1 #define PETSCMAT_DLL 2 3 #include "../src/mat/impls/baij/mpi/mpibaij.h" /*I "petscmat.h" I*/ 4 5 EXTERN PetscErrorCode MatSetUpMultiply_MPIBAIJ(Mat); 6 EXTERN PetscErrorCode DisAssemble_MPIBAIJ(Mat); 7 EXTERN PetscErrorCode MatIncreaseOverlap_MPIBAIJ(Mat,PetscInt,IS[],PetscInt); 8 EXTERN PetscErrorCode MatGetSubMatrices_MPIBAIJ(Mat,PetscInt,const IS[],const IS[],MatReuse,Mat *[]); 9 EXTERN PetscErrorCode MatGetValues_SeqBAIJ(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt [],PetscScalar []); 10 EXTERN PetscErrorCode MatSetValues_SeqBAIJ(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt [],const PetscScalar [],InsertMode); 11 EXTERN PetscErrorCode MatSetValuesBlocked_SeqBAIJ(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode); 12 EXTERN PetscErrorCode MatGetRow_SeqBAIJ(Mat,PetscInt,PetscInt*,PetscInt*[],PetscScalar*[]); 13 EXTERN PetscErrorCode MatRestoreRow_SeqBAIJ(Mat,PetscInt,PetscInt*,PetscInt*[],PetscScalar*[]); 14 EXTERN PetscErrorCode MatZeroRows_SeqBAIJ(Mat,PetscInt,const PetscInt[],PetscScalar); 15 16 #undef __FUNCT__ 17 #define __FUNCT__ "MatGetRowMaxAbs_MPIBAIJ" 18 PetscErrorCode MatGetRowMaxAbs_MPIBAIJ(Mat A,Vec v,PetscInt idx[]) 19 { 20 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 21 PetscErrorCode ierr; 22 PetscInt i,*idxb = 0; 23 PetscScalar *va,*vb; 24 Vec vtmp; 25 26 PetscFunctionBegin; 27 ierr = MatGetRowMaxAbs(a->A,v,idx);CHKERRQ(ierr); 28 ierr = VecGetArray(v,&va);CHKERRQ(ierr); 29 if (idx) { 30 for (i=0; i<A->rmap->n; i++) {if (PetscAbsScalar(va[i])) idx[i] += A->cmap->rstart;} 31 } 32 33 ierr = VecCreateSeq(PETSC_COMM_SELF,A->rmap->n,&vtmp);CHKERRQ(ierr); 34 if (idx) {ierr = PetscMalloc(A->rmap->n*sizeof(PetscInt),&idxb);CHKERRQ(ierr);} 35 ierr = MatGetRowMaxAbs(a->B,vtmp,idxb);CHKERRQ(ierr); 36 ierr = VecGetArray(vtmp,&vb);CHKERRQ(ierr); 37 38 for (i=0; i<A->rmap->n; i++){ 39 if (PetscAbsScalar(va[i]) < PetscAbsScalar(vb[i])) {va[i] = vb[i]; if (idx) idx[i] = A->cmap->bs*a->garray[idxb[i]/A->cmap->bs] + (idxb[i] % A->cmap->bs);} 40 } 41 42 ierr = VecRestoreArray(v,&va);CHKERRQ(ierr); 43 ierr = VecRestoreArray(vtmp,&vb);CHKERRQ(ierr); 44 if (idxb) {ierr = PetscFree(idxb);CHKERRQ(ierr);} 45 ierr = VecDestroy(vtmp);CHKERRQ(ierr); 46 PetscFunctionReturn(0); 47 } 48 49 EXTERN_C_BEGIN 50 #undef __FUNCT__ 51 #define __FUNCT__ "MatStoreValues_MPIBAIJ" 52 PetscErrorCode PETSCMAT_DLLEXPORT MatStoreValues_MPIBAIJ(Mat mat) 53 { 54 Mat_MPIBAIJ *aij = (Mat_MPIBAIJ *)mat->data; 55 PetscErrorCode ierr; 56 57 PetscFunctionBegin; 58 ierr = MatStoreValues(aij->A);CHKERRQ(ierr); 59 ierr = MatStoreValues(aij->B);CHKERRQ(ierr); 60 PetscFunctionReturn(0); 61 } 62 EXTERN_C_END 63 64 EXTERN_C_BEGIN 65 #undef __FUNCT__ 66 #define __FUNCT__ "MatRetrieveValues_MPIBAIJ" 67 PetscErrorCode PETSCMAT_DLLEXPORT MatRetrieveValues_MPIBAIJ(Mat mat) 68 { 69 Mat_MPIBAIJ *aij = (Mat_MPIBAIJ *)mat->data; 70 PetscErrorCode ierr; 71 72 PetscFunctionBegin; 73 ierr = MatRetrieveValues(aij->A);CHKERRQ(ierr); 74 ierr = MatRetrieveValues(aij->B);CHKERRQ(ierr); 75 PetscFunctionReturn(0); 76 } 77 EXTERN_C_END 78 79 /* 80 Local utility routine that creates a mapping from the global column 81 number to the local number in the off-diagonal part of the local 82 storage of the matrix. This is done in a non scable way since the 83 length of colmap equals the global matrix length. 84 */ 85 #undef __FUNCT__ 86 #define __FUNCT__ "CreateColmap_MPIBAIJ_Private" 87 PetscErrorCode CreateColmap_MPIBAIJ_Private(Mat mat) 88 { 89 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 90 Mat_SeqBAIJ *B = (Mat_SeqBAIJ*)baij->B->data; 91 PetscErrorCode ierr; 92 PetscInt nbs = B->nbs,i,bs=mat->rmap->bs; 93 94 PetscFunctionBegin; 95 #if defined (PETSC_USE_CTABLE) 96 ierr = PetscTableCreate(baij->nbs,&baij->colmap);CHKERRQ(ierr); 97 for (i=0; i<nbs; i++){ 98 ierr = PetscTableAdd(baij->colmap,baij->garray[i]+1,i*bs+1);CHKERRQ(ierr); 99 } 100 #else 101 ierr = PetscMalloc((baij->Nbs+1)*sizeof(PetscInt),&baij->colmap);CHKERRQ(ierr); 102 ierr = PetscLogObjectMemory(mat,baij->Nbs*sizeof(PetscInt));CHKERRQ(ierr); 103 ierr = PetscMemzero(baij->colmap,baij->Nbs*sizeof(PetscInt));CHKERRQ(ierr); 104 for (i=0; i<nbs; i++) baij->colmap[baij->garray[i]] = i*bs+1; 105 #endif 106 PetscFunctionReturn(0); 107 } 108 109 #define CHUNKSIZE 10 110 111 #define MatSetValues_SeqBAIJ_A_Private(row,col,value,addv) \ 112 { \ 113 \ 114 brow = row/bs; \ 115 rp = aj + ai[brow]; ap = aa + bs2*ai[brow]; \ 116 rmax = aimax[brow]; nrow = ailen[brow]; \ 117 bcol = col/bs; \ 118 ridx = row % bs; cidx = col % bs; \ 119 low = 0; high = nrow; \ 120 while (high-low > 3) { \ 121 t = (low+high)/2; \ 122 if (rp[t] > bcol) high = t; \ 123 else low = t; \ 124 } \ 125 for (_i=low; _i<high; _i++) { \ 126 if (rp[_i] > bcol) break; \ 127 if (rp[_i] == bcol) { \ 128 bap = ap + bs2*_i + bs*cidx + ridx; \ 129 if (addv == ADD_VALUES) *bap += value; \ 130 else *bap = value; \ 131 goto a_noinsert; \ 132 } \ 133 } \ 134 if (a->nonew == 1) goto a_noinsert; \ 135 if (a->nonew == -1) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) into matrix", row, col); \ 136 MatSeqXAIJReallocateAIJ(A,a->mbs,bs2,nrow,brow,bcol,rmax,aa,ai,aj,rp,ap,aimax,a->nonew,MatScalar); \ 137 N = nrow++ - 1; \ 138 /* shift up all the later entries in this row */ \ 139 for (ii=N; ii>=_i; ii--) { \ 140 rp[ii+1] = rp[ii]; \ 141 ierr = PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 142 } \ 143 if (N>=_i) { ierr = PetscMemzero(ap+bs2*_i,bs2*sizeof(MatScalar));CHKERRQ(ierr); } \ 144 rp[_i] = bcol; \ 145 ap[bs2*_i + bs*cidx + ridx] = value; \ 146 a_noinsert:; \ 147 ailen[brow] = nrow; \ 148 } 149 150 #define MatSetValues_SeqBAIJ_B_Private(row,col,value,addv) \ 151 { \ 152 brow = row/bs; \ 153 rp = bj + bi[brow]; ap = ba + bs2*bi[brow]; \ 154 rmax = bimax[brow]; nrow = bilen[brow]; \ 155 bcol = col/bs; \ 156 ridx = row % bs; cidx = col % bs; \ 157 low = 0; high = nrow; \ 158 while (high-low > 3) { \ 159 t = (low+high)/2; \ 160 if (rp[t] > bcol) high = t; \ 161 else low = t; \ 162 } \ 163 for (_i=low; _i<high; _i++) { \ 164 if (rp[_i] > bcol) break; \ 165 if (rp[_i] == bcol) { \ 166 bap = ap + bs2*_i + bs*cidx + ridx; \ 167 if (addv == ADD_VALUES) *bap += value; \ 168 else *bap = value; \ 169 goto b_noinsert; \ 170 } \ 171 } \ 172 if (b->nonew == 1) goto b_noinsert; \ 173 if (b->nonew == -1) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) into matrix", row, col); \ 174 MatSeqXAIJReallocateAIJ(B,b->mbs,bs2,nrow,brow,bcol,rmax,ba,bi,bj,rp,ap,bimax,b->nonew,MatScalar); \ 175 CHKMEMQ;\ 176 N = nrow++ - 1; \ 177 /* shift up all the later entries in this row */ \ 178 for (ii=N; ii>=_i; ii--) { \ 179 rp[ii+1] = rp[ii]; \ 180 ierr = PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));CHKERRQ(ierr); \ 181 } \ 182 if (N>=_i) { ierr = PetscMemzero(ap+bs2*_i,bs2*sizeof(MatScalar));CHKERRQ(ierr);} \ 183 rp[_i] = bcol; \ 184 ap[bs2*_i + bs*cidx + ridx] = value; \ 185 b_noinsert:; \ 186 bilen[brow] = nrow; \ 187 } 188 189 #undef __FUNCT__ 190 #define __FUNCT__ "MatSetValues_MPIBAIJ" 191 PetscErrorCode MatSetValues_MPIBAIJ(Mat mat,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode addv) 192 { 193 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 194 MatScalar value; 195 PetscTruth roworiented = baij->roworiented; 196 PetscErrorCode ierr; 197 PetscInt i,j,row,col; 198 PetscInt rstart_orig=mat->rmap->rstart; 199 PetscInt rend_orig=mat->rmap->rend,cstart_orig=mat->cmap->rstart; 200 PetscInt cend_orig=mat->cmap->rend,bs=mat->rmap->bs; 201 202 /* Some Variables required in the macro */ 203 Mat A = baij->A; 204 Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)(A)->data; 205 PetscInt *aimax=a->imax,*ai=a->i,*ailen=a->ilen,*aj=a->j; 206 MatScalar *aa=a->a; 207 208 Mat B = baij->B; 209 Mat_SeqBAIJ *b = (Mat_SeqBAIJ*)(B)->data; 210 PetscInt *bimax=b->imax,*bi=b->i,*bilen=b->ilen,*bj=b->j; 211 MatScalar *ba=b->a; 212 213 PetscInt *rp,ii,nrow,_i,rmax,N,brow,bcol; 214 PetscInt low,high,t,ridx,cidx,bs2=a->bs2; 215 MatScalar *ap,*bap; 216 217 PetscFunctionBegin; 218 for (i=0; i<m; i++) { 219 if (im[i] < 0) continue; 220 #if defined(PETSC_USE_DEBUG) 221 if (im[i] >= mat->rmap->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",im[i],mat->rmap->N-1); 222 #endif 223 if (im[i] >= rstart_orig && im[i] < rend_orig) { 224 row = im[i] - rstart_orig; 225 for (j=0; j<n; j++) { 226 if (in[j] >= cstart_orig && in[j] < cend_orig){ 227 col = in[j] - cstart_orig; 228 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 229 MatSetValues_SeqBAIJ_A_Private(row,col,value,addv); 230 /* ierr = MatSetValues_SeqBAIJ(baij->A,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 231 } else if (in[j] < 0) continue; 232 #if defined(PETSC_USE_DEBUG) 233 else if (in[j] >= mat->cmap->N) {SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[i],mat->cmap->N-1);} 234 #endif 235 else { 236 if (mat->was_assembled) { 237 if (!baij->colmap) { 238 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 239 } 240 #if defined (PETSC_USE_CTABLE) 241 ierr = PetscTableFind(baij->colmap,in[j]/bs + 1,&col);CHKERRQ(ierr); 242 col = col - 1; 243 #else 244 col = baij->colmap[in[j]/bs] - 1; 245 #endif 246 if (col < 0 && !((Mat_SeqBAIJ*)(baij->A->data))->nonew) { 247 ierr = DisAssemble_MPIBAIJ(mat);CHKERRQ(ierr); 248 col = in[j]; 249 /* Reinitialize the variables required by MatSetValues_SeqBAIJ_B_Private() */ 250 B = baij->B; 251 b = (Mat_SeqBAIJ*)(B)->data; 252 bimax=b->imax;bi=b->i;bilen=b->ilen;bj=b->j; 253 ba=b->a; 254 } else col += in[j]%bs; 255 } else col = in[j]; 256 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 257 MatSetValues_SeqBAIJ_B_Private(row,col,value,addv); 258 /* ierr = MatSetValues_SeqBAIJ(baij->B,1,&row,1,&col,&value,addv);CHKERRQ(ierr); */ 259 } 260 } 261 } else { 262 if (!baij->donotstash) { 263 if (roworiented) { 264 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n,in,v+i*n,PETSC_FALSE);CHKERRQ(ierr); 265 } else { 266 ierr = MatStashValuesCol_Private(&mat->stash,im[i],n,in,v+i,m,PETSC_FALSE);CHKERRQ(ierr); 267 } 268 } 269 } 270 } 271 PetscFunctionReturn(0); 272 } 273 274 #undef __FUNCT__ 275 #define __FUNCT__ "MatSetValuesBlocked_MPIBAIJ" 276 PetscErrorCode MatSetValuesBlocked_MPIBAIJ(Mat mat,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode addv) 277 { 278 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 279 const PetscScalar *value; 280 MatScalar *barray=baij->barray; 281 PetscTruth roworiented = baij->roworiented; 282 PetscErrorCode ierr; 283 PetscInt i,j,ii,jj,row,col,rstart=baij->rstartbs; 284 PetscInt rend=baij->rendbs,cstart=baij->cstartbs,stepval; 285 PetscInt cend=baij->cendbs,bs=mat->rmap->bs,bs2=baij->bs2; 286 287 PetscFunctionBegin; 288 if(!barray) { 289 ierr = PetscMalloc(bs2*sizeof(MatScalar),&barray);CHKERRQ(ierr); 290 baij->barray = barray; 291 } 292 293 if (roworiented) { 294 stepval = (n-1)*bs; 295 } else { 296 stepval = (m-1)*bs; 297 } 298 for (i=0; i<m; i++) { 299 if (im[i] < 0) continue; 300 #if defined(PETSC_USE_DEBUG) 301 if (im[i] >= baij->Mbs) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large, row %D max %D",im[i],baij->Mbs-1); 302 #endif 303 if (im[i] >= rstart && im[i] < rend) { 304 row = im[i] - rstart; 305 for (j=0; j<n; j++) { 306 /* If NumCol = 1 then a copy is not required */ 307 if ((roworiented) && (n == 1)) { 308 barray = (MatScalar*)v + i*bs2; 309 } else if((!roworiented) && (m == 1)) { 310 barray = (MatScalar*)v + j*bs2; 311 } else { /* Here a copy is required */ 312 if (roworiented) { 313 value = v + i*(stepval+bs)*bs + j*bs; 314 } else { 315 value = v + j*(stepval+bs)*bs + i*bs; 316 } 317 for (ii=0; ii<bs; ii++,value+=stepval) { 318 for (jj=0; jj<bs; jj++) { 319 *barray++ = *value++; 320 } 321 } 322 barray -=bs2; 323 } 324 325 if (in[j] >= cstart && in[j] < cend){ 326 col = in[j] - cstart; 327 ierr = MatSetValuesBlocked_SeqBAIJ(baij->A,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 328 } 329 else if (in[j] < 0) continue; 330 #if defined(PETSC_USE_DEBUG) 331 else if (in[j] >= baij->Nbs) {SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large, col %D max %D",in[j],baij->Nbs-1);} 332 #endif 333 else { 334 if (mat->was_assembled) { 335 if (!baij->colmap) { 336 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 337 } 338 339 #if defined(PETSC_USE_DEBUG) 340 #if defined (PETSC_USE_CTABLE) 341 { PetscInt data; 342 ierr = PetscTableFind(baij->colmap,in[j]+1,&data);CHKERRQ(ierr); 343 if ((data - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 344 } 345 #else 346 if ((baij->colmap[in[j]] - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 347 #endif 348 #endif 349 #if defined (PETSC_USE_CTABLE) 350 ierr = PetscTableFind(baij->colmap,in[j]+1,&col);CHKERRQ(ierr); 351 col = (col - 1)/bs; 352 #else 353 col = (baij->colmap[in[j]] - 1)/bs; 354 #endif 355 if (col < 0 && !((Mat_SeqBAIJ*)(baij->A->data))->nonew) { 356 ierr = DisAssemble_MPIBAIJ(mat);CHKERRQ(ierr); 357 col = in[j]; 358 } 359 } 360 else col = in[j]; 361 ierr = MatSetValuesBlocked_SeqBAIJ(baij->B,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 362 } 363 } 364 } else { 365 if (!baij->donotstash) { 366 if (roworiented) { 367 ierr = MatStashValuesRowBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 368 } else { 369 ierr = MatStashValuesColBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 370 } 371 } 372 } 373 } 374 PetscFunctionReturn(0); 375 } 376 377 #define HASH_KEY 0.6180339887 378 #define HASH(size,key,tmp) (tmp = (key)*HASH_KEY,(PetscInt)((size)*(tmp-(PetscInt)tmp))) 379 /* #define HASH(size,key) ((PetscInt)((size)*fmod(((key)*HASH_KEY),1))) */ 380 /* #define HASH(size,key,tmp) ((PetscInt)((size)*fmod(((key)*HASH_KEY),1))) */ 381 #undef __FUNCT__ 382 #define __FUNCT__ "MatSetValues_MPIBAIJ_HT" 383 PetscErrorCode MatSetValues_MPIBAIJ_HT(Mat mat,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode addv) 384 { 385 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 386 PetscTruth roworiented = baij->roworiented; 387 PetscErrorCode ierr; 388 PetscInt i,j,row,col; 389 PetscInt rstart_orig=mat->rmap->rstart; 390 PetscInt rend_orig=mat->rmap->rend,Nbs=baij->Nbs; 391 PetscInt h1,key,size=baij->ht_size,bs=mat->rmap->bs,*HT=baij->ht,idx; 392 PetscReal tmp; 393 MatScalar **HD = baij->hd,value; 394 #if defined(PETSC_USE_DEBUG) 395 PetscInt total_ct=baij->ht_total_ct,insert_ct=baij->ht_insert_ct; 396 #endif 397 398 PetscFunctionBegin; 399 400 for (i=0; i<m; i++) { 401 #if defined(PETSC_USE_DEBUG) 402 if (im[i] < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative row"); 403 if (im[i] >= mat->rmap->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",im[i],mat->rmap->N-1); 404 #endif 405 row = im[i]; 406 if (row >= rstart_orig && row < rend_orig) { 407 for (j=0; j<n; j++) { 408 col = in[j]; 409 if (roworiented) value = v[i*n+j]; else value = v[i+j*m]; 410 /* Look up PetscInto the Hash Table */ 411 key = (row/bs)*Nbs+(col/bs)+1; 412 h1 = HASH(size,key,tmp); 413 414 415 idx = h1; 416 #if defined(PETSC_USE_DEBUG) 417 insert_ct++; 418 total_ct++; 419 if (HT[idx] != key) { 420 for (idx=h1; (idx<size) && (HT[idx]!=key); idx++,total_ct++); 421 if (idx == size) { 422 for (idx=0; (idx<h1) && (HT[idx]!=key); idx++,total_ct++); 423 if (idx == h1) { 424 SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"(%D,%D) has no entry in the hash table", row, col); 425 } 426 } 427 } 428 #else 429 if (HT[idx] != key) { 430 for (idx=h1; (idx<size) && (HT[idx]!=key); idx++); 431 if (idx == size) { 432 for (idx=0; (idx<h1) && (HT[idx]!=key); idx++); 433 if (idx == h1) { 434 SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"(%D,%D) has no entry in the hash table", row, col); 435 } 436 } 437 } 438 #endif 439 /* A HASH table entry is found, so insert the values at the correct address */ 440 if (addv == ADD_VALUES) *(HD[idx]+ (col % bs)*bs + (row % bs)) += value; 441 else *(HD[idx]+ (col % bs)*bs + (row % bs)) = value; 442 } 443 } else { 444 if (!baij->donotstash) { 445 if (roworiented) { 446 ierr = MatStashValuesRow_Private(&mat->stash,im[i],n,in,v+i*n,PETSC_FALSE);CHKERRQ(ierr); 447 } else { 448 ierr = MatStashValuesCol_Private(&mat->stash,im[i],n,in,v+i,m,PETSC_FALSE);CHKERRQ(ierr); 449 } 450 } 451 } 452 } 453 #if defined(PETSC_USE_DEBUG) 454 baij->ht_total_ct = total_ct; 455 baij->ht_insert_ct = insert_ct; 456 #endif 457 PetscFunctionReturn(0); 458 } 459 460 #undef __FUNCT__ 461 #define __FUNCT__ "MatSetValuesBlocked_MPIBAIJ_HT" 462 PetscErrorCode MatSetValuesBlocked_MPIBAIJ_HT(Mat mat,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode addv) 463 { 464 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 465 PetscTruth roworiented = baij->roworiented; 466 PetscErrorCode ierr; 467 PetscInt i,j,ii,jj,row,col; 468 PetscInt rstart=baij->rstartbs; 469 PetscInt rend=mat->rmap->rend,stepval,bs=mat->rmap->bs,bs2=baij->bs2,nbs2=n*bs2; 470 PetscInt h1,key,size=baij->ht_size,idx,*HT=baij->ht,Nbs=baij->Nbs; 471 PetscReal tmp; 472 MatScalar **HD = baij->hd,*baij_a; 473 const PetscScalar *v_t,*value; 474 #if defined(PETSC_USE_DEBUG) 475 PetscInt total_ct=baij->ht_total_ct,insert_ct=baij->ht_insert_ct; 476 #endif 477 478 PetscFunctionBegin; 479 480 if (roworiented) { 481 stepval = (n-1)*bs; 482 } else { 483 stepval = (m-1)*bs; 484 } 485 for (i=0; i<m; i++) { 486 #if defined(PETSC_USE_DEBUG) 487 if (im[i] < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Negative row: %D",im[i]); 488 if (im[i] >= baij->Mbs) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",im[i],baij->Mbs-1); 489 #endif 490 row = im[i]; 491 v_t = v + i*nbs2; 492 if (row >= rstart && row < rend) { 493 for (j=0; j<n; j++) { 494 col = in[j]; 495 496 /* Look up into the Hash Table */ 497 key = row*Nbs+col+1; 498 h1 = HASH(size,key,tmp); 499 500 idx = h1; 501 #if defined(PETSC_USE_DEBUG) 502 total_ct++; 503 insert_ct++; 504 if (HT[idx] != key) { 505 for (idx=h1; (idx<size) && (HT[idx]!