1 /* 2 Routines to compute overlapping regions of a parallel MPI matrix 3 and to find submatrices that were shared across processors. 4 */ 5 #include <../src/mat/impls/aij/seq/aij.h> 6 #include <../src/mat/impls/aij/mpi/mpiaij.h> 7 #include <petscbt.h> 8 #include <petscsf.h> 9 10 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once(Mat,PetscInt,IS*); 11 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local(Mat,PetscInt,char**,PetscInt*,PetscInt**,PetscTable*); 12 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Receive(Mat,PetscInt,PetscInt**,PetscInt**,PetscInt*); 13 extern PetscErrorCode MatGetRow_MPIAIJ(Mat,PetscInt,PetscInt*,PetscInt**,PetscScalar**); 14 extern PetscErrorCode MatRestoreRow_MPIAIJ(Mat,PetscInt,PetscInt*,PetscInt**,PetscScalar**); 15 16 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once_Scalable(Mat,PetscInt,IS*); 17 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local_Scalable(Mat,PetscInt,IS*); 18 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Send_Scalable(Mat,PetscInt,PetscMPIInt,PetscMPIInt *,PetscInt *, PetscInt *,PetscInt **,PetscInt **); 19 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Receive_Scalable(Mat,PetscInt,IS*,PetscInt,PetscInt *); 20 21 22 PetscErrorCode MatIncreaseOverlap_MPIAIJ(Mat C,PetscInt imax,IS is[],PetscInt ov) 23 { 24 PetscErrorCode ierr; 25 PetscInt i; 26 27 PetscFunctionBegin; 28 if (ov < 0) SETERRQ(PetscObjectComm((PetscObject)C),PETSC_ERR_ARG_OUTOFRANGE,"Negative overlap specified"); 29 for (i=0; i<ov; ++i) { 30 ierr = MatIncreaseOverlap_MPIAIJ_Once(C,imax,is);CHKERRQ(ierr); 31 } 32 PetscFunctionReturn(0); 33 } 34 35 PetscErrorCode MatIncreaseOverlap_MPIAIJ_Scalable(Mat C,PetscInt imax,IS is[],PetscInt ov) 36 { 37 PetscErrorCode ierr; 38 PetscInt i; 39 40 PetscFunctionBegin; 41 if (ov < 0) SETERRQ(PetscObjectComm((PetscObject)C),PETSC_ERR_ARG_OUTOFRANGE,"Negative overlap specified"); 42 for (i=0; i<ov; ++i) { 43 ierr = MatIncreaseOverlap_MPIAIJ_Once_Scalable(C,imax,is);CHKERRQ(ierr); 44 } 45 PetscFunctionReturn(0); 46 } 47 48 49 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once_Scalable(Mat mat,PetscInt nidx,IS is[]) 50 { 51 PetscErrorCode ierr; 52 MPI_Comm comm; 53 PetscInt *length,length_i,tlength,*remoterows,nrrows,reducednrrows,*rrow_ranks,*rrow_isids,i,j,owner; 54 PetscInt *tosizes,*tosizes_temp,*toffsets,*fromsizes,*todata,*fromdata; 55 PetscInt nrecvrows,*sbsizes = 0,*sbdata = 0; 56 const PetscInt *indices_i,**indices; 57 PetscLayout rmap; 58 PetscMPIInt rank,size,*toranks,*fromranks,nto,nfrom; 59 PetscSF sf; 60 PetscSFNode *remote; 61 62 PetscFunctionBegin; 63 ierr = PetscObjectGetComm((PetscObject)mat,&comm);CHKERRQ(ierr); 64 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 65 ierr = MPI_Comm_size(comm,&size);CHKERRQ(ierr); 66 /* get row map to determine where rows should be going */ 67 ierr = MatGetLayouts(mat,&rmap,NULL);CHKERRQ(ierr); 68 /* retrieve IS data and put all together so that we 69 * can optimize communication 70 * */ 71 ierr = PetscCalloc2(nidx,(PetscInt ***)&indices,nidx,&length);CHKERRQ(ierr); 72 for (i=0,tlength=0; i<nidx; i++){ 73 ierr = ISGetLocalSize(is[i],&length[i]);CHKERRQ(ierr); 74 tlength += length[i]; 75 ierr = ISGetIndices(is[i],&indices[i]);CHKERRQ(ierr); 76 } 77 /* find these rows on remote processors */ 78 ierr = PetscCalloc3(tlength,&remoterows,tlength,&rrow_ranks,tlength,&rrow_isids);CHKERRQ(ierr); 79 ierr = PetscCalloc3(size,&toranks,2*size,&tosizes,size,&tosizes_temp);CHKERRQ(ierr); 80 nrrows = 0; 81 for (i=0; i<nidx; i++){ 82 length_i = length[i]; 83 indices_i = indices[i]; 84 for (j=0; j<length_i; j++){ 85 owner = -1; 86 ierr = PetscLayoutFindOwner(rmap,indices_i[j],&owner);CHKERRQ(ierr); 87 /* remote processors */ 88 if (owner != rank){ 89 tosizes_temp[owner]++; /* number of rows to owner */ 90 rrow_ranks[nrrows] = owner; /* processor */ 91 rrow_isids[nrrows] = i; /* is id */ 92 remoterows[nrrows++] = indices_i[j]; /* row */ 93 } 94 } 95 ierr = ISRestoreIndices(is[i],&indices[i]);CHKERRQ(ierr); 96 } 97 ierr = PetscFree2(indices,length);CHKERRQ(ierr); 98 /* test if we need to exchange messages 99 * generally speaking, we do not need to exchange 100 * data when overlap is 1 101 * */ 102 ierr = MPIU_Allreduce(&nrrows,&reducednrrows,1,MPIU_INT,MPIU_MAX,comm);CHKERRQ(ierr); 103 /* we do not have any messages 104 * It usually corresponds to overlap 1 105 * */ 106 if (!reducednrrows){ 107 ierr = PetscFree3(toranks,tosizes,tosizes_temp);CHKERRQ(ierr); 108 ierr = PetscFree3(remoterows,rrow_ranks,rrow_isids);CHKERRQ(ierr); 109 ierr = MatIncreaseOverlap_MPIAIJ_Local_Scalable(mat,nidx,is);CHKERRQ(ierr); 110 PetscFunctionReturn(0); 111 } 112 nto = 0; 113 /* send sizes and ranks for building a two-sided communcation */ 114 for (i=0; i<size; i++){ 115 if (tosizes_temp[i]){ 116 tosizes[nto*2] = tosizes_temp[i]*2; /* size */ 117 tosizes_temp[i] = nto; /* a map from processor to index */ 118 toranks[nto++] = i; /* processor */ 119 } 120 } 121 ierr = PetscCalloc1(nto+1,&toffsets);CHKERRQ(ierr); 122 for (i=0; i<nto; i++){ 123 toffsets[i+1] = toffsets[i]+tosizes[2*i]; /* offsets */ 124 tosizes[2*i+1] = toffsets[i]; /* offsets to send */ 125 } 126 /* send information to other processors */ 127 ierr = PetscCommBuildTwoSided(comm,2,MPIU_INT,nto,toranks,tosizes,&nfrom,&fromranks,&fromsizes);CHKERRQ(ierr); 128 nrecvrows = 0; 129 for (i=0; i<nfrom; i++) nrecvrows += fromsizes[2*i]; 130 ierr = PetscMalloc1(nrecvrows,&remote);CHKERRQ(ierr); 131 nrecvrows = 0; 132 for (i=0; i<nfrom; i++){ 133 for (j=0; j<fromsizes[2*i]; j++){ 134 remote[nrecvrows].rank = fromranks[i]; 135 remote[nrecvrows++].index = fromsizes[2*i+1]+j; 136 } 137 } 138 ierr = PetscSFCreate(comm,&sf);CHKERRQ(ierr); 139 ierr = PetscSFSetGraph(sf,nrecvrows,nrecvrows,NULL,PETSC_OWN_POINTER,remote,PETSC_OWN_POINTER);CHKERRQ(ierr); 140 /* use two-sided communication by default since OPENMPI has some bugs for one-sided one */ 141 ierr = PetscSFSetType(sf,PETSCSFBASIC);CHKERRQ(ierr); 142 ierr = PetscSFSetFromOptions(sf);CHKERRQ(ierr); 143 /* message pair <no of is, row> */ 144 ierr = PetscCalloc2(2*nrrows,&todata,nrecvrows,&fromdata);CHKERRQ(ierr); 145 for (i=0; i<nrrows; i++){ 146 owner = rrow_ranks[i]; /* processor */ 147 j = tosizes_temp[owner]; /* index */ 148 todata[toffsets[j]++] = rrow_isids[i]; 149 todata[toffsets[j]++] = remoterows[i]; 150 } 151 ierr = PetscFree3(toranks,tosizes,tosizes_temp);CHKERRQ(ierr); 152 ierr = PetscFree3(remoterows,rrow_ranks,rrow_isids);CHKERRQ(ierr); 153 ierr = PetscFree(toffsets);CHKERRQ(ierr); 154 ierr = PetscSFBcastBegin(sf,MPIU_INT,todata,fromdata);CHKERRQ(ierr); 155 ierr = PetscSFBcastEnd(sf,MPIU_INT,todata,fromdata);CHKERRQ(ierr); 156 ierr = PetscSFDestroy(&sf);CHKERRQ(ierr); 157 /* send rows belonging to the remote so that then we could get the overlapping data back */ 158 ierr = MatIncreaseOverlap_MPIAIJ_Send_Scalable(mat,nidx,nfrom,fromranks,fromsizes,fromdata,&sbsizes,&sbdata);CHKERRQ(ierr); 159 ierr = PetscFree2(todata,fromdata);CHKERRQ(ierr); 160 ierr = PetscFree(fromsizes);CHKERRQ(ierr); 161 ierr = PetscCommBuildTwoSided(comm,2,MPIU_INT,nfrom,fromranks,sbsizes,&nto,&toranks,&tosizes);CHKERRQ(ierr); 162 ierr = PetscFree(fromranks);CHKERRQ(ierr); 163 nrecvrows = 0; 164 for (i=0; i<nto; i++) nrecvrows += tosizes[2*i]; 165 ierr = PetscCalloc1(nrecvrows,&todata);CHKERRQ(ierr); 166 ierr = PetscMalloc1(nrecvrows,&remote);CHKERRQ(ierr); 167 nrecvrows = 0; 168 for (i=0; i<nto; i++){ 169 for (j=0; j<tosizes[2*i]; j++){ 170 remote[nrecvrows].rank = toranks[i]; 171 remote[nrecvrows++].index = tosizes[2*i+1]+j; 172 } 173 } 174 ierr = PetscSFCreate(comm,&sf);CHKERRQ(ierr); 175 ierr = PetscSFSetGraph(sf,nrecvrows,nrecvrows,NULL,PETSC_OWN_POINTER,remote,PETSC_OWN_POINTER);CHKERRQ(ierr); 176 /* use two-sided communication by default since OPENMPI has some bugs for one-sided one */ 177 ierr = PetscSFSetType(sf,PETSCSFBASIC);CHKERRQ(ierr); 178 ierr = PetscSFSetFromOptions(sf);CHKERRQ(ierr); 179 /* overlap communication and computation */ 180 ierr = PetscSFBcastBegin(sf,MPIU_INT,sbdata,todata);CHKERRQ(ierr); 181 ierr = MatIncreaseOverlap_MPIAIJ_Local_Scalable(mat,nidx,is);CHKERRQ(ierr); 182 ierr = PetscSFBcastEnd(sf,MPIU_INT,sbdata,todata);CHKERRQ(ierr); 183 ierr = PetscSFDestroy(&sf);CHKERRQ(ierr); 184 ierr = PetscFree2(sbdata,sbsizes);CHKERRQ(ierr); 185 ierr = MatIncreaseOverlap_MPIAIJ_Receive_Scalable(mat,nidx,is,nrecvrows,todata);CHKERRQ(ierr); 186 ierr = PetscFree(toranks);CHKERRQ(ierr); 187 ierr = PetscFree(tosizes);CHKERRQ(ierr); 188 ierr = PetscFree(todata);CHKERRQ(ierr); 189 PetscFunctionReturn(0); 190 } 191 192 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Receive_Scalable(Mat mat,PetscInt nidx, IS is[], PetscInt nrecvs, PetscInt *recvdata) 193 { 194 PetscInt *isz,isz_i,i,j,is_id, data_size; 195 PetscInt col,lsize,max_lsize,*indices_temp, *indices_i; 196 const PetscInt *indices_i_temp; 197 PetscErrorCode ierr; 198 199 PetscFunctionBegin; 200 max_lsize = 0; 201 ierr = PetscMalloc1(nidx,&isz);CHKERRQ(ierr); 202 for (i=0; i<nidx; i++){ 203 ierr = ISGetLocalSize(is[i],&lsize);CHKERRQ(ierr); 204 max_lsize = lsize>max_lsize ? lsize:max_lsize; 205 isz[i] = lsize; 206 } 207 ierr = PetscMalloc1((max_lsize+nrecvs)*nidx,&indices_temp);CHKERRQ(ierr); 208 for (i=0; i<nidx; i++){ 209 ierr = ISGetIndices(is[i],&indices_i_temp);CHKERRQ(ierr); 210 ierr = PetscMemcpy(indices_temp+i*(max_lsize+nrecvs),indices_i_temp, sizeof(PetscInt)*isz[i]);CHKERRQ(ierr); 211 ierr = ISRestoreIndices(is[i],&indices_i_temp);CHKERRQ(ierr); 212 ierr = ISDestroy(&is[i]);CHKERRQ(ierr); 213 } 214 /* retrieve information to get row id and its overlap */ 215 for (i=0; i<nrecvs; ){ 216 is_id = recvdata[i++]; 217 data_size = recvdata[i++]; 218 indices_i = indices_temp+(max_lsize+nrecvs)*is_id; 219 isz_i = isz[is_id]; 220 for (j=0; j< data_size; j++){ 221 col = recvdata[i++]; 222 indices_i[isz_i++] = col; 223 } 224 isz[is_id] = isz_i; 225 } 226 /* remove duplicate entities */ 227 for (i=0; i<nidx; i++){ 228 indices_i = indices_temp+(max_lsize+nrecvs)*i; 229 isz_i = isz[i]; 230 ierr = PetscSortRemoveDupsInt(&isz_i,indices_i);CHKERRQ(ierr); 231 ierr = ISCreateGeneral(PETSC_COMM_SELF,isz_i,indices_i,PETSC_COPY_VALUES,&is[i]);CHKERRQ(ierr); 232 } 233 ierr = PetscFree(isz);CHKERRQ(ierr); 234 ierr = PetscFree(indices_temp);CHKERRQ(ierr); 235 PetscFunctionReturn(0); 236 } 237 238 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Send_Scalable(Mat mat,PetscInt nidx, PetscMPIInt nfrom,PetscMPIInt *fromranks,PetscInt *fromsizes, PetscInt *fromrows, PetscInt **sbrowsizes, PetscInt **sbrows) 239 { 240 PetscLayout rmap,cmap; 241 PetscInt i,j,k,l,*rows_i,*rows_data_ptr,**rows_data,max_fszs,rows_pos,*rows_pos_i; 242 PetscInt is_id,tnz,an,bn,rstart,cstart,row,start,end,col,totalrows,*sbdata; 243 PetscInt *indv_counts,indvc_ij,*sbsizes,*indices_tmp,*offsets; 244 const PetscInt *gcols,*ai,*aj,*bi,*bj; 245 Mat amat,bmat; 246 PetscMPIInt rank; 247 PetscBool done; 248 MPI_Comm comm; 249 PetscErrorCode ierr; 250 251 PetscFunctionBegin; 252 ierr = PetscObjectGetComm((PetscObject)mat,&comm);CHKERRQ(ierr); 253 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 254 ierr = MatMPIAIJGetSeqAIJ(mat,&amat,&bmat,&gcols);CHKERRQ(ierr); 255 /* Even if the mat is symmetric, we still assume it is not symmetric */ 256 ierr = MatGetRowIJ(amat,0,PETSC_FALSE,PETSC_FALSE,&an,&ai,&aj,&done);CHKERRQ(ierr); 257 if (!done) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"can not get row IJ \n"); 258 ierr = MatGetRowIJ(bmat,0,PETSC_FALSE,PETSC_FALSE,&bn,&bi,&bj,&done);CHKERRQ(ierr); 259 if (!done) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"can not get row IJ \n"); 260 /* total number of nonzero values is used to estimate the memory usage in the next step */ 261 tnz = ai[an]+bi[bn]; 262 ierr = MatGetLayouts(mat,&rmap,&cmap);CHKERRQ(ierr); 263 ierr = PetscLayoutGetRange(rmap,&rstart,NULL);CHKERRQ(ierr); 264 ierr = PetscLayoutGetRange(cmap,&cstart,NULL);CHKERRQ(ierr); 265 /* to find the longest message */ 266 max_fszs = 0; 267 for (i=0; i<nfrom; i++) max_fszs = fromsizes[2*i]>max_fszs ? fromsizes[2*i]:max_fszs; 268 /* better way to estimate number of nonzero in the mat??? */ 269 ierr = PetscCalloc5(max_fszs*nidx,&rows_data_ptr,nidx,&rows_data,nidx,&rows_pos_i,nfrom*nidx,&indv_counts,tnz,&indices_tmp);CHKERRQ(ierr); 270 for (i=0; i<nidx; i++) rows_data[i] = rows_data_ptr+max_fszs*i; 271 rows_pos = 0; 272 totalrows = 0; 273 for (i=0; i<nfrom; i++){ 274 ierr = PetscMemzero(rows_pos_i,sizeof(PetscInt)*nidx);CHKERRQ(ierr); 275 /* group data together */ 276 for (j=0; j<fromsizes[2*i]; j+=2){ 277 is_id = fromrows[rows_pos++];/* no of is */ 278 rows_i = rows_data[is_id]; 279 rows_i[rows_pos_i[is_id]++] = fromrows[rows_pos++];/* row */ 280 } 281 /* estimate a space to avoid multiple allocations */ 282 for (j=0; j<nidx; j++){ 283 indvc_ij = 0; 284 rows_i = rows_data[j]; 285 for (l=0; l<rows_pos_i[j]; l++){ 286 row = rows_i[l]-rstart; 287 start = ai[row]; 288 end = ai[row+1]; 289 for (k=start; k<end; k++){ /* Amat */ 290 col = aj[k] + cstart; 291 indices_tmp[indvc_ij++] = col;/* do not count the rows from the original rank */ 292 } 293 start = bi[row]; 294 end = bi[row+1]; 295 for (k=start; k<end; k++) { /* Bmat */ 296 col = gcols[bj[k]]; 297 indices_tmp[indvc_ij++] = col; 298 } 299 } 300 ierr = PetscSortRemoveDupsInt(&indvc_ij,indices_tmp);CHKERRQ(ierr); 301 indv_counts[i*nidx+j] = indvc_ij; 302 totalrows += indvc_ij; 303 } 304 } 305 /* message triple <no of is, number of rows, rows> */ 306 ierr = PetscCalloc2(totalrows+nidx*nfrom*2,&sbdata,2*nfrom,&sbsizes);CHKERRQ(ierr); 307 totalrows = 0; 308 rows_pos = 0; 309 /* use this code again */ 310 for (i=0;i<nfrom;i++){ 311 ierr = PetscMemzero(rows_pos_i,sizeof(PetscInt)*nidx);CHKERRQ(ierr); 312 for (j=0; j<fromsizes[2*i]; j+=2){ 313 is_id = fromrows[rows_pos++]; 314 rows_i = rows_data[is_id]; 315 rows_i[rows_pos_i[is_id]++] = fromrows[rows_pos++]; 316 } 317 /* add data */ 318 for (j=0; j<nidx; j++){ 319 if (!indv_counts[i*nidx+j]) continue; 320 indvc_ij = 0; 321 sbdata[totalrows++] = j; 322 sbdata[totalrows++] = indv_counts[i*nidx+j]; 323 sbsizes[2*i] += 2; 324 rows_i = rows_data[j]; 325 for (l=0; l<rows_pos_i[j]; l++){ 326 row = rows_i[l]-rstart; 327 start = ai[row]; 328 end = ai[row+1]; 329 for (k=start; k<end; k++){ /* Amat */ 330 col = aj[k] + cstart; 331 indices_tmp[indvc_ij++] = col; 332 } 333 start = bi[row]; 334 end = bi[row+1]; 335 for (k=start; k<end; k++) { /* Bmat */ 336 col = gcols[bj[k]]; 337 indices_tmp[indvc_ij++] = col; 338 } 339 } 340 ierr = PetscSortRemoveDupsInt(&indvc_ij,indices_tmp);CHKERRQ(ierr); 341 sbsizes[2*i] += indvc_ij; 342 ierr = PetscMemcpy(sbdata+totalrows,indices_tmp,sizeof(PetscInt)*indvc_ij);CHKERRQ(ierr); 343 totalrows += indvc_ij; 344 } 345 } 346 ierr = PetscCalloc1(nfrom+1,&offsets);CHKERRQ(ierr); 347 for (i=0; i<nfrom; i++){ 348 offsets[i+1] = offsets[i] + sbsizes[2*i]; 349 sbsizes[2*i+1] = offsets[i]; 350 } 351 ierr = PetscFree(offsets);CHKERRQ(ierr); 352 if (sbrowsizes) *sbrowsizes = sbsizes; 353 if (sbrows) *sbrows = sbdata; 354 ierr = PetscFree5(rows_data_ptr,rows_data,rows_pos_i,indv_counts,indices_tmp);CHKERRQ(ierr); 355 ierr = MatRestoreRowIJ(amat,0,PETSC_FALSE,PETSC_FALSE,&an,&ai,&aj,&done);CHKERRQ(ierr); 356 ierr = MatRestoreRowIJ(bmat,0,PETSC_FALSE,PETSC_FALSE,&bn,&bi,&bj,&done);CHKERRQ(ierr); 