1 2 /* 3 Defines projective product routines where A is a MPIAIJ matrix 4 C = P^T * A * P 5 */ 6 7 #include <../src/mat/impls/aij/seq/aij.h> /*I "petscmat.h" I*/ 8 #include <../src/mat/utils/freespace.h> 9 #include <../src/mat/impls/aij/mpi/mpiaij.h> 10 #include <petscbt.h> 11 12 /* 13 #define DEBUG_MATPTAP 14 */ 15 16 extern PetscErrorCode MatDestroy_MPIAIJ(Mat); 17 #undef __FUNCT__ 18 #define __FUNCT__ "MatDestroy_MPIAIJ_PtAP" 19 PetscErrorCode MatDestroy_MPIAIJ_PtAP(Mat A) 20 { 21 PetscErrorCode ierr; 22 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data; 23 Mat_PtAPMPI *ptap=a->ptap; 24 25 PetscFunctionBegin; 26 if (ptap){ 27 ierr = PetscFree2(ptap->startsj_s,ptap->startsj_r);CHKERRQ(ierr); 28 ierr = PetscFree(ptap->bufa);CHKERRQ(ierr); 29 ierr = MatDestroy(&ptap->P_loc);CHKERRQ(ierr); 30 ierr = MatDestroy(&ptap->P_oth);CHKERRQ(ierr); 31 ierr = MatDestroy(&ptap->A_loc);CHKERRQ(ierr); /* used by MatTransposeMatMult() */ 32 if (ptap->api){ierr = PetscFree(ptap->api);CHKERRQ(ierr);} 33 if (ptap->apj){ierr = PetscFree(ptap->apj);CHKERRQ(ierr);} 34 } 35 if (ptap->merge) { 36 Mat_Merge_SeqsToMPI *merge=ptap->merge; 37 ierr = PetscFree(merge->id_r);CHKERRQ(ierr); 38 ierr = PetscFree(merge->len_s);CHKERRQ(ierr); 39 ierr = PetscFree(merge->len_r);CHKERRQ(ierr); 40 ierr = PetscFree(merge->bi);CHKERRQ(ierr); 41 ierr = PetscFree(merge->bj);CHKERRQ(ierr); 42 ierr = PetscFree(merge->buf_ri[0]);CHKERRQ(ierr); 43 ierr = PetscFree(merge->buf_ri);CHKERRQ(ierr); 44 ierr = PetscFree(merge->buf_rj[0]);CHKERRQ(ierr); 45 ierr = PetscFree(merge->buf_rj);CHKERRQ(ierr); 46 ierr = PetscFree(merge->coi);CHKERRQ(ierr); 47 ierr = PetscFree(merge->coj);CHKERRQ(ierr); 48 ierr = PetscFree(merge->owners_co);CHKERRQ(ierr); 49 ierr = PetscLayoutDestroy(&merge->rowmap);CHKERRQ(ierr); 50 ierr = merge->destroy(A);CHKERRQ(ierr); 51 ierr = PetscFree(ptap->merge);CHKERRQ(ierr); 52 } 53 ierr = PetscFree(ptap);CHKERRQ(ierr); 54 PetscFunctionReturn(0); 55 } 56 57 #undef __FUNCT__ 58 #define __FUNCT__ "MatDuplicate_MPIAIJ_MatPtAP" 59 PetscErrorCode MatDuplicate_MPIAIJ_MatPtAP(Mat A, MatDuplicateOption op, Mat *M) 60 { 61 PetscErrorCode ierr; 62 Mat_Merge_SeqsToMPI *merge; 63 PetscContainer container; 64 65 PetscFunctionBegin; 66 ierr = PetscObjectQuery((PetscObject)A,"MatMergeSeqsToMPI",(PetscObject *)&container);CHKERRQ(ierr); 67 if (container) { 68 ierr = PetscContainerGetPointer(container,(void **)&merge);CHKERRQ(ierr); 69 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Container does not exit"); 70 ierr = (*merge->duplicate)(A,op,M);CHKERRQ(ierr); 71 (*M)->ops->destroy = merge->destroy; /* =MatDestroy_MPIAIJ, *M doesn't duplicate A's container! */ 72 (*M)->ops->duplicate = merge->duplicate; /* =MatDuplicate_ MPIAIJ */ 73 PetscFunctionReturn(0); 74 } 75 76 #undef __FUNCT__ 77 #define __FUNCT__ "MatPtAPSymbolic_MPIAIJ" 78 PetscErrorCode MatPtAPSymbolic_MPIAIJ(Mat A,Mat P,PetscReal fill,Mat *C) 79 { 80 PetscErrorCode ierr; 81 82 PetscFunctionBegin; 83 if (!P->ops->ptapsymbolic_mpiaij) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_SUP,"Not implemented for A=%s and P=%s",((PetscObject)A)->type_name,((PetscObject)P)->type_name); 84 ierr = (*P->ops->ptapsymbolic_mpiaij)(A,P,fill,C);CHKERRQ(ierr); 85 PetscFunctionReturn(0); 86 } 87 88 #undef __FUNCT__ 89 #define __FUNCT__ "MatPtAPNumeric_MPIAIJ" 90 PetscErrorCode MatPtAPNumeric_MPIAIJ(Mat A,Mat P,Mat C) 91 { 92 PetscErrorCode ierr; 93 94 PetscFunctionBegin; 95 if (!P->ops->ptapnumeric_mpiaij) SETERRQ2(((PetscObject)A)->comm,PETSC_ERR_SUP,"Not implemented for A=%s and P=%s",((PetscObject)A)->type_name,((PetscObject)P)->type_name); 96 ierr = (*P->ops->ptapnumeric_mpiaij)(A,P,C);CHKERRQ(ierr); 97 PetscFunctionReturn(0); 98 } 99 100 #undef __FUNCT__ 101 #define __FUNCT__ "MatPtAPSymbolic_MPIAIJ_MPIAIJ" 102 PetscErrorCode MatPtAPSymbolic_MPIAIJ_MPIAIJ(Mat A,Mat P,PetscReal fill,Mat *C) 103 { 104 PetscErrorCode ierr; 105 Mat Cmpi; 106 Mat_PtAPMPI *ptap; 107 PetscFreeSpaceList free_space=PETSC_NULL,current_space=PETSC_NULL; 108 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data,*p=(Mat_MPIAIJ*)P->data,*c; 109 Mat_SeqAIJ *ad=(Mat_SeqAIJ*)(a->A)->data,*ao=(Mat_SeqAIJ*)(a->B)->data; 110 Mat_SeqAIJ *p_loc,*p_oth; 111 PetscInt *pi_loc,*pj_loc,*pi_oth,*pj_oth,*pdti,*pdtj,*poti,*potj,*ptJ; 112 PetscInt *adi=ad->i,*aj,*aoi=ao->i,nnz; 113 PetscInt nlnk,*lnk,*owners_co,*coi,*coj,i,k,pnz,row; 114 PetscInt