Actual source code: mpiov.c

  1: #define PETSCMAT_DLL

  3: /*
  4:    Routines to compute overlapping regions of a parallel MPI matrix
  5:   and to find submatrices that were shared across processors.
  6: */
 7:  #include src/mat/impls/aij/mpi/mpiaij.h
 8:  #include petscbt.h

 10: static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once(Mat,PetscInt,IS *);
 11: static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local(Mat,PetscInt,char **,PetscInt*,PetscInt**);
 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**);

 18: PetscErrorCode MatIncreaseOverlap_MPIAIJ(Mat C,PetscInt imax,IS is[],PetscInt ov)
 19: {
 21:   PetscInt       i;

 24:   if (ov < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative overlap specified");
 25:   for (i=0; i<ov; ++i) {
 26:     MatIncreaseOverlap_MPIAIJ_Once(C,imax,is);
 27:   }
 28:   return(0);
 29: }

 31: /*
 32:   Sample message format:
 33:   If a processor A wants processor B to process some elements corresponding
 34:   to index sets is[1],is[5]
 35:   mesg [0] = 2   (no of index sets in the mesg)
 36:   -----------  
 37:   mesg [1] = 1 => is[1]
 38:   mesg [2] = sizeof(is[1]);
 39:   -----------  
 40:   mesg [3] = 5  => is[5]
 41:   mesg [4] = sizeof(is[5]);
 42:   -----------
 43:   mesg [5] 
 44:   mesg [n]  datas[1]
 45:   -----------  
 46:   mesg[n+1]
 47:   mesg[m]  data(is[5])
 48:   -----------  
 49:   
 50:   Notes:
 51:   nrqs - no of requests sent (or to be sent out)
 52:   nrqr - no of requests recieved (which have to be or which have been processed
 53: */
 56: static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Once(Mat C,PetscInt imax,IS is[])
 57: {
 58:   Mat_MPIAIJ     *c = (Mat_MPIAIJ*)C->data;
 59:   PetscMPIInt    *w1,*w2,nrqr,*w3,*w4,*onodes1,*olengths1,*onodes2,*olengths2;
 60:   PetscInt       **idx,*n,*rtable,**data,len,*idx_i;
 62:   PetscMPIInt    size,rank,tag1,tag2;
 63:   PetscInt       m,i,j,k,**rbuf,row,proc,nrqs,msz,**outdat,**ptr;
 64:   PetscInt       *ctr,*pa,*tmp,*isz,*isz1,**xdata,**rbuf2;
 65:   PetscBT        *table;
 66:   MPI_Comm       comm;
 67:   MPI_Request    *s_waits1,*r_waits1,*s_waits2,*r_waits2;
 68:   MPI_Status     *s_status,*recv_status;

 71:   comm   = ((PetscObject)C)->comm;
 72:   size   = c->size;
 73:   rank   = c->rank;
 74:   m      = C->rmap.N;

 76:   PetscObjectGetNewTag((PetscObject)C,&tag1);
 77:   PetscObjectGetNewTag((PetscObject)C,&tag2);
 78: 
 79:   len    = (imax+1)*sizeof(PetscInt*)+ (imax + m)*sizeof(PetscInt);
 80:   PetscMalloc(len,&idx);
 81:   n      = (PetscInt*)(idx + imax);
 82:   rtable = n + imax;
 83: 
 84:   for (i=0; i<imax; i++) {
 85:     ISGetIndices(is[i],&idx[i]);
 86:     ISGetLocalSize(is[i],&n[i]);
 87:   }
 88: 
 89:   /* Create hash table for the mapping :row -> proc*/
 90:   for (i=0,j=0; i<size; i++) {
 91:     len = C->rmap.range[i+1];
 92:     for (; j<len; j++) {
 93:       rtable[j] = i;
 94:     }
 95:   }

 97:   /* evaluate communication - mesg to who,length of mesg, and buffer space
 98:      required. Based on this, buffers are allocated, and data copied into them*/
 99:   PetscMalloc(size*4*sizeof(PetscMPIInt),&w1);/*  mesg size */
100:   w2   = w1 + size;       /* if w2[i] marked, then a message to proc i*/
101:   w3   = w2 + size;       /* no of IS that needs to be sent to proc i */
102:   w4   = w3 + size;       /* temp work space used in determining w1, w2, w3 */
103:   PetscMemzero(w1,size*3*sizeof(PetscMPIInt)); /* initialise work vector*/
104:   for (i=0; i<imax; i++) {
105:     PetscMemzero(w4,size*sizeof(PetscMPIInt)); /* initialise work vector*/
106:     idx_i = idx[i];
107:     len   = n[i];
108:     for (j=0; j<len; j++) {
109:       row  = idx_i[j];
110:       if (row < 0) {
111:         SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Index set cannot have negative entries");
112:       }
113:       proc = rtable[row];
114:       w4[proc]++;
115:     }
116:     for (j=0; j<size; j++){
117:       if (w4[j]) { w1[j] += w4[j]; w3[j]++;}
118:     }
119:   }

121:   nrqs     = 0;              /* no of outgoing messages */
122:   msz      = 0;              /* total mesg length (for all proc */
123:   w1[rank] = 0;              /* no mesg sent to intself */
124:   w3[rank] = 0;
125:   for (i=0; i<size; i++) {
126:     if (w1[i])  {w2[i] = 1; nrqs++;} /* there exists a message to proc i */
127:   }
128:   /* pa - is list of processors to communicate with */
129:   PetscMalloc((nrqs+1)*sizeof(PetscInt),&pa);
130:   for (i=0,j=0; i<size; i++) {
131:     if (w1[i]) {pa[j] = i; j++;}
132:   }

134:   /* Each message would have a header = 1 + 2*(no of IS) + data */
135:   for (i=0; i<nrqs; i++) {
136:     j      = pa[i];
137:     w1[j] += w2[j] + 2*w3[j];
138:     msz   += w1[j];
139:   }

141:   /* Determine the number of messages to expect, their lengths, from from-ids */
142:   PetscGatherNumberOfMessages(comm,w2,w1,&nrqr);
143:   PetscGatherMessageLengths(comm,nrqs,nrqr,w1,&onodes1,&olengths1);

145:   /* Now post the Irecvs corresponding to these messages */
146:   PetscPostIrecvInt(comm,tag1,nrqr,onodes1,olengths1,&rbuf,&r_waits1);

148:   /* Allocate Memory for outgoing messages */
149:   len  = 2*size*sizeof(PetscInt*) + (size+msz)*sizeof(PetscInt);
150:   PetscMalloc(len,&outdat);
151:   ptr  = outdat + size;     /* Pointers to the data in outgoing buffers */
152:   PetscMemzero(outdat,2*size*sizeof(PetscInt*));
153:   tmp  = (PetscInt*)(outdat + 2*size);
154:   ctr  = tmp + msz;

156:   {
157:     PetscInt *iptr = tmp,ict  = 0;
158:     for (i=0; i<nrqs; i++) {
159:       j         = pa[i];
160:       iptr     +=  ict;
161:       outdat[j] = iptr;
162:       ict       = w1[j];
163:     }
164:   }

166:   /* Form the outgoing messages */
167:   /*plug in the headers*/
168:   for (i=0; i<nrqs; i++) {
169:     j            = pa[i];
170:     outdat[j][0] = 0;
171:     PetscMemzero(outdat[j]+1,2*w3[j]*sizeof(PetscInt));
172:     ptr[j]       = outdat[j] + 2*w3[j] + 1;
173:   }
174: 
175:   /* Memory for doing local proc's work*/
176:   {
177:     PetscInt  *d_p;
178:     char *t_p;

180:     len   = (imax)*(sizeof(PetscBT) + sizeof(PetscInt*)+ sizeof(PetscInt)) +
181:       (m)*imax*sizeof(PetscInt)  + (m/PETSC_BITS_PER_BYTE+1)*imax*sizeof(char) + 1;
182:     PetscMalloc(len,&table);
183:     PetscMemzero(table,len);
184:     data  = (PetscInt **)(table + imax);
185:     isz   = (PetscInt  *)(data  + imax);
186:     d_p   = (PetscInt  *)(isz   + imax);
187:     t_p   = (char *)(d_p   + m*imax);
188:     for (i=0; i<imax; i++) {
189:       table[i] = t_p + (m/PETSC_BITS_PER_BYTE+1)*i;
190:       data[i]  = d_p + (m)*i;
191:     }
192:   }

