1 2 /* 3 Defines the basic matrix operations for the AIJ (compressed row) 4 matrix storage format. 5 */ 6 7 8 #include <../src/mat/impls/aij/seq/aij.h> /*I "petscmat.h" I*/ 9 #include <petscblaslapack.h> 10 #include <petscbt.h> 11 #include <petsc/private/kernels/blocktranspose.h> 12 13 PetscErrorCode MatSeqAIJSetTypeFromOptions(Mat A) 14 { 15 PetscErrorCode ierr; 16 PetscBool flg; 17 char type[256]; 18 19 PetscFunctionBegin; 20 ierr = PetscObjectOptionsBegin((PetscObject)A); 21 ierr = PetscOptionsFList("-mat_seqaij_type","Matrix SeqAIJ type","MatSeqAIJSetType",MatSeqAIJList,"seqaij",type,256,&flg);CHKERRQ(ierr); 22 if (flg) { 23 ierr = MatSeqAIJSetType(A,type);CHKERRQ(ierr); 24 } 25 ierr = PetscOptionsEnd();CHKERRQ(ierr); 26 PetscFunctionReturn(0); 27 } 28 29 PetscErrorCode MatGetColumnNorms_SeqAIJ(Mat A,NormType type,PetscReal *norms) 30 { 31 PetscErrorCode ierr; 32 PetscInt i,m,n; 33 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)A->data; 34 35 PetscFunctionBegin; 36 ierr = MatGetSize(A,&m,&n);CHKERRQ(ierr); 37 ierr = PetscMemzero(norms,n*sizeof(PetscReal));CHKERRQ(ierr); 38 if (type == NORM_2) { 39 for (i=0; i<aij->i[m]; i++) { 40 norms[aij->j[i]] += PetscAbsScalar(aij->a[i]*aij->a[i]); 41 } 42 } else if (type == NORM_1) { 43 for (i=0; i<aij->i[m]; i++) { 44 norms[aij->j[i]] += PetscAbsScalar(aij->a[i]); 45 } 46 } else if (type == NORM_INFINITY) { 47 for (i=0; i<aij->i[m]; i++) { 48 norms[aij->j[i]] = PetscMax(PetscAbsScalar(aij->a[i]),norms[aij->j[i]]); 49 } 50 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Unknown NormType"); 51 52 if (type == NORM_2) { 53 for (i=0; i<n; i++) norms[i] = PetscSqrtReal(norms[i]); 54 } 55 PetscFunctionReturn(0); 56 } 57 58 PetscErrorCode MatFindOffBlockDiagonalEntries_SeqAIJ(Mat A,IS *is) 59 { 60 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 61 PetscInt i,m=A->rmap->n,cnt = 0, bs = A->rmap->bs; 62 const PetscInt *jj = a->j,*ii = a->i; 63 PetscInt *rows; 64 PetscErrorCode ierr; 65 66 PetscFunctionBegin; 67 for (i=0; i<m; i++) { 68 if ((ii[i] != ii[i+1]) && ((jj[ii[i]] < bs*(i/bs)) || (jj[ii[i+1]-1] > bs*((i+bs)/bs)-1))) { 69 cnt++; 70 } 71 } 72 ierr = PetscMalloc1(cnt,&rows);CHKERRQ(ierr); 73 cnt = 0; 74 for (i=0; i<m; i++) { 75 if ((ii[i] != ii[i+1]) && ((jj[ii[i]] < bs*(i/bs)) || (jj[ii[i+1]-1] > bs*((i+bs)/bs)-1))) { 76 rows[cnt] = i; 77 cnt++; 78 } 79 } 80 ierr = ISCreateGeneral(PETSC_COMM_SELF,cnt,rows,PETSC_OWN_POINTER,is);CHKERRQ(ierr); 81 PetscFunctionReturn(0); 82 } 83 84 PetscErrorCode MatFindZeroDiagonals_SeqAIJ_Private(Mat A,PetscInt *nrows,PetscInt **zrows) 85 { 86 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 87 const MatScalar *aa = a->a; 88 PetscInt i,m=A->rmap->n,cnt = 0; 89 const PetscInt *ii = a->i,*jj = a->j,*diag; 90 PetscInt *rows; 91 PetscErrorCode ierr; 92 93 PetscFunctionBegin; 94 ierr = MatMarkDiagonal_SeqAIJ(A);CHKERRQ(ierr); 95 diag = a->diag; 96 for (i=0; i<m; i++) { 97 if ((diag[i] >= ii[i+1]) || (jj[diag[i]] != i) || (aa[diag[i]] == 0.0)) { 98 cnt++; 99 } 100 } 101 ierr = PetscMalloc1(cnt,&rows);CHKERRQ(ierr); 102 cnt = 0; 103 for (i=0; i<m; i++) { 104 if ((diag[i] >= ii[i+1]) || (jj[diag[i]] != i) || (aa[diag[i]] == 0.0)) { 105 rows[cnt++] = i; 106 } 107 } 108 *nrows = cnt; 109 *zrows = rows; 110 PetscFunctionReturn(0); 111 } 112 113 PetscErrorCode MatFindZeroDiagonals_SeqAIJ(Mat A,IS *zrows) 114 { 115 PetscInt nrows,*rows; 116 PetscErrorCode ierr; 117 118 PetscFunctionBegin; 119 *zrows = NULL; 120 ierr = MatFindZeroDiagonals_SeqAIJ_Private(A,&nrows,&rows);CHKERRQ(ierr); 121 ierr = ISCreateGeneral(PetscObjectComm((PetscObject)A),nrows,rows,PETSC_OWN_POINTER,zrows);CHKERRQ(ierr); 122 PetscFunctionReturn(0); 123 } 124 125 PetscErrorCode MatFindNonzeroRows_SeqAIJ(Mat A,IS *keptrows) 126 { 127 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 128 const MatScalar *aa; 129 PetscInt m=A->rmap->n,cnt = 0; 130 const PetscInt *ii; 131 PetscInt n,i,j,*rows; 132 PetscErrorCode ierr; 133 134 PetscFunctionBegin; 135 *keptrows = 0; 136 ii = a->i; 137 for (i=0; i<m; i++) { 138 n = ii[i+1] - ii[i]; 139 if (!n) { 140 cnt++; 141 goto ok1; 142 } 143 aa = a->a + ii[i]; 144 for (j=0; j<n; j++) { 145 if (aa[j] != 0.0) goto ok1; 146 } 147 cnt++; 148 ok1:; 149 } 150 if (!cnt) PetscFunctionReturn(0); 151 ierr = PetscMalloc1(A->rmap->n-cnt,&rows);CHKERRQ(ierr); 152 cnt = 0; 153 for (i=0; i<m; i++) { 154 n = ii[i+1] - ii[i]; 155 if (!n) continue; 156 aa = a->a + ii[i]; 157 for (j=0; j<n; j++) { 158 if (aa[j] != 0.0) { 159 rows[cnt++] = i; 160 break; 161 } 162 } 163 } 164 ierr = ISCreateGeneral(PETSC_COMM_SELF,cnt,rows,PETSC_OWN_POINTER,keptrows);CHKERRQ(ierr); 165 PetscFunctionReturn(0); 166 } 167 168 PetscErrorCode MatDiagonalSet_SeqAIJ(Mat Y,Vec D,InsertMode is) 169 { 170 PetscErrorCode ierr; 171 Mat_SeqAIJ *aij = (Mat_SeqAIJ*) Y->data; 172 PetscInt i,m = Y->rmap->n; 173 const PetscInt *diag; 174 MatScalar *aa = aij->a; 175 const PetscScalar *v; 176 PetscBool missing; 177 178 PetscFunctionBegin; 179 if (Y->assembled) { 180 ierr = MatMissingDiagonal_SeqAIJ(Y,&missing,NULL);CHKERRQ(ierr); 181 if (!missing) { 182 diag = aij->diag; 183 ierr = VecGetArrayRead(D,&v);CHKERRQ(ierr); 184 if (is == INSERT_VALUES) { 185 for (i=0; i<m; i++) { 186 aa[diag[i]] = v[i]; 187 } 188 } else { 189 for (i=0; i<m; i++) { 190 aa[diag[i]] += v[i]; 191 } 192 } 193 ierr = VecRestoreArrayRead(D,&v);CHKERRQ(ierr); 194 PetscFunctionReturn(0); 195 } 196 ierr = MatSeqAIJInvalidateDiagonal(Y);CHKERRQ(ierr); 197 } 198 ierr = MatDiagonalSet_Default(Y,D,is);CHKERRQ(ierr); 199 PetscFunctionReturn(0); 200 } 201 202 PetscErrorCode MatGetRowIJ_SeqAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *m,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done) 203 { 204 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 205 PetscErrorCode ierr; 206 PetscInt i,ishift; 207 208 PetscFunctionBegin; 209 *m = A->rmap->n; 210 if (!ia) PetscFunctionReturn(0); 211 ishift = 0; 212 if (symmetric && !A->structurally_symmetric) { 213 ierr = MatToSymmetricIJ_SeqAIJ(A->rmap->n,a->i,a->j,PETSC_TRUE,ishift,oshift,(PetscInt**)ia,(PetscInt**)ja);CHKERRQ(ierr); 214 } else if (oshift == 1) { 215 PetscInt *tia; 216 PetscInt nz = a->i[A->rmap->n]; 217 /* malloc space and add 1 to i and j indices */ 218 ierr = PetscMalloc1(A->rmap->n+1,&tia);CHKERRQ(ierr); 219 for (i=0; i<A->rmap->n+1; i++) tia[i] = a->i[i] + 1; 220 *ia = tia; 221 if (ja) { 222 PetscInt *tja; 223 ierr = PetscMalloc1(nz+1,&tja);CHKERRQ(ierr); 224 for (i=0; i<nz; i++) tja[i] = a->j[i] + 1; 225 *ja = tja; 226 } 227 } else { 228 *ia = a->i; 229 if (ja) *ja = a->j; 230 } 231 PetscFunctionReturn(0); 232 } 233 234 PetscErrorCode MatRestoreRowIJ_SeqAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done) 235 { 236 PetscErrorCode ierr; 237 238 PetscFunctionBegin; 239 if (!ia) PetscFunctionReturn(0); 240 if ((symmetric && !A->structurally_symmetric) || oshift == 1) { 241 ierr = PetscFree(*ia);CHKERRQ(ierr); 242 if (ja) {ierr = PetscFree(*ja);CHKERRQ(ierr);} 243 } 244 PetscFunctionReturn(0); 245 } 246 247 PetscErrorCode MatGetColumnIJ_SeqAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *nn,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done) 248 { 249 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 250 PetscErrorCode ierr; 251 PetscInt i,*collengths,*cia,*cja,n = A->cmap->n,m = A->rmap->n; 252 PetscInt nz = a->i[m],row,*jj,mr,col; 253 254 PetscFunctionBegin; 255 *nn = n; 256 if (!ia) PetscFunctionReturn(0); 257 if (symmetric) { 258 ierr = MatToSymmetricIJ_SeqAIJ(A->rmap->n,a->i,a->j,PETSC_TRUE,0,oshift,(PetscInt**)ia,(PetscInt**)ja);CHKERRQ(ierr); 259 } else { 260 ierr = PetscCalloc1(n+1,&collengths);CHKERRQ(ierr); 261 ierr = PetscMalloc1(n+1,&cia);CHKERRQ(ierr); 262 ierr = PetscMalloc1(nz+1,&cja);CHKERRQ(ierr); 263 jj = a->j; 264 for (i=0; i<nz; i++) { 265 collengths[jj[i]]++; 266 } 267 cia[0] = oshift; 268 for (i=0; i<n; i++) { 269 cia[i+1] = cia[i] + collengths[i]; 270 } 271 ierr = PetscMemzero(collengths,n*sizeof(PetscInt));CHKERRQ(ierr); 272 jj = a->j; 273 for (row=0; row<m; row++) { 274 mr = a->i[row+1] - a->i[row]; 275 for (i=0; i<mr; i++) { 276 col = *jj++; 277 278 cja[cia[col] + collengths[col]++ - oshift] = row + oshift; 279 } 280 } 281 ierr = PetscFree(collengths);CHKERRQ(ierr); 282 *ia = cia; *ja = cja; 283 } 284 PetscFunctionReturn(0); 285 } 286 287 PetscErrorCode MatRestoreColumnIJ_SeqAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done) 288 { 289 PetscErrorCode ierr; 290 291 PetscFunctionBegin; 292 if (!ia) PetscFunctionReturn(0); 293 294 ierr = PetscFree(*ia);CHKERRQ(ierr); 295 ierr = PetscFree(*ja);CHKERRQ(ierr); 296 PetscFunctionReturn(0); 297 } 298 299 /* 300 MatGetColumnIJ_SeqAIJ_Color() and MatRestoreColumnIJ_SeqAIJ_Color() are customized from 301 MatGetColumnIJ_SeqAIJ() and MatRestoreColumnIJ_SeqAIJ() by adding an output 302 spidx[], index of a->a, to be used in MatTransposeColoringCreate_SeqAIJ() and MatFDColoringCreate_SeqXAIJ() 303 */ 304 PetscErrorCode MatGetColumnIJ_SeqAIJ_Color(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *nn,const PetscInt *ia[],const PetscInt *ja[],PetscInt *spidx[],PetscBool *done) 305 { 306 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 307 PetscErrorCode ierr; 308 PetscInt i,*collengths,*cia,*cja,n = A->cmap->n,m = A->rmap->n; 309 PetscInt nz = a->i[m],row,*jj,mr,col; 310 PetscInt *cspidx; 311 312 PetscFunctionBegin; 313 *nn = n; 314 if (!ia) PetscFunctionReturn(0); 315 316 ierr = PetscCalloc1(n+1,&collengths);CHKERRQ(ierr); 317 ierr = PetscMalloc1(n+1,&cia);CHKERRQ(ierr); 318 ierr = PetscMalloc1(nz+1,&cja);CHKERRQ(ierr); 319 ierr = PetscMalloc1(nz+1,&cspidx);CHKERRQ(ierr); 320 jj = a->j; 321 for (i=0; i<nz; i++) { 322 collengths[jj[i]]++; 323 } 324 cia[0] = oshift; 325 for (i=0; i<n; i++) { 326 cia[i+1] = cia[i] + collengths[i]; 327 } 328 ierr = PetscMemzero(collengths,n*sizeof(PetscInt));CHKERRQ(ierr); 329 jj = a->j; 330 for (row=0; row<m; row++) { 331 mr = a->i[row+1] - a->i[row]; 332 for (i=0; i<mr; i++) { 333 col = *jj++; 334 cspidx[cia[col] + collengths[col] - oshift] = a->i[row] + i; /* index of a->j */ 335 cja[cia[col] + collengths[col]++ - oshift] = row + oshift; 336 } 337 } 338 ierr = PetscFree(collengths);CHKERRQ(ierr); 339 *ia = cia; *ja = cja; 340 *spidx = cspidx; 341 PetscFunctionReturn(0); 342 } 343 344 PetscErrorCode MatRestoreColumnIJ_SeqAIJ_Color(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscInt *spidx[],PetscBool *done) 345 { 346 PetscErrorCode ierr; 347 348 PetscFunctionBegin; 349 ierr = MatRestoreColumnIJ_SeqAIJ(A,oshift,symmetric,inodecompressed,n,ia,ja,done);CHKERRQ(ierr); 350 ierr = PetscFree(*spidx);CHKERRQ(ierr); 351 PetscFunctionReturn(0); 352 } 353 354 PetscErrorCode MatSetValuesRow_SeqAIJ(Mat A,PetscInt row,const PetscScalar v[]) 355 { 356 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 357 PetscInt *ai = a->i; 358 PetscErrorCode ierr; 359 360 PetscFunctionBegin; 361 ierr = PetscMemcpy(a->a+ai[row],v,(ai[row+1]-ai[row])*sizeof(PetscScalar));CHKERRQ(ierr); 362 PetscFunctionReturn(0); 363 } 364 365 /* 366 MatSeqAIJSetValuesLocalFast - An optimized version of MatSetValuesLocal() for SeqAIJ matrices with several assumptions 367 368 - a single row of values is set with each call 369 - no row or column indices are negative or (in error) larger than the number of rows or columns 370 - the values are always added to the matrix, not set 371 - no new locations are introduced in the nonzero structure of the matrix 372 373 This does NOT assume the global column indices are sorted 374 375 */ 376 377 #include <petsc/private/isimpl.h> 378 PetscErrorCode MatSeqAIJSetValuesLocalFast(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode is) 379 { 380 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 381 PetscInt low,high,t,row,nrow,i,col,l; 382 const PetscInt *rp,*ai = a->i,*ailen = a->ilen,*aj = a->j; 383 PetscInt lastcol = -1; 384 MatScalar *ap,value,*aa = a->a; 385 const PetscInt *ridx = A->rmap->mapping->indices,*cidx = A->cmap->mapping->indices; 386 387 row = ridx[im[0]]; 388 rp = aj + ai[row]; 389 ap = aa + ai[row]; 390 nrow = ailen[row]; 391 low = 0; 392 high = nrow; 393 for (l=0; l<n; l++) { /* loop over added columns */ 394 col = cidx[in[l]]; 395 value = v[l]; 396 397 if (col <= lastcol) low = 0; 398 else high = nrow; 399 lastcol = col; 400 while (high-low > 5) { 401 t = (low+high)/2; 402 if (rp[t] > col) high = t; 403 else low = t; 404 } 405 for (i=low; i<high; i++) { 406 if (rp[i] == col) { 407 ap[i] += value; 408 low = i + 1; 409 break; 410 } 411 } 412 } 413 return 0; 414 } 415 416 PetscErrorCode MatSetValues_SeqAIJ(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode is) 417 { 418 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 419 PetscInt *rp,k,low,high,t,ii,row,nrow,i,col,l,rmax,N; 420 PetscInt *imax = a->imax,*ai = a->i,*ailen = a->ilen; 421 PetscErrorCode ierr; 422 PetscInt *aj = a->j,nonew = a->nonew,lastcol = -1; 423 MatScalar *ap=NULL,value=0.0,*aa = a->a; 424 PetscBool ignorezeroentries = a->ignorezeroentries; 425 PetscBool roworiented = a->roworiented; 426 427 PetscFunctionBegin; 428 for (k=0; k<m; k++) { /* loop over added rows */ 429 row = im[k]; 430 if (row < 0) continue; 431 #if defined(PETSC_USE_DEBUG) 432 if (row >= A->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",row,A->rmap->n-1); 433 #endif 434 rp = aj + ai[row]; 435 if (!A->structure_only) ap = aa + ai[row]; 436 rmax = imax[row]; nrow = ailen[row]; 437 low = 0; 438 high = nrow; 439 for (l=0; l<n; l++) { /* loop over added columns */ 440 if (in[l] < 0) continue; 441 #if defined(PETSC_USE_DEBUG) 442 if (in[l] >= A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[l],A->cmap->n-1); 443 #endif 444 col = in[l]; 445 if (!A->structure_only) { 446 if (roworiented) { 447 value = v[l + k*n]; 448 } else { 449 value = v[k + l*m]; 450 } 451 } else { /* A->structure_only */ 452 value = 1; /* avoid 'continue' below? */ 453 } 454 if ((value == 0.0 && ignorezeroentries) && (is == ADD_VALUES) && row != col) continue; 455 456 if (col <= lastcol) low = 0; 457 else high = nrow; 458 lastcol = col; 459 while (high-low > 5) { 460 t = (low+high)/2; 461 if (rp[t] > col) high = t; 462 else low = t; 463 } 464 for (i=low; i<high; i++) { 465 if (rp[i] > col) break; 466 if (rp[i] == col) { 467 if (!A->structure_only) { 468 if (is == ADD_VALUES) ap[i] += value; 469 else ap[i] = value; 470 } 471 low = i + 1; 472 goto noinsert; 473 } 474 } 475 if (value == 0.0 && ignorezeroentries && row != col) goto noinsert; 476 if (nonew == 1) goto noinsert; 477 if (nonew == -1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero at (%D,%D) in the matrix",row,col); 478 if (A->structure_only) { 479 MatSeqXAIJReallocateAIJ_structure_only(A,A->rmap->n,1,nrow,row,col,rmax,ai,aj,rp,imax,nonew,MatScalar); 480 } else { 481 MatSeqXAIJReallocateAIJ(A,A->rmap->n,1,nrow,row,col,rmax,aa,ai,aj,rp,ap,imax,nonew,MatScalar); 482 } 483 N = nrow++ - 1; a->nz++; high++; 484 /* shift up all the later entries in this row */ 485 for (ii=N; ii>=i; ii--) { 486 rp[ii+1] = rp[ii]; 487 if (!A->structure_only) ap[ii+1] = ap[ii]; 488 } 489 rp[i] = col; 490 if (!A->structure_only) ap[i] = value; 491 low = i + 1; 492 A->nonzerostate++; 493 noinsert:; 494 } 495 ailen[row] = nrow; 496 } 497 PetscFunctionReturn(0); 498 } 499 500 501 PetscErrorCode MatGetValues_SeqAIJ(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],PetscScalar v[]) 502 { 503 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 504 PetscInt *rp,k,low,high,t,row,nrow,i,col,l,*aj = a->j; 505 PetscInt *ai = a->i,*ailen = a->ilen; 506 MatScalar *ap,*aa = a->a; 507 508 PetscFunctionBegin; 509 for (k=0; k<m; k++) { /* loop over rows */ 510 row = im[k]; 511 if (row < 0) {v += n; continue;} /* SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative row: %D",row); */ 512 if (row >= A->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",row,A->rmap->n-1); 513 rp = aj + ai[row]; ap = aa + ai[row]; 514 nrow = ailen[row]; 515 for (l=0; l<n; l++) { /* loop over columns */ 516 if (in[l] < 0) {v++; continue;} /* SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative column: %D",in[l]); */ 517 if (in[l] >= A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[l],A->cmap->n-1); 518 col = in[l]; 519 high = nrow; low = 0; /* assume unsorted */ 520 while (high-low > 5) { 521 t = (low+high)/2; 522 if (rp[t] > col) high = t; 523 else low = t; 524 } 525 for (i=low; i<high; i++) { 526 if (rp[i] > col) break; 527 if (rp[i] == col) { 528 *v++ = ap[i]; 529 goto finished; 530 } 531 } 532 *v++ = 0.0; 533 finished:; 534 } 535 } 536 PetscFunctionReturn(0); 537 } 538 539 540 PetscErrorCode MatView_SeqAIJ_Binary(Mat A,PetscViewer viewer) 541 { 542 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 543 PetscErrorCode ierr; 544 PetscInt i,*col_lens; 545 int fd; 546 FILE *file; 547 548 PetscFunctionBegin; 549 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 550 ierr = PetscMalloc1(4+A->rmap->n,&col_lens);CHKERRQ(ierr); 551 552 col_lens[0] = MAT_FILE_CLASSID; 553 col_lens[1] = A->rmap->n; 554 col_lens[2] = A->cmap->n; 555 col_lens[3] = a->nz; 556 557 /* store lengths of each row and write (including header) to file */ 558 for (i=0; i<A->rmap->n; i++) { 559 col_lens[4+i] = a->i[i+1] - a->i[i]; 560 } 561 ierr = PetscBinaryWrite(fd,col_lens,4+A->rmap->n,PETSC_INT,PETSC_TRUE);CHKERRQ(ierr); 562 ierr = PetscFree(col_lens);CHKERRQ(ierr); 563 564 /* store column indices (zero start index) */ 565 ierr = PetscBinaryWrite(fd,a->j,a->nz,PETSC_INT,PETSC_FALSE);CHKERRQ(ierr); 566 567 /* store nonzero values */ 568 ierr = PetscBinaryWrite(fd,a->a,a->nz,PETSC_SCALAR,PETSC_FALSE);CHKERRQ(ierr); 569 570 ierr = PetscViewerBinaryGetInfoPointer(viewer,&file);CHKERRQ(ierr); 571 if (file) { 572 fprintf(file,"-matload_block_size %d\n",(int)PetscAbs(A->rmap->bs)); 573 } 574 PetscFunctionReturn(0); 575 } 576 577 static PetscErrorCode MatView_SeqAIJ_ASCII_structonly(Mat A,PetscViewer viewer) 578 { 579 PetscErrorCode ierr; 580 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 581 PetscInt i,k,m=A->rmap->N; 582 583 PetscFunctionBegin; 584 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 585 for (i=0; i<m; i++) { 586 ierr = PetscViewerASCIIPrintf(viewer,"row %D:",i);CHKERRQ(ierr); 587 for (k=a->i[i]; k<a->i[i+1]; k++) { 588 ierr = PetscViewerASCIIPrintf(viewer," (%D) ",a->j[k]);CHKERRQ(ierr); 589 } 590 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 591 } 592 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 593 PetscFunctionReturn(0); 594 } 595 596 extern PetscErrorCode MatSeqAIJFactorInfo_Matlab(Mat,PetscViewer); 597 598 PetscErrorCode MatView_SeqAIJ_ASCII(Mat A,PetscViewer viewer) 599 { 600 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 601 PetscErrorCode ierr; 602 PetscInt i,j,m = A->rmap->n; 603 const char *name; 604 PetscViewerFormat format; 605 606 PetscFunctionBegin; 607 if (A->structure_only) { 608 ierr = MatView_SeqAIJ_ASCII_structonly(A,viewer);CHKERRQ(ierr); 609 PetscFunctionReturn(0); 610 } 611 612 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 613 if (format == PETSC_VIEWER_ASCII_MATLAB) { 614 PetscInt nofinalvalue = 0; 615 if (m && ((a->i[m] == a->i[m-1]) || (a->j[a->nz-1] != A->cmap->n-1))) { 616 /* Need a dummy value to ensure the dimension of the matrix. */ 617 nofinalvalue = 1; 618 } 619 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 620 ierr = PetscViewerASCIIPrintf(viewer,"%% Size = %D %D \n",m,A->cmap->n);CHKERRQ(ierr); 621 ierr = PetscViewerASCIIPrintf(viewer,"%% Nonzeros = %D \n",a->nz);CHKERRQ(ierr); 622 #if defined(PETSC_USE_COMPLEX) 623 ierr = PetscViewerASCIIPrintf(viewer,"zzz = zeros(%D,4);\n",a->nz+nofinalvalue);CHKERRQ(ierr); 624 #else 625 ierr = PetscViewerASCIIPrintf(viewer,"zzz = zeros(%D,3);\n",a->nz+nofinalvalue);CHKERRQ(ierr); 626 #endif 627 ierr = PetscViewerASCIIPrintf(viewer,"zzz = [\n");CHKERRQ(ierr); 628 629 for (i=0; i<m; i++) { 630 for (j=a->i[i]; j<a->i[i+1]; j++) { 631 #if defined(PETSC_USE_COMPLEX) 632 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %18.16e %18.16e\n",i+1,a->j[j]+1,(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 633 #else 634 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %18.16e\n",i+1,a->j[j]+1,(double)a->a[j]);CHKERRQ(ierr); 635 #endif 636 } 637 } 638 if (nofinalvalue) { 639 #if defined(PETSC_USE_COMPLEX) 640 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %18.16e %18.16e\n",m,A->cmap->n,0.,0.);CHKERRQ(ierr); 641 #else 642 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %18.16e\n",m,A->cmap->n,0.0);CHKERRQ(ierr); 643 #endif 644 } 645 ierr = PetscObjectGetName((PetscObject)A,&name);CHKERRQ(ierr); 646 ierr = PetscViewerASCIIPrintf(viewer,"];\n %s = spconvert(zzz);\n",name);CHKERRQ(ierr); 647 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 648 } else if (format == PETSC_VIEWER_ASCII_FACTOR_INFO || format == PETSC_VIEWER_ASCII_INFO) { 649 PetscFunctionReturn(0); 650 } else if (format == PETSC_VIEWER_ASCII_COMMON) { 651 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 652 for (i=0; i<m; i++) { 653 ierr = PetscViewerASCIIPrintf(viewer,"row %D:",i);CHKERRQ(ierr); 654 for (j=a->i[i]; j<a->i[i+1]; j++) { 655 #if defined(PETSC_USE_COMPLEX) 656 if (PetscImaginaryPart(a->a[j]) > 0.0 && PetscRealPart(a->a[j]) != 0.0) { 657 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g + %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 658 } else if (PetscImaginaryPart(a->a[j]) < 0.0 && PetscRealPart(a->a[j]) != 0.0) { 659 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g - %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)-PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 660 } else if (PetscRealPart(a->a[j]) != 0.0) { 661 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)PetscRealPart(a->a[j]));CHKERRQ(ierr); 662 } 663 #else 664 if (a->a[j] != 0.0) {ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)a->a[j]);CHKERRQ(ierr);} 665 #endif 666 } 667 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 668 } 669 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 670 } else if (format == PETSC_VIEWER_ASCII_SYMMODU) { 671 PetscInt nzd=0,fshift=1,*sptr; 672 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 673 ierr = PetscMalloc1(m+1,&sptr);CHKERRQ(ierr); 674 for (i=0; i<m; i++) { 675 