357 PetscFunctionReturn(0); 358 } 359 360 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local_Scalable(Mat mat,PetscInt nidx, IS is[]) 361 { 362 const PetscInt *gcols,*ai,*aj,*bi,*bj, *indices; 363 PetscInt tnz,an,bn,i,j,row,start,end,rstart,cstart,col,k,*indices_temp; 364 PetscInt lsize,lsize_tmp,owner; 365 PetscMPIInt rank; 366 Mat amat,bmat; 367 PetscBool done; 368 PetscLayout cmap,rmap; 369 MPI_Comm comm; 370 PetscErrorCode ierr; 371 372 PetscFunctionBegin; 373 ierr = PetscObjectGetComm((PetscObject)mat,&comm);CHKERRQ(ierr); 374 ierr = MPI_Comm_rank(comm,&rank);CHKERRQ(ierr); 375 ierr = MatMPIAIJGetSeqAIJ(mat,&amat,&bmat,&gcols);CHKERRQ(ierr); 376 ierr = MatGetRowIJ(amat,0,PETSC_FALSE,PETSC_FALSE,&an,&ai,&aj,&done);CHKERRQ(ierr); 377 if (!done) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"can not get row IJ \n"); 378 ierr = MatGetRowIJ(bmat,0,PETSC_FALSE,PETSC_FALSE,&bn,&bi,&bj,&done);CHKERRQ(ierr); 379 if (!done) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"can not get row IJ \n"); 380 /* is it a safe way to compute number of nonzero values ? */ 381 tnz = ai[an]+bi[bn]; 382 ierr = MatGetLayouts(mat,&rmap,&cmap);CHKERRQ(ierr); 383 ierr = PetscLayoutGetRange(rmap,&rstart,NULL);CHKERRQ(ierr); 384 ierr = PetscLayoutGetRange(cmap,&cstart,NULL);CHKERRQ(ierr); 385 /* it is a better way to estimate memory than the old implementation 386 * where global size of matrix is used 387 * */ 388 ierr = PetscMalloc1(tnz,&indices_temp);CHKERRQ(ierr); 389 for (i=0; i<nidx; i++) { 390 ierr = ISGetLocalSize(is[i],&lsize);CHKERRQ(ierr); 391 ierr = ISGetIndices(is[i],&indices);CHKERRQ(ierr); 392 lsize_tmp = 0; 393 for (j=0; j<lsize; j++) { 394 owner = -1; 395 row = indices[j]; 396 ierr = PetscLayoutFindOwner(rmap,row,&owner);CHKERRQ(ierr); 397 if (owner != rank) continue; 398 /* local number */ 399 row -= rstart; 400 start = ai[row]; 401 end = ai[row+1]; 402 for (k=start; k<end; k++) { /* Amat */ 403 col = aj[k] + cstart; 404 indices_temp[lsize_tmp++] = col; 405 } 406 start = bi[row]; 407 end = bi[row+1]; 408 for (k=start; k<end; k++) { /* Bmat */ 409 col = gcols[bj[k]]; 410 indices_temp[lsize_tmp++] = col; 411 } 412 } 413 ierr = ISRestoreIndices(is[i],&indices);CHKERRQ(ierr); 414 ierr = ISDestroy(&is[i]);CHKERRQ(ierr); 415 ierr = PetscSortRemoveDupsInt(&lsize_tmp,indices_temp);CHKERRQ(ierr); 416 ierr = ISCreateGeneral(PETSC_COMM_SELF,lsize_tmp,indices_temp,PETSC_COPY_VALUES,&is[i]);CHKERRQ(ierr); 417 } 418 ierr = PetscFree(indices_temp);CHKERRQ(ierr); 419 ierr = MatRestoreRowIJ(amat,0,PETSC_FALSE,PETSC_FALSE,&an,&ai,&aj,&done);CHKERRQ(ierr); 420 ierr = MatRestoreRowIJ(bmat,0,PETSC_FALSE,PETSC_FALSE,&bn,&bi,&bj,&done);CHKERRQ(ierr); 421 PetscFunctionReturn(0); 422 } 423 424 425 /* 426 Sample message format: 427 If a processor A wants processor B to process some elements corresponding 428 to index sets is[1],is[5] 429 mesg [0] = 2 (no of index sets in the mesg) 430 ----------- 431 mesg [1] = 1 => is[1] 432 mesg [2] = sizeof(is[1]); 433 ----------- 434 mesg [3] = 5 => is[5] 435 mesg [4] = sizeof(is[5]); 436 ----------- 437 mesg [5] 438 mesg [n] datas[1] 439 ----------- 440 mesg[n+1] 441 mesg[m] data(is[5]) 442 ----------- 443 444 Notes: 445 nrqs - no of requests sent (or to be sent out) 446 nrqr - no of requests recieved (which have to be or which have been processed 447 */ 448 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once(Mat C,PetscInt imax,IS is[]) 449 { 450 Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data; 451 PetscMPIInt *w1,*w2,nrqr,*w3,*w4,*onodes1,*olengths1,*onodes2,*olengths2; 452 const PetscInt **idx,*idx_i; 453 PetscInt *n,**data,len; 454 #if defined(PETSC_USE_CTABLE) 455 PetscTable *table_data,table_data_i; 456 PetscInt *tdata,tcount,tcount_max; 457 #else 458 PetscInt *data_i,*d_p; 459 #endif 460 PetscErrorCode ierr; 461 PetscMPIInt size,rank,tag1,tag2; 462 PetscInt M,i,j,k,**rbuf,row,proc = 0,nrqs,msz,**outdat,**ptr; 463 PetscInt *ctr,*pa,*tmp,*isz,*isz1,**xdata,**rbuf2; 464 PetscBT *table; 465 MPI_Comm comm; 466 MPI_Request *s_waits1,*r_waits1,*s_waits2,*r_waits2; 467 MPI_Status *s_status,*recv_status; 468 char *t_p; 469 470 PetscFunctionBegin; 471 ierr = PetscObjectGetComm((PetscObject)C,&comm);CHKERRQ(ierr); 472 size = c->size; 473 rank = c->rank; 474 M = C->rmap->N; 475 476 ierr = PetscObjectGetNewTag((PetscObject)C,&tag1);CHKERRQ(ierr); 477 ierr = PetscObjectGetNewTag((PetscObject)C,&tag2);CHKERRQ(ierr); 478 479 ierr = PetscMalloc2(imax,&idx,imax,&n);CHKERRQ(ierr); 480 481 for (i=0; i<imax; i++) { 482 ierr = ISGetIndices(is[i],&idx[i]);CHKERRQ(ierr); 483 ierr = ISGetLocalSize(is[i],&n[i]);CHKERRQ(ierr); 484 } 485 486 /* evaluate communication - mesg to who,length of mesg, and buffer space 487 required. Based on this, buffers are allocated, and data copied into them */ 488 ierr = PetscMalloc4(size,&w1,size,&w2,size,&w3,size,&w4);CHKERRQ(ierr); 489 ierr = PetscMemzero(w1,size*sizeof(PetscMPIInt));CHKERRQ(ierr); /* initialise work vector*/ 490 ierr = PetscMemzero(w2,size*sizeof(PetscMPIInt));CHKERRQ(ierr); /* initialise work vector*/ 491 ierr = PetscMemzero(w3,size*sizeof(PetscMPIInt));CHKERRQ(ierr); /* initialise work vector*/ 492 for (i=0; i<imax; i++) { 493 ierr = PetscMemzero(w4,size*sizeof(PetscMPIInt));CHKERRQ(ierr); /* initialise work vector*/ 494 idx_i = idx[i]; 495 len = n[i]; 496 for (j=0; j<len; j++) { 497 row = idx_i[j]; 498 if (row < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Index set cannot have negative entries"); 499 ierr = PetscLayoutFindOwner(C->rmap,row,&proc);CHKERRQ(ierr); 500 w4[proc]++; 501 } 502 for (j=0; j<size; j++) { 503 if (w4[j]) { w1[j] += w4[j]; w3[j]++;} 504 } 505 } 506 507 nrqs = 0; /* no of outgoing messages */ 508 msz = 0; /* total mesg length (for all proc */ 509 w1[rank] = 0; /* no mesg sent to intself */ 510 w3[rank] = 0; 511 for (i=0; i<size; i++) { 512 if (w1[i]) {w2[i] = 1; nrqs++;} /* there exists a message to proc i */ 513 } 514 /* pa - is list of processors to communicate with */ 515 ierr = PetscMalloc1(nrqs+1,&pa);CHKERRQ(ierr); 516 for (i=0,j=0; i<size; i++) { 517 if (w1[i]) {pa[j] = i; j++;} 518 } 519 520 /* Each message would have a header = 1 + 2*(no of IS) + data */ 521 for (i=0; i<nrqs; i++) { 522 j = pa[i]; 523 w1[j] += w2[j] + 2*w3[j]; 524 msz += w1[j]; 525 } 526 527 /* Determine the number of messages to expect, their lengths, from from-ids */ 528 ierr = PetscGatherNumberOfMessages(comm,w2,w1,&nrqr);CHKERRQ(ierr); 529 ierr = PetscGatherMessageLengths(comm,nrqs,nrqr,w1,&onodes1,&olengths1);CHKERRQ(ierr); 530 531 /* Now post the Irecvs corresponding to these messages */ 532 ierr = PetscPostIrecvInt(comm,tag1,nrqr,onodes1,olengths1,&rbuf,&r_waits1);CHKERRQ(ierr); 533 534 /* Allocate Memory for outgoing messages */ 535 ierr = PetscMalloc4(size,&outdat,size,&ptr,msz,&tmp,size,&ctr);CHKERRQ(ierr); 536 ierr = PetscMemzero(outdat,size*sizeof(PetscInt*));CHKERRQ(ierr); 537 ierr = PetscMemzero(ptr,size*sizeof(PetscInt*));CHKERRQ(ierr); 538 539 { 540 PetscInt *iptr = tmp,ict = 0; 541 for (i=0; i<nrqs; i++) { 542 j = pa[i]; 543 iptr += ict; 544 outdat[j] = iptr; 545 ict = w1[j]; 546 } 547 } 548 549 /* Form the outgoing messages */ 550 /* plug in the headers */ 551 for (i=0; i<nrqs; i++) { 552 j = pa[i]; 553 outdat[j][0] = 0; 554 ierr = PetscMemzero(outdat[j]+1,2*w3[j]*sizeof(PetscInt));CHKERRQ(ierr); 555 ptr[j] = outdat[j] + 2*w3[j] + 1; 556 } 557 558 /* Memory for doing local proc's work */ 559 { 560 PetscInt M_BPB_imax = 0; 561 #if defined(PETSC_USE_CTABLE) 562 ierr = PetscIntMultError((M/PETSC_BITS_PER_BYTE+1),imax, &M_BPB_imax);CHKERRQ(ierr); 563 ierr = PetscMalloc1(imax,&table_data);CHKERRQ(ierr); 564 for (i=0; i<imax; i++) { 565 ierr = PetscTableCreate(n[i]+1,M+1,&table_data[i]);CHKERRQ(ierr); 566 } 567 ierr = PetscCalloc4(imax,&table, imax,&data, imax,&isz, M_BPB_imax,&t_p);CHKERRQ(ierr); 568 for (i=0; i<imax; i++) { 569 table[i] = t_p + (M/PETSC_BITS_PER_BYTE+1)*i; 570 } 571 #else 572 PetscInt Mimax = 0; 573 ierr = PetscIntMultError(M,imax, &Mimax);CHKERRQ(ierr); 574 ierr = PetscIntMultError((M/PETSC_BITS_PER_BYTE+1),imax, &M_BPB_imax);CHKERRQ(ierr); 575 ierr = PetscCalloc5(imax,&table, imax,&data, imax,&isz, Mimax,&d_p, M_BPB_imax,&t_p);CHKERRQ(ierr); 576 for (i=0; i<imax; i++) { 577 table[i] = t_p + (M/PETSC_BITS_PER_BYTE+1)*i; 578 data[i] = d_p + M*i; 579 } 580 #endif 581 } 582 583 /* Parse the IS and update local tables and the outgoing buf with the data */ 584 { 585 PetscInt n_i,isz_i,*outdat_j,ctr_j; 586 PetscBT table_i; 587 588 for (i=0; i<imax; i++) { 589 ierr = PetscMemzero(ctr,size*sizeof(PetscInt));CHKERRQ(ierr); 590 n_i = n[i]; 591 table_i = table[i]; 592 idx_i = idx[i]; 593 #if defined(PETSC_USE_CTABLE) 594 table_data_i = table_data[i]; 595 #else 596 data_i = data[i]; 597 #endif 598 isz_i = isz[i]; 599 for (j=0; j<n_i; j++) { /* parse the indices of each IS */ 600 row = idx_i[j]; 601 ierr = PetscLayoutFindOwner(C->rmap,row,&proc);CHKERRQ(ierr); 602 if (proc != rank) { /* copy to the outgoing buffer */ 603 ctr[proc]++; 604 *ptr[proc] = row; 605 ptr[proc]++; 606 } else if (!PetscBTLookupSet(table_i,row)) { 607 #if defined(PETSC_USE_CTABLE) 608 ierr = PetscTableAdd(table_data_i,row+1,isz_i+1,INSERT_VALUES);CHKERRQ(ierr); 609 #else 610 data_i[isz_i] = row; /* Update the local table */ 611 #endif 612 isz_i++; 613 } 614 } 615 /* Update the headers for the current IS */ 616 for (j=0; j<size; j++) { /* Can Optimise this loop by using pa[] */ 617 if ((ctr_j = ctr[j])) { 618 outdat_j = outdat[j]; 619 k = ++outdat_j[0]; 620 outdat_j[2*k] = ctr_j; 621 outdat_j[2*k-1] = i; 622 } 623 } 624 isz[i] = isz_i; 625 } 626 } 627 628 /* Now post the sends */ 629 ierr = PetscMalloc1(nrqs+1,&s_waits1);CHKERRQ(ierr); 630 for (i=0; i<nrqs; ++i) { 631 j = pa[i]; 632 ierr = MPI_Isend(outdat[j],w1[j],MPIU_INT,j,tag1,comm,s_waits1+i);CHKERRQ(ierr); 633 } 634 635 /* No longer need the original indices */ 636 for (i=0; i<imax; ++i) { 637 ierr = ISRestoreIndices(is[i],idx+i);CHKERRQ(ierr); 638 } 639 ierr = PetscFree2(idx,n);CHKERRQ(ierr); 640 641 for (i=0; i<imax; ++i) { 642 ierr = ISDestroy(&is[i]);CHKERRQ(ierr); 643 } 644 645 /* Do Local work */ 646 #if defined(PETSC_USE_CTABLE) 647 ierr = MatIncreaseOverlap_MPIAIJ_Local(C,imax,table,isz,NULL,table_data);CHKERRQ(ierr); 648 #else 649 ierr = MatIncreaseOverlap_MPIAIJ_Local(C,imax,table,isz,data,NULL);CHKERRQ(ierr); 650 #endif 651 652 /* Receive messages */ 653 ierr = PetscMalloc1(nrqr+1,&recv_status);CHKERRQ(ierr); 654 if (nrqr) {ierr = MPI_Waitall(nrqr,r_waits1,recv_status);CHKERRQ(ierr);} 655 656 ierr = PetscMalloc1(nrqs+1,&s_status);CHKERRQ(ierr); 657 if (nrqs) {ierr = MPI_Waitall(nrqs,s_waits1,s_status);CHKERRQ(ierr);} 658 659 /* Phase 1 sends are complete - deallocate buffers */ 660 ierr = PetscFree4(outdat,ptr,tmp,ctr);CHKERRQ(ierr); 661 ierr = PetscFree4(w1,w2,w3,w4);CHKERRQ(ierr); 662 663 ierr = PetscMalloc1(nrqr+1,&xdata);CHKERRQ(ierr); 664 ierr = PetscMalloc1(nrqr+1,&isz1);CHKERRQ(ierr); 665 ierr = MatIncreaseOverlap_MPIAIJ_Receive(C,nrqr,rbuf,xdata,isz1);CHKERRQ(ierr); 666 ierr = PetscFree(rbuf[0]);CHKERRQ(ierr); 667 ierr = PetscFree(rbuf);CHKERRQ(ierr); 668 669 670 /* Send the data back */ 671 /* Do a global reduction to know the buffer space req for incoming messages */ 672 { 673 PetscMPIInt *rw1; 674 675 ierr = PetscCalloc1(size,&rw1);CHKERRQ(ierr); 676 677 for (i=0; i<nrqr; ++i) { 678 proc = recv_status[i].MPI_SOURCE; 679 680 if (proc != onodes1[i]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"MPI_SOURCE mismatch"); 681 rw1[proc] = isz1[i]; 682 } 683 ierr = PetscFree(onodes1);CHKERRQ(ierr); 684 ierr = PetscFree(olengths1);CHKERRQ(ierr); 685 686 /* Determine the number of messages to expect, their lengths, from from-ids */ 687 ierr = PetscGatherMessageLengths(comm,nrqr,nrqs,rw1,&onodes2,&olengths2);CHKERRQ(ierr); 688 ierr = PetscFree(rw1);CHKERRQ(ierr); 689 } 690 /* Now post the Irecvs corresponding to these messages */ 691 ierr = PetscPostIrecvInt(comm,tag2,nrqs,onodes2,olengths2,&rbuf2,&r_waits2);CHKERRQ(ierr); 692 693 /* Now post the sends */ 694 ierr = PetscMalloc1(nrqr+1,&s_waits2);CHKERRQ(ierr); 695 for (i=0; i<nrqr; ++i) { 696 j = recv_status[i].MPI_SOURCE; 697 ierr = MPI_Isend(xdata[i],isz1[i],MPIU_INT,j,tag2,comm,s_waits2+i);CHKERRQ(ierr); 698 } 699 700 /* receive work done on other processors */ 701 { 702 PetscInt is_no,ct1,max,*rbuf2_i,isz_i,jmax; 703 PetscMPIInt idex; 704 PetscBT table_i; 705 MPI_Status *status2; 706 707 ierr = PetscMalloc1((PetscMax(nrqr,nrqs)+1),&status2);CHKERRQ(ierr); 708 for (i=0; i<nrqs; ++i) { 709 ierr = MPI_Waitany(nrqs,r_waits2,&idex,status2+i);CHKERRQ(ierr); 710 /* Process the message */ 711 rbuf2_i = rbuf2[idex]; 712 ct1 = 2*rbuf2_i[0]+1; 713 jmax = rbuf2[idex][0]; 714 for (j=1; j<=jmax; j++) { 715 max = rbuf2_i[2*j]; 716 is_no = rbuf2_i[2*j-1]; 717 isz_i = isz[is_no]; 718 table_i = table[is_no]; 719 #if defined(PETSC_USE_CTABLE) 720 table_data_i = table_data[is_no]; 721 #else 722 data_i = data[is_no]; 723 #endif 724 for (k=0; k<max; k++,ct1++) { 725 row = rbuf2_i[ct1]; 726 if (!PetscBTLookupSet(table_i,row)) { 727 #if defined(PETSC_USE_CTABLE) 728 ierr = PetscTableAdd(table_data_i,row+1,isz_i+1,INSERT_VALUES);CHKERRQ(ierr); 729 #else 730 data_i[isz_i] = row; 731 #endif 732 isz_i++; 733 } 734 } 735 isz[is_no] = isz_i; 736 } 737 } 738 739 if (nrqr) {ierr = MPI_Waitall(nrqr,s_waits2,status2);CHKERRQ(ierr);} 740 ierr = PetscFree(status2);CHKERRQ(ierr); 741 } 742 743 #if defined(PETSC_USE_CTABLE) 744 tcount_max = 0; 745 for (i=0; i<imax; ++i) { 746 table_data_i = table_data[i]; 747 ierr = PetscTableGetCount(table_data_i,&tcount);CHKERRQ(ierr); 748 if (tcount_max < tcount) tcount_max = tcount; 749 } 750 ierr = PetscMalloc1(tcount_max+1,&tdata);CHKERRQ(ierr); 751 #endif 752 753 for (i=0; i<imax; ++i) { 754 #if defined(PETSC_USE_CTABLE) 755 PetscTablePosition tpos; 756 table_data_i = table_data[i]; 757 758 ierr = PetscTableGetHeadPosition(table_data_i,&tpos);CHKERRQ(ierr); 759 while (tpos) { 760 ierr = PetscTableGetNext(table_data_i,&tpos,&k,&j);CHKERRQ(ierr); 761 tdata[--j] = --k; 762 } 763 ierr = ISCreateGeneral(PETSC_COMM_SELF,isz[i],tdata,PETSC_COPY_VALUES,is+i);CHKERRQ(ierr); 764 #else 765 ierr = ISCreateGeneral(PETSC_COMM_SELF,isz[i],data[i],PETSC_COPY_VALUES,is+i);CHKERRQ(ierr); 766 #endif 767 } 768 769 ierr = PetscFree(onodes2);CHKERRQ(ierr); 770 ierr = PetscFree(olengths2);CHKERRQ(ierr); 771 772 ierr = PetscFree(pa);CHKERRQ(ierr); 773 ierr = PetscFree(rbuf2[0]);CHKERRQ(ierr); 774 ierr = PetscFree(rbuf2);CHKERRQ(ierr); 775 ierr = PetscFree(s_waits1);CHKERRQ(ierr); 776 ierr = PetscFree(r_waits1);CHKERRQ(ierr); 777 ierr = PetscFree(s_waits2);CHKERRQ(ierr); 778 ierr = PetscFree(r_waits2);CHKERRQ(ierr); 779 ierr = PetscFree(s_status);CHKERRQ(ierr); 780 ierr = PetscFree(recv_status);CHKERRQ(ierr); 781 ierr = PetscFree(xdata[0]);CHKERRQ(ierr); 782 ierr = PetscFree(xdata);CHKERRQ(ierr); 783 ierr = PetscFree(isz1);CHKERRQ(ierr); 784 #if defined(PETSC_USE_CTABLE) 785 for (i=0; i<imax; i++) { 786 ierr = PetscTableDestroy((PetscTable*)&table_data[i]);CHKERRQ(ierr); 787 } 788 ierr = PetscFree(table_data);CHKERRQ(ierr); 789 ierr = PetscFree(tdata);CHKERRQ(ierr); 790 ierr = PetscFree4(table,data,isz,t_p);CHKERRQ(ierr); 791 #else 792 ierr = PetscFree5(table,data,isz,d_p,t_p);CHKERRQ(ierr); 793 #endif 794 PetscFunctionReturn(0); 795 } 796 797 /* 798 MatIncreaseOverlap_MPIAIJ_Local - Called by MatincreaseOverlap, to do 799 the work on the local processor. 