194:   /* Parse the IS and update local tables and the outgoing buf with the data*/
195:   {
196:     PetscInt     n_i,*data_i,isz_i,*outdat_j,ctr_j;
197:     PetscBT table_i;

199:     for (i=0; i<imax; i++) {
200:       PetscMemzero(ctr,size*sizeof(PetscInt));
201:       n_i     = n[i];
202:       table_i = table[i];
203:       idx_i   = idx[i];
204:       data_i  = data[i];
205:       isz_i   = isz[i];
206:       for (j=0;  j<n_i; j++) {  /* parse the indices of each IS */
207:         row  = idx_i[j];
208:         proc = rtable[row];
209:         if (proc != rank) { /* copy to the outgoing buffer */
210:           ctr[proc]++;
211:           *ptr[proc] = row;
212:           ptr[proc]++;
213:         } else { /* Update the local table */
214:           if (!PetscBTLookupSet(table_i,row)) { data_i[isz_i++] = row;}
215:         }
216:       }
217:       /* Update the headers for the current IS */
218:       for (j=0; j<size; j++) { /* Can Optimise this loop by using pa[] */
219:         if ((ctr_j = ctr[j])) {
220:           outdat_j        = outdat[j];
221:           k               = ++outdat_j[0];
222:           outdat_j[2*k]   = ctr_j;
223:           outdat_j[2*k-1] = i;
224:         }
225:       }
226:       isz[i] = isz_i;
227:     }
228:   }
229: 


232:   /*  Now  post the sends */
233:   PetscMalloc((nrqs+1)*sizeof(MPI_Request),&s_waits1);
234:   for (i=0; i<nrqs; ++i) {
235:     j    = pa[i];
236:     MPI_Isend(outdat[j],w1[j],MPIU_INT,j,tag1,comm,s_waits1+i);
237:   }
238: 
239:   /* No longer need the original indices*/
240:   for (i=0; i<imax; ++i) {
241:     ISRestoreIndices(is[i],idx+i);
242:   }
243:   PetscFree(idx);

245:   for (i=0; i<imax; ++i) {
246:     ISDestroy(is[i]);
247:   }
248: 
249:   /* Do Local work*/
250:   MatIncreaseOverlap_MPIAIJ_Local(C,imax,table,isz,data);

252:   /* Receive messages*/
253:   PetscMalloc((nrqr+1)*sizeof(MPI_Status),&recv_status);
254:   if (nrqr) {MPI_Waitall(nrqr,r_waits1,recv_status);}
255: 
256:   PetscMalloc((nrqs+1)*sizeof(MPI_Status),&s_status);
257:   if (nrqs) {MPI_Waitall(nrqs,s_waits1,s_status);}

259:   /* Phase 1 sends are complete - deallocate buffers */
260:   PetscFree(outdat);
261:   PetscFree(w1);

263:   PetscMalloc((nrqr+1)*sizeof(PetscInt*),&xdata);
264:   PetscMalloc((nrqr+1)*sizeof(PetscInt),&isz1);
265:   MatIncreaseOverlap_MPIAIJ_Receive(C,nrqr,rbuf,xdata,isz1);
266:   PetscFree(rbuf);

268: 
269:  /* Send the data back*/
270:   /* Do a global reduction to know the buffer space req for incoming messages*/
271:   {
272:     PetscMPIInt *rw1;
273: 
274:     PetscMalloc(size*sizeof(PetscMPIInt),&rw1);
275:     PetscMemzero(rw1,size*sizeof(PetscMPIInt));

277:     for (i=0; i<nrqr; ++i) {
278:       proc      = recv_status[i].MPI_SOURCE;
279:       if (proc != onodes1[i]) SETERRQ(PETSC_ERR_PLIB,"MPI_SOURCE mismatch");
280:       rw1[proc] = isz1[i];
281:     }
282:     PetscFree(onodes1);
283:     PetscFree(olengths1);

285:     /* Determine the number of messages to expect, their lengths, from from-ids */
286:     PetscGatherMessageLengths(comm,nrqr,nrqs,rw1,&onodes2,&olengths2);
287:     PetscFree(rw1);
288:   }
289:   /* Now post the Irecvs corresponding to these messages */
290:   PetscPostIrecvInt(comm,tag2,nrqs,onodes2,olengths2,&rbuf2,&r_waits2);

292:   /*  Now  post the sends */
293:   PetscMalloc((nrqr+1)*sizeof(MPI_Request),&s_waits2);
294:   for (i=0; i<nrqr; ++i) {
295:     j    = recv_status[i].MPI_SOURCE;
296:     MPI_Isend(xdata[i],isz1[i],MPIU_INT,j,tag2,comm,s_waits2+i);
297:   }

299:   /* receive work done on other processors*/
300:   {
301:     PetscInt    is_no,ct1,max,*rbuf2_i,isz_i,*data_i,jmax;
302:     PetscMPIInt idex;
303:     PetscBT     table_i;
304:     MPI_Status  *status2;
305: 
306:     PetscMalloc((PetscMax(nrqr,nrqs)+1)*sizeof(MPI_Status),&status2);
307:     for (i=0; i<nrqs; ++i) {
308:       MPI_Waitany(nrqs,r_waits2,&idex,status2+i);
309:       /* Process the message*/
310:       rbuf2_i = rbuf2[idex];
311:       ct1     = 2*rbuf2_i[0]+1;
312:       jmax    = rbuf2[idex][0];
313:       for (j=1; j<=jmax; j++) {
314:         max     = rbuf2_i[2*j];
315:         is_no   = rbuf2_i[2*j-1];
316:         isz_i   = isz[is_no];
317:         data_i  = data[is_no];
318:         table_i = table[is_no];
319:         for (k=0; k<max; k++,ct1++) {
320:           row = rbuf2_i[ct1];
321:           if (!PetscBTLookupSet(table_i,row)) { data_i[isz_i++] = row;}
322:         }
323:         isz[is_no] = isz_i;
324:       }
325:     }

327:     if (nrqr) {MPI_Waitall(nrqr,s_waits2,status2);}
328:     PetscFree(status2);
329:   }
330: 
331:   for (i=0; i<imax; ++i) {
332:     ISCreateGeneral(PETSC_COMM_SELF,isz[i],data[i],is+i);
333:   }
334: 
335:   PetscFree(onodes2);
336:   PetscFree(olengths2);

338:   PetscFree(pa);
339:   PetscFree(rbuf2);
340:   PetscFree(s_waits1);
341:   PetscFree(r_waits1);
342:   PetscFree(s_waits2);
343:   PetscFree(r_waits2);
344:   PetscFree(table);
345:   PetscFree(s_status);
346:   PetscFree(recv_status);
347:   PetscFree(xdata[0]);
348:   PetscFree(xdata);
349:   PetscFree(isz1);
350:   return(0);
351: }

355: /*  
356:    MatIncreaseOverlap_MPIAIJ_Local - Called by MatincreaseOverlap, to do 
357:        the work on the local processor.

359:      Inputs:
360:       C      - MAT_MPIAIJ;
361:       imax - total no of index sets processed at a time;
362:       table  - an array of char - size = m bits.
363:       
364:      Output:
365:       isz    - array containing the count of the solution elements correspondign
366:                to each index set;
367:       data   - pointer to the solutions
368: */
369: static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Local(Mat C,PetscInt imax,PetscBT *table,PetscInt *isz,PetscInt **data)
370: {
371:   Mat_MPIAIJ *c = (Mat_MPIAIJ*)C->data;
372:   Mat        A = c->A,B = c->B;
373:   Mat_SeqAIJ *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data;
374:   PetscInt   start,end,val,max,rstart,cstart,*ai,*aj;
375:   PetscInt   *bi,*bj,*garray,i,j,k,row,*data_i,isz_i;
376:   PetscBT    table_i;

379:   rstart = C->rmap.rstart;
380:   cstart = C->cmap.rstart;
381:   ai     = a->i;
382:   aj     = a->j;
383:   bi     = b->i;
384:   bj     = b->j;
385:   garray = c->garray;

387: 
388:   for (i=0; i<imax; i++) {
389:     data_i  = data[i];
390:     table_i = table[i];
391:     isz_i   = isz[i];
392:     for (j=0,max=isz[i]; j<max; j++) {
393:       row   = data_i[j] - rstart;
394:       start = ai[row];
395:       end   = ai[row+1];
396:       for (k=start; k<end; k++) { /* Amat */
397:         val = aj[k] + cstart;
398:         if (!PetscBTLookupSet(table_i,val)) { data_i[isz_i++] = val;}
399:       }
400:       start = bi[row];
401:       end   = bi[row+1];
402:       for (k=start; k<end; k++) { /* Bmat */
403:         val = garray[bj[k]];
404:         if (!PetscBTLookupSet(table_i,val)) { data_i[isz_i++] = val;}
405:       }
406:     }
407:     isz[i] = isz_i;
408:   }
409:   return(0);
410: }

414: /*     
415:       MatIncreaseOverlap_MPIAIJ_Receive - Process the recieved messages,
416:          and return the output

418:          Input:
419:            C    - the matrix
420:            nrqr - no of messages being processed.
421:            rbuf - an array of pointers to the recieved requests
422:            
423:          Output:
424:            xdata - array of messages to be sent back
425:            isz1  - size of each message

427:   For better efficiency perhaps we should malloc separately each xdata[i],
428: then if a remalloc is required we need only copy the data for that one row
429: rather then all previous rows as it is now where a single large chunck of 
430: memory is used.

