Actual source code: mpidense.c
1: #define PETSCMAT_DLL
3: /*
4: Basic functions for basic parallel dense matrices.
5: */
6:
7: #include src/mat/impls/dense/mpi/mpidense.h
11: PetscErrorCode MatLUFactorSymbolic_MPIDense(Mat A,IS row,IS col,MatFactorInfo *info,Mat *fact)
12: {
16: MatDuplicate(A,MAT_DO_NOT_COPY_VALUES,fact);
17: return(0);
18: }
22: PetscErrorCode MatCholeskyFactorSymbolic_MPIDense(Mat A,IS perm,MatFactorInfo *info,Mat *fact)
23: {
27: MatDuplicate(A,MAT_DO_NOT_COPY_VALUES,fact);
28: return(0);
29: }
33: /*@
35: MatDenseGetLocalMatrix - For a MATMPIDENSE or MATSEQDENSE matrix returns the sequential
36: matrix that represents the operator. For sequential matrices it returns itself.
38: Input Parameter:
39: . A - the Seq or MPI dense matrix
41: Output Parameter:
42: . B - the inner matrix
44: Level: intermediate
46: @*/
47: PetscErrorCode MatDenseGetLocalMatrix(Mat A,Mat *B)
48: {
49: Mat_MPIDense *mat = (Mat_MPIDense*)A->data;
51: PetscTruth flg;
52: PetscMPIInt size;
55: PetscTypeCompare((PetscObject)A,MATMPIDENSE,&flg);
56: if (!flg) { /* this check sucks! */
57: PetscTypeCompare((PetscObject)A,MATDENSE,&flg);
58: if (flg) {
59: MPI_Comm_size(A->comm,&size);
60: if (size == 1) flg = PETSC_FALSE;
61: }
62: }
63: if (flg) {
64: *B = mat->A;
65: } else {
66: *B = A;
67: }
68: return(0);
69: }
73: PetscErrorCode MatGetRow_MPIDense(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v)
74: {
75: Mat_MPIDense *mat = (Mat_MPIDense*)A->data;
77: PetscInt lrow,rstart = A->rmap.rstart,rend = A->rmap.rend;
80: if (row < rstart || row >= rend) SETERRQ(PETSC_ERR_SUP,"only local rows")
81: lrow = row - rstart;
82: MatGetRow(mat->A,lrow,nz,(const PetscInt **)idx,(const PetscScalar **)v);
83: return(0);
84: }
88: PetscErrorCode MatRestoreRow_MPIDense(Mat mat,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v)
89: {
93: if (idx) {PetscFree(*idx);}
94: if (v) {PetscFree(*v);}
95: return(0);
96: }
101: PetscErrorCode MatGetDiagonalBlock_MPIDense(Mat A,PetscTruth *iscopy,MatReuse reuse,Mat *B)
102: {
103: Mat_MPIDense *mdn = (Mat_MPIDense*)A->data;
105: PetscInt m = A->rmap.n,rstart = A->rmap.rstart;
106: PetscScalar *array;
107: MPI_Comm comm;
110: if (A->rmap.N != A->cmap.N) SETERRQ(PETSC_ERR_SUP,"Only square matrices supported.");
112: /* The reuse aspect is not implemented efficiently */
113: if (reuse) { MatDestroy(*B);}
115: PetscObjectGetComm((PetscObject)(mdn->A),&comm);
116: MatGetArray(mdn->A,&array);
117: MatCreate(comm,B);
118: MatSetSizes(*B,m,m,m,m);
119: MatSetType(*B,mdn->A->type_name);
120: MatSeqDenseSetPreallocation(*B,array+m*rstart);
121: MatRestoreArray(mdn->A,&array);
122: MatAssemblyBegin(*B,MAT_FINAL_ASSEMBLY);
123: MatAssemblyEnd(*B,MAT_FINAL_ASSEMBLY);
124:
125: *iscopy = PETSC_TRUE;
126: return(0);
127: }
132: PetscErrorCode MatSetValues_MPIDense(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],const PetscScalar v[],InsertMode addv)
133: {
134: Mat_MPIDense *A = (Mat_MPIDense*)mat->data;
136: PetscInt i,j,rstart = mat->rmap.rstart,rend = mat->rmap.rend,row;
137: PetscTruth roworiented = A->roworiented;
140: for (i=0; i<m; i++) {
141: if (idxm[i] < 0) continue;
142: if (idxm[i] >= mat->rmap.N) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Row too large");
143: if (idxm[i] >= rstart && idxm[i] < rend) {
144: row = idxm[i] - rstart;
145: if (roworiented) {
146: MatSetValues(A->A,1,&row,n,idxn,v+i*n,addv);
147: } else {
148: for (j=0; j<n; j++) {
149: if (idxn[j] < 0) continue;
150: if (idxn[j] >= mat->cmap.N) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Column too large");
151: MatSetValues(A->A,1,&row,1,&idxn[j],v+i+j*m,addv);
152: }
153: }
154: } else {
155: if (!A->donotstash) {
156: if (roworiented) {
157: MatStashValuesRow_Private(&mat->stash,idxm[i],n,idxn,v+i*n);