=key); idx++,total_ct++); 506 if (idx == size) { 507 for (idx=0; (idx<h1) && (HT[idx]!=key); idx++,total_ct++); 508 if (idx == h1) { 509 SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"(%D,%D) has no entry in the hash table", row, col); 510 } 511 } 512 } 513 #else 514 if (HT[idx] != key) { 515 for (idx=h1; (idx<size) && (HT[idx]!=key); idx++); 516 if (idx == size) { 517 for (idx=0; (idx<h1) && (HT[idx]!=key); idx++); 518 if (idx == h1) { 519 SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"(%D,%D) has no entry in the hash table", row, col); 520 } 521 } 522 } 523 #endif 524 baij_a = HD[idx]; 525 if (roworiented) { 526 /*value = v + i*(stepval+bs)*bs + j*bs;*/ 527 /* value = v + (i*(stepval+bs)+j)*bs; */ 528 value = v_t; 529 v_t += bs; 530 if (addv == ADD_VALUES) { 531 for (ii=0; ii<bs; ii++,value+=stepval) { 532 for (jj=ii; jj<bs2; jj+=bs) { 533 baij_a[jj] += *value++; 534 } 535 } 536 } else { 537 for (ii=0; ii<bs; ii++,value+=stepval) { 538 for (jj=ii; jj<bs2; jj+=bs) { 539 baij_a[jj] = *value++; 540 } 541 } 542 } 543 } else { 544 value = v + j*(stepval+bs)*bs + i*bs; 545 if (addv == ADD_VALUES) { 546 for (ii=0; ii<bs; ii++,value+=stepval,baij_a+=bs) { 547 for (jj=0; jj<bs; jj++) { 548 baij_a[jj] += *value++; 549 } 550 } 551 } else { 552 for (ii=0; ii<bs; ii++,value+=stepval,baij_a+=bs) { 553 for (jj=0; jj<bs; jj++) { 554 baij_a[jj] = *value++; 555 } 556 } 557 } 558 } 559 } 560 } else { 561 if (!baij->donotstash) { 562 if (roworiented) { 563 ierr = MatStashValuesRowBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 564 } else { 565 ierr = MatStashValuesColBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 566 } 567 } 568 } 569 } 570 #if defined(PETSC_USE_DEBUG) 571 baij->ht_total_ct = total_ct; 572 baij->ht_insert_ct = insert_ct; 573 #endif 574 PetscFunctionReturn(0); 575 } 576 577 #undef __FUNCT__ 578 #define __FUNCT__ "MatGetValues_MPIBAIJ" 579 PetscErrorCode MatGetValues_MPIBAIJ(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[]) 580 { 581 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 582 PetscErrorCode ierr; 583 PetscInt bs=mat->rmap->bs,i,j,bsrstart = mat->rmap->rstart,bsrend = mat->rmap->rend; 584 PetscInt bscstart = mat->cmap->rstart,bscend = mat->cmap->rend,row,col,data; 585 586 PetscFunctionBegin; 587 for (i=0; i<m; i++) { 588 if (idxm[i] < 0) continue; /* SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Negative row: %D",idxm[i]);*/ 589 if (idxm[i] >= mat->rmap->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",idxm[i],mat->rmap->N-1); 590 if (idxm[i] >= bsrstart && idxm[i] < bsrend) { 591 row = idxm[i] - bsrstart; 592 for (j=0; j<n; j++) { 593 if (idxn[j] < 0) continue; /* SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Negative column: %D",idxn[j]); */ 594 if (idxn[j] >= mat->cmap->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",idxn[j],mat->cmap->N-1); 595 if (idxn[j] >= bscstart && idxn[j] < bscend){ 596 col = idxn[j] - bscstart; 597 ierr = MatGetValues_SeqBAIJ(baij->A,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 598 } else { 599 if (!baij->colmap) { 600 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 601 } 602 #if defined (PETSC_USE_CTABLE) 603 ierr = PetscTableFind(baij->colmap,idxn[j]/bs+1,&data);CHKERRQ(ierr); 604 data --; 605 #else 606 data = baij->colmap[idxn[j]/bs]-1; 607 #endif 608 if((data < 0) || (baij->garray[data/bs] != idxn[j]/bs)) *(v+i*n+j) = 0.0; 609 else { 610 col = data + idxn[j]%bs; 611 ierr = MatGetValues_SeqBAIJ(baij->B,1,&row,1,&col,v+i*n+j);CHKERRQ(ierr); 612 } 613 } 614 } 615 } else { 616 SETERRQ(PETSC_ERR_SUP,"Only local values currently supported"); 617 } 618 } 619 PetscFunctionReturn(0); 620 } 621 622 #undef __FUNCT__ 623 #define __FUNCT__ "MatNorm_MPIBAIJ" 624 PetscErrorCode MatNorm_MPIBAIJ(Mat mat,NormType type,PetscReal *nrm) 625 { 626 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 627 Mat_SeqBAIJ *amat = (Mat_SeqBAIJ*)baij->A->data,*bmat = (Mat_SeqBAIJ*)baij->B->data; 628 PetscErrorCode ierr; 629 PetscInt i,j,bs2=baij->bs2,bs=baij->A->rmap->bs,nz,row,col; 630 PetscReal sum = 0.0; 631 MatScalar *v; 632 633 PetscFunctionBegin; 634 if (baij->size == 1) { 635 ierr = MatNorm(baij->A,type,nrm);CHKERRQ(ierr); 636 } else { 637 if (type == NORM_FROBENIUS) { 638 v = amat->a; 639 nz = amat->nz*bs2; 640 for (i=0; i<nz; i++) { 641 #if defined(PETSC_USE_COMPLEX) 642 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 643 #else 644 sum += (*v)*(*v); v++; 645 #endif 646 } 647 v = bmat->a; 648 nz = bmat->nz*bs2; 649 for (i=0; i<nz; i++) { 650 #if defined(PETSC_USE_COMPLEX) 651 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 652 #else 653 sum += (*v)*(*v); v++; 654 #endif 655 } 656 ierr = MPI_Allreduce(&sum,nrm,1,MPIU_REAL,MPI_SUM,((PetscObject)mat)->comm);CHKERRQ(ierr); 657 *nrm = sqrt(*nrm); 658 } else if (type == NORM_1) { /* max column sum */ 659 PetscReal *tmp,*tmp2; 660 PetscInt *jj,*garray=baij->garray,cstart=baij->rstartbs; 661 ierr = PetscMalloc((2*mat->cmap->N+1)*sizeof(PetscReal),&tmp);CHKERRQ(ierr); 662 tmp2 = tmp + mat->cmap->N; 663 ierr = PetscMemzero(tmp,mat->cmap->N*sizeof(PetscReal));CHKERRQ(ierr); 664 v = amat->a; jj = amat->j; 665 for (i=0; i<amat->nz; i++) { 666 for (j=0; j<bs; j++){ 667 col = bs*(cstart + *jj) + j; /* column index */ 668 for (row=0; row<bs; row++){ 669 tmp[col] += PetscAbsScalar(*v); v++; 670 } 671 } 672 jj++; 673 } 674 v = bmat->a; jj = bmat->j; 675 for (i=0; i<bmat->nz; i++) { 676 for (j=0; j<bs; j++){ 677 col = bs*garray[*jj] + j; 678 for (row=0; row<bs; row++){ 679 tmp[col] += PetscAbsScalar(*v); v++; 680 } 681 } 682 jj++; 683 } 684 ierr = MPI_Allreduce(tmp,tmp2,mat->cmap->N,MPIU_REAL,MPI_SUM,((PetscObject)mat)->comm);CHKERRQ(ierr); 685 *nrm = 0.0; 686 for (j=0; j<mat->cmap->N; j++) { 687 if (tmp2[j] > *nrm) *nrm = tmp2[j]; 688 } 689 ierr = PetscFree(tmp);CHKERRQ(ierr); 690 } else if (type == NORM_INFINITY) { /* max row sum */ 691 PetscReal *sums; 692 ierr = PetscMalloc(bs*sizeof(PetscReal),&sums);CHKERRQ(ierr) 693 sum = 0.0; 694 for (j=0; j<amat->mbs; j++) { 695 for (row=0; row<bs; row++) sums[row] = 0.0; 696 v = amat->a + bs2*amat->i[j]; 697 nz = amat->i[j+1]-amat->i[j]; 698 for (i=0; i<nz; i++) { 699 for (col=0; col<bs; col++){ 700 for (row=0; row<bs; row++){ 701 sums[row] += PetscAbsScalar(*v); v++; 702 } 703 } 704 } 705 v = bmat->a + bs2*bmat->i[j]; 706 nz = bmat->i[j+1]-bmat->i[j]; 707 for (i=0; i<nz; i++) { 708 for (col=0; col<bs; col++){ 709 for (row=0; row<bs; row++){ 710 sums[row] += PetscAbsScalar(*v); v++; 711 } 712 } 713 } 714 for (row=0; row<bs; row++){ 715 if (sums[row] > sum) sum = sums[row]; 716 } 717 } 718 ierr = MPI_Allreduce(&sum,nrm,1,MPIU_REAL,MPI_MAX,((PetscObject)mat)->comm);CHKERRQ(ierr); 719 ierr = PetscFree(sums);CHKERRQ(ierr); 720 } else { 721 SETERRQ(PETSC_ERR_SUP,"No support for this norm yet"); 722 } 723 } 724 PetscFunctionReturn(0); 725 } 726 727 /* 728 Creates the hash table, and sets the table 729 This table is created only once. 730 If new entried need to be added to the matrix 731 then the hash table has to be destroyed and 732 recreated. 733 */ 734 #undef __FUNCT__ 735 #define __FUNCT__ "MatCreateHashTable_MPIBAIJ_Private" 736 PetscErrorCode MatCreateHashTable_MPIBAIJ_Private(Mat mat,PetscReal factor) 737 { 738 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 739 Mat A = baij->A,B=baij->B; 740 Mat_SeqBAIJ *a=(Mat_SeqBAIJ *)A->data,*b=(Mat_SeqBAIJ *)B->data; 741 PetscInt i,j,k,nz=a->nz+b->nz,h1,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j; 742 PetscErrorCode ierr; 743 PetscInt size,bs2=baij->bs2,rstart=baij->rstartbs; 744 PetscInt cstart=baij->cstartbs,*garray=baij->garray,row,col,Nbs=baij->Nbs; 745 PetscInt *HT,key; 746 MatScalar **HD; 747 PetscReal tmp; 748 #if defined(PETSC_USE_INFO) 749 PetscInt ct=0,max=0; 750 #endif 751 752 PetscFunctionBegin; 753 baij->ht_size=(PetscInt)(factor*nz); 754 size = baij->ht_size; 755 756 if (baij->ht) { 757 PetscFunctionReturn(0); 758 } 759 760 /* Allocate Memory for Hash Table */ 761 ierr = PetscMalloc((size)*(sizeof(PetscInt)+sizeof(MatScalar*))+1,&baij->hd);CHKERRQ(ierr); 762 baij->ht = (PetscInt*)(baij->hd + size); 763 HD = baij->hd; 764 HT = baij->ht; 765 766 767 ierr = PetscMemzero(HD,size*(sizeof(PetscInt)+sizeof(PetscScalar*)));CHKERRQ(ierr); 768 769 770 /* Loop Over A */ 771 for (i=0; i<a->mbs; i++) { 772 for (j=ai[i]; j<ai[i+1]; j++) { 773 row = i+rstart; 774 col = aj[j]+cstart; 775 776 key = row*Nbs + col + 1; 777 h1 = HASH(size,key,tmp); 778 for (k=0; k<size; k++){ 779 if (!HT[(h1+k)%size]) { 780 HT[(h1+k)%size] = key; 781 HD[(h1+k)%size] = a->a + j*bs2; 782 break; 783 #if defined(PETSC_USE_INFO) 784 } else { 785 ct++; 786 #endif 787 } 788 } 789 #if defined(PETSC_USE_INFO) 790 if (k> max) max = k; 791 #endif 792 } 793 } 794 /* Loop Over B */ 795 for (i=0; i<b->mbs; i++) { 796 for (j=bi[i]; j<bi[i+1]; j++) { 797 row = i+rstart; 798 col = garray[bj[j]]; 799 key = row*Nbs + col + 1; 800 h1 = HASH(size,key,tmp); 801 for (k=0; k<size; k++){ 802 if (!HT[(h1+k)%size]) { 803 HT[(h1+k)%size] = key; 804 HD[(h1+k)%size] = b->a + j*bs2; 805 break; 806 #if defined(PETSC_USE_INFO) 807 } else { 808 ct++; 809 #endif 810 } 811 } 812 #if defined(PETSC_USE_INFO) 813 if (k> max) max = k; 814 #endif 815 } 816 } 817 818 /* Print Summary */ 819 #if defined(PETSC_USE_INFO) 820 for (i=0,j=0; i<size; i++) { 821 if (HT[i]) {j++;} 822 } 823 ierr = PetscInfo2(mat,"Average Search = %5.2f,max search = %D\n",(!j)? 0.0:((PetscReal)(ct+j))/j,max);CHKERRQ(ierr); 824 #endif 825 PetscFunctionReturn(0); 826 } 827 828 #undef __FUNCT__ 829 #define __FUNCT__ "MatAssemblyBegin_MPIBAIJ" 830 PetscErrorCode MatAssemblyBegin_MPIBAIJ(Mat mat,MatAssemblyType mode) 831 { 832 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 833 PetscErrorCode ierr; 834 PetscInt nstash,reallocs; 835 InsertMode addv; 836 837 PetscFunctionBegin; 838 if (baij->donotstash) { 839 PetscFunctionReturn(0); 840 } 841 842 /* make sure all processors are either in INSERTMODE or ADDMODE */ 843 ierr = MPI_Allreduce(&mat->insertmode,&addv,1,MPI_INT,MPI_BOR,((PetscObject)mat)->comm);CHKERRQ(ierr); 844 if (addv == (ADD_VALUES|INSERT_VALUES)) { 845 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Some processors inserted others added"); 846 } 847 mat->insertmode = addv; /* in case this processor had no cache */ 848 849 ierr = MatStashScatterBegin_Private(mat,&mat->stash,mat->rmap->range);CHKERRQ(ierr); 850 ierr = MatStashScatterBegin_Private(mat,&mat->bstash,baij->rangebs);CHKERRQ(ierr); 851 ierr = MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);CHKERRQ(ierr); 852 ierr = PetscInfo2(mat,"Stash has %D entries,uses %D mallocs.\n",nstash,reallocs);CHKERRQ(ierr); 853 ierr = MatStashGetInfo_Private(&mat->bstash,&nstash,&reallocs);CHKERRQ(ierr); 854 ierr = PetscInfo2(mat,"Block-Stash has %D entries, uses %D mallocs.\n",nstash,reallocs);CHKERRQ(ierr); 855 PetscFunctionReturn(0); 856 } 857 858 #undef __FUNCT__ 859 #define __FUNCT__ "MatAssemblyEnd_MPIBAIJ" 860 PetscErrorCode MatAssemblyEnd_MPIBAIJ(Mat mat,MatAssemblyType mode) 861 { 862 Mat_MPIBAIJ *baij=(Mat_MPIBAIJ*)mat->data; 863 Mat_SeqBAIJ *a=(Mat_SeqBAIJ*)baij->A->data; 864 PetscErrorCode ierr; 865 PetscInt i,j,rstart,ncols,flg,bs2=baij->bs2; 866 PetscInt *row,*col; 867 PetscTruth r1,r2,r3,other_disassembled; 868 MatScalar *val; 869 InsertMode addv = mat->insertmode; 870 PetscMPIInt n; 871 872 /* do not use 'b=(Mat_SeqBAIJ*)baij->B->data' as B can be reset in disassembly */ 873 PetscFunctionBegin; 874 if (!baij->donotstash) { 875 while (1) { 876 ierr = MatStashScatterGetMesg_Private(&mat->stash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 877 if (!flg) break; 878 879 for (i=0; i<n;) { 880 /* Now identify the consecutive vals belonging to the same row */ 881 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 882 if (j < n) ncols = j-i; 883 else ncols = n-i; 884 /* Now assemble all these values with a single function call */ 885 ierr = MatSetValues_MPIBAIJ(mat,1,row+i,ncols,col+i,val+i,addv);CHKERRQ(ierr); 886 i = j; 887 } 888 } 889 ierr = MatStashScatterEnd_Private(&mat->stash);CHKERRQ(ierr); 890 /* Now process the block-stash. Since the values are stashed column-oriented, 891 set the roworiented flag to column oriented, and after MatSetValues() 892 restore the original flags */ 893 r1 = baij->roworiented; 894 r2 = a->roworiented; 895 r3 = ((Mat_SeqBAIJ*)baij->B->data)->roworiented; 896 baij->roworiented = PETSC_FALSE; 897 a->roworiented = PETSC_FALSE; 898 (((Mat_SeqBAIJ*)baij->B->data))->roworiented = PETSC_FALSE; /* b->roworiented */ 899 while (1) { 900 ierr = MatStashScatterGetMesg_Private(&mat->bstash,&n,&row,&col,&val,&flg);CHKERRQ(ierr); 901 if (!flg) break; 902 903 for (i=0; i<n;) { 904 /* Now identify the consecutive vals belonging to the same row */ 905 for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; } 906 if (j < n) ncols = j-i; 907 else ncols = n-i; 908 ierr = MatSetValuesBlocked_MPIBAIJ(mat,1,row+i,ncols,col+i,val+i*bs2,addv);CHKERRQ(ierr); 909 i = j; 910 } 911 } 912 ierr = MatStashScatterEnd_Private(&mat->bstash);CHKERRQ(ierr); 913 baij->roworiented = r1; 914 a->roworiented = r2; 915 ((Mat_SeqBAIJ*)baij->B->data)->roworiented = r3; /* b->roworiented */ 916 } 917 918 ierr = MatAssemblyBegin(baij->A,mode);CHKERRQ(ierr); 919 ierr = MatAssemblyEnd(baij->A,mode);CHKERRQ(ierr); 920 921 /* determine if any processor has disassembled, if so we must 922 also disassemble ourselfs, in order that we may reassemble. */ 923 /* 924 if nonzero structure of submatrix B cannot change then we know that 925 no processor disassembled thus we can skip this stuff 926 */ 927 if (!((Mat_SeqBAIJ*)baij->B->data)->nonew) { 928 ierr = MPI_Allreduce(&mat->was_assembled,&other_disassembled,1,MPI_INT,MPI_PROD,((PetscObject)mat)->comm);CHKERRQ(ierr); 929 if (mat->was_assembled && !other_disassembled) { 930 ierr = DisAssemble_MPIBAIJ(mat);CHKERRQ(ierr); 931 } 932 } 933 934 if (!mat->was_assembled && mode == MAT_FINAL_ASSEMBLY) { 935 ierr = MatSetUpMultiply_MPIBAIJ(mat);CHKERRQ(ierr); 936 } 937 ((Mat_SeqBAIJ*)baij->B->data)->compressedrow.use = PETSC_TRUE; /* b->compressedrow.use */ 938 ierr = MatAssemblyBegin(baij->B,mode);CHKERRQ(ierr); 939 ierr = MatAssemblyEnd(baij->B,mode);CHKERRQ(ierr); 940 941 #if defined(PETSC_USE_INFO) 942 if (baij->ht && mode== MAT_FINAL_ASSEMBLY) { 943 ierr = PetscInfo1(mat,"Average Hash Table Search in MatSetValues = %5.2f\n",((PetscReal)baij->ht_total_ct)/baij->ht_insert_ct);CHKERRQ(ierr); 944 baij->ht_total_ct = 0; 945 baij->ht_insert_ct = 0; 946 } 947 #endif 948 if (baij->ht_flag && !baij->ht && mode == MAT_FINAL_ASSEMBLY) { 949 ierr = MatCreateHashTable_MPIBAIJ_Private(mat,baij->ht_fact);CHKERRQ(ierr); 950 mat->ops->setvalues = MatSetValues_MPIBAIJ_HT; 951 mat->ops->setvaluesblocked = MatSetValuesBlocked_MPIBAIJ_HT; 952 } 953 954 ierr = PetscFree(baij->rowvalues);CHKERRQ(ierr); 955 baij->rowvalues = 0; 956 PetscFunctionReturn(0); 957 } 958 959 #undef __FUNCT__ 960 #define __FUNCT__ "MatView_MPIBAIJ_ASCIIorDraworSocket" 961 static PetscErrorCode MatView_MPIBAIJ_ASCIIorDraworSocket(Mat mat,PetscViewer viewer) 962 { 963 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 964 PetscErrorCode ierr; 965 PetscMPIInt size = baij->size,rank = baij->rank; 966 PetscInt bs = mat->rmap->bs; 967 PetscTruth iascii,isdraw; 968 PetscViewer sviewer; 969 PetscViewerFormat format; 970 971 PetscFunctionBegin; 972 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr); 973 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 974 if (iascii) { 975 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 976 if (format == PETSC_VIEWER_ASCII_INFO_DETAIL) { 977 MatInfo info; 978 ierr = MPI_Comm_rank(((PetscObject)mat)->comm,&rank);CHKERRQ(ierr); 979 ierr = MatGetInfo(mat,MAT_LOCAL,&info);CHKERRQ(ierr); 980 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] Local rows %D nz %D nz alloced %D bs %D mem %D\n", 981 rank,mat->rmap->N,(PetscInt)info.nz_used*bs,(PetscInt)info.nz_allocated*bs, 982 mat->rmap->bs,(PetscInt)info.memory);CHKERRQ(ierr); 983 ierr = MatGetInfo(baij->A,MAT_LOCAL,&info);CHKERRQ(ierr); 984 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] on-diagonal part: nz %D \n",rank,(PetscInt)info.nz_used*bs);CHKERRQ(ierr); 985 ierr = MatGetInfo(baij->B,MAT_LOCAL,&info);CHKERRQ(ierr); 986 ierr = PetscViewerASCIISynchronizedPrintf(viewer,"[%d] off-diagonal part: nz %D \n",rank,(PetscInt)info.nz_used*bs);CHKERRQ(ierr); 987 ierr = PetscViewerFlush(viewer);CHKERRQ(ierr); 988 ierr = PetscViewerASCIIPrintf(viewer,"Information on VecScatter used in matrix-vector product: \n");CHKERRQ(ierr); 989 ierr = VecScatterView(baij->Mvctx,viewer);CHKERRQ(ierr); 990 PetscFunctionReturn(0); 991 } else if (format == PETSC_VIEWER_ASCII_INFO) { 992 ierr = PetscViewerASCIIPrintf(viewer," block size is %D\n",bs);CHKERRQ(ierr); 993 PetscFunctionReturn(0); 994 } else if (format == PETSC_VIEWER_ASCII_FACTOR_INFO) { 995 PetscFunctionReturn(0); 996 } 997 } 998 999 if (isdraw) { 1000 PetscDraw draw; 1001 PetscTruth isnull; 1002 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 1003 ierr = PetscDrawIsNull(draw,&isnull);CHKERRQ(ierr); if (isnull) PetscFunctionReturn(0); 1004 } 1005 1006 if (size == 1) { 1007 ierr = PetscObjectSetName((PetscObject)baij->A,((PetscObject)mat)->name);CHKERRQ(ierr); 1008 ierr = MatView(baij->A,viewer);CHKERRQ(ierr); 1009 } else { 1010 /* assemble the entire matrix onto first processor. */ 1011 Mat A; 1012 Mat_SeqBAIJ *Aloc; 1013 PetscInt M = mat->rmap->N,N = mat->cmap->N,*ai,*aj,col,i,j,k,*rvals,mbs = baij->mbs; 1014 MatScalar *a; 1015 1016 /* Here we are creating a temporary matrix, so will assume MPIBAIJ is acceptable */ 1017 /* Perhaps this should be the type of mat? */ 1018 ierr = MatCreate(((PetscObject)mat)->comm,&A);CHKERRQ(ierr); 1019 if (!rank) { 1020 ierr = MatSetSizes(A,M,N,M,N);CHKERRQ(ierr); 1021 } else { 1022 ierr = MatSetSizes(A,0,0,M,N);CHKERRQ(ierr); 1023 } 1024 ierr = MatSetType(A,MATMPIBAIJ);CHKERRQ(ierr); 1025 ierr = MatMPIBAIJSetPreallocation(A,mat->rmap->bs,0,PETSC_NULL,0,PETSC_NULL);CHKERRQ(ierr); 1026 ierr = PetscLogObjectParent(mat,A);CHKERRQ(ierr); 1027 1028 /* copy over the A part */ 1029 Aloc = (Mat_SeqBAIJ*)baij->A->data; 1030 ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1031 ierr = PetscMalloc(bs*sizeof(PetscInt),&rvals);CHKERRQ(ierr); 1032 1033 for (i=0; i<mbs; i++) { 1034 rvals[0] = bs*(baij->rstartbs + i); 1035 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 1036 for (j=ai[i]; j<ai[i+1]; j++) { 1037 col = (baij->cstartbs+aj[j])*bs; 1038 for (k=0; k<bs; k++) { 1039 ierr = MatSetValues_MPIBAIJ(A,bs,rvals,1,&col,a,INSERT_VALUES);CHKERRQ(ierr); 1040 col++; a += bs; 1041 } 1042 } 1043 } 1044 /* copy over the B part */ 1045 Aloc = (Mat_SeqBAIJ*)baij->B->data; 1046 ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1047 for (i=0; i<mbs; i++) { 1048 rvals[0] = bs*(baij->rstartbs + i); 1049 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 1050 for (j=ai[i]; j<ai[i+1]; j++) { 1051 col = baij->garray[aj[j]]*bs; 1052 for (k=0; k<bs; k++) { 1053 ierr = MatSetValues_MPIBAIJ(A,bs,rvals,1,&col,a,INSERT_VALUES);CHKERRQ(ierr); 1054 col++; a += bs; 1055 } 1056 } 1057 } 1058 ierr = PetscFree(rvals);CHKERRQ(ierr); 1059 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1060 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1061 /* 1062 Everyone has to call to draw the matrix since the graphics waits are 1063 synchronized across all processors that share the PetscDraw object 1064 */ 1065 ierr = PetscViewerGetSingleton(viewer,&sviewer);CHKERRQ(ierr); 1066 if (!rank) { 1067 ierr = PetscObjectSetName((PetscObject)((Mat_MPIBAIJ*)(A->data))->A,((PetscObject)mat)->name);CHKERRQ(ierr); 1068 ierr = MatView(((Mat_MPIBAIJ*)(A->data))->A,sviewer);CHKERRQ(ierr); 1069 } 1070 ierr = PetscViewerRestoreSingleton(viewer,&sviewer);CHKERRQ(ierr); 1071 ierr = MatDestroy(A);CHKERRQ(ierr); 1072 } 1073 PetscFunctionReturn(0); 1074 } 1075 1076 #undef __FUNCT__ 1077 #define __FUNCT__ "MatView_MPIBAIJ" 1078 PetscErrorCode MatView_MPIBAIJ(Mat mat,PetscViewer viewer) 1079 { 1080 PetscErrorCode ierr; 1081 PetscTruth iascii,isdraw,issocket,isbinary; 1082 1083 PetscFunctionBegin; 1084 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);CHKERRQ(ierr); 1085 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);CHKERRQ(ierr); 1086 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_SOCKET,&issocket);CHKERRQ(ierr); 1087 ierr = PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_BINARY,&isbinary);CHKERRQ(ierr); 1088 if (iascii || isdraw || issocket || isbinary) { 1089 ierr = MatView_MPIBAIJ_ASCIIorDraworSocket(mat,viewer);CHKERRQ(ierr); 1090 } else { 1091 SETERRQ1(PETSC_ERR_SUP,"Viewer type %s not supported by MPIBAIJ matrices",((PetscObject)viewer)->type_name); 1092 } 1093 PetscFunctionReturn(0); 1094 } 1095 1096 #undef __FUNCT__ 1097 #define __FUNCT__ "MatDestroy_MPIBAIJ" 1098 PetscErrorCode MatDestroy_MPIBAIJ(Mat mat) 1099 { 1100 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 1101 PetscErrorCode ierr; 1102 1103 PetscFunctionBegin; 1104 #if defined(PETSC_USE_LOG) 1105 PetscLogObjectState((PetscObject)mat,"Rows=%D,Cols=%D",mat->rmap->N,mat->cmap->N); 1106 #endif 1107 ierr = MatStashDestroy_Private(&mat->stash);CHKERRQ(ierr); 1108 ierr = MatStashDestroy_Private(&mat->bstash);CHKERRQ(ierr); 1109 ierr = MatDestroy(baij->A);CHKERRQ(ierr); 1110 ierr = MatDestroy(baij->B);CHKERRQ(ierr); 1111 #if defined (PETSC_USE_CTABLE) 1112 if (baij->colmap) {ierr = PetscTableDestroy(baij->colmap);CHKERRQ(ierr);} 1113 #else 1114 ierr = PetscFree(baij->colmap);CHKERRQ(ierr); 1115 #endif 1116 ierr = PetscFree(baij->garray);CHKERRQ(ierr); 1117 if (baij->lvec) {ierr = VecDestroy(baij->lvec);CHKERRQ(ierr);} 1118 if (baij->Mvctx) {ierr = VecScatterDestroy(baij->Mvctx);CHKERRQ(ierr);} 1119 ierr = PetscFree(baij->rowvalues);CHKERRQ(ierr); 1120 ierr = PetscFree(baij->barray);CHKERRQ(ierr); 1121 ierr = PetscFree(baij->hd);CHKERRQ(ierr); 1122 ierr = PetscFree(baij->rangebs);CHKERRQ(ierr); 1123 ierr = PetscFree(baij);CHKERRQ(ierr); 1124 1125 ierr = PetscObjectChangeTypeName((PetscObject)mat,0);CHKERRQ(ierr); 1126 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatStoreValues_C","",PETSC_NULL);CHKERRQ(ierr); 1127 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatRetrieveValues_C","",PETSC_NULL);CHKERRQ(ierr); 1128 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatGetDiagonalBlock_C","",PETSC_NULL);CHKERRQ(ierr); 1129 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatMPIBAIJSetPreallocation_C","",PETSC_NULL);CHKERRQ(ierr); 1130 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatMPIBAIJSetPreallocationCSR_C","",PETSC_NULL);CHKERRQ(ierr); 1131 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatDiagonalScaleLocal_C","",PETSC_NULL);CHKERRQ(ierr); 1132 ierr = PetscObjectComposeFunction((PetscObject)mat,"MatSetHashTableFactor_C","",PETSC_NULL);CHKERRQ(ierr); 1133 PetscFunctionReturn(0); 1134 } 1135 1136 #undef __FUNCT__ 1137 #define __FUNCT__ "MatMult_MPIBAIJ" 1138 PetscErrorCode MatMult_MPIBAIJ(Mat A,Vec xx,Vec yy) 1139 { 1140 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1141 PetscErrorCode ierr; 1142 PetscInt nt; 1143 1144 PetscFunctionBegin; 1145 ierr = VecGetLocalSize(xx,&nt);CHKERRQ(ierr); 1146 if (nt != A->cmap->n) { 1147 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible partition of A and xx"); 1148 } 1149 ierr = VecGetLocalSize(yy,&nt);CHKERRQ(ierr); 1150 if (nt != A->rmap->n) { 1151 SETERRQ(PETSC_ERR_ARG_SIZ,"Incompatible parition of A and yy"); 1152 } 1153 ierr = VecScatterBegin(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1154 ierr = (*a->A->ops->mult)(a->A,xx,yy);CHKERRQ(ierr); 1155 ierr = VecScatterEnd(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1156 ierr = (*a->B->ops->multadd)(a->B,a->lvec,yy,yy);CHKERRQ(ierr); 1157 PetscFunctionReturn(0); 1158 } 1159 1160 #undef __FUNCT__ 1161 #define __FUNCT__ "MatMultAdd_MPIBAIJ" 1162 PetscErrorCode MatMultAdd_MPIBAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1163 { 1164 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1165 PetscErrorCode ierr; 1166 1167 PetscFunctionBegin; 1168 ierr = VecScatterBegin(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1169 ierr = (*a->A->ops->multadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 1170 ierr = VecScatterEnd(a->Mvctx,xx,a->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1171 ierr = (*a->B->ops->multadd)(a->B,a->lvec,zz,zz);CHKERRQ(ierr); 1172 PetscFunctionReturn(0); 1173 } 1174 1175 #undef __FUNCT__ 1176 #define __FUNCT__ "MatMultTranspose_MPIBAIJ" 1177 PetscErrorCode MatMultTranspose_MPIBAIJ(Mat A,Vec xx,Vec yy) 1178 { 1179 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1180 PetscErrorCode ierr; 1181 PetscTruth merged; 1182 1183 PetscFunctionBegin; 1184 ierr = VecScatterGetMerged(a->Mvctx,&merged);CHKERRQ(ierr); 1185 /* do nondiagonal part */ 1186 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1187 if (!merged) { 1188 /* send it on its way */ 1189 ierr = VecScatterBegin(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1190 /* do local part */ 1191 ierr = (*a->A->ops->multtranspose)(a->A,xx,yy);CHKERRQ(ierr); 1192 /* receive remote parts: note this assumes the values are not actually */ 1193 /* inserted in yy until the next line */ 1194 ierr = VecScatterEnd(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1195 } else { 1196 /* do local part */ 1197 ierr = (*a->A->ops->multtranspose)(a->A,xx,yy);CHKERRQ(ierr); 1198 /* send it on its way */ 1199 ierr = VecScatterBegin(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1200 /* values actually were received in the Begin() but we need to call this nop */ 1201 ierr = VecScatterEnd(a->Mvctx,a->lvec,yy,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1202 } 1203 PetscFunctionReturn(0); 1204 } 1205 1206 #undef __FUNCT__ 1207 #define __FUNCT__ "MatMultTransposeAdd_MPIBAIJ" 1208 PetscErrorCode MatMultTransposeAdd_MPIBAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1209 { 1210 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1211 PetscErrorCode ierr; 1212 1213 PetscFunctionBegin; 1214 /* do nondiagonal part */ 1215 ierr = (*a->B->ops->multtranspose)(a->B,xx,a->lvec);CHKERRQ(ierr); 1216 /* send it on its way */ 1217 ierr = VecScatterBegin(a->Mvctx,a->lvec,zz,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1218 /* do local part */ 1219 ierr = (*a->A->ops->multtransposeadd)(a->A,xx,yy,zz);CHKERRQ(ierr); 1220 /* receive remote parts: note this assumes the values are not actually */ 1221 /* inserted in yy until the next line, which is true for my implementation*/ 1222 /* but is not perhaps always true. */ 1223 ierr = VecScatterEnd(a->Mvctx,a->lvec,zz,ADD_VALUES,SCATTER_REVERSE);CHKERRQ(ierr); 1224 PetscFunctionReturn(0); 1225 } 1226 1227 /* 1228 This only works correctly for square matrices where the subblock A->A is the 1229 diagonal block 1230 */ 1231 #undef __FUNCT__ 1232 #define __FUNCT__ "MatGetDiagonal_MPIBAIJ" 1233 PetscErrorCode MatGetDiagonal_MPIBAIJ(Mat A,Vec v) 1234 { 1235 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1236 PetscErrorCode ierr; 1237 1238 PetscFunctionBegin; 1239 if (A->rmap->N != A->cmap->N) SETERRQ(PETSC_ERR_SUP,"Supports only square matrix where A->A is diag block"); 1240 ierr = MatGetDiagonal(a->A,v);CHKERRQ(ierr); 1241 PetscFunctionReturn(0); 1242 } 1243 1244 #undef __FUNCT__ 1245 #define __FUNCT__ "MatScale_MPIBAIJ" 1246 PetscErrorCode MatScale_MPIBAIJ(Mat A,PetscScalar aa) 1247 { 1248 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1249 PetscErrorCode ierr; 1250 1251 PetscFunctionBegin; 1252 ierr = MatScale(a->A,aa);CHKERRQ(ierr); 1253 ierr = MatScale(a->B,aa);CHKERRQ(ierr); 1254 PetscFunctionReturn(0); 1255 } 1256 1257 #undef __FUNCT__ 1258 #define __FUNCT__ "MatGetRow_MPIBAIJ" 1259 PetscErrorCode MatGetRow_MPIBAIJ(Mat matin,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 1260 { 1261 Mat_MPIBAIJ *mat = (Mat_MPIBAIJ*)matin->data; 1262 PetscScalar *vworkA,*vworkB,**pvA,**pvB,*v_p; 1263 PetscErrorCode ierr; 1264 PetscInt bs = matin->rmap->bs,bs2 = mat->bs2,i,*cworkA,*cworkB,**pcA,**pcB; 1265 PetscInt nztot,nzA,nzB,lrow,brstart = matin->rmap->rstart,brend = matin->rmap->rend; 1266 PetscInt *cmap,*idx_p,cstart = mat->cstartbs; 1267 1268 PetscFunctionBegin; 1269 if (mat->getrowactive) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Already active"); 1270 mat->getrowactive = PETSC_TRUE; 1271 1272 if (!mat->rowvalues && (idx || v)) { 1273 /* 1274 allocate enough space to hold information from the longest row. 1275 */ 1276 Mat_SeqBAIJ *Aa = (Mat_SeqBAIJ*)mat->A->data,*Ba = (Mat_SeqBAIJ*)mat->B->data; 1277 PetscInt max = 1,mbs = mat->mbs,tmp; 1278 for (i=0; i<mbs; i++) { 1279 tmp = Aa->i[i+1] - Aa->i[i] + Ba->i[i+1] - Ba->i[i]; 1280 if (max < tmp) { max = tmp; } 1281 } 1282 ierr = PetscMalloc(max*bs2*(sizeof(PetscInt)+sizeof(PetscScalar)),&mat->rowvalues);CHKERRQ(ierr); 1283 mat->rowindices = (PetscInt*)(mat->rowvalues + max*bs2); 1284 } 1285 1286 if (row < brstart || row >= brend) SETERRQ(PETSC_ERR_SUP,"Only local rows") 1287 lrow = row - brstart; 1288 1289 pvA = &vworkA; pcA = &cworkA; pvB = &vworkB; pcB = &cworkB; 1290 if (!v) {pvA = 0; pvB = 0;} 1291 if (!idx) {pcA = 0; if (!v) pcB = 0;} 1292 ierr = (*mat->A->ops->getrow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1293 ierr = (*mat->B->ops->getrow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1294 nztot = nzA + nzB; 1295 1296 cmap = mat->garray; 1297 if (v || idx) { 1298 if (nztot) { 1299 /* Sort by increasing column numbers, assuming A and B already sorted */ 1300 PetscInt imark = -1; 1301 if (v) { 1302 *v = v_p = mat->rowvalues; 1303 for (i=0; i<nzB; i++) { 1304 if (cmap[cworkB[i]/bs] < cstart) v_p[i] = vworkB[i]; 1305 else break; 1306 } 1307 imark = i; 1308 for (i=0; i<nzA; i++) v_p[imark+i] = vworkA[i]; 1309 for (i=imark; i<nzB; i++) v_p[nzA+i] = vworkB[i]; 1310 } 1311 if (idx) { 1312 *idx = idx_p = mat->rowindices; 1313 if (imark > -1) { 1314 for (i=0; i<imark; i++) { 1315 idx_p[i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs; 1316 } 1317 } else { 1318 for (i=0; i<nzB; i++) { 1319 if (cmap[cworkB[i]/bs] < cstart) 1320 idx_p[i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs ; 1321 else break; 1322 } 1323 imark = i; 1324 } 1325 for (i=0; i<nzA; i++) idx_p[imark+i] = cstart*bs + cworkA[i]; 1326 for (i=imark; i<nzB; i++) idx_p[nzA+i] = cmap[cworkB[i]/bs]*bs + cworkB[i]%bs ; 1327 } 1328 } else { 1329 if (idx) *idx = 0; 1330 if (v) *v = 0; 1331 } 1332 } 1333 *nz = nztot; 1334 ierr = (*mat->A->ops->restorerow)(mat->A,lrow,&nzA,pcA,pvA);CHKERRQ(ierr); 1335 ierr = (*mat->B->ops->restorerow)(mat->B,lrow,&nzB,pcB,pvB);CHKERRQ(ierr); 1336 PetscFunctionReturn(0); 1337 } 1338 1339 #undef __FUNCT__ 1340 #define __FUNCT__ "MatRestoreRow_MPIBAIJ" 1341 PetscErrorCode MatRestoreRow_MPIBAIJ(Mat mat,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 1342 { 1343 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 1344 1345 PetscFunctionBegin; 1346 if (!baij->getrowactive) { 1347 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"MatGetRow not called"); 1348 } 1349 baij->getrowactive = PETSC_FALSE; 1350 PetscFunctionReturn(0); 1351 } 1352 1353 #undef __FUNCT__ 1354 #define __FUNCT__ "MatZeroEntries_MPIBAIJ" 1355 PetscErrorCode MatZeroEntries_MPIBAIJ(Mat A) 1356 { 1357 Mat_MPIBAIJ *l = (Mat_MPIBAIJ*)A->data; 1358 PetscErrorCode ierr; 1359 1360 PetscFunctionBegin; 1361 ierr = MatZeroEntries(l->A);CHKERRQ(ierr); 1362 ierr = MatZeroEntries(l->B);CHKERRQ(ierr); 1363 PetscFunctionReturn(0); 1364 } 1365 1366 #undef __FUNCT__ 1367 #define __FUNCT__ "MatGetInfo_MPIBAIJ" 1368 PetscErrorCode MatGetInfo_MPIBAIJ(Mat matin,MatInfoType flag,MatInfo *info) 1369 { 1370 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)matin->data; 1371 Mat A = a->A,B = a->B; 1372 PetscErrorCode ierr; 1373 PetscReal isend[5],irecv[5]; 1374 1375 PetscFunctionBegin; 1376 info->block_size = (PetscReal)matin->rmap->bs; 1377 ierr = MatGetInfo(A,MAT_LOCAL,info);CHKERRQ(ierr); 1378 isend[0] = info->nz_used; isend[1] = info->nz_allocated; isend[2] = info->nz_unneeded; 1379 isend[3] = info->memory; isend[4] = info->mallocs; 1380 ierr = MatGetInfo(B,MAT_LOCAL,info);CHKERRQ(ierr); 1381 isend[0] += info->nz_used; isend[1] += info->nz_allocated; isend[2] += info->nz_unneeded; 1382 isend[3] += info->memory; isend[4] += info->mallocs; 1383 if (flag == MAT_LOCAL) { 1384 info->nz_used = isend[0]; 1385 info->nz_allocated = isend[1]; 1386 info->nz_unneeded = isend[2]; 1387 info->memory = isend[3]; 1388 info->mallocs = isend[4]; 1389 } else if (flag == MAT_GLOBAL_MAX) { 1390 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_MAX,((PetscObject)matin)->comm);CHKERRQ(ierr); 1391 info->nz_used = irecv[0]; 1392 info->nz_allocated = irecv[1]; 1393 info->nz_unneeded = irecv[2]; 1394 info->memory = irecv[3]; 1395 info->mallocs = irecv[4]; 1396 } else if (flag == MAT_GLOBAL_SUM) { 1397 ierr = MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_SUM,((PetscObject)matin)->comm);CHKERRQ(ierr); 1398 info->nz_used = irecv[0]; 1399 info->nz_allocated = irecv[1]; 1400 info->nz_unneeded = irecv[2]; 1401 info->memory = irecv[3]; 1402 info->mallocs = irecv[4]; 1403 } else { 1404 SETERRQ1(PETSC_ERR_ARG_WRONG,"Unknown MatInfoType argument %d",(int)flag); 1405 } 1406 info->fill_ratio_given = 0; /* no parallel LU/ILU/Cholesky */ 1407 info->fill_ratio_needed = 0; 1408 info->factor_mallocs = 0; 1409 PetscFunctionReturn(0); 1410 } 1411 1412 #undef __FUNCT__ 1413 #define __FUNCT__ "MatSetOption_MPIBAIJ" 1414 PetscErrorCode MatSetOption_MPIBAIJ(Mat A,MatOption op,PetscTruth flg) 1415 { 1416 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1417 PetscErrorCode ierr; 1418 1419 PetscFunctionBegin; 1420 switch (op) { 1421 case MAT_NEW_NONZERO_LOCATIONS: 1422 case MAT_NEW_NONZERO_ALLOCATION_ERR: 1423 case MAT_UNUSED_NONZERO_LOCATION_ERR: 1424 case MAT_KEEP_NONZERO_PATTERN: 1425 case MAT_NEW_NONZERO_LOCATION_ERR: 1426 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1427 ierr = MatSetOption(a->B,op,flg);CHKERRQ(ierr); 1428 break; 1429 case MAT_ROW_ORIENTED: 1430 a->roworiented = flg; 1431 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1432 ierr = MatSetOption(a->B,op,flg);CHKERRQ(ierr); 1433 break; 1434 case MAT_NEW_DIAGONALS: 1435 ierr = PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);CHKERRQ(ierr); 1436 break; 1437 case MAT_IGNORE_OFF_PROC_ENTRIES: 1438 a->donotstash = flg; 1439 break; 1440 case MAT_USE_HASH_TABLE: 1441 a->ht_flag = flg; 1442 break; 1443 case MAT_SYMMETRIC: 1444 case MAT_STRUCTURALLY_SYMMETRIC: 1445 case MAT_HERMITIAN: 1446 case MAT_SYMMETRY_ETERNAL: 1447 ierr = MatSetOption(a->A,op,flg);CHKERRQ(ierr); 1448 break; 1449 default: 1450 SETERRQ1(PETSC_ERR_SUP,"unknown option %d",op); 1451 } 1452 PetscFunctionReturn(0); 1453 } 1454 1455 #undef __FUNCT__ 1456 #define __FUNCT__ "MatTranspose_MPIBAIJ(" 1457 PetscErrorCode MatTranspose_MPIBAIJ(Mat A,MatReuse reuse,Mat *matout) 1458 { 1459 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)A->data; 1460 Mat_SeqBAIJ *Aloc; 1461 Mat B; 1462 PetscErrorCode ierr; 1463 PetscInt M=A->rmap->N,N=A->cmap->N,*ai,*aj,i,*rvals,j,k,col; 1464 PetscInt bs=A->rmap->bs,mbs=baij->mbs; 1465 MatScalar *a; 1466 1467 PetscFunctionBegin; 1468 if (reuse == MAT_REUSE_MATRIX && A == *matout && M != N) SETERRQ(PETSC_ERR_ARG_SIZ,"Square matrix only for in-place"); 1469 if (reuse == MAT_INITIAL_MATRIX || *matout == A) { 1470 ierr = MatCreate(((PetscObject)A)->comm,&B);CHKERRQ(ierr); 1471 ierr = MatSetSizes(B,A->cmap->n,A->rmap->n,N,M);CHKERRQ(ierr); 1472 ierr = MatSetType(B,((PetscObject)A)->type_name);CHKERRQ(ierr); 1473 ierr = MatMPIBAIJSetPreallocation(B,A->rmap->bs,0,PETSC_NULL,0,PETSC_NULL);CHKERRQ(ierr); 1474 } else { 1475 B = *matout; 1476 } 1477 1478 /* copy over the A part */ 1479 Aloc = (Mat_SeqBAIJ*)baij->A->data; 1480 ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1481 ierr = PetscMalloc(bs*sizeof(PetscInt),&rvals);CHKERRQ(ierr); 1482 1483 for (i=0; i<mbs; i++) { 1484 rvals[0] = bs*(baij->rstartbs + i); 1485 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 1486 for (j=ai[i]; j<ai[i+1]; j++) { 1487 col = (baij->cstartbs+aj[j])*bs; 1488 for (k=0; k<bs; k++) { 1489 ierr = MatSetValues_MPIBAIJ(B,1,&col,bs,rvals,a,INSERT_VALUES);CHKERRQ(ierr); 1490 col++; a += bs; 1491 } 1492 } 1493 } 1494 /* copy over the B part */ 1495 Aloc = (Mat_SeqBAIJ*)baij->B->data; 1496 ai = Aloc->i; aj = Aloc->j; a = Aloc->a; 1497 for (i=0; i<mbs; i++) { 1498 rvals[0] = bs*(baij->rstartbs + i); 1499 for (j=1; j<bs; j++) { rvals[j] = rvals[j-1] + 1; } 1500 for (j=ai[i]; j<ai[i+1]; j++) { 1501 col = baij->garray[aj[j]]*bs; 1502 for (k=0; k<bs; k++) { 1503 ierr = MatSetValues_MPIBAIJ(B,1,&col,bs,rvals,a,INSERT_VALUES);CHKERRQ(ierr); 1504 col++; a += bs; 1505 } 1506 } 1507 } 1508 ierr = PetscFree(rvals);CHKERRQ(ierr); 1509 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1510 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1511 1512 if (reuse == MAT_INITIAL_MATRIX || *matout != A) { 1513 *matout = B; 1514 } else { 1515 ierr = MatHeaderCopy(A,B);CHKERRQ(ierr); 1516 } 1517 PetscFunctionReturn(0); 1518 } 1519 1520 #undef __FUNCT__ 1521 #define __FUNCT__ "MatDiagonalScale_MPIBAIJ" 1522 PetscErrorCode MatDiagonalScale_MPIBAIJ(Mat mat,Vec ll,Vec rr) 1523 { 1524 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 1525 Mat a = baij->A,b = baij->B; 1526 PetscErrorCode ierr; 1527 PetscInt s1,s2,s3; 1528 1529 PetscFunctionBegin; 1530 ierr = MatGetLocalSize(mat,&s2,&s3);CHKERRQ(ierr); 1531 if (rr) { 1532 ierr = VecGetLocalSize(rr,&s1);CHKERRQ(ierr); 1533 if (s1!