432: */
433: static PetscErrorCode MatIncreaseOverlap_MPIAIJ_Receive(Mat C,PetscInt nrqr,PetscInt **rbuf,PetscInt **xdata,PetscInt * isz1)
434: {
435:   Mat_MPIAIJ     *c = (Mat_MPIAIJ*)C->data;
436:   Mat            A = c->A,B = c->B;
437:   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)B->data;
439:   PetscMPIInt    rank;
440:   PetscInt       rstart,cstart,*ai,*aj,*bi,*bj,*garray,i,j,k;
441:   PetscInt       row,total_sz,ct,ct1,ct2,ct3,mem_estimate,oct2,l,start,end;
442:   PetscInt       val,max1,max2,m,no_malloc =0,*tmp,new_estimate,ctr;
443:   PetscInt       *rbuf_i,kmax,rbuf_0;
444:   PetscBT        xtable;

447:   rank   = c->rank;
448:   m      = C->rmap.N;
449:   rstart = C->rmap.rstart;
450:   cstart = C->cmap.rstart;
451:   ai     = a->i;
452:   aj     = a->j;
453:   bi     = b->i;
454:   bj     = b->j;
455:   garray = c->garray;
456: 
457: 
458:   for (i=0,ct=0,total_sz=0; i<nrqr; ++i) {
459:     rbuf_i  =  rbuf[i];
460:     rbuf_0  =  rbuf_i[0];
461:     ct     += rbuf_0;
462:     for (j=1; j<=rbuf_0; j++) { total_sz += rbuf_i[2*j]; }
463:   }
464: 
465:   if (C->rmap.n) max1 = ct*(a->nz + b->nz)/C->rmap.n;
466:   else      max1 = 1;
467:   mem_estimate = 3*((total_sz > max1 ? total_sz : max1)+1);
468:   PetscMalloc(mem_estimate*sizeof(PetscInt),&xdata[0]);
469:   ++no_malloc;
470:   PetscBTCreate(m,xtable);
471:   PetscMemzero(isz1,nrqr*sizeof(PetscInt));
472: 
473:   ct3 = 0;
474:   for (i=0; i<nrqr; i++) { /* for easch mesg from proc i */
475:     rbuf_i =  rbuf[i];
476:     rbuf_0 =  rbuf_i[0];
477:     ct1    =  2*rbuf_0+1;
478:     ct2    =  ct1;
479:     ct3    += ct1;
480:     for (j=1; j<=rbuf_0; j++) { /* for each IS from proc i*/
481:       PetscBTMemzero(m,xtable);
482:       oct2 = ct2;
483:       kmax = rbuf_i[2*j];
484:       for (k=0; k<kmax; k++,ct1++) {
485:         row = rbuf_i[ct1];
486:         if (!PetscBTLookupSet(xtable,row)) {
487:           if (!(ct3 < mem_estimate)) {
488:             new_estimate = (PetscInt)(1.5*mem_estimate)+1;
489:             PetscMalloc(new_estimate*sizeof(PetscInt),&tmp);
490:             PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));
491:             PetscFree(xdata[0]);
492:             xdata[0]     = tmp;
493:             mem_estimate = new_estimate; ++no_malloc;
494:             for (ctr=1; ctr<=i; ctr++) { xdata[ctr] = xdata[ctr-1] + isz1[ctr-1];}
495:           }
496:           xdata[i][ct2++] = row;
497:           ct3++;
498:         }
499:       }
500:       for (k=oct2,max2=ct2; k<max2; k++) {
501:         row   = xdata[i][k] - rstart;
502:         start = ai[row];
503:         end   = ai[row+1];
504:         for (l=start; l<end; l++) {
505:           val = aj[l] + cstart;
506:           if (!PetscBTLookupSet(xtable,val)) {
507:             if (!(ct3 < mem_estimate)) {
508:               new_estimate = (PetscInt)(1.5*mem_estimate)+1;
509:               PetscMalloc(new_estimate*sizeof(PetscInt),&tmp);
510:               PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));
511:               PetscFree(xdata[0]);
512:               xdata[0]     = tmp;
513:               mem_estimate = new_estimate; ++no_malloc;
514:               for (ctr=1; ctr<=i; ctr++) { xdata[ctr] = xdata[ctr-1] + isz1[ctr-1];}
515:             }
516:             xdata[i][ct2++] = val;
517:             ct3++;
518:           }
519:         }
520:         start = bi[row];
521:         end   = bi[row+1];
522:         for (l=start; l<end; l++) {
523:           val = garray[bj[l]];
524:           if (!PetscBTLookupSet(xtable,val)) {
525:             if (!(ct3 < mem_estimate)) {
526:               new_estimate = (PetscInt)(1.5*mem_estimate)+1;
527:               PetscMalloc(new_estimate*sizeof(PetscInt),&tmp);
528:               PetscMemcpy(tmp,xdata[0],mem_estimate*sizeof(PetscInt));
529:               PetscFree(xdata[0]);
530:               xdata[0]     = tmp;
531:               mem_estimate = new_estimate; ++no_malloc;
532:               for (ctr =1; ctr <=i; ctr++) { xdata[ctr] = xdata[ctr-1] + isz1[ctr-1];}
533:             }
534:             xdata[i][ct2++] = val;
535:             ct3++;
536:           }
537:         }
538:       }
539:       /* Update the header*/
540:       xdata[i][2*j]   = ct2 - oct2; /* Undo the vector isz1 and use only a var*/
541:       xdata[i][2*j-1] = rbuf_i[2*j-1];
542:     }
543:     xdata[i][0] = rbuf_0;
544:     xdata[i+1]  = xdata[i] + ct2;
545:     isz1[i]     = ct2; /* size of each message */
546:   }
547:   PetscBTDestroy(xtable);
548:   PetscInfo4(C,"Allocated %D bytes, required %D bytes, no of mallocs = %D\n",rank,mem_estimate,ct3,no_malloc);
549:   return(0);
550: }
551: /* -------------------------------------------------------------------------*/
552: EXTERN PetscErrorCode MatGetSubMatrices_MPIAIJ_Local(Mat,PetscInt,const IS[],const IS[],MatReuse,Mat*);
553: EXTERN PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat,MatAssemblyType);
554: /*
555:     Every processor gets the entire matrix
556: */
559: PetscErrorCode MatGetSubMatrix_MPIAIJ_All(Mat A,MatGetSubMatrixOption flag,MatReuse scall,Mat *Bin[])
560: {
561:   Mat            B;
562:   Mat_MPIAIJ     *a = (Mat_MPIAIJ *)A->data;
563:   Mat_SeqAIJ     *b,*ad = (Mat_SeqAIJ*)a->A->data,*bd = (Mat_SeqAIJ*)a->B->data;
565:   PetscMPIInt    size,rank,*recvcounts = 0,*displs = 0;
566:   PetscInt       sendcount,i,*rstarts = A->rmap.range,n,cnt,j;
567:   PetscInt       m,*b_sendj,*garray = a->garray,*lens,*jsendbuf,*a_jsendbuf,*b_jsendbuf;
568:   PetscScalar    *sendbuf,*recvbuf,*a_sendbuf,*b_sendbuf;