am=A->rmap->n,pN=P->cmap->N,pm=P->rmap->n,pn=P->cmap->n; 115 PetscBT lnkbt; 116 MPI_Comm comm=((PetscObject)A)->comm; 117 PetscMPIInt size,rank,tagi,tagj,*len_si,*len_s,*len_ri; 118 PetscInt **buf_rj,**buf_ri,**buf_ri_k; 119 PetscInt len,proc,*dnz,*onz,*owners; 120 PetscInt nzi,*bi,*bj; 121 PetscInt nrows,*buf_s,*buf_si,*buf_si_i,**nextrow,**nextci; 122 MPI_Request *swaits,*rwaits; 123 MPI_Status *sstatus,rstatus; 124 Mat_Merge_SeqsToMPI *merge; 125 PetscInt *api,*apj,*Jptr,apnz,*prmap=p->garray,pon,nspacedouble=0,j,ap_rmax=0; 126 PetscReal afill=1.0,afill_tmp; 127 PetscInt rstart = P->cmap->rstart,rmax; 128 PetscScalar *vals; 129 130 PetscFunctionBegin; 131 /* check if matrix local sizes are compatible */ 132 if (A->rmap->rstart != P->rmap->rstart || A->rmap->rend != P->rmap->rend){ 133 SETERRQ4(comm,PETSC_ERR_ARG_SIZ,"Matrix local dimensions are incompatible, Arow (%D, %D) != Prow (%D,%D)",A->rmap->rstart,A->rmap->rend,P->rmap->rstart,P->rmap->rend); 134 } 135 if (A->cmap->rstart != P->rmap->rstart || A->cmap->rend != P->rmap->rend){ 136 SETERRQ4(comm,PETSC_ERR_ARG_SIZ,"Matrix local dimensions are incompatible, Acol (%D, %D) != Prow (%D,%D)",A->cmap->rstart,A->cmap->rend,P->rmap->rstart,P->rmap->rend); 137 } 138 139 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 140 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 141 142 /* create struct Mat_PtAPMPI and attached it to C later */ 143 ierr = PetscNew(Mat_PtAPMPI,&ptap);CHKERRQ(ierr); 144 ptap->reuse = MAT_INITIAL_MATRIX; 145 146 147 /* get P_oth by taking rows of P (= non-zero cols of local A) from other processors */ 148 ierr = MatGetBrowsOfAoCols_MPIAIJ(A,P,MAT_INITIAL_MATRIX,&ptap->startsj_s,&ptap->startsj_r,&ptap->bufa,&ptap->P_oth);CHKERRQ(ierr); 149 #if defined(DEBUG_MATPTAP) 150 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] P_oth is done...\n",rank); 151 #endif 152 153 /* get P_loc by taking all local rows of P */ 154 ierr = MatMPIAIJGetLocalMat(P,MAT_INITIAL_MATRIX,&ptap->P_loc);CHKERRQ(ierr); 155 #if defined(DEBUG_MATPTAP) 156 ierr = MPI_Barrier(comm);CHKERRQ(ierr); 157 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] P_loc is done...\n",rank); 158 #endif 159 160 p_loc = (Mat_SeqAIJ*)(ptap->P_loc)->data; 161 p_oth = (Mat_SeqAIJ*)(ptap->P_oth)->data; 162 pi_loc = p_loc->i; pj_loc = p_loc->j; 163 pi_oth = p_oth->i; pj_oth = p_oth->j; 164 165 /* first, compute symbolic AP = A_loc*P = A_diag*P_loc + A_off*P_oth */ 166 /*-------------------------------------------------------------------*/ 167 ierr = PetscMalloc((am+1)*sizeof(PetscInt),&api);CHKERRQ(ierr); 168 api[0] = 0; 169 170 /* create and initialize a linked list */ 171 nlnk = pN+1; 172 #if defined(DEBUG_MATPTAP) 173 ierr = MPI_Barrier(comm);CHKERRQ(ierr); 174 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] call PetscLLCreate(), pN %D, P_rmax %D %D; Annz %D\n",rank,pN,p_loc->rmax,p_oth->rmax,adi[am]+aoi[am]); 175 #endif 176 ierr = PetscLLCreate(pN,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 177 #if defined(DEBUG_MATPTAP) 178 ierr = MPI_Barrier(comm);CHKERRQ(ierr); 179 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] call PetscLLCreate() is done\n",rank); 180 #endif 181 182 /* Initial FreeSpace size is fill*(nnz(A) + nnz(P)) */ 183 ierr = PetscFreeSpaceGet((PetscInt)(fill*(adi[am]+aoi[am]+pi_loc[pm])),&free_space);CHKERRQ(ierr); 184 current_space = free_space; 185 186 for (i=0; i<am; i++) { 187 apnz = 0; 188 /* diagonal portion of A */ 189 nzi = adi[i+1] - adi[i]; 190 aj = ad->j + adi[i]; 191 for (j=0; j<nzi; j++){ 192 row = aj[j]; 193 pnz = pi_loc[row+1] - pi_loc[row]; 194 Jptr = pj_loc + pi_loc[row]; 195 /* add non-zero cols of P into the sorted linked list lnk */ 196 ierr = PetscLLAddSorted(pnz,Jptr,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 197 apnz += nlnk; 198 } 199 /* off-diagonal portion of A */ 200 nzi = aoi[i+1] - aoi[i]; 201 aj = ao->j + aoi[i]; 202 for (j=0; j<nzi; j++){ 203 row = aj[j]; 204 pnz = pi_oth[row+1] - pi_oth[row]; 205 Jptr = pj_oth + pi_oth[row]; 206 ierr = PetscLLAddSorted(pnz,Jptr,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 207 apnz += nlnk; 208 } 209 210 api[i+1] = api[i] + apnz; 211 if (ap_rmax < apnz) ap_rmax = apnz; 212 213 /* if free space is not available, double the total space in the list */ 214 if (current_space->local_remaining<apnz) { 215 ierr = PetscFreeSpaceGet(apnz+current_space->total_array_size,¤t_space);CHKERRQ(ierr); 216 nspacedouble++; 