=s3) SETERRQ(PETSC_ERR_ARG_SIZ,"right vector non-conforming local size"); 1534 /* Overlap communication with computation. */ 1535 ierr = VecScatterBegin(baij->Mvctx,rr,baij->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1536 } 1537 if (ll) { 1538 ierr = VecGetLocalSize(ll,&s1);CHKERRQ(ierr); 1539 if (s1!=s2) SETERRQ(PETSC_ERR_ARG_SIZ,"left vector non-conforming local size"); 1540 ierr = (*b->ops->diagonalscale)(b,ll,PETSC_NULL);CHKERRQ(ierr); 1541 } 1542 /* scale the diagonal block */ 1543 ierr = (*a->ops->diagonalscale)(a,ll,rr);CHKERRQ(ierr); 1544 1545 if (rr) { 1546 /* Do a scatter end and then right scale the off-diagonal block */ 1547 ierr = VecScatterEnd(baij->Mvctx,rr,baij->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 1548 ierr = (*b->ops->diagonalscale)(b,PETSC_NULL,baij->lvec);CHKERRQ(ierr); 1549 } 1550 1551 PetscFunctionReturn(0); 1552 } 1553 1554 #undef __FUNCT__ 1555 #define __FUNCT__ "MatZeroRows_MPIBAIJ" 1556 PetscErrorCode MatZeroRows_MPIBAIJ(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag) 1557 { 1558 Mat_MPIBAIJ *l = (Mat_MPIBAIJ*)A->data; 1559 PetscErrorCode ierr; 1560 PetscMPIInt imdex,size = l->size,n,rank = l->rank; 1561 PetscInt i,*owners = A->rmap->range; 1562 PetscInt *nprocs,j,idx,nsends,row; 1563 PetscInt nmax,*svalues,*starts,*owner,nrecvs; 1564 PetscInt *rvalues,tag = ((PetscObject)A)->tag,count,base,slen,*source,lastidx = -1; 1565 PetscInt *lens,*lrows,*values,rstart_bs=A->rmap->rstart; 1566 MPI_Comm comm = ((PetscObject)A)->comm; 1567 MPI_Request *send_waits,*recv_waits; 1568 MPI_Status recv_status,*send_status; 1569 #if defined(PETSC_DEBUG) 1570 PetscTruth found = PETSC_FALSE; 1571 #endif 1572 1573 PetscFunctionBegin; 1574 /* first count number of contributors to each processor */ 1575 ierr = PetscMalloc(2*size*sizeof(PetscInt),&nprocs);CHKERRQ(ierr); 1576 ierr = PetscMemzero(nprocs,2*size*sizeof(PetscInt));CHKERRQ(ierr); 1577 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&owner);CHKERRQ(ierr); /* see note*/ 1578 j = 0; 1579 for (i=0; i<N; i++) { 1580 if (lastidx > (idx = rows[i])) j = 0; 1581 lastidx = idx; 1582 for (; j<size; j++) { 1583 if (idx >= owners[j] && idx < owners[j+1]) { 1584 nprocs[2*j]++; 1585 nprocs[2*j+1] = 1; 1586 owner[i] = j; 1587 #if defined(PETSC_DEBUG) 1588 found = PETSC_TRUE; 1589 #endif 1590 break; 1591 } 1592 } 1593 #if defined(PETSC_DEBUG) 1594 if (!found) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Index out of range"); 1595 found = PETSC_FALSE; 1596 #endif 1597 } 1598 nsends = 0; for (i=0; i<size; i++) { nsends += nprocs[2*i+1];} 1599 1600 /* inform other processors of number of messages and max length*/ 1601 ierr = PetscMaxSum(comm,nprocs,&nmax,&nrecvs);CHKERRQ(ierr); 1602 1603 /* post receives: */ 1604 ierr = PetscMalloc((nrecvs+1)*(nmax+1)*sizeof(PetscInt),&rvalues);CHKERRQ(ierr); 1605 ierr = PetscMalloc((nrecvs+1)*sizeof(MPI_Request),&recv_waits);CHKERRQ(ierr); 1606 for (i=0; i<nrecvs; i++) { 1607 ierr = MPI_Irecv(rvalues+nmax*i,nmax,MPIU_INT,MPI_ANY_SOURCE,tag,comm,recv_waits+i);CHKERRQ(ierr); 1608 } 1609 1610 /* do sends: 1611 1) starts[i] gives the starting index in svalues for stuff going to 1612 the ith processor 1613 */ 1614 ierr = PetscMalloc((N+1)*sizeof(PetscInt),&svalues);CHKERRQ(ierr); 1615 ierr = PetscMalloc((nsends+1)*sizeof(MPI_Request),&send_waits);CHKERRQ(ierr); 1616 ierr = PetscMalloc((size+1)*sizeof(PetscInt),&starts);CHKERRQ(ierr); 1617 starts[0] = 0; 1618 for (i=1; i<size; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 1619 for (i=0; i<N; i++) { 1620 svalues[starts[owner[i]]++] = rows[i]; 1621 } 1622 1623 starts[0] = 0; 1624 for (i=1; i<size+1; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];} 1625 count = 0; 1626 for (i=0; i<size; i++) { 1627 if (nprocs[2*i+1]) { 1628 ierr = MPI_Isend(svalues+starts[i],nprocs[2*i],MPIU_INT,i,tag,comm,send_waits+count++);CHKERRQ(ierr); 1629 } 1630 } 1631 ierr = PetscFree(starts);CHKERRQ(ierr); 1632 1633 base = owners[rank]; 1634 1635 /* wait on receives */ 1636 ierr = PetscMalloc(2*(nrecvs+1)*sizeof(PetscInt),&lens);CHKERRQ(ierr); 1637 source = lens + nrecvs; 1638 count = nrecvs; slen = 0; 1639 while (count) { 1640 ierr = MPI_Waitany(nrecvs,recv_waits,&imdex,&recv_status);CHKERRQ(ierr); 1641 /* unpack receives into our local space */ 1642 ierr = MPI_Get_count(&recv_status,MPIU_INT,&n);CHKERRQ(ierr); 1643 source[imdex] = recv_status.MPI_SOURCE; 1644 lens[imdex] = n; 1645 slen += n; 1646 count--; 1647 } 1648 ierr = PetscFree(recv_waits);CHKERRQ(ierr); 1649 1650 /* move the data into the send scatter */ 1651 ierr = PetscMalloc((slen+1)*sizeof(PetscInt),&lrows);CHKERRQ(ierr); 1652 count = 0; 1653 for (i=0; i<nrecvs; i++) { 1654 values = rvalues + i*nmax; 1655 for (j=0; j<lens[i]; j++) { 1656 lrows[count++] = values[j] - base; 1657 } 1658 } 1659 ierr = PetscFree(rvalues);CHKERRQ(ierr); 1660 ierr = PetscFree(lens);CHKERRQ(ierr); 1661 ierr = PetscFree(owner);CHKERRQ(ierr); 1662 ierr = PetscFree(nprocs);CHKERRQ(ierr); 1663 1664 /* actually zap the local rows */ 1665 /* 1666 Zero the required rows. If the "diagonal block" of the matrix 1667 is square and the user wishes to set the diagonal we use separate 1668 code so that MatSetValues() is not called for each diagonal allocating 1669 new memory, thus calling lots of mallocs and slowing things down. 1670 1671 */ 1672 /* must zero l->B before l->A because the (diag) case below may put values into l->B*/ 1673 ierr = MatZeroRows_SeqBAIJ(l->B,slen,lrows,0.0);CHKERRQ(ierr); 1674 if ((diag != 0.0) && (l->A->rmap->N == l->A->cmap->N)) { 1675 ierr = MatZeroRows_SeqBAIJ(l->A,slen,lrows,diag);CHKERRQ(ierr); 1676 } else if (diag != 0.0) { 1677 ierr = MatZeroRows_SeqBAIJ(l->A,slen,lrows,0.0);CHKERRQ(ierr); 1678 if (((Mat_SeqBAIJ*)l->A->data)->nonew) { 1679 SETERRQ(PETSC_ERR_SUP,"MatZeroRows() on rectangular matrices cannot be used with the Mat options \n\ 1680 MAT_NEW_NONZERO_LOCATIONS,MAT_NEW_NONZERO_LOCATION_ERR,MAT_NEW_NONZERO_ALLOCATION_ERR"); 1681 } 1682 for (i=0; i<slen; i++) { 1683 row = lrows[i] + rstart_bs; 1684 ierr = MatSetValues(A,1,&row,1,&row,&diag,INSERT_VALUES);CHKERRQ(ierr); 1685 } 1686 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1687 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1688 } else { 1689 ierr = MatZeroRows_SeqBAIJ(l->A,slen,lrows,0.0);CHKERRQ(ierr); 1690 } 1691 1692 ierr = PetscFree(lrows);CHKERRQ(ierr); 1693 1694 /* wait on sends */ 1695 if (nsends) { 1696 ierr = PetscMalloc(nsends*sizeof(MPI_Status),&send_status);CHKERRQ(ierr); 1697 ierr = MPI_Waitall(nsends,send_waits,send_status);CHKERRQ(ierr); 1698 ierr = PetscFree(send_status);CHKERRQ(ierr); 1699 } 1700 ierr = PetscFree(send_waits);CHKERRQ(ierr); 1701 ierr = PetscFree(svalues);CHKERRQ(ierr); 1702 1703 PetscFunctionReturn(0); 1704 } 1705 1706 #undef __FUNCT__ 1707 #define __FUNCT__ "MatSetUnfactored_MPIBAIJ" 1708 PetscErrorCode MatSetUnfactored_MPIBAIJ(Mat A) 1709 { 1710 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1711 PetscErrorCode ierr; 1712 1713 PetscFunctionBegin; 1714 ierr = MatSetUnfactored(a->A);CHKERRQ(ierr); 1715 PetscFunctionReturn(0); 1716 } 1717 1718 static PetscErrorCode MatDuplicate_MPIBAIJ(Mat,MatDuplicateOption,Mat *); 1719 1720 #undef __FUNCT__ 1721 #define __FUNCT__ "MatEqual_MPIBAIJ" 1722 PetscErrorCode MatEqual_MPIBAIJ(Mat A,Mat B,PetscTruth *flag) 1723 { 1724 Mat_MPIBAIJ *matB = (Mat_MPIBAIJ*)B->data,*matA = (Mat_MPIBAIJ*)A->data; 1725 Mat a,b,c,d; 1726 PetscTruth flg; 1727 PetscErrorCode ierr; 1728 1729 PetscFunctionBegin; 1730 a = matA->A; b = matA->B; 1731 c = matB->A; d = matB->B; 1732 1733 ierr = MatEqual(a,c,&flg);CHKERRQ(ierr); 1734 if (flg) { 1735 ierr = MatEqual(b,d,&flg);CHKERRQ(ierr); 1736 } 1737 ierr = MPI_Allreduce(&flg,flag,1,MPI_INT,MPI_LAND,((PetscObject)A)->comm);CHKERRQ(ierr); 1738 PetscFunctionReturn(0); 1739 } 1740 1741 #undef __FUNCT__ 1742 #define __FUNCT__ "MatCopy_MPIBAIJ" 1743 PetscErrorCode MatCopy_MPIBAIJ(Mat A,Mat B,MatStructure str) 1744 { 1745 PetscErrorCode ierr; 1746 Mat_MPIBAIJ *a = (Mat_MPIBAIJ *)A->data; 1747 Mat_MPIBAIJ *b = (Mat_MPIBAIJ *)B->data; 1748 1749 PetscFunctionBegin; 1750 /* If the two matrices don't have the same copy implementation, they aren't compatible for fast copy. */ 1751 if ((str != SAME_NONZERO_PATTERN) || (A->ops->copy != B->ops->copy)) { 1752 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 1753 } else { 1754 ierr = MatCopy(a->A,b->A,str);CHKERRQ(ierr); 1755 ierr = MatCopy(a->B,b->B,str);CHKERRQ(ierr); 1756 } 1757 PetscFunctionReturn(0); 1758 } 1759 1760 #undef __FUNCT__ 1761 #define __FUNCT__ "MatSetUpPreallocation_MPIBAIJ" 1762 PetscErrorCode MatSetUpPreallocation_MPIBAIJ(Mat A) 1763 { 1764 PetscErrorCode ierr; 1765 1766 PetscFunctionBegin; 1767 ierr = MatMPIBAIJSetPreallocation(A,-PetscMax(A->rmap->bs,1),PETSC_DEFAULT,0,PETSC_DEFAULT,0);CHKERRQ(ierr); 1768 PetscFunctionReturn(0); 1769 } 1770 1771 #include "petscblaslapack.h" 1772 #undef __FUNCT__ 1773 #define __FUNCT__ "MatAXPY_MPIBAIJ" 1774 PetscErrorCode MatAXPY_MPIBAIJ(Mat Y,PetscScalar a,Mat X,MatStructure str) 1775 { 1776 PetscErrorCode ierr; 1777 Mat_MPIBAIJ *xx=(Mat_MPIBAIJ *)X->data,*yy=(Mat_MPIBAIJ *)Y->data; 1778 PetscBLASInt bnz,one=1; 1779 Mat_SeqBAIJ *x,*y; 1780 1781 PetscFunctionBegin; 1782 if (str == SAME_NONZERO_PATTERN) { 1783 PetscScalar alpha = a; 1784 x = (Mat_SeqBAIJ *)xx->A->data; 1785 y = (Mat_SeqBAIJ *)yy->A->data; 1786 bnz = PetscBLASIntCast(x->nz); 1787 BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one); 1788 x = (Mat_SeqBAIJ *)xx->B->data; 1789 y = (Mat_SeqBAIJ *)yy->B->data; 1790 bnz = PetscBLASIntCast(x->nz); 1791 BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one); 1792 } else { 1793 ierr = MatAXPY_Basic(Y,a,X,str);CHKERRQ(ierr); 1794 } 1795 PetscFunctionReturn(0); 1796 } 1797 1798 #undef __FUNCT__ 1799 #define __FUNCT__ "MatRealPart_MPIBAIJ" 1800 PetscErrorCode MatRealPart_MPIBAIJ(Mat A) 1801 { 1802 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1803 PetscErrorCode ierr; 1804 1805 PetscFunctionBegin; 1806 ierr = MatRealPart(a->A);CHKERRQ(ierr); 1807 ierr = MatRealPart(a->B);CHKERRQ(ierr); 1808 PetscFunctionReturn(0); 1809 } 1810 1811 #undef __FUNCT__ 1812 #define __FUNCT__ "MatImaginaryPart_MPIBAIJ" 1813 PetscErrorCode MatImaginaryPart_MPIBAIJ(Mat A) 1814 { 1815 Mat_MPIBAIJ *a = (Mat_MPIBAIJ*)A->data; 1816 PetscErrorCode ierr; 1817 1818 PetscFunctionBegin; 1819 ierr = MatImaginaryPart(a->A);CHKERRQ(ierr); 1820 ierr = MatImaginaryPart(a->B);CHKERRQ(ierr); 1821 PetscFunctionReturn(0); 1822 } 1823 1824 #undef __FUNCT__ 1825 #define __FUNCT__ "MatGetSubMatrix_MPIBAIJ" 1826 PetscErrorCode MatGetSubMatrix_MPIBAIJ(Mat mat,IS isrow,IS iscol,MatReuse call,Mat *newmat) 1827 { 1828 PetscErrorCode ierr; 1829 IS iscol_local; 1830 PetscInt csize; 1831 1832 PetscFunctionBegin; 1833 ierr = ISGetLocalSize(iscol,&csize);CHKERRQ(ierr); 1834 if (call == MAT_REUSE_MATRIX) { 1835 ierr = PetscObjectQuery((PetscObject)*newmat,"ISAllGather",(PetscObject*)&iscol_local);CHKERRQ(ierr); 1836 if (!iscol_local) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Submatrix passed in was not used before, cannot reuse"); 1837 } else { 1838 ierr = ISAllGather(iscol,&iscol_local);CHKERRQ(ierr); 1839 } 1840 ierr = MatGetSubMatrix_MPIBAIJ_Private(mat,isrow,iscol_local,csize,call,newmat);CHKERRQ(ierr); 1841 if (call == MAT_INITIAL_MATRIX) { 1842 ierr = PetscObjectCompose((PetscObject)*newmat,"ISAllGather",(PetscObject)iscol_local);CHKERRQ(ierr); 1843 ierr = ISDestroy(iscol_local);CHKERRQ(ierr); 1844 } 1845 PetscFunctionReturn(0); 1846 } 1847 1848 #undef __FUNCT__ 1849 #define __FUNCT__ "MatGetSubMatrix_MPIBAIJ" 1850 /* 1851 Not great since it makes two copies of the submatrix, first an SeqBAIJ 1852 in local and then by concatenating the local matrices the end result. 1853 Writing it directly would be much like MatGetSubMatrices_MPIBAIJ() 1854 */ 1855 PetscErrorCode MatGetSubMatrix_MPIBAIJ_Private(Mat mat,IS isrow,IS iscol,PetscInt csize,MatReuse call,Mat *newmat) 1856 { 1857 PetscErrorCode ierr; 1858 PetscMPIInt rank,size; 1859 PetscInt i,m,n,rstart,row,rend,nz,*cwork,j,bs; 1860 PetscInt *ii,*jj,nlocal,*dlens,*olens,dlen,olen,jend,mglobal; 1861 Mat *local,M,Mreuse; 1862 MatScalar *vwork,*aa; 1863 MPI_Comm comm = ((PetscObject)mat)->comm; 1864 Mat_SeqBAIJ *aij; 1865 1866 1867 PetscFunctionBegin; 1868 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 1869 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 1870 1871 if (call == MAT_REUSE_MATRIX) { 1872 ierr = PetscObjectQuery((PetscObject)*newmat,"SubMatrix",(PetscObject *)&Mreuse);CHKERRQ(ierr); 1873 if (!Mreuse) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Submatrix passed in was not used before, cannot reuse"); 1874 local = &Mreuse; 1875 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_REUSE_MATRIX,&local);CHKERRQ(ierr); 1876 } else { 1877 ierr = MatGetSubMatrices(mat,1,&isrow,&iscol,MAT_INITIAL_MATRIX,&local);CHKERRQ(ierr); 1878 Mreuse = *local; 1879 ierr = PetscFree(local);CHKERRQ(ierr); 1880 } 1881 1882 /* 1883 m - number of local rows 1884 n - number of columns (same on all processors) 1885 rstart - first row in new global matrix generated 1886 */ 1887 ierr = MatGetBlockSize(mat,&bs);CHKERRQ(ierr); 1888 ierr = MatGetSize(Mreuse,&m,&n);CHKERRQ(ierr); 1889 m = m/bs; 1890 n = n/bs; 1891 1892 if (call == MAT_INITIAL_MATRIX) { 1893 aij = (Mat_SeqBAIJ*)(Mreuse)->data; 1894 ii = aij->i; 1895 jj = aij->j; 1896 1897 /* 1898 Determine the number of non-zeros in the diagonal and off-diagonal 1899 portions of the matrix in order to do correct preallocation 1900 */ 1901 1902 /* first get start and end of "diagonal" columns */ 1903 if (csize == PETSC_DECIDE) { 1904 ierr = ISGetSize(isrow,&mglobal);CHKERRQ(ierr); 1905 if (mglobal == n*bs) { /* square matrix */ 1906 nlocal = m; 1907 } else { 1908 nlocal = n/size + ((n % size) > rank); 1909 } 1910 } else { 1911 nlocal = csize/bs; 1912 } 1913 ierr = MPI_Scan(&nlocal,&rend,1,MPIU_INT,MPI_SUM,comm);CHKERRQ(ierr); 1914 rstart = rend - nlocal; 1915 if (rank == size - 1 && rend != n) { 1916 SETERRQ2(PETSC_ERR_ARG_SIZ,"Local column sizes %D do not add up to total number of columns %D",rend,n); 1917 } 1918 1919 /* next, compute all the lengths */ 1920 ierr = PetscMalloc((2*m+1)*sizeof(PetscInt),&dlens);CHKERRQ(ierr); 1921 olens = dlens + m; 1922 for (i=0; i<m; i++) { 1923 jend = ii[i+1] - ii[i]; 1924 olen = 0; 1925 dlen = 0; 1926 for (j=0; j<jend; j++) { 1927 if (*jj < rstart || *jj >= rend) olen++; 1928 else dlen++; 1929 jj++; 1930 } 1931 olens[i] = olen; 1932 dlens[i] = dlen; 1933 } 1934 ierr = MatCreate(comm,&M);CHKERRQ(ierr); 1935 ierr = MatSetSizes(M,bs*m,bs*nlocal,PETSC_DECIDE,bs*n);CHKERRQ(ierr); 1936 ierr = MatSetType(M,((PetscObject)mat)->type_name);CHKERRQ(ierr); 1937 ierr = MatMPIBAIJSetPreallocation(M,bs,0,dlens,0,olens);CHKERRQ(ierr); 1938 ierr = PetscFree(dlens);CHKERRQ(ierr); 1939 } else { 1940 PetscInt ml,nl; 1941 1942 M = *newmat; 1943 ierr = MatGetLocalSize(M,&ml,&nl);CHKERRQ(ierr); 1944 if (ml != m) SETERRQ(PETSC_ERR_ARG_SIZ,"Previous matrix must be same size/layout as request"); 1945 ierr = MatZeroEntries(M);CHKERRQ(ierr); 1946 /* 1947 The next two lines are needed so we may call MatSetValues_MPIAIJ() below directly, 1948 rather than the slower MatSetValues(). 