571:   MPI_Comm_size(((PetscObject)A)->comm,&size);
572:   MPI_Comm_rank(((PetscObject)A)->comm,&rank);

574:   if (scall == MAT_INITIAL_MATRIX) {
575:     /* ----------------------------------------------------------------
576:          Tell every processor the number of nonzeros per row
577:     */
578:     PetscMalloc(A->rmap.N*sizeof(PetscInt),&lens);
579:     for (i=A->rmap.rstart; i<A->rmap.rend; i++) {
580:       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];
581:     }
582:     sendcount = A->rmap.rend - A->rmap.rstart;
583:     PetscMalloc(2*size*sizeof(PetscMPIInt),&recvcounts);
584:     displs     = recvcounts + size;
585:     for (i=0; i<size; i++) {
586:       recvcounts[i] = A->rmap.range[i+1] - A->rmap.range[i];
587:       displs[i]     = A->rmap.range[i];
588:     }
589:     MPI_Allgatherv(lens+A->rmap.rstart,sendcount,MPIU_INT,lens,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);

591:     /* ---------------------------------------------------------------
592:          Create the sequential matrix of the same type as the local block diagonal
593:     */
594:     MatCreate(PETSC_COMM_SELF,&B);
595:     MatSetSizes(B,A->rmap.N,A->cmap.N,PETSC_DETERMINE,PETSC_DETERMINE);
596:     MatSetType(B,((PetscObject)a->A)->type_name);
597:     MatSeqAIJSetPreallocation(B,0,lens);
598:     PetscMalloc(sizeof(Mat),Bin);
599:     **Bin = B;
600:     b = (Mat_SeqAIJ *)B->data;

602:     /*--------------------------------------------------------------------
603:        Copy my part of matrix column indices over
604:     */
605:     sendcount  = ad->nz + bd->nz;
606:     jsendbuf   = b->j + b->i[rstarts[rank]];
607:     a_jsendbuf = ad->j;
608:     b_jsendbuf = bd->j;
609:     n          = A->rmap.rend - A->rmap.rstart;
610:     cnt        = 0;
611:     for (i=0; i<n; i++) {

613:       /* put in lower diagonal portion */
614:       m = bd->i[i+1] - bd->i[i];
615:       while (m > 0) {
616:         /* is it above diagonal (in bd (compressed) numbering) */
617:         if (garray[*b_jsendbuf] > A->rmap.rstart + i) break;
618:         jsendbuf[cnt++] = garray[*b_jsendbuf++];
619:         m--;
620:       }

622:       /* put in diagonal portion */
623:       for (j=ad->i[i]; j<ad->i[i+1]; j++) {
624:         jsendbuf[cnt++] = A->rmap.rstart + *a_jsendbuf++;
625:       }

627:       /* put in upper diagonal portion */
628:       while (m-- > 0) {
629:         jsendbuf[cnt++] = garray[*b_jsendbuf++];
630:       }
631:     }
632:     if (cnt != sendcount) SETERRQ2(PETSC_ERR_PLIB,"Corrupted PETSc matrix: nz given %D actual nz %D",sendcount,cnt);

634:     /*--------------------------------------------------------------------
635:        Gather all column indices to all processors
636:     */
637:     for (i=0; i<size; i++) {
638:       recvcounts[i] = 0;
639:       for (j=A->rmap.range[i]; j<A->rmap.range[i+1]; j++) {
640:         recvcounts[i] += lens[j];
641:       }
642:     }
643:     displs[0]  = 0;
644:     for (i=1; i<size; i++) {
645:       displs[i] = displs[i-1] + recvcounts[i-1];
646:     }
647:     MPI_Allgatherv(jsendbuf,sendcount,MPIU_INT,b->j,recvcounts,displs,MPIU_INT,((PetscObject)A)->comm);

649:     /*--------------------------------------------------------------------
650:         Assemble the matrix into useable form (note numerical values not yet set)
651:     */
652:     /* set the b->ilen (length of each row) values */
653:     PetscMemcpy(b->ilen,lens,A->rmap.N*sizeof(PetscInt));
654:     /* set the b->i indices */
655:     b->i[0] = 0;
656:     for (i=1; i<=A->rmap.N; i++) {
657:       b->i[i] = b->i[i-1] + lens[i-1];
658:     }
659:     PetscFree(lens);
660:     MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
661:     MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);

663:   } else {
664:     B  = **Bin;
665:     b = (Mat_SeqAIJ *)B->data;
666:   }

668:   /*--------------------------------------------------------------------
669:        Copy my part of matrix numerical values into the values location 
670:   */
671:   if (flag == MAT_GET_VALUES){
672:     sendcount = ad->nz + bd->nz;
673:     sendbuf   = b->a + b->i[rstarts[rank]];
674:     a_sendbuf = ad->a;
675:     b_sendbuf = bd->a;
676:     b_sendj   = bd->j;
677:     n         = A->rmap.rend - A->rmap.rstart;
678:     cnt       = 0;
679:     for (i=0; i<n; i++) {

681:       /* put in lower diagonal portion */
682:       m = bd->i[i+1] - bd->i[i];
683:       while (m > 0) {
684:         /* is it above diagonal (in bd (compressed) numbering) */
685:         if (garray[*b_sendj] > A->rmap.rstart + i) break;
686:         sendbuf[cnt++] = *b_sendbuf++;
687:         m--;
688:         b_sendj++;
689:       }

691:       /* put in diagonal portion */
692:       for (j=ad->i[i]; j<ad->i[i+1]; j++) {
693:         sendbuf[cnt++] = *a_sendbuf++;
694:       }

696:       /* put in upper diagonal portion */
697:       while (m-- > 0) {
698:         sendbuf[cnt++] = *b_sendbuf++;
699:         b_sendj++;
700:       }
701:     }
702:     if (cnt != sendcount) SETERRQ2(PETSC_ERR_PLIB,"Corrupted PETSc matrix: nz given %D actual nz %D",sendcount,cnt);
703: 
704:     /* ----------------------------------------------------------------- 
705:        Gather all numerical values to all processors 
706:     */
707:     if (!recvcounts) {
708:       PetscMalloc(2*size*sizeof(PetscInt),&recvcounts);
709:       displs = recvcounts + size;
710:     }
711:     for (i=0; i<size; i++) {
712:       recvcounts[i] = b->i[rstarts[i+1]] - b->i[rstarts[i]];
713:     }
714:     displs[0]  = 0;
715:     for (i=1; i<size; i++) {
716:       displs[i] = displs[i-1] + recvcounts[i-1];
717:     }
718:     recvbuf   = b->a;
719:     MPI_Allgatherv(sendbuf,sendcount,MPIU_SCALAR,recvbuf,recvcounts,displs,MPIU_SCALAR,((PetscObject)A)->comm);
720:   }  /* endof (flag == MAT_GET_VALUES) */
721:   PetscFree(recvcounts);

723:   if (A->symmetric){
724:     MatSetOption(B,MAT_SYMMETRIC,PETSC_TRUE);
725:   } else if (A->hermitian) {
726:     MatSetOption(B,MAT_HERMITIAN,PETSC_TRUE);
727:   } else if (A->structurally_symmetric) {
728:     MatSetOption(B,MAT_STRUCTURALLY_SYMMETRIC,PETSC_TRUE);
729:   }
730:   return(0);
731: }

735: PetscErrorCode MatGetSubMatrices_MPIAIJ(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submat[])
736: {
738:   PetscInt       nmax,nstages_local,nstages,i,pos,max_no,nrow,ncol;
739:   PetscTruth     rowflag,colflag,wantallmatrix = PETSC_FALSE,twantallmatrix;

742:   /*
743:        Check for special case each processor gets entire matrix
744:   */
745:   if (ismax == 1 && C->rmap.N == C->cmap.N) {
746:     ISIdentity(*isrow,&rowflag);
747:     ISIdentity(*iscol,&colflag);
748:     ISGetLocalSize(*isrow,&nrow);
749:     ISGetLocalSize(*iscol,&ncol);
750:     if (rowflag && colflag && nrow == C->rmap.N && ncol == C->cmap.N) {
751:       wantallmatrix = PETSC_TRUE;
752:       PetscOptionsGetTruth(((PetscObject)C)->prefix,"-use_fast_submatrix",&wantallmatrix,PETSC_NULL);
753:     }
754:   }
755:   MPI_Allreduce(&wantallmatrix,&twantallmatrix,1,MPI_INT,MPI_MIN,((PetscObject)C)->comm);
756:   if (twantallmatrix) {
757:     MatGetSubMatrix_MPIAIJ_All(C,MAT_GET_VALUES,scall,submat);
758:     return(0);
759:   }