217 } 218 219 /* Copy data into free space, then initialize lnk */ 220 ierr = PetscLLClean(pN,pN,apnz,lnk,current_space->array,lnkbt);CHKERRQ(ierr); 221 current_space->array += apnz; 222 current_space->local_used += apnz; 223 current_space->local_remaining -= apnz; 224 } 225 /* Allocate space for apj, initialize apj, and */ 226 /* destroy list of free space and other temporary array(s) */ 227 ierr = PetscMalloc((api[am]+1)*sizeof(PetscInt),&apj);CHKERRQ(ierr); 228 ierr = PetscFreeSpaceContiguous(&free_space,apj);CHKERRQ(ierr); 229 afill_tmp = (PetscReal)api[am]/(adi[am]+aoi[am]+pi_loc[pm]); 230 if (afill_tmp > afill) afill = afill_tmp; 231 232 /* determine symbolic Co=(p->B)^T*AP - send to others */ 233 /*----------------------------------------------------*/ 234 ierr = MatGetSymbolicTranspose_SeqAIJ(p->B,&poti,&potj);CHKERRQ(ierr); 235 236 /* then, compute symbolic Co = (p->B)^T*AP */ 237 pon = (p->B)->cmap->n; /* total num of rows to be sent to other processors 238 >= (num of nonzero rows of C_seq) - pn */ 239 ierr = PetscMalloc((pon+1)*sizeof(PetscInt),&coi);CHKERRQ(ierr); 240 coi[0] = 0; 241 242 /* set initial free space to be fill*(nnz(p->B) + nnz(AP)) */ 243 nnz = fill*(poti[pon] + api[am]); 244 ierr = PetscFreeSpaceGet(nnz,&free_space); 245 current_space = free_space; 246 247 for (i=0; i<pon; i++) { 248 nnz = 0; 249 pnz = poti[i+1] - poti[i]; 250 ptJ = potj + poti[i]; 251 for (j=0; j<pnz; j++){ 252 row = ptJ[j]; /* row of AP == col of Pot */ 253 apnz = api[row+1] - api[row]; 254 Jptr = apj + api[row]; 255 /* add non-zero cols of AP into the sorted linked list lnk */ 256 ierr = PetscLLAddSorted(apnz,Jptr,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 257 nnz += nlnk; 258 } 259 260 /* If free space is not available, double the total space in the list */ 261 if (current_space->local_remaining<nnz) { 262 ierr = PetscFreeSpaceGet(nnz+current_space->total_array_size,¤t_space);CHKERRQ(ierr); 263 nspacedouble++; 264 } 265 266 /* Copy data into free space, and zero out denserows */ 267 ierr = PetscLLClean(pN,pN,nnz,lnk,current_space->array,lnkbt);CHKERRQ(ierr); 268 current_space->array += nnz; 269 current_space->local_used += nnz; 270 current_space->local_remaining -= nnz; 271 coi[i+1] = coi[i] + nnz; 272 } 273 ierr = PetscMalloc((coi[pon]+1)*sizeof(PetscInt),&coj);CHKERRQ(ierr); 274 ierr = PetscFreeSpaceContiguous(&free_space,coj);CHKERRQ(ierr); 275 afill_tmp = (PetscReal)coi[pon]/(poti[pon] + api[am]); 276 if (afill_tmp > afill) afill = afill_tmp; 277 ierr = MatRestoreSymbolicTranspose_SeqAIJ(p->B,&poti,&potj);CHKERRQ(ierr); 278 279 #if defined(DEBUG_MATPTAP) 280 ierr = MPI_Barrier(comm);CHKERRQ(ierr); 281 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] local Co is done\n",rank); 282 #endif 283 284 /* send j-array (coj) of Co to other processors */ 285 /*----------------------------------------------*/ 286 /* determine row ownership */ 287 ierr = PetscNew(Mat_Merge_SeqsToMPI,&merge);CHKERRQ(ierr); 288 ierr = PetscLayoutCreate(comm,&merge->rowmap);CHKERRQ(ierr); 289 merge->rowmap->n = pn; 290 merge->rowmap->bs = 1; 291 ierr = PetscLayoutSetUp(merge->rowmap);CHKERRQ(ierr); 292 owners = merge->rowmap->range; 293 294 /* determine the number of messages to send, their lengths */ 295 ierr = PetscMalloc(size*sizeof(PetscMPIInt),&len_si);CHKERRQ(ierr); 296 ierr = PetscMemzero(len_si,size*sizeof(PetscMPIInt));CHKERRQ(ierr); 297 ierr = PetscMalloc(size*sizeof(PetscMPIInt),&merge->len_s);CHKERRQ(ierr); 298 len_s = merge->len_s; 299 merge->nsend = 0; 300 301 ierr = PetscMalloc((size+2)*sizeof(PetscInt),&owners_co);CHKERRQ(ierr); 302 ierr = PetscMemzero(len_s,size*sizeof(PetscMPIInt));CHKERRQ(ierr); 303 304 proc = 0; 305 for (i=0; i<pon; i++){ 306 while (prmap[i] >= owners[proc+1]) proc++; 307 len_si[proc]++; /* num of rows in Co to be sent to [proc] */ 308 len_s[proc] += coi[i+1] - coi[i]; 309 } 310 311 len = 0; /* max length of buf_si[] */ 312 owners_co[0] = 0; 313 for (proc=0; proc<size; proc++){ 314 owners_co[proc+1] = owners_co[proc] + len_si[proc]; 315 if (len_si[proc]){ 316 merge->nsend++; 317 len_si[proc] = 2*(len_si[proc] + 1); 318 len += len_si[proc]; 319 } 320 } 321 322 /* determine the number and length of messages to receive for coi and coj */ 323 ierr = PetscGatherNumberOfMessages(comm,PETSC_NULL,len_s,&merge->nrecv);CHKERRQ(ierr); 324 ierr = PetscGatherMessageLengths2(comm,merge->nsend,merge->nrecv,len_s,len_si,&merge->id_r,&merge->len_r,&len_ri);CHKERRQ(ierr); 