800 801 Inputs: 802 C - MAT_MPIAIJ; 803 imax - total no of index sets processed at a time; 804 table - an array of char - size = m bits. 805 806 Output: 807 isz - array containing the count of the solution elements corresponding 808 to each index set; 809 data or table_data - pointer to the solutions 810 */ 811 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local(Mat C,PetscInt imax,PetscBT *table,PetscInt *isz,PetscInt **data,PetscTable *table_data) 812 { 813 Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data; 814 Mat A = c->A,B = c->B; 815 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data; 816 PetscInt start,end,val,max,rstart,cstart,*ai,*aj; 817 PetscInt *bi,*bj,*garray,i,j,k,row,isz_i; 818 PetscBT table_i; 819 #if defined(PETSC_USE_CTABLE) 820 PetscTable table_data_i; 821 PetscErrorCode ierr; 822 PetscTablePosition tpos; 823 PetscInt tcount,*tdata; 824 #else 825 PetscInt *data_i; 826 #endif 827 828 PetscFunctionBegin; 829 rstart = C->rmap->rstart; 830 cstart = C->cmap->rstart; 831 ai = a->i; 832 aj = a->j; 833 bi = b->i; 834 bj = b->j; 835 garray = c->garray; 836 837 for (i=0; i<imax; i++) { 838 #if defined(PETSC_USE_CTABLE) 839 /* copy existing entries of table_data_i into tdata[] */ 840 table_data_i = table_data[i]; 841 ierr = PetscTableGetCount(table_data_i,&tcount);CHKERRQ(ierr); 842 if (tcount != isz[i]) SETERRQ3(PETSC_COMM_SELF,0," tcount %d != isz[%d] %d",tcount,i,isz[i]); 843 844 ierr = PetscMalloc1(tcount,&tdata);CHKERRQ(ierr); 845 ierr = PetscTableGetHeadPosition(table_data_i,&tpos);CHKERRQ(ierr); 846 while (tpos) { 847 ierr = PetscTableGetNext(table_data_i,&tpos,&row,&j);CHKERRQ(ierr); 848 tdata[--j] = --row; 849 if (j > tcount - 1) SETERRQ2(PETSC_COMM_SELF,0," j %d >= tcount %d",j,tcount); 850 } 851 #else 852 data_i = data[i]; 853 #endif 854 table_i = table[i]; 855 isz_i = isz[i]; 856 max = isz[i]; 857 858 for (j=0; j<max; j++) { 859 #if defined(PETSC_USE_CTABLE) 860 row = tdata[j] - rstart; 861 #else 862 row = data_i[j] - rstart; 863 #endif 864 start = ai[row]; 865 end = ai[row+1]; 866 for (k=start; k<end; k++) { /* Amat */ 867 val = aj[k] + cstart; 868 if (!PetscBTLookupSet(table_i,val)) { 869 #if defined(PETSC_USE_CTABLE) 870 ierr = PetscTableAdd(table_data_i,val+1,isz_i+1,INSERT_VALUES);CHKERRQ(ierr); 871 #else 872 data_i[isz_i] = val; 873 #endif 874 isz_i++; 875 } 876 } 877 start = bi[row]; 878 end = bi[row+1]; 879 for (k=start; k<end; k++) { /* Bmat */ 880 val = garray[bj[k]]; 881 if (!PetscBTLookupSet(table_i,val)) { 882 #if defined(PETSC_USE_CTABLE) 883 ierr = PetscTableAdd(table_data_i,val+1,isz_i+1,INSERT_VALUES);CHKERRQ(ierr); 884 #else 885 data_i[isz_i] = val; 886 #endif 887 isz_i++; 888 } 889 } 890 } 891 isz[i] = isz_i; 892 893 #if defined(PETSC_USE_CTABLE) 894 ierr = PetscFree(tdata);CHKERRQ(ierr); 895 #endif 896 } 897 PetscFunctionReturn(0); 898 } 899 900 /* 901 MatIncreaseOverlap_MPIAIJ_Receive - Process the recieved messages, 902 and return the output 903 904 Input: 905 C - the matrix 906 nrqr - no of messages being processed. 907 rbuf - an array of pointers to the recieved requests 908 909 Output: 910 xdata - array of messages to be sent back 911 isz1 - size of each message 912 913 For better efficiency perhaps we should malloc separately each xdata[i], 914 then if a remalloc is required we need only copy the data for that one row 915 rather then all previous rows as it is now where a single large chunck of 916 memory is used. 917 918 */ 919 static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Receive(Mat C,PetscInt nrqr,PetscInt **rbuf,PetscInt **xdata,PetscInt * isz1) 920 { 921 Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data; 922 Mat A = c->A,B = c->B; 923 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data; 924 PetscErrorCode ierr; 925 PetscInt rstart,cstart,*ai,*aj,*bi,*bj,*garray,i,j,k; 926 PetscInt row,total_sz,ct,ct1,ct2,ct3,mem_estimate,oct2,l,start,end; 927 PetscInt val,max1,max2,m,no_malloc =0,*tmp,new_estimate,ctr; 928 PetscInt *rbuf_i,kmax,rbuf_0; 929 PetscBT xtable; 930 931 PetscFunctionBegin; 932 m = C->rmap->N; 933 rstart = C->rmap->rstart; 934 cstart = C->cmap->rstart; 935 ai = a->i; 936 aj = a->j; 937 bi = b->i; 938 bj = b->j; 939 garray = c->garray; 940 941 942 for (i=0,ct=0,total_sz=0; i<nrqr; ++i) { 943 rbuf_i = rbuf[i]; 944 rbuf_0 = rbuf_i[0]; 945 ct += rbuf_0; 946 for (j=1; j<=rbuf_0; j++) total_sz += rbuf_i[2*j]; 947 } 948 949 if (C->rmap->n) max1 = ct*(a->nz + b->nz)/C->rmap->n; 950 else max1 = 1; 951 mem_estimate = 3*((total_sz > max1 ? total_sz : max1)+1); 952 ierr = PetscMalloc1(mem_estimate,&xdata[0]);CHKERRQ(ierr); 953 ++no_malloc; 954 ierr = PetscBTCreate(m,&xtable);CHKERRQ(ierr); 955 ierr = PetscMemzero(isz1,nrqr*sizeof(PetscInt));CHKERRQ(ierr); 956 957 ct3 = 0; 958 for (i=0; i<nrqr; i++) { /* for easch mesg from proc i */ 959 rbuf_i = rbuf[i]; 960 rbuf_0 = rbuf_i[0]; 961 ct1 = 2*rbuf_0+1; 962 ct2 = ct1; 963 ct3 += ct1; 964 for (j=1; j<=rbuf_0; j++) { /* for each IS from proc i*/ 965 ierr = PetscBTMemzero(m,xtable);CHKERRQ(ierr); 966 oct2 = ct2; 967 kmax = rbuf_i[2*j]; 968 for (k=0; k<kmax; k++,ct1++) { 969 row = rbuf_i[ct1]; 970 if (!PetscBTLookupSet(xtable,row)) { 971 if (!(ct3 < mem_estimate)) { 972 new_estimate = (PetscInt)(1.5*mem_estimate)+1; 973 ierr = PetscMalloc1(new_estimate,&tmp);CHKERRQ(ierr); 974 ierr = PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));CHKERRQ(ierr); 975 ierr = PetscFree(xdata[0]);CHKERRQ(ierr); 976 xdata[0] = tmp; 977 mem_estimate = new_estimate; ++no_malloc; 978 for (ctr=1; ctr<=i; ctr++) xdata[ctr] = xdata[ctr-1] + isz1[ctr-1]; 979 } 980 xdata[i][ct2++] = row; 981 ct3++; 982 } 983 } 984 for (k=oct2,max2=ct2; k<max2; k++) { 985 row = xdata[i][k] - rstart; 986 start = ai[row]; 987 end = ai[row+1]; 988 for (l=start; l<end; l++) { 989 val = aj[l] + cstart; 990 if (!PetscBTLookupSet(xtable,val)) { 991 if (!(ct3 < mem_estimate)) { 992 new_estimate = (PetscInt)(1.5*mem_estimate)+1; 993 ierr = PetscMalloc1(new_estimate,&tmp);CHKERRQ(ierr); 994 ierr = PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));CHKERRQ(ierr); 995 ierr = PetscFree(xdata[0]);CHKERRQ(ierr); 996 xdata[0] = tmp; 997 mem_estimate = new_estimate; ++no_malloc; 998 for (ctr=1; ctr<=i; ctr++) xdata[ctr] = xdata[ctr-1] + isz1[ctr-1]; 999 } 1000 xdata[i][ct2++] = val; 1001 ct3++; 1002 } 1003 } 1004 start = bi[row]; 1005 end = bi[row+1]; 1006 for (l=start; l<end; l++) { 1007 val = garray[bj[l]]; 1008 if (!PetscBTLookupSet(xtable,val)) { 1009 if (!(ct3 < mem_estimate)) { 1010 new_estimate = (PetscInt)(1.5*mem_estimate)+1; 1011 ierr = PetscMalloc1(new_estimate,&tmp);CHKERRQ(ierr); 1012 ierr = PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));CHKERRQ(ierr); 1013 ierr = PetscFree(xdata[0]);CHKERRQ(ierr); 1014 xdata[0] = tmp; 1015 mem_estimate = new_estimate; ++no_malloc; 1016 for (ctr =1; ctr <=i; ctr++) xdata[ctr] = xdata[ctr-1] + isz1[ctr-1]; 1017 } 1018 xdata[i][ct2++] = val; 1019 ct3++; 1020 } 1021 } 1022 } 1023 /* Update the header*/ 1024 xdata[i][2*j] = ct2 - oct2; /* Undo the vector isz1 and use only a var*/ 1025 xdata[i][2*j-1] = rbuf_i[2*j-1]; 1026 } 1027 xdata[i][0] = rbuf_0; 1028 xdata[i+1] = xdata[i] + ct2; 1029 isz1[i] = ct2; /* size of each message */ 1030 } 1031 ierr = PetscBTDestroy(&xtable);CHKERRQ(ierr); 1032 ierr = PetscInfo3(C,"Allocated %D bytes, required %D bytes, no of mallocs = %D\n",mem_estimate,ct3,no_malloc);CHKERRQ(ierr); 1033 PetscFunctionReturn(0); 1034 } 1035 /* -------------------------------------------------------------------------*/ 1036 extern PetscErrorCode MatGetSubMatrices_MPIAIJ_Local(Mat,PetscInt,const IS[],const IS[],MatReuse,Mat*); 1037 extern PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat,MatAssemblyType); 1038 /* 1039 Every processor gets the entire matrix 1040 */ 1041 PetscErrorCode MatGetSubMatrix_MPIAIJ_All(Mat A,MatGetSubMatrixOption flag,MatReuse scall,Mat *Bin[]) 1042 { 1043 Mat B; 1044 Mat_MPIAIJ *a = (Mat_MPIAIJ*)A->data; 1045 Mat_SeqAIJ *b,*ad = (Mat_SeqAIJ*)a->A->data,*bd = (Mat_SeqAIJ*)a->B->data; 1046 PetscErrorCode ierr; 1047 PetscMPIInt size,rank,*recvcounts = 0,*displs = 0; 1048 PetscInt sendcount,i,*rstarts = A->rmap->range,n,cnt,j; 1049 PetscInt m,*b_sendj,*garray = a->garray,*lens,*jsendbuf,*a_jsendbuf,*b_jsendbuf; 1050 MatScalar *sendbuf,*recvbuf,*a_sendbuf,*b_sendbuf; 1051 1052 PetscFunctionBegin; 1053 ierr = MPI_Comm_size(PetscObjectComm((PetscObject)A),&size);CHKERRQ(ierr); 1054 ierr = MPI_Comm_rank(PetscObjectComm((PetscObject)A),&rank);CHKERRQ(ierr); 1055 1056 if (scall == MAT_INITIAL_MATRIX) { 1057 /* ---------------------------------------------------------------- 1058 Tell every processor the number of nonzeros per row 1059 */ 1060 ierr = PetscMalloc1(A->rmap->N,&lens);CHKERRQ(ierr); 1061 for (i=A->rmap->rstart; i<A->rmap->rend; i++) { 1062 lens[i] = ad->i[i-A->rmap->rstart+1] - ad->i[i-A->rmap->rstart] + bd->i[i-A->rmap->rstart+1] - bd->i[i-A->rmap->rstart]; 1063 } 1064 ierr = PetscMalloc2(size,&recvcounts,size,&displs);CHKERRQ(ierr); 1065 for (i=0; i<size; i++) { 1066 recvcounts[i] = A->rmap->range[i+1] - A->rmap->range[i]; 1067 displs[i] = A->rmap->range[i]; 1068 } 1069 #if defined(PETSC_HAVE_MPI_IN_PLACE) 1070 ierr = MPI_Allgatherv(MPI_IN_PLACE,0,MPI_DATATYPE_NULL,lens,recvcounts,displs,MPIU_INT,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1071 #else 1072 sendcount = A->rmap->rend - A->rmap->rstart; 1073 ierr = MPI_Allgatherv(lens+A->rmap->rstart,sendcount,MPIU_INT,lens,recvcounts,displs,MPIU_INT,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1074 #endif 1075 /* --------------------------------------------------------------- 1076 Create the sequential matrix of the same type as the local block diagonal 1077 */ 1078 ierr = MatCreate(PETSC_COMM_SELF,&B);CHKERRQ(ierr); 1079 ierr = MatSetSizes(B,A->rmap->N,A->cmap->N,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 1080 ierr = MatSetBlockSizesFromMats(B,A,A);CHKERRQ(ierr); 1081 ierr = MatSetType(B,((PetscObject)a->A)->type_name);CHKERRQ(ierr); 1082 ierr = MatSeqAIJSetPreallocation(B,0,lens);CHKERRQ(ierr); 1083 ierr = PetscMalloc1(1,Bin);CHKERRQ(ierr); 1084 **Bin = B; 1085 b = (Mat_SeqAIJ*)B->data; 1086 1087 /*-------------------------------------------------------------------- 1088 Copy my part of matrix column indices over 1089 */ 1090 sendcount = ad->nz + bd->nz; 1091 jsendbuf = b->j + b->i[rstarts[rank]]; 1092 a_jsendbuf = ad->j; 1093 b_jsendbuf = bd->j; 1094 n = A->rmap->rend - A->rmap->rstart; 1095 cnt = 0; 1096 for (i=0; i<n; i++) { 1097 1098 /* put in lower diagonal portion */ 1099 m = bd->i[i+1] - bd->i[i]; 1100 while (m > 0) { 1101 /* is it above diagonal (in bd (compressed) numbering) */ 1102 if (garray[*b_jsendbuf] > A->rmap->rstart + i) break; 1103 jsendbuf[cnt++] = garray[*b_jsendbuf++]; 1104 m--; 1105 } 1106 1107 /* put in diagonal portion */ 1108 for (j=ad->i[i]; j<ad->i[i+1]; j++) { 1109 jsendbuf[cnt++] = A->rmap->rstart + *a_jsendbuf++; 1110 } 1111 1112 /* put in upper diagonal portion */ 1113 while (m-- > 0) { 1114 jsendbuf[cnt++] = garray[*b_jsendbuf++]; 1115 } 1116 } 1117 if (cnt != sendcount) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Corrupted PETSc matrix: nz given %D actual nz %D",sendcount,cnt); 1118 1119 /*-------------------------------------------------------------------- 1120 Gather all column indices to all processors 1121 */ 1122 for (i=0; i<size; i++) { 1123 recvcounts[i] = 0; 1124 for (j=A->rmap->range[i]; j<A->rmap->range[i+1]; j++) { 1125 recvcounts[i] += lens[j]; 1126 } 1127 } 1128 displs[0] = 0; 1129 for (i=1; i<size; i++) { 1130 displs[i] = displs[i-1] + recvcounts[i-1]; 1131 } 1132 #if defined(PETSC_HAVE_MPI_IN_PLACE) 1133 ierr = MPI_Allgatherv(MPI_IN_PLACE,0,MPI_DATATYPE_NULL,b->j,recvcounts,displs,MPIU_INT,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1134 #else 1135 ierr = MPI_Allgatherv(jsendbuf,sendcount,MPIU_INT,b->j,recvcounts,displs,MPIU_INT,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1136 #endif 1137 /*-------------------------------------------------------------------- 1138 Assemble the matrix into useable form (note numerical values not yet set) 1139 */ 1140 /* set the b->ilen (length of each row) values */ 1141 ierr = PetscMemcpy(b->ilen,lens,A->rmap->N*sizeof(PetscInt));CHKERRQ(ierr); 1142 /* set the b->i indices */ 1143 b->i[0] = 0; 1144 for (i=1; i<=A->rmap->N; i++) { 1145 b->i[i] = b->i[i-1] + lens[i-1]; 1146 } 1147 ierr = PetscFree(lens);CHKERRQ(ierr); 1148 ierr = MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1149 ierr = MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1150 1151 } else { 1152 B = **Bin; 1153 b = (Mat_SeqAIJ*)B->data; 1154 } 1155 1156 /*-------------------------------------------------------------------- 1157 Copy my part of matrix numerical values into the values location 1158 */ 1159 if (flag == MAT_GET_VALUES) { 1160 sendcount = ad->nz + bd->nz; 1161 sendbuf = b->a + b->i[rstarts[rank]]; 1162 a_sendbuf = ad->a; 1163 b_sendbuf = bd->a; 1164 b_sendj = bd->j; 1165 n = A->rmap->rend - A->rmap->rstart; 1166 cnt = 0; 1167 for (i=0; i<n; i++) { 1168 1169 /* put in lower diagonal portion */ 1170 m = bd->i[i+1] - bd->i[i]; 1171 while (m > 0) { 1172 /* is it above diagonal (in bd (compressed) numbering) */ 1173 if (garray[*b_sendj] > A->rmap->rstart + i) break; 1174 sendbuf[cnt++] = *b_sendbuf++; 1175 m--; 1176 b_sendj++; 1177 } 1178 1179 /* put in diagonal portion */ 1180 for (j=ad->i[i]; j<ad->i[i+1]; j++) { 1181 sendbuf[cnt++] = *a_sendbuf++; 1182 } 1183 1184 /* put in upper diagonal portion */ 1185 while (m-- > 0) { 1186 sendbuf[cnt++] = *b_sendbuf++; 1187 b_sendj++; 1188 } 1189 } 1190 if (cnt != sendcount) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Corrupted PETSc matrix: nz given %D actual nz %D",sendcount,cnt); 1191 1192 /* ----------------------------------------------------------------- 1193 Gather all numerical values to all processors 1194 */ 1195 if (!recvcounts) { 1196 ierr = PetscMalloc2(size,&recvcounts,size,&displs);CHKERRQ(ierr); 1197 } 1198 for (i=0; i<size; i++) { 1199 recvcounts[i] = b->i[rstarts[i+1]] - b->i[rstarts[i]]; 1200 } 1201 displs[0] = 0; 1202 for (i=1; i<size; i++) { 1203 displs[i] = displs[i-1] + recvcounts[i-1]; 1204 } 1205 recvbuf = b->a; 1206 #if defined(PETSC_HAVE_MPI_IN_PLACE) 1207 ierr = MPI_Allgatherv(MPI_IN_PLACE,0,MPI_DATATYPE_NULL,recvbuf,recvcounts,displs,MPIU_SCALAR,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1208 #else 1209 ierr = MPI_Allgatherv(sendbuf,sendcount,MPIU_SCALAR,recvbuf,recvcounts,displs,MPIU_SCALAR,PetscObjectComm((PetscObject)A));CHKERRQ(ierr); 1210 #endif 1211 } /* endof (flag == MAT_GET_VALUES) */ 1212 ierr = PetscFree2(recvcounts,displs);CHKERRQ(ierr); 1213 1214 if (A->symmetric) { 1215 ierr = MatSetOption(B,MAT_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 1216 } else if (A->hermitian) { 1217 ierr = MatSetOption(B,MAT_HERMITIAN,PETSC_TRUE);CHKERRQ(ierr); 1218 } else if (A->structurally_symmetric) { 1219 ierr = MatSetOption(B,MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);CHKERRQ(ierr); 1220 } 1221 PetscFunctionReturn(0); 1222 } 1223 1224 PetscErrorCode MatGetSubMatrices_MPIAIJ_SingleIS_Local(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,PetscBool allcolumns,Mat *submats) 1225 { 1226 Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data; 1227 Mat submat,A = c->A,B = c->B; 1228 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data,*subc; 1229 PetscInt *ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j,nzA,nzB; 1230 PetscInt cstart = C->cmap->rstart,cend = C->cmap->rend,rstart = C->rmap->rstart,*bmap = c->garray; 1231 const PetscInt *icol,*irow; 1232 PetscInt nrow,ncol,start; 1233 PetscErrorCode ierr; 1234 PetscMPIInt rank,size,tag1,tag2,tag3,tag4,*w1,*w2,nrqr; 1235 PetscInt **sbuf1,**sbuf2,i,j,k,l,ct1,ct2,ct3,**rbuf1,row,proc; 1236 PetscInt nrqs=0,msz,**ptr,*req_size,*ctr,*pa,*tmp,tcol,*iptr; 1237 PetscInt **rbuf3,*req_source1,*req_source2,**sbuf_aj,**rbuf2,max1,nnz; 1238 PetscInt *lens,rmax,ncols,*cols,Crow; 1239 #if defined(PETSC_USE_CTABLE) 1240 PetscTable cmap,rmap; 1241 PetscInt *cmap_loc,*rmap_loc; 1242 #else 1243 PetscInt *cmap,*rmap; 1244 #endif 1245 PetscInt ctr_j,*sbuf1_j,*sbuf_aj_i,*rbuf1_i,kmax,*sbuf1_i,*rbuf2_i,*rbuf3_i; 1246 PetscInt *cworkB,lwrite,*subcols,*row2proc; 1247 PetscScalar *vworkA,*vworkB,*a_a = a->a,*b_a = b->a,*subvals=NULL; 1248 MPI_Request *s_waits1,*r_waits1,*s_waits2,*r_waits2,*r_waits3; 1249 MPI_Request *r_waits4,*s_waits3 = NULL,*s_waits4; 1250 MPI_Status *r_status1,*r_status2,*s_status1,*s_status3 = NULL,*s_status2; 1251 MPI_Status *r_status3 = NULL,*r_status4,*s_status4; 1252 MPI_Comm comm; 1253 PetscScalar **rbuf4,**sbuf_aa,*vals,*sbuf_aa_i,*rbuf4_i; 1254 PetscMPIInt *onodes1,*olengths1,idex,end; 1255 Mat_SubMat *smatis1; 1256 PetscBool isrowsorted; 1257 1258 PetscFunctionBegin; 1259 if (ismax != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"This routine only works when all processes have ismax=1"); 1260 1261 ierr = PetscObjectGetComm((PetscObject)C,&comm);CHKERRQ(ierr); 1262 size = c->size; 1263 rank = c->rank; 1264 1265 ierr = ISSorted(isrow[0],&isrowsorted);CHKERRQ(ierr); 1266 if (!isrowsorted) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"isrow[0] must be sorted"); 1267 1268 ierr = ISGetIndices(isrow[0],&irow);CHKERRQ(ierr); 1269 ierr = ISGetLocalSize(isrow[0],&nrow);CHKERRQ(ierr); 1270 if (allcolumns) { 1271 icol = NULL; 1272 ncol = C->cmap->N; 1273 } else { 1274 ierr = ISGetIndices(iscol[0],&icol);CHKERRQ(ierr); 1275 ierr = ISGetLocalSize(iscol[0],&ncol);CHKERRQ(ierr); 1276 } 1277 1278 if (scall == MAT_INITIAL_MATRIX) { 1279 PetscInt *sbuf2_i,*cworkA,lwrite,ctmp; 1280 1281 /* Get some new tags to keep the communication clean */ 1282 tag1 = ((PetscObject)C)->tag; 1283 ierr = PetscObjectGetNewTag((PetscObject)C,&tag2);CHKERRQ(ierr); 1284 ierr = PetscObjectGetNewTag((PetscObject)C,&tag3);CHKERRQ(ierr); 1285 1286 /* evaluate communication - mesg to who, length of mesg, and buffer space 1287 required. Based on this, buffers are allocated, and data copied into them */ 1288 ierr = PetscCalloc2(size,&w1,size,&w2);CHKERRQ(ierr); 1289 ierr = PetscMalloc1(nrow,&row2proc);CHKERRQ(ierr); 1290 1291 /* w1[proc] = num of rows owned by proc -- to be requested */ 1292 proc = 0; 1293 nrqs = 0; /* num of outgoing messages */ 1294 for (j=0; j<nrow; j++) { 1295 row = irow[j]; /* sorted! */ 1296 while (row >= C->rmap->range[proc+1]) proc++; 1297 w1[proc]++; 1298 row2proc[j] = proc; /* map row index to proc */ 1299 1300 if (proc != rank && !w2[proc]) { 1301 w2[proc] = 1; nrqs++; 1302 } 1303 } 1304 w1[rank] = 0; /* rows owned by self will not be requested */ 1305 1306 ierr = PetscMalloc1(nrqs+1,&pa);CHKERRQ(ierr); /*(proc -array)*/ 1307 for (proc=0,j=0; proc<size; proc++) { 1308 if (w1[proc]) { pa[j++] = proc;} 1309 } 1310 1311 /* Each message would have a header = 1 + 2*(num of IS) + data (here,num of IS = 1) */ 1312 msz = 0; /* total mesg length (for all procs) */ 1313 for (i=0; i<nrqs; i++) { 1314 proc = pa[i]; 1315 w1[proc] += 3; 1316 msz += w1[proc]; 1317 } 1318 ierr = PetscInfo2(0,"Number of outgoing messages %D Total message length %D\n",nrqs,msz);CHKERRQ(ierr); 1319 1320 /* Determine nrqr, the number of messages to expect, their lengths, from from-ids */ 1321 /* if w2[proc]=1, a message of length w1[proc] will be sent to proc; */ 1322 ierr = PetscGatherNumberOfMessages(comm,w2,w1,&nrqr);CHKERRQ(ierr); 1323 1324 /* Input: nrqs: nsend; nrqr: nrecv; w1: msg length to be sent; 1325 Output: onodes1: recv node-ids; olengths1: corresponding recv message length */ 1326 ierr = PetscGatherMessageLengths(comm,nrqs,nrqr,w1,&onodes1,&olengths1);CHKERRQ(ierr); 1327 1328 /* Now post the Irecvs corresponding to these messages */ 1329 ierr = PetscPostIrecvInt(comm,tag1,nrqr,onodes1,olengths1,&rbuf1,&r_waits1);CHKERRQ(ierr); 1330 1331 ierr = PetscFree(onodes1);CHKERRQ(ierr); 1332 ierr = PetscFree(olengths1);CHKERRQ(ierr); 1333 1334 /* Allocate Memory for outgoing messages */ 1335 ierr = PetscMalloc4(size,&sbuf1,size,&ptr,2*msz,&tmp,size,&ctr);CHKERRQ(ierr); 1336 ierr = PetscMemzero(sbuf1,size*sizeof(PetscInt*));CHKERRQ(ierr); 1337 ierr = PetscMemzero(ptr,size*sizeof(PetscInt*));CHKERRQ(ierr); 1338 1339 /* subf1[pa[0]] = tmp, subf1[pa[i]] = subf1[pa[i-1]] + w1[pa[i-1]] */ 1340 iptr = tmp; 1341 for (i=0; i<nrqs; i++) { 1342 proc = pa[i]; 1343 sbuf1[proc] = iptr; 1344 iptr += w1[proc]; 1345 } 1346 1347 /* Form the outgoing messages */ 1348 /* Initialize the header space */ 1349 for (i=0; i<nrqs; i++) { 1350 proc = pa[i]; 1351 ierr = PetscMemzero(sbuf1[proc],3*sizeof(PetscInt));CHKERRQ(ierr); 1352 ptr[proc] = sbuf1[proc] + 3; 1353 } 1354 1355 /* Parse the isrow and copy data into outbuf */ 1356 ierr = PetscMemzero(ctr,size*sizeof(PetscInt));CHKERRQ(ierr); 1357 for (j=0; j<nrow; j++) { /* parse the indices of each IS */ 1358 proc = row2proc[j]; 1359 if (proc != rank) { /* copy to the outgoing buf*/ 1360 *ptr[proc] = irow[j]; 1361 ctr[proc]++; ptr[proc]++; 1362 } 1363 } 1364 1365 /* Update the headers for the current IS */ 1366 for (j=0; j<size; j++) { /* Can Optimise this loop too */ 1367 if ((ctr_j = ctr[j])) { 1368 sbuf1_j = sbuf1[j]; 1369 k = ++sbuf1_j[0]; 1370 sbuf1_j[2*k] = ctr_j; 1371 sbuf1_j[2*k-1] = 0; 1372 } 1373 } 1374 1375 /* Now post the sends */ 1376 ierr = PetscMalloc1(nrqs+1,&s_waits1);CHKERRQ(ierr); 1377 for (i=0; i<nrqs; ++i) { 1378 proc = pa[i]; 1379 ierr = MPI_Isend(sbuf1[proc],w1[proc],MPIU_INT,proc,tag1,comm,s_waits1+i);CHKERRQ(ierr); 1380 } 1381 1382 /* Post Receives to capture the buffer size */ 1383 ierr = PetscMalloc4(nrqs+1,&r_status2,nrqr+1,&s_waits2,nrqs+1,&r_waits2,nrqr+1,&s_status2);CHKERRQ(ierr); 1384 ierr = PetscMalloc3(nrqs+1,&req_source2,nrqs+1,&rbuf2,nrqs+1,&rbuf3);CHKERRQ(ierr); 1385 1386 rbuf2[0] = tmp + msz; 1387 for (i=1; i<nrqs; ++i) rbuf2[i] = rbuf2[i-1] + w1[pa[i-1]]; 1388 1389 for (i=0; i<nrqs; ++i) { 1390 proc = pa[i]; 1391 ierr = MPI_Irecv(rbuf2[i],w1[proc],MPIU_INT,proc,tag2,comm,r_waits2+i);CHKERRQ(ierr); 1392 } 1393 1394 ierr = PetscFree2(w1,w2);CHKERRQ(ierr); 1395 1396 /* Send to other procs the buf size they should allocate */ 1397 /* Receive messages*/ 1398 ierr = PetscMalloc1(nrqr+1,&r_status1);CHKERRQ(ierr); 1399 ierr = PetscMalloc3(nrqr,&sbuf2,nrqr,&req_size,nrqr,&req_source1);CHKERRQ(ierr); 1400 1401 ierr = MPI_Waitall(nrqr,r_waits1,r_status1);CHKERRQ(ierr); 1402 for (i=0; i<nrqr; ++i) { 1403 req_size[i] = 0; 1404 rbuf1_i = rbuf1[i]; 1405 start = 2*rbuf1_i[0] + 1; 1406 ierr = MPI_Get_count(r_status1+i,MPIU_INT,&end);CHKERRQ(ierr); 1407 ierr = PetscMalloc1(end+1,&sbuf2[i]);CHKERRQ(ierr); 1408 sbuf2_i = sbuf2[i]; 1409 for (j=start; j<end; j++) { 1410 k = rbuf1_i[j] - rstart; 1411 ncols = ai[k+1] - ai[k] + bi[k+1] - bi[k]; 1412 sbuf2_i[j] = ncols; 1413 req_size[i] += ncols; 1414 } 1415 req_source1[i] = r_status1[i].MPI_SOURCE; 1416 1417 /* form the header */ 1418 sbuf2_i[0] = req_size[i]; 1419 for (j=1; j<start; j++) sbuf2_i[j] = rbuf1_i[j]; 1420 1421 ierr = MPI_Isend(sbuf2_i,end,MPIU_INT,req_source1[i],tag2,comm,s_waits2+i);CHKERRQ(ierr); 1422 } 1423 1424 ierr = PetscFree(r_status1);CHKERRQ(ierr); 1425 ierr = PetscFree(r_waits1);CHKERRQ(ierr); 1426 1427 /* rbuf2 is received, Post recv column indices a->j */ 1428 ierr = MPI_Waitall(nrqs,r_waits2,r_status2);CHKERRQ(ierr); 1429 1430 ierr = PetscMalloc4(nrqs+1,&r_waits3,nrqr+1,&s_waits3,nrqs+1,&r_status3,nrqr+1,&s_status3);CHKERRQ(ierr); 1431 for (i=0; i<nrqs; ++i) { 1432 ierr = PetscMalloc1(rbuf2[i][0]+1,&rbuf3[i]);CHKERRQ(ierr); 1433 req_source2[i] = r_status2[i].MPI_SOURCE; 1434 ierr = MPI_Irecv(rbuf3[i],rbuf2[i][0],MPIU_INT,req_source2[i],tag3,comm,r_waits3+i);CHKERRQ(ierr); 1435 } 1436 1437 /* Wait on sends1 and sends2 */ 1438 ierr = PetscMalloc1(nrqs+1,&s_status1);CHKERRQ(ierr); 1439 ierr = MPI_Waitall(nrqs,s_waits1,s_status1);CHKERRQ(ierr); 1440 ierr = PetscFree(s_waits1);CHKERRQ(ierr); 1441 ierr = PetscFree(s_status1);CHKERRQ(ierr); 1442 1443 ierr = MPI_Waitall(nrqr,s_waits2,s_status2);CHKERRQ(ierr); 1444 ierr = PetscFree4(r_status2,s_waits2,r_waits2,s_status2);CHKERRQ(ierr); 1445 1446 /* Now allocate sending buffers for a->j, and send them off */ 1447 ierr = PetscMalloc1(nrqr+1,&sbuf_aj);CHKERRQ(ierr); 1448 for (i=0,j=0; i<nrqr; i++) j += req_size[i]; 1449 ierr = PetscMalloc1(j+1,&sbuf_aj[0]);CHKERRQ(ierr); 1450 for (i=1; i<nrqr; i++) sbuf_aj[i] = sbuf_aj[i-1] + req_size[i-1]; 1451 1452 for (i=0; i<nrqr; i++) { /* for each requested message */ 1453 rbuf1_i = rbuf1[i]; 1454 sbuf_aj_i = sbuf_aj[i]; 1455 ct1 = 2*rbuf1_i[0] + 1; 1456 ct2 = 0; 1457 /* max1=rbuf1_i[0]; if (max1 != 1) SETERRQ1(PETSC_COMM_SELF,0,"max1 %d != 1",max1); */ 1458 1459 kmax = rbuf1[i][2]; 1460 for (k=0; k<kmax; k++,ct1++) { /* for each row */ 1461 row = rbuf1_i[ct1] - rstart; 1462 nzA = ai[row+1] - ai[row]; 1463 nzB = bi[row+1] - bi[row]; 1464 ncols = nzA + nzB; 1465 cworkA = aj + ai[row]; cworkB = bj + bi[row]; 1466 1467 /* load the column indices for this row into cols*/ 1468 cols = sbuf_aj_i + ct2; 1469 1470 lwrite = 0; 1471 for (l=0; l<nzB; l++) { 1472 if ((ctmp = bmap[cworkB[l]]) < cstart) cols[lwrite++] = ctmp; 1473 } 1474 for (l=0; l<nzA; l++) cols[lwrite++] = cstart + cworkA[l]; 1475 for (l=0; l<nzB; l++) { 1476 if ((ctmp = bmap[cworkB[l]]) >= cend) cols[lwrite++] = ctmp; 1477 } 1478 1479 ct2 += ncols; 1480 } 1481 ierr = MPI_Isend(sbuf_aj_i,req_size[i],MPIU_INT,req_source1[i],tag3,comm,s_waits3+i);CHKERRQ(ierr); 1482 } 1483 1484 /* create column map (cmap): global col of C -> local col of submat */ 1485 #if defined(PETSC_USE_CTABLE) 1486 if (!allcolumns) { 1487 ierr = PetscTableCreate(ncol+1,C->cmap->N+1,&cmap);CHKERRQ(ierr); 1488 ierr = PetscCalloc1(C->cmap->n,&cmap_loc);CHKERRQ(ierr); 1489 for (j=0; j<ncol; j++) { /* use array cmap_loc[] for local col indices */ 1490 if (icol[j] >= cstart && icol[j] <cend) { 1491 cmap_loc[icol[j] - cstart] = j+1; 1492 } else { /* use PetscTable for non-local col indices */ 1493 ierr = PetscTableAdd(cmap,icol[j]+1,j+1,INSERT_VALUES);CHKERRQ(ierr); 1494 } 1495 } 1496 } else { 1497 cmap = NULL; 1498 cmap_loc = NULL; 1499 } 1500 ierr = PetscCalloc1(C->rmap->n,&rmap_loc);CHKERRQ(ierr); 1501 #else 1502 if (!allcolumns) { 1503 ierr = PetscCalloc1(C->cmap->N,&cmap);CHKERRQ(ierr); 1504 for (j=0; j<ncol; j++) cmap[icol[j]] = j+1; 1505 } else { 1506 cmap = NULL; 1507 } 1508 #endif 1509 1510 /* Create lens for MatSeqAIJSetPreallocation() */ 1511 ierr = PetscCalloc1(nrow,&lens);CHKERRQ(ierr); 1512 1513 /* Compute lens from local part of C */ 1514 for (j=0; j<nrow; j++) { 1515 row = irow[j]; 1516 proc = row2proc[j]; 1517 if (proc == rank) { 1518 /* diagonal part A = c->A */ 1519 ncols = ai[row-rstart+1] - ai[row-rstart]; 1520 cols = aj + ai[row-rstart]; 1521 if (!allcolumns) { 1522 for (k=0; k<ncols; k++) { 1523 #if defined(PETSC_USE_CTABLE) 1524 tcol = cmap_loc[cols[k]]; 1525 #else 1526 tcol = cmap[cols[k]+cstart]; 1527 #endif 1528 if (tcol) lens[j]++; 1529 } 1530 } else { /* allcolumns */ 1531 lens[j] = ncols; 1532 } 1533 1534 /* off-diagonal part B = c->B */ 1535 ncols = bi[row-rstart+1] - bi[row-rstart]; 1536 cols = bj + bi[row-rstart]; 1537 if (!allcolumns) { 1538 for (k=0; k<ncols; k++) { 1539 #if defined(PETSC_USE_CTABLE) 1540 ierr = PetscTableFind(cmap,bmap[cols[k]]+1,&tcol);CHKERRQ(ierr); 1541 #else 1542 tcol = cmap[bmap[cols[k]]]; 1543 #endif 1544 if (tcol) lens[j]++; 1545 } 1546 } else { /* allcolumns */ 1547 lens[j] += ncols; 1548 } 1549 } 1550 } 1551 1552 /* Create row map (rmap): global row of C -> local row of submat */ 1553 #if defined(PETSC_USE_CTABLE) 1554 ierr = PetscTableCreate(nrow+1,C->rmap->N+1,&rmap);CHKERRQ(ierr); 1555 for (j=0; j<nrow; j++) { 1556 row = irow[j]; 1557 proc = row2proc[j]; 1558 if (proc == rank) { /* a local row */ 1559 rmap_loc[row - rstart] = j; 1560 } else { 1561 ierr = PetscTableAdd(rmap,irow[j]+1,j+1,INSERT_VALUES);CHKERRQ(ierr); 1562 } 1563 } 1564 #else 1565 ierr = PetscCalloc1(C->rmap->N,&rmap);CHKERRQ(ierr); 1566 for (j=0; j<nrow; j++) { 1567 rmap[irow[j]] = j; 1568 } 1569 #endif 1570 1571 /* Update lens from offproc data */ 1572 /* recv a->j is done */ 1573 ierr = MPI_Waitall(nrqs,r_waits3,r_status3);CHKERRQ(ierr); 1574 for (i=0; i<nrqs; i++) { 1575 proc = pa[i]; 1576 sbuf1_i = sbuf1[proc]; 1577 /* jmax = sbuf1_i[0]; if (jmax != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"jmax !=1"); */ 1578 ct1 = 2 + 1; 1579 ct2 = 0; 1580 rbuf2_i = rbuf2[i]; /* received length of C->j */ 1581 rbuf3_i = rbuf3[i]; /* received C->j */ 1582 1583 /* is_no = sbuf1_i[2*j-1]; if (is_no != 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"is_no !