sptr[i] = nzd+1; 676 for (j=a->i[i]; j<a->i[i+1]; j++) { 677 if (a->j[j] >= i) { 678 #if defined(PETSC_USE_COMPLEX) 679 if (PetscImaginaryPart(a->a[j]) != 0.0 || PetscRealPart(a->a[j]) != 0.0) nzd++; 680 #else 681 if (a->a[j] != 0.0) nzd++; 682 #endif 683 } 684 } 685 } 686 sptr[m] = nzd+1; 687 ierr = PetscViewerASCIIPrintf(viewer," %D %D\n\n",m,nzd);CHKERRQ(ierr); 688 for (i=0; i<m+1; i+=6) { 689 if (i+4<m) { 690 ierr = PetscViewerASCIIPrintf(viewer," %D %D %D %D %D %D\n",sptr[i],sptr[i+1],sptr[i+2],sptr[i+3],sptr[i+4],sptr[i+5]);CHKERRQ(ierr); 691 } else if (i+3<m) { 692 ierr = PetscViewerASCIIPrintf(viewer," %D %D %D %D %D\n",sptr[i],sptr[i+1],sptr[i+2],sptr[i+3],sptr[i+4]);CHKERRQ(ierr); 693 } else if (i+2<m) { 694 ierr = PetscViewerASCIIPrintf(viewer," %D %D %D %D\n",sptr[i],sptr[i+1],sptr[i+2],sptr[i+3]);CHKERRQ(ierr); 695 } else if (i+1<m) { 696 ierr = PetscViewerASCIIPrintf(viewer," %D %D %D\n",sptr[i],sptr[i+1],sptr[i+2]);CHKERRQ(ierr); 697 } else if (i<m) { 698 ierr = PetscViewerASCIIPrintf(viewer," %D %D\n",sptr[i],sptr[i+1]);CHKERRQ(ierr); 699 } else { 700 ierr = PetscViewerASCIIPrintf(viewer," %D\n",sptr[i]);CHKERRQ(ierr); 701 } 702 } 703 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 704 ierr = PetscFree(sptr);CHKERRQ(ierr); 705 for (i=0; i<m; i++) { 706 for (j=a->i[i]; j<a->i[i+1]; j++) { 707 if (a->j[j] >= i) {ierr = PetscViewerASCIIPrintf(viewer," %D ",a->j[j]+fshift);CHKERRQ(ierr);} 708 } 709 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 710 } 711 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 712 for (i=0; i<m; i++) { 713 for (j=a->i[i]; j<a->i[i+1]; j++) { 714 if (a->j[j] >= i) { 715 #if defined(PETSC_USE_COMPLEX) 716 if (PetscImaginaryPart(a->a[j]) != 0.0 || PetscRealPart(a->a[j]) != 0.0) { 717 ierr = PetscViewerASCIIPrintf(viewer," %18.16e %18.16e ",(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 718 } 719 #else 720 if (a->a[j] != 0.0) {ierr = PetscViewerASCIIPrintf(viewer," %18.16e ",(double)a->a[j]);CHKERRQ(ierr);} 721 #endif 722 } 723 } 724 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 725 } 726 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 727 } else if (format == PETSC_VIEWER_ASCII_DENSE) { 728 PetscInt cnt = 0,jcnt; 729 PetscScalar value; 730 #if defined(PETSC_USE_COMPLEX) 731 PetscBool realonly = PETSC_TRUE; 732 733 for (i=0; i<a->i[m]; i++) { 734 if (PetscImaginaryPart(a->a[i]) != 0.0) { 735 realonly = PETSC_FALSE; 736 break; 737 } 738 } 739 #endif 740 741 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 742 for (i=0; i<m; i++) { 743 jcnt = 0; 744 for (j=0; j<A->cmap->n; j++) { 745 if (jcnt < a->i[i+1]-a->i[i] && j == a->j[cnt]) { 746 value = a->a[cnt++]; 747 jcnt++; 748 } else { 749 value = 0.0; 750 } 751 #if defined(PETSC_USE_COMPLEX) 752 if (realonly) { 753 ierr = PetscViewerASCIIPrintf(viewer," %7.5e ",(double)PetscRealPart(value));CHKERRQ(ierr); 754 } else { 755 ierr = PetscViewerASCIIPrintf(viewer," %7.5e+%7.5e i ",(double)PetscRealPart(value),(double)PetscImaginaryPart(value));CHKERRQ(ierr); 756 } 757 #else 758 ierr = PetscViewerASCIIPrintf(viewer," %7.5e ",(double)value);CHKERRQ(ierr); 759 #endif 760 } 761 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 762 } 763 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 764 } else if (format == PETSC_VIEWER_ASCII_MATRIXMARKET) { 765 PetscInt fshift=1; 766 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 767 #if defined(PETSC_USE_COMPLEX) 768 ierr = PetscViewerASCIIPrintf(viewer,"%%%%MatrixMarket matrix coordinate complex general\n");CHKERRQ(ierr); 769 #else 770 ierr = PetscViewerASCIIPrintf(viewer,"%%%%MatrixMarket matrix coordinate real general\n");CHKERRQ(ierr); 771 #endif 772 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %D\n", m, A->cmap->n, a->nz);CHKERRQ(ierr); 773 for (i=0; i<m; i++) { 774 for (j=a->i[i]; j<a->i[i+1]; j++) { 775 #if defined(PETSC_USE_COMPLEX) 776 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %g %g\n", i+fshift,a->j[j]+fshift,(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 777 #else 778 ierr = PetscViewerASCIIPrintf(viewer,"%D %D %g\n", i+fshift, a->j[j]+fshift, (double)a->a[j]);CHKERRQ(ierr); 779 #endif 780 } 781 } 782 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 783 } else { 784 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);CHKERRQ(ierr); 785 if (A->factortype) { 786 for (i=0; i<m; i++) { 787 ierr = PetscViewerASCIIPrintf(viewer,"row %D:",i);CHKERRQ(ierr); 788 /* L part */ 789 for (j=a->i[i]; j<a->i[i+1]; j++) { 790 #if defined(PETSC_USE_COMPLEX) 791 if (PetscImaginaryPart(a->a[j]) > 0.0) { 792 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g + %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 793 } else if (PetscImaginaryPart(a->a[j]) < 0.0) { 794 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g - %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)(-PetscImaginaryPart(a->a[j])));CHKERRQ(ierr); 795 } else { 796 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)PetscRealPart(a->a[j]));CHKERRQ(ierr); 797 } 798 #else 799 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)a->a[j]);CHKERRQ(ierr); 800 #endif 801 } 802 /* diagonal */ 803 j = a->diag[i]; 804 #if defined(PETSC_USE_COMPLEX) 805 if (PetscImaginaryPart(a->a[j]) > 0.0) { 806 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g + %g i)",a->j[j],(double)PetscRealPart(1.0/a->a[j]),(double)PetscImaginaryPart(1.0/a->a[j]));CHKERRQ(ierr); 807 } else if (PetscImaginaryPart(a->a[j]) < 0.0) { 808 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g - %g i)",a->j[j],(double)PetscRealPart(1.0/a->a[j]),(double)(-PetscImaginaryPart(1.0/a->a[j])));CHKERRQ(ierr); 809 } else { 810 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)PetscRealPart(1.0/a->a[j]));CHKERRQ(ierr); 811 } 812 #else 813 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)(1.0/a->a[j]));CHKERRQ(ierr); 814 #endif 815 816 /* U part */ 817 for (j=a->diag[i+1]+1; j<a->diag[i]; j++) { 818 #if defined(PETSC_USE_COMPLEX) 819 if (PetscImaginaryPart(a->a[j]) > 0.0) { 820 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g + %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 821 } else if (PetscImaginaryPart(a->a[j]) < 0.0) { 822 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g - %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)(-PetscImaginaryPart(a->a[j])));CHKERRQ(ierr); 823 } else { 824 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)PetscRealPart(a->a[j]));CHKERRQ(ierr); 825 } 826 #else 827 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)a->a[j]);CHKERRQ(ierr); 828 #endif 829 } 830 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 831 } 832 } else { 833 for (i=0; i<m; i++) { 834 ierr = PetscViewerASCIIPrintf(viewer,"row %D:",i);CHKERRQ(ierr); 835 for (j=a->i[i]; j<a->i[i+1]; j++) { 836 #if defined(PETSC_USE_COMPLEX) 837 if (PetscImaginaryPart(a->a[j]) > 0.0) { 838 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g + %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 839 } else if (PetscImaginaryPart(a->a[j]) < 0.0) { 840 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g - %g i)",a->j[j],(double)PetscRealPart(a->a[j]),(double)-PetscImaginaryPart(a->a[j]));CHKERRQ(ierr); 841 } else { 842 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)PetscRealPart(a->a[j]));CHKERRQ(ierr); 843 } 844 #else 845 ierr = PetscViewerASCIIPrintf(viewer," (%D, %g) ",a->j[j],(double)a->a[j]);CHKERRQ(ierr); 846 #endif 847 } 848 ierr = PetscViewerASCIIPrintf(viewer,"\n");CHKERRQ(ierr); 849 } 850 } 851 ierr = PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);CHKERRQ(ierr); 852 } 853 ierr = PetscViewerFlush(viewer);CHKERRQ(ierr); 854 PetscFunctionReturn(0); 855 } 856 857 #include <petscdraw.h> 858 PetscErrorCode MatView_SeqAIJ_Draw_Zoom(PetscDraw draw,void *Aa) 859 { 860 Mat A = (Mat) Aa; 861 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 862 PetscErrorCode ierr; 863 PetscInt i,j,m = A->rmap->n; 864 int color; 865 PetscReal xl,yl,xr,yr,x_l,x_r,y_l,y_r; 866 PetscViewer viewer; 867 PetscViewerFormat format; 868 869 PetscFunctionBegin; 870 ierr = PetscObjectQuery((PetscObject)A,"Zoomviewer",(PetscObject*)&viewer);CHKERRQ(ierr); 871 ierr = PetscViewerGetFormat(viewer,&format);CHKERRQ(ierr); 872 ierr = PetscDrawGetCoordinates(draw,&xl,&yl,&xr,&yr);CHKERRQ(ierr); 873 874 /* loop over matrix elements drawing boxes */ 875 876 if (format != PETSC_VIEWER_DRAW_CONTOUR) { 877 ierr = PetscDrawCollectiveBegin(draw);CHKERRQ(ierr); 878 /* Blue for negative, Cyan for zero and Red for positive */ 879 color = PETSC_DRAW_BLUE; 880 for (i=0; i<m; i++) { 881 y_l = m - i - 1.0; y_r = y_l + 1.0; 882 for (j=a->i[i]; j<a->i[i+1]; j++) { 883 x_l = a->j[j]; x_r = x_l + 1.0; 884 if (PetscRealPart(a->a[j]) >= 0.) continue; 885 ierr = PetscDrawRectangle(draw,x_l,y_l,x_r,y_r,color,color,color,color);CHKERRQ(ierr); 886 } 887 } 888 color = PETSC_DRAW_CYAN; 889 for (i=0; i<m; i++) { 890 y_l = m - i - 1.0; y_r = y_l + 1.0; 891 for (j=a->i[i]; j<a->i[i+1]; j++) { 892 x_l = a->j[j]; x_r = x_l + 1.0; 893 if (a->a[j] != 0.) continue; 894 ierr = PetscDrawRectangle(draw,x_l,y_l,x_r,y_r,color,color,color,color);CHKERRQ(ierr); 895 } 896 } 897 color = PETSC_DRAW_RED; 898 for (i=0; i<m; i++) { 899 y_l = m - i - 1.0; y_r = y_l + 1.0; 900 for (j=a->i[i]; j<a->i[i+1]; j++) { 901 x_l = a->j[j]; x_r = x_l + 1.0; 902 if (PetscRealPart(a->a[j]) <= 0.) continue; 903 ierr = PetscDrawRectangle(draw,x_l,y_l,x_r,y_r,color,color,color,color);CHKERRQ(ierr); 904 } 905 } 906 ierr = PetscDrawCollectiveEnd(draw);CHKERRQ(ierr); 907 } else { 908 /* use contour shading to indicate magnitude of values */ 909 /* first determine max of all nonzero values */ 910 PetscReal minv = 0.0, maxv = 0.0; 911 PetscInt nz = a->nz, count = 0; 912 PetscDraw popup; 913 914 for (i=0; i<nz; i++) { 915 if (PetscAbsScalar(a->a[i]) > maxv) maxv = PetscAbsScalar(a->a[i]); 916 } 917 if (minv >= maxv) maxv = minv + PETSC_SMALL; 918 ierr = PetscDrawGetPopup(draw,&popup);CHKERRQ(ierr); 919 ierr = PetscDrawScalePopup(popup,minv,maxv);CHKERRQ(ierr); 920 921 ierr = PetscDrawCollectiveBegin(draw);CHKERRQ(ierr); 922 for (i=0; i<m; i++) { 923 y_l = m - i - 1.0; 924 y_r = y_l + 1.0; 925 for (j=a->i[i]; j<a->i[i+1]; j++) { 926 x_l = a->j[j]; 927 x_r = x_l + 1.0; 928 color = PetscDrawRealToColor(PetscAbsScalar(a->a[count]),minv,maxv); 929 ierr = PetscDrawRectangle(draw,x_l,y_l,x_r,y_r,color,color,color,color);CHKERRQ(ierr); 930 count++; 931 } 932 } 933 ierr = PetscDrawCollectiveEnd(draw);CHKERRQ(ierr); 934 } 935 PetscFunctionReturn(0); 936 } 937 938 #include <petscdraw.h> 939 PetscErrorCode MatView_SeqAIJ_Draw(Mat A,PetscViewer viewer) 940 { 941 PetscErrorCode ierr; 942 PetscDraw draw; 943 PetscReal xr,yr,xl,yl,h,w; 944 PetscBool isnull; 945 946 PetscFunctionBegin; 947 ierr = PetscViewerDrawGetDraw(viewer,0,&draw);CHKERRQ(ierr); 948 ierr = PetscDrawIsNull(draw,&isnull);CHKERRQ(ierr); 949 if (isnull) PetscFunctionReturn(0); 950 951 xr = A->cmap->n; yr = A->rmap->n; h = yr/10.0; w = xr/10.0; 952 xr += w; yr += h; xl = -w; yl = -h; 953 ierr = PetscDrawSetCoordinates(draw,xl,yl,xr,yr);CHKERRQ(ierr); 954 ierr = PetscObjectCompose((PetscObject)A,"Zoomviewer",(PetscObject)viewer);CHKERRQ(ierr); 955 ierr = PetscDrawZoom(draw,MatView_SeqAIJ_Draw_Zoom,A);CHKERRQ(ierr); 956 ierr = PetscObjectCompose((PetscObject)A,"Zoomviewer",NULL);CHKERRQ(ierr); 957 ierr = PetscDrawSave(draw);CHKERRQ(ierr); 958 PetscFunctionReturn(0); 959 } 960 961 PetscErrorCode MatView_SeqAIJ(Mat A,PetscViewer viewer) 962 { 963 PetscErrorCode ierr; 964 PetscBool iascii,isbinary,isdraw; 965 966 PetscFunctionBegin; 967 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);CHKERRQ(ierr); 968 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 969 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);CHKERRQ(ierr); 970 if (iascii) { 971 ierr = MatView_SeqAIJ_ASCII(A,viewer);CHKERRQ(ierr); 972 } else if (isbinary) { 973 ierr = MatView_SeqAIJ_Binary(A,viewer);CHKERRQ(ierr); 974 } else if (isdraw) { 975 ierr = MatView_SeqAIJ_Draw(A,viewer);CHKERRQ(ierr); 976 } 977 ierr = MatView_SeqAIJ_Inode(A,viewer);CHKERRQ(ierr); 978 PetscFunctionReturn(0); 979 } 980 981 PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat A,MatAssemblyType mode) 982 { 983 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 984 PetscErrorCode ierr; 985 PetscInt fshift = 0,i,j,*ai = a->i,*aj = a->j,*imax = a->imax; 986 PetscInt m = A->rmap->n,*ip,N,*ailen = a->ilen,rmax = 0; 987 MatScalar *aa = a->a,*ap; 988 PetscReal ratio = 0.6; 989 990 PetscFunctionBegin; 991 if (mode == MAT_FLUSH_ASSEMBLY) PetscFunctionReturn(0); 992 993 if (m) rmax = ailen[0]; /* determine row with most nonzeros */ 994 for (i=1; i<m; i++) { 995 /* move each row back by the amount of empty slots (fshift) before it*/ 996 fshift += imax[i-1] - ailen[i-1]; 997 rmax = PetscMax(rmax,ailen[i]); 998 if (fshift) { 999 ip = aj + ai[i]; 1000 ap = aa + ai[i]; 1001 N = ailen[i]; 1002 for (j=0; j<N; j++) { 1003 ip[j-fshift] = ip[j]; 1004 if (!A->structure_only) ap[j-fshift] = ap[j]; 1005 } 1006 } 1007 ai[i] = ai[i-1] + ailen[i-1]; 1008 } 1009 if (m) { 1010 fshift += imax[m-1] - ailen[m-1]; 1011 ai[m] = ai[m-1] + ailen[m-1]; 1012 } 1013 1014 /* reset ilen and imax for each row */ 1015 a->nonzerorowcnt = 0; 1016 if (A->structure_only) { 1017 ierr = PetscFree2(a->imax,a->ilen);CHKERRQ(ierr); 1018 } else { /* !A->structure_only */ 1019 for (i=0; i<m; i++) { 1020 ailen[i] = imax[i] = ai[i+1] - ai[i]; 1021 a->nonzerorowcnt += ((ai[i+1] - ai[i]) > 0); 1022 } 1023 } 1024 a->nz = ai[m]; 1025 if (fshift && a->nounused == -1) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_PLIB, "Unused space detected in matrix: %D X %D, %D unneeded", m, A->cmap->n, fshift); 1026 1027 ierr = MatMarkDiagonal_SeqAIJ(A);CHKERRQ(ierr); 1028 ierr = PetscInfo4(A,"Matrix size: %D X %D; storage space: %D unneeded,%D used\n",m,A->cmap->n,fshift,a->nz);CHKERRQ(ierr); 1029 ierr = PetscInfo1(A,"Number of mallocs during MatSetValues() is %D\n",a->reallocs);CHKERRQ(ierr); 1030 ierr = PetscInfo1(A,"Maximum nonzeros in any row is %D\n",rmax);CHKERRQ(ierr); 1031 1032 A->info.mallocs += a->reallocs; 1033 a->reallocs = 0; 1034 A->info.nz_unneeded = (PetscReal)fshift; 1035 a->rmax = rmax; 1036 1037 if (!A->structure_only) { 1038 ierr = MatCheckCompressedRow(A,a->nonzerorowcnt,&a->compressedrow,a->i,m,ratio);CHKERRQ(ierr); 1039 } 1040 ierr = MatAssemblyEnd_SeqAIJ_Inode(A,mode);CHKERRQ(ierr); 1041 ierr = MatSeqAIJInvalidateDiagonal(A);CHKERRQ(ierr); 1042 PetscFunctionReturn(0); 1043 } 1044 1045 PetscErrorCode MatRealPart_SeqAIJ(Mat A) 1046 { 1047 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1048 PetscInt i,nz = a->nz; 1049 MatScalar *aa = a->a; 1050 PetscErrorCode ierr; 1051 1052 PetscFunctionBegin; 1053 for (i=0; i<nz; i++) aa[i] = PetscRealPart(aa[i]); 1054 ierr = MatSeqAIJInvalidateDiagonal(A);CHKERRQ(ierr); 1055 PetscFunctionReturn(0); 1056 } 1057 1058 PetscErrorCode MatImaginaryPart_SeqAIJ(Mat A) 1059 { 1060 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1061 PetscInt i,nz = a->nz; 1062 MatScalar *aa = a->a; 1063 PetscErrorCode ierr; 1064 1065 PetscFunctionBegin; 1066 for (i=0; i<nz; i++) aa[i] = PetscImaginaryPart(aa[i]); 1067 ierr = MatSeqAIJInvalidateDiagonal(A);CHKERRQ(ierr); 1068 PetscFunctionReturn(0); 1069 } 1070 1071 PetscErrorCode MatZeroEntries_SeqAIJ(Mat A) 1072 { 1073 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1074 PetscErrorCode ierr; 1075 1076 PetscFunctionBegin; 1077 ierr = PetscMemzero(a->a,(a->i[A->rmap->n])*sizeof(PetscScalar));CHKERRQ(ierr); 1078 ierr = MatSeqAIJInvalidateDiagonal(A);CHKERRQ(ierr); 1079 PetscFunctionReturn(0); 1080 } 1081 1082 PetscErrorCode MatDestroy_SeqAIJ(Mat A) 1083 { 1084 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1085 PetscErrorCode ierr; 1086 1087 PetscFunctionBegin; 1088 #if defined(PETSC_USE_LOG) 1089 PetscLogObjectState((PetscObject)A,"Rows=%D, Cols=%D, NZ=%D",A->rmap->n,A->cmap->n,a->nz); 1090 #endif 1091 ierr = MatSeqXAIJFreeAIJ(A,&a->a,&a->j,&a->i);CHKERRQ(ierr); 1092 ierr = ISDestroy(&a->row);CHKERRQ(ierr); 1093 ierr = ISDestroy(&a->col);CHKERRQ(ierr); 1094 ierr = PetscFree(a->diag);CHKERRQ(ierr); 1095 ierr = PetscFree(a->ibdiag);CHKERRQ(ierr); 1096 ierr = PetscFree2(a->imax,a->ilen);CHKERRQ(ierr); 1097 ierr = PetscFree(a->ipre);CHKERRQ(ierr); 1098 ierr = PetscFree3(a->idiag,a->mdiag,a->ssor_work);CHKERRQ(ierr); 1099 ierr = PetscFree(a->solve_work);CHKERRQ(ierr); 1100 ierr = ISDestroy(&a->icol);CHKERRQ(ierr); 1101 ierr = PetscFree(a->saved_values);CHKERRQ(ierr); 1102 ierr = ISColoringDestroy(&a->coloring);CHKERRQ(ierr); 1103 ierr = PetscFree2(a->compressedrow.i,a->compressedrow.rindex);CHKERRQ(ierr); 1104 ierr = PetscFree(a->matmult_abdense);CHKERRQ(ierr); 1105 1106 ierr = MatDestroy_SeqAIJ_Inode(A);CHKERRQ(ierr); 1107 ierr = PetscFree(A->data);CHKERRQ(ierr); 1108 1109 ierr = PetscObjectChangeTypeName((PetscObject)A,0);CHKERRQ(ierr); 1110 ierr = PetscObjectComposeFunction((PetscObject)A,"MatSeqAIJSetColumnIndices_C",NULL);CHKERRQ(ierr); 1111 ierr = PetscObjectComposeFunction((PetscObject)A,"MatStoreValues_C",NULL);CHKERRQ(ierr); 1112 ierr = PetscObjectComposeFunction((PetscObject)A,"MatRetrieveValues_C",NULL);CHKERRQ(ierr); 1113 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_seqsbaij_C",NULL);CHKERRQ(ierr); 1114 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_seqbaij_C",NULL);CHKERRQ(ierr); 1115 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_seqaijperm_C",NULL);CHKERRQ(ierr); 1116 #if defined(PETSC_HAVE_ELEMENTAL) 1117 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_elemental_C",NULL);CHKERRQ(ierr); 1118 #endif 1119 #if defined(PETSC_HAVE_HYPRE) 1120 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_hypre_C",NULL);CHKERRQ(ierr); 1121 ierr = PetscObjectComposeFunction((PetscObject)A,"MatMatMatMult_transpose_seqaij_seqaij_C",NULL);CHKERRQ(ierr); 1122 #endif 1123 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_seqdense_C",NULL);CHKERRQ(ierr); 1124 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_seqsell_C",NULL);CHKERRQ(ierr); 1125 ierr = PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqaij_is_C",NULL);CHKERRQ(ierr); 1126 ierr = PetscObjectComposeFunction((PetscObject)A,"MatIsTranspose_C",NULL);CHKERRQ(ierr); 1127 ierr = PetscObjectComposeFunction((PetscObject)A,"MatSeqAIJSetPreallocation_C",NULL);CHKERRQ(ierr); 1128 ierr = PetscObjectComposeFunction((PetscObject)A,"MatResetPreallocation_C",NULL);CHKERRQ(ierr); 1129 ierr = PetscObjectComposeFunction((PetscObject)A,"MatSeqAIJSetPreallocationCSR_C",NULL);CHKERRQ(ierr); 1130 ierr = PetscObjectComposeFunction((PetscObject)A,"MatReorderForNonzeroDiagonal_C",NULL);CHKERRQ(ierr); 1131 ierr = PetscObjectComposeFunction((PetscObject)A,"MatPtAP_is_seqaij_C",NULL);CHKERRQ(ierr); 1132 PetscFunctionReturn(0); 1133 } 1134 1135 PetscErrorCode MatSetOption_SeqAIJ(Mat A,MatOption op,PetscBool flg) 1136 { 1137 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1138 PetscErrorCode ierr; 1139 1140 PetscFunctionBegin; 1141 switch (op) { 1142 case MAT_ROW_ORIENTED: 1143 a->roworiented = flg; 1144 break; 1145 case MAT_KEEP_NONZERO_PATTERN: 1146 a->keepnonzeropattern = flg; 1147 break; 1148 case MAT_NEW_NONZERO_LOCATIONS: 1149 a->nonew = (flg ? 