1949 */ 1950 M->was_assembled = PETSC_TRUE; 1951 M->assembled = PETSC_FALSE; 1952 } 1953 ierr = MatSetOption(M,MAT_ROW_ORIENTED,PETSC_FALSE);CHKERRQ(ierr); 1954 ierr = MatGetOwnershipRange(M,&rstart,&rend);CHKERRQ(ierr); 1955 aij = (Mat_SeqBAIJ*)(Mreuse)->data; 1956 ii = aij->i; 1957 jj = aij->j; 1958 aa = aij->a; 1959 for (i=0; i<m; i++) { 1960 row = rstart/bs + i; 1961 nz = ii[i+1] - ii[i]; 1962 cwork = jj; jj += nz; 1963 vwork = aa; aa += nz; 1964 ierr = MatSetValuesBlocked_MPIBAIJ(M,1,&row,nz,cwork,vwork,INSERT_VALUES);CHKERRQ(ierr); 1965 } 1966 1967 ierr = MatAssemblyBegin(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1968 ierr = MatAssemblyEnd(M,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1969 *newmat = M; 1970 1971 /* save submatrix used in processor for next request */ 1972 if (call == MAT_INITIAL_MATRIX) { 1973 ierr = PetscObjectCompose((PetscObject)M,"SubMatrix",(PetscObject)Mreuse);CHKERRQ(ierr); 1974 ierr = PetscObjectDereference((PetscObject)Mreuse);CHKERRQ(ierr); 1975 } 1976 1977 PetscFunctionReturn(0); 1978 } 1979 1980 #undef __FUNCT__ 1981 #define __FUNCT__ "MatPermute_MPIBAIJ" 1982 PetscErrorCode MatPermute_MPIBAIJ(Mat A,IS rowp,IS colp,Mat *B) 1983 { 1984 MPI_Comm comm,pcomm; 1985 PetscInt first,local_size,nrows; 1986 const PetscInt *rows; 1987 PetscMPIInt size; 1988 IS crowp,growp,irowp,lrowp,lcolp,icolp; 1989 PetscErrorCode ierr; 1990 1991 PetscFunctionBegin; 1992 ierr = PetscObjectGetComm((PetscObject)A,&comm);CHKERRQ(ierr); 1993 /* make a collective version of 'rowp' */ 1994 ierr = PetscObjectGetComm((PetscObject)rowp,&pcomm);CHKERRQ(ierr); 1995 if (pcomm==comm) { 1996 crowp = rowp; 1997 } else { 1998 ierr = ISGetSize(rowp,&nrows);CHKERRQ(ierr); 1999 ierr = ISGetIndices(rowp,&rows);CHKERRQ(ierr); 2000 ierr = ISCreateGeneral(comm,nrows,rows,&crowp);CHKERRQ(ierr); 2001 ierr = ISRestoreIndices(rowp,&rows);CHKERRQ(ierr); 2002 } 2003 /* collect the global row permutation and invert it */ 2004 ierr = ISAllGather(crowp,&growp);CHKERRQ(ierr); 2005 ierr = ISSetPermutation(growp);CHKERRQ(ierr); 2006 if (pcomm!=comm) { 2007 ierr = ISDestroy(crowp);CHKERRQ(ierr); 2008 } 2009 ierr = ISInvertPermutation(growp,PETSC_DECIDE,&irowp);CHKERRQ(ierr); 2010 /* get the local target indices */ 2011 ierr = MatGetOwnershipRange(A,&first,PETSC_NULL);CHKERRQ(ierr); 2012 ierr = MatGetLocalSize(A,&local_size,PETSC_NULL);CHKERRQ(ierr); 2013 ierr = ISGetIndices(irowp,&rows);CHKERRQ(ierr); 2014 ierr = ISCreateGeneral(MPI_COMM_SELF,local_size,rows+first,&lrowp);CHKERRQ(ierr); 2015 ierr = ISRestoreIndices(irowp,&rows);CHKERRQ(ierr); 2016 ierr = ISDestroy(irowp);CHKERRQ(ierr); 2017 /* the column permutation is so much easier; 2018 make a local version of 'colp' and invert it */ 2019 ierr = PetscObjectGetComm((PetscObject)colp,&pcomm);CHKERRQ(ierr); 2020 ierr = MPI_Comm_size(pcomm,&size);CHKERRQ(ierr); 2021 if (size==1) { 2022 lcolp = colp; 2023 } else { 2024 ierr = ISGetSize(colp,&nrows);CHKERRQ(ierr); 2025 ierr = ISGetIndices(colp,&rows);CHKERRQ(ierr); 2026 ierr = ISCreateGeneral(MPI_COMM_SELF,nrows,rows,&lcolp);CHKERRQ(ierr); 2027 } 2028 ierr = ISSetPermutation(lcolp);CHKERRQ(ierr); 2029 ierr = ISInvertPermutation(lcolp,PETSC_DECIDE,&icolp);CHKERRQ(ierr); 2030 ierr = ISSetPermutation(icolp);CHKERRQ(ierr); 2031 if (size>1) { 2032 ierr = ISRestoreIndices(colp,&rows);CHKERRQ(ierr); 2033 ierr = ISDestroy(lcolp);CHKERRQ(ierr); 2034 } 2035 /* now we just get the submatrix */ 2036 ierr = MatGetSubMatrix_MPIBAIJ_Private(A,lrowp,icolp,local_size,MAT_INITIAL_MATRIX,B);CHKERRQ(ierr); 2037 /* clean up */ 2038 ierr = ISDestroy(lrowp);CHKERRQ(ierr); 2039 ierr = ISDestroy(icolp);CHKERRQ(ierr); 2040 PetscFunctionReturn(0); 2041 } 2042 2043 #undef __FUNCT__ 2044 #define __FUNCT__ "MatGetGhosts_MPIBAIJ" 2045 PetscErrorCode PETSCMAT_DLLEXPORT MatGetGhosts_MPIBAIJ(Mat mat,PetscInt *nghosts,const PetscInt *ghosts[]) 2046 { 2047 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*) mat->data; 2048 Mat_SeqBAIJ *B = (Mat_SeqBAIJ*)baij->B->data; 2049 2050 PetscFunctionBegin; 2051 if (nghosts) { *nghosts = B->nbs;} 2052 if (ghosts) {*ghosts = baij->garray;} 2053 PetscFunctionReturn(0); 2054 } 2055 2056 EXTERN PetscErrorCode CreateColmap_MPIBAIJ_Private(Mat); 2057 2058 #undef __FUNCT__ 2059 #define __FUNCT__ "MatFDColoringCreate_MPIBAIJ" 2060 /* 2061 This routine is almost identical to MatFDColoringCreate_MPIBAIJ()! 2062 */ 2063 PetscErrorCode MatFDColoringCreate_MPIBAIJ(Mat mat,ISColoring iscoloring,MatFDColoring c) 2064 { 2065 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 2066 PetscErrorCode ierr; 2067 PetscMPIInt size,*ncolsonproc,*disp,nn; 2068 PetscInt bs,i,n,nrows,j,k,m,*rows = 0,*A_ci,*A_cj,ncols,col; 2069 const PetscInt *is; 2070 PetscInt nis = iscoloring->n,nctot,*cols,*B_ci,*B_cj; 2071 PetscInt *rowhit,M,cstart,cend,colb; 2072 PetscInt *columnsforrow,l; 2073 IS *isa; 2074 PetscTruth done,flg; 2075 ISLocalToGlobalMapping map = mat->bmapping; 2076 PetscInt *ltog = (map ? map->indices : (PetscInt*) PETSC_NULL) ,ctype=c->ctype; 2077 2078 PetscFunctionBegin; 2079 if (!mat->assembled) { 2080 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Matrix must be assembled first; MatAssemblyBegin/End();"); 2081 } 2082 if (ctype == IS_COLORING_GHOSTED && !map) SETERRQ(PETSC_ERR_ARG_INCOMP,"When using ghosted differencing matrix must have local to global mapping provided with MatSetLocalToGlobalMappingBlock"); 2083 2084 ierr = ISColoringGetIS(iscoloring,PETSC_IGNORE,&isa);CHKERRQ(ierr); 2085 2086 ierr = MatGetBlockSize(mat,&bs);CHKERRQ(ierr); 2087 M = mat->rmap->n/bs; 2088 cstart = mat->cmap->rstart/bs; 2089 cend = mat->cmap->rend/bs; 2090 c->M = mat->rmap->N/bs; /* set the global rows and columns and local rows */ 2091 c->N = mat->cmap->N/bs; 2092 c->m = mat->rmap->n/bs; 2093 c->rstart = mat->rmap->rstart/bs; 2094 2095 c->ncolors = nis; 2096 ierr = PetscMalloc(nis*sizeof(PetscInt),&c->ncolumns);CHKERRQ(ierr); 2097 ierr = PetscMalloc(nis*sizeof(PetscInt*),&c->columns);CHKERRQ(ierr); 2098 ierr = PetscMalloc(nis*sizeof(PetscInt),&c->nrows);CHKERRQ(ierr); 2099 ierr = PetscMalloc(nis*sizeof(PetscInt*),&c->rows);CHKERRQ(ierr); 2100 ierr = PetscMalloc(nis*sizeof(PetscInt*),&c->columnsforrow);CHKERRQ(ierr); 2101 ierr = PetscLogObjectMemory(c,5*nis*sizeof(PetscInt));CHKERRQ(ierr); 2102 2103 /* Allow access to data structures of local part of matrix */ 2104 if (!baij->colmap) { 2105 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 2106 } 2107 ierr = MatGetColumnIJ(baij->A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&done);CHKERRQ(ierr); 2108 ierr = MatGetColumnIJ(baij->B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&done);CHKERRQ(ierr); 2109 2110 ierr = PetscMalloc((M+1)*sizeof(PetscInt),&rowhit);CHKERRQ(ierr); 2111 ierr = PetscMalloc((M+1)*sizeof(PetscInt),&columnsforrow);CHKERRQ(ierr); 2112 2113 for (i=0; i<nis; i++) { 2114 ierr = ISGetLocalSize(isa[i],&n);CHKERRQ(ierr); 2115 ierr = ISGetIndices(isa[i],&is);CHKERRQ(ierr); 2116 c->ncolumns[i] = n; 2117 if (n) { 2118 ierr = PetscMalloc(n*sizeof(PetscInt),&c->columns[i]);CHKERRQ(ierr); 2119 ierr = PetscLogObjectMemory(c,n*sizeof(PetscInt));CHKERRQ(ierr); 2120 ierr = PetscMemcpy(c->columns[i],is,n*sizeof(PetscInt));CHKERRQ(ierr); 2121 } else { 2122 c->columns[i] = 0; 2123 } 2124 2125 if (ctype == IS_COLORING_GLOBAL){ 2126 /* Determine the total (parallel) number of columns of this color */ 2127 ierr = MPI_Comm_size(((PetscObject)mat)->comm,&size);CHKERRQ(ierr); 2128 ierr = PetscMalloc(2*size*sizeof(PetscInt*),&ncolsonproc);CHKERRQ(ierr); 2129 disp = ncolsonproc + size; 2130 2131 nn = PetscMPIIntCast(n); 2132 ierr = MPI_Allgather(&nn,1,MPI_INT,ncolsonproc,1,MPI_INT,((PetscObject)mat)->comm);CHKERRQ(ierr); 2133 nctot = 0; for (j=0; j<size; j++) {nctot += ncolsonproc[j];} 2134 if (!nctot) { 2135 ierr = PetscInfo(mat,"Coloring of matrix has some unneeded colors with no corresponding rows\n");CHKERRQ(ierr); 2136 } 2137 2138 disp[0] = 0; 2139 for (j=1; j<size; j++) { 2140 disp[j] = disp[j-1] + ncolsonproc[j-1]; 2141 } 2142 2143 /* Get complete list of columns for color on each processor */ 2144 ierr = PetscMalloc((nctot+1)*sizeof(PetscInt),&cols);CHKERRQ(ierr); 2145 ierr = MPI_Allgatherv((void*)is,n,MPIU_INT,cols,ncolsonproc,disp,MPIU_INT,((PetscObject)mat)->comm);CHKERRQ(ierr); 2146 ierr = PetscFree(ncolsonproc);CHKERRQ(ierr); 2147 } else if (ctype == IS_COLORING_GHOSTED){ 2148 /* Determine local number of columns of this color on this process, including ghost points */ 2149 nctot = n; 2150 ierr = PetscMalloc((nctot+1)*sizeof(PetscInt),&cols);CHKERRQ(ierr); 2151 ierr = PetscMemcpy(cols,is,n*sizeof(PetscInt));CHKERRQ(ierr); 2152 } else { 2153 SETERRQ(PETSC_ERR_SUP,"Not provided for this MatFDColoring type"); 2154 } 2155 2156 /* 2157 Mark all rows affect by these columns 2158 */ 2159 /* Temporary option to allow for debugging/testing */ 2160 flg = PETSC_FALSE; 2161 ierr = PetscOptionsGetTruth(PETSC_NULL,"-matfdcoloring_slow",&flg,PETSC_NULL);CHKERRQ(ierr); 2162 if (!flg) {/*-----------------------------------------------------------------------------*/ 2163 /* crude, fast version */ 2164 ierr = PetscMemzero(rowhit,M*sizeof(PetscInt));CHKERRQ(ierr); 2165 /* loop over columns*/ 2166 for (j=0; j<nctot; j++) { 2167 if (ctype == IS_COLORING_GHOSTED) { 2168 col = ltog[cols[j]]; 2169 } else { 2170 col = cols[j]; 2171 } 2172 if (col >= cstart && col < cend) { 2173 /* column is in diagonal block of matrix */ 2174 rows = A_cj + A_ci[col-cstart]; 2175 m = A_ci[col-cstart+1] - A_ci[col-cstart]; 2176 } else { 2177 #if defined (PETSC_USE_CTABLE) 2178 ierr = PetscTableFind(baij->colmap,col+1,&colb);CHKERRQ(ierr) 2179 colb --; 2180 #else 2181 colb = baij->colmap[col] - 1; 2182 #endif 2183 if (colb == -1) { 2184 m = 0; 2185 } else { 2186 colb = colb/bs; 2187 rows = B_cj + B_ci[colb]; 2188 m = B_ci[colb+1] - B_ci[colb]; 2189 } 2190 } 2191 /* loop over columns marking them in rowhit */ 2192 for (k=0; k<m; k++) { 2193 rowhit[*rows++] = col + 1; 2194 } 2195 } 2196 2197 /* count the number of hits */ 2198 nrows = 0; 2199 for (j=0; j<M; j++) { 2200 if (rowhit[j]) nrows++; 2201 } 2202 c->nrows[i] = nrows; 2203 ierr = PetscMalloc((nrows+1)*sizeof(PetscInt),&c->rows[i]);CHKERRQ(ierr); 2204 ierr = PetscMalloc((nrows+1)*sizeof(PetscInt),&c->columnsforrow[i]);CHKERRQ(ierr); 2205 ierr = PetscLogObjectMemory(c,2*(nrows+1)*sizeof(PetscInt));CHKERRQ(ierr); 2206 nrows = 0; 2207 for (j=0; j<M; j++) { 2208 if (rowhit[j]) { 2209 c->rows[i][nrows] = j; 2210 c->columnsforrow[i][nrows] = rowhit[j] - 1; 2211 nrows++; 2212 } 2213 } 2214 } else {/*-------------------------------------------------------------------------------*/ 2215 /* slow version, using rowhit as a linked list */ 2216 PetscInt currentcol,fm,mfm; 2217 rowhit[M] = M; 2218 nrows = 0; 2219 /* loop over columns*/ 2220 for (j=0; j<nctot; j++) { 2221 if (ctype == IS_COLORING_GHOSTED) { 2222 col = ltog[cols[j]]; 2223 } else { 2224 col = cols[j]; 2225 } 2226 if (col >= cstart && col < cend) { 2227 /* column is in diagonal block of matrix */ 2228 rows = A_cj + A_ci[col-cstart]; 2229 m = A_ci[col-cstart+1] - A_ci[col-cstart]; 2230 } else { 2231 #if defined (PETSC_USE_CTABLE) 2232 ierr = PetscTableFind(baij->colmap,col+1,&colb);CHKERRQ(ierr); 2233 colb --; 2234 #else 2235 colb = baij->colmap[col] - 1; 2236 #endif 2237 if (colb == -1) { 2238 m = 0; 2239 } else { 2240 colb = colb/bs; 2241 rows = B_cj + B_ci[colb]; 2242 m = B_ci[colb+1] - B_ci[colb]; 2243 } 2244 } 2245 2246 /* loop over columns marking them in rowhit */ 2247 fm = M; /* fm points to first entry in linked list */ 2248 for (k=0; k<m; k++) { 2249 currentcol = *rows++; 2250 /* is it already in the list? */ 2251 do { 2252 mfm = fm; 2253 fm = rowhit[fm]; 2254 } while (fm < currentcol); 2255 /* not in list so add it */ 2256 if (fm != currentcol) { 2257 nrows++; 2258 columnsforrow[currentcol] = col; 2259 /* next three lines insert new entry into linked list */ 2260 rowhit[mfm] = currentcol; 2261 rowhit[currentcol] = fm; 2262 fm = currentcol; 2263 /* fm points to present position in list since we know the columns are sorted */ 2264 } else { 2265 SETERRQ(PETSC_ERR_PLIB,"Invalid coloring of matrix detected"); 2266 } 2267 } 2268 } 2269 c->nrows[i] = nrows; 2270 ierr = PetscMalloc((nrows+1)*sizeof(PetscInt),&c->rows[i]);CHKERRQ(ierr); 2271 ierr = PetscMalloc((nrows+1)*sizeof(PetscInt),&c->columnsforrow[i]);CHKERRQ(ierr); 2272 ierr = PetscLogObjectMemory(c,(nrows+1)*sizeof(PetscInt));CHKERRQ(ierr); 2273 /* now store the linked list of rows into c->rows[i] */ 2274 nrows = 0; 2275 fm = rowhit[M]; 2276 do { 2277 c->rows[i][nrows] = fm; 2278 c->columnsforrow[i][nrows++] = columnsforrow[fm]; 2279 fm = rowhit[fm]; 2280 } while (fm < M); 2281 } /* ---------------------------------------------------------------------------------------*/ 2282 ierr = PetscFree(cols);CHKERRQ(ierr); 2283 } 2284 2285 /* Optimize by adding the vscale, and scaleforrow[][] fields */ 2286 /* 2287 vscale will contain the "diagonal" on processor scalings followed by the off processor 2288 */ 2289 if (ctype == IS_COLORING_GLOBAL) { 2290 PetscInt *garray; 2291 ierr = PetscMalloc(baij->B->cmap->n*sizeof(PetscInt),&garray);CHKERRQ(ierr); 2292 for (i=0; i<baij->B->cmap->n/bs; i++) { 2293 for (j=0; j<bs; j++) { 2294 garray[i*bs+j] = bs*baij->garray[i]+j; 2295 } 2296 } 2297 ierr = VecCreateGhost(((PetscObject)mat)->comm,baij->A->rmap->n,PETSC_DETERMINE,baij->B->cmap->n,garray,&c->vscale);CHKERRQ(ierr); 2298 ierr = PetscFree(garray);CHKERRQ(ierr); 2299 CHKMEMQ; 2300 ierr = PetscMalloc(c->ncolors*sizeof(PetscInt*),&c->vscaleforrow);CHKERRQ(ierr); 2301 for (k=0; k<c->ncolors; k++) { 2302 ierr = PetscMalloc((c->nrows[k]+1)*sizeof(PetscInt),&c->vscaleforrow[k]);CHKERRQ(ierr); 2303 for (l=0; l<c->nrows[k]; l++) { 2304 col = c->columnsforrow[k][l]; 2305 if (col >= cstart && col < cend) { 2306 /* column is in diagonal block of matrix */ 2307 colb = col - cstart; 2308 } else { 2309 /* column is in "off-processor" part */ 2310 #if defined (PETSC_USE_CTABLE) 2311 ierr = PetscTableFind(baij->colmap,col+1,&colb);CHKERRQ(ierr); 2312 colb --; 2313 #else 2314 colb = baij->colmap[col] - 1; 2315 #endif 2316 colb = colb/bs; 2317 colb += cend - cstart; 2318 } 2319 c->vscaleforrow[k][l] = colb; 2320 } 2321 } 2322 } else if (ctype == IS_COLORING_GHOSTED) { 2323 /* Get gtol mapping */ 2324 PetscInt N = mat->cmap->N, *gtol; 2325 ierr = PetscMalloc((N+1)*sizeof(PetscInt),>ol);CHKERRQ(ierr); 2326 for (i=0; i<N; i++) gtol[i] = -1; 2327 for (i=0; i<map->n; i++) gtol[ltog[i]] = i; 2328 2329 c->vscale = 0; /* will be created in MatFDColoringApply() */ 2330 ierr = PetscMalloc(c->ncolors*sizeof(PetscInt*),&c->vscaleforrow);CHKERRQ(ierr); 2331 for (k=0; k<c->ncolors; k++) { 2332 ierr = PetscMalloc((c->nrows[k]+1)*sizeof(PetscInt),&c->vscaleforrow[k]);CHKERRQ(ierr); 2333 for (l=0; l<c->nrows[k]; l++) { 2334 col = c->columnsforrow[k][l]; /* global column index */ 2335 c->vscaleforrow[k][l] = gtol[col]; /* local column index */ 2336 } 2337 } 2338 ierr = PetscFree(gtol);CHKERRQ(ierr); 2339 } 2340 ierr = ISColoringRestoreIS(iscoloring,&isa);CHKERRQ(ierr); 2341 2342 ierr = PetscFree(rowhit);CHKERRQ(ierr); 2343 ierr = PetscFree(columnsforrow);CHKERRQ(ierr); 2344 ierr = MatRestoreColumnIJ(baij->A,0,PETSC_FALSE,PETSC_FALSE,&ncols,&A_ci,&A_cj,&done);CHKERRQ(ierr); 2345 ierr = MatRestoreColumnIJ(baij->B,0,PETSC_FALSE,PETSC_FALSE,&ncols,&B_ci,&B_cj,&done);CHKERRQ(ierr); 2346 CHKMEMQ; 2347 PetscFunctionReturn(0); 2348 } 2349 2350 #undef __FUNCT__ 2351 #define __FUNCT__ "MatGetSeqNonzerostructure_MPIBAIJ" 2352 PetscErrorCode MatGetSeqNonzerostructure_MPIBAIJ(Mat A,Mat *newmat) 2353 { 2354 Mat B; 2355 Mat_MPIBAIJ *a = (Mat_MPIBAIJ *)A->data; 2356 Mat_SeqBAIJ *ad = (Mat_SeqBAIJ*)a->A->data,*bd = (Mat_SeqBAIJ*)a->B->data; 2357 Mat_SeqAIJ *b; 2358 PetscErrorCode ierr; 2359 PetscMPIInt size,rank,*recvcounts = 0,*displs = 0; 2360 PetscInt sendcount,i,*rstarts = A->rmap->range,n,cnt,j,bs = A->rmap->bs; 2361 PetscInt m,*garray = a->garray,*lens,*jsendbuf,*a_jsendbuf,*b_jsendbuf; 2362 2363 PetscFunctionBegin; 2364 ierr = MPI_Comm_size(((PetscObject)A)->comm,&size);CHKERRQ(ierr); 2365 ierr = MPI_Comm_rank(((PetscObject)A)->comm,&rank);CHKERRQ(ierr); 2366 2367 /* ---------------------------------------------------------------- 2368 Tell every processor the number of nonzeros per row 2369 */ 2370 ierr = PetscMalloc((A->rmap->N/bs)*sizeof(PetscInt),&lens);CHKERRQ(ierr); 2371 for (i=A->rmap->rstart/bs; i<A->rmap->rend/bs; i++) { 2372 lens[i] = ad->i[i-A->rmap->rstart/bs+1] - ad->i[i-A->rmap->rstart/bs] + bd->i[i-A->rmap->rstart/bs+1] - bd->i[i-A->rmap->rstart/bs]; 2373 } 2374 sendcount = A->rmap->rend/bs - A->rmap->rstart/bs; 2375 ierr = PetscMalloc(2*size*sizeof(PetscMPIInt),&recvcounts);CHKERRQ(ierr); 2376 displs = recvcounts + size; 2377 for (i=0; i<size; i++) { 2378 recvcounts[i] = A->rmap->range[i+1]/bs - A->rmap->range[i]/bs; 2379 displs[i] = A->rmap->range[i]/bs; 2380 } 2381 #if defined(PETSC_HAVE_MPI_IN_PLACE) 2382 ierr = MPI_Allgatherv(MPI_IN_PLACE,0,MPI_DATATYPE_NULL,lens,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);CHKERRQ(ierr); 2383 #else 2384 ierr = MPI_Allgatherv(lens+A->rmap->rstart/bs,sendcount,MPIU_INT,lens,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);CHKERRQ(ierr); 2385 #endif 2386 /* --------------------------------------------------------------- 2387 Create the sequential matrix of the same type as the local block diagonal 2388 */ 2389 ierr = MatCreate(PETSC_COMM_SELF,&B);CHKERRQ(ierr); 2390 ierr = MatSetSizes(B,A->rmap->N/bs,A->cmap->N/bs,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 2391 ierr = MatSetType(B,MATSEQAIJ);CHKERRQ(ierr); 2392 ierr = MatSeqAIJSetPreallocation(B,0,lens);CHKERRQ(ierr); 2393 b = (Mat_SeqAIJ *)B->data; 2394 2395 /*-------------------------------------------------------------------- 2396 Copy my part of matrix column indices over 2397 */ 2398 sendcount = ad->nz + bd->nz; 2399 jsendbuf = b->j + b->i[rstarts[rank]/bs]; 2400 a_jsendbuf = ad->j; 2401 b_jsendbuf = bd->j; 2402 n = A->rmap->rend/bs - A->rmap->rstart/bs; 2403 cnt = 0; 2404 for (i=0; i<n; i++) { 2405 2406 /* put in lower diagonal portion */ 2407 m = bd->i[i+1] - bd->i[i]; 2408 while (m > 0) { 2409 /* is it above diagonal (in bd (compressed) numbering) */ 2410 if (garray[*b_jsendbuf] > A->rmap->rstart/bs + i) break; 2411 jsendbuf[cnt++] = garray[*b_jsendbuf++]; 2412 m--; 2413 } 2414 2415 /* put in diagonal portion */ 2416 for (j=ad->i[i]; j<ad->i[i+1]; j++) { 2417 jsendbuf[cnt++] = A->rmap->rstart/bs + *a_jsendbuf++; 2418 } 2419 2420 /* put in upper diagonal portion */ 2421 while (m-- > 0) { 2422 jsendbuf[cnt++] = garray[*b_jsendbuf++]; 2423 } 2424 } 2425 if (cnt != sendcount) SETERRQ2(PETSC_ERR_PLIB,"Corrupted PETSc matrix: nz given %D actual nz %D",sendcount,cnt); 2426 2427 /*-------------------------------------------------------------------- 2428 Gather all column indices to all processors 2429 */ 2430 for (i=0; i<size; i++) { 2431 recvcounts[i] = 0; 2432 for (j=A->rmap->range[i]/bs; j<A->rmap->range[i+1]/bs; j++) { 2433 recvcounts[i] += lens[j]; 2434 } 2435 } 2436 displs[0] = 0; 2437 for (i=1; i<size; i++) { 2438 displs[i] = displs[i-1] + recvcounts[i-1]; 2439 } 2440 #if defined(PETSC_HAVE_MPI_IN_PLACE) 2441 ierr = MPI_Allgatherv(MPI_IN_PLACE,0,MPI_DATATYPE_NULL,b->j,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);CHKERRQ(ierr); 2442 #else 2443 ierr = MPI_Allgatherv(jsendbuf,sendcount,MPIU_INT,b->j,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);CHKERRQ(ierr); 2444 #endif 2445 /*-------------------------------------------------------------------- 2446 Assemble the matrix into useable form (note numerical values not yet set) 2447 */ 2448 /* set the b->ilen (length of each row) values */ 2449 ierr = PetscMemcpy(b->ilen,lens,(A->rmap->N/bs)*sizeof(PetscInt));CHKERRQ(ierr); 2450 /* set the b->i indices */ 2451 b->i[0] = 0; 2452 for (i=1; i<=A->rmap->N/bs; i++) { 2453 b->i[i] = b->i[i-1] + lens[i-1]; 2454 } 2455 ierr = PetscFree(lens);CHKERRQ(ierr); 2456 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2457 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2458 ierr = PetscFree(recvcounts);CHKERRQ(ierr); 2459 2460 if (A->symmetric){ 2461 ierr = MatSetOption(B,MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 2462 } else if (A->hermitian) { 2463 ierr = MatSetOption(B,MAT_HERMITIAN,PETSC_TRUE);CHKERRQ(ierr); 2464 } else if (A->structurally_symmetric) { 2465 ierr = MatSetOption(B,MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 2466 } 2467 *newmat = B; 2468 PetscFunctionReturn(0); 2469 } 2470 2471 #undef __FUNCT__ 2472 #define __FUNCT__ "MatSOR_MPIBAIJ" 2473 PetscErrorCode MatSOR_MPIBAIJ(Mat matin,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx) 2474 { 2475 Mat_MPIBAIJ *mat = (Mat_MPIBAIJ*)matin->data; 2476 PetscErrorCode ierr; 2477 Vec bb1 = 0; 2478 2479 PetscFunctionBegin; 2480 if (its > 1 || ~flag & SOR_ZERO_INITIAL_GUESS) { 2481 ierr = VecDuplicate(bb,&bb1);CHKERRQ(ierr); 2482 } 2483 2484 if (flag == SOR_APPLY_UPPER) { 2485 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 2486 PetscFunctionReturn(0); 2487 } 2488 2489 if ((flag & SOR_LOCAL_SYMMETRIC_SWEEP) == SOR_LOCAL_SYMMETRIC_SWEEP){ 2490 if (flag & SOR_ZERO_INITIAL_GUESS) { 2491 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 2492 its--; 2493 } 2494 2495 while (its--) { 2496 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2497 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2498 2499 /* update rhs: bb1 = bb - B*x */ 2500 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 2501 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 2502 2503 /* local sweep */ 2504 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_SYMMETRIC_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 2505 } 2506 } else if (flag & SOR_LOCAL_FORWARD_SWEEP){ 2507 if (flag & SOR_ZERO_INITIAL_GUESS) { 2508 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 2509 its--; 2510 } 2511 while (its--) { 2512 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2513 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2514 2515 /* update rhs: bb1 = bb - B*x */ 2516 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 2517 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 2518 2519 /* local sweep */ 2520 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_FORWARD_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 2521 } 2522 } else if (flag & SOR_LOCAL_BACKWARD_SWEEP){ 2523 if (flag & SOR_ZERO_INITIAL_GUESS) { 2524 ierr = (*mat->A->ops->sor)(mat->A,bb,omega,flag,fshift,lits,1,xx);CHKERRQ(ierr); 2525 its--; 2526 } 2527 while (its--) { 2528 ierr = VecScatterBegin(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2529 ierr = VecScatterEnd(mat->Mvctx,xx,mat->lvec,INSERT_VALUES,SCATTER_FORWARD);CHKERRQ(ierr); 2530 2531 /* update rhs: bb1 = bb - B*x */ 2532 ierr = VecScale(mat->lvec,-1.0);CHKERRQ(ierr); 2533 ierr = (*mat->B->ops->multadd)(mat->B,mat->lvec,bb,bb1);CHKERRQ(ierr); 2534 2535 /* local sweep */ 2536 ierr = (*mat->A->ops->sor)(mat->A,bb1,omega,SOR_BACKWARD_SWEEP,fshift,lits,1,xx);CHKERRQ(ierr); 2537 } 2538 } else { 2539 SETERRQ(PETSC_ERR_SUP,"Parallel version of SOR requested not supported"); 2540 } 2541 2542 if (bb1) {ierr = VecDestroy(bb1);CHKERRQ(ierr);} 2543 PetscFunctionReturn(0); 2544 } 2545 2546 extern PetscErrorCode PETSCMAT_DLLEXPORT MatFDColoringApply_BAIJ(Mat,MatFDColoring,Vec,MatStructure*,void*); 2547 2548 2549 /* -------------------------------------------------------------------*/ 2550 static struct _MatOps MatOps_Values = { 2551 MatSetValues_MPIBAIJ, 2552 MatGetRow_MPIBAIJ, 2553 MatRestoreRow_MPIBAIJ, 2554 MatMult_MPIBAIJ, 2555 /* 4*/ MatMultAdd_MPIBAIJ, 2556 MatMultTranspose_MPIBAIJ, 2557 MatMultTransposeAdd_MPIBAIJ, 2558 0, 2559 0, 2560 0, 2561 /*10*/ 0, 2562 0, 2563 0, 2564 MatSOR_MPIBAIJ, 2565 MatTranspose_MPIBAIJ, 2566 /*15*/ MatGetInfo_MPIBAIJ, 2567 MatEqual_MPIBAIJ, 2568 MatGetDiagonal_MPIBAIJ, 2569 MatDiagonalScale_MPIBAIJ, 2570 MatNorm_MPIBAIJ, 2571 /*20*/ MatAssemblyBegin_MPIBAIJ, 2572 MatAssemblyEnd_MPIBAIJ, 2573 MatSetOption_MPIBAIJ, 2574 MatZeroEntries_MPIBAIJ, 2575 /*24*/ MatZeroRows_MPIBAIJ, 2576 0, 2577 0, 2578 0, 2579 0, 2580 /*29*/ MatSetUpPreallocation_MPIBAIJ, 2581 0, 2582 0, 2583 0, 2584 0, 2585 /*34*/ MatDuplicate_MPIBAIJ, 2586 0, 2587 0, 2588 0, 2589 0, 2590 /*39*/ MatAXPY_MPIBAIJ, 2591 MatGetSubMatrices_MPIBAIJ, 2592 MatIncreaseOverlap_MPIBAIJ, 2593 MatGetValues_MPIBAIJ, 2594 MatCopy_MPIBAIJ, 2595 /*44*/ 0, 2596 MatScale_MPIBAIJ, 2597 0, 2598 0, 2599 0, 2600 /*49*/ 0, 2601 0, 2602 0, 2603 0, 2604 0, 2605 /*54*/ MatFDColoringCreate_MPIBAIJ, 2606 0, 2607 MatSetUnfactored_MPIBAIJ, 2608 MatPermute_MPIBAIJ, 2609 MatSetValuesBlocked_MPIBAIJ, 2610 /*59*/ MatGetSubMatrix_MPIBAIJ, 2611 MatDestroy_MPIBAIJ, 2612 MatView_MPIBAIJ, 2613 0, 2614 0, 2615 /*64*/ 0, 2616 0, 2617 0, 2618 0, 2619 0, 2620 /*69*/ MatGetRowMaxAbs_MPIBAIJ, 2621 0, 2622 0, 2623 0, 2624 0, 2625 /*74*/ 0, 2626 MatFDColoringApply_BAIJ, 2627 0, 2628 0, 2629 0, 2630 /*79*/ 0, 2631 0, 2632 0, 2633 0, 2634 MatLoad_MPIBAIJ, 2635 /*84*/ 0, 2636 0, 2637 0, 2638 0, 2639 0, 2640 /*89*/ 0, 2641 0, 2642 0, 2643 0, 2644 0, 2645 /*94*/ 0, 2646 0, 2647 0, 2648 0, 2649 0, 2650 /*99*/ 0, 2651 0, 2652 0, 2653 0, 2654 0, 2655 /*104*/0, 2656 MatRealPart_MPIBAIJ, 2657 MatImaginaryPart_MPIBAIJ, 2658 0, 2659 0, 2660 /*109*/0, 2661 0, 2662 0, 2663 0, 2664 0, 2665 /*114*/MatGetSeqNonzerostructure_MPIBAIJ, 2666 0, 2667 MatGetGhosts_MPIBAIJ 2668 }; 2669 2670 EXTERN_C_BEGIN 2671 #undef __FUNCT__ 2672 #define __FUNCT__ "MatGetDiagonalBlock_MPIBAIJ" 2673 PetscErrorCode PETSCMAT_DLLEXPORT MatGetDiagonalBlock_MPIBAIJ(Mat A,PetscTruth *iscopy,MatReuse reuse,Mat *a) 2674 { 2675 PetscFunctionBegin; 2676 *a = ((Mat_MPIBAIJ *)A->data)->A; 2677 *iscopy = PETSC_FALSE; 2678 PetscFunctionReturn(0); 2679 } 2680 EXTERN_C_END 2681 2682 EXTERN_C_BEGIN 2683 extern PetscErrorCode PETSCMAT_DLLEXPORT MatConvert_MPIBAIJ_MPISBAIJ(Mat, MatType,MatReuse,Mat*); 2684 EXTERN_C_END 2685 2686 EXTERN_C_BEGIN 2687 #undef __FUNCT__ 2688 #define __FUNCT__ "MatMPIBAIJSetPreallocationCSR_MPIBAIJ" 2689 PetscErrorCode MatMPIBAIJSetPreallocationCSR_MPIBAIJ(Mat B,PetscInt bs,const PetscInt ii[],const PetscInt jj[],const PetscScalar V[]) 2690 { 2691 PetscInt m,rstart,cstart,cend; 2692 PetscInt i,j,d,nz,nz_max=0,*d_nnz=0,*o_nnz=0; 2693 const PetscInt *JJ=0; 2694 PetscScalar *values=0; 2695 PetscErrorCode ierr; 2696 2697 PetscFunctionBegin; 2698 2699 if (bs < 1) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Invalid block size specified, must be positive but it is %D",bs); 2700 ierr = PetscLayoutSetBlockSize(B->rmap,bs);CHKERRQ(ierr); 2701 ierr = PetscLayoutSetBlockSize(B->cmap,bs);CHKERRQ(ierr); 2702 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 2703 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 2704 m = B->rmap->n/bs; 2705 rstart = B->rmap->rstart/bs; 2706 cstart = B->cmap->rstart/bs; 2707 cend = B->cmap->rend/bs; 2708 2709 if (ii[0]) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"ii[0] must be 0 but it is %D",ii[0]); 2710 ierr = PetscMalloc((2*m+1)*sizeof(PetscInt),&d_nnz);CHKERRQ(ierr); 2711 o_nnz = d_nnz + m; 2712 for (i=0; i<m; i++) { 2713 nz = ii[i+1] - ii[i]; 2714 if (nz < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Local row %D has a negative number of columns %D",i,nz); 2715 nz_max = PetscMax(nz_max,nz); 2716 JJ = jj + ii[i]; 2717 for (j=0; j<nz; j++) { 2718 if (*JJ >= cstart) break; 2719 JJ++; 2720 } 2721 d = 0; 2722 for (; j<nz; j++) { 2723 if (*JJ++ >= cend) break; 2724 d++; 2725 } 2726 d_nnz[i] = d; 2727 o_nnz[i] = nz - d; 2728 } 2729 ierr = MatMPIBAIJSetPreallocation(B,bs,0,d_nnz,0,o_nnz);CHKERRQ(ierr); 2730 ierr = PetscFree(d_nnz);CHKERRQ(ierr); 2731 2732 values = (PetscScalar*)V; 2733 if (!values) { 2734 ierr = PetscMalloc(bs*bs*(nz_max+1)*sizeof(PetscScalar),&values);CHKERRQ(ierr); 2735 ierr = PetscMemzero(values,bs*bs*nz_max*sizeof(PetscScalar));CHKERRQ(ierr); 2736 } 2737 for (i=0; i<m; i++) { 2738 PetscInt row = i + rstart; 2739 PetscInt ncols = ii[i+1] - ii[i]; 2740 const PetscInt *icols = jj + ii[i]; 2741 const PetscScalar *svals = values + (V ? (bs*bs*ii[i]) : 0); 2742 ierr = MatSetValuesBlocked_MPIBAIJ(B,1,&row,ncols,icols,svals,INSERT_VALUES);CHKERRQ(ierr); 2743 } 2744 2745 if (!V) { ierr = PetscFree(values);CHKERRQ(ierr); } 2746 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2747 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2748 2749 PetscFunctionReturn(0); 2750 } 2751 EXTERN_C_END 2752 2753 #undef __FUNCT__ 2754 #define __FUNCT__ "MatMPIBAIJSetPreallocationCSR" 2755 /*@C 2756 MatMPIBAIJSetPreallocationCSR - Allocates memory for a sparse parallel matrix in AIJ format 2757 (the default parallel PETSc format). 2758 2759 Collective on MPI_Comm 2760 2761 Input Parameters: 2762 + A - the matrix 2763 . i - the indices into j for the start of each local row (starts with zero) 2764 . j - the column indices for each local row (starts with zero) these must be sorted for each row 2765 - v - optional values in the matrix 2766 2767 Level: developer 2768 2769 .keywords: matrix, aij, compressed row, sparse, parallel 2770 2771 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatMPIBAIJSetPreallocation(), MatCreateMPIAIJ(), MPIAIJ 2772 @*/ 2773 PetscErrorCode PETSCMAT_DLLEXPORT MatMPIBAIJSetPreallocationCSR(Mat B,PetscInt bs,const PetscInt i[],const PetscInt j[], const PetscScalar v[]) 2774 { 2775 PetscErrorCode ierr,(*f)(Mat,PetscInt,const PetscInt[],const PetscInt[],const PetscScalar[]); 2776 2777 PetscFunctionBegin; 2778 ierr = PetscObjectQueryFunction((PetscObject)B,"MatMPIBAIJSetPreallocationCSR_C",(void (**)(void))&f);CHKERRQ(ierr); 2779 if (f) { 2780 ierr = (*f)(B,bs,i,j,v);CHKERRQ(ierr); 2781 } 2782 PetscFunctionReturn(0); 2783 } 2784 2785 EXTERN_C_BEGIN 2786 #undef __FUNCT__ 2787 #define __FUNCT__ "MatMPIBAIJSetPreallocation_MPIBAIJ" 2788 PetscErrorCode PETSCMAT_DLLEXPORT MatMPIBAIJSetPreallocation_MPIBAIJ(Mat B,PetscInt bs,PetscInt d_nz,PetscInt *d_nnz,PetscInt o_nz,PetscInt *o_nnz) 2789 { 2790 Mat_MPIBAIJ *b; 2791 PetscErrorCode ierr; 2792 PetscInt i, newbs = PetscAbs(bs); 2793 2794 PetscFunctionBegin; 2795 if (bs < 0) { 2796 ierr = PetscOptionsBegin(((PetscObject)B)->comm,((PetscObject)B)->prefix,"Options for MPIBAIJ matrix","Mat");CHKERRQ(ierr); 2797 ierr = PetscOptionsInt("-mat_block_size","Set the blocksize used to store the matrix","MatMPIBAIJSetPreallocation",newbs,&newbs,PETSC_NULL);CHKERRQ(ierr); 2798 ierr = PetscOptionsEnd();CHKERRQ(ierr); 2799 bs = PetscAbs(bs); 2800 } 2801 if ((d_nnz || o_nnz) && newbs != bs) { 2802 SETERRQ(PETSC_ERR_ARG_WRONG,"Cannot change blocksize from command line if setting d_nnz or o_nnz"); 2803 } 2804 bs = newbs; 2805 2806 2807 if (bs < 1) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Invalid block size specified, must be positive"); 2808 if (d_nz == PETSC_DEFAULT || d_nz == PETSC_DECIDE) d_nz = 5; 2809 if (o_nz == PETSC_DEFAULT || o_nz == PETSC_DECIDE) o_nz = 2; 2810 if (d_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"d_nz cannot be less than 0: value %D",d_nz); 2811 if (o_nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"o_nz cannot be less than 0: value %D",o_nz); 2812 2813 ierr = PetscLayoutSetBlockSize(B->rmap,bs);CHKERRQ(ierr); 2814 ierr = PetscLayoutSetBlockSize(B->cmap,bs);CHKERRQ(ierr); 2815 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 2816 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 2817 2818 if (d_nnz) { 2819 for (i=0; i<B->rmap->n/bs; i++) { 2820 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]); 2821 } 2822 } 2823 if (o_nnz) { 2824 for (i=0; i<B->rmap->n/bs; i++) { 2825 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]); 2826 } 2827 } 2828 2829 b = (Mat_MPIBAIJ*)B->data; 2830 b->bs2 = bs*bs; 2831 b->mbs = B->rmap->n/bs; 2832 b->nbs = B->cmap->n/bs; 2833 b->Mbs = B->rmap->N/bs; 2834 b->Nbs = B->cmap->N/bs; 2835 2836 for (i=0; i<=b->size; i++) { 2837 b->rangebs[i] = B->rmap->range[i]/bs; 2838 } 2839 b->rstartbs = B->rmap->rstart/bs; 2840 b->rendbs = B->rmap->rend/bs; 2841 b->cstartbs = B->cmap->rstart/bs; 2842 b->cendbs = B->cmap->rend/bs; 2843 2844 if (!B->preallocated) { 2845 ierr = MatCreate(PETSC_COMM_SELF,&b->A);CHKERRQ(ierr); 2846 ierr = MatSetSizes(b->A,B->rmap->n,B->cmap->n,B->rmap->n,B->cmap->n);CHKERRQ(ierr); 2847 ierr = MatSetType(b->A,MATSEQBAIJ);CHKERRQ(ierr); 2848 ierr = PetscLogObjectParent(B,b->A);CHKERRQ(ierr); 2849 ierr = MatCreate(PETSC_COMM_SELF,&b->B);CHKERRQ(ierr); 2850 ierr = MatSetSizes(b->B,B->rmap->n,B->cmap->N,B->rmap->n,B->cmap->N);CHKERRQ(ierr); 2851 ierr = MatSetType(b->B,MATSEQBAIJ);CHKERRQ(ierr); 2852 ierr = PetscLogObjectParent(B,b->B);CHKERRQ(ierr); 2853 ierr = MatStashCreate_Private(((PetscObject)B)->comm,bs,&B->bstash);CHKERRQ(ierr); 2854 } 2855 2856 ierr = MatSeqBAIJSetPreallocation(b->A,bs,d_nz,d_nnz);CHKERRQ(ierr); 2857 ierr = MatSeqBAIJSetPreallocation(b->B,bs,o_nz,o_nnz);CHKERRQ(ierr); 2858 B->preallocated = PETSC_TRUE; 2859 PetscFunctionReturn(0); 2860 } 2861 EXTERN_C_END 2862 2863 EXTERN_C_BEGIN 2864 EXTERN PetscErrorCode PETSCMAT_DLLEXPORT MatDiagonalScaleLocal_MPIBAIJ(Mat,Vec); 2865 EXTERN PetscErrorCode PETSCMAT_DLLEXPORT MatSetHashTableFactor_MPIBAIJ(Mat,PetscReal); 2866 EXTERN_C_END 2867 2868 2869 EXTERN_C_BEGIN 2870 #undef __FUNCT__ 2871 #define __FUNCT__ "MatConvert_MPIBAIJ_MPIAdj" 2872 PetscErrorCode PETSCMAT_DLLEXPORT MatConvert_MPIBAIJ_MPIAdj(Mat B, const MatType newtype,MatReuse reuse,Mat *adj) 2873 { 2874 Mat_MPIBAIJ *b = (Mat_MPIBAIJ*)B->data; 2875 PetscErrorCode ierr; 2876 Mat_SeqBAIJ *d = (Mat_SeqBAIJ*) b->A->data,*o = (Mat_SeqBAIJ*) b->B->data; 2877 PetscInt M = B->rmap->n/B->rmap->bs,i,*ii,*jj,cnt,j,k,rstart = B->rmap->rstart/B->rmap->bs; 2878 const PetscInt *id = d->i, *jd = d->j, *io = o->i, *jo = o->j, *garray = b->garray; 2879 2880 PetscFunctionBegin; 2881 ierr = PetscMalloc((M+1)*sizeof(PetscInt),&ii);CHKERRQ(ierr); 2882 ii[0] = 0; 2883 CHKMEMQ; 2884 for (i=0; i<M; i++) { 2885 if ((id[i+1] - id[i]) < 0) SETERRQ3(PETSC_ERR_PLIB,"Indices wrong %D %D %D",i,id[i],id[i+1]); 2886 if ((io[i+1] - io[i]) < 0) SETERRQ3(PETSC_ERR_PLIB,"Indices wrong %D %D %D",i,io[i],io[i+1]); 2887 ii[i+1] = ii[i] + id[i+1] - id[i] + io[i+1] - io[i]; 2888 /* remove one from count of matrix has diagonal */ 2889 for (j=id[i]; j<id[i+1]; j++) { 2890 if (jd[j] == i) {ii[i+1]--;break;} 2891 } 2892 CHKMEMQ; 2893 } 2894 ierr = PetscMalloc(ii[M]*sizeof(PetscInt),&jj);CHKERRQ(ierr); 2895 cnt = 0; 2896 for (i=0; i<M; i++) { 2897 for (j=io[i]; j<io[i+1]; j++) { 2898 if (garray[jo[j]] > rstart) break; 2899 jj[cnt++] = garray[jo[j]]; 2900 CHKMEMQ; 2901 } 2902 for (k=id[i]; k<id[i+1]; k++) { 2903 if (jd[k] != i) { 2904 jj[cnt++] = rstart + jd[k]; 2905 CHKMEMQ; 2906 } 2907 } 2908 for (;j<io[i+1]; j++) { 2909 jj[cnt++] = garray[jo[j]]; 2910 CHKMEMQ; 2911 } 2912 } 2913 ierr = MatCreateMPIAdj(((PetscObject)B)->comm,M,B->cmap->N/B->rmap->bs,ii,jj,PETSC_NULL,adj);CHKERRQ(ierr); 2914 PetscFunctionReturn(0); 2915 } 2916 EXTERN_C_END 2917 2918 /*MC 2919 MATMPIBAIJ - MATMPIBAIJ = "mpibaij" - A matrix type to be used for distributed block sparse matrices. 