,"%g ",work[i]); 2493 PetscPrintf(PETSC_COMM_SELF,"[%d]: \n"); 2494 */ 2495 2496 /* send values to its owners */ 2497 for (dest=rank+1; dest<size; dest++){ 2498 svalues = work + rowners_bs[dest]; 2499 count = rowners_bs[dest+1]-rowners_bs[dest]; 2500 ierr = MPI_Send(svalues,count,MPIU_REAL,dest,rank,A->comm);CHKERRQ(ierr); 2501 /* 2502 PetscSynchronizedPrintf(A->comm,"[%d] sends %d values to [%d]: %g, %g, %g, %g\n",rank,count,dest,svalues[0],svalues[1],svalues[2],svalues[3]); 2503 PetscSynchronizedFlush(A->comm); 2504 */ 2505 } 2506 } 2507 2508 /* receive values */ 2509 if (rank){ 2510 rvalues = work; 2511 count = rowners_bs[rank+1]-rowners_bs[rank]; 2512 for (source=0; source<rank; source++){ 2513 ierr = MPI_Recv(rvalues,count,MPIU_REAL,MPI_ANY_SOURCE,MPI_ANY_TAG,A->comm,&stat);CHKERRQ(ierr); 2514 /* process values */ 2515 for (i=0; i<count; i++){ 2516 if (PetscRealPart(va[i]) < rvalues[i]) va[i] = rvalues[i]; 2517 } 2518 /* 2519 PetscSynchronizedPrintf(A->comm,"[%d] received %d values from [%d]: %g, %g, %g, %g \n",rank,count,stat.MPI_SOURCE,rvalues[0],rvalues[1],rvalues[2],rvalues[3]); 2520 PetscSynchronizedFlush(A->comm); 2521 */ 2522 } 2523 } 2524 2525 ierr = VecRestoreArray(v,&va);CHKERRQ(ierr); 2526 ierr = PetscFree(work);CHKERRQ(ierr); 2527 PetscFunctionReturn(0); 2528 } 2529 2530 #undef __FUNCT__ 2531 #define __FUNCT__ "MatRelax_MPISBAIJ" 2532 PetscErrorCode MatRelax_MPISBAIJ(Mat matin,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,int its,int lits,Vec xx) 2533 { 2534 Mat_MPISBAIJ *mat = (Mat_MPISBAIJ*)matin->data; 2535 PetscErrorCode ierr; 2536 int mbs=mat->mbs,bs=mat->bs; 2537 PetscScalar mone=-1.0,*x,*b,*ptr,zero=0.0; 2538 Vec bb1; 2539 2540 PetscFunctionBegin; 2541 if (its <= 0 || lits <= 0) SETERRQ2(PETSC_ERR_ARG_WRONG,"Relaxation requires global its %d and local its %d both positive",its,lits); 2542 if (bs > 1) 2543 SETERRQ(PETSC_ERR_SUP,"SSOR for block size > 1 is not yet implemented"); 2544 2545 if ((flag & SOR_LOCAL_SYMMETRIC_SWEEP) == SOR_LOCAL_SYMMETRIC_SWEEP){ 2546 if ( flag & SOR_ZERO_INITIAL_GUESS ) { 2547 ierr = (*mat->A->ops->relax)(mat->A,bb,omega,flag,fshift,lits,lits,xx);CHKERRQ(ierr); 2548 its--; 2549 } 2550 2551 ierr = VecDuplicate(bb,&bb1);CHKERRQ(ierr); 2552 while (its--){ 2553 2554 /* lower triangular part: slvec0b = - B^T*xx */ 2555 ierr = (*mat->B->ops->multtranspose)(mat->B,xx,mat->slvec0b);CHKERRQ(ierr); 2556 2557 /* copy xx into slvec0a */ 2558 ierr = VecGetArray(mat->slvec0,&ptr);CHKERRQ(ierr); 2559 ierr = VecGetArray(xx,&x);CHKERRQ(ierr); 2560 ierr = PetscMemcpy(ptr,x,bs*mbs*sizeof(MatScalar));CHKERRQ(ierr); 2561 ierr = VecRestoreArray(mat->slvec0,&ptr);CHKERRQ(ierr); 2562 2563 ierr = VecScale(&mone,mat->slvec0);CHKERRQ(ierr); 2564 2565 /* copy bb into slvec1a */ 2566 ierr = VecGetArray(mat->slvec1,&ptr);CHKERRQ(ierr); 2567 ierr = VecGetArray(bb,&b);CHKERRQ(ierr); 2568 ierr = PetscMemcpy(ptr,b,bs*mbs*sizeof(MatScalar));CHKERRQ(ierr); 2569 ierr = VecRestoreArray(mat->slvec1,&ptr);CHKERRQ(ierr); 2570 2571 /* set slvec1b = 0 */ 2572 ierr = VecSet(&zero,mat->slvec1b);CHKERRQ(ierr); 2573 2574 ierr = VecScatterBegin(mat->slvec0,mat->slvec1,ADD_VALUES,SCATTER_FORWARD,mat->sMvctx);CHKERRQ(ierr); 2575 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 2576 ierr = VecRestoreArray(bb,&b);CHKERRQ(ierr); 2577 ierr = VecScatterEnd(mat->slvec0,mat->slvec1,ADD_VALUES,SCATTER_FORWARD,mat->sMvctx);CHKERRQ(ierr); 2578 2579 /* upper triangular part: bb1 = bb1 - B*x */ 2580 ierr = (*mat->B->ops->multadd)(mat->B,mat->slvec1b,mat->slvec1a,bb1);CHKERRQ(ierr); 2581 2582 /* local diagonal sweep */ 2583 ierr = (*mat->A->ops->relax)(mat->A,bb1,omega,SOR_SYMMETRIC_SWEEP,fshift,lits,lits,xx);CHKERRQ(ierr); 2584 } 2585 ierr = VecDestroy(bb1);CHKERRQ(ierr); 2586 } else { 2587 SETERRQ(PETSC_ERR_SUP,"MatSORType is not supported for SBAIJ matrix format"); 2588 } 2589 PetscFunctionReturn(0); 2590 } 2591 2592 #undef __FUNCT__ 2593 #define __FUNCT__ "MatRelax_MPISBAIJ_2comm" 2594 PetscErrorCode MatRelax_MPISBAIJ_2comm(Mat matin,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,int its,int lits,Vec xx) 2595 { 2596 Mat_MPISBAIJ *mat = (Mat_MPISBAIJ*)matin->data; 2597 PetscErrorCode ierr; 2598 PetscScalar mone=-1.0; 2599 Vec lvec1,bb1; 2600 2601 PetscFunctionBegin; 2602 if (its <= 0 || lits <= 0) SETERRQ2(PETSC_ERR_ARG_WRONG,"Relaxation requires global its %d and local its %d both positive",its,lits); 2603 if (mat->bs > 1) 2604 SETERRQ(PETSC_ERR_SUP,"SSOR for block size > 1 is not yet implemented"); 2605 2606 if ((flag & SOR_LOCAL_SYMMETRIC_SWEEP) == SOR_LOCAL_SYMMETRIC_SWEEP){ 2607 if ( flag & SOR_ZERO_INITIAL_GUESS ) { 2608 ierr = (*mat->A->ops->relax)(mat->A,bb,omega,flag,fshift,lits,lits,xx);CHKERRQ(ierr); 2609 its--; 2610 } 2611 2612 ierr = VecDuplicate(mat->lvec,&lvec1);CHKERRQ(ierr); 2613 ierr = VecDuplicate(bb,&bb1);CHKERRQ(ierr); 2614 while (its--){ 2615 ierr = VecScatterBegin(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 2616 2617 /* lower diagonal part: bb1 = bb - B^T*xx */ 2618 ierr = (*mat->B->ops->multtranspose)(mat->B,xx,lvec1);CHKERRQ(ierr); 2619 ierr = VecScale(&mone,lvec1);CHKERRQ(ierr); 2620 2621 ierr = VecScatterEnd(xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD,mat->Mvctx);CHKERRQ(ierr); 2622 ierr = VecCopy(bb,bb1);CHKERRQ(ierr); 2623 ierr = VecScatterBegin(lvec1,bb1,ADD_VALUES,SCATTER_REVERSE,mat->Mvctx);CHKERRQ(ierr); 2624 2625 /* upper diagonal part: bb1 = bb1 - B*x */ 2626 ierr = VecScale(&mone,mat->lvec);CHKERRQ(ierr); 2627 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb1,bb1);CHKERRQ(ierr); 2628 2629 ierr = VecScatterEnd(lvec1,bb1,ADD_VALUES,SCATTER_REVERSE,mat->Mvctx);CHKERRQ(ierr); 2630 2631 /* diagonal sweep */ 2632 ierr = (*mat->A->ops->relax)(mat->A,bb1,omega,SOR_SYMMETRIC_SWEEP,fshift,lits,lits,xx);CHKERRQ(ierr); 2633 } 2634 ierr = VecDestroy(lvec1);CHKERRQ(ierr); 2635 ierr = VecDestroy(bb1);CHKERRQ(ierr); 2636 } else { 2637 SETERRQ(PETSC_ERR_SUP,"MatSORType is not supported for SBAIJ matrix format"); 2638 } 2639 PetscFunctionReturn(0); 2640 } 2641 2642