0 : 1); 1150 break; 1151 case MAT_NEW_NONZERO_LOCATION_ERR: 1152 a->nonew = (flg ? -1 : 0); 1153 break; 1154 case MAT_NEW_NONZERO_ALLOCATION_ERR: 1155 a->nonew = (flg ? -2 : 0); 1156 break; 1157 case MAT_UNUSED_NONZERO_LOCATION_ERR: 1158 a->nounused = (flg ? -1 : 0); 1159 break; 1160 case MAT_IGNORE_ZERO_ENTRIES: 1161 a->ignorezeroentries = flg; 1162 break; 1163 case MAT_SPD: 1164 case MAT_SYMMETRIC: 1165 case MAT_STRUCTURALLY_SYMMETRIC: 1166 case MAT_HERMITIAN: 1167 case MAT_SYMMETRY_ETERNAL: 1168 case MAT_STRUCTURE_ONLY: 1169 /* These options are handled directly by MatSetOption() */ 1170 break; 1171 case MAT_NEW_DIAGONALS: 1172 case MAT_IGNORE_OFF_PROC_ENTRIES: 1173 case MAT_USE_HASH_TABLE: 1174 ierr = PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);CHKERRQ(ierr); 1175 break; 1176 case MAT_USE_INODES: 1177 /* Not an error because MatSetOption_SeqAIJ_Inode handles this one */ 1178 break; 1179 case MAT_SUBMAT_SINGLEIS: 1180 A->submat_singleis = flg; 1181 break; 1182 default: 1183 SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"unknown option %d",op); 1184 } 1185 ierr = MatSetOption_SeqAIJ_Inode(A,op,flg);CHKERRQ(ierr); 1186 PetscFunctionReturn(0); 1187 } 1188 1189 PetscErrorCode MatGetDiagonal_SeqAIJ(Mat A,Vec v) 1190 { 1191 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1192 PetscErrorCode ierr; 1193 PetscInt i,j,n,*ai=a->i,*aj=a->j,nz; 1194 PetscScalar *aa=a->a,*x,zero=0.0; 1195 1196 PetscFunctionBegin; 1197 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 1198 if (n != A->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector"); 1199 1200 if (A->factortype == MAT_FACTOR_ILU || A->factortype == MAT_FACTOR_LU) { 1201 PetscInt *diag=a->diag; 1202 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 1203 for (i=0; i<n; i++) x[i] = 1.0/aa[diag[i]]; 1204 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 1205 PetscFunctionReturn(0); 1206 } 1207 1208 ierr = VecSet(v,zero);CHKERRQ(ierr); 1209 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 1210 for (i=0; i<n; i++) { 1211 nz = ai[i+1] - ai[i]; 1212 if (!nz) x[i] = 0.0; 1213 for (j=ai[i]; j<ai[i+1]; j++) { 1214 if (aj[j] == i) { 1215 x[i] = aa[j]; 1216 break; 1217 } 1218 } 1219 } 1220 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 1221 PetscFunctionReturn(0); 1222 } 1223 1224 #include <../src/mat/impls/aij/seq/ftn-kernels/fmult.h> 1225 PetscErrorCode MatMultTransposeAdd_SeqAIJ(Mat A,Vec xx,Vec zz,Vec yy) 1226 { 1227 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1228 PetscScalar *y; 1229 const PetscScalar *x; 1230 PetscErrorCode ierr; 1231 PetscInt m = A->rmap->n; 1232 #if !defined(PETSC_USE_FORTRAN_KERNEL_MULTTRANSPOSEAIJ) 1233 const MatScalar *v; 1234 PetscScalar alpha; 1235 PetscInt n,i,j; 1236 const PetscInt *idx,*ii,*ridx=NULL; 1237 Mat_CompressedRow cprow = a->compressedrow; 1238 PetscBool usecprow = cprow.use; 1239 #endif 1240 1241 PetscFunctionBegin; 1242 if (zz != yy) {ierr = VecCopy(zz,yy);CHKERRQ(ierr);} 1243 ierr = VecGetArrayRead(xx,&x);CHKERRQ(ierr); 1244 ierr = VecGetArray(yy,&y);CHKERRQ(ierr); 1245 1246 #if defined(PETSC_USE_FORTRAN_KERNEL_MULTTRANSPOSEAIJ) 1247 fortranmulttransposeaddaij_(&m,x,a->i,a->j,a->a,y); 1248 #else 1249 if (usecprow) { 1250 m = cprow.nrows; 1251 ii = cprow.i; 1252 ridx = cprow.rindex; 1253 } else { 1254 ii = a->i; 1255 } 1256 for (i=0; i<m; i++) { 1257 idx = a->j + ii[i]; 1258 v = a->a + ii[i]; 1259 n = ii[i+1] - ii[i]; 1260 if (usecprow) { 1261 alpha = x[ridx[i]]; 1262 } else { 1263 alpha = x[i]; 1264 } 1265 for (j=0; j<n; j++) y[idx[j]] += alpha*v[j]; 1266 } 1267 #endif 1268 ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr); 1269 ierr = VecRestoreArrayRead(xx,&x);CHKERRQ(ierr); 1270 ierr = VecRestoreArray(yy,&y);CHKERRQ(ierr); 1271 PetscFunctionReturn(0); 1272 } 1273 1274 PetscErrorCode MatMultTranspose_SeqAIJ(Mat A,Vec xx,Vec yy) 1275 { 1276 PetscErrorCode ierr; 1277 1278 PetscFunctionBegin; 1279 ierr = VecSet(yy,0.0);CHKERRQ(ierr); 1280 ierr = MatMultTransposeAdd_SeqAIJ(A,xx,yy,yy);CHKERRQ(ierr); 1281 PetscFunctionReturn(0); 1282 } 1283 1284 #include <../src/mat/impls/aij/seq/ftn-kernels/fmult.h> 1285 1286 PetscErrorCode MatMult_SeqAIJ(Mat A,Vec xx,Vec yy) 1287 { 1288 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1289 PetscScalar *y; 1290 const PetscScalar *x; 1291 const MatScalar *aa; 1292 PetscErrorCode ierr; 1293 PetscInt m=A->rmap->n; 1294 const PetscInt *aj,*ii,*ridx=NULL; 1295 PetscInt n,i; 1296 PetscScalar sum; 1297 PetscBool usecprow=a->compressedrow.use; 1298 1299 #if defined(PETSC_HAVE_PRAGMA_DISJOINT) 1300 #pragma disjoint(*x,*y,*aa) 1301 #endif 1302 1303 PetscFunctionBegin; 1304 ierr = VecGetArrayRead(xx,&x);CHKERRQ(ierr); 1305 ierr = VecGetArray(yy,&y);CHKERRQ(ierr); 1306 ii = a->i; 1307 if (usecprow) { /* use compressed row format */ 1308 ierr = PetscMemzero(y,m*sizeof(PetscScalar));CHKERRQ(ierr); 1309 m = a->compressedrow.nrows; 1310 ii = a->compressedrow.i; 1311 ridx = a->compressedrow.rindex; 1312 for (i=0; i<m; i++) { 1313 n = ii[i+1] - ii[i]; 1314 aj = a->j + ii[i]; 1315 aa = a->a + ii[i]; 1316 sum = 0.0; 1317 PetscSparseDensePlusDot(sum,x,aa,aj,n); 1318 /* for (j=0; j<n; j++) sum += (*aa++)*x[*aj++]; */ 1319 y[*ridx++] = sum; 1320 } 1321 } else { /* do not use compressed row format */ 1322 #if defined(PETSC_USE_FORTRAN_KERNEL_MULTAIJ) 1323 aj = a->j; 1324 aa = a->a; 1325 fortranmultaij_(&m,x,ii,aj,aa,y); 1326 #else 1327 for (i=0; i<m; i++) { 1328 n = ii[i+1] - ii[i]; 1329 aj = a->j + ii[i]; 1330 aa = a->a + ii[i]; 1331 sum = 0.0; 1332 PetscSparseDensePlusDot(sum,x,aa,aj,n); 1333 y[i] = sum; 1334 } 1335 #endif 1336 } 1337 ierr = PetscLogFlops(2.0*a->nz - a->nonzerorowcnt);CHKERRQ(ierr); 1338 ierr = VecRestoreArrayRead(xx,&x);CHKERRQ(ierr); 1339 ierr = VecRestoreArray(yy,&y);CHKERRQ(ierr); 1340 PetscFunctionReturn(0); 1341 } 1342 1343 PetscErrorCode MatMultMax_SeqAIJ(Mat A,Vec xx,Vec yy) 1344 { 1345 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1346 PetscScalar *y; 1347 const PetscScalar *x; 1348 const MatScalar *aa; 1349 PetscErrorCode ierr; 1350 PetscInt m=A->rmap->n; 1351 const PetscInt *aj,*ii,*ridx=NULL; 1352 PetscInt n,i,nonzerorow=0; 1353 PetscScalar sum; 1354 PetscBool usecprow=a->compressedrow.use; 1355 1356 #if defined(PETSC_HAVE_PRAGMA_DISJOINT) 1357 #pragma disjoint(*x,*y,*aa) 1358 #endif 1359 1360 PetscFunctionBegin; 1361 ierr = VecGetArrayRead(xx,&x);CHKERRQ(ierr); 1362 ierr = VecGetArray(yy,&y);CHKERRQ(ierr); 1363 if (usecprow) { /* use compressed row format */ 1364 m = a->compressedrow.nrows; 1365 ii = a->compressedrow.i; 1366 ridx = a->compressedrow.rindex; 1367 for (i=0; i<m; i++) { 1368 n = ii[i+1] - ii[i]; 1369 aj = a->j + ii[i]; 1370 aa = a->a + ii[i]; 1371 sum = 0.0; 1372 nonzerorow += (n>0); 1373 PetscSparseDenseMaxDot(sum,x,aa,aj,n); 1374 /* for (j=0; j<n; j++) sum += (*aa++)*x[*aj++]; */ 1375 y[*ridx++] = sum; 1376 } 1377 } else { /* do not use compressed row format */ 1378 ii = a->i; 1379 for (i=0; i<m; i++) { 1380 n = ii[i+1] - ii[i]; 1381 aj = a->j + ii[i]; 1382 aa = a->a + ii[i]; 1383 sum = 0.0; 1384 nonzerorow += (n>0); 1385 PetscSparseDenseMaxDot(sum,x,aa,aj,n); 1386 y[i] = sum; 1387 } 1388 } 1389 ierr = PetscLogFlops(2.0*a->nz - nonzerorow);CHKERRQ(ierr); 1390 ierr = VecRestoreArrayRead(xx,&x);CHKERRQ(ierr); 1391 ierr = VecRestoreArray(yy,&y);CHKERRQ(ierr); 1392 PetscFunctionReturn(0); 1393 } 1394 1395 PetscErrorCode MatMultAddMax_SeqAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1396 { 1397 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1398 PetscScalar *y,*z; 1399 const PetscScalar *x; 1400 const MatScalar *aa; 1401 PetscErrorCode ierr; 1402 PetscInt m = A->rmap->n,*aj,*ii; 1403 PetscInt n,i,*ridx=NULL; 1404 PetscScalar sum; 1405 PetscBool usecprow=a->compressedrow.use; 1406 1407 PetscFunctionBegin; 1408 ierr = VecGetArrayRead(xx,&x);CHKERRQ(ierr); 1409 ierr = VecGetArrayPair(yy,zz,&y,&z);CHKERRQ(ierr); 1410 if (usecprow) { /* use compressed row format */ 1411 if (zz != yy) { 1412 ierr = PetscMemcpy(z,y,m*sizeof(PetscScalar));CHKERRQ(ierr); 1413 } 1414 m = a->compressedrow.nrows; 1415 ii = a->compressedrow.i; 1416 ridx = a->compressedrow.rindex; 1417 for (i=0; i<m; i++) { 1418 n = ii[i+1] - ii[i]; 1419 aj = a->j + ii[i]; 1420 aa = a->a + ii[i]; 1421 sum = y[*ridx]; 1422 PetscSparseDenseMaxDot(sum,x,aa,aj,n); 1423 z[*ridx++] = sum; 1424 } 1425 } else { /* do not use compressed row format */ 1426 ii = a->i; 1427 for (i=0; i<m; i++) { 1428 n = ii[i+1] - ii[i]; 1429 aj = a->j + ii[i]; 1430 aa = a->a + ii[i]; 1431 sum = y[i]; 1432 PetscSparseDenseMaxDot(sum,x,aa,aj,n); 1433 z[i] = sum; 1434 } 1435 } 1436 ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr); 1437 ierr = VecRestoreArrayRead(xx,&x);CHKERRQ(ierr); 1438 ierr = VecRestoreArrayPair(yy,zz,&y,&z);CHKERRQ(ierr); 1439 PetscFunctionReturn(0); 1440 } 1441 1442 #include <../src/mat/impls/aij/seq/ftn-kernels/fmultadd.h> 1443 PetscErrorCode MatMultAdd_SeqAIJ(Mat A,Vec xx,Vec yy,Vec zz) 1444 { 1445 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1446 PetscScalar *y,*z; 1447 const PetscScalar *x; 1448 const MatScalar *aa; 1449 PetscErrorCode ierr; 1450 const PetscInt *aj,*ii,*ridx=NULL; 1451 PetscInt m = A->rmap->n,n,i; 1452 PetscScalar sum; 1453 PetscBool usecprow=a->compressedrow.use; 1454 1455 PetscFunctionBegin; 1456 ierr = VecGetArrayRead(xx,&x);CHKERRQ(ierr); 1457 ierr = VecGetArrayPair(yy,zz,&y,&z);CHKERRQ(ierr); 1458 if (usecprow) { /* use compressed row format */ 1459 if (zz != yy) { 1460 ierr = PetscMemcpy(z,y,m*sizeof(PetscScalar));CHKERRQ(ierr); 1461 } 1462 m = a->compressedrow.nrows; 1463 ii = a->compressedrow.i; 1464 ridx = a->compressedrow.rindex; 1465 for (i=0; i<m; i++) { 1466 n = ii[i+1] - ii[i]; 1467 aj = a->j + ii[i]; 1468 aa = a->a + ii[i]; 1469 sum = y[*ridx]; 1470 PetscSparseDensePlusDot(sum,x,aa,aj,n); 1471 z[*ridx++] = sum; 1472 } 1473 } else { /* do not use compressed row format */ 1474 ii = a->i; 1475 #if defined(PETSC_USE_FORTRAN_KERNEL_MULTADDAIJ) 1476 aj = a->j; 1477 aa = a->a; 1478 fortranmultaddaij_(&m,x,ii,aj,aa,y,z); 1479 #else 1480 for (i=0; i<m; i++) { 1481 n = ii[i+1] - ii[i]; 1482 aj = a->j + ii[i]; 1483 aa = a->a + ii[i]; 1484 sum = y[i]; 1485 PetscSparseDensePlusDot(sum,x,aa,aj,n); 1486 z[i] = sum; 1487 } 1488 #endif 1489 } 1490 ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr); 1491 ierr = VecRestoreArrayRead(xx,&x);CHKERRQ(ierr); 1492 ierr = VecRestoreArrayPair(yy,zz,&y,&z);CHKERRQ(ierr); 1493 PetscFunctionReturn(0); 1494 } 1495 1496 /* 1497 Adds diagonal pointers to sparse matrix structure. 1498 */ 1499 PetscErrorCode MatMarkDiagonal_SeqAIJ(Mat A) 1500 { 1501 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1502 PetscErrorCode ierr; 1503 PetscInt i,j,m = A->rmap->n; 1504 1505 PetscFunctionBegin; 1506 if (!a->diag) { 1507 ierr = PetscMalloc1(m,&a->diag);CHKERRQ(ierr); 1508 ierr = PetscLogObjectMemory((PetscObject)A, m*sizeof(PetscInt));CHKERRQ(ierr); 1509 } 1510 for (i=0; i<A->rmap->n; i++) { 1511 a->diag[i] = a->i[i+1]; 1512 for (j=a->i[i]; j<a->i[i+1]; j++) { 1513 if (a->j[j] == i) { 1514 a->diag[i] = j; 1515 break; 1516 } 1517 } 1518 } 1519 PetscFunctionReturn(0); 1520 } 1521 1522 PetscErrorCode MatShift_SeqAIJ(Mat A,PetscScalar v) 1523 { 1524 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1525 const PetscInt *diag = (const PetscInt*)a->diag; 1526 const PetscInt *ii = (const PetscInt*) a->i; 1527 PetscInt i,*mdiag = NULL; 1528 PetscErrorCode ierr; 1529 PetscInt cnt = 0; /* how many diagonals are missing */ 1530 1531 PetscFunctionBegin; 1532 if (!A->preallocated || !a->nz) { 1533 ierr = MatSeqAIJSetPreallocation(A,1,NULL);CHKERRQ(ierr); 1534 ierr = MatShift_Basic(A,v);CHKERRQ(ierr); 1535 PetscFunctionReturn(0); 1536 } 1537 1538 if (a->diagonaldense) { 1539 cnt = 0; 1540 } else { 1541 ierr = PetscCalloc1(A->rmap->n,&mdiag);CHKERRQ(ierr); 1542 for (i=0; i<A->rmap->n; i++) { 1543 if (diag[i] >= ii[i+1]) { 1544 cnt++; 1545 mdiag[i] = 1; 1546 } 1547 } 1548 } 1549 if (!cnt) { 1550 ierr = MatShift_Basic(A,v);CHKERRQ(ierr); 1551 } else { 1552 PetscScalar *olda = a->a; /* preserve pointers to current matrix nonzeros structure and values */ 1553 PetscInt *oldj = a->j, *oldi = a->i; 1554 PetscBool singlemalloc = a->singlemalloc,free_a = a->free_a,free_ij = a->free_ij; 1555 1556 a->a = NULL; 1557 a->j = NULL; 1558 a->i = NULL; 1559 /* increase the values in imax for each row where a diagonal is being inserted then reallocate the matrix data structures */ 1560 for (i=0; i<A->rmap->n; i++) { 1561 a->imax[i] += mdiag[i]; 1562 a->imax[i] = PetscMin(a->imax[i],A->cmap->n); 1563 } 1564 ierr = MatSeqAIJSetPreallocation_SeqAIJ(A,0,a->imax);CHKERRQ(ierr); 1565 1566 /* copy old values into new matrix data structure */ 1567 for (i=0; i<A->rmap->n; i++) { 1568 ierr = MatSetValues(A,1,&i,a->imax[i] - mdiag[i],&oldj[oldi[i]],&olda[oldi[i]],ADD_VALUES);CHKERRQ(ierr); 1569 if (i < A->cmap->n) { 1570 ierr = MatSetValue(A,i,i,v,ADD_VALUES);CHKERRQ(ierr); 1571 } 1572 } 1573 ierr = MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1574 ierr = MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1575 if (singlemalloc) { 1576 ierr = PetscFree3(olda,oldj,oldi);CHKERRQ(ierr); 1577 } else { 1578 if (free_a) {ierr = PetscFree(olda);CHKERRQ(ierr);} 1579 if (free_ij) {ierr = PetscFree(oldj);CHKERRQ(ierr);} 1580 if (free_ij) {ierr = PetscFree(oldi);CHKERRQ(ierr);} 1581 } 1582 } 1583 ierr = PetscFree(mdiag);CHKERRQ(ierr); 1584 a->diagonaldense = PETSC_TRUE; 1585 PetscFunctionReturn(0); 1586 } 1587 1588 /* 1589 Checks for missing diagonals 1590 */ 1591 PetscErrorCode MatMissingDiagonal_SeqAIJ(Mat A,PetscBool *missing,PetscInt *d) 1592 { 1593 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1594 PetscInt *diag,*ii = a->i,i; 1595 PetscErrorCode ierr; 1596 1597 PetscFunctionBegin; 1598 *missing = PETSC_FALSE; 1599 if (A->rmap->n > 0 && !ii) { 1600 *missing = PETSC_TRUE; 1601 if (d) *d = 0; 1602 ierr = PetscInfo(A,"Matrix has no entries therefore is missing diagonal\n");CHKERRQ(ierr); 1603 } else { 1604 diag = a->diag; 1605 for (i=0; i<A->rmap->n; i++) { 1606 if (diag[i] >= ii[i+1]) { 1607 *missing = PETSC_TRUE; 1608 if (d) *d = i; 1609 ierr = PetscInfo1(A,"Matrix is missing diagonal number %D\n",i);CHKERRQ(ierr); 1610 break; 1611 } 1612 } 1613 } 1614 PetscFunctionReturn(0); 1615 } 1616 1617 #include <petscblaslapack.h> 1618 #include <petsc/private/kernels/blockinvert.h> 1619 1620 /* 1621 Note that values is allocated externally by the PC and then passed into this routine 1622 */ 1623 PetscErrorCode MatInvertVariableBlockDiagonal_SeqAIJ(Mat A,PetscInt nblocks,const PetscInt *bsizes,PetscScalar *diag) 1624 { 1625 PetscErrorCode ierr; 1626 PetscInt n = A->rmap->n, i, ncnt = 0, *indx,j,bsizemax = 0,*v_pivots; 1627 PetscBool allowzeropivot,zeropivotdetected=PETSC_FALSE; 1628 const PetscReal shift = 0.0; 1629 PetscInt ipvt[5]; 1630 PetscScalar work[25],*v_work; 1631 1632 PetscFunctionBegin; 1633 allowzeropivot = PetscNot(A->erroriffailure); 1634 for (i=0; i<nblocks; i++) ncnt += bsizes[i]; 1635 if (ncnt != n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Total blocksizes %D doesn't match number matrix rows %D",ncnt,n); 1636 for (i=0; i<nblocks; i++) { 1637 bsizemax = PetscMax(bsizemax,bsizes[i]); 1638 } 1639 ierr = PetscMalloc1(bsizemax,&indx);CHKERRQ(ierr); 1640 if (bsizemax > 7) { 1641 ierr = PetscMalloc2(bsizemax,&v_work,bsizemax,&v_pivots);CHKERRQ(ierr); 1642 } 1643 ncnt = 0; 1644 for (i=0; i<nblocks; i++) { 1645 for (j=0; j<bsizes[i]; j++) indx[j] = ncnt+j; 1646 ierr = MatGetValues(A,bsizes[i],indx,bsizes[i],indx,diag);CHKERRQ(ierr); 1647 switch (bsizes[i]) { 1648 case 1: 1649 *diag = 1.0/(*diag); 1650 break; 1651 case 2: 1652 ierr = PetscKernel_A_gets_inverse_A_2(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1653 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1654 ierr = PetscKernel_A_gets_transpose_A_2(diag);CHKERRQ(ierr); 1655 break; 1656 case 3: 1657 ierr = PetscKernel_A_gets_inverse_A_3(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1658 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1659 ierr = PetscKernel_A_gets_transpose_A_3(diag);CHKERRQ(ierr); 1660 break; 1661 case 4: 1662 ierr = PetscKernel_A_gets_inverse_A_4(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1663 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1664 ierr = PetscKernel_A_gets_transpose_A_4(diag);CHKERRQ(ierr); 1665 break; 1666 case 5: 1667 ierr = PetscKernel_A_gets_inverse_A_5(diag,ipvt,work,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1668 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1669 ierr = PetscKernel_A_gets_transpose_A_5(diag);CHKERRQ(ierr); 1670 break; 1671 case 6: 1672 ierr = PetscKernel_A_gets_inverse_A_6(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1673 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1674 ierr = PetscKernel_A_gets_transpose_A_6(diag);CHKERRQ(ierr); 1675 break; 1676 case 7: 1677 ierr = PetscKernel_A_gets_inverse_A_7(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1678 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1679 ierr = PetscKernel_A_gets_transpose_A_7(diag);CHKERRQ(ierr); 1680 break; 1681 default: 1682 ierr = PetscKernel_A_gets_inverse_A(bsizes[i],diag,v_pivots,v_work,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 1683 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1684 ierr = PetscKernel_A_gets_transpose_A_N(diag,bsizes[i]);CHKERRQ(ierr); 1685 } 1686 ncnt += bsizes[i]; 1687 diag += bsizes[i]*bsizes[i]; 1688 } 1689 if (bsizemax > 7) { 1690 ierr = PetscFree2(v_work,v_pivots);CHKERRQ(ierr); 1691 } 1692 ierr = PetscFree(indx);CHKERRQ(ierr); 1693 PetscFunctionReturn(0); 1694 } 1695 1696 /* 1697 Negative shift indicates do not generate an error if there is a zero diagonal, just invert it anyways 1698 */ 1699 PetscErrorCode MatInvertDiagonal_SeqAIJ(Mat A,PetscScalar omega,PetscScalar fshift) 1700 { 1701 Mat_SeqAIJ *a = (Mat_SeqAIJ*) A->data; 1702 PetscErrorCode ierr; 1703 PetscInt i,*diag,m = A->rmap->n; 1704 MatScalar *v = a->a; 1705 PetscScalar *idiag,*mdiag; 1706 1707 PetscFunctionBegin; 1708 if (a->idiagvalid) PetscFunctionReturn(0); 1709 ierr = MatMarkDiagonal_SeqAIJ(A);CHKERRQ(ierr); 1710 diag = a->diag; 1711 if (!a->idiag) { 1712 ierr = PetscMalloc3(m,&a->idiag,m,&a->mdiag,m,&a->ssor_work);CHKERRQ(ierr); 1713 ierr = PetscLogObjectMemory((PetscObject)A, 3*m*sizeof(PetscScalar));CHKERRQ(ierr); 1714 v = a->a; 1715 } 1716 mdiag = a->mdiag; 1717 idiag = a->idiag; 1718 1719 if (omega == 1.0 && PetscRealPart(fshift) <= 0.0) { 1720 for (i=0; i<m; i++) { 1721 mdiag[i] = v[diag[i]]; 1722 if (!PetscAbsScalar(mdiag[i])) { /* zero diagonal */ 1723 if (PetscRealPart(fshift)) { 1724 ierr = PetscInfo1(A,"Zero diagonal on row %D\n",i);CHKERRQ(ierr); 1725 A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 1726 A->factorerror_zeropivot_value = 0.0; 1727 A->factorerror_zeropivot_row = i; 1728 } else SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Zero diagonal on row %D",i); 1729 } 1730 idiag[i] = 1.0/v[diag[i]]; 1731 } 1732 ierr = PetscLogFlops(m);CHKERRQ(ierr); 1733 } else { 1734 for (i=0; i<m; i++) { 1735 mdiag[i] = v[diag[i]]; 1736 idiag[i] = omega/(fshift + v[diag[i]]); 1737 } 1738 ierr = PetscLogFlops(2.0*m);CHKERRQ(ierr); 1739 } 1740 a->idiagvalid = PETSC_TRUE; 1741 PetscFunctionReturn(0); 1742 } 1743 1744 #include <../src/mat/impls/aij/seq/ftn-kernels/frelax.h> 1745 PetscErrorCode MatSOR_SeqAIJ(Mat A,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx) 1746 { 1747 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1748 PetscScalar *x,d,sum,*t,scale; 1749 const MatScalar *v,*idiag=0,*mdiag; 1750 const PetscScalar *b, *bs,*xb, *ts; 1751 PetscErrorCode ierr; 1752 PetscInt n,m = A->rmap->n,i; 1753 const PetscInt *idx,*diag; 1754 1755 PetscFunctionBegin; 1756 its = its*lits; 1757 1758 if (fshift != a->fshift || omega != a->omega) a->idiagvalid = PETSC_FALSE; /* must recompute idiag[] */ 1759 if (!a->idiagvalid) {ierr = MatInvertDiagonal_SeqAIJ(A,omega,fshift);CHKERRQ(ierr);} 1760 a->fshift = fshift; 1761 a->omega = omega; 1762 1763 diag = a->diag; 1764 t = a->ssor_work; 1765 idiag = a->idiag; 1766 mdiag = a->mdiag; 1767 1768 ierr = VecGetArray(xx,&x);CHKERRQ(ierr); 1769 ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr); 1770 /* We count flops by assuming the upper triangular and lower triangular parts have the same number of nonzeros */ 1771 if (flag == SOR_APPLY_UPPER) { 1772 /* apply (U + D/omega) to the vector */ 1773 bs = b; 1774 for (i=0; i<m; i++) { 1775 d = fshift + mdiag[i]; 1776 n = a->i[i+1] - diag[i] - 1; 1777 idx = a->j + diag[i] + 1; 1778 v = a->a + diag[i] + 1; 1779 sum = b[i]*d/omega; 1780 PetscSparseDensePlusDot(sum,bs,v,idx,n); 1781 x[i] = sum; 1782 } 1783 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 1784 ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr); 1785 ierr = PetscLogFlops(a->nz);CHKERRQ(ierr); 1786 PetscFunctionReturn(0); 1787 } 1788 1789 if (flag == SOR_APPLY_LOWER) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"SOR_APPLY_LOWER is not implemented"); 1790 else if (flag & SOR_EISENSTAT) { 1791 /* Let A = L + U + D; where L is lower trianglar, 1792 U is upper triangular, E = D/omega; This routine applies 1793 1794 (L + E)^{-1} A (U + E)^{-1} 1795 1796 to a vector efficiently using Eisenstat's trick. 1797 */ 1798 scale = (2.0/omega) - 1.0; 1799 1800 /* x = (E + U)^{-1} b */ 1801 for (i=m-1; i>=0; i--) { 1802 n = a->i[i+1] - diag[i] - 1; 1803 idx = a->j + diag[i] + 1; 1804 v = a->a + diag[i] + 1; 1805 sum = b[i]; 1806 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1807 x[i] = sum*idiag[i]; 1808 } 1809 1810 /* t = b - (2*E - D)x */ 1811 v = a->a; 1812 for (i=0; i<m; i++) t[i] = b[i] - scale*(v[*diag++])*x[i]; 1813 1814 /* t = (E + L)^{-1}t */ 1815 ts = t; 1816 diag = a->diag; 1817 for (i=0; i<m; i++) { 1818 n = diag[i] - a->i[i]; 1819 idx = a->j + a->i[i]; 1820 v = a->a + a->i[i]; 1821 sum = t[i]; 1822 PetscSparseDenseMinusDot(sum,ts,v,idx,n); 1823 t[i] = sum*idiag[i]; 1824 /* x = x + t */ 1825 x[i] += t[i]; 1826 } 1827 1828 ierr = PetscLogFlops(6.0*m-1 + 2.0*a->nz);CHKERRQ(ierr); 1829 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 1830 ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr); 1831 PetscFunctionReturn(0); 1832 } 1833 if (flag & SOR_ZERO_INITIAL_GUESS) { 1834 if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP) { 1835 for (i=0; i<m; i++) { 1836 n = diag[i] - a->i[i]; 1837 idx = a->j + a->i[i]; 1838 v = a->a + a->i[i]; 1839 sum = b[i]; 1840 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1841 t[i] = sum; 1842 x[i] = sum*idiag[i]; 1843 } 1844 xb = t; 1845 ierr = PetscLogFlops(a->nz);CHKERRQ(ierr); 1846 } else xb = b; 1847 if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP) { 1848 for (i=m-1; i>=0; i--) { 1849 n = a->i[i+1] - diag[i] - 1; 1850 idx = a->j + diag[i] + 1; 1851 v = a->a + diag[i] + 1; 1852 sum = xb[i]; 1853 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1854 if (xb == b) { 1855 x[i] = sum*idiag[i]; 1856 } else { 1857 x[i] = (1-omega)*x[i] + sum*idiag[i]; /* omega in idiag */ 1858 } 1859 } 1860 ierr = PetscLogFlops(a->nz);CHKERRQ(ierr); /* assumes 1/2 in upper */ 1861 } 1862 its--; 1863 } 1864 while (its--) { 1865 if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP) { 1866 for (i=0; i<m; i++) { 1867 /* lower */ 1868 n = diag[i] - a->i[i]; 1869 idx = a->j + a->i[i]; 1870 v = a->a + a->i[i]; 1871 sum = b[i]; 1872 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1873 t[i] = sum; /* save application of the lower-triangular part */ 1874 /* upper */ 1875 n = a->i[i+1] - diag[i] - 1; 1876 idx = a->j + diag[i] + 1; 1877 v = a->a + diag[i] + 1; 1878 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1879 x[i] = (1. - omega)*x[i] + sum*idiag[i]; /* omega in idiag */ 1880 } 1881 xb = t; 1882 ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr); 1883 } else xb = b; 1884 if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP) { 1885 for (i=m-1; i>=0; i--) { 1886 sum = xb[i]; 1887 if (xb == b) { 1888 /* whole matrix (no checkpointing available) */ 1889 n = a->i[i+1] - a->i[i]; 1890 idx = a->j + a->i[i]; 1891 v = a->a + a->i[i]; 1892 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1893 x[i] = (1. - omega)*x[i] + (sum + mdiag[i]*x[i])*idiag[i]; 1894 } else { /* lower-triangular part has been saved, so only apply upper-triangular */ 1895 n = a->i[i+1] - diag[i] - 1; 1896 idx = a->j + diag[i] + 1; 1897 v = a->a + diag[i] + 1; 1898 PetscSparseDenseMinusDot(sum,x,v,idx,n); 1899 x[i] = (1. - omega)*x[i] + sum*idiag[i]; /* omega in idiag */ 1900 } 1901 } 1902 if (xb == b) { 1903 ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr); 1904 } else { 1905 ierr = PetscLogFlops(a->nz);CHKERRQ(ierr); /* assumes 1/2 in upper */ 1906 } 1907 } 1908 } 1909 ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr); 1910 ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr); 1911 PetscFunctionReturn(0); 1912 } 1913 1914 1915 PetscErrorCode MatGetInfo_SeqAIJ(Mat A,MatInfoType flag,MatInfo *info) 1916 { 1917 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1918 1919 PetscFunctionBegin; 1920 info->block_size = 1.0; 1921 info->nz_allocated = (double)a->maxnz; 1922 info->nz_used = (double)a->nz; 1923 info->nz_unneeded = (double)(a->maxnz - a->nz); 1924 info->assemblies = (double)A->num_ass; 1925 info->mallocs = (double)A->info.mallocs; 1926 info->memory = ((PetscObject)A)->mem; 1927 if (A->factortype) { 1928 info->fill_ratio_given = A->info.fill_ratio_given; 1929 info->fill_ratio_needed = A->info.fill_ratio_needed; 1930 info->factor_mallocs = A->info.factor_mallocs; 1931 } else { 1932 info->fill_ratio_given = 0; 1933 info->fill_ratio_needed = 0; 1934 info->factor_mallocs = 0; 1935 } 1936 PetscFunctionReturn(0); 1937 } 1938 1939 PetscErrorCode MatZeroRows_SeqAIJ(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 1940 { 1941 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 1942 PetscInt i,m = A->rmap->n - 1; 1943 PetscErrorCode ierr; 1944 const PetscScalar *xx; 1945 PetscScalar *bb; 1946 PetscInt d = 0; 1947 1948 PetscFunctionBegin; 1949 if (x && b) { 1950 ierr = VecGetArrayRead(x,&xx);CHKERRQ(ierr); 1951 ierr = VecGetArray(b,&bb);CHKERRQ(ierr); 1952 for (i=0; i<N; i++) { 1953 if (rows[i] < 0 || rows[i] > m) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range", rows[i]); 1954 if (rows[i] >= A->cmap->n) continue; 1955 bb[rows[i]] = diag*xx[rows[i]]; 1956 } 1957 ierr = VecRestoreArrayRead(x,&xx);CHKERRQ(ierr); 1958 ierr = VecRestoreArray(b,&bb);CHKERRQ(ierr); 1959 } 1960 1961 if (a->keepnonzeropattern) { 1962 for (i=0; i<N; i++) { 1963 if (rows[i] < 0 || rows[i] > m) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range", rows[i]); 1964 ierr = PetscMemzero(&a->a[a->i[rows[i]]],a->ilen[rows[i]]*sizeof(PetscScalar));CHKERRQ(ierr); 1965 } 1966 if (diag != 0.0) { 1967 for (i=0; i<N; i++) { 1968 d = rows[i]; 1969 if (rows[i] >= A->cmap->n) continue; 1970 if (a->diag[d] >= a->i[d+1]) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry in the zeroed row %D",d); 1971 } 1972 for (i=0; i<N; i++) { 1973 if (rows[i] >= A->cmap->n) continue; 1974 a->a[a->diag[rows[i]]] = diag; 1975 } 1976 } 1977 } else { 1978 if (diag != 0.0) { 1979 for (i=0; i<N; i++) { 1980 if (rows[i] < 0 || rows[i] > m) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range", rows[i]); 1981 if (a->ilen[rows[i]] > 0) { 1982 if (rows[i] >= A->cmap->n) { 1983 a->ilen[rows[i]] = 0; 1984 } else { 1985 a->ilen[rows[i]] = 1; 1986 a->a[a->i[rows[i]]] = diag; 1987 a->j[a->i[rows[i]]] = rows[i]; 1988 } 1989 } else if (rows[i] < A->cmap->n) { /* in case row was completely empty */ 1990 ierr = MatSetValues_SeqAIJ(A,1,&rows[i],1,&rows[i],&diag,INSERT_VALUES);CHKERRQ(ierr); 1991 } 1992 } 1993 } else { 1994 for (i=0; i<N; i++) { 1995 if (rows[i] < 0 || rows[i] > m) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range", rows[i]); 1996 a->ilen[rows[i]] = 0; 1997 } 1998 } 1999 A->nonzerostate++; 2000 } 2001 ierr = (*A->ops->assemblyend)(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2002 PetscFunctionReturn(0); 2003 } 2004 2005 PetscErrorCode MatZeroRowsColumns_SeqAIJ(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag,Vec x,Vec b) 2006 { 2007 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2008 PetscInt i,j,m = A->rmap->n - 1,d = 0; 2009 PetscErrorCode ierr; 2010 PetscBool missing,*zeroed,vecs = PETSC_FALSE; 2011 const PetscScalar *xx; 2012 PetscScalar *bb; 2013 2014 PetscFunctionBegin; 2015 if (x && b) { 2016 ierr = VecGetArrayRead(x,&xx);CHKERRQ(ierr); 2017 ierr = VecGetArray(b,&bb);CHKERRQ(ierr); 2018 vecs = PETSC_TRUE; 2019 } 2020 ierr = PetscCalloc1(A->rmap->n,&zeroed);CHKERRQ(ierr); 2021 for (i=0; i<N; i++) { 2022 if (rows[i] < 0 || rows[i] > m) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range", rows[i]); 2023 ierr = PetscMemzero(&a->a[a->i[rows[i]]],a->ilen[rows[i]]*sizeof(PetscScalar));CHKERRQ(ierr); 2024 2025 zeroed[rows[i]] = PETSC_TRUE; 2026 } 2027 for (i=0; i<A->rmap->n; i++) { 2028 if (!zeroed[i]) { 2029 for (j=a->i[i]; j<a->i[i+1]; j++) { 2030 if (a->j[j] < A->rmap->n && zeroed[a->j[j]]) { 2031 if (vecs) bb[i] -= a->a[j]*xx[a->j[j]]; 2032 a->a[j] = 0.0; 2033 } 2034 } 2035 } else if (vecs && i < A->cmap->N) bb[i] = diag*xx[i]; 2036 } 2037 if (x && b) { 2038 ierr = VecRestoreArrayRead(x,&xx);CHKERRQ(ierr); 2039 ierr = VecRestoreArray(b,&bb);CHKERRQ(ierr); 2040 } 2041 ierr = PetscFree(zeroed);CHKERRQ(ierr); 2042 if (diag != 0.0) { 2043 ierr = MatMissingDiagonal_SeqAIJ(A,&missing,&d);CHKERRQ(ierr); 2044 if (missing) { 2045 for (i=0; i<N; i++) { 2046 if (rows[i] >= A->cmap->N) continue; 2047 if (a->nonew && rows[i] >= d) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry in row %D (%D)",d,rows[i]); 2048 ierr = MatSetValues_SeqAIJ(A,1,&rows[i],1,&rows[i],&diag,INSERT_VALUES);CHKERRQ(ierr); 2049 } 2050 } else { 2051 for (i=0; i<N; i++) { 2052 a->a[a->diag[rows[i]]] = diag; 2053 } 2054 } 2055 } 2056 ierr = (*A->ops->assemblyend)(A,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2057 PetscFunctionReturn(0); 2058 } 2059 2060 PetscErrorCode MatGetRow_SeqAIJ(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 2061 { 2062 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2063 PetscInt *itmp; 2064 2065 PetscFunctionBegin; 2066 if (row < 0 || row >= A->rmap->n) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row %D out of range",row); 2067 2068 *nz = a->i[row+1] - a->i[row]; 2069 if (v) *v = a->a + a->i[row]; 2070 if (idx) { 2071 itmp = a->j + a->i[row]; 2072 if (*nz) *idx = itmp; 2073 else *idx = 0; 2074 } 2075 PetscFunctionReturn(0); 2076 } 2077 2078 /* remove this function? */ 2079 PetscErrorCode MatRestoreRow_SeqAIJ(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v) 2080 { 2081 PetscFunctionBegin; 2082 PetscFunctionReturn(0); 2083 } 2084 2085 PetscErrorCode MatNorm_SeqAIJ(Mat A,NormType type,PetscReal *nrm) 2086 { 2087 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2088 MatScalar *v = a->a; 2089 PetscReal sum = 0.0; 2090 PetscErrorCode ierr; 2091 PetscInt i,j; 2092 2093 PetscFunctionBegin; 2094 if (type == NORM_FROBENIUS) { 2095 #if defined(PETSC_USE_REAL___FP16) 2096 PetscBLASInt one = 1,nz = a->nz; 2097 *nrm = BLASnrm2_(&nz,v,&one); 2098 #else 2099 for (i=0; i<a->nz; i++) { 2100 sum += PetscRealPart(PetscConj(*v)*(*v)); v++; 2101 } 2102 *nrm = PetscSqrtReal(sum); 2103 #endif 2104 ierr = PetscLogFlops(2*a->nz);CHKERRQ(ierr); 2105 } else if (type == NORM_1) { 2106 PetscReal *tmp; 2107 PetscInt *jj = a->j; 2108 ierr = PetscCalloc1(A->cmap->n+1,&tmp);CHKERRQ(ierr); 2109 *nrm = 0.0; 2110 for (j=0; j<a->nz; j++) { 2111 tmp[*jj++] += PetscAbsScalar(*v); v++; 2112 } 2113 for (j=0; j<A->cmap->n; j++) { 2114 if (tmp[j] > *nrm) *nrm = tmp[j]; 2115 } 2116 ierr = PetscFree(tmp);CHKERRQ(ierr); 2117 ierr = PetscLogFlops(PetscMax(a->nz-1,0));CHKERRQ(ierr); 2118 } else if (type == NORM_INFINITY) { 2119 *nrm = 0.0; 2120 for (j=0; j<A->rmap->n; j++) { 2121 v = a->a + a->i[j]; 2122 sum = 0.0; 2123 for (i=0; i<a->i[j+1]-a->i[j]; i++) { 2124 sum += PetscAbsScalar(*v); v++; 2125 } 2126 if (sum > *nrm) *nrm = sum; 2127 } 2128 ierr = PetscLogFlops(PetscMax(a->nz-1,0));CHKERRQ(ierr); 2129 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support for two norm"); 2130 PetscFunctionReturn(0); 2131 } 2132 2133 /* Merged from MatGetSymbolicTranspose_SeqAIJ() - replace MatGetSymbolicTranspose_SeqAIJ()? */ 2134 PetscErrorCode MatTransposeSymbolic_SeqAIJ(Mat A,Mat *B) 2135 { 2136 PetscErrorCode ierr; 2137 PetscInt i,j,anzj; 2138 Mat_SeqAIJ *a=(Mat_SeqAIJ*)A->data,*b; 2139 PetscInt an=A->cmap->N,am=A->rmap->N; 2140 PetscInt *ati,*atj,*atfill,*ai=a->i,*aj=a->j; 2141 2142 PetscFunctionBegin; 2143 /* Allocate space for symbolic transpose info and work array */ 2144 ierr = PetscCalloc1(an+1,&ati);CHKERRQ(ierr); 2145 ierr = PetscMalloc1(ai[am],&atj);CHKERRQ(ierr); 2146 ierr = PetscMalloc1(an,&atfill);CHKERRQ(ierr); 2147 2148 /* Walk through aj and count ## of non-zeros in each row of A^T. */ 2149 /* Note: offset by 1 for fast conversion into csr format. */ 2150 for (i=0;i<ai[am];i++) ati[aj[i]+1] += 1; 2151 /* Form ati for csr format of A^T. */ 2152 for (i=0;i<an;i++) ati[i+1] += ati[i]; 2153 2154 /* Copy ati into atfill so we have locations of the next free space in atj */ 2155 ierr = PetscMemcpy(atfill,ati,an*sizeof(PetscInt));CHKERRQ(ierr); 2156 2157 /* Walk through A row-wise and mark nonzero entries of A^T. */ 2158 for (i=0;i<am;i++) { 2159 anzj = ai[i+1] - ai[i]; 2160 for (j=0;j<anzj;j++) { 2161 atj[atfill[*aj]] = i; 2162 atfill[*aj++] += 1; 2163 } 2164 } 2165 2166 /* Clean up temporary space and complete requests. */ 2167 ierr = PetscFree(atfill);CHKERRQ(ierr); 2168 ierr = MatCreateSeqAIJWithArrays(PetscObjectComm((PetscObject)A),an,am,ati,atj,NULL,B);CHKERRQ(ierr); 2169 ierr = MatSetBlockSizes(*B,PetscAbs(A->cmap->bs),PetscAbs(A->rmap->bs));CHKERRQ(ierr); 2170 2171 b = (Mat_SeqAIJ*)((*B)->data); 2172 b->free_a = PETSC_FALSE; 2173 b->free_ij = PETSC_TRUE; 2174 b->nonew = 0; 2175 PetscFunctionReturn(0); 2176 } 2177 2178 PetscErrorCode MatTranspose_SeqAIJ(Mat A,MatReuse reuse,Mat *B) 2179 { 2180 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2181 Mat C; 2182 PetscErrorCode ierr; 2183 PetscInt i,*aj = a->j,*ai = a->i,m = A->rmap->n,len,*col; 2184 MatScalar *array = a->a; 2185 2186 PetscFunctionBegin; 2187 if (reuse == MAT_INITIAL_MATRIX || reuse == MAT_INPLACE_MATRIX) { 2188 ierr = PetscCalloc1(1+A->cmap->n,&col);CHKERRQ(ierr); 2189 2190 for (i=0; i<ai[m]; i++) col[aj[i]] += 1; 2191 ierr = MatCreate(PetscObjectComm((PetscObject)A),&C);CHKERRQ(ierr); 2192 ierr = MatSetSizes(C,A->cmap->n,m,A->cmap->n,m);CHKERRQ(ierr); 2193 ierr = MatSetBlockSizes(C,PetscAbs(A->cmap->bs),PetscAbs(A->rmap->bs));CHKERRQ(ierr); 2194 ierr = MatSetType(C,((PetscObject)A)->type_name);CHKERRQ(ierr); 2195 ierr = MatSeqAIJSetPreallocation_SeqAIJ(C,0,col);CHKERRQ(ierr); 2196 ierr = PetscFree(col);CHKERRQ(ierr); 2197 } else { 2198 C = *B; 2199 } 2200 2201 for (i=0; i<m; i++) { 2202 len = ai[i+1]-ai[i]; 2203 ierr = MatSetValues_SeqAIJ(C,len,aj,1,&i,array,INSERT_VALUES);CHKERRQ(ierr); 2204 array += len; 2205 aj += len; 2206 } 2207 ierr = MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2208 ierr = MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2209 2210 if (reuse == MAT_INITIAL_MATRIX || reuse == MAT_REUSE_MATRIX) { 2211 *B = C; 2212 } else { 2213 ierr = MatHeaderMerge(A,&C);CHKERRQ(ierr); 2214 } 2215 PetscFunctionReturn(0); 2216 } 2217 2218 PetscErrorCode MatIsTranspose_SeqAIJ(Mat A,Mat B,PetscReal tol,PetscBool *f) 2219 { 2220 Mat_SeqAIJ *aij = (Mat_SeqAIJ*) A->data,*bij = (Mat_SeqAIJ*) B->data; 2221 PetscInt *adx,*bdx,*aii,*bii,*aptr,*bptr; 2222 MatScalar *va,*vb; 2223 PetscErrorCode ierr; 2224 PetscInt ma,na,mb,nb, i; 2225 2226 PetscFunctionBegin; 2227 ierr = MatGetSize(A,&ma,&na);CHKERRQ(ierr); 2228 ierr = MatGetSize(B,&mb,&nb);CHKERRQ(ierr); 2229 if (ma!=nb || na!=mb) { 2230 *f = PETSC_FALSE; 2231 PetscFunctionReturn(0); 2232 } 2233 aii = aij->i; bii = bij->i; 2234 adx = aij->j; bdx = bij->j; 2235 va = aij->a; vb = bij->a; 2236 ierr = PetscMalloc1(ma,&aptr);CHKERRQ(ierr); 2237 ierr = PetscMalloc1(mb,&bptr);CHKERRQ(ierr); 2238 for (i=0; i<ma; i++) aptr[i] = aii[i]; 2239 for (i=0; i<mb; i++) bptr[i] = bii[i]; 2240 2241 *f = PETSC_TRUE; 2242 for (i=0; i<ma; i++) { 2243 while (aptr[i]<aii[i+1]) { 2244 PetscInt idc,idr; 2245 PetscScalar vc,vr; 2246 /* column/row index/value */ 2247 idc = adx[aptr[i]]; 2248 idr = bdx[bptr[idc]]; 2249 vc = va[aptr[i]]; 2250 vr = vb[bptr[idc]]; 2251 if (i!=idr || PetscAbsScalar(vc-vr) > tol) { 2252 *f = PETSC_FALSE; 2253 goto done; 2254 } else { 2255 aptr[i]++; 2256 if (B || i!=idc) bptr[idc]++; 2257 } 2258 } 2259 } 2260 done: 2261 ierr = PetscFree(aptr);CHKERRQ(ierr); 2262 ierr = PetscFree(bptr);CHKERRQ(ierr); 2263 PetscFunctionReturn(0); 2264 } 2265 2266 PetscErrorCode MatIsHermitianTranspose_SeqAIJ(Mat A,Mat B,PetscReal tol,PetscBool *f) 2267 { 2268 Mat_SeqAIJ *aij = (Mat_SeqAIJ*) A->data,*bij = (Mat_SeqAIJ*) B->data; 2269 PetscInt *adx,*bdx,*aii,*bii,*aptr,*bptr; 2270 MatScalar *va,*vb; 2271 PetscErrorCode ierr; 2272 PetscInt ma,na,mb,nb, i; 2273 2274 PetscFunctionBegin; 2275 ierr = MatGetSize(A,&ma,&na);CHKERRQ(ierr); 2276 ierr = MatGetSize(B,&mb,&nb);CHKERRQ(ierr); 2277 if (ma!=nb || na!=mb) { 2278 *f = PETSC_FALSE; 2279 PetscFunctionReturn(0); 2280 } 2281 aii = aij->i; bii = bij->i; 2282 adx = aij->j; bdx = bij->j; 2283 va = aij->a; vb = bij->a; 2284 ierr = PetscMalloc1(ma,&aptr);CHKERRQ(ierr); 2285 ierr = PetscMalloc1(mb,&bptr);CHKERRQ(ierr); 2286 for (i=0; i<ma; i++) aptr[i] = aii[i]; 2287 for (i=0; i<mb; i++) bptr[i] = bii[i]; 2288 2289 *f = PETSC_TRUE; 2290 for (i=0; i<ma; i++) { 2291 while (aptr[i]<aii[i+1]) { 2292 PetscInt idc,idr; 2293 PetscScalar vc,vr; 2294 /* column/row index/value */ 2295 idc = adx[aptr[i]]; 2296 idr = bdx[bptr[idc]]; 2297 vc = va[aptr[i]]; 2298 vr = vb[bptr[idc]]; 2299 if (i!=idr || PetscAbsScalar(vc-PetscConj(vr)) > tol) { 2300 *f = PETSC_FALSE; 2301 goto done; 2302 } else { 2303 aptr[i]++; 2304 if (B || i!=idc) bptr[idc]++; 2305 } 2306 } 2307 } 2308 done: 2309 ierr = PetscFree(aptr);CHKERRQ(ierr); 2310 ierr = PetscFree(bptr);CHKERRQ(ierr); 2311 PetscFunctionReturn(0); 2312 } 2313 2314 PetscErrorCode MatIsSymmetric_SeqAIJ(Mat A,PetscReal tol,PetscBool *f) 2315 { 2316 PetscErrorCode ierr; 2317 2318 PetscFunctionBegin; 2319 ierr = MatIsTranspose_SeqAIJ(A,A,tol,f);CHKERRQ(ierr); 2320 PetscFunctionReturn(0); 2321 } 2322 2323 PetscErrorCode MatIsHermitian_SeqAIJ(Mat A,PetscReal tol,PetscBool *f) 2324 { 2325 PetscErrorCode ierr; 2326 2327 PetscFunctionBegin; 2328 ierr = MatIsHermitianTranspose_SeqAIJ(A,A,tol,f);CHKERRQ(ierr); 2329 PetscFunctionReturn(0); 2330 } 2331 2332 PetscErrorCode MatDiagonalScale_SeqAIJ(Mat A,Vec ll,Vec rr) 2333 { 2334 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2335 const PetscScalar *l,*r; 2336 PetscScalar x; 2337 MatScalar *v; 2338 PetscErrorCode ierr; 2339 PetscInt i,j,m = A->rmap->n,n = A->cmap->n,M,nz = a->nz; 2340 const PetscInt *jj; 2341 2342 PetscFunctionBegin; 2343 if (ll) { 2344 /* The local size is used so that VecMPI can be passed to this routine 2345 by MatDiagonalScale_MPIAIJ */ 2346 ierr = VecGetLocalSize(ll,&m);CHKERRQ(ierr); 2347 if (m != A->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Left scaling vector wrong length"); 2348 ierr = VecGetArrayRead(ll,&l);CHKERRQ(ierr); 2349 v = a->a; 2350 for (i=0; i<m; i++) { 2351 x = l[i]; 2352 M = a->i[i+1] - a->i[i]; 2353 for (j=0; j<M; j++) (*v++) *= x; 2354 } 2355 ierr = VecRestoreArrayRead(ll,&l);CHKERRQ(ierr); 2356 ierr = PetscLogFlops(nz);CHKERRQ(ierr); 2357 } 2358 if (rr) { 2359 ierr = VecGetLocalSize(rr,&n);CHKERRQ(ierr); 2360 if (n != A->cmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Right scaling vector wrong length"); 2361 ierr = VecGetArrayRead(rr,&r);CHKERRQ(ierr); 2362 v = a->a; jj = a->j; 2363 for (i=0; i<nz; i++) (*v++) *= r[*jj++]; 2364 ierr = VecRestoreArrayRead(rr,&r);CHKERRQ(ierr); 2365 ierr = PetscLogFlops(nz);CHKERRQ(ierr); 2366 } 2367 ierr = MatSeqAIJInvalidateDiagonal(A);CHKERRQ(ierr); 2368 PetscFunctionReturn(0); 2369 } 2370 2371 PetscErrorCode MatCreateSubMatrix_SeqAIJ(Mat A,IS isrow,IS iscol,PetscInt csize,MatReuse scall,Mat *B) 2372 { 2373 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*c; 2374 PetscErrorCode ierr; 2375 PetscInt *smap,i,k,kstart,kend,oldcols = A->cmap->n,*lens; 2376 PetscInt row,mat_i,*mat_j,tcol,first,step,*mat_ilen,sum,lensi; 2377 const PetscInt *irow,*icol; 2378 PetscInt nrows,ncols; 2379 PetscInt *starts,*j_new,*i_new,*aj = a->j,*ai = a->i,ii,*ailen = a->ilen; 2380 MatScalar *a_new,*mat_a; 2381 Mat C; 2382 PetscBool stride; 2383 2384 PetscFunctionBegin; 2385 2386 ierr = ISGetIndices(isrow,&irow);CHKERRQ(ierr); 2387 ierr = ISGetLocalSize(isrow,&nrows);CHKERRQ(ierr); 2388 ierr = ISGetLocalSize(iscol,&ncols);CHKERRQ(ierr); 2389 2390 ierr = PetscObjectTypeCompare((PetscObject)iscol,ISSTRIDE,&stride);CHKERRQ(ierr); 2391 if (stride) { 2392 ierr = ISStrideGetInfo(iscol,&first,&step);CHKERRQ(ierr); 2393 } else { 2394 first = 0; 2395 step = 0; 2396 } 2397 if (stride && step == 1) { 2398 /* special case of contiguous rows */ 2399 ierr = PetscMalloc2(nrows,&lens,nrows,&starts);CHKERRQ(ierr); 2400 /* loop over new rows determining lens and starting points */ 2401 for (i=0; i<nrows; i++) { 2402 kstart = ai[irow[i]]; 2403 kend = kstart + ailen[irow[i]]; 2404 starts[i] = kstart; 2405 for (k=kstart; k<kend; k++) { 2406 if (aj[k] >= first) { 2407 starts[i] = k; 2408 break; 2409 } 2410 } 2411 sum = 0; 2412 while (k < kend) { 2413 if (aj[k++] >= first+ncols) break; 2414 sum++; 2415 } 2416 lens[i] = sum; 2417 } 2418 /* create submatrix */ 2419 if (scall == MAT_REUSE_MATRIX) { 2420 PetscInt n_cols,n_rows; 2421 ierr = MatGetSize(*B,&n_rows,&n_cols);CHKERRQ(ierr); 2422 if (n_rows != nrows || n_cols != ncols) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Reused submatrix wrong size"); 2423 ierr = MatZeroEntries(*B);CHKERRQ(ierr); 2424 C = *B; 2425 } else { 2426 PetscInt rbs,cbs; 2427 ierr = MatCreate(PetscObjectComm((PetscObject)A),&C);CHKERRQ(ierr); 2428 ierr = MatSetSizes(C,nrows,ncols,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 2429 ierr = ISGetBlockSize(isrow,&rbs);CHKERRQ(ierr); 2430 ierr = ISGetBlockSize(iscol,&cbs);CHKERRQ(ierr); 2431 ierr = MatSetBlockSizes(C,rbs,cbs);CHKERRQ(ierr); 2432 ierr = MatSetType(C,((PetscObject)A)->type_name);CHKERRQ(ierr); 2433 ierr = MatSeqAIJSetPreallocation_SeqAIJ(C,0,lens);CHKERRQ(ierr); 2434 } 2435 c = (Mat_SeqAIJ*)C->data; 2436 2437 /* loop over rows inserting into submatrix */ 2438 a_new = c->a; 2439 j_new = c->j; 2440 i_new = c->i; 2441 2442 for (i=0; i<nrows; i++) { 2443 ii = starts[i]; 2444 lensi = lens[i]; 2445 for (k=0; k<lensi; k++) { 2446 *j_new++ = aj[ii+k] - first; 2447 } 2448 ierr = PetscMemcpy(a_new,a->a + starts[i],lensi*sizeof(PetscScalar));CHKERRQ(ierr); 2449 a_new += lensi; 2450 i_new[i+1] = i_new[i] + lensi; 2451 c->ilen[i] = lensi; 2452 } 2453 ierr = PetscFree2(lens,starts);CHKERRQ(ierr); 2454 } else { 2455 ierr = ISGetIndices(iscol,&icol);CHKERRQ(ierr); 2456 ierr = PetscCalloc1(oldcols,&smap);CHKERRQ(ierr); 2457 ierr = PetscMalloc1(1+nrows,&lens);CHKERRQ(ierr); 2458 for (i=0; i<ncols; i++) { 2459 #if defined(PETSC_USE_DEBUG) 2460 if (icol[i] >= oldcols) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Requesting column beyond largest column icol[%D] %D <= A->cmap->n %D",i,icol[i],oldcols); 2461 #endif 2462 smap[icol[i]] = i+1; 2463 } 2464 2465 /* determine lens of each row */ 2466 for (i=0; i<nrows; i++) { 2467 kstart = ai[irow[i]]; 2468 kend = kstart + a->ilen[irow[i]]; 2469 lens[i] = 0; 2470 for (k=kstart; k<kend; k++) { 2471 if (smap[aj[k]]) { 2472 lens[i]++; 2473 } 2474 } 2475 } 2476 /* Create and fill new matrix */ 2477 if (scall == MAT_REUSE_MATRIX) { 2478 PetscBool equal; 2479 2480 c = (Mat_SeqAIJ*)((*B)->data); 2481 if ((*B)->rmap->n != nrows || (*B)->cmap->n != ncols) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong size"); 2482 ierr = PetscMemcmp(c->ilen,lens,(*B)->rmap->n*sizeof(PetscInt),&equal);CHKERRQ(ierr); 2483 if (!equal) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong no of nonzeros"); 2484 ierr = PetscMemzero(c->ilen,(*B)->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 2485 C = *B; 2486 } else { 2487 PetscInt rbs,cbs; 2488 ierr = MatCreate(PetscObjectComm((PetscObject)A),&C);CHKERRQ(ierr); 2489 ierr = MatSetSizes(C,nrows,ncols,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 2490 ierr = ISGetBlockSize(isrow,&rbs);CHKERRQ(ierr); 2491 ierr = ISGetBlockSize(iscol,&cbs);CHKERRQ(ierr); 2492 ierr = MatSetBlockSizes(C,rbs,cbs);CHKERRQ(ierr); 2493 ierr = MatSetType(C,((PetscObject)A)->type_name);CHKERRQ(ierr); 2494 ierr = MatSeqAIJSetPreallocation_SeqAIJ(C,0,lens);CHKERRQ(ierr); 2495 } 2496 c = (Mat_SeqAIJ*)(C->data); 2497 for (i=0; i<nrows; i++) { 2498 row = irow[i]; 2499 kstart = ai[row]; 2500 kend = kstart + a->ilen[row]; 2501 mat_i = c->i[i]; 2502 mat_j = c->j + mat_i; 2503 mat_a = c->a + mat_i; 2504 mat_ilen = c->ilen + i; 2505 for (k=kstart; k<kend; k++) { 2506 if ((tcol=smap[a->j[k]])) { 2507 *mat_j++ = tcol - 1; 2508 *mat_a++ = a->a[k]; 2509 (*mat_ilen)++; 2510 2511 } 2512 } 2513 } 2514 /* Free work space */ 2515 ierr = ISRestoreIndices(iscol,&icol);CHKERRQ(ierr); 2516 ierr = PetscFree(smap);CHKERRQ(ierr); 2517 ierr = PetscFree(lens);CHKERRQ(ierr); 2518 /* sort */ 2519 for (i = 0; i < nrows; i++) { 2520 PetscInt ilen; 2521 2522 mat_i = c->i[i]; 2523 mat_j = c->j + mat_i; 2524 mat_a = c->a + mat_i; 2525 ilen = c->ilen[i]; 2526 ierr = PetscSortIntWithScalarArray(ilen,mat_j,mat_a);CHKERRQ(ierr); 2527 } 2528 } 2529 ierr = MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2530 ierr = MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2531 2532 ierr = ISRestoreIndices(isrow,&irow);CHKERRQ(ierr); 2533 *B = C; 2534 PetscFunctionReturn(0); 2535 } 2536 2537 PetscErrorCode MatGetMultiProcBlock_SeqAIJ(Mat mat,MPI_Comm subComm,MatReuse scall,Mat *subMat) 2538 { 2539 PetscErrorCode ierr; 2540 Mat B; 2541 2542 PetscFunctionBegin; 2543 if (scall == MAT_INITIAL_MATRIX) { 2544 ierr = MatCreate(subComm,&B);CHKERRQ(ierr); 2545 ierr = MatSetSizes(B,mat->rmap->n,mat->cmap->n,mat->rmap->n,mat->cmap->n);CHKERRQ(ierr); 2546 ierr = MatSetBlockSizesFromMats(B,mat,mat);CHKERRQ(ierr); 2547 ierr = MatSetType(B,MATSEQAIJ);CHKERRQ(ierr); 2548 ierr = MatDuplicateNoCreate_SeqAIJ(B,mat,MAT_COPY_VALUES,PETSC_TRUE);CHKERRQ(ierr); 2549 *subMat = B; 2550 } else { 2551 ierr = MatCopy_SeqAIJ(mat,*subMat,SAME_NONZERO_PATTERN);CHKERRQ(ierr); 2552 } 2553 PetscFunctionReturn(0); 2554 } 2555 2556 PetscErrorCode MatILUFactor_SeqAIJ(Mat inA,IS row,IS col,const MatFactorInfo *info) 2557 { 2558 Mat_SeqAIJ *a = (Mat_SeqAIJ*)inA->data; 2559 PetscErrorCode ierr; 2560 Mat outA; 2561 PetscBool row_identity,col_identity; 2562 2563 PetscFunctionBegin; 2564 if (info->levels != 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only levels=0 supported for in-place ilu"); 2565 2566 ierr = ISIdentity(row,&row_identity);CHKERRQ(ierr); 2567 ierr = ISIdentity(col,&col_identity);CHKERRQ(ierr); 2568 2569 outA = inA; 2570 outA->factortype = MAT_FACTOR_LU; 2571 ierr = PetscFree(inA->solvertype);CHKERRQ(ierr); 2572 ierr = PetscStrallocpy(MATSOLVERPETSC,&inA->solvertype);CHKERRQ(ierr); 2573 2574 ierr = PetscObjectReference((PetscObject)row);CHKERRQ(ierr); 2575 ierr = ISDestroy(&a->row);CHKERRQ(ierr); 2576 2577 a->row = row; 2578 2579 ierr = PetscObjectReference((PetscObject)col);CHKERRQ(ierr); 