2920 2921 Options Database Keys: 2922 + -mat_type mpibaij - sets the matrix type to "mpibaij" during a call to MatSetFromOptions() 2923 . -mat_block_size <bs> - set the blocksize used to store the matrix 2924 - -mat_use_hash_table <fact> 2925 2926 Level: beginner 2927 2928 .seealso: MatCreateMPIBAIJ 2929 M*/ 2930 2931 EXTERN_C_BEGIN 2932 #undef __FUNCT__ 2933 #define __FUNCT__ "MatCreate_MPIBAIJ" 2934 PetscErrorCode PETSCMAT_DLLEXPORT MatCreate_MPIBAIJ(Mat B) 2935 { 2936 Mat_MPIBAIJ *b; 2937 PetscErrorCode ierr; 2938 PetscTruth flg; 2939 2940 PetscFunctionBegin; 2941 ierr = PetscNewLog(B,Mat_MPIBAIJ,&b);CHKERRQ(ierr); 2942 B->data = (void*)b; 2943 2944 2945 ierr = PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 2946 B->mapping = 0; 2947 B->assembled = PETSC_FALSE; 2948 2949 B->insertmode = NOT_SET_VALUES; 2950 ierr = MPI_Comm_rank(((PetscObject)B)->comm,&b->rank);CHKERRQ(ierr); 2951 ierr = MPI_Comm_size(((PetscObject)B)->comm,&b->size);CHKERRQ(ierr); 2952 2953 /* build local table of row and column ownerships */ 2954 ierr = PetscMalloc((b->size+1)*sizeof(PetscInt),&b->rangebs);CHKERRQ(ierr); 2955 2956 /* build cache for off array entries formed */ 2957 ierr = MatStashCreate_Private(((PetscObject)B)->comm,1,&B->stash);CHKERRQ(ierr); 2958 b->donotstash = PETSC_FALSE; 2959 b->colmap = PETSC_NULL; 2960 b->garray = PETSC_NULL; 2961 b->roworiented = PETSC_TRUE; 2962 2963 /* stuff used in block assembly */ 2964 b->barray = 0; 2965 2966 /* stuff used for matrix vector multiply */ 2967 b->lvec = 0; 2968 b->Mvctx = 0; 2969 2970 /* stuff for MatGetRow() */ 2971 b->rowindices = 0; 2972 b->rowvalues = 0; 2973 b->getrowactive = PETSC_FALSE; 2974 2975 /* hash table stuff */ 2976 b->ht = 0; 2977 b->hd = 0; 2978 b->ht_size = 0; 2979 b->ht_flag = PETSC_FALSE; 2980 b->ht_fact = 0; 2981 b->ht_total_ct = 0; 2982 b->ht_insert_ct = 0; 2983 2984 ierr = PetscOptionsBegin(((PetscObject)B)->comm,PETSC_NULL,"Options for loading MPIBAIJ matrix 1","Mat");CHKERRQ(ierr); 2985 ierr = PetscOptionsTruth("-mat_use_hash_table","Use hash table to save memory in constructing matrix","MatSetOption",PETSC_FALSE,&flg,PETSC_NULL);CHKERRQ(ierr); 2986 if (flg) { 2987 PetscReal fact = 1.39; 2988 ierr = MatSetOption(B,MAT_USE_HASH_TABLE,PETSC_TRUE);CHKERRQ(ierr); 2989 ierr = PetscOptionsReal("-mat_use_hash_table","Use hash table factor","MatMPIBAIJSetHashTableFactor",fact,&fact,PETSC_NULL);CHKERRQ(ierr); 2990 if (fact <= 1.0) fact = 1.39; 2991 ierr = MatMPIBAIJSetHashTableFactor(B,fact);CHKERRQ(ierr); 2992 ierr = PetscInfo1(B,"Hash table Factor used %5.2f\n",fact);CHKERRQ(ierr); 2993 } 2994 ierr = PetscOptionsEnd();CHKERRQ(ierr); 2995 2996 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatConvert_mpibaij_mpiadj_C", 2997 "MatConvert_MPIBAIJ_MPIAdj", 2998 MatConvert_MPIBAIJ_MPIAdj);CHKERRQ(ierr); 2999 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatStoreValues_C", 3000 "MatStoreValues_MPIBAIJ", 3001 MatStoreValues_MPIBAIJ);CHKERRQ(ierr); 3002 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatRetrieveValues_C", 3003 "MatRetrieveValues_MPIBAIJ", 3004 MatRetrieveValues_MPIBAIJ);CHKERRQ(ierr); 3005 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatGetDiagonalBlock_C", 3006 "MatGetDiagonalBlock_MPIBAIJ", 3007 MatGetDiagonalBlock_MPIBAIJ);CHKERRQ(ierr); 3008 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMPIBAIJSetPreallocation_C", 3009 "MatMPIBAIJSetPreallocation_MPIBAIJ", 3010 MatMPIBAIJSetPreallocation_MPIBAIJ);CHKERRQ(ierr); 3011 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatMPIBAIJSetPreallocationCSR_C", 3012 "MatMPIBAIJSetPreallocationCSR_MPIBAIJ", 3013 MatMPIBAIJSetPreallocationCSR_MPIBAIJ);CHKERRQ(ierr); 3014 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatDiagonalScaleLocal_C", 3015 "MatDiagonalScaleLocal_MPIBAIJ", 3016 MatDiagonalScaleLocal_MPIBAIJ);CHKERRQ(ierr); 3017 ierr = PetscObjectComposeFunctionDynamic((PetscObject)B,"MatSetHashTableFactor_C", 3018 "MatSetHashTableFactor_MPIBAIJ", 3019 MatSetHashTableFactor_MPIBAIJ);CHKERRQ(ierr); 3020 ierr = PetscObjectChangeTypeName((PetscObject)B,MATMPIBAIJ);CHKERRQ(ierr); 3021 PetscFunctionReturn(0); 3022 } 3023 EXTERN_C_END 3024 3025 /*MC 3026 MATBAIJ - MATBAIJ = "baij" - A matrix type to be used for block sparse matrices. 3027 3028 This matrix type is identical to MATSEQBAIJ when constructed with a single process communicator, 3029 and MATMPIBAIJ otherwise. 3030 3031 Options Database Keys: 3032 . -mat_type baij - sets the matrix type to "baij" during a call to MatSetFromOptions() 3033 3034 Level: beginner 3035 3036 .seealso: MatCreateMPIBAIJ(),MATSEQBAIJ,MATMPIBAIJ, MatMPIBAIJSetPreallocation(), MatMPIBAIJSetPreallocationCSR() 3037 M*/ 3038 3039 EXTERN_C_BEGIN 3040 #undef __FUNCT__ 3041 #define __FUNCT__ "MatCreate_BAIJ" 3042 PetscErrorCode PETSCMAT_DLLEXPORT MatCreate_BAIJ(Mat A) 3043 { 3044 PetscErrorCode ierr; 3045 PetscMPIInt size; 3046 3047 PetscFunctionBegin; 3048 ierr = MPI_Comm_size(((PetscObject)A)->comm,&size);CHKERRQ(ierr); 3049 if (size == 1) { 3050 ierr = MatSetType(A,MATSEQBAIJ);CHKERRQ(ierr); 3051 } else { 3052 ierr = MatSetType(A,MATMPIBAIJ);CHKERRQ(ierr); 3053 } 3054 PetscFunctionReturn(0); 3055 } 3056 EXTERN_C_END 3057 3058 #undef __FUNCT__ 3059 #define __FUNCT__ "MatMPIBAIJSetPreallocation" 3060 /*@C 3061 MatMPIBAIJSetPreallocation - Allocates memory for a sparse parallel matrix in block AIJ format 3062 (block compressed row). For good matrix assembly performance 3063 the user should preallocate the matrix storage by setting the parameters 3064 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 3065 performance can be increased by more than a factor of 50. 3066 3067 Collective on Mat 3068 3069 Input Parameters: 3070 + A - the matrix 3071 . bs - size of blockk 3072 . d_nz - number of block nonzeros per block row in diagonal portion of local 3073 submatrix (same for all local rows) 3074 . d_nnz - array containing the number of block nonzeros in the various block rows 3075 of the in diagonal portion of the local (possibly different for each block 3076 row) or PETSC_NULL. You must leave room for the diagonal entry even if it is zero. 3077 . o_nz - number of block nonzeros per block row in the off-diagonal portion of local 3078 submatrix (same for all local rows). 3079 - o_nnz - array containing the number of nonzeros in the various block rows of the 3080 off-diagonal portion of the local submatrix (possibly different for 3081 each block row) or PETSC_NULL. 3082 3083 If the *_nnz parameter is given then the *_nz parameter is ignored 3084 3085 Options Database Keys: 3086 + -mat_block_size - size of the blocks to use 3087 - -mat_use_hash_table <fact> 3088 3089 Notes: 3090 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one processor 3091 than it must be used on all processors that share the object for that argument. 3092 3093 Storage Information: 3094 For a square global matrix we define each processor's diagonal portion 3095 to be its local rows and the corresponding columns (a square submatrix); 3096 each processor's off-diagonal portion encompasses the remainder of the 3097 local matrix (a rectangular submatrix). 3098 3099 The user can specify preallocated storage for the diagonal part of 3100 the local submatrix with either d_nz or d_nnz (not both). Set 3101 d_nz=PETSC_DEFAULT and d_nnz=PETSC_NULL for PETSc to control dynamic 3102 memory allocation. Likewise, specify preallocated storage for the 3103 off-diagonal part of the local submatrix with o_nz or o_nnz (not both). 3104 3105 Consider a processor that owns rows 3, 4 and 5 of a parallel matrix. In 3106 the figure below we depict these three local rows and all columns (0-11). 3107 3108 .vb 3109 0 1 2 3 4 5 6 7 8 9 10 11 3110 ------------------- 3111 row 3 | o o o d d d o o o o o o 3112 row 4 | o o o d d d o o o o o o 3113 row 5 | o o o d d d o o o o o o 3114 ------------------- 3115 .ve 3116 3117 Thus, any entries in the d locations are stored in the d (diagonal) 3118 submatrix, and any entries in the o locations are stored in the 3119 o (off-diagonal) submatrix. Note that the d and the o submatrices are 3120 stored simply in the MATSEQBAIJ format for compressed row storage. 3121 3122 Now d_nz should indicate the number of block nonzeros per row in the d matrix, 3123 and o_nz should indicate the number of block nonzeros per row in the o matrix. 3124 In general, for PDE problems in which most nonzeros are near the diagonal, 3125 one expects d_nz >> o_nz. For large problems you MUST preallocate memory 3126 or you will get TERRIBLE performance; see the users' manual chapter on 3127 matrices. 3128 3129 You can call MatGetInfo() to get information on how effective the preallocation was; 3130 for example the fields mallocs,nz_allocated,nz_used,nz_unneeded; 3131 You can also run with the option -info and look for messages with the string 3132 malloc in them to see if additional memory allocation was needed. 3133 3134 Level: intermediate 3135 3136 .keywords: matrix, block, aij, compressed row, sparse, parallel 3137 3138 .seealso: MatCreate(), MatCreateSeqBAIJ(), MatSetValues(), MatCreateMPIBAIJ(), MatMPIBAIJSetPreallocationCSR() 3139 @*/ 3140 PetscErrorCode PETSCMAT_DLLEXPORT MatMPIBAIJSetPreallocation(Mat B,PetscInt bs,PetscInt d_nz,const PetscInt d_nnz[],PetscInt o_nz,const PetscInt o_nnz[]) 3141 { 3142 PetscErrorCode ierr,(*f)(Mat,PetscInt,PetscInt,const PetscInt[],PetscInt,const PetscInt[]); 3143 3144 PetscFunctionBegin; 3145 ierr = PetscObjectQueryFunction((PetscObject)B,"MatMPIBAIJSetPreallocation_C",(void (**)(void))&f);CHKERRQ(ierr); 3146 if (f) { 3147 ierr = (*f)(B,bs,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 3148 } 3149 PetscFunctionReturn(0); 3150 } 3151 3152 #undef __FUNCT__ 3153 #define __FUNCT__ "MatCreateMPIBAIJ" 3154 /*@C 3155 MatCreateMPIBAIJ - Creates a sparse parallel matrix in block AIJ format 3156 (block compressed row). For good matrix assembly performance 3157 the user should preallocate the matrix storage by setting the parameters 3158 d_nz (or d_nnz) and o_nz (or o_nnz). By setting these parameters accurately, 3159 performance can be increased by more than a factor of 50. 3160 3161 Collective on MPI_Comm 3162 3163 Input Parameters: 3164 + comm - MPI communicator 3165 . bs - size of blockk 3166 . m - number of local rows (or PETSC_DECIDE to have calculated if M is given) 3167 This value should be the same as the local size used in creating the 3168 y vector for the matrix-vector product y = Ax. 3169 . n - number of local columns (or PETSC_DECIDE to have calculated if N is given) 3170 This value should be the same as the local size used in creating the 3171 x vector for the matrix-vector product y = Ax. 3172 . M - number of global rows (or PETSC_DETERMINE to have calculated if m is given) 3173 . N - number of global columns (or PETSC_DETERMINE to have calculated if n is given) 3174 . d_nz - number of nonzero blocks per block row in diagonal portion of local 3175 submatrix (same for all local rows) 3176 . d_nnz - array containing the number of nonzero blocks in the various block rows 3177 of the in diagonal portion of the local (possibly different for each block 3178 row) or PETSC_NULL. You must leave room for the diagonal entry even if it is zero. 3179 . o_nz - number of nonzero blocks per block row in the off-diagonal portion of local 3180 submatrix (same for all local rows). 3181 - o_nnz - array containing the number of nonzero blocks in the various block rows of the 3182 off-diagonal portion of the local submatrix (possibly different for 3183 each block row) or PETSC_NULL. 3184 3185 Output Parameter: 3186 . A - the matrix 3187 3188 Options Database Keys: 3189 + -mat_block_size - size of the blocks to use 3190 - -mat_use_hash_table <fact> 3191 3192 It is recommended that one use the MatCreate(), MatSetType() and/or MatSetFromOptions(), 3193 MatXXXXSetPreallocation() paradgm instead of this routine directly. 3194 [MatXXXXSetPreallocation() is, for example, MatSeqAIJSetPreallocation] 3195 3196 Notes: 3197 If the *_nnz parameter is given then the *_nz parameter is ignored 3198 3199 A nonzero block is any block that as 1 or more nonzeros in it 3200 3201 The user MUST specify either the local or global matrix dimensions 3202 (possibly both). 3203 3204 If PETSC_DECIDE or PETSC_DETERMINE is used for a particular argument on one processor 3205 than it must be used on all processors that share the object for that argument. 3206 3207 Storage Information: 3208 For a square global matrix we define each processor's diagonal portion 3209 to be its local rows and the corresponding columns (a square submatrix); 3210 each processor's off-diagonal portion encompasses the remainder of the 3211 local matrix (a rectangular submatrix). 3212 3213 The user can specify preallocated storage for the diagonal part of 3214 the local submatrix with either d_nz or d_nnz (not both). Set 3215 d_nz=PETSC_DEFAULT and d_nnz=PETSC_NULL for PETSc to control dynamic 3216 memory allocation. Likewise, specify preallocated storage for the 3217 off-diagonal part of the local submatrix with o_nz or o_nnz (not both). 3218 3219 Consider a processor that owns rows 3, 4 and 5 of a parallel matrix. In 3220 the figure below we depict these three local rows and all columns (0-11). 3221 3222 .vb 3223 0 1 2 3 4 5 6 7 8 9 10 11 3224 ------------------- 3225 row 3 | o o o d d d o o o o o o 3226 row 4 | o o o d d d o o o o o o 3227 row 5 | o o o d d d o o o o o o 3228 ------------------- 3229 .ve 3230 3231 Thus, any entries in the d locations are stored in the d (diagonal) 3232 submatrix, and any entries in the o locations are stored in the 3233 o (off-diagonal) submatrix. Note that the d and the o submatrices are 3234 stored simply in the MATSEQBAIJ format for compressed row storage. 3235 3236 Now d_nz should indicate the number of block nonzeros per row in the d matrix, 3237 and o_nz should indicate the number of block nonzeros per row in the o matrix. 3238 In general, for PDE problems in which most nonzeros are near the diagonal, 3239 one expects d_nz >> o_nz. For large problems you MUST preallocate memory 3240 or you will get TERRIBLE performance; see the users' manual chapter on 3241 matrices. 3242 3243 Level: intermediate 3244 3245 .keywords: matrix, block, aij, compressed row, sparse, parallel 3246 3247 .seealso: MatCreate(), MatCreateSeqBAIJ(), MatSetValues(), MatCreateMPIBAIJ(), MatMPIBAIJSetPreallocation(), MatMPIBAIJSetPreallocationCSR() 3248 @*/ 3249 PetscErrorCode PETSCMAT_DLLEXPORT MatCreateMPIBAIJ(MPI_Comm comm,PetscInt bs,PetscInt m,PetscInt n,PetscInt M,PetscInt N,PetscInt d_nz,const PetscInt d_nnz[],PetscInt o_nz,const PetscInt o_nnz[],Mat *A) 3250 { 3251 PetscErrorCode ierr; 3252 PetscMPIInt size; 3253 3254 PetscFunctionBegin; 3255 ierr = MatCreate(comm,A);CHKERRQ(ierr); 3256 ierr = MatSetSizes(*A,m,n,M,N);CHKERRQ(ierr); 3257 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 3258 if (size > 1) { 3259 ierr = MatSetType(*A,MATMPIBAIJ);CHKERRQ(ierr); 3260 ierr = MatMPIBAIJSetPreallocation(*A,bs,d_nz,d_nnz,o_nz,o_nnz);CHKERRQ(ierr); 3261 } else { 3262 ierr = MatSetType(*A,MATSEQBAIJ);CHKERRQ(ierr); 3263 ierr = MatSeqBAIJSetPreallocation(*A,bs,d_nz,d_nnz);CHKERRQ(ierr); 3264 } 3265 PetscFunctionReturn(0); 3266 } 3267 3268 #undef __FUNCT__ 3269 #define __FUNCT__ "MatDuplicate_MPIBAIJ" 3270 static PetscErrorCode MatDuplicate_MPIBAIJ(Mat matin,MatDuplicateOption cpvalues,Mat *newmat) 3271 { 3272 Mat mat; 3273 Mat_MPIBAIJ *a,*oldmat = (Mat_MPIBAIJ*)matin->data; 3274 PetscErrorCode ierr; 3275 PetscInt len=0; 3276 3277 PetscFunctionBegin; 3278 *newmat = 0; 3279 ierr = MatCreate(((PetscObject)matin)->comm,&mat);CHKERRQ(ierr); 3280 ierr = MatSetSizes(mat,matin->rmap->n,matin->cmap->n,matin->rmap->N,matin->cmap->N);CHKERRQ(ierr); 3281 ierr = MatSetType(mat,((PetscObject)matin)->type_name);CHKERRQ(ierr); 3282 ierr = PetscMemcpy(mat->ops,matin->ops,sizeof(struct _MatOps));CHKERRQ(ierr); 3283 3284 mat->factor = matin->factor; 3285 mat->preallocated = PETSC_TRUE; 3286 mat->assembled = PETSC_TRUE; 3287 mat->insertmode = NOT_SET_VALUES; 3288 3289 a = (Mat_MPIBAIJ*)mat->data; 3290 mat->rmap->bs = matin->rmap->bs; 3291 a->bs2 = oldmat->bs2; 3292 a->mbs = oldmat->mbs; 3293 a->nbs = oldmat->nbs; 3294 a->Mbs = oldmat->Mbs; 3295 a->Nbs = oldmat->Nbs; 3296 3297 ierr = PetscLayoutCopy(matin->rmap,&mat->rmap);CHKERRQ(ierr); 3298 ierr = PetscLayoutCopy(matin->cmap,&mat->cmap);CHKERRQ(ierr); 3299 3300 a->size = oldmat->size; 3301 a->rank = oldmat->rank; 3302 a->donotstash = oldmat->donotstash; 3303 a->roworiented = oldmat->roworiented; 3304 a->rowindices = 0; 3305 a->rowvalues = 0; 3306 a->getrowactive = PETSC_FALSE; 3307 a->barray = 0; 3308 a->rstartbs = oldmat->rstartbs; 3309 a->rendbs = oldmat->rendbs; 3310 a->cstartbs = oldmat->cstartbs; 3311 a->cendbs = oldmat->cendbs; 3312 3313 /* hash table stuff */ 3314 a->ht = 0; 3315 a->hd = 0; 3316 a->ht_size = 0; 3317 a->ht_flag = oldmat->ht_flag; 3318 a->ht_fact = oldmat->ht_fact; 3319 a->ht_total_ct = 0; 3320 a->ht_insert_ct = 0; 3321 3322 ierr = PetscMemcpy(a->rangebs,oldmat->rangebs,(a->size+1)*sizeof(PetscInt));CHKERRQ(ierr); 3323 ierr = MatStashCreate_Private(((PetscObject)matin)->comm,1,&mat->stash);CHKERRQ(ierr); 3324 ierr = MatStashCreate_Private(((PetscObject)matin)->comm,matin->rmap->bs,&mat->bstash);CHKERRQ(ierr); 3325 if (oldmat->colmap) { 3326 #if defined (PETSC_USE_CTABLE) 3327 ierr = PetscTableCreateCopy(oldmat->colmap,&a->colmap);CHKERRQ(ierr); 3328 #else 3329 ierr = PetscMalloc((a->Nbs)*sizeof(PetscInt),&a->colmap);CHKERRQ(ierr); 3330 ierr = PetscLogObjectMemory(mat,(a->Nbs)*sizeof(PetscInt));CHKERRQ(ierr); 3331 ierr = PetscMemcpy(a->colmap,oldmat->colmap,(a->Nbs)*sizeof(PetscInt));CHKERRQ(ierr); 