158: } else {
159: MatStashValuesCol_Private(&mat->stash,idxm[i],n,idxn,v+i,m);
160: }
161: }
162: }
163: }
164: return(0);
165: }
169: PetscErrorCode MatGetValues_MPIDense(Mat mat,PetscInt m,const PetscInt idxm[],PetscInt n,const PetscInt idxn[],PetscScalar v[])
170: {
171: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
173: PetscInt i,j,rstart = mat->rmap.rstart,rend = mat->rmap.rend,row;
176: for (i=0; i<m; i++) {
177: if (idxm[i] < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative row");
178: if (idxm[i] >= mat->rmap.N) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Row too large");
179: if (idxm[i] >= rstart && idxm[i] < rend) {
180: row = idxm[i] - rstart;
181: for (j=0; j<n; j++) {
182: if (idxn[j] < 0) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Negative column");
183: if (idxn[j] >= mat->cmap.N) {
184: SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Column too large");
185: }
186: MatGetValues(mdn->A,1,&row,1,&idxn[j],v+i*n+j);
187: }
188: } else {
189: SETERRQ(PETSC_ERR_SUP,"Only local values currently supported");
190: }
191: }
192: return(0);
193: }
197: PetscErrorCode MatGetArray_MPIDense(Mat A,PetscScalar *array[])
198: {
199: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
203: MatGetArray(a->A,array);
204: return(0);
205: }
209: static PetscErrorCode MatGetSubMatrix_MPIDense(Mat A,IS isrow,IS iscol,PetscInt cs,MatReuse scall,Mat *B)
210: {
211: Mat_MPIDense *mat = (Mat_MPIDense*)A->data,*newmatd;
212: Mat_SeqDense *lmat = (Mat_SeqDense*)mat->A->data;
214: PetscInt i,j,*irow,*icol,rstart,rend,nrows,ncols,nlrows,nlcols;
215: PetscScalar *av,*bv,*v = lmat->v;
216: Mat newmat;
219: ISGetIndices(isrow,&irow);
220: ISGetIndices(iscol,&icol);
221: ISGetLocalSize(isrow,&nrows);
222: ISGetLocalSize(iscol,&ncols);
224: /* No parallel redistribution currently supported! Should really check each index set
225: to comfirm that it is OK. ... Currently supports only submatrix same partitioning as
226: original matrix! */
228: MatGetLocalSize(A,&nlrows,&nlcols);
229: MatGetOwnershipRange(A,&rstart,&rend);
230:
231: /* Check submatrix call */
232: if (scall == MAT_REUSE_MATRIX) {
233: /* SETERRQ(PETSC_ERR_ARG_SIZ,"Reused submatrix wrong size"); */
234: /* Really need to test rows and column sizes! */
235: newmat = *B;
236: } else {
237: /* Create and fill new matrix */
238: MatCreate(A->comm,&newmat);
239: MatSetSizes(newmat,nrows,cs,PETSC_DECIDE,ncols);
240: MatSetType(newmat,A->type_name);
241: MatMPIDenseSetPreallocation(newmat,PETSC_NULL);
242: }
244: /* Now extract the data pointers and do the copy, column at a time */
245: newmatd = (Mat_MPIDense*)newmat->data;
246: bv = ((Mat_SeqDense *)newmatd->A->data)->v;
247:
248: for (i=0; i<ncols; i++) {
249: av = v + nlrows*icol[i];
250: for (j=0; j<nrows; j++) {
251: *bv++ = av[irow[j] - rstart];
252: }
253: }
255: /* Assemble the matrices so that the correct flags are set */
256: MatAssemblyBegin(newmat,MAT_FINAL_ASSEMBLY);
257: MatAssemblyEnd(newmat,MAT_FINAL_ASSEMBLY);
259: /* Free work space */
260: ISRestoreIndices(isrow,&irow);
261: ISRestoreIndices(iscol,&icol);
262: *B = newmat;
263: return(0);
264: }
268: PetscErrorCode MatRestoreArray_MPIDense(Mat A,PetscScalar *array[])
269: {
271: return(0);
272: }
276: PetscErrorCode MatAssemblyBegin_MPIDense(Mat mat,MatAssemblyType mode)
277: {
278: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
279: MPI_Comm comm = mat->comm;
281: PetscInt nstash,reallocs;
282: InsertMode addv;
285: /* make sure all processors are either in INSERTMODE or ADDMODE */
286: MPI_Allreduce(&mat->insertmode,&addv,1,MPI_INT,MPI_BOR,comm);
287: if (addv == (ADD_VALUES|INSERT_VALUES)) {
288: SETERRQ(PETSC_ERR_ARG_WRONGSTATE,"Cannot mix adds/inserts on different procs");
289: }