761:   /* Allocate memory to hold all the submatrices */
762:   if (scall != MAT_REUSE_MATRIX) {
763:     PetscMalloc((ismax+1)*sizeof(Mat),submat);
764:   }
765:   /* Determine the number of stages through which submatrices are done */
766:   nmax          = 20*1000000 / (C->cmap.N * sizeof(PetscInt));
767:   if (!nmax) nmax = 1;
768:   nstages_local = ismax/nmax + ((ismax % nmax)?1:0);

770:   /* Make sure every processor loops through the nstages */
771:   MPI_Allreduce(&nstages_local,&nstages,1,MPIU_INT,MPI_MAX,((PetscObject)C)->comm);

773:   for (i=0,pos=0; i<nstages; i++) {
774:     if (pos+nmax <= ismax) max_no = nmax;
775:     else if (pos == ismax) max_no = 0;
776:     else                   max_no = ismax-pos;
777:     MatGetSubMatrices_MPIAIJ_Local(C,max_no,isrow+pos,iscol+pos,scall,*submat+pos);
778:     pos += max_no;
779:   }
780:   return(0);
781: }
782: /* -------------------------------------------------------------------------*/
785: PetscErrorCode MatGetSubMatrices_MPIAIJ_Local(Mat C,PetscInt ismax,const IS isrow[],const IS iscol[],MatReuse scall,Mat *submats)
786: {
787:   Mat_MPIAIJ     *c = (Mat_MPIAIJ*)C->data;
788:   Mat            A = c->A;
789:   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data,*b = (Mat_SeqAIJ*)c->B->data,*mat;
790:   PetscInt       **irow,**icol,*nrow,*ncol,start;
792:   PetscMPIInt    rank,size,tag0,tag1,tag2,tag3,*w1,*w2,*w3,*w4,nrqr;
793:   PetscInt       **sbuf1,**sbuf2,i,j,k,l,ct1,ct2,**rbuf1,row,proc;
794:   PetscInt       nrqs,msz,**ptr,*req_size,*ctr,*pa,*tmp,tcol;
795:   PetscInt       **rbuf3,*req_source,**sbuf_aj,**rbuf2,max1,max2,**rmap;
796:   PetscInt       **cmap,**lens,is_no,ncols,*cols,mat_i,*mat_j,tmp2,jmax,*irow_i;
797:   PetscInt       len,ctr_j,*sbuf1_j,*sbuf_aj_i,*rbuf1_i,kmax,*cmap_i,*lens_i;
798:   PetscInt       *rmap_i;
799:   MPI_Request    *s_waits1,*r_waits1,*s_waits2,*r_waits2,*r_waits3;
800:   MPI_Request    *r_waits4,*s_waits3,*s_waits4;
801:   MPI_Status     *r_status1,*r_status2,*s_status1,*s_status3,*s_status2;
802:   MPI_Status     *r_status3,*r_status4,*s_status4;
803:   MPI_Comm       comm;
804:   PetscScalar    **rbuf4,**sbuf_aa,*vals,*mat_a,*sbuf_aa_i;
805:   PetscTruth     sorted;
806:   PetscMPIInt    *onodes1,*olengths1;
807:   PetscMPIInt    idex,idex2,end;

810:   comm   = ((PetscObject)C)->comm;
811:   tag0   = ((PetscObject)C)->tag;
812:   size   = c->size;
813:   rank   = c->rank;
814: 
815:   /* Get some new tags to keep the communication clean */
816:   PetscObjectGetNewTag((PetscObject)C,&tag1);
817:   PetscObjectGetNewTag((PetscObject)C,&tag2);
818:   PetscObjectGetNewTag((PetscObject)C,&tag3);

820:     /* Check if the col indices are sorted */
821:   for (i=0; i<ismax; i++) {
822:     ISSorted(isrow[i],&sorted);
823:     /*if (!sorted) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"isrow is not sorted");*/
824:     ISSorted(iscol[i],&sorted);
825:     /*    if (!sorted) SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"iscol is not sorted"); */
826:   }

828:   len    = (2*ismax+1)*(sizeof(PetscInt*)+ sizeof(PetscInt));
829:   PetscMalloc(len,&irow);
830:   icol   = irow + ismax;
831:   nrow   = (PetscInt*)(icol + ismax);
832:   ncol   = nrow + ismax;

834:   for (i=0; i<ismax; i++) {
835:     ISGetIndices(isrow[i],&irow[i]);
836:     ISGetIndices(iscol[i],&icol[i]);
837:     ISGetLocalSize(isrow[i],&nrow[i]);
838:     ISGetLocalSize(iscol[i],&ncol[i]);
839:   }

841:   /* evaluate communication - mesg to who, length of mesg, and buffer space
842:      required. Based on this, buffers are allocated, and data copied into them*/
843:   PetscMalloc(size*4*sizeof(PetscMPIInt),&w1); /* mesg size */
844:   w2     = w1 + size;      /* if w2[i] marked, then a message to proc i*/
845:   w3     = w2 + size;      /* no of IS that needs to be sent to proc i */
846:   w4     = w3 + size;      /* temp work space used in determining w1, w2, w3 */
847:   PetscMemzero(w1,size*3*sizeof(PetscMPIInt)); /* initialize work vector*/
848:   for (i=0; i<ismax; i++) {
849:     PetscMemzero(w4,size*sizeof(PetscMPIInt)); /* initialize work vector*/
850:     jmax   = nrow[i];
851:     irow_i = irow[i];
852:     for (j=0; j<jmax; j++) {
853:       l = 0;
854:       row  = irow_i[j];
855:       while (row >= C->rmap.range[l+1]) l++;
856:       proc = l;
857:       w4[proc]++;
858:     }
859:     for (j=0; j<size; j++) {
860:       if (w4[j]) { w1[j] += w4[j];  w3[j]++;}
861:     }
862:   }
863: 
864:   nrqs     = 0;              /* no of outgoing messages */
865:   msz      = 0;              /* total mesg length (for all procs) */
866:   w1[rank] = 0;              /* no mesg sent to self */
867:   w3[rank] = 0;
868:   for (i=0; i<size; i++) {
869:     if (w1[i])  { w2[i] = 1; nrqs++;} /* there exists a message to proc i */
870:   }
871:   PetscMalloc((nrqs+1)*sizeof(PetscInt),&pa); /*(proc -array)*/
872:   for (i=0,j=0; i<size; i++) {
873:     if (w1[i]) { pa[j] = i; j++; }
874:   }

876:   /* Each message would have a header = 1 + 2*(no of IS) + data */
877:   for (i=0; i<nrqs; i++) {
878:     j     = pa[i];
879:     w1[j] += w2[j] + 2* w3[j];
880:     msz   += w1[j];
881:   }

883:   /* Determine the number of messages to expect, their lengths, from from-ids */
884:   PetscGatherNumberOfMessages(comm,w2,w1,&nrqr);
885:   PetscGatherMessageLengths(comm,nrqs,nrqr,w1,&onodes1,&olengths1);

887:   /* Now post the Irecvs corresponding to these messages */
888:   PetscPostIrecvInt(comm,tag0,nrqr,onodes1,olengths1,&rbuf1,&r_waits1);
889: 
890:   PetscFree(onodes1);
891:   PetscFree(olengths1);
892: 
893:   /* Allocate Memory for outgoing messages */
894:   len      = 2*size*sizeof(PetscInt*) + 2*msz*sizeof(PetscInt) + size*sizeof(PetscInt);
895:   PetscMalloc(len,&sbuf1);
896:   ptr      = sbuf1 + size;   /* Pointers to the data in outgoing buffers */
897:   PetscMemzero(sbuf1,2*size*sizeof(PetscInt*));
898:   /* allocate memory for outgoing data + buf to receive the first reply */
899:   tmp      = (PetscInt*)(ptr + size);
900:   ctr      = tmp + 2*msz;