=0"); */ 1584 max1 = sbuf1_i[2]; 1585 for (k=0; k<max1; k++,ct1++) { 1586 #if defined(PETSC_USE_CTABLE) 1587 ierr = PetscTableFind(rmap,sbuf1_i[ct1]+1,&row);CHKERRQ(ierr); 1588 row--; 1589 if (row < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"row not found in table"); 1590 #else 1591 row = rmap[sbuf1_i[ct1]]; /* the row index in submat */ 1592 #endif 1593 /* Now, store row index of submat in sbuf1_i[ct1] */ 1594 sbuf1_i[ct1] = row; 1595 1596 nnz = rbuf2_i[ct1]; 1597 if (!allcolumns) { 1598 for (l=0; l<nnz; l++,ct2++) { 1599 #if defined(PETSC_USE_CTABLE) 1600 if (rbuf3_i[ct2] >= cstart && rbuf3_i[ct2] <cend) { 1601 tcol = cmap_loc[rbuf3_i[ct2] - cstart]; 1602 } else { 1603 ierr = PetscTableFind(cmap,rbuf3_i[ct2]+1,&tcol);CHKERRQ(ierr); 1604 } 1605 #else 1606 tcol = cmap[rbuf3_i[ct2]]; /* column index in submat */ 1607 #endif 1608 if (tcol) lens[row]++; 1609 } 1610 } else { /* allcolumns */ 1611 lens[row] += nnz; 1612 } 1613 } 1614 } 1615 ierr = MPI_Waitall(nrqr,s_waits3,s_status3);CHKERRQ(ierr); 1616 ierr = PetscFree4(r_waits3,s_waits3,r_status3,s_status3);CHKERRQ(ierr); 1617 1618 /* Create the submatrices */ 1619 ierr = MatCreate(PETSC_COMM_SELF,&submat);CHKERRQ(ierr); 1620 ierr = MatSetSizes(submat,nrow,ncol,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 1621 1622 ierr = ISGetBlockSize(isrow[0],&i);CHKERRQ(ierr); 1623 ierr = ISGetBlockSize(iscol[0],&j);CHKERRQ(ierr); 1624 ierr = MatSetBlockSizes(submat,i,j);CHKERRQ(ierr); 1625 ierr = MatSetType(submat,((PetscObject)A)->type_name);CHKERRQ(ierr); 1626 ierr = MatSeqAIJSetPreallocation(submat,0,lens);CHKERRQ(ierr); 1627 1628 /* create struct Mat_SubMat and attached it to submat */ 1629 ierr = PetscNew(&smatis1);CHKERRQ(ierr); 1630 subc = (Mat_SeqAIJ*)submat->data; 1631 subc->submatis1 = smatis1; 1632 1633 smatis1->id = 0; 1634 smatis1->nrqs = nrqs; 1635 smatis1->nrqr = nrqr; 1636 smatis1->rbuf1 = rbuf1; 1637 smatis1->rbuf2 = rbuf2; 1638 smatis1->rbuf3 = rbuf3; 1639 smatis1->sbuf2 = sbuf2; 1640 smatis1->req_source2 = req_source2; 1641 1642 smatis1->sbuf1 = sbuf1; 1643 smatis1->ptr = ptr; 1644 smatis1->tmp = tmp; 1645 smatis1->ctr = ctr; 1646 1647 smatis1->pa = pa; 1648 smatis1->req_size = req_size; 1649 smatis1->req_source1 = req_source1; 1650 1651 smatis1->allcolumns = allcolumns; 1652 smatis1->singleis = PETSC_TRUE; 1653 smatis1->row2proc = row2proc; 1654 smatis1->rmap = rmap; 1655 smatis1->cmap = cmap; 1656 #if defined(PETSC_USE_CTABLE) 1657 smatis1->rmap_loc = rmap_loc; 1658 smatis1->cmap_loc = cmap_loc; 1659 #endif 1660 1661 smatis1->destroy = submat->ops->destroy; 1662 submat->ops->destroy = MatDestroy_SeqAIJ_Submatrices; 1663 submat->factortype = C->factortype; 1664 1665 /* compute rmax */ 1666 rmax = 0; 1667 for (i=0; i<nrow; i++) rmax = PetscMax(rmax,lens[i]); 1668 1669 } else { /* scall == MAT_REUSE_MATRIX */ 1670 submat = submats[0]; 1671 if (submat->rmap->n != nrow || submat->cmap->n != ncol) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong size"); 1672 1673 subc = (Mat_SeqAIJ*)submat->data; 1674 rmax = subc->rmax; 1675 smatis1 = subc->submatis1; 1676 nrqs = smatis1->nrqs; 1677 nrqr = smatis1->nrqr; 1678 rbuf1 = smatis1->rbuf1; 1679 rbuf2 = smatis1->rbuf2; 1680 rbuf3 = smatis1->rbuf3; 1681 req_source2 = smatis1->req_source2; 1682 1683 sbuf1 = smatis1->sbuf1; 1684 sbuf2 = smatis1->sbuf2; 1685 ptr = smatis1->ptr; 1686 tmp = smatis1->tmp; 1687 ctr = smatis1->ctr; 1688 1689 pa = smatis1->pa; 1690 req_size = smatis1->req_size; 1691 req_source1 = smatis1->req_source1; 1692 1693 allcolumns = smatis1->allcolumns; 1694 row2proc = smatis1->row2proc; 1695 rmap = smatis1->rmap; 1696 cmap = smatis1->cmap; 1697 #if defined(PETSC_USE_CTABLE) 1698 rmap_loc = smatis1->rmap_loc; 1699 cmap_loc = smatis1->cmap_loc; 1700 #endif 1701 } 1702 1703 /* Post recv matrix values */ 1704 ierr = PetscMalloc3(nrqs+1,&rbuf4, rmax,&subcols, rmax,&subvals);CHKERRQ(ierr); 1705 ierr = PetscMalloc4(nrqs+1,&r_waits4,nrqr+1,&s_waits4,nrqs+1,&r_status4,nrqr+1,&s_status4);CHKERRQ(ierr); 1706 ierr = PetscObjectGetNewTag((PetscObject)C,&tag4);CHKERRQ(ierr); 1707 for (i=0; i<nrqs; ++i) { 1708 ierr = PetscMalloc1(rbuf2[i][0]+1,&rbuf4[i]);CHKERRQ(ierr); 1709 ierr = MPI_Irecv(rbuf4[i],rbuf2[i][0],MPIU_SCALAR,req_source2[i],tag4,comm,r_waits4+i);CHKERRQ(ierr); 1710 } 1711 1712 /* Allocate sending buffers for a->a, and send them off */ 1713 ierr = PetscMalloc1(nrqr+1,&sbuf_aa);CHKERRQ(ierr); 1714 for (i=0,j=0; i<nrqr; i++) j += req_size[i]; 1715 ierr = PetscMalloc1(j+1,&sbuf_aa[0]);CHKERRQ(ierr); 1716 for (i=1; i<nrqr; i++) sbuf_aa[i] = sbuf_aa[i-1] + req_size[i-1]; 1717 1718 for (i=0; i<nrqr; i++) { 1719 rbuf1_i = rbuf1[i]; 1720 sbuf_aa_i = sbuf_aa[i]; 1721 ct1 = 2*rbuf1_i[0]+1; 1722 ct2 = 0; 1723 /* max1=rbuf1_i[0]; if (max1 != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"max1 !=1"); */ 1724 1725 kmax = rbuf1_i[2]; 1726 for (k=0; k<kmax; k++,ct1++) { 1727 row = rbuf1_i[ct1] - rstart; 1728 nzA = ai[row+1] - ai[row]; 1729 nzB = bi[row+1] - bi[row]; 1730 ncols = nzA + nzB; 1731 cworkB = bj + bi[row]; 1732 vworkA = a_a + ai[row]; 1733 vworkB = b_a + bi[row]; 1734 1735 /* load the column values for this row into vals*/ 1736 vals = sbuf_aa_i + ct2; 1737 1738 lwrite = 0; 1739 for (l=0; l<nzB; l++) { 1740 if ((bmap[cworkB[l]]) < cstart) vals[lwrite++] = vworkB[l]; 1741 } 1742 for (l=0; l<nzA; l++) vals[lwrite++] = vworkA[l]; 1743 for (l=0; l<nzB; l++) { 1744 if ((bmap[cworkB[l]]) >= cend) vals[lwrite++] = vworkB[l]; 1745 } 1746 1747 ct2 += ncols; 1748 } 1749 ierr = MPI_Isend(sbuf_aa_i,req_size[i],MPIU_SCALAR,req_source1[i],tag4,comm,s_waits4+i);CHKERRQ(ierr); 1750 } 1751 1752 /* Assemble submat */ 1753 /* First assemble the local rows */ 1754 for (j=0; j<nrow; j++) { 1755 row = irow[j]; 1756 proc = row2proc[j]; 1757 if (proc == rank) { 1758 Crow = row - rstart; /* local row index of C */ 1759 #if defined(PETSC_USE_CTABLE) 1760 row = rmap_loc[Crow]; /* row index of submat */ 1761 #else 1762 row = rmap[row]; 1763 #endif 1764 1765 if (allcolumns) { 1766 /* diagonal part A = c->A */ 1767 ncols = ai[Crow+1] - ai[Crow]; 1768 cols = aj + ai[Crow]; 1769 vals = a->a + ai[Crow]; 1770 i = 0; 1771 for (k=0; k<ncols; k++) { 1772 subcols[i] = cols[k] + cstart; 1773 subvals[i++] = vals[k]; 1774 } 1775 1776 /* off-diagonal part B = c->B */ 1777 ncols = bi[Crow+1] - bi[Crow]; 1778 cols = bj + bi[Crow]; 1779 vals = b->a + bi[Crow]; 1780 for (k=0; k<ncols; k++) { 1781 subcols[i] = bmap[cols[k]]; 1782 subvals[i++] = vals[k]; 1783 } 1784 1785 ierr = MatSetValues_SeqAIJ(submat,1,&row,i,subcols,subvals,INSERT_VALUES);CHKERRQ(ierr); 1786 1787 } else { /* !allcolumns */ 1788 #if defined(PETSC_USE_CTABLE) 1789 /* diagonal part A = c->A */ 1790 ncols = ai[Crow+1] - ai[Crow]; 1791 cols = aj + ai[Crow]; 1792 vals = a->a + ai[Crow]; 1793 i = 0; 1794 for (k=0; k<ncols; k++) { 1795 tcol = cmap_loc[cols[k]]; 1796 if (tcol) { 1797 subcols[i] = --tcol; 1798 subvals[i++] = vals[k]; 1799 } 1800 } 1801 1802 /* off-diagonal part B = c->B */ 1803 ncols = bi[Crow+1] - bi[Crow]; 1804 cols = bj + bi[Crow]; 1805 vals = b->a + bi[Crow]; 1806 for (k=0; k<ncols; k++) { 1807 ierr = PetscTableFind(cmap,bmap[cols[k]]+1,&tcol);CHKERRQ(ierr); 1808 if (tcol) { 1809 subcols[i] = --tcol; 1810 subvals[i++] = vals[k]; 1811 } 1812 } 1813 #else 1814 /* diagonal part A = c->A */ 1815 ncols = ai[Crow+1] - ai[Crow]; 1816 cols = aj + ai[Crow]; 1817 vals = a->a + ai[Crow]; 1818 i = 0; 1819 for (k=0; k<ncols; k++) { 1820 tcol = cmap[cols[k]+cstart]; 1821 if (tcol) { 1822 subcols[i] = --tcol; 1823 subvals[i++] = vals[k]; 1824 } 1825 } 1826 1827 /* off-diagonal part B = c->B */ 1828 ncols = bi[Crow+1] - bi[Crow]; 1829 cols = bj + bi[Crow]; 1830 vals = b->a + bi[Crow]; 1831 for (k=0; k<ncols; k++) { 1832 tcol = cmap[bmap[cols[k]]]; 1833 if (tcol) { 1834 subcols[i] = --tcol; 1835 subvals[i++] = vals[k]; 1836 } 1837 } 1838 #endif 1839 ierr = MatSetValues_SeqAIJ(submat,1,&row,i,subcols,subvals,INSERT_VALUES);CHKERRQ(ierr); 1840 } 1841 } 1842 } 1843 1844 /* Now assemble the off-proc rows */ 1845 for (i=0; i<nrqs; i++) { /* for each requested message */ 1846 /* recv values from other processes */ 1847 ierr = MPI_Waitany(nrqs,r_waits4,&idex,r_status4+i);CHKERRQ(ierr); 1848 proc = pa[idex]; 1849 sbuf1_i = sbuf1[proc]; 1850 /* jmax = sbuf1_i[0]; if (jmax != 1)SETERRQ1(PETSC_COMM_SELF,0,"jmax %d != 1",jmax); */ 1851 ct1 = 2 + 1; 1852 ct2 = 0; /* count of received C->j */ 1853 ct3 = 0; /* count of received C->j that will be inserted into submat */ 1854 rbuf2_i = rbuf2[idex]; /* int** received length of C->j from other processes */ 1855 rbuf3_i = rbuf3[idex]; /* int** received C->j from other processes */ 1856 rbuf4_i = rbuf4[idex]; /* scalar** received C->a from other processes */ 1857 1858 /* is_no = sbuf1_i[2*j-1]; if (is_no != 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"is_no !=0"); */ 1859 max1 = sbuf1_i[2]; /* num of rows */ 1860 for (k=0; k<max1; k++,ct1++) { /* for each recved row */ 1861 row = sbuf1_i[ct1]; /* row index of submat */ 1862 if (!allcolumns) { 1863 idex = 0; 1864 if (scall == MAT_INITIAL_MATRIX) { 1865 nnz = rbuf2_i[ct1]; /* num of C entries in this row */ 1866 for (l=0; l<nnz; l++,ct2++) { /* for each recved column */ 1867 #if defined(PETSC_USE_CTABLE) 1868 if (rbuf3_i[ct2] >= cstart && rbuf3_i[ct2] <cend) { 1869 tcol = cmap_loc[rbuf3_i[ct2] - cstart]; 1870 } else { 1871 ierr = PetscTableFind(cmap,rbuf3_i[ct2]+1,&tcol);CHKERRQ(ierr); 1872 } 1873 #else 1874 tcol = cmap[rbuf3_i[ct2]]; 1875 #endif 1876 if (tcol) { 1877 subcols[idex] = --tcol; 1878 subvals[idex++] = rbuf4_i[ct2]; 1879 1880 /* We receive an entire column of C, but a subset of it needs to be inserted into submat. 1881 For reuse, we replace received C->j with index that should be inserted to submat */ 1882 rbuf3_i[ct3++] = ct2; 1883 } 1884 } 1885 ierr = MatSetValues_SeqAIJ(submat,1,&row,idex,subcols,subvals,INSERT_VALUES);CHKERRQ(ierr); 1886 1887 } else { /* scall == MAT_REUSE_MATRIX */ 1888 submat = submats[0]; 1889 subc = (Mat_SeqAIJ*)submat->data; 1890 1891 nnz = subc->i[row+1] - subc->i[row]; /* num of submat entries in this row */ 1892 for (l=0; l<nnz; l++) { 1893 ct2 = rbuf3_i[ct3++]; /* index of rbuf4_i[] which needs to be inserted into submat */ 1894 subvals[idex++] = rbuf4_i[ct2]; 1895 } 1896 1897 bj = subc->j + subc->i[row]; 1898 ierr = MatSetValues_SeqAIJ(submat,1,&row,nnz,bj,subvals,INSERT_VALUES);CHKERRQ(ierr); 1899 } 1900 } else { /* allcolumns */ 1901 nnz = rbuf2_i[ct1]; /* num of C entries in this row */ 1902 ierr = MatSetValues_SeqAIJ(submat,1,&row,nnz,rbuf3_i+ct2,rbuf4_i+ct2,INSERT_VALUES);CHKERRQ(ierr); 1903 ct2 += nnz; 1904 } 1905 } 1906 } 1907 1908 /* sending a->a are done */ 1909 ierr = MPI_Waitall(nrqr,s_waits4,s_status4);CHKERRQ(ierr); 1910 ierr = PetscFree4(r_waits4,s_waits4,r_status4,s_status4);CHKERRQ(ierr); 1911 1912 ierr = MatAssemblyBegin(submat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1913 ierr = MatAssemblyEnd(submat,MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 1914 submats[0] = submat; 1915 1916 /* Restore the indices */ 1917 ierr = ISRestoreIndices(isrow[0],&irow);CHKERRQ(ierr); 1918 if (!allcolumns) { 1919 ierr = ISRestoreIndices(iscol[0],&icol);CHKERRQ(ierr); 1920 } 1921 1922 /* Destroy allocated memory */ 1923 for (i=0; i<nrqs; ++i) { 1924 ierr = PetscFree3(rbuf4[i],subcols,subvals);CHKERRQ(ierr); 1925 } 1926 ierr = PetscFree3(rbuf4,subcols,subvals);CHKERRQ(ierr); 1927 ierr = PetscFree(sbuf_aa[0]);CHKERRQ(ierr); 1928 ierr = PetscFree(sbuf_aa);CHKERRQ(ierr); 1929 1930 if (scall == MAT_INITIAL_MATRIX) { 1931 ierr = PetscFree(lens);CHKERRQ(ierr); 1932 ierr = PetscFree(sbuf_aj[0]);CHKERRQ(ierr); 1933 ierr = PetscFree(sbuf_aj);CHKERRQ(ierr); 1934 } 1935 PetscFunctionReturn(0); 1936 } 1937 1938 PetscErrorCode MatGetSubMatrices_MPIAIJ_SingleIS(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submat[]) 1939 { 1940 PetscErrorCode ierr; 1941 PetscInt ncol; 1942 PetscBool colflag,allcolumns=PETSC_FALSE; 1943 1944 PetscFunctionBegin; 1945 /* Allocate memory to hold all the submatrices */ 1946 if (scall == MAT_INITIAL_MATRIX) { 1947 ierr = PetscCalloc1(2,submat);CHKERRQ(ierr); 1948 } 1949 1950 /* Check for special case: each processor gets entire matrix columns */ 1951 ierr = ISIdentity(iscol[0],&colflag);CHKERRQ(ierr); 1952 ierr = ISGetLocalSize(iscol[0],&ncol);CHKERRQ(ierr); 1953 if (colflag && ncol == C->cmap->N) allcolumns = PETSC_TRUE; 1954 1955 ierr = MatGetSubMatrices_MPIAIJ_SingleIS_Local(C,ismax,isrow,iscol,scall,allcolumns,*submat);CHKERRQ(ierr); 1956 PetscFunctionReturn(0); 1957 } 1958 1959 PetscErrorCode MatGetSubMatrices_MPIAIJ(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submat[]) 1960 { 1961 PetscErrorCode ierr; 1962 PetscInt nmax,nstages,i,pos,max_no,nrow,ncol,in[2],out[2]; 1963 PetscBool rowflag,colflag,wantallmatrix=PETSC_FALSE; 1964 Mat_SeqAIJ *subc; 1965 Mat_SubMat *smat; 1966 1967 PetscFunctionBegin; 1968 /* Check for special case: each processor has a single IS */ 1969 if (C->submat_singleis) { /* flag is set in PCSetUp_ASM() to skip MPIU_Allreduce() */ 1970 ierr = MatGetSubMatrices_MPIAIJ_SingleIS(C,ismax,isrow,iscol,scall,submat);CHKERRQ(ierr); 1971 C->submat_singleis = PETSC_FALSE; /* resume its default value in case C will be used for non-singlis */ 1972 PetscFunctionReturn(0); 1973 } 1974 1975 /* Collect global wantallmatrix and nstages */ 1976 if (!C->cmap->N) nmax=20*1000000/sizeof(PetscInt); 1977 else nmax = 20*1000000 / (C->cmap->N * sizeof(PetscInt)); 1978 if (!nmax) nmax = 1; 1979 1980 if (scall == MAT_INITIAL_MATRIX) { 1981 /* Collect global wantallmatrix and nstages */ 1982 if (ismax == 1 && C->rmap->N == C->cmap->N) { 1983 ierr = ISIdentity(*isrow,&rowflag);CHKERRQ(ierr); 1984 ierr = ISIdentity(*iscol,&colflag);CHKERRQ(ierr); 1985 ierr = ISGetLocalSize(*isrow,&nrow);CHKERRQ(ierr); 1986 ierr = ISGetLocalSize(*iscol,&ncol);CHKERRQ(ierr); 1987 if (rowflag && colflag && nrow == C->rmap->N && ncol == C->cmap->N) { 1988 wantallmatrix = PETSC_TRUE; 1989 1990 ierr = PetscOptionsGetBool(((PetscObject)C)->options,((PetscObject)C)->prefix,"-use_fast_submatrix",&wantallmatrix,NULL);CHKERRQ(ierr); 1991 } 1992 } 1993 1994 /* Determine the number of stages through which submatrices are done 1995 Each stage will extract nmax submatrices. 1996 nmax is determined by the matrix column dimension. 1997 If the original matrix has 20M columns, only one submatrix per stage is allowed, etc. 1998 */ 1999 nstages = ismax/nmax + ((ismax % nmax) ? 1 : 0); /* local nstages */ 2000 2001 in[0] = -1*(PetscInt)wantallmatrix; 2002 in[1] = nstages; 2003 ierr = MPIU_Allreduce(in,out,2,MPIU_INT,MPI_MAX,PetscObjectComm((PetscObject)C));CHKERRQ(ierr); 2004 wantallmatrix = (PetscBool)(-out[0]); 2005 nstages = out[1]; /* Make sure every processor loops through the global nstages */ 2006 2007 } else { /* MAT_REUSE_MATRIX */ 2008 if (ismax) { 2009 subc = (Mat_SeqAIJ*)(*submat)[0]->data; 2010 smat = subc->submatis1; 2011 } else { /* (*submat)[0] is a dummy matrix */ 2012 smat = (Mat_SubMat*)(*submat)[0]->data; 2013 } 2014 if (!smat) { 2015 /* smat is not generated by MatGetSubMatrix_MPIAIJ_All(...,MAT_INITIAL_MATRIX,...) */ 2016 wantallmatrix = PETSC_TRUE; 2017 } else if (smat->singleis) { 2018 ierr = MatGetSubMatrices_MPIAIJ_SingleIS(C,ismax,isrow,iscol,scall,submat);CHKERRQ(ierr); 2019 PetscFunctionReturn(0); 2020 } else { 2021 nstages = smat->nstages; 2022 } 2023 } 2024 2025 if (wantallmatrix) { 2026 ierr = MatGetSubMatrix_MPIAIJ_All(C,MAT_GET_VALUES,scall,submat);CHKERRQ(ierr); 2027 PetscFunctionReturn(0); 2028 } 2029 2030 /* Allocate memory to hold all the submatrices and dummy submatrices */ 2031 if (scall == MAT_INITIAL_MATRIX) { 2032 ierr = PetscCalloc1(ismax+nstages,submat);CHKERRQ(ierr); 2033 } 2034 2035 for (i=0,pos=0; i<nstages; i++) { 2036 if (pos+nmax <= ismax) max_no = nmax; 2037 else if (pos == ismax) max_no = 0; 2038 else max_no = ismax-pos; 2039 2040 ierr = MatGetSubMatrices_MPIAIJ_Local(C,max_no,isrow+pos,iscol+pos,scall,*submat+pos);CHKERRQ(ierr); 2041 if (!max_no && scall == MAT_INITIAL_MATRIX) { /* submat[pos] is a dummy matrix */ 2042 smat = (Mat_SubMat*)(*submat)[pos]->data; 2043 smat->nstages = nstages; 2044 } 2045 pos += max_no; 2046 } 2047 2048 if (ismax && scall == MAT_INITIAL_MATRIX) { 2049 /* save nstages for reuse */ 2050 subc = (Mat_SeqAIJ*)(*submat)[0]->data; 2051 smat = subc->submatis1; 2052 smat->nstages = nstages; 2053 } 2054 PetscFunctionReturn(0); 2055 } 2056 2057 /* -------------------------------------------------------------------------*/ 2058 PetscErrorCode MatGetSubMatrices_MPIAIJ_Local(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submats) 2059 { 2060 Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data; 2061 Mat A = c->A; 2062 Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)c->B->data,*subc; 2063 const PetscInt **icol,**irow; 2064 PetscInt *nrow,*ncol,start; 2065 PetscErrorCode ierr; 2066 PetscMPIInt rank,size,tag0,tag2,tag3,tag4,*w1,*w2,*w3,*w4,nrqr; 2067 PetscInt **sbuf1,**sbuf2,i,j,k,l,ct1,ct2,**rbuf1,row,proc=-1; 2068 PetscInt nrqs=0,msz,**ptr=NULL,*req_size=NULL,*ctr=NULL,*pa,*tmp=NULL,tcol; 2069 PetscInt **rbuf3=NULL,*req_source1=NULL,*req_source2,**sbuf_aj,**rbuf2=NULL,max1,max2; 