325 326 /* post the Irecv and Isend of coj */ 327 ierr = PetscCommGetNewTag(comm,&tagj);CHKERRQ(ierr); 328 ierr = PetscPostIrecvInt(comm,tagj,merge->nrecv,merge->id_r,merge->len_r,&buf_rj,&rwaits);CHKERRQ(ierr); 329 ierr = PetscMalloc((merge->nsend+1)*sizeof(MPI_Request),&swaits);CHKERRQ(ierr); 330 for (proc=0, k=0; proc<size; proc++){ 331 if (!len_s[proc]) continue; 332 i = owners_co[proc]; 333 ierr = MPI_Isend(coj+coi[i],len_s[proc],MPIU_INT,proc,tagj,comm,swaits+k);CHKERRQ(ierr); 334 k++; 335 } 336 337 /* receives and sends of coj are complete */ 338 ierr = PetscMalloc(size*sizeof(MPI_Status),&sstatus);CHKERRQ(ierr); 339 for (i=0; i<merge->nrecv; i++){ 340 ierr = MPI_Waitany(merge->nrecv,rwaits,&j,&rstatus);CHKERRQ(ierr); 341 } 342 ierr = PetscFree(rwaits);CHKERRQ(ierr); 343 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,swaits,sstatus);CHKERRQ(ierr);} 344 345 /* send and recv coi */ 346 /*-------------------*/ 347 ierr = PetscCommGetNewTag(comm,&tagi);CHKERRQ(ierr); 348 ierr = PetscPostIrecvInt(comm,tagi,merge->nrecv,merge->id_r,len_ri,&buf_ri,&rwaits);CHKERRQ(ierr); 349 ierr = PetscMalloc((len+1)*sizeof(PetscInt),&buf_s);CHKERRQ(ierr); 350 buf_si = buf_s; /* points to the beginning of k-th msg to be sent */ 351 for (proc=0,k=0; proc<size; proc++){ 352 if (!len_s[proc]) continue; 353 /* form outgoing message for i-structure: 354 buf_si[0]: nrows to be sent 355 [1:nrows]: row index (global) 356 [nrows+1:2*nrows+1]: i-structure index 357 */ 358 /*-------------------------------------------*/ 359 nrows = len_si[proc]/2 - 1; 360 buf_si_i = buf_si + nrows+1; 361 buf_si[0] = nrows; 362 buf_si_i[0] = 0; 363 nrows = 0; 364 for (i=owners_co[proc]; i<owners_co[proc+1]; i++){ 365 nzi = coi[i+1] - coi[i]; 366 buf_si_i[nrows+1] = buf_si_i[nrows] + nzi; /* i-structure */ 367 buf_si[nrows+1] =prmap[i] -owners[proc]; /* local row index */ 368 nrows++; 369 } 370 ierr = MPI_Isend(buf_si,len_si[proc],MPIU_INT,proc,tagi,comm,swaits+k);CHKERRQ(ierr); 371 k++; 372 buf_si += len_si[proc]; 373 } 374 i = merge->nrecv; 375 while (i--) { 376 ierr = MPI_Waitany(merge->nrecv,rwaits,&j,&rstatus);CHKERRQ(ierr); 377 } 378 ierr = PetscFree(rwaits);CHKERRQ(ierr); 379 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,swaits,sstatus);CHKERRQ(ierr);} 380 /* 381 ierr = PetscInfo2(A,"nsend: %d, nrecv: %d\n",merge->nsend,merge->nrecv);CHKERRQ(ierr); 382 for (i=0; i<merge->nrecv; i++){ 383 ierr = PetscInfo3(A,"recv len_ri=%d, len_rj=%d from [%d]\n",len_ri[i],merge->len_r[i],merge->id_r[i]);CHKERRQ(ierr); 384 } 385 */ 386 ierr = PetscFree(len_si);CHKERRQ(ierr); 387 ierr = PetscFree(len_ri);CHKERRQ(ierr); 388 ierr = PetscFree(swaits);CHKERRQ(ierr); 389 ierr = PetscFree(sstatus);CHKERRQ(ierr); 390 ierr = PetscFree(buf_s);CHKERRQ(ierr); 391 392 #if defined(DEBUG_MATPTAP) 393 ierr = MPI_Barrier(comm);CHKERRQ(ierr); 394 if (!rank) ierr = PetscPrintf(PETSC_COMM_SELF,"[%d] Co is done\n",rank); 395 #endif 396 397 /* compute the local portion of C (mpi mat) */ 398 /*------------------------------------------*/ 399 ierr = MatGetSymbolicTranspose_SeqAIJ(p->A,&pdti,&pdtj);CHKERRQ(ierr); 400 401 /* allocate bi array and free space for accumulating nonzero column info */ 402 ierr = PetscMalloc((pn+1)*sizeof(PetscInt),&bi);CHKERRQ(ierr); 403 bi[0] = 0; 404 405 /* set initial free space to be fill*(nnz(P) + nnz(AP)) */ 406 nnz = fill*(pi_loc[pm] + api[am]); 407 ierr = PetscFreeSpaceGet(nnz,&free_space); 408 current_space = free_space; 409 410 ierr = PetscMalloc3(merge->nrecv,PetscInt**,&buf_ri_k,merge->nrecv,PetscInt*,&nextrow,merge->nrecv,PetscInt*,&nextci);CHKERRQ(ierr); 411 for (k=0; k<merge->nrecv; k++){ 412 buf_ri_k[k] = buf_ri[k]; /* beginning of k-th recved i-structure */ 413 nrows = *buf_ri_k[k]; 414 nextrow[k] = buf_ri_k[k] + 1; /* next row number of k-th recved i-structure */ 415 nextci[k] = buf_ri_k[k] + (nrows + 1);/* poins to the next i-structure of k-th recved i-structure */ 416 } 417 ierr = MatPreallocateInitialize(comm,pn,pn,dnz,onz);CHKERRQ(ierr); 418 rmax = 0; 419 for (i=0; i<pn; i++) { 420 /* add pdt[i,:]*AP into lnk */ 421 nnz = 0; 422 pnz = pdti[i+1] - pdti[i]; 423 ptJ = pdtj + pdti[i]; 424 for (j=0; j<pnz; j++){ 425 row = ptJ[j]; /* row of AP == col of Pt */ 426 apnz = api[row+1] - api[row]; 427 Jptr = apj + api[row]; 428 /* add non-zero cols of AP into the sorted linked list lnk */ 429 ierr = PetscLLAddSorted(apnz,Jptr,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 430 nnz += nlnk; 431 } 432 /* add received col data into lnk */ 433 for (k=0; k<merge->nrecv; k++){ /* k-th received message */ 434 if (i == *nextrow[k]) { /* i-th row */ 435 nzi = *(nextci[k]+1) - *nextci[k]; 436 Jptr = buf_rj[k] + *nextci[k]; 437 ierr = PetscLLAddSorted(nzi,Jptr,pN,nlnk,lnk,lnkbt);CHKERRQ(ierr); 438 nnz += nlnk; 439 nextrow[k]++; nextci[k]++; 440 } 441 } 442 443 /* if free space is not available, make more free space */ 444 if (current_space->local_remaining<nnz) { 445 ierr = PetscFreeSpaceGet(nnz+current_space->total_array_size,¤t_space);CHKERRQ(ierr); 446 nspacedouble++; 447 } 448 /* copy data into free space, then initialize lnk */ 449 ierr = PetscLLClean(pN,pN,nnz,lnk,current_space->array,lnkbt);CHKERRQ(ierr); 450 ierr = MatPreallocateSet(i+owners[rank],nnz,current_space->array,dnz,onz);CHKERRQ(ierr); 451 current_space->array += nnz; 452 current_space->local_used += nnz; 453 current_space->local_remaining -= nnz; 454 bi[i+1] = bi[i] + nnz; 455 if (nnz > rmax) rmax = nnz; 456 } 457 ierr = MatRestoreSymbolicTranspose_SeqAIJ(p->A,&pdti,&pdtj);CHKERRQ(ierr); 458 ierr = PetscFree3(buf_ri_k,nextrow,nextci);CHKERRQ(ierr); 459 460 ierr = PetscMalloc((bi[pn]+1)*sizeof(PetscInt),&bj);CHKERRQ(ierr); 461 ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr); 462 afill_tmp = (PetscReal)bi[pn]/(pi_loc[pm] + api[am]); 463 if (afill_tmp > afill) afill = afill_tmp; 464 ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr); 465 466 /* create symbolic parallel matrix Cmpi - why cannot be assembled in Numeric part - check runex55_SA ? */ 467 /*------------------------------------------------------------------------------------------------------*/ 468 ierr = PetscMalloc((rmax+1)*sizeof(PetscScalar),&vals);CHKERRQ(ierr); 469 ierr = PetscMemzero(vals,rmax*sizeof(PetscScalar));CHKERRQ(ierr); 470 471 ierr = MatCreate(comm,&Cmpi);CHKERRQ(ierr); 472 ierr = MatSetSizes(Cmpi,pn,pn,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 473 ierr = MatSetType(Cmpi,MATMPIAIJ);CHKERRQ(ierr); 474 ierr = MatMPIAIJSetPreallocation(Cmpi,0,dnz,0,onz);CHKERRQ(ierr); 475 ierr = MatPreallocateFinalize(dnz,onz);CHKERRQ(ierr); 476 ierr = MatSetBlockSize(Cmpi,1);CHKERRQ(ierr); 477 for (i=0; i<pn; i++){ 478 row = i + rstart; 479 nnz = bi[i+1] - bi[i]; 480 Jptr = bj + bi[i]; 481 ierr = MatSetValues(Cmpi,1,&row,nnz,Jptr,vals,INSERT_VALUES);CHKERRQ(ierr); 482 } 483 ierr = MatAssemblyBegin(Cmpi,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 484 ierr = MatAssemblyEnd(Cmpi,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 485 ierr = PetscFree(vals);CHKERRQ(ierr); 486 487 merge->bi = bi; 488 merge->bj = bj; 489 merge->coi = coi; 490 merge->coj = coj; 491 merge->buf_ri = buf_ri; 492 merge->buf_rj = buf_rj; 493 merge->owners_co = owners_co; 494 merge->destroy = Cmpi->ops->destroy; 495 merge->duplicate = Cmpi->ops->duplicate; 496 497 /* Cmpi is not ready for use - assembly will be done by MatPtAPNumeric() */ 498 /* Cmpi->assembled = PETSC_FALSE; */ 499 Cmpi->ops->destroy = MatDestroy_MPIAIJ_PtAP; 500 Cmpi->ops->duplicate = MatDuplicate_MPIAIJ_MatPtAP; 501 502 /* attach the supporting struct to Cmpi for reuse */ 503 c = (Mat_MPIAIJ*)Cmpi->data; 504 c->ptap = ptap; 505 ptap->api = api; 506 ptap->apj = apj; 507 ptap->merge = merge; 508 ptap->rmax = ap_rmax; 509 510 *C = Cmpi; 511 #if defined(PETSC_USE_INFO) 512 if (bi[pn] != 0) { 513 ierr = PetscInfo3(Cmpi,"Reallocs %D; Fill ratio: given %G needed %G.\n",nspacedouble,fill,afill);CHKERRQ(ierr); 514 ierr = PetscInfo1(Cmpi,"Use MatPtAP(A,P,MatReuse,%G,&C) for best performance.\n",afill);CHKERRQ(ierr); 515 } else { 516 ierr = PetscInfo(Cmpi,"Empty matrix product\n");CHKERRQ(ierr); 517 } 518 #endif 519 PetscFunctionReturn(0); 520 } 521 522 #undef __FUNCT__ 523 #define __FUNCT__ "MatPtAPNumeric_MPIAIJ_MPIAIJ" 524 PetscErrorCode MatPtAPNumeric_MPIAIJ_MPIAIJ(Mat A,Mat P,Mat C) 525 { 526 PetscErrorCode ierr; 527 Mat_Merge_SeqsToMPI *merge; 528 Mat_MPIAIJ *a=(Mat_MPIAIJ*)A->data,*p=(Mat_MPIAIJ*)P->data,*c=(Mat_MPIAIJ*)C->data; 529 Mat_SeqAIJ *ad=(Mat_SeqAIJ*)(a->A)->data,*ao=(Mat_SeqAIJ*)(a->B)->data; 530 Mat_SeqAIJ *pd=(Mat_SeqAIJ*)(p->A)->data,*po=(Mat_SeqAIJ*)(p->B)->data; 531 Mat_SeqAIJ *p_loc,*p_oth; 532 Mat_PtAPMPI *ptap; 533 PetscInt *adi=ad->i,*aoi=ao->i,*adj,*aoj,*apJ,nextp; 534 PetscInt *pi_loc,*pj_loc,*pi_oth,*pj_oth,*pJ,*pj; 535 PetscInt