902:   {
903:     PetscInt *iptr = tmp,ict = 0;
904:     for (i=0; i<nrqs; i++) {
905:       j         = pa[i];
906:       iptr     += ict;
907:       sbuf1[j]  = iptr;
908:       ict       = w1[j];
909:     }
910:   }

912:   /* Form the outgoing messages */
913:   /* Initialize the header space */
914:   for (i=0; i<nrqs; i++) {
915:     j           = pa[i];
916:     sbuf1[j][0] = 0;
917:     PetscMemzero(sbuf1[j]+1,2*w3[j]*sizeof(PetscInt));
918:     ptr[j]      = sbuf1[j] + 2*w3[j] + 1;
919:   }
920: 
921:   /* Parse the isrow and copy data into outbuf */
922:   for (i=0; i<ismax; i++) {
923:     PetscMemzero(ctr,size*sizeof(PetscInt));
924:     irow_i = irow[i];
925:     jmax   = nrow[i];
926:     for (j=0; j<jmax; j++) {  /* parse the indices of each IS */
927:       l = 0;
928:       row  = irow_i[j];
929:       while (row >= C->rmap.range[l+1]) l++;
930:       proc = l;
931:       if (proc != rank) { /* copy to the outgoing buf*/
932:         ctr[proc]++;
933:         *ptr[proc] = row;
934:         ptr[proc]++;
935:       }
936:     }
937:     /* Update the headers for the current IS */
938:     for (j=0; j<size; j++) { /* Can Optimise this loop too */
939:       if ((ctr_j = ctr[j])) {
940:         sbuf1_j        = sbuf1[j];
941:         k              = ++sbuf1_j[0];
942:         sbuf1_j[2*k]   = ctr_j;
943:         sbuf1_j[2*k-1] = i;
944:       }
945:     }
946:   }

948:   /*  Now  post the sends */
949:   PetscMalloc((nrqs+1)*sizeof(MPI_Request),&s_waits1);
950:   for (i=0; i<nrqs; ++i) {
951:     j    = pa[i];
952:     MPI_Isend(sbuf1[j],w1[j],MPIU_INT,j,tag0,comm,s_waits1+i);
953:   }

955:   /* Post Receives to capture the buffer size */
956:   PetscMalloc((nrqs+1)*sizeof(MPI_Request),&r_waits2);
957:   PetscMalloc((nrqs+1)*sizeof(PetscInt*),&rbuf2);
958:   rbuf2[0] = tmp + msz;
959:   for (i=1; i<nrqs; ++i) {
960:     rbuf2[i] = rbuf2[i-1]+w1[pa[i-1]];
961:   }
962:   for (i=0; i<nrqs; ++i) {
963:     j    = pa[i];
964:     MPI_Irecv(rbuf2[i],w1[j],MPIU_INT,j,tag1,comm,r_waits2+i);
965:   }

967:   /* Send to other procs the buf size they should allocate */
968: 

970:   /* Receive messages*/
971:   PetscMalloc((nrqr+1)*sizeof(MPI_Request),&s_waits2);
972:   PetscMalloc((nrqr+1)*sizeof(MPI_Status),&r_status1);
973:   len         = 2*nrqr*sizeof(PetscInt) + (nrqr+1)*sizeof(PetscInt*);
974:   PetscMalloc(len,&sbuf2);
975:   req_size    = (PetscInt*)(sbuf2 + nrqr);
976:   req_source  = req_size + nrqr;
977: 
978:   {
979:     Mat_SeqAIJ  *sA = (Mat_SeqAIJ*)c->A->data,*sB = (Mat_SeqAIJ*)c->B->data;
980:     PetscInt    *sAi = sA->i,*sBi = sB->i,id,rstart = C->rmap.rstart;
981:     PetscInt    *sbuf2_i;

983:     for (i=0; i<nrqr; ++i) {
984:       MPI_Waitany(nrqr,r_waits1,&idex,r_status1+i);
985:       req_size[idex] = 0;
986:       rbuf1_i         = rbuf1[idex];
987:       start           = 2*rbuf1_i[0] + 1;
988:       MPI_Get_count(r_status1+i,MPIU_INT,&end);
989:       PetscMalloc((end+1)*sizeof(PetscInt),&sbuf2[idex]);
990:       sbuf2_i         = sbuf2[idex];
991:       for (j=start; j<end; j++) {
992:         id               = rbuf1_i[j] - rstart;
993:         ncols            = sAi[id+1] - sAi[id] + sBi[id+1] - sBi[id];
994:         sbuf2_i[j]       = ncols;
995:         req_size[idex] += ncols;
996:       }
997:       req_source[idex] = r_status1[i].MPI_SOURCE;
998:       /* form the header */
999:       sbuf2_i[0]   = req_size[idex];
1000:       for (j=1; j<start; j++) { sbuf2_i[j] = rbuf1_i[j]; }
1001:       MPI_Isend(sbuf2_i,end,MPIU_INT,req_source[idex],tag1,comm,s_waits2+i);
1002:     }
1003:   }
1004:   PetscFree(r_status1);
1005:   PetscFree(r_waits1);

1007:   /*  recv buffer sizes */
1008:   /* Receive messages*/
1009: 
1010:   PetscMalloc((nrqs+1)*sizeof(PetscInt*),&rbuf3);
1011:   PetscMalloc((nrqs+1)*sizeof(PetscScalar*),&rbuf4);
1012:   PetscMalloc((nrqs+1)*sizeof(MPI_Request),&r_waits3);
1013:   PetscMalloc((nrqs+1)*sizeof(MPI_Request),&r_waits4);
1014:   PetscMalloc((nrqs+1)*sizeof(MPI_Status),&r_status2);

1016:   for (i=0; i<nrqs; ++i) {
1017:     MPI_Waitany(nrqs,r_waits2,&idex,r_status2+i);
1018:     PetscMalloc((rbuf2[idex][0]+1)*sizeof(PetscInt),&rbuf3[idex]);
1019:     PetscMalloc((rbuf2[idex][0]+1)*sizeof(PetscScalar),&rbuf4[idex]);
1020:     MPI_Irecv(rbuf3[idex],rbuf2[idex][0],MPIU_INT,r_status2[i].MPI_SOURCE,tag2,comm,r_waits3+idex);
1021:     MPI_Irecv(rbuf4[idex],rbuf2[idex][0],MPIU_SCALAR,r_status2[i].MPI_SOURCE,tag3,comm,r_waits4+idex);
1022:   }
1023:   PetscFree(r_status2);
1024:   PetscFree(r_waits2);
1025: 
1026:   /* Wait on sends1 and sends2 */
1027:   PetscMalloc((nrqs+1)*sizeof(MPI_Status),&s_status1);
1028:   PetscMalloc((nrqr+1)*sizeof(MPI_Status),&s_status2);

1030:   if (nrqs) {MPI_Waitall(nrqs,s_waits1,s_status1);}
1031:   if (nrqr) {MPI_Waitall(nrqr,s_waits2,s_status2);}
1032:   PetscFree(s_status1);
1033:   PetscFree(s_status2);
1034:   PetscFree(s_waits1);
1035:   PetscFree(s_waits2);

1037:   /* Now allocate buffers for a->j, and send them off */
1038:   PetscMalloc((nrqr+1)*sizeof(PetscInt*),&sbuf_aj);
1039:   for (i=0,j=0; i<nrqr; i++) j += req_size[i];
1040:   PetscMalloc((j+1)*sizeof(PetscInt),&sbuf_aj[0]);
1041:   for (i=1; i<nrqr; i++)  sbuf_aj[i] = sbuf_aj[i-1] + req_size[i-1];
1042: 
1043:   PetscMalloc((nrqr+1)*sizeof(MPI_Request),&s_waits3);
1044:   {
1045:     PetscInt nzA,nzB,*a_i = a->i,*b_i = b->i,lwrite;
1046:     PetscInt *cworkA,*cworkB,cstart = C->cmap.rstart,rstart = C->rmap.rstart,*bmap = c->garray;
1047:     PetscInt cend = C->cmap.rend;
1048:     PetscInt *a_j = a->j,*b_j = b->j,ctmp;