2070 PetscInt **lens,is_no,ncols,*cols,mat_i,*mat_j,tmp2,jmax; 2071 #if defined(PETSC_USE_CTABLE) 2072 PetscTable *cmap,cmap_i=NULL,*rmap,rmap_i; 2073 #else 2074 PetscInt **cmap,*cmap_i=NULL,**rmap,*rmap_i; 2075 #endif 2076 const PetscInt *irow_i; 2077 PetscInt ctr_j,*sbuf1_j,*sbuf_aj_i,*rbuf1_i,kmax,*lens_i; 2078 MPI_Request *s_waits1,*r_waits1,*s_waits2,*r_waits2,*r_waits3; 2079 MPI_Request *r_waits4,*s_waits3,*s_waits4; 2080 MPI_Status *r_status1,*r_status2,*s_status1,*s_status3,*s_status2; 2081 MPI_Status *r_status3,*r_status4,*s_status4; 2082 MPI_Comm comm; 2083 PetscScalar **rbuf4,*rbuf4_i,**sbuf_aa,*vals,*mat_a,*imat_a,*sbuf_aa_i; 2084 PetscMPIInt *onodes1,*olengths1,end; 2085 PetscInt **row2proc,*row2proc_i,ilen_row,*imat_ilen,*imat_j,*imat_i,old_row; 2086 Mat_SubMat *smat_i; 2087 PetscBool *issorted,*allcolumns,colflag,iscsorted=PETSC_TRUE; 2088 PetscInt *sbuf1_i,*rbuf2_i,*rbuf3_i,ilen; 2089 2090 PetscFunctionBegin; 2091 ierr = PetscObjectGetComm((PetscObject)C,&comm);CHKERRQ(ierr); 2092 size = c->size; 2093 rank = c->rank; 2094 2095 ierr = PetscMalloc4(ismax,&row2proc,ismax,&cmap,ismax,&rmap,ismax+1,&allcolumns);CHKERRQ(ierr); 2096 ierr = PetscMalloc5(ismax,&irow,ismax,&icol,ismax,&nrow,ismax,&ncol,ismax,&issorted);CHKERRQ(ierr); 2097 2098 for (i=0; i<ismax; i++) { 2099 ierr = ISSorted(iscol[i],&issorted[i]);CHKERRQ(ierr); 2100 if (!issorted[i]) iscsorted = issorted[i]; 2101 2102 ierr = ISSorted(isrow[i],&issorted[i]);CHKERRQ(ierr); 2103 2104 ierr = ISGetIndices(isrow[i],&irow[i]);CHKERRQ(ierr); 2105 ierr = ISGetLocalSize(isrow[i],&nrow[i]);CHKERRQ(ierr); 2106 2107 /* Check for special case: allcolumn */ 2108 ierr = ISIdentity(iscol[i],&colflag);CHKERRQ(ierr); 2109 ierr = ISGetLocalSize(iscol[i],&ncol[i]);CHKERRQ(ierr); 2110 if (colflag && ncol[i] == C->cmap->N) { 2111 allcolumns[i] = PETSC_TRUE; 2112 icol[i] = NULL; 2113 } else { 2114 allcolumns[i] = PETSC_FALSE; 2115 ierr = ISGetIndices(iscol[i],&icol[i]);CHKERRQ(ierr); 2116 } 2117 } 2118 2119 if (scall == MAT_REUSE_MATRIX) { 2120 /* Assumes new rows are same length as the old rows */ 2121 for (i=0; i<ismax; i++) { 2122 if (!submats[i]) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_NULL,"submats[%D] is null, cannot reuse",i); 2123 subc = (Mat_SeqAIJ*)submats[i]->data; 2124 if ((submats[i]->rmap->n != nrow[i]) || (submats[i]->cmap->n != ncol[i])) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong size"); 2125 2126 /* Initial matrix as if empty */ 2127 ierr = PetscMemzero(subc->ilen,submats[i]->rmap->n*sizeof(PetscInt));CHKERRQ(ierr); 2128 2129 smat_i = subc->submatis1; 2130 2131 nrqs = smat_i->nrqs; 2132 nrqr = smat_i->nrqr; 2133 rbuf1 = smat_i->rbuf1; 2134 rbuf2 = smat_i->rbuf2; 2135 rbuf3 = smat_i->rbuf3; 2136 req_source2 = smat_i->req_source2; 2137 2138 sbuf1 = smat_i->sbuf1; 2139 sbuf2 = smat_i->sbuf2; 2140 ptr = smat_i->ptr; 2141 tmp = smat_i->tmp; 2142 ctr = smat_i->ctr; 2143 2144 pa = smat_i->pa; 2145 req_size = smat_i->req_size; 2146 req_source1 = smat_i->req_source1; 2147 2148 allcolumns[i] = smat_i->allcolumns; 2149 row2proc[i] = smat_i->row2proc; 2150 rmap[i] = smat_i->rmap; 2151 cmap[i] = smat_i->cmap; 2152 } 2153 2154 if (!ismax){ /* Get dummy submatrices and retrieve struct submatis1 */ 2155 if (!submats[0]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_NULL,"submats are null, cannot reuse"); 2156 smat_i = (Mat_SubMat*)submats[0]->data; 2157 2158 nrqs = smat_i->nrqs; 2159 nrqr = smat_i->nrqr; 2160 rbuf1 = smat_i->rbuf1; 2161 rbuf2 = smat_i->rbuf2; 2162 rbuf3 = smat_i->rbuf3; 2163 req_source2 = smat_i->req_source2; 2164 2165 sbuf1 = smat_i->sbuf1; 2166 sbuf2 = smat_i->sbuf2; 2167 ptr = smat_i->ptr; 2168 tmp = smat_i->tmp; 2169 ctr = smat_i->ctr; 2170 2171 pa = smat_i->pa; 2172 req_size = smat_i->req_size; 2173 req_source1 = smat_i->req_source1; 2174 2175 allcolumns[0] = PETSC_TRUE; 2176 } 2177 } else { /* scall == MAT_INITIAL_MATRIX */ 2178 /* Get some new tags to keep the communication clean */ 2179 ierr = PetscObjectGetNewTag((PetscObject)C,&tag2);CHKERRQ(ierr); 2180 ierr = PetscObjectGetNewTag((PetscObject)C,&tag3);CHKERRQ(ierr); 2181 2182 /* evaluate communication - mesg to who, length of mesg, and buffer space 2183 required. Based on this, buffers are allocated, and data copied into them*/ 2184 ierr = PetscCalloc4(size,&w1,size,&w2,size,&w3,size,&w4);CHKERRQ(ierr); /* mesg size, initialize work vectors */ 2185 2186 for (i=0; i<ismax; i++) { 2187 jmax = nrow[i]; 2188 irow_i = irow[i]; 2189 2190 ierr = PetscMalloc1(jmax,&row2proc_i);CHKERRQ(ierr); 2191 row2proc[i] = row2proc_i; 2192 2193 if (issorted[i]) proc = 0; 2194 for (j=0; j<jmax; j++) { 2195 if (!issorted[i]) proc = 0; 2196 row = irow_i[j]; 2197 while (row >= C->rmap->range[proc+1]) proc++; 2198 w4[proc]++; 2199 row2proc_i[j] = proc; /* map row index to proc */ 2200 } 2201 for (j=0; j<size; j++) { 2202 if (w4[j]) { w1[j] += w4[j]; w3[j]++; w4[j] = 0;} 2203 } 2204 } 2205 2206 nrqs = 0; /* no of outgoing messages */ 2207 msz = 0; /* total mesg length (for all procs) */ 2208 w1[rank] = 0; /* no mesg sent to self */ 2209 w3[rank] = 0; 2210 for (i=0; i<size; i++) { 2211 if (w1[i]) { w2[i] = 1; nrqs++;} /* there exists a message to proc i */ 2212 } 2213 ierr = PetscMalloc1(nrqs+1,&pa);CHKERRQ(ierr); /*(proc -array)*/ 2214 for (i=0,j=0; i<size; i++) { 2215 if (w1[i]) { pa[j] = i; j++; } 2216 } 2217 2218 /* Each message would have a header = 1 + 2*(no of IS) + data */ 2219 for (i=0; i<nrqs; i++) { 2220 j = pa[i]; 2221 w1[j] += w2[j] + 2* w3[j]; 2222 msz += w1[j]; 2223 } 2224 ierr = PetscInfo2(0,"Number of outgoing messages %D Total message length %D\n",nrqs,msz);CHKERRQ(ierr); 2225 2226 /* Determine the number of messages to expect, their lengths, from from-ids */ 2227 ierr = PetscGatherNumberOfMessages(comm,w2,w1,&nrqr);CHKERRQ(ierr); 2228 ierr = PetscGatherMessageLengths(comm,nrqs,nrqr,w1,&onodes1,&olengths1);CHKERRQ(ierr); 2229 2230 /* Now post the Irecvs corresponding to these messages */ 2231 tag0 = ((PetscObject)C)->tag; 2232 ierr = PetscPostIrecvInt(comm,tag0,nrqr,onodes1,olengths1,&rbuf1,&r_waits1);CHKERRQ(ierr); 2233 2234 ierr = PetscFree(onodes1);CHKERRQ(ierr); 2235 ierr = PetscFree(olengths1);CHKERRQ(ierr); 2236 2237 /* Allocate Memory for outgoing messages */ 2238 ierr = PetscMalloc4(size,&sbuf1,size,&ptr,2*msz,&tmp,size,&ctr);CHKERRQ(ierr); 2239 ierr = PetscMemzero(sbuf1,size*sizeof(PetscInt*));CHKERRQ(ierr); 2240 ierr = PetscMemzero(ptr,size*sizeof(PetscInt*));CHKERRQ(ierr); 2241 2242 { 2243 PetscInt *iptr = tmp; 2244 k = 0; 2245 for (i=0; i<nrqs; i++) { 2246 j = pa[i]; 2247 iptr += k; 2248 sbuf1[j] = iptr; 2249 k = w1[j]; 2250 } 2251 } 2252 2253 /* Form the outgoing messages. Initialize the header space */ 2254 for (i=0; i<nrqs; i++) { 2255 j = pa[i]; 2256 sbuf1[j][0] = 0; 2257 ierr = PetscMemzero(sbuf1[j]+1,2*w3[j]*sizeof(PetscInt));CHKERRQ(ierr); 2258 ptr[j] = sbuf1[j] + 2*w3[j] + 1; 2259 } 2260 2261 /* Parse the isrow and copy data into outbuf */ 2262 for (i=0; i<ismax; i++) { 2263 row2proc_i = row2proc[i]; 2264 ierr = PetscMemzero(ctr,size*sizeof(PetscInt));CHKERRQ(ierr); 2265 irow_i = irow[i]; 2266 jmax = nrow[i]; 2267 for (j=0; j<jmax; j++) { /* parse the indices of each IS */ 2268 proc = row2proc_i[j]; 2269 if (proc != rank) { /* copy to the outgoing buf*/ 2270 ctr[proc]++; 2271 *ptr[proc] = irow_i[j]; 2272 ptr[proc]++; 2273 } 2274 } 2275 /* Update the headers for the current IS */ 2276 for (j=0; j<size; j++) { /* Can Optimise this loop too */ 2277 if ((ctr_j = ctr[j])) { 2278 sbuf1_j = sbuf1[j]; 2279 k = ++sbuf1_j[0]; 2280 sbuf1_j[2*k] = ctr_j; 2281 sbuf1_j[2*k-1] = i; 2282 } 2283 } 2284 } 2285 2286 /* Now post the sends */ 2287 ierr = PetscMalloc1(nrqs+1,&s_waits1);CHKERRQ(ierr); 2288 for (i=0; i<nrqs; ++i) { 2289 j = pa[i]; 2290 ierr = MPI_Isend(sbuf1[j],w1[j],MPIU_INT,j,tag0,comm,s_waits1+i);CHKERRQ(ierr); 2291 } 2292 2293 /* Post Receives to capture the buffer size */ 2294 ierr = PetscMalloc1(nrqs+1,&r_waits2);CHKERRQ(ierr); 2295 ierr = PetscMalloc3(nrqs+1,&req_source2,nrqs+1,&rbuf2,nrqs+1,&rbuf3);CHKERRQ(ierr); 2296 rbuf2[0] = tmp + msz; 2297 for (i=1; i<nrqs; ++i) { 2298 rbuf2[i] = rbuf2[i-1]+w1[pa[i-1]]; 2299 } 2300 for (i=0; i<nrqs; ++i) { 2301 j = pa[i]; 2302 ierr = MPI_Irecv(rbuf2[i],w1[j],MPIU_INT,j,tag2,comm,r_waits2+i);CHKERRQ(ierr); 2303 } 2304 2305 /* Send to other procs the buf size they should allocate */ 2306 /* Receive messages*/ 2307 ierr = PetscMalloc1(nrqr+1,&s_waits2);CHKERRQ(ierr); 2308 ierr = PetscMalloc1(nrqr+1,&r_status1);CHKERRQ(ierr); 2309 ierr = PetscMalloc3(nrqr,&sbuf2,nrqr,&req_size,nrqr,&req_source1);CHKERRQ(ierr); 2310 { 2311 PetscInt *sAi = a->i,*sBi = b->i,id,rstart = C->rmap->rstart; 2312 PetscInt *sbuf2_i; 2313 2314 ierr = MPI_Waitall(nrqr,r_waits1,r_status1);CHKERRQ(ierr); 2315 for (i=0; i<nrqr; ++i) { 2316 req_size[i] = 0; 2317 rbuf1_i = rbuf1[i]; 2318 start = 2*rbuf1_i[0] + 1; 2319 ierr = MPI_Get_count(r_status1+i,MPIU_INT,&end);CHKERRQ(ierr); 2320 ierr = PetscMalloc1(end+1,&sbuf2[i]);CHKERRQ(ierr); 2321 sbuf2_i = sbuf2[i]; 2322 for (j=start; j<end; j++) { 2323 id = rbuf1_i[j] - rstart; 2324 ncols = sAi[id+1] - sAi[id] + sBi[id+1] - sBi[id]; 2325 sbuf2_i[j] = ncols; 2326 req_size[i] += ncols; 2327 } 2328 req_source1[i] = r_status1[i].MPI_SOURCE; 2329 /* form the header */ 2330 sbuf2_i[0] = req_size[i]; 2331 for (j=1; j<start; j++) sbuf2_i[j] = rbuf1_i[j]; 2332 2333 ierr = MPI_Isend(sbuf2_i,end,MPIU_INT,req_source1[i],tag2,comm,s_waits2+i);CHKERRQ(ierr); 2334 } 2335 } 2336 ierr = PetscFree(r_status1);CHKERRQ(ierr); 2337 ierr = PetscFree(r_waits1);CHKERRQ(ierr); 2338 ierr = PetscFree4(w1,w2,w3,w4);CHKERRQ(ierr); 2339 2340 /* Receive messages*/ 2341 ierr = PetscMalloc1(nrqs+1,&r_waits3);CHKERRQ(ierr); 2342 ierr = PetscMalloc1(nrqs+1,&r_status2);CHKERRQ(ierr); 2343 2344 ierr = MPI_Waitall(nrqs,r_waits2,r_status2);CHKERRQ(ierr); 2345 for (i=0; i<nrqs; ++i) { 2346 ierr = PetscMalloc1(rbuf2[i][0]+1,&rbuf3[i]);CHKERRQ(ierr); 2347 req_source2[i] = r_status2[i].MPI_SOURCE; 2348 ierr = MPI_Irecv(rbuf3[i],rbuf2[i][0],MPIU_INT,req_source2[i],tag3,comm,r_waits3+i);CHKERRQ(ierr); 2349 } 2350 ierr = PetscFree(r_status2);CHKERRQ(ierr); 2351 ierr = PetscFree(r_waits2);CHKERRQ(ierr); 2352 2353 /* Wait on sends1 and sends2 */ 2354 ierr = PetscMalloc1(nrqs+1,&s_status1);CHKERRQ(ierr); 2355 ierr = PetscMalloc1(nrqr+1,&s_status2);CHKERRQ(ierr); 2356 2357 if (nrqs) {ierr = MPI_Waitall(nrqs,s_waits1,s_status1);CHKERRQ(ierr);} 2358 if (nrqr) {ierr = MPI_Waitall(nrqr,s_waits2,s_status2);CHKERRQ(ierr);} 2359 ierr = PetscFree(s_status1);CHKERRQ(ierr); 2360 ierr = PetscFree(s_status2);CHKERRQ(ierr); 2361 ierr = PetscFree(s_waits1);CHKERRQ(ierr); 2362 ierr = PetscFree(s_waits2);CHKERRQ(ierr); 2363 2364 /* Now allocate sending buffers for a->j, and send them off */ 2365 ierr = PetscMalloc1(nrqr+1,&sbuf_aj);CHKERRQ(ierr); 2366 for (i=0,j=0; i<nrqr; i++) j += req_size[i]; 2367 ierr = PetscMalloc1(j+1,&sbuf_aj[0]);CHKERRQ(ierr); 2368 for (i=1; i<nrqr; i++) sbuf_aj[i] = sbuf_aj[i-1] + req_size[i-1]; 2369 2370 ierr = PetscMalloc1(nrqr+1,&s_waits3);CHKERRQ(ierr); 2371 { 2372 PetscInt nzA,nzB,*a_i = a->i,*b_i = b->i,lwrite; 2373 PetscInt *cworkA,*cworkB,cstart = C->cmap->rstart,rstart = C->rmap->rstart,*bmap = c->garray; 2374 PetscInt cend = C->cmap->rend; 2375 PetscInt *a_j = a->j,*b_j = b->j,ctmp; 2376 2377 for (i=0; i<nrqr; i++) { 2378 rbuf1_i = rbuf1[i]; 2379 sbuf_aj_i = sbuf_aj[i]; 2380 ct1 = 2*rbuf1_i[0] + 1; 2381 ct2 = 0; 2382 for (j=1,max1=rbuf1_i[0]; j<=max1; j++) { 2383 kmax = rbuf1[i][2*j]; 2384 for (k=0; k<kmax; k++,ct1++) { 2385 row = rbuf1_i[ct1] - rstart; 2386 nzA = a_i[row+1] - a_i[row]; nzB = b_i[row+1] - b_i[row]; 2387 ncols = nzA + nzB; 2388 cworkA = a_j + a_i[row]; cworkB = b_j + b_i[row]; 2389 2390 /* load the column indices for this row into cols */ 2391 cols = sbuf_aj_i + ct2; 2392 2393 lwrite = 0; 2394 for (l=0; l<nzB; l++) { 2395 if ((ctmp = bmap[cworkB[l]]) < cstart) cols[lwrite++] = ctmp; 2396 } 2397 for (l=0; l<nzA; l++) cols[lwrite++] = cstart + cworkA[l]; 2398 for (l=0; l<nzB; l++) { 2399 if ((ctmp = bmap[cworkB[l]]) >= cend) cols[lwrite++] = ctmp; 2400 } 2401 2402 ct2 += ncols; 2403 } 2404 } 2405 ierr = MPI_Isend(sbuf_aj_i,req_size[i],MPIU_INT,req_source1[i],tag3,comm,s_waits3+i);CHKERRQ(ierr); 2406 } 2407 } 2408 ierr = PetscMalloc2(nrqs+1,&r_status3,nrqr+1,&s_status3);CHKERRQ(ierr); 2409 2410 /* create col map: global col of C -> local col of submatrices */ 2411 { 2412 const PetscInt *icol_i; 2413 #if defined(PETSC_USE_CTABLE) 2414 for (i=0; i<ismax; i++) { 2415 if (!allcolumns[i]) { 2416 ierr = PetscTableCreate(ncol[i]+1,C->cmap->N+1,&cmap[i]);CHKERRQ(ierr); 2417 2418 jmax = ncol[i]; 2419 icol_i = icol[i]; 2420 cmap_i = cmap[i]; 2421 for (j=0; j<jmax; j++) { 2422 ierr = PetscTableAdd(cmap[i],icol_i[j]+1,j+1,INSERT_VALUES);CHKERRQ(ierr); 2423 } 2424 } else cmap[i] = NULL; 2425 } 2426 #else 2427 for (i=0; i<ismax; i++) { 2428 if (!allcolumns[i]) { 2429 ierr = PetscCalloc1(C->cmap->N,&cmap[i]);CHKERRQ(ierr); 2430 jmax = ncol[i]; 2431 icol_i = icol[i]; 2432 cmap_i = cmap[i]; 2433 for (j=0; j<jmax; j++) { 2434 cmap_i[icol_i[j]] = j+1; 2435 } 2436 } else cmap[i] = NULL; 2437 } 2438 #endif 2439 } 2440 2441 /* Create lens which is required for MatCreate... */ 2442 for (i=0,j=0; i<ismax; i++) j += nrow[i]; 2443 ierr = PetscMalloc1(ismax,&lens);CHKERRQ(ierr); 2444 2445 if (ismax) { 2446 ierr = PetscCalloc1(j,&lens[0]);CHKERRQ(ierr); 2447 } 2448 for (i=1; i<ismax; i++) lens[i] = lens[i-1] + nrow[i-1]; 2449 2450 /* Update lens from local data */ 2451 for (i=0; i<ismax; i++) { 2452 row2proc_i = row2proc[i]; 2453 jmax = nrow[i]; 2454 if (!allcolumns[i]) cmap_i = cmap[i]; 2455 irow_i = irow[i]; 2456 lens_i = lens[i]; 2457 for (j=0; j<jmax; j++) { 2458 row = irow_i[j]; 2459 proc = row2proc_i[j]; 2460 if (proc == rank) { 2461 ierr = MatGetRow_MPIAIJ(C,row,&ncols,&cols,0);CHKERRQ(ierr); 2462 if (!allcolumns[i]) { 2463 for (k=0; k<ncols; k++) { 2464 #if defined(PETSC_USE_CTABLE) 2465 ierr = PetscTableFind(cmap_i,cols[k]+1,&tcol);CHKERRQ(ierr); 2466 #else 2467 tcol = cmap_i[cols[k]]; 2468 #endif 2469 if (tcol) lens_i[j]++; 2470 } 2471 } else { /* allcolumns */ 2472 lens_i[j] = ncols; 2473 } 2474 ierr = MatRestoreRow_MPIAIJ(C,row,&ncols,&cols,0);CHKERRQ(ierr); 2475 } 2476 } 2477 } 2478 2479 /* Create row map: global row of C -> local row of submatrices */ 2480 #if defined(PETSC_USE_CTABLE) 2481 for (i=0; i<ismax; i++) { 2482 ierr = PetscTableCreate(nrow[i]+1,C->rmap->N+1,&rmap[i]);CHKERRQ(ierr); 2483 irow_i = irow[i]; 2484 jmax = nrow[i]; 2485 for (j=0; j<jmax; j++) { 2486 ierr = PetscTableAdd(rmap[i],irow_i[j]+1,j+1,INSERT_VALUES);CHKERRQ(ierr); 2487 } 2488 } 2489 #else 2490 for (i=0; i<ismax; i++) { 2491 ierr = PetscCalloc1(C->rmap->N,&rmap[i]);CHKERRQ(ierr); 2492 rmap_i = rmap[i]; 2493 irow_i = irow[i]; 2494 jmax = nrow[i]; 2495 for (j=0; j<jmax; j++) { 2496 rmap_i[irow_i[j]] = j; 2497 } 2498 } 2499 #endif 2500 2501 /* Update lens from offproc data */ 2502 { 2503 PetscInt *rbuf2_i,*rbuf3_i,*sbuf1_i; 2504 2505 ierr = MPI_Waitall(nrqs,r_waits3,r_status3);CHKERRQ(ierr); 2506 for (tmp2=0; tmp2<nrqs; tmp2++) { 2507 sbuf1_i = sbuf1[pa[tmp2]]; 2508 jmax = sbuf1_i[0]; 2509 ct1 = 2*jmax+1; 2510 ct2 = 0; 2511 rbuf2_i = rbuf2[tmp2]; 2512 rbuf3_i = rbuf3[tmp2]; 2513 for (j=1; j<=jmax; j++) { 2514 is_no = sbuf1_i[2*j-1]; 