2580 ierr = ISDestroy(&a->col);CHKERRQ(ierr); 2581 2582 a->col = col; 2583 2584 /* Create the inverse permutation so that it can be used in MatLUFactorNumeric() */ 2585 ierr = ISDestroy(&a->icol);CHKERRQ(ierr); 2586 ierr = ISInvertPermutation(col,PETSC_DECIDE,&a->icol);CHKERRQ(ierr); 2587 ierr = PetscLogObjectParent((PetscObject)inA,(PetscObject)a->icol);CHKERRQ(ierr); 2588 2589 if (!a->solve_work) { /* this matrix may have been factored before */ 2590 ierr = PetscMalloc1(inA->rmap->n+1,&a->solve_work);CHKERRQ(ierr); 2591 ierr = PetscLogObjectMemory((PetscObject)inA, (inA->rmap->n+1)*sizeof(PetscScalar));CHKERRQ(ierr); 2592 } 2593 2594 ierr = MatMarkDiagonal_SeqAIJ(inA);CHKERRQ(ierr); 2595 if (row_identity && col_identity) { 2596 ierr = MatLUFactorNumeric_SeqAIJ_inplace(outA,inA,info);CHKERRQ(ierr); 2597 } else { 2598 ierr = MatLUFactorNumeric_SeqAIJ_InplaceWithPerm(outA,inA,info);CHKERRQ(ierr); 2599 } 2600 PetscFunctionReturn(0); 2601 } 2602 2603 PetscErrorCode MatScale_SeqAIJ(Mat inA,PetscScalar alpha) 2604 { 2605 Mat_SeqAIJ *a = (Mat_SeqAIJ*)inA->data; 2606 PetscScalar oalpha = alpha; 2607 PetscErrorCode ierr; 2608 PetscBLASInt one = 1,bnz; 2609 2610 PetscFunctionBegin; 2611 ierr = PetscBLASIntCast(a->nz,&bnz);CHKERRQ(ierr); 2612 PetscStackCallBLAS("BLASscal",BLASscal_(&bnz,&oalpha,a->a,&one)); 2613 ierr = PetscLogFlops(a->nz);CHKERRQ(ierr); 2614 ierr = MatSeqAIJInvalidateDiagonal(inA);CHKERRQ(ierr); 2615 PetscFunctionReturn(0); 2616 } 2617 2618 PetscErrorCode MatDestroySubMatrix_Private(Mat_SubSppt *submatj) 2619 { 2620 PetscErrorCode ierr; 2621 PetscInt i; 2622 2623 PetscFunctionBegin; 2624 if (!submatj->id) { /* delete data that are linked only to submats[id=0] */ 2625 ierr = PetscFree4(submatj->sbuf1,submatj->ptr,submatj->tmp,submatj->ctr);CHKERRQ(ierr); 2626 2627 for (i=0; i<submatj->nrqr; ++i) { 2628 ierr = PetscFree(submatj->sbuf2[i]);CHKERRQ(ierr); 2629 } 2630 ierr = PetscFree3(submatj->sbuf2,submatj->req_size,submatj->req_source1);CHKERRQ(ierr); 2631 2632 if (submatj->rbuf1) { 2633 ierr = PetscFree(submatj->rbuf1[0]);CHKERRQ(ierr); 2634 ierr = PetscFree(submatj->rbuf1);CHKERRQ(ierr); 2635 } 2636 2637 for (i=0; i<submatj->nrqs; ++i) { 2638 ierr = PetscFree(submatj->rbuf3[i]);CHKERRQ(ierr); 2639 } 2640 ierr = PetscFree3(submatj->req_source2,submatj->rbuf2,submatj->rbuf3);CHKERRQ(ierr); 2641 ierr = PetscFree(submatj->pa);CHKERRQ(ierr); 2642 } 2643 2644 #if defined(PETSC_USE_CTABLE) 2645 ierr = PetscTableDestroy((PetscTable*)&submatj->rmap);CHKERRQ(ierr); 2646 if (submatj->cmap_loc) {ierr = PetscFree(submatj->cmap_loc);CHKERRQ(ierr);} 2647 ierr = PetscFree(submatj->rmap_loc);CHKERRQ(ierr); 2648 #else 2649 ierr = PetscFree(submatj->rmap);CHKERRQ(ierr); 2650 #endif 2651 2652 if (!submatj->allcolumns) { 2653 #if defined(PETSC_USE_CTABLE) 2654 ierr = PetscTableDestroy((PetscTable*)&submatj->cmap);CHKERRQ(ierr); 2655 #else 2656 ierr = PetscFree(submatj->cmap);CHKERRQ(ierr); 2657 #endif 2658 } 2659 ierr = PetscFree(submatj->row2proc);CHKERRQ(ierr); 2660 2661 ierr = PetscFree(submatj);CHKERRQ(ierr); 2662 PetscFunctionReturn(0); 2663 } 2664 2665 PetscErrorCode MatDestroySubMatrix_SeqAIJ(Mat C) 2666 { 2667 PetscErrorCode ierr; 2668 Mat_SeqAIJ *c = (Mat_SeqAIJ*)C->data; 2669 Mat_SubSppt *submatj = c->submatis1; 2670 2671 PetscFunctionBegin; 2672 ierr = (*submatj->destroy)(C);CHKERRQ(ierr); 2673 ierr = MatDestroySubMatrix_Private(submatj);CHKERRQ(ierr); 2674 PetscFunctionReturn(0); 2675 } 2676 2677 PetscErrorCode MatDestroySubMatrices_SeqAIJ(PetscInt n,Mat *mat[]) 2678 { 2679 PetscErrorCode ierr; 2680 PetscInt i; 2681 Mat C; 2682 Mat_SeqAIJ *c; 2683 Mat_SubSppt *submatj; 2684 2685 PetscFunctionBegin; 2686 for (i=0; i<n; i++) { 2687 C = (*mat)[i]; 2688 c = (Mat_SeqAIJ*)C->data; 2689 submatj = c->submatis1; 2690 if (submatj) { 2691 if (--((PetscObject)C)->refct <= 0) { 2692 ierr = (*submatj->destroy)(C);CHKERRQ(ierr); 2693 ierr = MatDestroySubMatrix_Private(submatj);CHKERRQ(ierr); 2694 ierr = PetscFree(C->defaultvectype);CHKERRQ(ierr); 2695 ierr = PetscLayoutDestroy(&C->rmap);CHKERRQ(ierr); 2696 ierr = PetscLayoutDestroy(&C->cmap);CHKERRQ(ierr); 2697 ierr = PetscHeaderDestroy(&C);CHKERRQ(ierr); 2698 } 2699 } else { 2700 ierr = MatDestroy(&C);CHKERRQ(ierr); 2701 } 2702 } 2703 2704 /* Destroy Dummy submatrices created for reuse */ 2705 ierr = MatDestroySubMatrices_Dummy(n,mat);CHKERRQ(ierr); 2706 2707 ierr = PetscFree(*mat);CHKERRQ(ierr); 2708 PetscFunctionReturn(0); 2709 } 2710 2711 PetscErrorCode MatCreateSubMatrices_SeqAIJ(Mat A,PetscInt n,const IS irow[],const IS icol[],MatReuse scall,Mat *B[]) 2712 { 2713 PetscErrorCode ierr; 2714 PetscInt i; 2715 2716 PetscFunctionBegin; 2717 if (scall == MAT_INITIAL_MATRIX) { 2718 ierr = PetscCalloc1(n+1,B);CHKERRQ(ierr); 2719 } 2720 2721 for (i=0; i<n; i++) { 2722 ierr = MatCreateSubMatrix_SeqAIJ(A,irow[i],icol[i],PETSC_DECIDE,scall,&(*B)[i]);CHKERRQ(ierr); 2723 } 2724 PetscFunctionReturn(0); 2725 } 2726 2727 PetscErrorCode MatIncreaseOverlap_SeqAIJ(Mat A,PetscInt is_max,IS is[],PetscInt ov) 2728 { 2729 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2730 PetscErrorCode ierr; 2731 PetscInt row,i,j,k,l,m,n,*nidx,isz,val; 2732 const PetscInt *idx; 2733 PetscInt start,end,*ai,*aj; 2734 PetscBT table; 2735 2736 PetscFunctionBegin; 2737 m = A->rmap->n; 2738 ai = a->i; 2739 aj = a->j; 2740 2741 if (ov < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"illegal negative overlap value used"); 2742 2743 ierr = PetscMalloc1(m+1,&nidx);CHKERRQ(ierr); 2744 ierr = PetscBTCreate(m,&table);CHKERRQ(ierr); 2745 2746 for (i=0; i<is_max; i++) { 2747 /* Initialize the two local arrays */ 2748 isz = 0; 2749 ierr = PetscBTMemzero(m,table);CHKERRQ(ierr); 2750 2751 /* Extract the indices, assume there can be duplicate entries */ 2752 ierr = ISGetIndices(is[i],&idx);CHKERRQ(ierr); 2753 ierr = ISGetLocalSize(is[i],&n);CHKERRQ(ierr); 2754 2755 /* Enter these into the temp arrays. I.e., mark table[row], enter row into new index */ 2756 for (j=0; j<n; ++j) { 2757 if (!PetscBTLookupSet(table,idx[j])) nidx[isz++] = idx[j]; 2758 } 2759 ierr = ISRestoreIndices(is[i],&idx);CHKERRQ(ierr); 2760 ierr = ISDestroy(&is[i]);CHKERRQ(ierr); 2761 2762 k = 0; 2763 for (j=0; j<ov; j++) { /* for each overlap */ 2764 n = isz; 2765 for (; k<n; k++) { /* do only those rows in nidx[k], which are not done yet */ 2766 row = nidx[k]; 2767 start = ai[row]; 2768 end = ai[row+1]; 2769 for (l = start; l<end; l++) { 2770 val = aj[l]; 2771 if (!PetscBTLookupSet(table,val)) nidx[isz++] = val; 2772 } 2773 } 2774 } 2775 ierr = ISCreateGeneral(PETSC_COMM_SELF,isz,nidx,PETSC_COPY_VALUES,(is+i));CHKERRQ(ierr); 2776 } 2777 ierr = PetscBTDestroy(&table);CHKERRQ(ierr); 2778 ierr = PetscFree(nidx);CHKERRQ(ierr); 2779 PetscFunctionReturn(0); 2780 } 2781 2782 /* -------------------------------------------------------------- */ 2783 PetscErrorCode MatPermute_SeqAIJ(Mat A,IS rowp,IS colp,Mat *B) 2784 { 2785 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2786 PetscErrorCode ierr; 2787 PetscInt i,nz = 0,m = A->rmap->n,n = A->cmap->n; 2788 const PetscInt *row,*col; 2789 PetscInt *cnew,j,*lens; 2790 IS icolp,irowp; 2791 PetscInt *cwork = NULL; 2792 PetscScalar *vwork = NULL; 2793 2794 PetscFunctionBegin; 2795 ierr = ISInvertPermutation(rowp,PETSC_DECIDE,&irowp);CHKERRQ(ierr); 2796 ierr = ISGetIndices(irowp,&row);CHKERRQ(ierr); 2797 ierr = ISInvertPermutation(colp,PETSC_DECIDE,&icolp);CHKERRQ(ierr); 2798 ierr = ISGetIndices(icolp,&col);CHKERRQ(ierr); 2799 2800 /* determine lengths of permuted rows */ 2801 ierr = PetscMalloc1(m+1,&lens);CHKERRQ(ierr); 2802 for (i=0; i<m; i++) lens[row[i]] = a->i[i+1] - a->i[i]; 2803 ierr = MatCreate(PetscObjectComm((PetscObject)A),B);CHKERRQ(ierr); 2804 ierr = MatSetSizes(*B,m,n,m,n);CHKERRQ(ierr); 2805 ierr = MatSetBlockSizesFromMats(*B,A,A);CHKERRQ(ierr); 2806 ierr = MatSetType(*B,((PetscObject)A)->type_name);CHKERRQ(ierr); 2807 ierr = MatSeqAIJSetPreallocation_SeqAIJ(*B,0,lens);CHKERRQ(ierr); 2808 ierr = PetscFree(lens);CHKERRQ(ierr); 2809 2810 ierr = PetscMalloc1(n,&cnew);CHKERRQ(ierr); 2811 for (i=0; i<m; i++) { 2812 ierr = MatGetRow_SeqAIJ(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 2813 for (j=0; j<nz; j++) cnew[j] = col[cwork[j]]; 2814 ierr = MatSetValues_SeqAIJ(*B,1,&row[i],nz,cnew,vwork,INSERT_VALUES);CHKERRQ(ierr); 2815 ierr = MatRestoreRow_SeqAIJ(A,i,&nz,&cwork,&vwork);CHKERRQ(ierr); 2816 } 2817 ierr = PetscFree(cnew);CHKERRQ(ierr); 2818 2819 (*B)->assembled = PETSC_FALSE; 2820 2821 ierr = MatAssemblyBegin(*B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2822 ierr = MatAssemblyEnd(*B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2823 ierr = ISRestoreIndices(irowp,&row);CHKERRQ(ierr); 2824 ierr = ISRestoreIndices(icolp,&col);CHKERRQ(ierr); 2825 ierr = ISDestroy(&irowp);CHKERRQ(ierr); 2826 ierr = ISDestroy(&icolp);CHKERRQ(ierr); 2827 PetscFunctionReturn(0); 2828 } 2829 2830 PetscErrorCode MatCopy_SeqAIJ(Mat A,Mat B,MatStructure str) 2831 { 2832 PetscErrorCode ierr; 2833 2834 PetscFunctionBegin; 2835 /* If the two matrices have the same copy implementation, use fast copy. */ 2836 if (str == SAME_NONZERO_PATTERN && (A->ops->copy == B->ops->copy)) { 2837 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2838 Mat_SeqAIJ *b = (Mat_SeqAIJ*)B->data; 2839 2840 if (a->i[A->rmap->n] != b->i[B->rmap->n]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Number of nonzeros in two matrices are different"); 2841 ierr = PetscMemcpy(b->a,a->a,(a->i[A->rmap->n])*sizeof(PetscScalar));CHKERRQ(ierr); 2842 ierr = PetscObjectStateIncrease((PetscObject)B);CHKERRQ(ierr); 2843 } else { 2844 ierr = MatCopy_Basic(A,B,str);CHKERRQ(ierr); 2845 } 2846 PetscFunctionReturn(0); 2847 } 2848 2849 PetscErrorCode MatSetUp_SeqAIJ(Mat A) 2850 { 2851 PetscErrorCode ierr; 2852 2853 PetscFunctionBegin; 2854 ierr = MatSeqAIJSetPreallocation_SeqAIJ(A,PETSC_DEFAULT,0);CHKERRQ(ierr); 2855 PetscFunctionReturn(0); 2856 } 2857 2858 PetscErrorCode MatSeqAIJGetArray_SeqAIJ(Mat A,PetscScalar *array[]) 2859 { 2860 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2861 2862 PetscFunctionBegin; 2863 *array = a->a; 2864 PetscFunctionReturn(0); 2865 } 2866 2867 PetscErrorCode MatSeqAIJRestoreArray_SeqAIJ(Mat A,PetscScalar *array[]) 2868 { 2869 PetscFunctionBegin; 2870 PetscFunctionReturn(0); 2871 } 2872 2873 /* 2874 Computes the number of nonzeros per row needed for preallocation when X and Y 2875 have different nonzero structure. 2876 */ 2877 PetscErrorCode MatAXPYGetPreallocation_SeqX_private(PetscInt m,const PetscInt *xi,const PetscInt *xj,const PetscInt *yi,const PetscInt *yj,PetscInt *nnz) 2878 { 2879 PetscInt i,j,k,nzx,nzy; 2880 2881 PetscFunctionBegin; 2882 /* Set the number of nonzeros in the new matrix */ 2883 for (i=0; i<m; i++) { 2884 const PetscInt *xjj = xj+xi[i],*yjj = yj+yi[i]; 2885 nzx = xi[i+1] - xi[i]; 2886 nzy = yi[i+1] - yi[i]; 2887 nnz[i] = 0; 2888 for (j=0,k=0; j<nzx; j++) { /* Point in X */ 2889 for (; k<nzy && yjj[k]<xjj[j]; k++) nnz[i]++; /* Catch up to X */ 2890 if (k<nzy && yjj[k]==xjj[j]) k++; /* Skip duplicate */ 2891 nnz[i]++; 2892 } 2893 for (; k<nzy; k++) nnz[i]++; 2894 } 2895 PetscFunctionReturn(0); 2896 } 2897 2898 PetscErrorCode MatAXPYGetPreallocation_SeqAIJ(Mat Y,Mat X,PetscInt *nnz) 2899 { 2900 PetscInt m = Y->rmap->N; 2901 Mat_SeqAIJ *x = (Mat_SeqAIJ*)X->data; 2902 Mat_SeqAIJ *y = (Mat_SeqAIJ*)Y->data; 2903 PetscErrorCode ierr; 2904 2905 PetscFunctionBegin; 2906 /* Set the number of nonzeros in the new matrix */ 2907 ierr = MatAXPYGetPreallocation_SeqX_private(m,x->i,x->j,y->i,y->j,nnz);CHKERRQ(ierr); 2908 PetscFunctionReturn(0); 2909 } 2910 2911 PetscErrorCode MatAXPY_SeqAIJ(Mat Y,PetscScalar a,Mat X,MatStructure str) 2912 { 2913 PetscErrorCode ierr; 2914 Mat_SeqAIJ *x = (Mat_SeqAIJ*)X->data,*y = (Mat_SeqAIJ*)Y->data; 2915 PetscBLASInt one=1,bnz; 2916 2917 PetscFunctionBegin; 2918 ierr = PetscBLASIntCast(x->nz,&bnz);CHKERRQ(ierr); 2919 if (str == SAME_NONZERO_PATTERN) { 2920 PetscScalar alpha = a; 2921 PetscStackCallBLAS("BLASaxpy",BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one)); 2922 ierr = MatSeqAIJInvalidateDiagonal(Y);CHKERRQ(ierr); 2923 ierr = PetscObjectStateIncrease((PetscObject)Y);CHKERRQ(ierr); 2924 } else if (str == SUBSET_NONZERO_PATTERN) { /* nonzeros of X is a subset of Y's */ 2925 ierr = MatAXPY_Basic(Y,a,X,str);CHKERRQ(ierr); 2926 } else { 2927 Mat B; 2928 PetscInt *nnz; 2929 ierr = PetscMalloc1(Y->rmap->N,&nnz);CHKERRQ(ierr); 2930 ierr = MatCreate(PetscObjectComm((PetscObject)Y),&B);CHKERRQ(ierr); 2931 ierr = PetscObjectSetName((PetscObject)B,((PetscObject)Y)->name);CHKERRQ(ierr); 2932 ierr = MatSetSizes(B,Y->rmap->n,Y->cmap->n,Y->rmap->N,Y->cmap->N);CHKERRQ(ierr); 2933 ierr = MatSetBlockSizesFromMats(B,Y,Y);CHKERRQ(ierr); 2934 ierr = MatSetType(B,(MatType) ((PetscObject)Y)->type_name);CHKERRQ(ierr); 2935 ierr = MatAXPYGetPreallocation_SeqAIJ(Y,X,nnz);CHKERRQ(ierr); 2936 ierr = MatSeqAIJSetPreallocation(B,0,nnz);CHKERRQ(ierr); 2937 ierr = MatAXPY_BasicWithPreallocation(B,Y,a,X,str);CHKERRQ(ierr); 2938 ierr = MatHeaderReplace(Y,&B);CHKERRQ(ierr); 2939 ierr = PetscFree(nnz);CHKERRQ(ierr); 2940 } 2941 PetscFunctionReturn(0); 2942 } 2943 2944 PetscErrorCode MatConjugate_SeqAIJ(Mat mat) 2945 { 2946 #if defined(PETSC_USE_COMPLEX) 2947 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)mat->data; 2948 PetscInt i,nz; 2949 PetscScalar *a; 2950 2951 PetscFunctionBegin; 2952 nz = aij->nz; 2953 a = aij->a; 2954 for (i=0; i<nz; i++) a[i] = PetscConj(a[i]); 2955 #else 2956 PetscFunctionBegin; 2957 #endif 2958 PetscFunctionReturn(0); 2959 } 2960 2961 PetscErrorCode MatGetRowMaxAbs_SeqAIJ(Mat A,Vec v,PetscInt idx[]) 2962 { 2963 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2964 PetscErrorCode ierr; 2965 PetscInt i,j,m = A->rmap->n,*ai,*aj,ncols,n; 2966 PetscReal atmp; 2967 PetscScalar *x; 2968 MatScalar *aa; 2969 2970 PetscFunctionBegin; 2971 if (A->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 2972 aa = a->a; 2973 ai = a->i; 2974 aj = a->j; 2975 2976 ierr = VecSet(v,0.0);CHKERRQ(ierr); 2977 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 2978 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 2979 if (n != A->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector"); 2980 for (i=0; i<m; i++) { 2981 ncols = ai[1] - ai[0]; ai++; 2982 x[i] = 0.0; 2983 for (j=0; j<ncols; j++) { 2984 atmp = PetscAbsScalar(*aa); 2985 if (PetscAbsScalar(x[i]) < atmp) {x[i] = atmp; if (idx) idx[i] = *aj;} 2986 aa++; aj++; 2987 } 2988 } 2989 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 2990 PetscFunctionReturn(0); 2991 } 2992 2993 PetscErrorCode MatGetRowMax_SeqAIJ(Mat A,Vec v,PetscInt idx[]) 2994 { 2995 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 2996 PetscErrorCode ierr; 2997 PetscInt i,j,m = A->rmap->n,*ai,*aj,ncols,n; 2998 PetscScalar *x; 2999 MatScalar *aa; 3000 3001 PetscFunctionBegin; 3002 if (A->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3003 aa = a->a; 3004 ai = a->i; 3005 aj = a->j; 3006 3007 ierr = VecSet(v,0.0);CHKERRQ(ierr); 3008 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 3009 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 3010 if (n != A->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector"); 3011 for (i=0; i<m; i++) { 3012 ncols = ai[1] - ai[0]; ai++; 3013 if (ncols == A->cmap->n) { /* row is dense */ 3014 x[i] = *aa; if (idx) idx[i] = 0; 3015 } else { /* row is sparse so already KNOW maximum is 0.0 or higher */ 3016 x[i] = 0.0; 3017 if (idx) { 3018 idx[i] = 0; /* in case ncols is zero */ 3019 for (j=0;j<ncols;j++) { /* find first implicit 0.0 in the row */ 3020 if (aj[j] > j) { 3021 idx[i] = j; 3022 break; 3023 } 3024 } 3025 } 3026 } 3027 for (j=0; j<ncols; j++) { 3028 if (PetscRealPart(x[i]) < PetscRealPart(*aa)) {x[i] = *aa; if (idx) idx[i] = *aj;} 3029 aa++; aj++; 3030 } 3031 } 3032 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 3033 PetscFunctionReturn(0); 3034 } 3035 3036 PetscErrorCode MatGetRowMinAbs_SeqAIJ(Mat A,Vec v,PetscInt idx[]) 3037 { 3038 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 3039 PetscErrorCode ierr; 3040 PetscInt i,j,m = A->rmap->n,*ai,*aj,ncols,n; 3041 PetscReal atmp; 3042 PetscScalar *x; 3043 MatScalar *aa; 3044 3045 PetscFunctionBegin; 3046 if (A->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3047 aa = a->a; 3048 ai = a->i; 3049 aj = a->j; 3050 3051 ierr = VecSet(v,0.0);CHKERRQ(ierr); 3052 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 3053 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 3054 if (n != A->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector, %D vs. %D rows", A->rmap->n, n); 3055 for (i=0; i<m; i++) { 3056 ncols = ai[1] - ai[0]; ai++; 3057 if (ncols) { 3058 /* Get first nonzero */ 3059 for (j = 0; j < ncols; j++) { 3060 atmp = PetscAbsScalar(aa[j]); 3061 if (atmp > 1.0e-12) { 3062 x[i] = atmp; 3063 if (idx) idx[i] = aj[j]; 3064 break; 3065 } 3066 } 3067 if (j == ncols) {x[i] = PetscAbsScalar(*aa); if (idx) idx[i] = *aj;} 3068 } else { 3069 x[i] = 0.0; if (idx) idx[i] = 0; 3070 } 3071 for (j = 0; j < ncols; j++) { 3072 atmp = PetscAbsScalar(*aa); 3073 if (atmp > 1.0e-12 && PetscAbsScalar(x[i]) > atmp) {x[i] = atmp; if (idx) idx[i] = *aj;} 3074 aa++; aj++; 3075 } 3076 } 3077 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 3078 PetscFunctionReturn(0); 3079 } 3080 3081 PetscErrorCode MatGetRowMin_SeqAIJ(Mat A,Vec v,PetscInt idx[]) 3082 { 3083 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 3084 PetscErrorCode ierr; 3085 PetscInt i,j,m = A->rmap->n,ncols,n; 3086 const PetscInt *ai,*aj; 3087 PetscScalar *x; 3088 const MatScalar *aa; 3089 3090 PetscFunctionBegin; 3091 if (A->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3092 aa = a->a; 3093 ai = a->i; 3094 aj = a->j; 3095 3096 ierr = VecSet(v,0.0);CHKERRQ(ierr); 3097 ierr = VecGetArray(v,&x);CHKERRQ(ierr); 3098 ierr = VecGetLocalSize(v,&n);CHKERRQ(ierr); 3099 if (n != A->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector"); 3100 for (i=0; i<m; i++) { 3101 ncols = ai[1] - ai[0]; ai++; 3102 if (ncols == A->cmap->n) { /* row is dense */ 3103 x[i] = *aa; if (idx) idx[i] = 0; 3104 } else { /* row is sparse so already KNOW minimum is 0.0 or lower */ 3105 x[i] = 0.0; 3106 if (idx) { /* find first implicit 0.0 in the row */ 3107 idx[i] = 0; /* in case ncols is zero */ 3108 for (j=0; j<ncols; j++) { 3109 if (aj[j] > j) { 3110 idx[i] = j; 3111 break; 3112 } 3113 } 3114 } 3115 } 3116 for (j=0; j<ncols; j++) { 3117 if (PetscRealPart(x[i]) > PetscRealPart(*aa)) {x[i] = *aa; if (idx) idx[i] = *aj;} 3118 aa++; aj++; 3119 } 3120 } 3121 ierr = VecRestoreArray(v,&x);CHKERRQ(ierr); 3122 PetscFunctionReturn(0); 3123 } 3124 3125 PetscErrorCode MatInvertBlockDiagonal_SeqAIJ(Mat A,const PetscScalar **values) 3126 { 3127 Mat_SeqAIJ *a = (Mat_SeqAIJ*) A->data; 3128 PetscErrorCode ierr; 3129 PetscInt i,bs = PetscAbs(A->rmap->bs),mbs = A->rmap->n/bs,ipvt[5],bs2 = bs*bs,*v_pivots,ij[7],*IJ,j; 3130 MatScalar *diag,work[25],*v_work; 3131 const PetscReal shift = 0.0; 3132 PetscBool allowzeropivot,zeropivotdetected=PETSC_FALSE; 3133 3134 PetscFunctionBegin; 3135 allowzeropivot = PetscNot(A->erroriffailure); 3136 if (a->ibdiagvalid) { 3137 if (values) *values = a->ibdiag; 3138 PetscFunctionReturn(0); 3139 } 3140 ierr = MatMarkDiagonal_SeqAIJ(A);CHKERRQ(ierr); 3141 if (!a->ibdiag) { 3142 ierr = PetscMalloc1(bs2*mbs,&a->ibdiag);CHKERRQ(ierr); 3143 ierr = PetscLogObjectMemory((PetscObject)A,bs2*mbs*sizeof(PetscScalar));CHKERRQ(ierr); 3144 } 3145 diag = a->ibdiag; 3146 if (values) *values = a->ibdiag; 3147 /* factor and invert each block */ 3148 switch (bs) { 3149 case 1: 3150 for (i=0; i<mbs; i++) { 3151 ierr = MatGetValues(A,1,&i,1,&i,diag+i);CHKERRQ(ierr); 3152 if (PetscAbsScalar(diag[i] + shift) < PETSC_MACHINE_EPSILON) { 3153 if (allowzeropivot) { 3154 A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3155 A->factorerror_zeropivot_value = PetscAbsScalar(diag[i]); 3156 A->factorerror_zeropivot_row = i; 3157 ierr = PetscInfo3(A,"Zero pivot, row %D pivot %g tolerance %g\n",i,(double)PetscAbsScalar(diag[i]),(double)PETSC_MACHINE_EPSILON);CHKERRQ(ierr); 3158 } else SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Zero pivot, row %D pivot %g tolerance %g",i,(double)PetscAbsScalar(diag[i]),(double)PETSC_MACHINE_EPSILON); 3159 } 3160 diag[i] = (PetscScalar)1.0 / (diag[i] + shift); 3161 } 3162 break; 3163 case 2: 3164 for (i=0; i<mbs; i++) { 3165 ij[0] = 2*i; ij[1] = 2*i + 1; 3166 ierr = MatGetValues(A,2,ij,2,ij,diag);CHKERRQ(ierr); 3167 ierr = PetscKernel_A_gets_inverse_A_2(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3168 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3169 ierr = PetscKernel_A_gets_transpose_A_2(diag);CHKERRQ(ierr); 3170 diag += 4; 3171 } 3172 break; 3173 case 3: 3174 for (i=0; i<mbs; i++) { 3175 ij[0] = 3*i; ij[1] = 3*i + 1; ij[2] = 3*i + 2; 3176 ierr = MatGetValues(A,3,ij,3,ij,diag);CHKERRQ(ierr); 3177 ierr = PetscKernel_A_gets_inverse_A_3(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3178 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3179 ierr = PetscKernel_A_gets_transpose_A_3(diag);CHKERRQ(ierr); 3180 diag += 9; 3181 } 3182 break; 3183 case 4: 3184 for (i=0; i<mbs; i++) { 3185 ij[0] = 4*i; ij[1] = 4*i + 1; ij[2] = 4*i + 2; ij[3] = 4*i + 3; 3186 ierr = MatGetValues(A,4,ij,4,ij,diag);CHKERRQ(ierr); 3187 ierr = PetscKernel_A_gets_inverse_A_4(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3188 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3189 ierr = PetscKernel_A_gets_transpose_A_4(diag);CHKERRQ(ierr); 3190 diag += 16; 3191 } 3192 break; 3193 case 5: 3194 for (i=0; i<mbs; i++) { 3195 ij[0] = 5*i; ij[1] = 5*i + 1; ij[2] = 5*i + 2; ij[3] = 5*i + 3; ij[4] = 5*i + 4; 3196 ierr = MatGetValues(A,5,ij,5,ij,diag);CHKERRQ(ierr); 3197 ierr = PetscKernel_A_gets_inverse_A_5(diag,ipvt,work,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3198 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3199 ierr = PetscKernel_A_gets_transpose_A_5(diag);CHKERRQ(ierr); 3200 diag += 25; 3201 } 3202 break; 3203 case 6: 3204 for (i=0; i<mbs; i++) { 3205 ij[0] = 6*i; ij[1] = 6*i + 1; ij[2] = 6*i + 2; ij[3] = 6*i + 3; ij[4] = 6*i + 4; ij[5] = 6*i + 5; 3206 ierr = MatGetValues(A,6,ij,6,ij,diag);CHKERRQ(ierr); 