i,j,k,anz,pnz,apnz,nextap,row,*cj; 536 MatScalar *ada,*aoa,*apa,*pa,*ca,*pa_loc,*pa_oth,valtmp; 537 PetscInt am=A->rmap->n,cm=C->rmap->n,pon=(p->B)->cmap->n; 538 MPI_Comm comm=((PetscObject)C)->comm; 539 PetscMPIInt size,rank,taga,*len_s; 540 PetscInt *owners,proc,nrows,**buf_ri_k,**nextrow,**nextci; 541 PetscInt **buf_ri,**buf_rj; 542 PetscInt cnz=0,*bj_i,*bi,*bj,bnz,nextcj; /* bi,bj,ba: local array of C(mpi mat) */ 543 MPI_Request *s_waits,*r_waits; 544 MPI_Status *status; 545 MatScalar **abuf_r,*ba_i,*pA,*coa,*ba; 546 PetscInt *api,*apj,*coi,*coj; 547 PetscInt *poJ=po->j,*pdJ=pd->j,pcstart=P->cmap->rstart,pcend=P->cmap->rend; 548 PetscInt sparse_axpy; 549 550 PetscFunctionBegin; 551 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 552 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 553 554 ptap = c->ptap; 555 merge = ptap->merge; 556 557 /* 1) get P_oth = ptap->P_oth and P_loc = ptap->P_loc */ 558 /*--------------------------------------------------*/ 559 if (ptap->reuse == MAT_INITIAL_MATRIX){ 560 ptap->reuse = MAT_REUSE_MATRIX; 561 } else { /* update numerical values of P_oth and P_loc */ 562 ierr = MatGetBrowsOfAoCols_MPIAIJ(A,P,MAT_REUSE_MATRIX,&ptap->startsj_s,&ptap->startsj_r,&ptap->bufa,&ptap->P_oth);CHKERRQ(ierr); 563 ierr = MatMPIAIJGetLocalMat(P,MAT_REUSE_MATRIX,&ptap->P_loc);CHKERRQ(ierr); 564 } 565 566 /* 2) compute numeric C_seq = P_loc^T*A_loc*P - dominating part */ 567 /*--------------------------------------------------------------*/ 568 /* get data from symbolic products */ 569 p_loc = (Mat_SeqAIJ*)(ptap->P_loc)->data; 570 p_oth = (Mat_SeqAIJ*)(ptap->P_oth)->data; 571 pi_loc=p_loc->i; pj_loc=p_loc->j; pJ=pj_loc; pa_loc=p_loc->a; 572 pi_oth=p_oth->i; pj_oth=p_oth->j; pa_oth=p_oth->a; 573 574 coi = merge->coi; coj = merge->coj; 575 ierr = PetscMalloc((coi[pon]+1)*sizeof(MatScalar),&coa);CHKERRQ(ierr); 576 ierr = PetscMemzero(coa,coi[pon]*sizeof(MatScalar));CHKERRQ(ierr); 577 578 bi = merge->bi; bj = merge->bj; 579 owners = merge->rowmap->range; 580 ierr = PetscMalloc((bi[cm]+1)*sizeof(MatScalar),&ba);CHKERRQ(ierr); 581 ierr = PetscMemzero(ba,bi[cm]*sizeof(MatScalar));CHKERRQ(ierr); 582 583 api = ptap->api; apj = ptap->apj; 584 /* flag 'sparse_axpy' determines which implementations to be used: 585 0: do dense axpy in MatPtAPNumeric() - fastest, but requires storage of a dense array apa; (default) 586 1: do one sparse axpy - uses same memory as sparse_axpy=0 and might execute less flops 587 (apnz vs. cnz in the outerproduct), slower than case '0' when cnz is not too large than apnz; 588 2: do two sparse axpy in MatPtAPNumeric() - slowest, uses a sparse array apa */ 589 /* set default sparse_axpy */ 590 sparse_axpy = 0; 591 ierr = PetscOptionsGetInt(PETSC_NULL,"-matptap_sparseaxpy",&sparse_axpy,PETSC_NULL);CHKERRQ(ierr); 592 if (sparse_axpy == 0){ /* Do not perform sparse axpy */ 593 /*--------------------------------------------------*/ 594 /* malloc apa to store dense row A[i,:]*P */ 595 ierr = PetscMalloc((P->cmap->N)*sizeof(PetscScalar),&apa);CHKERRQ(ierr); 596 ierr = PetscMemzero(apa,P->cmap->N*sizeof(PetscScalar));CHKERRQ(ierr); 597 598 for (i=0; i<am; i++) { 599 /* 2-a) form i-th sparse row of A_loc*P = Ad*P_loc + Ao*P_oth */ 600 /*------------------------------------------------------------*/ 601 apJ = apj + api[i]; 602 603 /* diagonal portion of A */ 604 anz = adi[i+1] - adi[i]; 605 adj = ad->j + adi[i]; 606 ada = ad->a + adi[i]; 607 for (j=0; j<anz; j++) { 608 row = adj[j]; 609 pnz = pi_loc[row+1] - pi_loc[row]; 610 pj = pj_loc + pi_loc[row]; 611 pa = pa_loc + pi_loc[row]; 612 613 /* perform dense axpy */ 614 valtmp = ada[j]; 615 for (k=0; k<pnz; k++){ 616 apa[pj[k]] += valtmp*pa[k]; 617 } 618 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 619 } 620 621 /* off-diagonal portion of A */ 622 anz = aoi[i+1] - aoi[i]; 623 aoj = ao->j + aoi[i]; 624 aoa = ao->a + aoi[i]; 625 for (j=0; j<anz; j++) { 626 row = aoj[j]; 627 pnz = pi_oth[row+1] - pi_oth[row]; 628 pj = pj_oth + pi_oth[row]; 629 pa = pa_oth + pi_oth[row]; 630 631 /* perform dense axpy */ 632 valtmp = aoa[j]; 633 for (k=0; k<pnz; k++){ 634 apa[pj[k]] += valtmp*pa[k]; 635 } 636 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 637 } 638 639 /* 2-b) Compute Cseq = P_loc[i,:]^T*AP[i,:] using outer product */ 640 /*--------------------------------------------------------------*/ 641 apnz = api[i+1] - api[i]; 642 /* put the value into Co=(p->B)^T*AP (off-diagonal part, send to others) */ 643 pnz = po->i[i+1] - po->i[i]; 644 poJ = po->j + po->i[i]; 645 pA = po->a + po->i[i]; 646 for (j=0; j<pnz; j++){ 647 row = poJ[j]; 648 cnz = coi[row+1] - coi[row]; 649 cj = coj + coi[row]; 650 ca = coa + coi[row]; 651 /* perform dense axpy */ 652 valtmp = pA[j]; 653 for (k=0; k<cnz; k++) { 654 ca[k] += valtmp*apa[cj[k]]; 655 } 656 ierr = PetscLogFlops(2.0*cnz);CHKERRQ(ierr); 657 } 658 659 /* put the value into Cd (diagonal part) */ 660 pnz = pd->i[i+1] - pd->i[i]; 661 pdJ = pd->j + pd->i[i]; 662 pA = pd->a + pd->i[i]; 663 for (j=0; j<pnz; j++){ 664 row = pdJ[j]; 665 cnz = bi[row+1] - bi[row]; 666 cj = bj + bi[row]; 667 ca = ba + bi[row]; 668 /* perform dense axpy */ 669 valtmp = pA[j]; 670 for (k=0; k<cnz; k++) { 671 ca[k] += valtmp*apa[cj[k]]; 672 } 673 ierr = PetscLogFlops(2.0*cnz);CHKERRQ(ierr); 674 } 675 676 /* zero the current row of A*P */ 677 for (k=0; k<apnz; k++) apa[apJ[k]] = 0.0; 678 } 679 } else if (sparse_axpy == 1){ /* Perform one sparse axpy */ 680 /*------------------------------------------------------*/ 681 /* malloc apa to store dense row A[i,:]*P */ 682 ierr = PetscMalloc((P->cmap->N)*sizeof(PetscScalar),&apa);CHKERRQ(ierr); 683 ierr = PetscMemzero(apa,P->cmap->N*sizeof(PetscScalar));CHKERRQ(ierr); 684 685 for (i=0; i<am; i++) { 686 /* 2-a) form i-th sparse row of A_loc*P = Ad*P_loc + Ao*P_oth */ 687 /*------------------------------------------------------------*/ 688 apJ = apj + api[i]; 689 690 /* diagonal portion of A */ 691 anz = adi[i+1] - adi[i]; 692 adj = ad->j + adi[i]; 693 ada = ad->a + adi[i]; 694 for (j=0; j<anz; j++) { 695 row = adj[j]; 696 pnz = pi_loc[row+1] - pi_loc[row]; 697 pj = pj_loc + pi_loc[row]; 698 pa = pa_loc + pi_loc[row]; 699 700 /* perform dense axpy */ 701 valtmp = ada[j]; 702 for (k=0; k<pnz; k++){ 703 apa[pj[k]] += valtmp*pa[k]; 704 } 705 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 706 } 707 708 /* off-diagonal portion of A */ 709 anz = aoi[i+1] - aoi[i]; 710 aoj = ao->j + aoi[i]; 711 aoa = ao->a + aoi[i]; 712 for (j=0; j<anz; j++) { 713 row = aoj[j]; 714 pnz = pi_oth[row+1] - pi_oth[row]; 715 pj = pj_oth + pi_oth[row]; 716 pa = pa_oth + pi_oth[row]; 717 718 /* perform dense axpy */ 719 valtmp = aoa[j]; 720 for (k=0; k<pnz; k++){ 721 apa[pj[k]] += valtmp*pa[k]; 722 } 723 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 724 } 725 726 /* 2-b) Compute Cseq = P_loc[i,:]^T*AP[i,:] using outer product */ 727 /*--------------------------------------------------------------*/ 728 apnz = api[i+1] - api[i]; 729 /* put the value into Co=(p->B)^T*AP (off-diagonal part, send to others) */ 730 pnz = po->i[i+1] - po->i[i]; 731 poJ = po->j + po->i[i]; 732 pA = po->a + po->i[i]; 733 for (j=0; j<pnz; j++){ 734 row = poJ[j]; 735 cj = coj + coi[row]; 736 ca = coa + coi[row]; 737 valtmp = pA[j]; 738 /* perform sparse axpy */ 739 nextap = 0; 740 for (k=0; nextap<apnz; k++) { 741 if (cj[k]==apJ[nextap]) { /* global column index */ 742 ca[k] += valtmp*apa[cj[k]]; nextap++; 743 } 744 } 745 ierr = PetscLogFlops(2.0*apnz);CHKERRQ(ierr); 746 } 747 748 /* put the value into Cd (diagonal part) */ 749 pnz = pd->i[i+1] - pd->i[i]; 750 pdJ = pd->j + pd->i[i]; 751 pA = pd->a + pd->i[i]; 752 for (j=0; j<pnz; j++){ 753 row = pdJ[j]; 754 cj = bj + bi[row]; 755 ca = ba + bi[row]; 756 valtmp = pA[j]; 757 /* perform sparse axpy */ 758 nextap = 0; 759 for (k=0; nextap<apnz; k++) { 760 if (cj[k]==apJ[nextap]) { /* global column index */ 761 ca[k] += valtmp*apa[cj[k]]; 762 nextap++; 763 } 764 } 765 ierr = PetscLogFlops(2.0*apnz);CHKERRQ(ierr); 766 } 767 768 /* zero the current row of A*P */ 769 for (k=0; k<apnz; k++) apa[apJ[k]] = 0.0; 770 } 771 } else if (sparse_axpy == 2){/* Perform two sparse axpy */ 772 /*----------------------------------------------------*/ 773 /* malloc apa to store sparse row A[i,:]*P */ 774 ierr = PetscMalloc((ptap->rmax+1)*sizeof(MatScalar),&apa);CHKERRQ(ierr); 775 ierr = PetscMemzero(apa,ptap->rmax*sizeof(MatScalar));CHKERRQ(ierr); 776 777 pA=pa_loc; 778 for (i=0; i<am; i++) { 779 /* form i-th sparse row of A*P */ 780 apnz = api[i+1] - api[i]; 781 apJ = apj + api[i]; 782 /* diagonal portion of A */ 783 anz = adi[i+1] - adi[i]; 784 adj = ad->j + adi[i]; 785 ada = ad->a + adi[i]; 786 for (j=0; j<anz; j++) { 787 row = adj[j]; 788 pnz = pi_loc[row+1] - pi_loc[row]; 789 pj = pj_loc + pi_loc[row]; 790 pa = pa_loc + pi_loc[row]; 791 valtmp = ada[j]; 792 nextp = 0; 793 for (k=0; nextp<pnz; k++) { 794 if (apJ[k] == pj[nextp]) { /* col of AP == col of P */ 795 apa[k] += valtmp*pa[nextp++]; 796 } 797 } 798 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 799 } 800 /* off-diagonal portion of A */ 801 anz = aoi[i+1] - aoi[i]; 802 aoj = ao->j + aoi[i]; 803 aoa = ao->a + aoi[i]; 804 for (j=0; j<anz; j++) { 805 row = aoj[j]; 806 pnz = pi_oth[row+1] - pi_oth[row]; 807 pj = pj_oth + pi_oth[row]; 808 pa = pa_oth + pi_oth[row]; 809 valtmp = aoa[j]; 810 nextp = 0; 811 for (k=0; nextp<pnz; k++) { 812 if (apJ[k] == pj[nextp]) { /* col of AP == col of P */ 813 apa[k] += valtmp*pa[nextp++]; 814 } 815 } 816 ierr = PetscLogFlops(2.0*pnz);CHKERRQ(ierr); 817 } 818 819 /* 2-b) Compute Cseq = P_loc[i,:]^T*AP[i,:] using outer product */ 820 /*--------------------------------------------------------------*/ 821 pnz = pi_loc[i+1] - pi_loc[i]; 822 pJ = pj_loc + pi_loc[i]; 823 for (j=0; j<pnz; j++) { 824 nextap = 0; 825 row = pJ[j]; /* global index */ 826 if (row < pcstart || row >=pcend) { /* put the value into Co */ 827 row = *poJ; 828 cj = coj + coi[row]; 829 ca = coa + coi[row]; poJ++; 830 } else { /* put the value into Cd */ 831 row = *pdJ; 832 cj = bj + bi[row]; 833 ca = ba + bi[row]; pdJ++; 834 } 835 valtmp = pA[j]; 836 for (k=0; nextap<apnz; k++) { 837 if (cj[k]==apJ[nextap]) ca[k] += valtmp*apa[nextap++]; 838 } 839 ierr = PetscLogFlops(2.0*apnz);CHKERRQ(ierr); 840 } 841 pA += pnz; 842 /* zero the current row info for A*P */ 843 ierr = PetscMemzero(apa,apnz*sizeof(MatScalar));CHKERRQ(ierr); 844 } 845 } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"sparse_axpy only takes values 0, 1 and 2"); 846 ierr = PetscFree(apa);CHKERRQ(ierr); 847 848 /* 3) send and recv matrix values coa */ 849 /*------------------------------------*/ 850 buf_ri = merge->buf_ri; 851 buf_rj = merge->buf_rj; 852 len_s = merge->len_s; 853 ierr = PetscCommGetNewTag(comm,&taga);CHKERRQ(ierr); 854 ierr = PetscPostIrecvScalar(comm,taga,merge->nrecv,merge->id_r,merge->len_r,&abuf_r,&r_waits);CHKERRQ(ierr); 855 856 ierr = PetscMalloc2(merge->nsend+1,MPI_Request,&s_waits,size,MPI_Status,&status);CHKERRQ(ierr); 857 for (proc=0,k=0; proc<size; proc++){ 858 if (!len_s[proc]) continue; 859 i = merge->owners_co[proc]; 860 ierr = MPI_Isend(coa+coi[i],len_s[proc],MPIU_MATSCALAR,proc,taga,comm,s_waits+k);CHKERRQ(ierr); 861 k++; 862 } 863 if (merge->nrecv) {ierr = MPI_Waitall(merge->nrecv,r_waits,status);CHKERRQ(ierr);} 864 if (merge->nsend) {ierr = MPI_Waitall(merge->nsend,s_waits,status);CHKERRQ(ierr);} 865 866 ierr = PetscFree2(s_waits,status);CHKERRQ(ierr); 867 ierr = PetscFree(r_waits);CHKERRQ(ierr); 868 ierr = PetscFree(coa);CHKERRQ(ierr); 869 870 /* 4) insert local Cseq and received values into Cmpi */ 871 /*------------------------------------------------------*/ 872 ierr = PetscMalloc3(merge->nrecv,PetscInt**,&buf_ri_k,merge->nrecv,PetscInt*,&nextrow,merge->nrecv,PetscInt*,&nextci);CHKERRQ(ierr); 873 for (k=0; k<merge->nrecv; k++){ 874 buf_ri_k[k] = buf_ri[k]; /* beginning of k-th recved i-structure */ 875 nrows = *(buf_ri_k[k]); 876 nextrow[k] = buf_ri_k[k]+1; /* next row number of k-th recved i-structure */ 877 nextci[k] = buf_ri_k[k] + (nrows + 1);/* poins to the next i-structure of k-th recved i-structure */ 878 } 879 880 for (i=0; i<cm; i++) { 881 row = owners[rank] + i; /* global row index of C_seq */ 882 bj_i = bj + bi[i]; /* col indices of the i-th row of C */ 883 ba_i = ba + bi[i]; 884 bnz = bi[i+1] - bi[i]; 885 /* add received vals into ba */ 886 for (k=0; k<merge->nrecv; k++){ /* k-th received message */ 887 /* i-th row */ 888 if (i == *nextrow[k]) { 889 cnz = *(nextci[k]+1) - *nextci[k]; 890 cj = buf_rj[k] + *(nextci[k]); 891 ca = abuf_r[k] + *(nextci[k]); 892 nextcj = 0; 893 for (j=0; nextcj<cnz; j++){ 894 if (bj_i[j] == cj[nextcj]){ /* bcol == ccol */ 895 ba_i[j] += ca[nextcj++]; 896 } 897 } 898 nextrow[k]++; nextci[k]++; 899 ierr = PetscLogFlops(2.0*cnz);CHKERRQ(ierr); 900 } 901 } 902 ierr = MatSetValues(C,1,&row,bnz,bj_i,ba_i,INSERT_VALUES);CHKERRQ(ierr); 903 } 904 ierr = MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 905 ierr = MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 906 907 ierr = PetscFree(ba);CHKERRQ(ierr); 908 ierr = PetscFree(abuf_r[0]);CHKERRQ(ierr); 909 ierr = PetscFree(abuf_r);CHKERRQ(ierr); 910 ierr = PetscFree3(buf_ri_k,nextrow,nextci);CHKERRQ(ierr); 911 PetscFunctionReturn(0); 912 } 913