2515 max1 = sbuf1_i[2*j]; 2516 lens_i = lens[is_no]; 2517 if (!allcolumns[is_no]) cmap_i = cmap[is_no]; 2518 rmap_i = rmap[is_no]; 2519 for (k=0; k<max1; k++,ct1++) { 2520 #if defined(PETSC_USE_CTABLE) 2521 ierr = PetscTableFind(rmap_i,sbuf1_i[ct1]+1,&row);CHKERRQ(ierr); 2522 row--; 2523 if (row < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"row not found in table"); 2524 #else 2525 row = rmap_i[sbuf1_i[ct1]]; /* the val in the new matrix to be */ 2526 #endif 2527 max2 = rbuf2_i[ct1]; 2528 for (l=0; l<max2; l++,ct2++) { 2529 if (!allcolumns[is_no]) { 2530 #if defined(PETSC_USE_CTABLE) 2531 ierr = PetscTableFind(cmap_i,rbuf3_i[ct2]+1,&tcol);CHKERRQ(ierr); 2532 #else 2533 tcol = cmap_i[rbuf3_i[ct2]]; 2534 #endif 2535 if (tcol) lens_i[row]++; 2536 } else { /* allcolumns */ 2537 lens_i[row]++; /* lens_i[row] += max2 ? */ 2538 } 2539 } 2540 } 2541 } 2542 } 2543 } 2544 ierr = PetscFree(r_waits3);CHKERRQ(ierr); 2545 if (nrqr) {ierr = MPI_Waitall(nrqr,s_waits3,s_status3);CHKERRQ(ierr);} 2546 ierr = PetscFree2(r_status3,s_status3);CHKERRQ(ierr); 2547 ierr = PetscFree(s_waits3);CHKERRQ(ierr); 2548 2549 /* Create the submatrices */ 2550 for (i=0; i<ismax; i++) { 2551 PetscInt rbs,cbs; 2552 2553 ierr = ISGetBlockSize(isrow[i],&rbs);CHKERRQ(ierr); 2554 ierr = ISGetBlockSize(iscol[i],&cbs);CHKERRQ(ierr); 2555 2556 ierr = MatCreate(PETSC_COMM_SELF,submats+i);CHKERRQ(ierr); 2557 ierr = MatSetSizes(submats[i],nrow[i],ncol[i],PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 2558 2559 ierr = MatSetBlockSizes(submats[i],rbs,cbs);CHKERRQ(ierr); 2560 ierr = MatSetType(submats[i],((PetscObject)A)->type_name);CHKERRQ(ierr); 2561 ierr = MatSeqAIJSetPreallocation(submats[i],0,lens[i]);CHKERRQ(ierr); 2562 2563 /* create struct Mat_SubMat and attached it to submat */ 2564 ierr = PetscNew(&smat_i);CHKERRQ(ierr); 2565 subc = (Mat_SeqAIJ*)submats[i]->data; 2566 subc->submatis1 = smat_i; 2567 2568 smat_i->destroy = submats[i]->ops->destroy; 2569 submats[i]->ops->destroy = MatDestroy_SeqAIJ_Submatrices; 2570 submats[i]->factortype = C->factortype; 2571 2572 smat_i->id = i; 2573 smat_i->nrqs = nrqs; 2574 smat_i->nrqr = nrqr; 2575 smat_i->rbuf1 = rbuf1; 2576 smat_i->rbuf2 = rbuf2; 2577 smat_i->rbuf3 = rbuf3; 2578 smat_i->sbuf2 = sbuf2; 2579 smat_i->req_source2 = req_source2; 2580 2581 smat_i->sbuf1 = sbuf1; 2582 smat_i->ptr = ptr; 2583 smat_i->tmp = tmp; 2584 smat_i->ctr = ctr; 2585 2586 smat_i->pa = pa; 2587 smat_i->req_size = req_size; 2588 smat_i->req_source1 = req_source1; 2589 2590 smat_i->allcolumns = allcolumns[i]; 2591 smat_i->singleis = PETSC_FALSE; 2592 smat_i->row2proc = row2proc[i]; 2593 smat_i->rmap = rmap[i]; 2594 smat_i->cmap = cmap[i]; 2595 } 2596 2597 if (!ismax) { /* Create dummy submats[0] for reuse struct subc */ 2598 ierr = MatCreate(PETSC_COMM_SELF,&submats[0]);CHKERRQ(ierr); 2599 ierr = MatSetSizes(submats[0],0,0,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 2600 ierr = MatSetType(submats[0],MATDUMMY);CHKERRQ(ierr); 2601 2602 /* create struct Mat_SubMat and attached it to submat */ 2603 ierr = PetscNewLog(submats[0],&smat_i);CHKERRQ(ierr); 2604 submats[0]->data = (void*)smat_i; 2605 2606 smat_i->destroy = submats[0]->ops->destroy; 2607 submats[0]->ops->destroy = MatDestroy_Dummy_Submatrices; 2608 submats[0]->factortype = C->factortype; 2609 2610 smat_i->id = i; 2611 smat_i->nrqs = nrqs; 2612 smat_i->nrqr = nrqr; 2613 smat_i->rbuf1 = rbuf1; 2614 smat_i->rbuf2 = rbuf2; 2615 smat_i->rbuf3 = rbuf3; 2616 smat_i->sbuf2 = sbuf2; 2617 smat_i->req_source2 = req_source2; 2618 2619 smat_i->sbuf1 = sbuf1; 2620 smat_i->ptr = ptr; 2621 smat_i->tmp = tmp; 2622 smat_i->ctr = ctr; 2623 2624 smat_i->pa = pa; 2625 smat_i->req_size = req_size; 2626 smat_i->req_source1 = req_source1; 2627 2628 smat_i->allcolumns = PETSC_TRUE; 2629 smat_i->singleis = PETSC_FALSE; 2630 smat_i->row2proc = NULL; 2631 smat_i->rmap = NULL; 2632 smat_i->cmap = NULL; 2633 } 2634 2635 if (ismax) {ierr = PetscFree(lens[0]);CHKERRQ(ierr);} 2636 ierr = PetscFree(lens);CHKERRQ(ierr); 2637 ierr = PetscFree(sbuf_aj[0]);CHKERRQ(ierr); 2638 ierr = PetscFree(sbuf_aj);CHKERRQ(ierr); 2639 2640 } /* endof scall == MAT_INITIAL_MATRIX */ 2641 2642 /* Post recv matrix values */ 2643 ierr = PetscObjectGetNewTag((PetscObject)C,&tag4);CHKERRQ(ierr); 2644 ierr = PetscMalloc1(nrqs+1,&rbuf4);CHKERRQ(ierr); 2645 ierr = PetscMalloc1(nrqs+1,&r_waits4);CHKERRQ(ierr); 2646 ierr = PetscMalloc1(nrqs+1,&r_status4);CHKERRQ(ierr); 2647 ierr = PetscMalloc1(nrqr+1,&s_status4);CHKERRQ(ierr); 2648 for (i=0; i<nrqs; ++i) { 2649 ierr = PetscMalloc1(rbuf2[i][0]+1,&rbuf4[i]);CHKERRQ(ierr); 2650 ierr = MPI_Irecv(rbuf4[i],rbuf2[i][0],MPIU_SCALAR,req_source2[i],tag4,comm,r_waits4+i);CHKERRQ(ierr); 2651 } 2652 2653 /* Allocate sending buffers for a->a, and send them off */ 2654 ierr = PetscMalloc1(nrqr+1,&sbuf_aa);CHKERRQ(ierr); 2655 for (i=0,j=0; i<nrqr; i++) j += req_size[i]; 2656 ierr = PetscMalloc1(j+1,&sbuf_aa[0]);CHKERRQ(ierr); 2657 for (i=1; i<nrqr; i++) sbuf_aa[i] = sbuf_aa[i-1] + req_size[i-1]; 2658 2659 ierr = PetscMalloc1(nrqr+1,&s_waits4);CHKERRQ(ierr); 2660 { 2661 PetscInt nzA,nzB,*a_i = a->i,*b_i = b->i, *cworkB,lwrite; 2662 PetscInt cstart = C->cmap->rstart,rstart = C->rmap->rstart,*bmap = c->garray; 2663 PetscInt cend = C->cmap->rend; 2664 PetscInt *b_j = b->j; 2665 PetscScalar *vworkA,*vworkB,*a_a = a->a,*b_a = b->a; 2666 2667 for (i=0; i<nrqr; i++) { 2668 rbuf1_i = rbuf1[i]; 2669 sbuf_aa_i = sbuf_aa[i]; 2670 ct1 = 2*rbuf1_i[0]+1; 2671 ct2 = 0; 2672 for (j=1,max1=rbuf1_i[0]; j<=max1; j++) { 2673 kmax = rbuf1_i[2*j]; 2674 for (k=0; k<kmax; k++,ct1++) { 2675 row = rbuf1_i[ct1] - rstart; 2676 nzA = a_i[row+1] - a_i[row]; nzB = b_i[row+1] - b_i[row]; 2677 ncols = nzA + nzB; 2678 cworkB = b_j + b_i[row]; 2679 vworkA = a_a + a_i[row]; 2680 vworkB = b_a + b_i[row]; 2681 2682 /* load the column values for this row into vals*/ 2683 vals = sbuf_aa_i+ct2; 2684 2685 lwrite = 0; 2686 for (l=0; l<nzB; l++) { 2687 if ((bmap[cworkB[l]]) < cstart) vals[lwrite++] = vworkB[l]; 2688 } 2689 for (l=0; l<nzA; l++) vals[lwrite++] = vworkA[l]; 2690 for (l=0; l<nzB; l++) { 2691 if ((bmap[cworkB[l]]) >= cend) vals[lwrite++] = vworkB[l]; 2692 } 2693 2694 ct2 += ncols; 2695 } 2696 } 2697 ierr = MPI_Isend(sbuf_aa_i,req_size[i],MPIU_SCALAR,req_source1[i],tag4,comm,s_waits4+i);CHKERRQ(ierr); 2698 } 2699 } 2700 2701 /* Assemble the matrices */ 2702 /* First assemble the local rows */ 2703 for (i=0; i<ismax; i++) { 2704 row2proc_i = row2proc[i]; 2705 subc = (Mat_SeqAIJ*)submats[i]->data; 2706 imat_ilen = subc->ilen; 2707 imat_j = subc->j; 2708 imat_i = subc->i; 2709 imat_a = subc->a; 2710 2711 if (!allcolumns[i]) cmap_i = cmap[i]; 2712 rmap_i = rmap[i]; 2713 irow_i = irow[i]; 2714 jmax = nrow[i]; 2715 for (j=0; j<jmax; j++) { 2716 row = irow_i[j]; 2717 proc = row2proc_i[j]; 2718 if (proc == rank) { 2719 old_row = row; 2720 #if defined(PETSC_USE_CTABLE) 2721 ierr = PetscTableFind(rmap_i,row+1,&row);CHKERRQ(ierr); 2722 row--; 2723 #else 2724 row = rmap_i[row]; 2725 #endif 2726 ilen_row = imat_ilen[row]; 2727 ierr = MatGetRow_MPIAIJ(C,old_row,&ncols,&cols,&vals);CHKERRQ(ierr); 2728 mat_i = imat_i[row]; 2729 mat_a = imat_a + mat_i; 2730 mat_j = imat_j + mat_i; 2731 if (!allcolumns[i]) { 2732 for (k=0; k<ncols; k++) { 2733 #if defined(PETSC_USE_CTABLE) 2734 ierr = PetscTableFind(cmap_i,cols[k]+1,&tcol);CHKERRQ(ierr); 2735 #else 2736 tcol = cmap_i[cols[k]]; 2737 #endif 2738 if (tcol) { 2739 *mat_j++ = tcol - 1; 2740 *mat_a++ = vals[k]; 2741 ilen_row++; 2742 } 2743 } 2744 } else { /* allcolumns */ 2745 for (k=0; k<ncols; k++) { 2746 *mat_j++ = cols[k]; /* global col index! */ 2747 *mat_a++ = vals[k]; 2748 ilen_row++; 2749 } 2750 } 2751 ierr = MatRestoreRow_MPIAIJ(C,old_row,&ncols,&cols,&vals);CHKERRQ(ierr); 2752 2753 imat_ilen[row] = ilen_row; 2754 } 2755 } 2756 } 2757 2758 /* Now assemble the off proc rows */ 2759 ierr = MPI_Waitall(nrqs,r_waits4,r_status4);CHKERRQ(ierr); 2760 for (tmp2=0; tmp2<nrqs; tmp2++) { 2761 sbuf1_i = sbuf1[pa[tmp2]]; 2762 jmax = sbuf1_i[0]; 2763 ct1 = 2*jmax + 1; 2764 ct2 = 0; 2765 rbuf2_i = rbuf2[tmp2]; 2766 rbuf3_i = rbuf3[tmp2]; 2767 rbuf4_i = rbuf4[tmp2]; 2768 for (j=1; j<=jmax; j++) { 2769 is_no = sbuf1_i[2*j-1]; 2770 rmap_i = rmap[is_no]; 2771 if (!allcolumns[is_no]) cmap_i = cmap[is_no]; 2772 subc = (Mat_SeqAIJ*)submats[is_no]->data; 2773 imat_ilen = subc->ilen; 2774 imat_j = subc->j; 2775 imat_i = subc->i; 2776 imat_a = subc->a; 2777 max1 = sbuf1_i[2*j]; 2778 for (k=0; k<max1; k++,ct1++) { 2779 row = sbuf1_i[ct1]; 2780 #if defined(PETSC_USE_CTABLE) 2781 ierr = PetscTableFind(rmap_i,row+1,&row);CHKERRQ(ierr); 2782 row--; 2783 #else 2784 row = rmap_i[row]; 2785 #endif 2786 ilen = imat_ilen[row]; 2787 mat_i = imat_i[row]; 2788 mat_a = imat_a + mat_i; 2789 mat_j = imat_j + mat_i; 2790 max2 = rbuf2_i[ct1]; 2791 if (!allcolumns[is_no]) { 2792 for (l=0; l<max2; l++,ct2++) { 2793 #if defined(PETSC_USE_CTABLE) 2794 ierr = PetscTableFind(cmap_i,rbuf3_i[ct2]+1,&tcol);CHKERRQ(ierr); 2795 #else 2796 tcol = cmap_i[rbuf3_i[ct2]]; 2797 #endif 2798 if (tcol) { 2799 *mat_j++ = tcol - 1; 2800 *mat_a++ = rbuf4_i[ct2]; 2801 ilen++; 2802 } 2803 } 2804 } else { /* allcolumns */ 2805 for (l=0; l<max2; l++,ct2++) { 2806 *mat_j++ = rbuf3_i[ct2]; /* same global column index of C */ 2807 *mat_a++ = rbuf4_i[ct2]; 2808 ilen++; 2809 } 2810 } 2811 imat_ilen[row] = ilen; 2812 } 2813 } 2814 } 2815 2816 if (!iscsorted) { /* sort column indices of the rows */ 2817 for (i=0; i<ismax; i++) { 2818 subc = (Mat_SeqAIJ*)submats[i]->data; 2819 imat_j = subc->j; 2820 imat_i = subc->i; 2821 imat_a = subc->a; 2822 imat_ilen = subc->ilen; 2823 2824 if (allcolumns[i]) continue; 2825 jmax = nrow[i]; 2826 for (j=0; j<jmax; j++) { 2827 mat_i = imat_i[j]; 2828 mat_a = imat_a + mat_i; 2829 mat_j = imat_j + mat_i; 2830 ierr = PetscSortIntWithScalarArray(imat_ilen[j],mat_j,mat_a);CHKERRQ(ierr); 2831 } 2832 } 2833 } 2834 2835 ierr = PetscFree(r_status4);CHKERRQ(ierr); 2836 ierr = PetscFree(r_waits4);CHKERRQ(ierr); 2837 if (nrqr) {ierr = MPI_Waitall(nrqr,s_waits4,s_status4);CHKERRQ(ierr);} 2838 ierr = PetscFree(s_waits4);CHKERRQ(ierr); 2839 ierr = PetscFree(s_status4);CHKERRQ(ierr); 2840 2841 /* Restore the indices */ 2842 for (i=0; i<ismax; i++) { 2843 ierr = ISRestoreIndices(isrow[i],irow+i);CHKERRQ(ierr); 2844 if (!allcolumns[i]) { 2845 ierr = ISRestoreIndices(iscol[i],icol+i);CHKERRQ(ierr); 2846 } 2847 } 2848 2849 for (i=0; i<ismax; i++) { 2850 ierr = MatAssemblyBegin(submats[i],MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2851 ierr = MatAssemblyEnd(submats[i],MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 2852 } 2853 2854 /* Destroy allocated memory */ 2855 ierr = PetscFree(sbuf_aa[0]);CHKERRQ(ierr); 2856 ierr = PetscFree(sbuf_aa);CHKERRQ(ierr); 2857 ierr = PetscFree5(irow,icol,nrow,ncol,issorted);CHKERRQ(ierr); 2858 2859 for (i=0; i<nrqs; ++i) { 2860 ierr = PetscFree(rbuf4[i]);CHKERRQ(ierr); 2861 } 2862 ierr = PetscFree(rbuf4);CHKERRQ(ierr); 2863 2864 ierr = PetscFree4(row2proc,cmap,rmap,allcolumns);CHKERRQ(ierr); 2865 PetscFunctionReturn(0); 2866 } 2867 2868 /* 2869 Permute A & B into C's *local* index space using rowemb,dcolemb for A and rowemb,ocolemb for B. 2870 Embeddings are supposed to be injections and the above implies that the range of rowemb is a subset 2871 of [0,m), dcolemb is in [0,n) and ocolemb is in [N-n). 2872 If pattern == DIFFERENT_NONZERO_PATTERN, C is preallocated according to A&B. 2873 After that B's columns are mapped into C's global column space, so that C is in the "disassembled" 2874 state, and needs to be "assembled" later by compressing B's column space. 2875 2876 This function may be called in lieu of preallocation, so C should not be expected to be preallocated. 2877 Following this call, C->A & C->B have been created, even if empty. 2878 */ 2879 PetscErrorCode MatSetSeqMats_MPIAIJ(Mat C,IS rowemb,IS dcolemb,IS ocolemb,MatStructure pattern,Mat A,Mat B) 2880 { 2881 /* If making this function public, change the error returned in this function away from _PLIB. */ 2882 PetscErrorCode ierr; 2883 Mat_MPIAIJ *aij; 2884 Mat_SeqAIJ *Baij; 2885 PetscBool seqaij,Bdisassembled; 2886 PetscInt m,n,*nz,i,j,ngcol,col,rstart,rend,shift,count; 2887 PetscScalar v; 2888 const PetscInt *rowindices,*colindices; 2889 2890 PetscFunctionBegin; 2891 /* Check to make sure the component matrices (and embeddings) are compatible with C. */ 2892 if (A) { 2893 ierr = PetscObjectTypeCompare((PetscObject)A,MATSEQAIJ,&seqaij);CHKERRQ(ierr); 2894 if (!seqaij) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Diagonal matrix is of wrong type"); 2895 if (rowemb) { 2896 ierr = ISGetLocalSize(rowemb,&m);CHKERRQ(ierr); 2897 if (m != A->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Row IS of size %D is incompatible with diag matrix row size %D",m,A->rmap->n); 2898 } else { 2899 if (C->rmap->n != A->rmap->n) { 2900 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Diag seq matrix is row-incompatible with the MPIAIJ matrix"); 2901 } 2902 } 2903 if (dcolemb) { 2904 ierr = ISGetLocalSize(dcolemb,&n);CHKERRQ(ierr); 2905 if (n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Diag col IS of size %D is incompatible with diag matrix col size %D",n,A->cmap->n); 2906 } else { 2907 if (C->cmap->n != A->cmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Diag seq matrix is col-incompatible with the MPIAIJ matrix"); 2908 } 2909 } 2910 if (B) { 2911 ierr = PetscObjectTypeCompare((PetscObject)B,MATSEQAIJ,&seqaij);CHKERRQ(ierr); 2912 if (!seqaij) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Off-diagonal matrix is of wrong type"); 2913 if (rowemb) { 2914 ierr = ISGetLocalSize(rowemb,&m);CHKERRQ(ierr); 2915 if (m != B->rmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Row IS of size %D is incompatible with off-diag matrix row size %D",m,A->rmap->n); 2916 } else { 2917 if (C->rmap->n != B->rmap->n) { 2918 SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Off-diag seq matrix is row-incompatible with the MPIAIJ matrix"); 2919 } 2920 } 2921 if (ocolemb) { 2922 ierr = ISGetLocalSize(ocolemb,&n);CHKERRQ(ierr); 2923 if (n != B->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Off-diag col IS of size %D is incompatible with off-diag matrix col size %D",n,B->cmap->n); 2924 } else { 2925 if (C->cmap->N - C->cmap->n != B->cmap->n) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Off-diag seq matrix is col-incompatible with the MPIAIJ matrix"); 2926 } 2927 } 2928 2929 aij = (Mat_MPIAIJ*)C->data; 2930 if (!aij->A) { 2931 /* Mimic parts of MatMPIAIJSetPreallocation() */ 2932 ierr = MatCreate(PETSC_COMM_SELF,&aij->A);CHKERRQ(ierr); 2933 ierr = MatSetSizes(aij->A,C->rmap->n,C->cmap->n,C->rmap->n,C->cmap->n);CHKERRQ(ierr); 2934 ierr = MatSetBlockSizesFromMats(aij->A,C,C);CHKERRQ(ierr); 2935 ierr = MatSetType(aij->A,MATSEQAIJ);CHKERRQ(ierr); 2936 ierr = PetscLogObjectParent((PetscObject)C,(PetscObject)aij->A);CHKERRQ(ierr); 2937 } 2938 if (A) { 2939 ierr = MatSetSeqMat_SeqAIJ(aij->A,rowemb,dcolemb,pattern,A);CHKERRQ(ierr); 2940 } else { 2941 ierr = MatSetUp(aij->A);CHKERRQ(ierr); 2942 } 2943 if (B) { /* Destroy the old matrix or the column map, depending on the sparsity pattern. */ 2944 /* 2945 If pattern == DIFFERENT_NONZERO_PATTERN, we reallocate B and 2946 need to "disassemble" B -- convert it to using C's global indices. 