3207 ierr = PetscKernel_A_gets_inverse_A_6(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3208 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3209 ierr = PetscKernel_A_gets_transpose_A_6(diag);CHKERRQ(ierr); 3210 diag += 36; 3211 } 3212 break; 3213 case 7: 3214 for (i=0; i<mbs; i++) { 3215 ij[0] = 7*i; ij[1] = 7*i + 1; ij[2] = 7*i + 2; ij[3] = 7*i + 3; ij[4] = 7*i + 4; ij[5] = 7*i + 5; ij[5] = 7*i + 6; 3216 ierr = MatGetValues(A,7,ij,7,ij,diag);CHKERRQ(ierr); 3217 ierr = PetscKernel_A_gets_inverse_A_7(diag,shift,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3218 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3219 ierr = PetscKernel_A_gets_transpose_A_7(diag);CHKERRQ(ierr); 3220 diag += 49; 3221 } 3222 break; 3223 default: 3224 ierr = PetscMalloc3(bs,&v_work,bs,&v_pivots,bs,&IJ);CHKERRQ(ierr); 3225 for (i=0; i<mbs; i++) { 3226 for (j=0; j<bs; j++) { 3227 IJ[j] = bs*i + j; 3228 } 3229 ierr = MatGetValues(A,bs,IJ,bs,IJ,diag);CHKERRQ(ierr); 3230 ierr = PetscKernel_A_gets_inverse_A(bs,diag,v_pivots,v_work,allowzeropivot,&zeropivotdetected);CHKERRQ(ierr); 3231 if (zeropivotdetected) A->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT; 3232 ierr = PetscKernel_A_gets_transpose_A_N(diag,bs);CHKERRQ(ierr); 3233 diag += bs2; 3234 } 3235 ierr = PetscFree3(v_work,v_pivots,IJ);CHKERRQ(ierr); 3236 } 3237 a->ibdiagvalid = PETSC_TRUE; 3238 PetscFunctionReturn(0); 3239 } 3240 3241 static PetscErrorCode MatSetRandom_SeqAIJ(Mat x,PetscRandom rctx) 3242 { 3243 PetscErrorCode ierr; 3244 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)x->data; 3245 PetscScalar a; 3246 PetscInt m,n,i,j,col; 3247 3248 PetscFunctionBegin; 3249 if (!x->assembled) { 3250 ierr = MatGetSize(x,&m,&n);CHKERRQ(ierr); 3251 for (i=0; i<m; i++) { 3252 for (j=0; j<aij->imax[i]; j++) { 3253 ierr = PetscRandomGetValue(rctx,&a);CHKERRQ(ierr); 3254 col = (PetscInt)(n*PetscRealPart(a)); 3255 ierr = MatSetValues(x,1,&i,1,&col,&a,ADD_VALUES);CHKERRQ(ierr); 3256 } 3257 } 3258 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Not yet coded"); 3259 ierr = MatAssemblyBegin(x,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3260 ierr = MatAssemblyEnd(x,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3261 PetscFunctionReturn(0); 3262 } 3263 3264 /* -------------------------------------------------------------------*/ 3265 static struct _MatOps MatOps_Values = { MatSetValues_SeqAIJ, 3266 MatGetRow_SeqAIJ, 3267 MatRestoreRow_SeqAIJ, 3268 MatMult_SeqAIJ, 3269 /* 4*/ MatMultAdd_SeqAIJ, 3270 MatMultTranspose_SeqAIJ, 3271 MatMultTransposeAdd_SeqAIJ, 3272 0, 3273 0, 3274 0, 3275 /* 10*/ 0, 3276 MatLUFactor_SeqAIJ, 3277 0, 3278 MatSOR_SeqAIJ, 3279 MatTranspose_SeqAIJ_FAST, 3280 /*1 5*/ MatGetInfo_SeqAIJ, 3281 MatEqual_SeqAIJ, 3282 MatGetDiagonal_SeqAIJ, 3283 MatDiagonalScale_SeqAIJ, 3284 MatNorm_SeqAIJ, 3285 /* 20*/ 0, 3286 MatAssemblyEnd_SeqAIJ, 3287 MatSetOption_SeqAIJ, 3288 MatZeroEntries_SeqAIJ, 3289 /* 24*/ MatZeroRows_SeqAIJ, 3290 0, 3291 0, 3292 0, 3293 0, 3294 /* 29*/ MatSetUp_SeqAIJ, 3295 0, 3296 0, 3297 0, 3298 0, 3299 /* 34*/ MatDuplicate_SeqAIJ, 3300 0, 3301 0, 3302 MatILUFactor_SeqAIJ, 3303 0, 3304 /* 39*/ MatAXPY_SeqAIJ, 3305 MatCreateSubMatrices_SeqAIJ, 3306 MatIncreaseOverlap_SeqAIJ, 3307 MatGetValues_SeqAIJ, 3308 MatCopy_SeqAIJ, 3309 /* 44*/ MatGetRowMax_SeqAIJ, 3310 MatScale_SeqAIJ, 3311 MatShift_SeqAIJ, 3312 MatDiagonalSet_SeqAIJ, 3313 MatZeroRowsColumns_SeqAIJ, 3314 /* 49*/ MatSetRandom_SeqAIJ, 3315 MatGetRowIJ_SeqAIJ, 3316 MatRestoreRowIJ_SeqAIJ, 3317 MatGetColumnIJ_SeqAIJ, 3318 MatRestoreColumnIJ_SeqAIJ, 3319 /* 54*/ MatFDColoringCreate_SeqXAIJ, 3320 0, 3321 0, 3322 MatPermute_SeqAIJ, 3323 0, 3324 /* 59*/ 0, 3325 MatDestroy_SeqAIJ, 3326 MatView_SeqAIJ, 3327 0, 3328 MatMatMatMult_SeqAIJ_SeqAIJ_SeqAIJ, 3329 /* 64*/ MatMatMatMultSymbolic_SeqAIJ_SeqAIJ_SeqAIJ, 3330 MatMatMatMultNumeric_SeqAIJ_SeqAIJ_SeqAIJ, 3331 0, 3332 0, 3333 0, 3334 /* 69*/ MatGetRowMaxAbs_SeqAIJ, 3335 MatGetRowMinAbs_SeqAIJ, 3336 0, 3337 0, 3338 0, 3339 /* 74*/ 0, 3340 MatFDColoringApply_AIJ, 3341 0, 3342 0, 3343 0, 3344 /* 79*/ MatFindZeroDiagonals_SeqAIJ, 3345 0, 3346 0, 3347 0, 3348 MatLoad_SeqAIJ, 3349 /* 84*/ MatIsSymmetric_SeqAIJ, 3350 MatIsHermitian_SeqAIJ, 3351 0, 3352 0, 3353 0, 3354 /* 89*/ MatMatMult_SeqAIJ_SeqAIJ, 3355 MatMatMultSymbolic_SeqAIJ_SeqAIJ, 3356 MatMatMultNumeric_SeqAIJ_SeqAIJ, 3357 MatPtAP_SeqAIJ_SeqAIJ, 3358 MatPtAPSymbolic_SeqAIJ_SeqAIJ_SparseAxpy, 3359 /* 94*/ MatPtAPNumeric_SeqAIJ_SeqAIJ_SparseAxpy, 3360 MatMatTransposeMult_SeqAIJ_SeqAIJ, 3361 MatMatTransposeMultSymbolic_SeqAIJ_SeqAIJ, 3362 MatMatTransposeMultNumeric_SeqAIJ_SeqAIJ, 3363 0, 3364 /* 99*/ 0, 3365 0, 3366 0, 3367 MatConjugate_SeqAIJ, 3368 0, 3369 /*104*/ MatSetValuesRow_SeqAIJ, 3370 MatRealPart_SeqAIJ, 3371 MatImaginaryPart_SeqAIJ, 3372 0, 3373 0, 3374 /*109*/ MatMatSolve_SeqAIJ, 3375 0, 3376 MatGetRowMin_SeqAIJ, 3377 0, 3378 MatMissingDiagonal_SeqAIJ, 3379 /*114*/ 0, 3380 0, 3381 0, 3382 0, 3383 0, 3384 /*119*/ 0, 3385 0, 3386 0, 3387 0, 3388 MatGetMultiProcBlock_SeqAIJ, 3389 /*124*/ MatFindNonzeroRows_SeqAIJ, 3390 MatGetColumnNorms_SeqAIJ, 3391 MatInvertBlockDiagonal_SeqAIJ, 3392 MatInvertVariableBlockDiagonal_SeqAIJ, 3393 0, 3394 /*129*/ 0, 3395 MatTransposeMatMult_SeqAIJ_SeqAIJ, 3396 MatTransposeMatMultSymbolic_SeqAIJ_SeqAIJ, 3397 MatTransposeMatMultNumeric_SeqAIJ_SeqAIJ, 3398 MatTransposeColoringCreate_SeqAIJ, 3399 /*134*/ MatTransColoringApplySpToDen_SeqAIJ, 3400 MatTransColoringApplyDenToSp_SeqAIJ, 3401 MatRARt_SeqAIJ_SeqAIJ, 3402 MatRARtSymbolic_SeqAIJ_SeqAIJ, 3403 MatRARtNumeric_SeqAIJ_SeqAIJ, 3404 /*139*/0, 3405 0, 3406 0, 3407 MatFDColoringSetUp_SeqXAIJ, 3408 MatFindOffBlockDiagonalEntries_SeqAIJ, 3409 /*144*/MatCreateMPIMatConcatenateSeqMat_SeqAIJ, 3410 MatDestroySubMatrices_SeqAIJ 3411 }; 3412 3413 PetscErrorCode MatSeqAIJSetColumnIndices_SeqAIJ(Mat mat,PetscInt *indices) 3414 { 3415 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)mat->data; 3416 PetscInt i,nz,n; 3417 3418 PetscFunctionBegin; 3419 nz = aij->maxnz; 3420 n = mat->rmap->n; 3421 for (i=0; i<nz; i++) { 3422 aij->j[i] = indices[i]; 3423 } 3424 aij->nz = nz; 3425 for (i=0; i<n; i++) { 3426 aij->ilen[i] = aij->imax[i]; 3427 } 3428 PetscFunctionReturn(0); 3429 } 3430 3431 /* 3432 * When a sparse matrix has many zero columns, we should compact them out to save the space 3433 * This happens in MatPtAPSymbolic_MPIAIJ_MPIAIJ_scalable() 3434 * */ 3435 PetscErrorCode MatSeqAIJCompactOutExtraColumns_SeqAIJ(Mat mat, ISLocalToGlobalMapping *mapping) 3436 { 3437 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)mat->data; 3438 PetscTable gid1_lid1; 3439 PetscTablePosition tpos; 3440 PetscInt gid,lid,i,j,ncols,ec; 3441 PetscInt *garray; 3442 PetscErrorCode ierr; 3443 3444 PetscFunctionBegin; 3445 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3446 PetscValidPointer(mapping,2); 3447 /* use a table */ 3448 ierr = PetscTableCreate(mat->rmap->n,mat->cmap->N+1,&gid1_lid1);CHKERRQ(ierr); 3449 ec = 0; 3450 for (i=0; i<mat->rmap->n; i++) { 3451 ncols = aij->i[i+1] - aij->i[i]; 3452 for (j=0; j<ncols; j++) { 3453 PetscInt data,gid1 = aij->j[aij->i[i] + j] + 1; 3454 ierr = PetscTableFind(gid1_lid1,gid1,&data);CHKERRQ(ierr); 3455 if (!data) { 3456 /* one based table */ 3457 ierr = PetscTableAdd(gid1_lid1,gid1,++ec,INSERT_VALUES);CHKERRQ(ierr); 3458 } 3459 } 3460 } 3461 /* form array of columns we need */ 3462 ierr = PetscMalloc1(ec+1,&garray);CHKERRQ(ierr); 3463 ierr = PetscTableGetHeadPosition(gid1_lid1,&tpos);CHKERRQ(ierr); 3464 while (tpos) { 3465 ierr = PetscTableGetNext(gid1_lid1,&tpos,&gid,&lid);CHKERRQ(ierr); 3466 gid--; 3467 lid--; 3468 garray[lid] = gid; 3469 } 3470 ierr = PetscSortInt(ec,garray);CHKERRQ(ierr); /* sort, and rebuild */ 3471 ierr = PetscTableRemoveAll(gid1_lid1);CHKERRQ(ierr); 3472 for (i=0; i<ec; i++) { 3473 ierr = PetscTableAdd(gid1_lid1,garray[i]+1,i+1,INSERT_VALUES);CHKERRQ(ierr); 3474 } 3475 /* compact out the extra columns in B */ 3476 for (i=0; i<mat->rmap->n; i++) { 3477 ncols = aij->i[i+1] - aij->i[i]; 3478 for (j=0; j<ncols; j++) { 3479 PetscInt gid1 = aij->j[aij->i[i] + j] + 1; 3480 ierr = PetscTableFind(gid1_lid1,gid1,&lid);CHKERRQ(ierr); 3481 lid--; 3482 aij->j[aij->i[i] + j] = lid; 3483 } 3484 } 3485 mat->cmap->n = mat->cmap->N = ec; 3486 mat->cmap->bs = 1; 3487 3488 ierr = PetscTableDestroy(&gid1_lid1);CHKERRQ(ierr); 3489 ierr = PetscLayoutSetUp((mat->cmap));CHKERRQ(ierr); 3490 ierr = ISLocalToGlobalMappingCreate(PETSC_COMM_SELF,mat->cmap->bs,mat->cmap->n,garray,PETSC_OWN_POINTER,mapping);CHKERRQ(ierr); 3491 ierr = ISLocalToGlobalMappingSetType(*mapping,ISLOCALTOGLOBALMAPPINGHASH);CHKERRQ(ierr); 3492 PetscFunctionReturn(0); 3493 } 3494 3495 /*@ 3496 MatSeqAIJSetColumnIndices - Set the column indices for all the rows 3497 in the matrix. 3498 3499 Input Parameters: 3500 + mat - the SeqAIJ matrix 3501 - indices - the column indices 3502 3503 Level: advanced 3504 3505 Notes: 3506 This can be called if you have precomputed the nonzero structure of the 3507 matrix and want to provide it to the matrix object to improve the performance 3508 of the MatSetValues() operation. 3509 3510 You MUST have set the correct numbers of nonzeros per row in the call to 3511 MatCreateSeqAIJ(), and the columns indices MUST be sorted. 3512 3513 MUST be called before any calls to MatSetValues(); 3514 3515 The indices should start with zero, not one. 3516 3517 @*/ 3518 PetscErrorCode MatSeqAIJSetColumnIndices(Mat mat,PetscInt *indices) 3519 { 3520 PetscErrorCode ierr; 3521 3522 PetscFunctionBegin; 3523 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3524 PetscValidPointer(indices,2); 3525 ierr = PetscUseMethod(mat,"MatSeqAIJSetColumnIndices_C",(Mat,PetscInt*),(mat,indices));CHKERRQ(ierr); 3526 PetscFunctionReturn(0); 3527 } 3528 3529 /* ----------------------------------------------------------------------------------------*/ 3530 3531 PetscErrorCode MatStoreValues_SeqAIJ(Mat mat) 3532 { 3533 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)mat->data; 3534 PetscErrorCode ierr; 3535 size_t nz = aij->i[mat->rmap->n]; 3536 3537 PetscFunctionBegin; 3538 if (!aij->nonew) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatSetOption(A,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE);first"); 3539 3540 /* allocate space for values if not already there */ 3541 if (!aij->saved_values) { 3542 ierr = PetscMalloc1(nz+1,&aij->saved_values);CHKERRQ(ierr); 3543 ierr = PetscLogObjectMemory((PetscObject)mat,(nz+1)*sizeof(PetscScalar));CHKERRQ(ierr); 3544 } 3545 3546 /* copy values over */ 3547 ierr = PetscMemcpy(aij->saved_values,aij->a,nz*sizeof(PetscScalar));CHKERRQ(ierr); 3548 PetscFunctionReturn(0); 3549 } 3550 3551 /*@ 3552 MatStoreValues - Stashes a copy of the matrix values; this allows, for 3553 example, reuse of the linear part of a Jacobian, while recomputing the 3554 nonlinear portion. 3555 3556 Collect on Mat 3557 3558 Input Parameters: 3559 . mat - the matrix (currently only AIJ matrices support this option) 3560 3561 Level: advanced 3562 3563 Common Usage, with SNESSolve(): 3564 $ Create Jacobian matrix 3565 $ Set linear terms into matrix 3566 $ Apply boundary conditions to matrix, at this time matrix must have 3567 $ final nonzero structure (i.e. setting the nonlinear terms and applying 3568 $ boundary conditions again will not change the nonzero structure 3569 $ ierr = MatSetOption(mat,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE); 3570 $ ierr = MatStoreValues(mat); 3571 $ Call SNESSetJacobian() with matrix 3572 $ In your Jacobian routine 3573 $ ierr = MatRetrieveValues(mat); 3574 $ Set nonlinear terms in matrix 3575 3576 Common Usage without SNESSolve(), i.e. when you handle nonlinear solve yourself: 3577 $ // build linear portion of Jacobian 3578 $ ierr = MatSetOption(mat,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE); 3579 $ ierr = MatStoreValues(mat); 3580 $ loop over nonlinear iterations 3581 $ ierr = MatRetrieveValues(mat); 3582 $ // call MatSetValues(mat,...) to set nonliner portion of Jacobian 3583 $ // call MatAssemblyBegin/End() on matrix 3584 $ Solve linear system with Jacobian 3585 $ endloop 3586 3587 Notes: 3588 Matrix must already be assemblied before calling this routine 3589 Must set the matrix option MatSetOption(mat,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE); before 3590 calling this routine. 3591 3592 When this is called multiple times it overwrites the previous set of stored values 3593 and does not allocated additional space. 3594 3595 .seealso: MatRetrieveValues() 3596 3597 @*/ 3598 PetscErrorCode MatStoreValues(Mat mat) 3599 { 3600 PetscErrorCode ierr; 3601 3602 PetscFunctionBegin; 3603 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3604 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3605 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3606 ierr = PetscUseMethod(mat,"MatStoreValues_C",(Mat),(mat));CHKERRQ(ierr); 3607 PetscFunctionReturn(0); 3608 } 3609 3610 PetscErrorCode MatRetrieveValues_SeqAIJ(Mat mat) 3611 { 3612 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)mat->data; 3613 PetscErrorCode ierr; 3614 PetscInt nz = aij->i[mat->rmap->n]; 3615 3616 PetscFunctionBegin; 3617 if (!aij->nonew) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatSetOption(A,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE);first"); 3618 if (!aij->saved_values) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatStoreValues(A);first"); 3619 /* copy values over */ 3620 ierr = PetscMemcpy(aij->a,aij->saved_values,nz*sizeof(PetscScalar));CHKERRQ(ierr); 3621 PetscFunctionReturn(0); 3622 } 3623 3624 /*@ 3625 MatRetrieveValues - Retrieves the copy of the matrix values; this allows, for 3626 example, reuse of the linear part of a Jacobian, while recomputing the 3627 nonlinear portion. 3628 3629 Collect on Mat 3630 3631 Input Parameters: 3632 . mat - the matrix (currently only AIJ matrices support this option) 3633 3634 Level: advanced 3635 3636 .seealso: MatStoreValues() 3637 3638 @*/ 3639 PetscErrorCode MatRetrieveValues(Mat mat) 3640 { 3641 PetscErrorCode ierr; 3642 3643 PetscFunctionBegin; 3644 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 3645 if (!mat->assembled) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for unassembled matrix"); 3646 if (mat->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix"); 3647 ierr = PetscUseMethod(mat,"MatRetrieveValues_C",(Mat),(mat));CHKERRQ(ierr); 3648 PetscFunctionReturn(0); 3649 } 3650 3651 3652 /* --------------------------------------------------------------------------------*/ 3653 /*@C 3654 MatCreateSeqAIJ - Creates a sparse matrix in AIJ (compressed row) format 3655 (the default parallel PETSc format). For good matrix assembly performance 3656 the user should preallocate the matrix storage by setting the parameter nz 3657 (or the array nnz). By setting these parameters accurately, performance 3658 during matrix assembly can be increased by more than a factor of 50. 3659 3660 Collective on MPI_Comm 3661 3662 Input Parameters: 3663 + comm - MPI communicator, set to PETSC_COMM_SELF 3664 . m - number of rows 3665 . n - number of columns 3666 . nz - number of nonzeros per row (same for all rows) 3667 - nnz - array containing the number of nonzeros in the various rows 3668 (possibly different for each row) or NULL 3669 3670 Output Parameter: 3671 . A - the matrix 3672 3673 It is recommended that one use the MatCreate(), MatSetType() and/or MatSetFromOptions(), 3674 MatXXXXSetPreallocation() paradgm instead of this routine directly. 3675 [MatXXXXSetPreallocation() is, for example, MatSeqAIJSetPreallocation] 3676 3677 Notes: 3678 If nnz is given then nz is ignored 3679 3680 The AIJ format (also called the Yale sparse matrix format or 3681 compressed row storage), is fully compatible with standard Fortran 77 3682 storage. That is, the stored row and column indices can begin at 3683 either one (as in Fortran) or zero. See the users' manual for details. 3684 3685 Specify the preallocated storage with either nz or nnz (not both). 3686 Set nz=PETSC_DEFAULT and nnz=NULL for PETSc to control dynamic memory 3687 allocation. For large problems you MUST preallocate memory or you 3688 will get TERRIBLE performance, see the users' manual chapter on matrices. 3689 3690 By default, this format uses inodes (identical nodes) when possible, to 3691 improve numerical efficiency of matrix-vector products and solves. We 3692 search for consecutive rows with the same nonzero structure, thereby 3693 reusing matrix information to achieve increased efficiency. 3694 3695 Options Database Keys: 3696 + -mat_no_inode - Do not use inodes 3697 - -mat_inode_limit <limit> - Sets inode limit (max limit=5) 3698 3699 Level: intermediate 3700 3701 .seealso: MatCreate(), MatCreateAIJ(), MatSetValues(), MatSeqAIJSetColumnIndices(), MatCreateSeqAIJWithArrays() 3702 3703 @*/ 3704 PetscErrorCode MatCreateSeqAIJ(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt nz,const PetscInt nnz[],Mat *A) 3705 { 3706 PetscErrorCode ierr; 3707 3708 PetscFunctionBegin; 3709 ierr = MatCreate(comm,A);CHKERRQ(ierr); 3710 ierr = MatSetSizes(*A,m,n,m,n);CHKERRQ(ierr); 3711 ierr = MatSetType(*A,MATSEQAIJ);CHKERRQ(ierr); 3712 ierr = MatSeqAIJSetPreallocation_SeqAIJ(*A,nz,nnz);CHKERRQ(ierr); 3713 PetscFunctionReturn(0); 3714 } 3715 3716 /*@C 3717 MatSeqAIJSetPreallocation - For good matrix assembly performance 3718 the user should preallocate the matrix storage by setting the parameter nz 3719 (or the array nnz). By setting these parameters accurately, performance 3720 during matrix assembly can be increased by more than a factor of 50. 3721 3722 Collective on MPI_Comm 3723 3724 Input Parameters: 3725 + B - The matrix 3726 . nz - number of nonzeros per row (same for all rows) 3727 - nnz - array containing the number of nonzeros in the various rows 3728 (possibly different for each row) or NULL 3729 3730 Notes: 3731 If nnz is given then nz is ignored 3732 3733 The AIJ format (also called the Yale sparse matrix format or 3734 compressed row storage), is fully compatible with standard Fortran 77 3735 storage. That is, the stored row and column indices can begin at 3736 either one (as in Fortran) or zero. See the users' manual for details. 3737 3738 Specify the preallocated storage with either nz or nnz (not both). 3739 Set nz=PETSC_DEFAULT and nnz=NULL for PETSc to control dynamic memory 3740 allocation. For large problems you MUST preallocate memory or you 3741 will get TERRIBLE performance, see the users' manual chapter on matrices. 3742 3743 You can call MatGetInfo() to get information on how effective the preallocation was; 3744 for example the fields mallocs,nz_allocated,nz_used,nz_unneeded; 3745 You can also run with the option -info and look for messages with the string 3746 malloc in them to see if additional memory allocation was needed. 3747 3748 Developers: Use nz of MAT_SKIP_ALLOCATION to not allocate any space for the matrix 3749 entries or columns indices 3750 3751 By default, this format uses inodes (identical nodes) when possible, to 3752 improve numerical efficiency of matrix-vector products and solves. We 3753 search for consecutive rows with the same nonzero structure, thereby 3754 reusing matrix information to achieve increased efficiency. 3755 3756 Options Database Keys: 3757 + -mat_no_inode - Do not use inodes 3758 - -mat_inode_limit <limit> - Sets inode limit (max limit=5) 3759 3760 Level: intermediate 3761 3762 .seealso: MatCreate(), MatCreateAIJ(), MatSetValues(), MatSeqAIJSetColumnIndices(), MatCreateSeqAIJWithArrays(), MatGetInfo() 3763 3764 @*/ 3765 PetscErrorCode MatSeqAIJSetPreallocation(Mat B,PetscInt nz,const PetscInt nnz[]) 3766 { 3767 PetscErrorCode ierr; 3768 3769 PetscFunctionBegin; 3770 PetscValidHeaderSpecific(B,MAT_CLASSID,1); 3771 PetscValidType(B,1); 3772 ierr = PetscTryMethod(B,"MatSeqAIJSetPreallocation_C",(Mat,PetscInt,const PetscInt[]),(B,nz,nnz));CHKERRQ(ierr); 3773 PetscFunctionReturn(0); 3774 } 3775 3776 PetscErrorCode MatSeqAIJSetPreallocation_SeqAIJ(Mat B,PetscInt nz,const PetscInt *nnz) 3777 { 3778 Mat_SeqAIJ *b; 3779 PetscBool skipallocation = PETSC_FALSE,realalloc = PETSC_FALSE; 3780 PetscErrorCode ierr; 3781 PetscInt i; 3782 3783 PetscFunctionBegin; 3784 if (nz >= 0 || nnz) realalloc = PETSC_TRUE; 3785 if (nz == MAT_SKIP_ALLOCATION) { 3786 skipallocation = PETSC_TRUE; 3787 nz = 0; 3788 } 3789 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 3790 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 3791 3792 if (nz == PETSC_DEFAULT || nz == PETSC_DECIDE) nz = 5; 3793 if (nz < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nz cannot be less than 0: value %D",nz); 3794 if (nnz) { 3795 for (i=0; i<B->rmap->n; i++) { 3796 if (nnz[i] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be less than 0: local row %D value %D",i,nnz[i]); 3797 if (nnz[i] > B->cmap->n) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be greater than row length: local row %D value %d rowlength %D",i,nnz[i],B->cmap->n); 3798 } 3799 } 3800 3801 B->preallocated = PETSC_TRUE; 3802 3803 b = (Mat_SeqAIJ*)B->data; 3804 3805 if (!skipallocation) { 3806 if (!b->imax) { 3807 ierr = PetscMalloc2(B->rmap->n,&b->imax,B->rmap->n,&b->ilen);CHKERRQ(ierr); 3808 ierr = PetscLogObjectMemory((PetscObject)B,2*B->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 3809 } 3810 if (!b->ipre) { 3811 ierr = PetscMalloc1(B->rmap->n,&b->ipre);CHKERRQ(ierr); 3812 ierr = PetscLogObjectMemory((PetscObject)B,B->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 3813 } 3814 if (!nnz) { 3815 if (nz == PETSC_DEFAULT || nz == PETSC_DECIDE) nz = 10; 3816 else if (nz < 0) nz = 1; 3817 for (i=0; i<B->rmap->n; i++) b->imax[i] = nz; 3818 nz = nz*B->rmap->n; 3819 } else { 3820 nz = 0; 3821 for (i=0; i<B->rmap->n; i++) {b->imax[i] = nnz[i]; nz += nnz[i];} 3822 } 3823 /* b->ilen will count nonzeros in each row so far. */ 3824 for (i=0; i<B->rmap->n; i++) b->ilen[i] = 0; 3825 3826 /* allocate the matrix space */ 3827 /* FIXME: should B's old memory be unlogged? */ 3828 ierr = MatSeqXAIJFreeAIJ(B,&b->a,&b->j,&b->i);CHKERRQ(ierr); 3829 if (B->structure_only) { 3830 ierr = PetscMalloc1(nz,&b->j);CHKERRQ(ierr); 3831 ierr = PetscMalloc1(B->rmap->n+1,&b->i);CHKERRQ(ierr); 3832 ierr = PetscLogObjectMemory((PetscObject)B,(B->rmap->n+1)*sizeof(PetscInt)+nz*sizeof(PetscInt));CHKERRQ(ierr); 3833 } else { 3834 ierr = PetscMalloc3(nz,&b->a,nz,&b->j,B->rmap->n+1,&b->i);CHKERRQ(ierr); 3835 ierr = PetscLogObjectMemory((PetscObject)B,(B->rmap->n+1)*sizeof(PetscInt)+nz*(sizeof(PetscScalar)+sizeof(PetscInt)));CHKERRQ(ierr); 3836 } 3837 b->i[0] = 0; 3838 for (i=1; i<B->rmap->n+1; i++) { 3839 b->i[i] = b->i[i-1] + b->imax[i-1]; 3840 } 3841 if (B->structure_only) { 3842 b->singlemalloc = PETSC_FALSE; 3843 b->free_a = PETSC_FALSE; 3844 } else { 3845 b->singlemalloc = PETSC_TRUE; 3846 b->free_a = PETSC_TRUE; 3847 } 3848 b->free_ij = PETSC_TRUE; 3849 } else { 3850 b->free_a = PETSC_FALSE; 3851 b->free_ij = PETSC_FALSE; 3852 } 3853 3854 if (b->ipre && nnz != b->ipre && b->imax) { 3855 /* reserve user-requested sparsity */ 3856 ierr = PetscMemcpy(b->ipre,b->imax,B->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 3857 } 3858 3859 3860 b->nz = 0; 3861 b->maxnz = nz; 3862 B->info.nz_unneeded = (double)b->maxnz; 3863 if (realalloc) { 3864 ierr = MatSetOption(B,MAT_NEW_NONZERO_ALLOCATION_ERR,PETSC_TRUE);CHKERRQ(ierr); 3865 } 3866 B->was_assembled = PETSC_FALSE; 3867 B->assembled = PETSC_FALSE; 3868 PetscFunctionReturn(0); 3869 } 3870 3871 3872 PetscErrorCode MatResetPreallocation_SeqAIJ(Mat A) 3873 { 3874 Mat_SeqAIJ *a; 3875 PetscInt i; 3876 PetscErrorCode ierr; 3877 3878 PetscFunctionBegin; 3879 PetscValidHeaderSpecific(A,MAT_CLASSID,1); 3880 a = (Mat_SeqAIJ*)A->data; 3881 /* if no saved info, we error out */ 3882 if (!a->ipre) SETERRQ(PETSC_COMM_WORLD,PETSC_ERR_ARG_NULL,"No saved preallocation info \n"); 3883 3884 if (!a->i || !a->j || !a->a || !a->imax || !a->ilen) SETERRQ(PETSC_COMM_WORLD,PETSC_ERR_ARG_NULL,"Memory info is incomplete, and can not reset preallocation \n"); 3885 3886 ierr = PetscMemcpy(a->imax,a->ipre,A->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 3887 ierr = PetscMemzero(a->ilen,A->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 3888 a->i[0] = 0; 3889 for (i=1; i<A->rmap->n+1; i++) { 3890 a->i[i] = a->i[i-1] + a->imax[i-1]; 3891 } 3892 A->preallocated = PETSC_TRUE; 3893 a->nz = 0; 3894 a->maxnz = a->i[A->rmap->n]; 3895 A->info.nz_unneeded = (double)a->maxnz; 3896 A->was_assembled = PETSC_FALSE; 3897 A->assembled = PETSC_FALSE; 3898 PetscFunctionReturn(0); 3899 } 3900 3901 /*@ 3902 MatSeqAIJSetPreallocationCSR - Allocates memory for a sparse sequential matrix in AIJ format. 