1050:     for (i=0; i<nrqr; i++) {
1051:       rbuf1_i   = rbuf1[i];
1052:       sbuf_aj_i = sbuf_aj[i];
1053:       ct1       = 2*rbuf1_i[0] + 1;
1054:       ct2       = 0;
1055:       for (j=1,max1=rbuf1_i[0]; j<=max1; j++) {
1056:         kmax = rbuf1[i][2*j];
1057:         for (k=0; k<kmax; k++,ct1++) {
1058:           row    = rbuf1_i[ct1] - rstart;
1059:           nzA    = a_i[row+1] - a_i[row];     nzB = b_i[row+1] - b_i[row];
1060:           ncols  = nzA + nzB;
1061:           cworkA = a_j + a_i[row]; cworkB = b_j + b_i[row];

1063:           /* load the column indices for this row into cols*/
1064:           cols  = sbuf_aj_i + ct2;
1065: 
1066:           lwrite = 0;
1067:           for (l=0; l<nzB; l++) {
1068:             if ((ctmp = bmap[cworkB[l]]) < cstart)  cols[lwrite++] = ctmp;
1069:           }
1070:           for (l=0; l<nzA; l++)   cols[lwrite++] = cstart + cworkA[l];
1071:           for (l=0; l<nzB; l++) {
1072:             if ((ctmp = bmap[cworkB[l]]) >= cend)  cols[lwrite++] = ctmp;
1073:           }

1075:           ct2 += ncols;
1076:         }
1077:       }
1078:       MPI_Isend(sbuf_aj_i,req_size[i],MPIU_INT,req_source[i],tag2,comm,s_waits3+i);
1079:     }
1080:   }
1081:   PetscMalloc((nrqs+1)*sizeof(MPI_Status),&r_status3);
1082:   PetscMalloc((nrqr+1)*sizeof(MPI_Status),&s_status3);

1084:   /* Allocate buffers for a->a, and send them off */
1085:   PetscMalloc((nrqr+1)*sizeof(PetscScalar*),&sbuf_aa);
1086:   for (i=0,j=0; i<nrqr; i++) j += req_size[i];
1087:   PetscMalloc((j+1)*sizeof(PetscScalar),&sbuf_aa[0]);
1088:   for (i=1; i<nrqr; i++)  sbuf_aa[i] = sbuf_aa[i-1] + req_size[i-1];
1089: 
1090:   PetscMalloc((nrqr+1)*sizeof(MPI_Request),&s_waits4);
1091:   {
1092:     PetscInt    nzA,nzB,*a_i = a->i,*b_i = b->i, *cworkB,lwrite;
1093:     PetscInt    cstart = C->cmap.rstart,rstart = C->rmap.rstart,*bmap = c->garray;
1094:     PetscInt    cend = C->cmap.rend;
1095:     PetscInt    *b_j = b->j;
1096:     PetscScalar *vworkA,*vworkB,*a_a = a->a,*b_a = b->a;
1097: 
1098:     for (i=0; i<nrqr; i++) {
1099:       rbuf1_i   = rbuf1[i];
1100:       sbuf_aa_i = sbuf_aa[i];
1101:       ct1       = 2*rbuf1_i[0]+1;
1102:       ct2       = 0;
1103:       for (j=1,max1=rbuf1_i[0]; j<=max1; j++) {
1104:         kmax = rbuf1_i[2*j];
1105:         for (k=0; k<kmax; k++,ct1++) {
1106:           row    = rbuf1_i[ct1] - rstart;
1107:           nzA    = a_i[row+1] - a_i[row];     nzB = b_i[row+1] - b_i[row];
1108:           ncols  = nzA + nzB;
1109:           cworkB = b_j + b_i[row];
1110:           vworkA = a_a + a_i[row];
1111:           vworkB = b_a + b_i[row];

1113:           /* load the column values for this row into vals*/
1114:           vals  = sbuf_aa_i+ct2;
1115: 
1116:           lwrite = 0;
1117:           for (l=0; l<nzB; l++) {
1118:             if ((bmap[cworkB[l]]) < cstart)  vals[lwrite++] = vworkB[l];
1119:           }
1120:           for (l=0; l<nzA; l++)   vals[lwrite++] = vworkA[l];
1121:           for (l=0; l<nzB; l++) {
1122:             if ((bmap[cworkB[l]]) >= cend)  vals[lwrite++] = vworkB[l];
1123:           }
1124: 
1125:           ct2 += ncols;
1126:         }
1127:       }
1128:       MPI_Isend(sbuf_aa_i,req_size[i],MPIU_SCALAR,req_source[i],tag3,comm,s_waits4+i);
1129:     }
1130:   }
1131:   PetscMalloc((nrqs+1)*sizeof(MPI_Status),&r_status4);
1132:   PetscMalloc((nrqr+1)*sizeof(MPI_Status),&s_status4);
1133:   PetscFree(rbuf1);

1135:   /* Form the matrix */
1136:   /* create col map */
1137:   {
1138:     PetscInt *icol_i;
1139: 
1140:     len     = (1+ismax)*sizeof(PetscInt*)+ (1+ismax*C->cmap.N)*sizeof(PetscInt);
1141:     PetscMalloc(len,&cmap);
1142:     cmap[0] = (PetscInt*)(cmap + ismax);
1143:     PetscMemzero(cmap[0],(1+ismax*C->cmap.N)*sizeof(PetscInt));
1144:     for (i=1; i<ismax; i++) { cmap[i] = cmap[i-1] + C->cmap.N; }
1145:     for (i=0; i<ismax; i++) {
1146:       jmax   = ncol[i];
1147:       icol_i = icol[i];
1148:       cmap_i = cmap[i];
1149:       for (j=0; j<jmax; j++) {
1150:         cmap_i[icol_i[j]] = j+1;
1151:       }
1152:     }
1153:   }

1155:   /* Create lens which is required for MatCreate... */
1156:   for (i=0,j=0; i<ismax; i++) { j += nrow[i]; }
1157:   len     = (1+ismax)*sizeof(PetscInt*)+ j*sizeof(PetscInt);
1158:   PetscMalloc(len,&lens);
1159:   lens[0] = (PetscInt*)(lens + ismax);
1160:   PetscMemzero(lens[0],j*sizeof(PetscInt));
1161:   for (i=1; i<ismax; i++) { lens[i] = lens[i-1] + nrow[i-1]; }
1162: 
1163:   /* Update lens from local data */
1164:   for (i=0; i<ismax; i++) {
1165:     jmax   = nrow[i];
1166:     cmap_i = cmap[i];
1167:     irow_i = irow[i];
1168:     lens_i = lens[i];
1169:     for (j=0; j<jmax; j++) {
1170:       l = 0;
1171:       row  = irow_i[j];
1172:       while (row >= C->rmap.range[l+1]) l++;
1173:       proc = l;
1174:       if (proc == rank) {
1175:         MatGetRow_MPIAIJ(C,row,&ncols,&cols,0);
1176:         for (k=0; k<ncols; k++) {
1177:           if (cmap_i[cols[k]]) { lens_i[j]++;}
1178:         }
1179:         MatRestoreRow_MPIAIJ(C,row,&ncols,&cols,0);
1180:       }
1181:     }
1182:   }
1183: 
1184:   /* Create row map*/
1185:   len     = (1+ismax)*sizeof(PetscInt*)+ ismax*C->rmap.N*sizeof(PetscInt);
1186:   PetscMalloc(len,&rmap);
1187:   rmap[0] = (PetscInt*)(rmap + ismax);
1188:   PetscMemzero(rmap[0],ismax*C->rmap.N*sizeof(PetscInt));
1189:   for (i=1; i<ismax; i++) { rmap[i] = rmap[i-1] + C->rmap.N;}
1190:   for (i=0; i<ismax; i++) {
1191:     rmap_i = rmap[i];
1192:     irow_i = irow[i];
1193:     jmax   = nrow[i];
1194:     for (j=0; j<jmax; j++) {
1195:       rmap_i[irow_i[j]] = j;
1196:     }
1197:   }
1198: 
1199:   /* Update lens from offproc data */
1200:   {
1201:     PetscInt *rbuf2_i,*rbuf3_i,*sbuf1_i;