2947 To insert the values we take the safer, albeit more expensive, route of MatSetValues(). 2948 2949 If pattern == SUBSET_NONZERO_PATTERN, we do not "disassemble" B and do not reallocate; 2950 we MatZeroValues(B) first, so there may be a bunch of zeros that, perhaps, could be compacted out. 2951 2952 TODO: Put B's values into aij->B's aij structure in place using the embedding ISs? 2953 At least avoid calling MatSetValues() and the implied searches? 2954 */ 2955 2956 if (B && pattern == DIFFERENT_NONZERO_PATTERN) { 2957 #if defined(PETSC_USE_CTABLE) 2958 ierr = PetscTableDestroy(&aij->colmap);CHKERRQ(ierr); 2959 #else 2960 ierr = PetscFree(aij->colmap);CHKERRQ(ierr); 2961 /* A bit of a HACK: ideally we should deal with case aij->B all in one code block below. */ 2962 if (aij->B) { 2963 ierr = PetscLogObjectMemory((PetscObject)C,-aij->B->cmap->n*sizeof(PetscInt));CHKERRQ(ierr); 2964 } 2965 #endif 2966 ngcol = 0; 2967 if (aij->lvec) { 2968 ierr = VecGetSize(aij->lvec,&ngcol);CHKERRQ(ierr); 2969 } 2970 if (aij->garray) { 2971 ierr = PetscFree(aij->garray);CHKERRQ(ierr); 2972 ierr = PetscLogObjectMemory((PetscObject)C,-ngcol*sizeof(PetscInt));CHKERRQ(ierr); 2973 } 2974 ierr = VecDestroy(&aij->lvec);CHKERRQ(ierr); 2975 ierr = VecScatterDestroy(&aij->Mvctx);CHKERRQ(ierr); 2976 } 2977 if (aij->B && B && pattern == DIFFERENT_NONZERO_PATTERN) { 2978 ierr = MatDestroy(&aij->B);CHKERRQ(ierr); 2979 } 2980 if (aij->B && B && pattern == SUBSET_NONZERO_PATTERN) { 2981 ierr = MatZeroEntries(aij->B);CHKERRQ(ierr); 2982 } 2983 } 2984 Bdisassembled = PETSC_FALSE; 2985 if (!aij->B) { 2986 ierr = MatCreate(PETSC_COMM_SELF,&aij->B);CHKERRQ(ierr); 2987 ierr = PetscLogObjectParent((PetscObject)C,(PetscObject)aij->B);CHKERRQ(ierr); 2988 ierr = MatSetSizes(aij->B,C->rmap->n,C->cmap->N,C->rmap->n,C->cmap->N);CHKERRQ(ierr); 2989 ierr = MatSetBlockSizesFromMats(aij->B,B,B);CHKERRQ(ierr); 2990 ierr = MatSetType(aij->B,MATSEQAIJ);CHKERRQ(ierr); 2991 Bdisassembled = PETSC_TRUE; 2992 } 2993 if (B) { 2994 Baij = (Mat_SeqAIJ*)B->data; 2995 if (pattern == DIFFERENT_NONZERO_PATTERN) { 2996 ierr = PetscMalloc1(B->rmap->n,&nz);CHKERRQ(ierr); 2997 for (i=0; i<B->rmap->n; i++) { 2998 nz[i] = Baij->i[i+1] - Baij->i[i]; 2999 } 3000 ierr = MatSeqAIJSetPreallocation(aij->B,0,nz);CHKERRQ(ierr); 3001 ierr = PetscFree(nz);CHKERRQ(ierr); 3002 } 3003 3004 ierr = PetscLayoutGetRange(C->rmap,&rstart,&rend);CHKERRQ(ierr); 3005 shift = rend-rstart; 3006 count = 0; 3007 rowindices = NULL; 3008 colindices = NULL; 3009 if (rowemb) { 3010 ierr = ISGetIndices(rowemb,&rowindices);CHKERRQ(ierr); 3011 } 3012 if (ocolemb) { 3013 ierr = ISGetIndices(ocolemb,&colindices);CHKERRQ(ierr); 3014 } 3015 for (i=0; i<B->rmap->n; i++) { 3016 PetscInt row; 3017 row = i; 3018 if (rowindices) row = rowindices[i]; 3019 for (j=Baij->i[i]; j<Baij->i[i+1]; j++) { 3020 col = Baij->j[count]; 3021 if (colindices) col = colindices[col]; 3022 if (Bdisassembled && col>=rstart) col += shift; 3023 v = Baij->a[count]; 3024 ierr = MatSetValues(aij->B,1,&row,1,&col,&v,INSERT_VALUES);CHKERRQ(ierr); 3025 ++count; 3026 } 3027 } 3028 /* No assembly for aij->B is necessary. */ 3029 /* FIXME: set aij->B's nonzerostate correctly. */ 3030 } else { 3031 ierr = MatSetUp(aij->B);CHKERRQ(ierr); 3032 } 3033 C->preallocated = PETSC_TRUE; 3034 C->was_assembled = PETSC_FALSE; 3035 C->assembled = PETSC_FALSE; 3036 /* 3037 C will need to be assembled so that aij->B can be compressed into local form in MatSetUpMultiply_MPIAIJ(). 3038 Furthermore, its nonzerostate will need to be based on that of aij->A's and aij->B's. 3039 */ 3040 PetscFunctionReturn(0); 3041 } 3042 3043 /* 3044 B uses local indices with column indices ranging between 0 and N-n; they must be interpreted using garray. 3045 */ 3046 PetscErrorCode MatGetSeqMats_MPIAIJ(Mat C,Mat *A,Mat *B) 3047 { 3048 Mat_MPIAIJ *aij = (Mat_MPIAIJ*)C->data; 3049 3050 PetscFunctionBegin; 3051 PetscValidPointer(A,2); 3052 PetscValidPointer(B,3); 3053 /* FIXME: make sure C is assembled */ 3054 *A = aij->A; 3055 *B = aij->B; 3056 /* Note that we don't incref *A and *B, so be careful! */ 3057 PetscFunctionReturn(0); 3058 } 3059 3060 /* 3061 Extract MPI submatrices encoded by pairs of IS that may live on subcomms of C. 3062 NOT SCALABLE due to the use of ISGetNonlocalIS() (see below). 3063 */ 3064 PetscErrorCode MatGetSubMatricesMPI_MPIXAIJ(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submat[], 3065 PetscErrorCode(*getsubmats_seq)(Mat,PetscInt,const IS[],const IS[],MatReuse,Mat**), 3066 PetscErrorCode(*getlocalmats)(Mat,Mat*,Mat*), 3067 PetscErrorCode(*setseqmat)(Mat,IS,IS,MatStructure,Mat), 3068 PetscErrorCode(*setseqmats)(Mat,IS,IS,IS,MatStructure,Mat,Mat)) 3069 { 3070 PetscErrorCode ierr; 3071 PetscMPIInt isize,flag; 3072 PetscInt i,ii,cismax,ispar; 3073 Mat *A,*B; 3074 IS *isrow_p,*iscol_p,*cisrow,*ciscol,*ciscol_p; 3075 3076 PetscFunctionBegin; 3077 if (!ismax) PetscFunctionReturn(0); 3078 3079 for (i = 0, cismax = 0; i < ismax; ++i) { 3080 PetscMPIInt isize; 3081 ierr = MPI_Comm_compare(((PetscObject)isrow[i])->comm,((PetscObject)iscol[i])->comm,&flag);CHKERRQ(ierr); 3082 if (flag != MPI_IDENT) SETERRQ(PETSC_COMM_SELF, PETSC_ERR_ARG_WRONG, "Row and column index sets must have the same communicator"); 3083 ierr = MPI_Comm_size(((PetscObject)isrow[i])->comm, &isize);CHKERRQ(ierr); 3084 if (isize > 1) ++cismax; 3085 } 3086 3087 /* 3088 If cismax is zero on all C's ranks, then and only then can we use purely sequential matrix extraction. 3089 ispar counts the number of parallel ISs across C's comm. 3090 */ 3091 ierr = MPIU_Allreduce(&cismax,&ispar,1,MPIU_INT,MPI_MAX,PetscObjectComm((PetscObject)C));CHKERRQ(ierr); 3092 if (!ispar) { /* Sequential ISs only across C's comm, so can call the sequential matrix extraction subroutine. */ 3093 ierr = (*getsubmats_seq)(C,ismax,isrow,iscol,scall,submat);CHKERRQ(ierr); 3094 PetscFunctionReturn(0); 3095 } 3096 3097 /* if (ispar) */ 3098 /* 3099 Construct the "complements" -- the off-processor indices -- of the iscol ISs for parallel ISs only. 3100 These are used to extract the off-diag portion of the resulting parallel matrix. 3101 The row IS for the off-diag portion is the same as for the diag portion, 3102 so we merely alias (without increfing) the row IS, while skipping those that are sequential. 3103 */ 3104 ierr = PetscMalloc2(cismax,&cisrow,cismax,&ciscol);CHKERRQ(ierr); 3105 ierr = PetscMalloc1(cismax,&ciscol_p);CHKERRQ(ierr); 3106 for (i = 0, ii = 0; i < ismax; ++i) { 3107 ierr = MPI_Comm_size(((PetscObject)isrow[i])->comm,&isize);CHKERRQ(ierr); 3108 if (isize > 1) { 3109 /* 3110 TODO: This is the part that's ***NOT SCALABLE***. 3111 To fix this we need to extract just the indices of C's nonzero columns 3112 that lie on the intersection of isrow[i] and ciscol[ii] -- the nonlocal 3113 part of iscol[i] -- without actually computing ciscol[ii]. This also has 3114 to be done without serializing on the IS list, so, most likely, it is best 3115 done by rewriting MatGetSubMatrices_MPIAIJ() directly. 3116 */ 3117 ierr = ISGetNonlocalIS(iscol[i],&(ciscol[ii]));CHKERRQ(ierr); 3118 /* Now we have to 3119 (a) make sure ciscol[ii] is sorted, since, even if the off-proc indices 3120 were sorted on each rank, concatenated they might no longer be sorted; 3121 (b) Use ISSortPermutation() to construct ciscol_p, the mapping from the 3122 indices in the nondecreasing order to the original index positions. 3123 If ciscol[ii] is strictly increasing, the permutation IS is NULL. 3124 */ 3125 ierr = ISSortPermutation(ciscol[ii],PETSC_FALSE,ciscol_p+ii);CHKERRQ(ierr); 3126 ierr = ISSort(ciscol[ii]);CHKERRQ(ierr); 3127 ++ii; 3128 } 3129 } 3130 ierr = PetscMalloc2(ismax,&isrow_p,ismax,&iscol_p);CHKERRQ(ierr); 3131 for (i = 0, ii = 0; i < ismax; ++i) { 3132 PetscInt j,issize; 3133 const PetscInt *indices; 3134 3135 /* 3136 Permute the indices into a nondecreasing order. Reject row and col indices with duplicates. 3137 */ 3138 ierr = ISSortPermutation(isrow[i],PETSC_FALSE,isrow_p+i);CHKERRQ(ierr); 3139 ierr = ISSort(isrow[i]);CHKERRQ(ierr); 3140 ierr = ISGetLocalSize(isrow[i],&issize);CHKERRQ(ierr); 3141 ierr = ISGetIndices(isrow[i],&indices);CHKERRQ(ierr); 3142 for (j = 1; j < issize; ++j) { 3143 if (indices[j] == indices[j-1]) { 3144 SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Repeated indices in row IS %D: indices at %D and %D are both %D",i,j-1,j,indices[j]); 3145 } 3146 } 3147 ierr = ISRestoreIndices(isrow[i],&indices);CHKERRQ(ierr); 3148 3149 3150 ierr = ISSortPermutation(iscol[i],PETSC_FALSE,iscol_p+i);CHKERRQ(ierr); 3151 ierr = ISSort(iscol[i]);CHKERRQ(ierr); 3152 ierr = ISGetLocalSize(iscol[i],&issize);CHKERRQ(ierr); 3153 ierr = ISGetIndices(iscol[i],&indices);CHKERRQ(ierr); 3154 for (j = 1; j < issize; ++j) { 3155 if (indices[j-1] == indices[j]) { 3156 SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Repeated indices in col IS %D: indices at %D and %D are both %D",i,j-1,j,indices[j]); 3157 } 3158 } 3159 ierr = ISRestoreIndices(iscol[i],&indices);CHKERRQ(ierr); 3160 ierr = MPI_Comm_size(((PetscObject)isrow[i])->comm,&isize);CHKERRQ(ierr); 3161 if (isize > 1) { 3162 cisrow[ii] = isrow[i]; 3163 ++ii; 3164 } 3165 } 3166 /* 3167 Allocate the necessary arrays to hold the resulting parallel matrices as well as the intermediate 3168 array of sequential matrices underlying the resulting parallel matrices. 3169 Which arrays to allocate is based on the value of MatReuse scall and whether ISs are sorted and/or 3170 contain duplicates. 3171 3172 There are as many diag matrices as there are original index sets. There are only as many parallel 3173 and off-diag matrices, as there are parallel (comm size > 1) index sets. 3174 3175 ARRAYS that can hold Seq matrices get allocated in any event -- either here or by getsubmats_seq(): 3176 - If the array of MPI matrices already exists and is being reused, we need to allocate the array 3177 and extract the underlying seq matrices into it to serve as placeholders, into which getsubmats_seq 3178 will deposite the extracted diag and off-diag parts. Thus, we allocate the A&B arrays and fill them 3179 with A[i] and B[ii] extracted from the corresponding MPI submat. 3180 - However, if the rows, A's column indices or B's column indices are not sorted, the extracted A[i] & B[ii] 3181 will have a different order from what getsubmats_seq expects. To handle this case -- indicated 3182 by a nonzero isrow_p[i], iscol_p[i], or ciscol_p[ii] -- we duplicate A[i] --> AA[i], B[ii] --> BB[ii] 3183 (retrieve composed AA[i] or BB[ii]) and reuse them here. AA[i] and BB[ii] are then used to permute its 3184 values into A[i] and B[ii] sitting inside the corresponding submat. 3185 - If no reuse is taking place then getsubmats_seq will allocate the A&B arrays and create the corresponding 3186 A[i], B[ii], AA[i] or BB[ii] matrices. 3187 */ 3188 /* Parallel matrix array is allocated here only if no reuse is taking place. If reused, it is passed in by the caller. */ 3189 if (scall == MAT_INITIAL_MATRIX) { 3190 ierr = PetscMalloc1(ismax,submat);CHKERRQ(ierr); 3191 } 3192 3193 /* Now obtain the sequential A and B submatrices separately. */ 3194 /* scall=MAT_REUSE_MATRIX is not handled yet, because getsubmats_seq() requires reuse of A and B */ 3195 ierr = (*getsubmats_seq)(C,ismax,isrow,iscol,MAT_INITIAL_MATRIX,&A);CHKERRQ(ierr); 3196 ierr = (*getsubmats_seq)(C,cismax,cisrow,ciscol,MAT_INITIAL_MATRIX,&B);CHKERRQ(ierr); 3197 3198 /* 3199 If scall == MAT_REUSE_MATRIX AND the permutations are NULL, we are done, since the sequential 3200 matrices A & B have been extracted directly into the parallel matrices containing them, or 3201 simply into the sequential matrix identical with the corresponding A (if isize == 1). 3202 Note that in that case colmap doesn't need to be rebuilt, since the matrices are expected 3203 to have the same sparsity pattern. 3204 Otherwise, A and/or B have to be properly embedded into C's index spaces and the correct colmap 3205 must be constructed for C. This is done by setseqmat(s). 3206 */ 3207 for (i = 0, ii = 0; i < ismax; ++i) { 3208 /* 3209 TODO: cache ciscol, permutation ISs and maybe cisrow? What about isrow & iscol? 3210 That way we can avoid sorting and computing permutations when reusing. 3211 To this end: 3212 - remove the old cache, if it exists, when extracting submatrices with MAT_INITIAL_MATRIX 3213 - if caching arrays to hold the ISs, make and compose a container for them so that it can 3214 be destroyed upon destruction of C (use PetscContainerUserDestroy() to clear out the contents). 3215 */ 3216 MatStructure pattern; 3217 pattern = DIFFERENT_NONZERO_PATTERN; 3218 3219 ierr = MPI_Comm_size(((PetscObject)isrow[i])->comm,&isize);CHKERRQ(ierr); 3220 /* Construct submat[i] from the Seq pieces A (and B, if necessary). */ 3221 if (isize > 1) { 3222 if (scall == MAT_INITIAL_MATRIX) { 3223 ierr = MatCreate(((PetscObject)isrow[i])->comm,(*submat)+i);CHKERRQ(ierr); 3224 ierr = MatSetSizes((*submat)[i],A[i]->rmap->n,A[i]->cmap->n,PETSC_DETERMINE,PETSC_DETERMINE);CHKERRQ(ierr); 3225 ierr = MatSetType((*submat)[i],MATMPIAIJ);CHKERRQ(ierr); 3226 ierr = PetscLayoutSetUp((*submat)[i]->rmap);CHKERRQ(ierr); 3227 ierr = PetscLayoutSetUp((*submat)[i]->cmap);CHKERRQ(ierr); 3228 } 3229 /* 3230 For each parallel isrow[i], insert the extracted sequential matrices into the parallel matrix. 3231 */ 3232 { 3233 Mat AA,BB; 3234 AA = A[i]; 3235 BB = B[ii]; 3236 if (AA || BB) { 3237 ierr = setseqmats((*submat)[i],isrow_p[i],iscol_p[i],ciscol_p[ii],pattern,AA,BB);CHKERRQ(ierr); 3238 ierr = MatAssemblyBegin((*submat)[i],MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3239 ierr = MatAssemblyEnd((*submat)[i],MAT_FINAL_ASSEMBLY);CHKERRQ(ierr); 3240 } 3241 3242 ierr = MatDestroy(&AA);CHKERRQ(ierr); 3243 } 3244 ierr = ISDestroy(ciscol+ii);CHKERRQ(ierr); 3245 ierr = ISDestroy(ciscol_p+ii);CHKERRQ(ierr); 3246 ++ii; 3247 } else { /* if (isize == 1) */ 3248 if (scall == MAT_REUSE_MATRIX) { 3249 ierr = MatDestroy(&(*submat)[i]);CHKERRQ(ierr); 3250 } 3251 if (isrow_p[i] || iscol_p[i]) { 3252 ierr = MatDuplicate(A[i],MAT_DO_NOT_COPY_VALUES,(*submat)+i);CHKERRQ(ierr); 3253 ierr = setseqmat((*submat)[i],isrow_p[i],iscol_p[i],pattern,A[i]);CHKERRQ(ierr); 3254 /* Otherwise A is extracted straight into (*submats)[i]. */ 3255 /* TODO: Compose A[i] on (*submat([i] for future use, if ((isrow_p[i] || iscol_p[i]) && MAT_INITIAL_MATRIX). */ 3256 ierr = MatDestroy(A+i);CHKERRQ(ierr); 3257 } else (*submat)[i] = A[i]; 3258 } 3259 ierr = ISDestroy(&isrow_p[i]);CHKERRQ(ierr); 3260 ierr = ISDestroy(&iscol_p[i]);CHKERRQ(ierr); 3261 } 3262 ierr = PetscFree2(cisrow,ciscol);CHKERRQ(ierr); 3263 ierr = PetscFree2(isrow_p,iscol_p);CHKERRQ(ierr); 3264 ierr = PetscFree(ciscol_p);CHKERRQ(ierr); 3265 ierr = PetscFree(A);CHKERRQ(ierr); 3266 ierr = MatDestroySubMatrices(cismax,&B);CHKERRQ(ierr); 3267 PetscFunctionReturn(0); 3268 } 3269 3270 PetscErrorCode MatGetSubMatricesMPI_MPIAIJ(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submat[]) 3271 { 3272 PetscErrorCode ierr; 3273 3274 PetscFunctionBegin; 3275 ierr = MatGetSubMatricesMPI_MPIXAIJ(C,ismax,isrow,iscol,scall,submat,MatGetSubMatrices_MPIAIJ,MatGetSeqMats_MPIAIJ,MatSetSeqMat_SeqAIJ,MatSetSeqMats_MPIAIJ);CHKERRQ(ierr); 3276 PetscFunctionReturn(0); 3277 } 3278