3903 3904 Input Parameters: 3905 + B - the matrix 3906 . i - the indices into j for the start of each row (starts with zero) 3907 . j - the column indices for each row (starts with zero) these must be sorted for each row 3908 - v - optional values in the matrix 3909 3910 Level: developer 3911 3912 The i,j,v values are COPIED with this routine; to avoid the copy use MatCreateSeqAIJWithArrays() 3913 3914 .keywords: matrix, aij, compressed row, sparse, sequential 3915 3916 .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatSeqAIJSetPreallocation(), MatCreateSeqAIJ(), MATSEQAIJ 3917 @*/ 3918 PetscErrorCode MatSeqAIJSetPreallocationCSR(Mat B,const PetscInt i[],const PetscInt j[],const PetscScalar v[]) 3919 { 3920 PetscErrorCode ierr; 3921 3922 PetscFunctionBegin; 3923 PetscValidHeaderSpecific(B,MAT_CLASSID,1); 3924 PetscValidType(B,1); 3925 ierr = PetscTryMethod(B,"MatSeqAIJSetPreallocationCSR_C",(Mat,const PetscInt[],const PetscInt[],const PetscScalar[]),(B,i,j,v));CHKERRQ(ierr); 3926 PetscFunctionReturn(0); 3927 } 3928 3929 PetscErrorCode MatSeqAIJSetPreallocationCSR_SeqAIJ(Mat B,const PetscInt Ii[],const PetscInt J[],const PetscScalar v[]) 3930 { 3931 PetscInt i; 3932 PetscInt m,n; 3933 PetscInt nz; 3934 PetscInt *nnz, nz_max = 0; 3935 PetscScalar *values; 3936 PetscErrorCode ierr; 3937 3938 PetscFunctionBegin; 3939 if (Ii[0]) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE, "Ii[0] must be 0 it is %D", Ii[0]); 3940 3941 ierr = PetscLayoutSetUp(B->rmap);CHKERRQ(ierr); 3942 ierr = PetscLayoutSetUp(B->cmap);CHKERRQ(ierr); 3943 3944 ierr = MatGetSize(B, &m, &n);CHKERRQ(ierr); 3945 ierr = PetscMalloc1(m+1, &nnz);CHKERRQ(ierr); 3946 for (i = 0; i < m; i++) { 3947 nz = Ii[i+1]- Ii[i]; 3948 nz_max = PetscMax(nz_max, nz); 3949 if (nz < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE, "Local row %D has a negative number of columns %D", i, nnz); 3950 nnz[i] = nz; 3951 } 3952 ierr = MatSeqAIJSetPreallocation(B, 0, nnz);CHKERRQ(ierr); 3953 ierr = PetscFree(nnz);CHKERRQ(ierr); 3954 3955 if (v) { 3956 values = (PetscScalar*) v; 3957 } else { 3958 ierr = PetscCalloc1(nz_max, &values);CHKERRQ(ierr); 3959 } 3960 3961 for (i = 0; i < m; i++) { 3962 nz = Ii[i+1] - Ii[i]; 3963 ierr = MatSetValues_SeqAIJ(B, 1, &i, nz, J+Ii[i], values + (v ? Ii[i] : 0), INSERT_VALUES);CHKERRQ(ierr); 3964 } 3965 3966 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3967 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3968 3969 if (!v) { 3970 ierr = PetscFree(values);CHKERRQ(ierr); 3971 } 3972 ierr = MatSetOption(B,MAT_NEW_NONZERO_LOCATION_ERR,PETSC_TRUE);CHKERRQ(ierr); 3973 PetscFunctionReturn(0); 3974 } 3975 3976 #include <../src/mat/impls/dense/seq/dense.h> 3977 #include <petsc/private/kernels/petscaxpy.h> 3978 3979 /* 3980 Computes (B'*A')' since computing B*A directly is untenable 3981 3982 n p p 3983 ( ) ( ) ( ) 3984 m ( A ) * n ( B ) = m ( C ) 3985 ( ) ( ) ( ) 3986 3987 */ 3988 PetscErrorCode MatMatMultNumeric_SeqDense_SeqAIJ(Mat A,Mat B,Mat C) 3989 { 3990 PetscErrorCode ierr; 3991 Mat_SeqDense *sub_a = (Mat_SeqDense*)A->data; 3992 Mat_SeqAIJ *sub_b = (Mat_SeqAIJ*)B->data; 3993 Mat_SeqDense *sub_c = (Mat_SeqDense*)C->data; 3994 PetscInt i,n,m,q,p; 3995 const PetscInt *ii,*idx; 3996 const PetscScalar *b,*a,*a_q; 3997 PetscScalar *c,*c_q; 3998 3999 PetscFunctionBegin; 4000 m = A->rmap->n; 4001 n = A->cmap->n; 4002 p = B->cmap->n; 4003 a = sub_a->v; 4004 b = sub_b->a; 4005 c = sub_c->v; 4006 ierr = PetscMemzero(c,m*p*sizeof(PetscScalar));CHKERRQ(ierr); 4007 4008 ii = sub_b->i; 4009 idx = sub_b->j; 4010 for (i=0; i<n; i++) { 4011 q = ii[i+1] - ii[i]; 4012 while (q-->0) { 4013 c_q = c + m*(*idx); 4014 a_q = a + m*i; 4015 PetscKernelAXPY(c_q,*b,a_q,m); 4016 idx++; 4017 b++; 4018 } 4019 } 4020 PetscFunctionReturn(0); 4021 } 4022 4023 PetscErrorCode MatMatMultSymbolic_SeqDense_SeqAIJ(Mat A,Mat B,PetscReal fill,Mat *C) 4024 { 4025 PetscErrorCode ierr; 4026 PetscInt m=A->rmap->n,n=B->cmap->n; 4027 Mat Cmat; 4028 4029 PetscFunctionBegin; 4030 if (A->cmap->n != B->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"A->cmap->n %D != B->rmap->n %D\n",A->cmap->n,B->rmap->n); 4031 ierr = MatCreate(PetscObjectComm((PetscObject)A),&Cmat);CHKERRQ(ierr); 4032 ierr = MatSetSizes(Cmat,m,n,m,n);CHKERRQ(ierr); 4033 ierr = MatSetBlockSizesFromMats(Cmat,A,B);CHKERRQ(ierr); 4034 ierr = MatSetType(Cmat,MATSEQDENSE);CHKERRQ(ierr); 4035 ierr = MatSeqDenseSetPreallocation(Cmat,NULL);CHKERRQ(ierr); 4036 4037 Cmat->ops->matmultnumeric = MatMatMultNumeric_SeqDense_SeqAIJ; 4038 4039 *C = Cmat; 4040 PetscFunctionReturn(0); 4041 } 4042 4043 /* ----------------------------------------------------------------*/ 4044 PETSC_INTERN PetscErrorCode MatMatMult_SeqDense_SeqAIJ(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) 4045 { 4046 PetscErrorCode ierr; 4047 4048 PetscFunctionBegin; 4049 if (scall == MAT_INITIAL_MATRIX) { 4050 ierr = PetscLogEventBegin(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 4051 ierr = MatMatMultSymbolic_SeqDense_SeqAIJ(A,B,fill,C);CHKERRQ(ierr); 4052 ierr = PetscLogEventEnd(MAT_MatMultSymbolic,A,B,0,0);CHKERRQ(ierr); 4053 } 4054 ierr = PetscLogEventBegin(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 4055 ierr = MatMatMultNumeric_SeqDense_SeqAIJ(A,B,*C);CHKERRQ(ierr); 4056 ierr = PetscLogEventEnd(MAT_MatMultNumeric,A,B,0,0);CHKERRQ(ierr); 4057 PetscFunctionReturn(0); 4058 } 4059 4060 4061 /*MC 4062 MATSEQAIJ - MATSEQAIJ = "seqaij" - A matrix type to be used for sequential sparse matrices, 4063 based on compressed sparse row format. 4064 4065 Options Database Keys: 4066 . -mat_type seqaij - sets the matrix type to "seqaij" during a call to MatSetFromOptions() 4067 4068 Level: beginner 4069 4070 .seealso: MatCreateSeqAIJ(), MatSetFromOptions(), MatSetType(), MatCreate(), MatType 4071 M*/ 4072 4073 /*MC 4074 MATAIJ - MATAIJ = "aij" - A matrix type to be used for sparse matrices. 4075 4076 This matrix type is identical to MATSEQAIJ when constructed with a single process communicator, 4077 and MATMPIAIJ otherwise. As a result, for single process communicators, 4078 MatSeqAIJSetPreallocation is supported, and similarly MatMPIAIJSetPreallocation is supported 4079 for communicators controlling multiple processes. It is recommended that you call both of 4080 the above preallocation routines for simplicity. 4081 4082 Options Database Keys: 4083 . -mat_type aij - sets the matrix type to "aij" during a call to MatSetFromOptions() 4084 4085 Developer Notes: 4086 Subclasses include MATAIJCUSPARSE, MATAIJPERM, MATAIJSELL, MATAIJMKL, MATAIJCRL, and also automatically switches over to use inodes when 4087 enough exist. 4088 4089 Level: beginner 4090 4091 .seealso: MatCreateAIJ(), MatCreateSeqAIJ(), MATSEQAIJ,MATMPIAIJ 4092 M*/ 4093 4094 /*MC 4095 MATAIJCRL - MATAIJCRL = "aijcrl" - A matrix type to be used for sparse matrices. 4096 4097 This matrix type is identical to MATSEQAIJCRL when constructed with a single process communicator, 4098 and MATMPIAIJCRL otherwise. As a result, for single process communicators, 4099 MatSeqAIJSetPreallocation() is supported, and similarly MatMPIAIJSetPreallocation() is supported 4100 for communicators controlling multiple processes. It is recommended that you call both of 4101 the above preallocation routines for simplicity. 4102 4103 Options Database Keys: 4104 . -mat_type aijcrl - sets the matrix type to "aijcrl" during a call to MatSetFromOptions() 4105 4106 Level: beginner 4107 4108 .seealso: MatCreateMPIAIJCRL,MATSEQAIJCRL,MATMPIAIJCRL, MATSEQAIJCRL, MATMPIAIJCRL 4109 M*/ 4110 4111 PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqAIJCRL(Mat,MatType,MatReuse,Mat*); 4112 #if defined(PETSC_HAVE_ELEMENTAL) 4113 PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_Elemental(Mat,MatType,MatReuse,Mat*); 4114 #endif 4115 #if defined(PETSC_HAVE_HYPRE) 4116 PETSC_INTERN PetscErrorCode MatConvert_AIJ_HYPRE(Mat A,MatType,MatReuse,Mat*); 4117 PETSC_INTERN PetscErrorCode MatMatMatMult_Transpose_AIJ_AIJ(Mat,Mat,Mat,MatReuse,PetscReal,Mat*); 4118 #endif 4119 PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqDense(Mat,MatType,MatReuse,Mat*); 4120 4121 PETSC_EXTERN PetscErrorCode MatConvert_SeqAIJ_SeqSELL(Mat,MatType,MatReuse,Mat*); 4122 PETSC_INTERN PetscErrorCode MatConvert_XAIJ_IS(Mat,MatType,MatReuse,Mat*); 4123 PETSC_INTERN PetscErrorCode MatPtAP_IS_XAIJ(Mat,Mat,MatReuse,PetscReal,Mat*); 4124 4125 /*@C 4126 MatSeqAIJGetArray - gives access to the array where the data for a MATSEQAIJ matrix is stored 4127 4128 Not Collective 4129 4130 Input Parameter: 4131 . mat - a MATSEQAIJ matrix 4132 4133 Output Parameter: 4134 . array - pointer to the data 4135 4136 Level: intermediate 4137 4138 .seealso: MatSeqAIJRestoreArray(), MatSeqAIJGetArrayF90() 4139 @*/ 4140 PetscErrorCode MatSeqAIJGetArray(Mat A,PetscScalar **array) 4141 { 4142 PetscErrorCode ierr; 4143 4144 PetscFunctionBegin; 4145 ierr = PetscUseMethod(A,"MatSeqAIJGetArray_C",(Mat,PetscScalar**),(A,array));CHKERRQ(ierr); 4146 PetscFunctionReturn(0); 4147 } 4148 4149 /*@C 4150 MatSeqAIJGetMaxRowNonzeros - returns the maximum number of nonzeros in any row 4151 4152 Not Collective 4153 4154 Input Parameter: 4155 . mat - a MATSEQAIJ matrix 4156 4157 Output Parameter: 4158 . nz - the maximum number of nonzeros in any row 4159 4160 Level: intermediate 4161 4162 .seealso: MatSeqAIJRestoreArray(), MatSeqAIJGetArrayF90() 4163 @*/ 4164 PetscErrorCode MatSeqAIJGetMaxRowNonzeros(Mat A,PetscInt *nz) 4165 { 4166 Mat_SeqAIJ *aij = (Mat_SeqAIJ*)A->data; 4167 4168 PetscFunctionBegin; 4169 *nz = aij->rmax; 4170 PetscFunctionReturn(0); 4171 } 4172 4173 /*@C 4174 MatSeqAIJRestoreArray - returns access to the array where the data for a MATSEQAIJ matrix is stored obtained by MatSeqAIJGetArray() 4175 4176 Not Collective 4177 4178 Input Parameters: 4179 . mat - a MATSEQAIJ matrix 4180 . array - pointer to the data 4181 4182 Level: intermediate 4183 4184 .seealso: MatSeqAIJGetArray(), MatSeqAIJRestoreArrayF90() 4185 @*/ 4186 PetscErrorCode MatSeqAIJRestoreArray(Mat A,PetscScalar **array) 4187 { 4188 PetscErrorCode ierr; 4189 4190 PetscFunctionBegin; 4191 ierr = PetscUseMethod(A,"MatSeqAIJRestoreArray_C",(Mat,PetscScalar**),(A,array));CHKERRQ(ierr); 4192 PetscFunctionReturn(0); 4193 } 4194 4195 #if defined(PETSC_HAVE_CUDA) 4196 PETSC_EXTERN PetscErrorCode MatConvert_SeqAIJ_SeqAIJCUSPARSE(Mat); 4197 #endif 4198 4199 PETSC_EXTERN PetscErrorCode MatCreate_SeqAIJ(Mat B) 4200 { 4201 Mat_SeqAIJ *b; 4202 PetscErrorCode ierr; 4203 PetscMPIInt size; 4204 4205 PetscFunctionBegin; 4206 ierr = MPI_Comm_size(PetscObjectComm((PetscObject)B),&size);CHKERRQ(ierr); 4207 if (size > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Comm must be of size 1"); 4208 4209 ierr = PetscNewLog(B,&b);CHKERRQ(ierr); 4210 4211 B->data = (void*)b; 4212 4213 ierr = PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));CHKERRQ(ierr); 4214 4215 b->row = 0; 4216 b->col = 0; 4217 b->icol = 0; 4218 b->reallocs = 0; 4219 b->ignorezeroentries = PETSC_FALSE; 4220 b->roworiented = PETSC_TRUE; 4221 b->nonew = 0; 4222 b->diag = 0; 4223 b->solve_work = 0; 4224 B->spptr = 0; 4225 b->saved_values = 0; 4226 b->idiag = 0; 4227 b->mdiag = 0; 4228 b->ssor_work = 0; 4229 b->omega = 1.0; 4230 b->fshift = 0.0; 4231 b->idiagvalid = PETSC_FALSE; 4232 b->ibdiagvalid = PETSC_FALSE; 4233 b->keepnonzeropattern = PETSC_FALSE; 4234 4235 ierr = PetscObjectChangeTypeName((PetscObject)B,MATSEQAIJ);CHKERRQ(ierr); 4236 ierr = PetscObjectComposeFunction((PetscObject)B,"MatSeqAIJGetArray_C",MatSeqAIJGetArray_SeqAIJ);CHKERRQ(ierr); 4237 ierr = PetscObjectComposeFunction((PetscObject)B,"MatSeqAIJRestoreArray_C",MatSeqAIJRestoreArray_SeqAIJ);CHKERRQ(ierr); 4238 4239 #if defined(PETSC_HAVE_MATLAB_ENGINE) 4240 ierr = PetscObjectComposeFunction((PetscObject)B,"PetscMatlabEnginePut_C",MatlabEnginePut_SeqAIJ);CHKERRQ(ierr); 4241 ierr = PetscObjectComposeFunction((PetscObject)B,"PetscMatlabEngineGet_C",MatlabEngineGet_SeqAIJ);CHKERRQ(ierr); 4242 #endif 4243 4244 ierr = PetscObjectComposeFunction((PetscObject)B,"MatSeqAIJSetColumnIndices_C",MatSeqAIJSetColumnIndices_SeqAIJ);CHKERRQ(ierr); 4245 ierr = PetscObjectComposeFunction((PetscObject)B,"MatStoreValues_C",MatStoreValues_SeqAIJ);CHKERRQ(ierr); 4246 ierr = PetscObjectComposeFunction((PetscObject)B,"MatRetrieveValues_C",MatRetrieveValues_SeqAIJ);CHKERRQ(ierr); 4247 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqsbaij_C",MatConvert_SeqAIJ_SeqSBAIJ);CHKERRQ(ierr); 4248 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqbaij_C",MatConvert_SeqAIJ_SeqBAIJ);CHKERRQ(ierr); 4249 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqaijperm_C",MatConvert_SeqAIJ_SeqAIJPERM);CHKERRQ(ierr); 4250 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqaijsell_C",MatConvert_SeqAIJ_SeqAIJSELL);CHKERRQ(ierr); 4251 #if defined(PETSC_HAVE_MKL_SPARSE) 4252 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqaijmkl_C",MatConvert_SeqAIJ_SeqAIJMKL);CHKERRQ(ierr); 4253 #endif 4254 #if defined(PETSC_HAVE_CUDA) 4255 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqaijcusparse_C",MatConvert_SeqAIJ_SeqAIJCUSPARSE);CHKERRQ(ierr); 4256 #endif 4257 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqaijcrl_C",MatConvert_SeqAIJ_SeqAIJCRL);CHKERRQ(ierr); 4258 #if defined(PETSC_HAVE_ELEMENTAL) 4259 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_elemental_C",MatConvert_SeqAIJ_Elemental);CHKERRQ(ierr); 4260 #endif 4261 #if defined(PETSC_HAVE_HYPRE) 4262 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_hypre_C",MatConvert_AIJ_HYPRE);CHKERRQ(ierr); 4263 ierr = PetscObjectComposeFunction((PetscObject)B,"MatMatMatMult_transpose_seqaij_seqaij_C",MatMatMatMult_Transpose_AIJ_AIJ);CHKERRQ(ierr); 4264 #endif 4265 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqdense_C",MatConvert_SeqAIJ_SeqDense);CHKERRQ(ierr); 4266 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_seqsell_C",MatConvert_SeqAIJ_SeqSELL);CHKERRQ(ierr); 4267 ierr = PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaij_is_C",MatConvert_XAIJ_IS);CHKERRQ(ierr); 4268 ierr = PetscObjectComposeFunction((PetscObject)B,"MatIsTranspose_C",MatIsTranspose_SeqAIJ);CHKERRQ(ierr); 4269 ierr = PetscObjectComposeFunction((PetscObject)B,"MatIsHermitianTranspose_C",MatIsTranspose_SeqAIJ);CHKERRQ(ierr); 4270 ierr = PetscObjectComposeFunction((PetscObject)B,"MatSeqAIJSetPreallocation_C",MatSeqAIJSetPreallocation_SeqAIJ);CHKERRQ(ierr); 4271 ierr = PetscObjectComposeFunction((PetscObject)B,"MatResetPreallocation_C",MatResetPreallocation_SeqAIJ);CHKERRQ(ierr); 4272 ierr = PetscObjectComposeFunction((PetscObject)B,"MatSeqAIJSetPreallocationCSR_C",MatSeqAIJSetPreallocationCSR_SeqAIJ);CHKERRQ(ierr); 4273 ierr = PetscObjectComposeFunction((PetscObject)B,"MatReorderForNonzeroDiagonal_C",MatReorderForNonzeroDiagonal_SeqAIJ);CHKERRQ(ierr); 4274 ierr = PetscObjectComposeFunction((PetscObject)B,"MatMatMult_seqdense_seqaij_C",MatMatMult_SeqDense_SeqAIJ);CHKERRQ(ierr); 4275 ierr = PetscObjectComposeFunction((PetscObject)B,"MatMatMultSymbolic_seqdense_seqaij_C",MatMatMultSymbolic_SeqDense_SeqAIJ);CHKERRQ(ierr); 4276 ierr = PetscObjectComposeFunction((PetscObject)B,"MatMatMultNumeric_seqdense_seqaij_C",MatMatMultNumeric_SeqDense_SeqAIJ);CHKERRQ(ierr); 4277 ierr = PetscObjectComposeFunction((PetscObject)B,"MatPtAP_is_seqaij_C",MatPtAP_IS_XAIJ);CHKERRQ(ierr); 4278 ierr = MatCreate_SeqAIJ_Inode(B);CHKERRQ(ierr); 4279 ierr = PetscObjectChangeTypeName((PetscObject)B,MATSEQAIJ);CHKERRQ(ierr); 4280 ierr = MatSeqAIJSetTypeFromOptions(B);CHKERRQ(ierr); /* this allows changing the matrix subtype to say MATSEQAIJPERM */ 4281 PetscFunctionReturn(0); 4282 } 4283 4284 /* 4285 Given a matrix generated with MatGetFactor() duplicates all the information in A into B 4286 */ 4287 PetscErrorCode MatDuplicateNoCreate_SeqAIJ(Mat C,Mat A,MatDuplicateOption cpvalues,PetscBool mallocmatspace) 4288 { 4289 Mat_SeqAIJ *c,*a = (Mat_SeqAIJ*)A->data; 4290 PetscErrorCode ierr; 4291 PetscInt i,m = A->rmap->n; 4292 4293 PetscFunctionBegin; 4294 c = (Mat_SeqAIJ*)C->data; 4295 4296 C->factortype = A->factortype; 4297 c->row = 0; 4298 c->col = 0; 4299 c->icol = 0; 4300 c->reallocs = 0; 4301 4302 C->assembled = PETSC_TRUE; 4303 4304 ierr = PetscLayoutReference(A->rmap,&C->rmap);CHKERRQ(ierr); 4305 ierr = PetscLayoutReference(A->cmap,&C->cmap);CHKERRQ(ierr); 4306 4307 ierr = PetscMalloc2(m,&c->imax,m,&c->ilen);CHKERRQ(ierr); 4308 ierr = PetscLogObjectMemory((PetscObject)C, 2*m*sizeof(PetscInt));CHKERRQ(ierr); 4309 for (i=0; i<m; i++) { 4310 c->imax[i] = a->imax[i]; 4311 c->ilen[i] = a->ilen[i]; 4312 } 4313 4314 /* allocate the matrix space */ 4315 if (mallocmatspace) { 4316 ierr = PetscMalloc3(a->i[m],&c->a,a->i[m],&c->j,m+1,&c->i);CHKERRQ(ierr); 4317 ierr = PetscLogObjectMemory((PetscObject)C, a->i[m]*(sizeof(PetscScalar)+sizeof(PetscInt))+(m+1)*sizeof(PetscInt));CHKERRQ(ierr); 4318 4319 c->singlemalloc = PETSC_TRUE; 4320 4321 ierr = PetscMemcpy(c->i,a->i,(m+1)*sizeof(PetscInt));CHKERRQ(ierr); 4322 if (m > 0) { 4323 ierr = PetscMemcpy(c->j,a->j,(a->i[m])*sizeof(PetscInt));CHKERRQ(ierr); 4324 if (cpvalues == MAT_COPY_VALUES) { 4325 ierr = PetscMemcpy(c->a,a->a,(a->i[m])*sizeof(PetscScalar));CHKERRQ(ierr); 4326 } else { 4327 ierr = PetscMemzero(c->a,(a->i[m])*sizeof(PetscScalar));CHKERRQ(ierr); 4328 } 4329 } 4330 } 4331 4332 c->ignorezeroentries = a->ignorezeroentries; 4333 c->roworiented = a->roworiented; 4334 c->nonew = a->nonew; 4335 if (a->diag) { 4336 ierr = PetscMalloc1(m+1,&c->diag);CHKERRQ(ierr); 4337 ierr = PetscLogObjectMemory((PetscObject)C,(m+1)*sizeof(PetscInt));CHKERRQ(ierr); 4338 for (i=0; i<m; i++) { 4339 c->diag[i] = a->diag[i]; 4340 } 4341 } else c->diag = 0; 4342 4343 c->solve_work = 0; 4344 c->saved_values = 0; 4345 c->idiag = 0; 4346 c->ssor_work = 0; 4347 c->keepnonzeropattern = a->keepnonzeropattern; 4348 c->free_a = PETSC_TRUE; 4349 c->free_ij = PETSC_TRUE; 4350 4351 c->rmax = a->rmax; 4352 c->nz = a->nz; 4353 c->maxnz = a->nz; /* Since we allocate exactly the right amount */ 4354 C->preallocated = PETSC_TRUE; 4355 4356 c->compressedrow.use = a->compressedrow.use; 4357 c->compressedrow.nrows = a->compressedrow.nrows; 4358 if (a->compressedrow.use) { 4359 i = a->compressedrow.nrows; 4360 ierr = PetscMalloc2(i+1,&c->compressedrow.i,i,&c->compressedrow.rindex);CHKERRQ(ierr); 4361 ierr = PetscMemcpy(c->compressedrow.i,a->compressedrow.i,(i+1)*sizeof(PetscInt));CHKERRQ(ierr); 4362 ierr = PetscMemcpy(c->compressedrow.rindex,a->compressedrow.rindex,i*sizeof(PetscInt));CHKERRQ(ierr); 4363 } else { 4364 c->compressedrow.use = PETSC_FALSE; 4365 c->compressedrow.i = NULL; 4366 c->compressedrow.rindex = NULL; 4367 } 4368 c->nonzerorowcnt = a->nonzerorowcnt; 4369 C->nonzerostate = A->nonzerostate; 4370 4371 ierr = MatDuplicate_SeqAIJ_Inode(A,cpvalues,&C);CHKERRQ(ierr); 4372 ierr = PetscFunctionListDuplicate(((PetscObject)A)->qlist,&((PetscObject)C)->qlist);CHKERRQ(ierr); 4373 PetscFunctionReturn(0); 4374 } 4375 4376 PetscErrorCode MatDuplicate_SeqAIJ(Mat A,MatDuplicateOption cpvalues,Mat *B) 4377 { 4378 PetscErrorCode ierr; 4379 4380 PetscFunctionBegin; 4381 ierr = MatCreate(PetscObjectComm((PetscObject)A),B);CHKERRQ(ierr); 4382 ierr = MatSetSizes(*B,A->rmap->n,A->cmap->n,A->rmap->n,A->cmap->n);CHKERRQ(ierr); 4383 if (!(A->rmap->n % A->rmap->bs) && !