1203:     for (tmp2=0; tmp2<nrqs; tmp2++) {
1204:       MPI_Waitany(nrqs,r_waits3,&idex2,r_status3+tmp2);
1205:       idex   = pa[idex2];
1206:       sbuf1_i = sbuf1[idex];
1207:       jmax    = sbuf1_i[0];
1208:       ct1     = 2*jmax+1;
1209:       ct2     = 0;
1210:       rbuf2_i = rbuf2[idex2];
1211:       rbuf3_i = rbuf3[idex2];
1212:       for (j=1; j<=jmax; j++) {
1213:         is_no   = sbuf1_i[2*j-1];
1214:         max1    = sbuf1_i[2*j];
1215:         lens_i  = lens[is_no];
1216:         cmap_i  = cmap[is_no];
1217:         rmap_i  = rmap[is_no];
1218:         for (k=0; k<max1; k++,ct1++) {
1219:           row  = rmap_i[sbuf1_i[ct1]]; /* the val in the new matrix to be */
1220:           max2 = rbuf2_i[ct1];
1221:           for (l=0; l<max2; l++,ct2++) {
1222:             if (cmap_i[rbuf3_i[ct2]]) {
1223:               lens_i[row]++;
1224:             }
1225:           }
1226:         }
1227:       }
1228:     }
1229:   }
1230:   PetscFree(r_status3);
1231:   PetscFree(r_waits3);
1232:   if (nrqr) {MPI_Waitall(nrqr,s_waits3,s_status3);}
1233:   PetscFree(s_status3);
1234:   PetscFree(s_waits3);

1236:   /* Create the submatrices */
1237:   if (scall == MAT_REUSE_MATRIX) {
1238:     PetscTruth flag;

1240:     /*
1241:         Assumes new rows are same length as the old rows,hence bug!
1242:     */
1243:     for (i=0; i<ismax; i++) {
1244:       mat = (Mat_SeqAIJ *)(submats[i]->data);
1245:       if ((submats[i]->rmap.n != nrow[i]) || (submats[i]->cmap.n != ncol[i])) {
1246:         SETERRQ(PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong size");
1247:       }
1248:       PetscMemcmp(mat->ilen,lens[i],submats[i]->rmap.n*sizeof(PetscInt),&flag);
1249:       if (!flag) {
1250:         SETERRQ(PETSC_ERR_ARG_SIZ,"Cannot reuse matrix. wrong no of nonzeros");
1251:       }
1252:       /* Initial matrix as if empty */
1253:       PetscMemzero(mat->ilen,submats[i]->rmap.n*sizeof(PetscInt));
1254:       submats[i]->factor = C->factor;
1255:     }
1256:   } else {
1257:     for (i=0; i<ismax; i++) {
1258:       MatCreate(PETSC_COMM_SELF,submats+i);
1259:       MatSetSizes(submats[i],nrow[i],ncol[i],PETSC_DETERMINE,PETSC_DETERMINE);
1260:       MatSetType(submats[i],((PetscObject)A)->type_name);
1261:       MatSeqAIJSetPreallocation(submats[i],0,lens[i]);
1262:     }
1263:   }

1265:   /* Assemble the matrices */
1266:   /* First assemble the local rows */
1267:   {
1268:     PetscInt    ilen_row,*imat_ilen,*imat_j,*imat_i,old_row;
1269:     PetscScalar *imat_a;
1270: 
1271:     for (i=0; i<ismax; i++) {
1272:       mat       = (Mat_SeqAIJ*)submats[i]->data;
1273:       imat_ilen = mat->ilen;
1274:       imat_j    = mat->j;
1275:       imat_i    = mat->i;
1276:       imat_a    = mat->a;
1277:       cmap_i    = cmap[i];
1278:       rmap_i    = rmap[i];
1279:       irow_i    = irow[i];
1280:       jmax      = nrow[i];
1281:       for (j=0; j<jmax; j++) {
1282:         l = 0;
1283:         row      = irow_i[j];
1284:         while (row >= C->rmap.range[l+1]) l++;
1285:         proc = l;
1286:         if (proc == rank) {
1287:           old_row  = row;
1288:           row      = rmap_i[row];
1289:           ilen_row = imat_ilen[row];
1290:           MatGetRow_MPIAIJ(C,old_row,&ncols,&cols,&vals);
1291:           mat_i    = imat_i[row] ;
1292:           mat_a    = imat_a + mat_i;
1293:           mat_j    = imat_j + mat_i;
1294:           for (k=0; k<ncols; k++) {
1295:             if ((tcol = cmap_i[cols[k]])) {
1296:               *mat_j++ = tcol - 1;
1297:               *mat_a++ = vals[k];
1298:               ilen_row++;
1299:             }
1300:           }
1301:           MatRestoreRow_MPIAIJ(C,old_row,&ncols,&cols,&vals);
1302:           imat_ilen[row] = ilen_row;
1303:         }
1304:       }
1305:     }
1306:   }

1308:   /*   Now assemble the off proc rows*/
1309:   {
1310:     PetscInt    *sbuf1_i,*rbuf2_i,*rbuf3_i,*imat_ilen,ilen;
1311:     PetscInt    *imat_j,*imat_i;
1312:     PetscScalar *imat_a,*rbuf4_i;

1314:     for (tmp2=0; tmp2<nrqs; tmp2++) {
1315:       MPI_Waitany(nrqs,r_waits4,&idex2,r_status4+tmp2);
1316:       idex   = pa[idex2];
1317:       sbuf1_i = sbuf1[idex];
1318:       jmax    = sbuf1_i[0];
1319:       ct1     = 2*jmax + 1;
1320:       ct2     = 0;
1321:       rbuf2_i = rbuf2[idex2];
1322:       rbuf3_i = rbuf3[idex2];
1323:       rbuf4_i = rbuf4[idex2];
1324:       for (j=1; j<=jmax; j++) {
1325:         is_no     = sbuf1_i[2*j-1];
1326:         rmap_i    = rmap[is_no];
1327:         cmap_i    = cmap[is_no];
1328:         mat       = (Mat_SeqAIJ*)submats[is_no]->data;
1329:         imat_ilen = mat->ilen;
1330:         imat_j    = mat->j;
1331:         imat_i    = mat->i;
1332:         imat_a    = mat->a;
1333:         max1      = sbuf1_i[2*j];
1334:         for (k=0; k<max1; k++,ct1++) {
1335:           row   = sbuf1_i[ct1];
1336:           row   = rmap_i[row];
1337:           ilen  = imat_ilen[row];
1338:           mat_i = imat_i[row] ;
1339:           mat_a = imat_a + mat_i;
1340:           mat_j = imat_j + mat_i;
1341:           max2 = rbuf2_i[ct1];
1342:           for (l=0; l<max2; l++,ct2++) {
1343:             if ((tcol = cmap_i[rbuf3_i[ct2]])) {
1344:               *mat_j++ = tcol - 1;
1345:               *mat_a++ = rbuf4_i[ct2];
1346:               ilen++;
1347:             }
1348:           }
1349:           imat_ilen[row] = ilen;
1350:         }
1351:       }
1352:     }
1353:   }
1354:   PetscFree(r_status4);
1355:   PetscFree(r_waits4);
1356:   if (nrqr) {MPI_Waitall(nrqr,s_waits4,s_status4);}
1357:   PetscFree(s_waits4);
1358:   PetscFree(s_status4);

1360:   /* Restore the indices */
1361:   for (i=0; i<ismax; i++) {
1362:     ISRestoreIndices(isrow[i],irow+i);
1363:     ISRestoreIndices(iscol[i],icol+i);
1364:   }

1366:   /* Destroy allocated memory */
1367:   PetscFree(irow);
1368:   PetscFree(w1);
1369:   PetscFree(pa);

1371:   PetscFree(sbuf1);
1372:   PetscFree(rbuf2);
1373:   for (i=0; i<nrqr; ++i) {
1374:     PetscFree(sbuf2[i]);
1375:   }
1376:   for (i=0; i<nrqs; ++i) {
1377:     PetscFree(rbuf3[i]);
1378:     PetscFree(rbuf4[i]);
1379:   }

1381:   PetscFree(sbuf2);
1382:   PetscFree(rbuf3);
1383:   PetscFree(rbuf4);
1384:   PetscFree(sbuf_aj[0]);
1385:   PetscFree(sbuf_aj);
1386:   PetscFree(sbuf_aa[0]);
1387:   PetscFree(sbuf_aa);
1388: 
1389:   PetscFree(cmap);
1390:   PetscFree(rmap);
1391:   PetscFree(lens);

1393:   for (i=0; i<ismax; i++) {
1394:     MatAssemblyBegin(submats[i],MAT_FINAL_ASSEMBLY);
1395:     MatAssemblyEnd(submats[i],MAT_FINAL_ASSEMBLY);
1396:   }
1397:   return(0);
1398: }