3332 #endif 3333 } else a->colmap = 0; 3334 3335 if (oldmat->garray && (len = ((Mat_SeqBAIJ*)(oldmat->B->data))->nbs)) { 3336 ierr = PetscMalloc(len*sizeof(PetscInt),&a->garray);CHKERRQ(ierr); 3337 ierr = PetscLogObjectMemory(mat,len*sizeof(PetscInt));CHKERRQ(ierr); 3338 ierr = PetscMemcpy(a->garray,oldmat->garray,len*sizeof(PetscInt));CHKERRQ(ierr); 3339 } else a->garray = 0; 3340 3341 ierr = VecDuplicate(oldmat->lvec,&a->lvec);CHKERRQ(ierr); 3342 ierr = PetscLogObjectParent(mat,a->lvec);CHKERRQ(ierr); 3343 ierr = VecScatterCopy(oldmat->Mvctx,&a->Mvctx);CHKERRQ(ierr); 3344 ierr = PetscLogObjectParent(mat,a->Mvctx);CHKERRQ(ierr); 3345 3346 ierr = MatDuplicate(oldmat->A,cpvalues,&a->A);CHKERRQ(ierr); 3347 ierr = PetscLogObjectParent(mat,a->A);CHKERRQ(ierr); 3348 ierr = MatDuplicate(oldmat->B,cpvalues,&a->B);CHKERRQ(ierr); 3349 ierr = PetscLogObjectParent(mat,a->B);CHKERRQ(ierr); 3350 ierr = PetscFListDuplicate(((PetscObject)matin)->qlist,&((PetscObject)mat)->qlist);CHKERRQ(ierr); 3351 *newmat = mat; 3352 3353 PetscFunctionReturn(0); 3354 } 3355 3356 #include "petscsys.h" 3357 3358 #undef __FUNCT__ 3359 #define __FUNCT__ "MatLoad_MPIBAIJ" 3360 PetscErrorCode MatLoad_MPIBAIJ(PetscViewer viewer, const MatType type,Mat *newmat) 3361 { 3362 Mat A; 3363 PetscErrorCode ierr; 3364 int fd; 3365 PetscInt i,nz,j,rstart,rend; 3366 PetscScalar *vals,*buf; 3367 MPI_Comm comm = ((PetscObject)viewer)->comm; 3368 MPI_Status status; 3369 PetscMPIInt rank,size,maxnz; 3370 PetscInt header[4],*rowlengths = 0,M,N,m,*rowners,*cols; 3371 PetscInt *locrowlens = PETSC_NULL,*procsnz = PETSC_NULL,*browners = PETSC_NULL; 3372 PetscInt jj,*mycols,*ibuf,bs=1,Mbs,mbs,extra_rows,mmax; 3373 PetscMPIInt tag = ((PetscObject)viewer)->tag; 3374 PetscInt *dlens = PETSC_NULL,*odlens = PETSC_NULL,*mask = PETSC_NULL,*masked1 = PETSC_NULL,*masked2 = PETSC_NULL,rowcount,odcount; 3375 PetscInt dcount,kmax,k,nzcount,tmp,mend; 3376 3377 PetscFunctionBegin; 3378 ierr = PetscOptionsBegin(comm,PETSC_NULL,"Options for loading MPIBAIJ matrix 2","Mat");CHKERRQ(ierr); 3379 ierr = PetscOptionsInt("-matload_block_size","Set the blocksize used to store the matrix","MatLoad",bs,&bs,PETSC_NULL);CHKERRQ(ierr); 3380 ierr = PetscOptionsEnd();CHKERRQ(ierr); 3381 3382 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 3383 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 3384 if (!rank) { 3385 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 3386 ierr = PetscBinaryRead(fd,(char *)header,4,PETSC_INT);CHKERRQ(ierr); 3387 if (header[0] != MAT_FILE_COOKIE) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"not matrix object"); 3388 } 3389 3390 ierr = MPI_Bcast(header+1,3,MPIU_INT,0,comm);CHKERRQ(ierr); 3391 M = header[1]; N = header[2]; 3392 3393 if (M != N) SETERRQ(PETSC_ERR_SUP,"Can only do square matrices"); 3394 3395 /* 3396 This code adds extra rows to make sure the number of rows is 3397 divisible by the blocksize 3398 */ 3399 Mbs = M/bs; 3400 extra_rows = bs - M + bs*Mbs; 3401 if (extra_rows == bs) extra_rows = 0; 3402 else Mbs++; 3403 if (extra_rows && !rank) { 3404 ierr = PetscInfo(viewer,"Padding loaded matrix to match blocksize\n");CHKERRQ(ierr); 3405 } 3406 3407 /* determine ownership of all rows */ 3408 mbs = Mbs/size + ((Mbs % size) > rank); 3409 m = mbs*bs; 3410 ierr = PetscMalloc2(size+1,PetscInt,&rowners,size+1,PetscInt,&browners);CHKERRQ(ierr); 3411 ierr = MPI_Allgather(&mbs,1,MPIU_INT,rowners+1,1,MPIU_INT,comm);CHKERRQ(ierr); 3412 3413 /* process 0 needs enough room for process with most rows */ 3414 if (!rank) { 3415 mmax = rowners[1]; 3416 for (i=2; i<size; i++) { 3417 mmax = PetscMax(mmax,rowners[i]); 3418 } 3419 mmax*=bs; 3420 } else mmax = m; 3421 3422 rowners[0] = 0; 3423 for (i=2; i<=size; i++) rowners[i] += rowners[i-1]; 3424 for (i=0; i<=size; i++) browners[i] = rowners[i]*bs; 3425 rstart = rowners[rank]; 3426 rend = rowners[rank+1]; 3427 3428 /* distribute row lengths to all processors */ 3429 ierr = PetscMalloc((mmax+1)*sizeof(PetscInt),&locrowlens);CHKERRQ(ierr); 3430 if (!rank) { 3431 mend = m; 3432 if (size == 1) mend = mend - extra_rows; 3433 ierr = PetscBinaryRead(fd,locrowlens,mend,PETSC_INT);CHKERRQ(ierr); 3434 for (j=mend; j<m; j++) locrowlens[j] = 1; 3435 ierr = PetscMalloc(m*sizeof(PetscInt),&rowlengths);CHKERRQ(ierr); 3436 ierr = PetscMalloc(size*sizeof(PetscInt),&procsnz);CHKERRQ(ierr); 3437 ierr = PetscMemzero(procsnz,size*sizeof(PetscInt));CHKERRQ(ierr); 3438 for (j=0; j<m; j++) { 3439 procsnz[0] += locrowlens[j]; 3440 } 3441 for (i=1; i<size; i++) { 3442 mend = browners[i+1] - browners[i]; 3443 if (i == size-1) mend = mend - extra_rows; 3444 ierr = PetscBinaryRead(fd,rowlengths,mend,PETSC_INT);CHKERRQ(ierr); 3445 for (j=mend; j<browners[i+1] - browners[i]; j++) rowlengths[j] = 1; 3446 /* calculate the number of nonzeros on each processor */ 3447 for (j=0; j<browners[i+1]-browners[i]; j++) { 3448 procsnz[i] += rowlengths[j]; 3449 } 3450 ierr = MPI_Send(rowlengths,browners[i+1]-browners[i],MPIU_INT,i,tag,comm);CHKERRQ(ierr); 3451 } 3452 ierr = PetscFree(rowlengths);CHKERRQ(ierr); 3453 } else { 3454 ierr = MPI_Recv(locrowlens,m,MPIU_INT,0,tag,comm,&status);CHKERRQ(ierr); 3455 } 3456 3457 if (!rank) { 3458 /* determine max buffer needed and allocate it */ 3459 maxnz = procsnz[0]; 3460 for (i=1; i<size; i++) { 3461 maxnz = PetscMax(maxnz,procsnz[i]); 3462 } 3463 ierr = PetscMalloc(maxnz*sizeof(PetscInt),&cols);CHKERRQ(ierr); 3464 3465 /* read in my part of the matrix column indices */ 3466 nz = procsnz[0]; 3467 ierr = PetscMalloc((nz+1)*sizeof(PetscInt),&ibuf);CHKERRQ(ierr); 3468 mycols = ibuf; 3469 if (size == 1) nz -= extra_rows; 3470 ierr = PetscBinaryRead(fd,mycols,nz,PETSC_INT);CHKERRQ(ierr); 3471 if (size == 1) for (i=0; i< extra_rows; i++) { mycols[nz+i] = M+i; } 3472 3473 /* read in every ones (except the last) and ship off */ 3474 for (i=1; i<size-1; i++) { 3475 nz = procsnz[i]; 3476 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 3477 ierr = MPI_Send(cols,nz,MPIU_INT,i,tag,comm);CHKERRQ(ierr); 3478 } 3479 /* read in the stuff for the last proc */ 3480 if (size != 1) { 3481 nz = procsnz[size-1] - extra_rows; /* the extra rows are not on the disk */ 3482 ierr = PetscBinaryRead(fd,cols,nz,PETSC_INT);CHKERRQ(ierr); 3483 for (i=0; i<extra_rows; i++) cols[nz+i] = M+i; 3484 ierr = MPI_Send(cols,nz+extra_rows,MPIU_INT,size-1,tag,comm);CHKERRQ(ierr); 3485 } 3486 ierr = PetscFree(cols);CHKERRQ(ierr); 3487 } else { 3488 /* determine buffer space needed for message */ 3489 nz = 0; 3490 for (i=0; i<m; i++) { 3491 nz += locrowlens[i]; 3492 } 3493 ierr = PetscMalloc((nz+1)*sizeof(PetscInt),&ibuf);CHKERRQ(ierr); 3494 mycols = ibuf; 3495 /* receive message of column indices*/ 3496 ierr = MPI_Recv(mycols,nz,MPIU_INT,0,tag,comm,&status);CHKERRQ(ierr); 3497 ierr = MPI_Get_count(&status,MPIU_INT,&maxnz);CHKERRQ(ierr); 3498 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 3499 } 3500 3501 /* loop over local rows, determining number of off diagonal entries */ 3502 ierr = PetscMalloc2(rend-rstart,PetscInt,&dlens,rend-rstart,PetscInt,&odlens);CHKERRQ(ierr); 3503 ierr = PetscMalloc3(Mbs,PetscInt,&mask,Mbs,PetscInt,&masked1,Mbs,PetscInt,&masked2);CHKERRQ(ierr); 3504 ierr = PetscMemzero(mask,Mbs*sizeof(PetscInt));CHKERRQ(ierr); 3505 ierr = PetscMemzero(masked1,Mbs*sizeof(PetscInt));CHKERRQ(ierr); 3506 ierr = PetscMemzero(masked2,Mbs*sizeof(PetscInt));CHKERRQ(ierr); 3507 rowcount = 0; nzcount = 0; 3508 for (i=0; i<mbs; i++) { 3509 dcount = 0; 3510 odcount = 0; 3511 for (j=0; j<bs; j++) { 3512 kmax = locrowlens[rowcount]; 3513 for (k=0; k<kmax; k++) { 3514 tmp = mycols[nzcount++]/bs; 3515 if (!mask[tmp]) { 3516 mask[tmp] = 1; 3517 if (tmp < rstart || tmp >= rend) masked2[odcount++] = tmp; 3518 else masked1[dcount++] = tmp; 3519 } 3520 } 3521 rowcount++; 3522 } 3523 3524 dlens[i] = dcount; 3525 odlens[i] = odcount; 3526 3527 /* zero out the mask elements we set */ 3528 for (j=0; j<dcount; j++) mask[masked1[j]] = 0; 3529 for (j=0; j<odcount; j++) mask[masked2[j]] = 0; 3530 } 3531 3532 /* create our matrix */ 3533 ierr = MatCreate(comm,&A);CHKERRQ(ierr); 3534 ierr = MatSetSizes(A,m,m,M+extra_rows,N+extra_rows);CHKERRQ(ierr); 3535 ierr = MatSetType(A,type);CHKERRQ(ierr) 3536 ierr = MatMPIBAIJSetPreallocation(A,bs,0,dlens,0,odlens);CHKERRQ(ierr); 3537 3538 if (!rank) { 3539 ierr = PetscMalloc((maxnz+1)*sizeof(PetscScalar),&buf);CHKERRQ(ierr); 3540 /* read in my part of the matrix numerical values */ 3541 nz = procsnz[0]; 3542 vals = buf; 3543 mycols = ibuf; 3544 if (size == 1) nz -= extra_rows; 3545 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 3546 if (size == 1) for (i=0; i< extra_rows; i++) { vals[nz+i] = 1.0; } 3547 3548 /* insert into matrix */ 3549 jj = rstart*bs; 3550 for (i=0; i<m; i++) { 3551 ierr = MatSetValues_MPIBAIJ(A,1,&jj,locrowlens[i],mycols,vals,INSERT_VALUES);CHKERRQ(ierr); 3552 mycols += locrowlens[i]; 3553 vals += locrowlens[i]; 3554 jj++; 3555 } 3556 /* read in other processors (except the last one) and ship out */ 3557 for (i=1; i<size-1; i++) { 3558 nz = procsnz[i]; 3559 vals = buf; 3560 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 3561 ierr = MPI_Send(vals,nz,MPIU_SCALAR,i,((PetscObject)A)->tag,comm);CHKERRQ(ierr); 3562 } 3563 /* the last proc */ 3564 if (size != 1){ 3565 nz = procsnz[i] - extra_rows; 3566 vals = buf; 3567 ierr = PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);CHKERRQ(ierr); 3568 for (i=0; i<extra_rows; i++) vals[nz+i] = 1.0; 3569 ierr = MPI_Send(vals,nz+extra_rows,MPIU_SCALAR,size-1,((PetscObject)A)->tag,comm);CHKERRQ(ierr); 3570 } 3571 ierr = PetscFree(procsnz);CHKERRQ(ierr); 3572 } else { 3573 /* receive numeric values */ 3574 ierr = PetscMalloc((nz+1)*sizeof(PetscScalar),&buf);CHKERRQ(ierr); 3575 3576 /* receive message of values*/ 3577 vals = buf; 3578 mycols = ibuf; 3579 ierr = MPI_Recv(vals,nz,MPIU_SCALAR,0,((PetscObject)A)->tag,comm,&status);CHKERRQ(ierr); 3580 ierr = MPI_Get_count(&status,MPIU_SCALAR,&maxnz);CHKERRQ(ierr); 3581 if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file"); 3582 3583 /* insert into matrix */ 3584 jj = rstart*bs; 3585 for (i=0; i<m; i++) { 3586 ierr = MatSetValues_MPIBAIJ(A,1,&jj,locrowlens[i],mycols,vals,INSERT_VALUES);CHKERRQ(ierr); 3587 mycols += locrowlens[i]; 3588 vals += locrowlens[i]; 3589 jj++; 3590 } 3591 } 3592 ierr = PetscFree(locrowlens);CHKERRQ(ierr); 3593 ierr = PetscFree(buf);CHKERRQ(ierr); 3594 ierr = PetscFree(ibuf);CHKERRQ(ierr); 3595 ierr = PetscFree2(rowners,browners);CHKERRQ(ierr); 3596 ierr = PetscFree2(dlens,odlens);CHKERRQ(ierr); 3597 ierr = PetscFree3(mask,masked1,masked2);CHKERRQ(ierr); 3598 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3599 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3600 3601 *newmat = A; 3602 PetscFunctionReturn(0); 3603 } 3604 3605 #undef __FUNCT__ 3606 #define __FUNCT__ "MatMPIBAIJSetHashTableFactor" 3607 /*@ 3608 MatMPIBAIJSetHashTableFactor - Sets the factor required to compute the size of the HashTable. 3609 3610 Input Parameters: 3611 . mat - the matrix 3612 . fact - factor 3613 3614 Collective on Mat 3615 3616 Level: advanced 3617 3618 Notes: 3619 This can also be set by the command line option: -mat_use_hash_table <fact> 3620 3621 .keywords: matrix, hashtable, factor, HT 3622 3623 .seealso: MatSetOption() 3624 @*/ 3625 PetscErrorCode PETSCMAT_DLLEXPORT MatMPIBAIJSetHashTableFactor(Mat mat,PetscReal fact) 3626 { 3627 PetscErrorCode ierr,(*f)(Mat,PetscReal); 3628 3629 PetscFunctionBegin; 3630 ierr = PetscObjectQueryFunction((PetscObject)mat,"MatSetHashTableFactor_C",(void (**)(void))&f);CHKERRQ(ierr); 3631 if (f) { 3632 ierr = (*f)(mat,fact);CHKERRQ(ierr); 3633 } 3634 PetscFunctionReturn(0); 3635 } 3636 3637 EXTERN_C_BEGIN 3638 #undef __FUNCT__ 3639 #define __FUNCT__ "MatSetHashTableFactor_MPIBAIJ" 3640 PetscErrorCode PETSCMAT_DLLEXPORT MatSetHashTableFactor_MPIBAIJ(Mat mat,PetscReal fact) 3641 { 3642 Mat_MPIBAIJ *baij; 3643 3644 PetscFunctionBegin; 3645 baij = (Mat_MPIBAIJ*)mat->data; 3646 baij->ht_fact = fact; 3647 PetscFunctionReturn(0); 3648 } 3649 EXTERN_C_END 3650 3651 #undef __FUNCT__ 3652 #define __FUNCT__ "MatMPIBAIJGetSeqBAIJ" 3653 PetscErrorCode PETSCMAT_DLLEXPORT MatMPIBAIJGetSeqBAIJ(Mat A,Mat *Ad,Mat *Ao,PetscInt *colmap[]) 3654 { 3655 Mat_MPIBAIJ *a = (Mat_MPIBAIJ *)A->data; 3656 PetscFunctionBegin; 3657 *Ad = a->A; 3658 *Ao = a->B; 3659 *colmap = a->garray; 3660 PetscFunctionReturn(0); 3661 } 3662 3663 /* 3664 Special version for direct calls from Fortran (to eliminate two function call overheads 3665 */ 3666 #if defined(PETSC_HAVE_FORTRAN_CAPS) 3667 #define matmpibaijsetvaluesblocked_ MATMPIBAIJSETVALUESBLOCKED 3668 #elif !defined(PETSC_HAVE_FORTRAN_UNDERSCORE) 3669 #define matmpibaijsetvaluesblocked_ matmpibaijsetvaluesblocked 3670 #endif 3671 3672 #undef __FUNCT__ 3673 #define __FUNCT__ "matmpibiajsetvaluesblocked" 3674 /*@C 3675 MatMPIBAIJSetValuesBlocked - Direct Fortran call to replace call to MatSetValuesBlocked() 3676 3677 Collective on Mat 3678 3679 Input Parameters: 3680 + mat - the matrix 3681 . min - number of input rows 3682 . im - input rows 3683 . nin - number of input columns 3684 . in - input columns 3685 . v - numerical values input 3686 - addvin - INSERT_VALUES or ADD_VALUES 3687 3688 Notes: This has a complete copy of MatSetValuesBlocked_MPIBAIJ() which is terrible code un-reuse. 3689 3690 Level: advanced 3691 3692 .seealso: MatSetValuesBlocked() 3693 @*/ 3694 PetscErrorCode matmpibaijsetvaluesblocked_(Mat *matin,PetscInt *min,const PetscInt im[],PetscInt *nin,const PetscInt in[],const MatScalar v[],InsertMode *addvin) 3695 { 3696 /* convert input arguments to C version */ 3697 Mat mat = *matin; 3698 PetscInt m = *min, n = *nin; 3699 InsertMode addv = *addvin; 3700 3701 Mat_MPIBAIJ *baij = (Mat_MPIBAIJ*)mat->data; 3702 const MatScalar *value; 3703 MatScalar *barray=baij->barray; 3704 PetscTruth roworiented = baij->roworiented; 3705 PetscErrorCode ierr; 3706 PetscInt i,j,ii,jj,row,col,rstart=baij->rstartbs; 3707 PetscInt rend=baij->rendbs,cstart=baij->cstartbs,stepval; 3708 PetscInt cend=baij->cendbs,bs=mat->rmap->bs,bs2=baij->bs2; 3709 3710 PetscFunctionBegin; 3711 /* tasks normally handled by MatSetValuesBlocked() */ 3712 if (mat->insertmode == NOT_SET_VALUES) { 3713 mat->insertmode = addv; 3714 } 3715 #if defined(PETSC_USE_DEBUG) 3716 else if (mat->insertmode != addv) { 3717 SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix add values and insert values"); 3718 } 3719 if (mat->factor) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3720 #endif 3721 if (mat->assembled) { 3722 mat->was_assembled = PETSC_TRUE; 3723 mat->assembled = PETSC_FALSE; 3724 } 3725 ierr = PetscLogEventBegin(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 3726 3727 3728 if(!barray) { 3729 ierr = PetscMalloc(bs2*sizeof(MatScalar),&barray);CHKERRQ(ierr); 3730 baij->barray = barray; 3731 } 3732 3733 if (roworiented) { 3734 stepval = (n-1)*bs; 3735 } else { 3736 stepval = (m-1)*bs; 3737 } 3738 for (i=0; i<m; i++) { 3739 if (im[i] < 0) continue; 3740 #if defined(PETSC_USE_DEBUG) 3741 if (im[i] >= baij->Mbs) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large, row %D max %D",im[i],baij->Mbs-1); 3742 #endif 3743 if (im[i] >= rstart && im[i] < rend) { 3744 row = im[i] - rstart; 3745 for (j=0; j<n; j++) { 3746 /* If NumCol = 1 then a copy is not required */ 3747 if ((roworiented) && (n == 1)) { 3748 barray = (MatScalar*)v + i*bs2; 3749 } else if((!roworiented) && (m == 1)) { 3750 barray = (MatScalar*)v + j*bs2; 3751 } else { /* Here a copy is required */ 3752 if (roworiented) { 3753 value = v + i*(stepval+bs)*bs + j*bs; 3754 } else { 3755 value = v + j*(stepval+bs)*bs + i*bs; 3756 } 3757 for (ii=0; ii<bs; ii++,value+=stepval) { 3758 for (jj=0; jj<bs; jj++) { 3759 *barray++ = *value++; 3760 } 3761 } 3762 barray -=bs2; 3763 } 3764 3765 if (in[j] >= cstart && in[j] < cend){ 3766 col = in[j] - cstart; 3767 ierr = MatSetValuesBlocked_SeqBAIJ(baij->A,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 3768 } 3769 else if (in[j] < 0) continue; 3770 #if defined(PETSC_USE_DEBUG) 3771 else if (in[j] >= baij->Nbs) {SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large, col %D max %D",in[j],baij->Nbs-1);} 3772 #endif 3773 else { 3774 if (mat->was_assembled) { 3775 if (!baij->colmap) { 3776 ierr = CreateColmap_MPIBAIJ_Private(mat);CHKERRQ(ierr); 3777 } 3778 3779 #if defined(PETSC_USE_DEBUG) 3780 #if defined (PETSC_USE_CTABLE) 3781 { PetscInt data; 3782 ierr = PetscTableFind(baij->colmap,in[j]+1,&data);CHKERRQ(ierr); 3783 if ((data - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 3784 } 3785 #else 3786 if ((baij->colmap[in[j]] - 1) % bs) SETERRQ(PETSC_ERR_PLIB,"Incorrect colmap"); 3787 #endif 3788 #endif 3789 #if defined (PETSC_USE_CTABLE) 3790 ierr = PetscTableFind(baij->colmap,in[j]+1,&col);CHKERRQ(ierr); 3791 col = (col - 1)/bs; 3792 #else 3793 col = (baij->colmap[in[j]] - 1)/bs; 3794 #endif 3795 if (col < 0 && !((Mat_SeqBAIJ*)(baij->A->data))->nonew) { 3796 ierr = DisAssemble_MPIBAIJ(mat);CHKERRQ(ierr); 3797 col = in[j]; 3798 } 3799 } 3800 else col = in[j]; 3801 ierr = MatSetValuesBlocked_SeqBAIJ(baij->B,1,&row,1,&col,barray,addv);CHKERRQ(ierr); 3802 } 3803 } 3804 } else { 3805 if (!baij->donotstash) { 3806 if (roworiented) { 3807 ierr = MatStashValuesRowBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 3808 } else { 3809 ierr = MatStashValuesColBlocked_Private(&mat->bstash,im[i],n,in,v,m,n,i);CHKERRQ(ierr); 3810 } 3811 } 3812 } 3813 } 3814 3815 /* task normally handled by MatSetValuesBlocked() */ 3816 ierr = PetscLogEventEnd(MAT_SetValues,mat,0,0,0);CHKERRQ(ierr); 3817 PetscFunctionReturn(0); 3818 } 3819