(A->cmap->n % A->cmap->bs)) { 4384 ierr = MatSetBlockSizesFromMats(*B,A,A);CHKERRQ(ierr); 4385 } 4386 ierr = MatSetType(*B,((PetscObject)A)->type_name);CHKERRQ(ierr); 4387 ierr = MatDuplicateNoCreate_SeqAIJ(*B,A,cpvalues,PETSC_TRUE);CHKERRQ(ierr); 4388 PetscFunctionReturn(0); 4389 } 4390 4391 PetscErrorCode MatLoad_SeqAIJ(Mat newMat, PetscViewer viewer) 4392 { 4393 PetscBool isbinary, ishdf5; 4394 PetscErrorCode ierr; 4395 4396 PetscFunctionBegin; 4397 PetscValidHeaderSpecific(newMat,MAT_CLASSID,1); 4398 PetscValidHeaderSpecific(viewer,PETSC_VIEWER_CLASSID,2); 4399 /* force binary viewer to load .info file if it has not yet done so */ 4400 ierr = PetscViewerSetUp(viewer);CHKERRQ(ierr); 4401 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);CHKERRQ(ierr); 4402 ierr = PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERHDF5, &ishdf5);CHKERRQ(ierr); 4403 if (isbinary) { 4404 ierr = MatLoad_SeqAIJ_Binary(newMat,viewer);CHKERRQ(ierr); 4405 } else if (ishdf5) { 4406 #if defined(PETSC_HAVE_HDF5) 4407 ierr = MatLoad_AIJ_HDF5(newMat,viewer);CHKERRQ(ierr); 4408 #else 4409 SETERRQ(PetscObjectComm((PetscObject)newMat),PETSC_ERR_SUP,"HDF5 not supported in this build.\nPlease reconfigure using --download-hdf5"); 4410 #endif 4411 } else { 4412 SETERRQ2(PetscObjectComm((PetscObject)newMat),PETSC_ERR_SUP,"Viewer type %s not yet supported for reading %s matrices",((PetscObject)viewer)->type_name,((PetscObject)newMat)->type_name); 4413 } 4414 PetscFunctionReturn(0); 4415 } 4416 4417 PetscErrorCode MatLoad_SeqAIJ_Binary(Mat newMat, PetscViewer viewer) 4418 { 4419 Mat_SeqAIJ *a; 4420 PetscErrorCode ierr; 4421 PetscInt i,sum,nz,header[4],*rowlengths = 0,M,N,rows,cols; 4422 int fd; 4423 PetscMPIInt size; 4424 MPI_Comm comm; 4425 PetscInt bs = newMat->rmap->bs; 4426 4427 PetscFunctionBegin; 4428 ierr = PetscObjectGetComm((PetscObject)viewer,&comm);CHKERRQ(ierr); 4429 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 4430 if (size > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"view must have one processor"); 4431 4432 ierr = PetscOptionsBegin(comm,NULL,"Options for loading SEQAIJ matrix","Mat");CHKERRQ(ierr); 4433 ierr = PetscOptionsInt("-matload_block_size","Set the blocksize used to store the matrix","MatLoad",bs,&bs,NULL);CHKERRQ(ierr); 4434 ierr = PetscOptionsEnd();CHKERRQ(ierr); 4435 if (bs < 0) bs = 1; 4436 ierr = MatSetBlockSize(newMat,bs);CHKERRQ(ierr); 4437 4438 ierr = PetscViewerBinaryGetDescriptor(viewer,&fd);CHKERRQ(ierr); 4439 ierr = PetscBinaryRead(fd,header,4,PETSC_INT);CHKERRQ(ierr); 4440 if (header[0] != MAT_FILE_CLASSID) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"not matrix object in file"); 4441 M = header[1]; N = header[2]; nz = header[3]; 4442 4443 if (nz < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"Matrix stored in special format on disk,cannot load as SeqAIJ"); 4444 4445 /* read in row lengths */ 4446 ierr = PetscMalloc1(M,&rowlengths);CHKERRQ(ierr); 4447 ierr = PetscBinaryRead(fd,rowlengths,M,PETSC_INT);CHKERRQ(ierr); 4448 4449 /* check if sum of rowlengths is same as nz */ 4450 for (i=0,sum=0; i< M; i++) sum +=rowlengths[i]; 4451 if (sum != nz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_FILE_READ,"Inconsistant matrix data in file. no-nonzeros = %dD, sum-row-lengths = %D\n",nz,sum); 4452 4453 /* set global size if not set already*/ 4454 if (newMat->rmap->n < 0 && newMat->rmap->N < 0 && newMat->cmap->n < 0 && newMat->cmap->N < 0) { 4455 ierr = MatSetSizes(newMat,PETSC_DECIDE,PETSC_DECIDE,M,N);CHKERRQ(ierr); 4456 } else { 4457 /* if sizes and type are already set, check if the matrix global sizes are correct */ 4458 ierr = MatGetSize(newMat,&rows,&cols);CHKERRQ(ierr); 4459 if (rows < 0 && cols < 0) { /* user might provide local size instead of global size */ 4460 ierr = MatGetLocalSize(newMat,&rows,&cols);CHKERRQ(ierr); 4461 } 4462 if (M != rows || N != cols) SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED, "Matrix in file of different length (%D, %D) than the input matrix (%D, %D)",M,N,rows,cols); 4463 } 4464 ierr = MatSeqAIJSetPreallocation_SeqAIJ(newMat,0,rowlengths);CHKERRQ(ierr); 4465 a = (Mat_SeqAIJ*)newMat->data; 4466 4467 ierr = PetscBinaryRead(fd,a->j,nz,PETSC_INT);CHKERRQ(ierr); 4468 4469 /* read in nonzero values */ 4470 ierr = PetscBinaryRead(fd,a->a,nz,PETSC_SCALAR);CHKERRQ(ierr); 4471 4472 /* set matrix "i" values */ 4473 a->i[0] = 0; 4474 for (i=1; i<= M; i++) { 4475 a->i[i] = a->i[i-1] + rowlengths[i-1]; 4476 a->ilen[i-1] = rowlengths[i-1]; 4477 } 4478 ierr = PetscFree(rowlengths);CHKERRQ(ierr); 4479 4480 ierr = MatAssemblyBegin(newMat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4481 ierr = MatAssemblyEnd(newMat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4482 PetscFunctionReturn(0); 4483 } 4484 4485 PetscErrorCode MatEqual_SeqAIJ(Mat A,Mat B,PetscBool * flg) 4486 { 4487 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data; 4488 PetscErrorCode ierr; 4489 #if defined(PETSC_USE_COMPLEX) 4490 PetscInt k; 4491 #endif 4492 4493 PetscFunctionBegin; 4494 /* If the matrix dimensions are not equal,or no of nonzeros */ 4495 if ((A->rmap->n != B->rmap->n) || (A->cmap->n != B->cmap->n) ||(a->nz != b->nz)) { 4496 *flg = PETSC_FALSE; 4497 PetscFunctionReturn(0); 4498 } 4499 4500 /* if the a->i are the same */ 4501 ierr = PetscMemcmp(a->i,b->i,(A->rmap->n+1)*sizeof(PetscInt),flg);CHKERRQ(ierr); 4502 if (!*flg) PetscFunctionReturn(0); 4503 4504 /* if a->j are the same */ 4505 ierr = PetscMemcmp(a->j,b->j,(a->nz)*sizeof(PetscInt),flg);CHKERRQ(ierr); 4506 if (!*flg) PetscFunctionReturn(0); 4507 4508 /* if a->a are the same */ 4509 #if defined(PETSC_USE_COMPLEX) 4510 for (k=0; k<a->nz; k++) { 4511 if (PetscRealPart(a->a[k]) != PetscRealPart(b->a[k]) || PetscImaginaryPart(a->a[k]) != PetscImaginaryPart(b->a[k])) { 4512 *flg = PETSC_FALSE; 4513 PetscFunctionReturn(0); 4514 } 4515 } 4516 #else 4517 ierr = PetscMemcmp(a->a,b->a,(a->nz)*sizeof(PetscScalar),flg);CHKERRQ(ierr); 4518 #endif 4519 PetscFunctionReturn(0); 4520 } 4521 4522 /*@ 4523 MatCreateSeqAIJWithArrays - Creates an sequential AIJ matrix using matrix elements (in CSR format) 4524 provided by the user. 4525 4526 Collective on MPI_Comm 4527 4528 Input Parameters: 4529 + comm - must be an MPI communicator of size 1 4530 . m - number of rows 4531 . n - number of columns 4532 . i - row indices; that is i[0] = 0, i[row] = i[row-1] + number of elements in that row of the matrix 4533 . j - column indices 4534 - a - matrix values 4535 4536 Output Parameter: 4537 . mat - the matrix 4538 4539 Level: intermediate 4540 4541 Notes: 4542 The i, j, and a arrays are not copied by this routine, the user must free these arrays 4543 once the matrix is destroyed and not before 4544 4545 You cannot set new nonzero locations into this matrix, that will generate an error. 4546 4547 The i and j indices are 0 based 4548 4549 The format which is used for the sparse matrix input, is equivalent to a 4550 row-major ordering.. i.e for the following matrix, the input data expected is 4551 as shown 4552 4553 $ 1 0 0 4554 $ 2 0 3 4555 $ 4 5 6 4556 $ 4557 $ i = {0,1,3,6} [size = nrow+1 = 3+1] 4558 $ j = {0,0,2,0,1,2} [size = 6]; values must be sorted for each row 4559 $ v = {1,2,3,4,5,6} [size = 6] 4560 4561 4562 .seealso: MatCreate(), MatCreateAIJ(), MatCreateSeqAIJ(), MatCreateMPIAIJWithArrays(), MatMPIAIJSetPreallocationCSR() 4563 4564 @*/ 4565 PetscErrorCode MatCreateSeqAIJWithArrays(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt i[],PetscInt j[],PetscScalar a[],Mat *mat) 4566 { 4567 PetscErrorCode ierr; 4568 PetscInt ii; 4569 Mat_SeqAIJ *aij; 4570 #if defined(PETSC_USE_DEBUG) 4571 PetscInt jj; 4572 #endif 4573 4574 PetscFunctionBegin; 4575 if (m > 0 && i[0]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"i (row indices) must start with 0"); 4576 ierr = MatCreate(comm,mat);CHKERRQ(ierr); 4577 ierr = MatSetSizes(*mat,m,n,m,n);CHKERRQ(ierr); 4578 /* ierr = MatSetBlockSizes(*mat,,);CHKERRQ(ierr); */ 4579 ierr = MatSetType(*mat,MATSEQAIJ);CHKERRQ(ierr); 4580 ierr = MatSeqAIJSetPreallocation_SeqAIJ(*mat,MAT_SKIP_ALLOCATION,0);CHKERRQ(ierr); 4581 aij = (Mat_SeqAIJ*)(*mat)->data; 4582 ierr = PetscMalloc2(m,&aij->imax,m,&aij->ilen);CHKERRQ(ierr); 4583 4584 aij->i = i; 4585 aij->j = j; 4586 aij->a = a; 4587 aij->singlemalloc = PETSC_FALSE; 4588 aij->nonew = -1; /*this indicates that inserting a new value in the matrix that generates a new nonzero is an error*/ 4589 aij->free_a = PETSC_FALSE; 4590 aij->free_ij = PETSC_FALSE; 4591 4592 for (ii=0; ii<m; ii++) { 4593 aij->ilen[ii] = aij->imax[ii] = i[ii+1] - i[ii]; 4594 #if defined(PETSC_USE_DEBUG) 4595 if (i[ii+1] - i[ii] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative row length in i (row indices) row = %D length = %D",ii,i[ii+1] - i[ii]); 4596 for (jj=i[ii]+1; jj<i[ii+1]; jj++) { 4597 if (j[jj] < j[jj-1]) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column entry number %D (actual column %D) in row %D is not sorted",jj-i[ii],j[jj],ii); 4598 if (j[jj] == j[jj-1]) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column entry number %D (actual column %D) in row %D is identical to previous entry",jj-i[ii],j[jj],ii); 4599 } 4600 #endif 4601 } 4602 #if defined(PETSC_USE_DEBUG) 4603 for (ii=0; ii<aij->i[m]; ii++) { 4604 if (j[ii] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative column index at location = %D index = %D",ii,j[ii]); 4605 if (j[ii] > n - 1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column index to large at location = %D index = %D",ii,j[ii]); 4606 } 4607 #endif 4608 4609 ierr = MatAssemblyBegin(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4610 ierr = MatAssemblyEnd(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4611 PetscFunctionReturn(0); 4612 } 4613 /*@C 4614 MatCreateSeqAIJFromTriple - Creates an sequential AIJ matrix using matrix elements (in COO format) 4615 provided by the user. 4616 4617 Collective on MPI_Comm 4618 4619 Input Parameters: 4620 + comm - must be an MPI communicator of size 1 4621 . m - number of rows 4622 . n - number of columns 4623 . i - row indices 4624 . j - column indices 4625 . a - matrix values 4626 . nz - number of nonzeros 4627 - idx - 0 or 1 based 4628 4629 Output Parameter: 4630 . mat - the matrix 4631 4632 Level: intermediate 4633 4634 Notes: 4635 The i and j indices are 0 based 4636 4637 The format which is used for the sparse matrix input, is equivalent to a 4638 row-major ordering.. i.e for the following matrix, the input data expected is 4639 as shown: 4640 4641 1 0 0 4642 2 0 3 4643 4 5 6 4644 4645 i = {0,1,1,2,2,2} 4646 j = {0,0,2,0,1,2} 4647 v = {1,2,3,4,5,6} 4648 4649 4650 .seealso: MatCreate(), MatCreateAIJ(), MatCreateSeqAIJ(), MatCreateSeqAIJWithArrays(), MatMPIAIJSetPreallocationCSR() 4651 4652 @*/ 4653 PetscErrorCode MatCreateSeqAIJFromTriple(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt i[],PetscInt j[],PetscScalar a[],Mat *mat,PetscInt nz,PetscBool idx) 4654 { 4655 PetscErrorCode ierr; 4656 PetscInt ii, *nnz, one = 1,row,col; 4657 4658 4659 PetscFunctionBegin; 4660 ierr = PetscCalloc1(m,&nnz);CHKERRQ(ierr); 4661 for (ii = 0; ii < nz; ii++) { 4662 nnz[i[ii] - !!idx] += 1; 4663 } 4664 ierr = MatCreate(comm,mat);CHKERRQ(ierr); 4665 ierr = MatSetSizes(*mat,m,n,m,n);CHKERRQ(ierr); 4666 ierr = MatSetType(*mat,MATSEQAIJ);CHKERRQ(ierr); 4667 ierr = MatSeqAIJSetPreallocation_SeqAIJ(*mat,0,nnz);CHKERRQ(ierr); 4668 for (ii = 0; ii < nz; ii++) { 4669 if (idx) { 4670 row = i[ii] - 1; 4671 col = j[ii] - 1; 4672 } else { 4673 row = i[ii]; 4674 col = j[ii]; 4675 } 4676 ierr = MatSetValues(*mat,one,&row,one,&col,&a[ii],ADD_VALUES);CHKERRQ(ierr); 4677 } 4678 ierr = MatAssemblyBegin(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4679 ierr = MatAssemblyEnd(*mat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 4680 ierr = PetscFree(nnz);CHKERRQ(ierr); 4681 PetscFunctionReturn(0); 4682 } 4683 4684 PetscErrorCode MatSeqAIJInvalidateDiagonal(Mat A) 4685 { 4686 Mat_SeqAIJ *a=(Mat_SeqAIJ*)A->data; 4687 PetscErrorCode ierr; 4688 4689 PetscFunctionBegin; 4690 a->idiagvalid = PETSC_FALSE; 4691 a->ibdiagvalid = PETSC_FALSE; 4692 4693 ierr = MatSeqAIJInvalidateDiagonal_Inode(A);CHKERRQ(ierr); 4694 PetscFunctionReturn(0); 4695 } 4696 4697 PetscErrorCode MatCreateMPIMatConcatenateSeqMat_SeqAIJ(MPI_Comm comm,Mat inmat,PetscInt n,MatReuse scall,Mat *outmat) 4698 { 4699 PetscErrorCode ierr; 4700 PetscMPIInt size; 4701 4702 PetscFunctionBegin; 4703 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 4704 if (size == 1) { 4705 if (scall == MAT_INITIAL_MATRIX) { 4706 ierr = MatDuplicate(inmat,MAT_COPY_VALUES,outmat);CHKERRQ(ierr); 4707 } else { 4708 ierr = MatCopy(inmat,*outmat,SAME_NONZERO_PATTERN);CHKERRQ(ierr); 4709 } 4710 } else { 4711 ierr = MatCreateMPIMatConcatenateSeqMat_MPIAIJ(comm,inmat,n,scall,outmat);CHKERRQ(ierr); 4712 } 4713 PetscFunctionReturn(0); 4714 } 4715 4716 /* 4717 Permute A into C's *local* index space using rowemb,colemb. 4718 The embedding are supposed to be injections and the above implies that the range of rowemb is a subset 4719 of [0,m), colemb is in [0,n). 4720 If pattern == DIFFERENT_NONZERO_PATTERN, C is preallocated according to A. 4721 */ 4722 PetscErrorCode MatSetSeqMat_SeqAIJ(Mat C,IS rowemb,IS colemb,MatStructure pattern,Mat B) 4723 { 4724 /* If making this function public, change the error returned in this function away from _PLIB. */ 4725 PetscErrorCode ierr; 4726 Mat_SeqAIJ *Baij; 4727 PetscBool seqaij; 4728 PetscInt m,n,*nz,i,j,count; 4729 PetscScalar v; 4730 const PetscInt *rowindices,*colindices; 4731 4732 PetscFunctionBegin; 4733 if (!B) PetscFunctionReturn(0); 4734 /* Check to make sure the target matrix (and embeddings) are compatible with C and each other. */ 4735 ierr = PetscObjectBaseTypeCompare((PetscObject)B,MATSEQAIJ,&seqaij);CHKERRQ(ierr); 4736 if (!seqaij) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Input matrix is of wrong type"); 4737 if (rowemb) { 4738 ierr = ISGetLocalSize(rowemb,&m);CHKERRQ(ierr); 4739 if (m != B->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Row IS of size %D is incompatible with matrix row size %D",m,B->rmap->n); 4740 } else { 4741 if (C->rmap->n != B->rmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Input matrix is row-incompatible with the target matrix"); 4742 } 4743 if (colemb) { 4744 ierr = ISGetLocalSize(colemb,&n);CHKERRQ(ierr); 4745 if (n != B->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Diag col IS of size %D is incompatible with input matrix col size %D",n,B->cmap->n); 4746 } else { 4747 if (C->cmap->n != B->cmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Input matrix is col-incompatible with the target matrix"); 4748 } 4749 4750 Baij = (Mat_SeqAIJ*)(B->data); 4751 if (pattern == DIFFERENT_NONZERO_PATTERN) { 4752 ierr = PetscMalloc1(B->rmap->n,&nz);CHKERRQ(ierr); 4753 for (i=0; i<B->rmap->n; i++) { 4754 nz[i] = Baij->i[i+1] - Baij->i[i]; 4755 } 4756 ierr = MatSeqAIJSetPreallocation(C,0,nz);CHKERRQ(ierr); 4757 ierr = PetscFree(nz);CHKERRQ(ierr); 4758 } 4759 if (pattern == SUBSET_NONZERO_PATTERN) { 4760 ierr = MatZeroEntries(C);CHKERRQ(ierr); 4761 } 4762 count = 0; 4763 rowindices = NULL; 4764 colindices = NULL; 4765 if (rowemb) { 4766 ierr = ISGetIndices(rowemb,&rowindices);CHKERRQ(ierr); 4767 } 4768 if (colemb) { 4769 ierr = ISGetIndices(colemb,&colindices);CHKERRQ(ierr); 4770 } 4771 for (i=0; i<B->rmap->n; i++) { 4772 PetscInt row; 4773 row = i; 4774 if (rowindices) row = rowindices[i]; 4775 for (j=Baij->i[i]; j<Baij->i[i+1]; j++) { 4776 PetscInt col; 4777 col = Baij->j[count]; 4778 if (colindices) col = colindices[col]; 4779 v = Baij->a[count]; 4780 ierr = MatSetValues(C,1,&row,1,&col,&v,INSERT_VALUES);CHKERRQ(ierr); 4781 ++count; 4782 } 4783 } 4784 /* FIXME: set C's nonzerostate correctly. */ 4785 /* Assembly for C is necessary. */ 4786 C->preallocated = PETSC_TRUE; 4787 C->assembled = PETSC_TRUE; 4788 C->was_assembled = PETSC_FALSE; 4789 PetscFunctionReturn(0); 4790 } 4791 4792 PetscFunctionList MatSeqAIJList = NULL; 4793 4794 /*@C 4795 MatSeqAIJSetType - Converts a MATSEQAIJ matrix to a subtype 4796 4797 Collective on Mat 4798 4799 Input Parameters: 4800 + mat - the matrix object 4801 - matype - matrix type 4802 4803 Options Database Key: 4804 . -mat_seqai_type <method> - for example seqaijcrl 4805 4806 4807 Level: intermediate 4808 4809 .keywords: Mat, MatType, set, method 4810 4811 .seealso: PCSetType(), VecSetType(), MatCreate(), MatType, Mat 4812 @*/ 4813 PetscErrorCode MatSeqAIJSetType(Mat mat, MatType matype) 4814 { 4815 PetscErrorCode ierr,(*r)(Mat,MatType,MatReuse,Mat*); 4816 PetscBool sametype; 4817 4818 PetscFunctionBegin; 4819 PetscValidHeaderSpecific(mat,MAT_CLASSID,1); 4820 ierr = PetscObjectTypeCompare((PetscObject)mat,matype,&sametype);CHKERRQ(ierr); 4821 if (sametype) PetscFunctionReturn(0); 4822 4823 ierr = PetscFunctionListFind(MatSeqAIJList,matype,&r);CHKERRQ(ierr); 4824 if (!r) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_UNKNOWN_TYPE,"Unknown Mat type given: %s",matype); 4825 ierr = (*r)(mat,matype,MAT_INPLACE_MATRIX,&mat);CHKERRQ(ierr); 4826 PetscFunctionReturn(0); 4827 } 4828 4829 4830 /*@C 4831 MatSeqAIJRegister - - Adds a new sub-matrix type for sequential AIJ matrices 4832 4833 Not Collective 4834 4835 Input Parameters: 4836 + name - name of a new user-defined matrix type, for example MATSEQAIJCRL 4837 - function - routine to convert to subtype 4838 4839 Notes: 4840 MatSeqAIJRegister() may be called multiple times to add several user-defined solvers. 4841 4842 4843 Then, your matrix can be chosen with the procedural interface at runtime via the option 4844 $ -mat_seqaij_type my_mat 4845 4846 Level: advanced 4847 4848 .keywords: Mat, register 4849 4850 .seealso: MatSeqAIJRegisterAll() 4851 4852 4853 Level: advanced 4854 @*/ 4855 PetscErrorCode MatSeqAIJRegister(const char sname[],PetscErrorCode (*function)(Mat,MatType,MatReuse,Mat *)) 4856 { 4857 PetscErrorCode ierr; 4858 4859 PetscFunctionBegin; 4860 ierr = MatInitializePackage();CHKERRQ(ierr); 4861 ierr = PetscFunctionListAdd(&MatSeqAIJList,sname,function);CHKERRQ(ierr); 4862 PetscFunctionReturn(0); 4863 } 4864 4865 PetscBool MatSeqAIJRegisterAllCalled = PETSC_FALSE; 4866 4867 /*@C 4868 MatSeqAIJRegisterAll - Registers all of the matrix subtypes of SeqAIJ 4869 4870 Not Collective 4871 4872 Level: advanced 4873 4874 Developers Note: CUSP and CUSPARSE do not yet support the MatConvert_SeqAIJ..() paradigm and thus cannot be registered here 4875 4876 .keywords: KSP, register, all 4877 4878 .seealso: MatRegisterAll(), MatSeqAIJRegister() 4879 @*/ 4880 PetscErrorCode MatSeqAIJRegisterAll(void) 4881 { 4882 PetscErrorCode ierr; 4883 4884 PetscFunctionBegin; 4885 if (MatSeqAIJRegisterAllCalled) PetscFunctionReturn(0); 4886 MatSeqAIJRegisterAllCalled = PETSC_TRUE; 4887 4888 ierr = MatSeqAIJRegister(MATSEQAIJCRL, MatConvert_SeqAIJ_SeqAIJCRL);CHKERRQ(ierr); 4889 ierr = MatSeqAIJRegister(MATSEQAIJPERM, MatConvert_SeqAIJ_SeqAIJPERM);CHKERRQ(ierr); 4890 ierr = MatSeqAIJRegister(MATSEQAIJSELL, MatConvert_SeqAIJ_SeqAIJSELL);CHKERRQ(ierr); 4891 #if defined(PETSC_HAVE_MKL_SPARSE) 4892 ierr = MatSeqAIJRegister(MATSEQAIJMKL, MatConvert_SeqAIJ_SeqAIJMKL);CHKERRQ(ierr); 4893 #endif 4894 #if defined(PETSC_HAVE_VIENNACL) && defined(PETSC_HAVE_VIENNACL_NO_CUDA) 4895 ierr = MatSeqAIJRegister(MATMPIAIJVIENNACL, MatConvert_SeqAIJ_SeqAIJViennaCL);CHKERRQ(ierr); 4896 #endif 4897 PetscFunctionReturn(0); 4898 } 4899 4900 /* 4901 Special version for direct calls from Fortran 4902 */ 4903 #include <petsc/private/fortranimpl.h> 4904 #if defined(PETSC_HAVE_FORTRAN_CAPS) 4905 #define matsetvaluesseqaij_ MATSETVALUESSEQAIJ 4906 #elif !defined(PETSC_HAVE_FORTRAN_UNDERSCORE) 4907 #define matsetvaluesseqaij_ matsetvaluesseqaij 4908 #endif 4909 4910 /* Change these macros so can be used in void function */ 4911 #undef CHKERRQ 4912 #define CHKERRQ(ierr) CHKERRABORT(PetscObjectComm((PetscObject)A),ierr) 4913 #undef SETERRQ2 4914 #define SETERRQ2(comm,ierr,b,c,d) CHKERRABORT(comm,ierr) 4915 #undef SETERRQ3 4916 #define SETERRQ3(comm,ierr,b,c,d,e) CHKERRABORT(comm,ierr) 4917 4918 PETSC_EXTERN void PETSC_STDCALL matsetvaluesseqaij_(Mat *AA,PetscInt *mm,const PetscInt im[],PetscInt *nn,const PetscInt in[],const PetscScalar v[],InsertMode *isis, PetscErrorCode *_ierr) 4919 { 4920 Mat A = *AA; 4921 PetscInt m = *mm, n = *nn; 4922 InsertMode is = *isis; 4923 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data; 4924 PetscInt *rp,k,low,high,t,ii,row,nrow,i,col,l,rmax,N; 4925 PetscInt *imax,*ai,*ailen; 4926 PetscErrorCode ierr; 4927 PetscInt *aj,nonew = a->nonew,lastcol = -1; 4928 MatScalar *ap,value,*aa; 4929 PetscBool ignorezeroentries = a->ignorezeroentries; 4930 PetscBool roworiented = a->roworiented; 4931 4932 PetscFunctionBegin; 4933 MatCheckPreallocated(A,1); 4934 imax = a->imax; 4935 ai = a->i; 4936 ailen = a->ilen; 4937 aj = a->j; 4938 aa = a->a; 4939 4940 for (k=0; k<m; k++) { /* loop over added rows */ 4941 row = im[k]; 4942 if (row < 0) continue; 4943 #if defined(PETSC_USE_DEBUG) 4944 if (row >= A->rmap->n) SETERRABORT(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_OUTOFRANGE,"Row too large"); 4945 #endif 4946 rp = aj + ai[row]; ap = aa + ai[row]; 4947 rmax = imax[row]; nrow = ailen[row]; 4948 low = 0; 4949 high = nrow; 4950 for (l=0; l<n; l++) { /* loop over added columns */ 4951 if (in[l] < 0) continue; 4952 #if defined(PETSC_USE_DEBUG) 4953 if (in[l] >= A->cmap->n) SETERRABORT(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_OUTOFRANGE,"Column too large"); 4954 #endif 4955 col = in[l]; 4956 if (roworiented) value = v[l + k*n]; 4957 else value = v[k + l*m]; 4958 4959 if (value == 0.0 && ignorezeroentries && (is == ADD_VALUES)) continue; 4960 4961 if (col <= lastcol) low = 0; 4962 else high = nrow; 4963 lastcol = col; 4964 while (high-low > 5) { 4965 t = (low+high)/2; 4966 if (rp[t] > col) high = t; 4967 else low = t; 4968 } 4969 for (i=low; i<high; i++) { 4970 if (rp[i] > col) break; 4971 if (rp[i] == col) { 4972 if (is == ADD_VALUES) ap[i] += value; 4973 else ap[i] = value; 4974 goto noinsert; 4975 } 4976 } 4977 if (value == 0.0 && ignorezeroentries) goto noinsert; 4978 if (nonew == 1) goto noinsert; 4979 if (nonew == -1) SETERRABORT(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero in the matrix"); 4980 MatSeqXAIJReallocateAIJ(A,A->rmap->n,1,nrow,row,col,rmax,aa,ai,aj,rp,ap,imax,nonew,MatScalar); 4981 N = nrow++ - 1; a->nz++; high++; 4982 /* shift up all the later entries in this row */ 4983 for (ii=N; ii>=i; ii--) { 4984 rp[ii+1] = rp[ii]; 4985 ap[ii+1] = ap[ii]; 4986 } 4987 rp[i] = col; 4988 ap[i] = value; 4989 A->nonzerostate++; 4990 noinsert:; 4991 low = i + 1; 4992 } 4993 ailen[row] = nrow; 4994 } 4995 PetscFunctionReturnVoid(); 4996 } 4997