290: mat->insertmode = addv; /* in case this processor had no cache */
292: MatStashScatterBegin_Private(&mat->stash,mat->rmap.range);
293: MatStashGetInfo_Private(&mat->stash,&nstash,&reallocs);
294: PetscInfo2(mdn->A,"Stash has %D entries, uses %D mallocs.\n",nstash,reallocs);
295: return(0);
296: }
300: PetscErrorCode MatAssemblyEnd_MPIDense(Mat mat,MatAssemblyType mode)
301: {
302: Mat_MPIDense *mdn=(Mat_MPIDense*)mat->data;
303: PetscErrorCode ierr;
304: PetscInt i,*row,*col,flg,j,rstart,ncols;
305: PetscMPIInt n;
306: PetscScalar *val;
307: InsertMode addv=mat->insertmode;
310: /* wait on receives */
311: while (1) {
312: MatStashScatterGetMesg_Private(&mat->stash,&n,&row,&col,&val,&flg);
313: if (!flg) break;
314:
315: for (i=0; i<n;) {
316: /* Now identify the consecutive vals belonging to the same row */
317: for (j=i,rstart=row[j]; j<n; j++) { if (row[j] != rstart) break; }
318: if (j < n) ncols = j-i;
319: else ncols = n-i;
320: /* Now assemble all these values with a single function call */
321: MatSetValues_MPIDense(mat,1,row+i,ncols,col+i,val+i,addv);
322: i = j;
323: }
324: }
325: MatStashScatterEnd_Private(&mat->stash);
326:
327: MatAssemblyBegin(mdn->A,mode);
328: MatAssemblyEnd(mdn->A,mode);
330: if (!mat->was_assembled && mode == MAT_FINAL_ASSEMBLY) {
331: MatSetUpMultiply_MPIDense(mat);
332: }
333: return(0);
334: }
338: PetscErrorCode MatZeroEntries_MPIDense(Mat A)
339: {
341: Mat_MPIDense *l = (Mat_MPIDense*)A->data;
344: MatZeroEntries(l->A);
345: return(0);
346: }
348: /* the code does not do the diagonal entries correctly unless the
349: matrix is square and the column and row owerships are identical.
350: This is a BUG. The only way to fix it seems to be to access
351: mdn->A and mdn->B directly and not through the MatZeroRows()
352: routine.
353: */
356: PetscErrorCode MatZeroRows_MPIDense(Mat A,PetscInt N,const PetscInt rows[],PetscScalar diag)
357: {
358: Mat_MPIDense *l = (Mat_MPIDense*)A->data;
360: PetscInt i,*owners = A->rmap.range;
361: PetscInt *nprocs,j,idx,nsends;
362: PetscInt nmax,*svalues,*starts,*owner,nrecvs;
363: PetscInt *rvalues,tag = A->tag,count,base,slen,*source;
364: PetscInt *lens,*lrows,*values;
365: PetscMPIInt n,imdex,rank = l->rank,size = l->size;
366: MPI_Comm comm = A->comm;
367: MPI_Request *send_waits,*recv_waits;
368: MPI_Status recv_status,*send_status;
369: PetscTruth found;
372: /* first count number of contributors to each processor */
373: PetscMalloc(2*size*sizeof(PetscInt),&nprocs);
374: PetscMemzero(nprocs,2*size*sizeof(PetscInt));
375: PetscMalloc((N+1)*sizeof(PetscInt),&owner); /* see note*/
376: for (i=0; i<N; i++) {
377: idx = rows[i];
378: found = PETSC_FALSE;
379: for (j=0; j<size; j++) {
380: if (idx >= owners[j] && idx < owners[j+1]) {
381: nprocs[2*j]++; nprocs[2*j+1] = 1; owner[i] = j; found = PETSC_TRUE; break;
382: }
383: }
384: if (!found) SETERRQ(PETSC_ERR_ARG_OUTOFRANGE,"Index out of range");
385: }
386: nsends = 0; for (i=0; i<size; i++) { nsends += nprocs[2*i+1];}
388: /* inform other processors of number of messages and max length*/
389: PetscMaxSum(comm,nprocs,&nmax,&nrecvs);
391: /* post receives: */
392: PetscMalloc((nrecvs+1)*(nmax+1)*sizeof(PetscInt),&rvalues);
393: PetscMalloc((nrecvs+1)*sizeof(MPI_Request),&recv_waits);
394: for (i=0; i<nrecvs; i++) {
395: MPI_Irecv(rvalues+nmax*i,nmax,MPIU_INT,MPI_ANY_SOURCE,tag,comm,recv_waits+i);
396: }
398: /* do sends:
399: 1) starts[i] gives the starting index in svalues for stuff going to
400: the ith processor
401: */
402: PetscMalloc((N+1)*sizeof(PetscInt),&svalues);
403: PetscMalloc((nsends+1)*sizeof(MPI_Request),&send_waits);
404: PetscMalloc((size+1)*sizeof(PetscInt),&starts);
405: starts[0] = 0;
406: for (i=1; i<size; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];}
407: for (i=0; i<N; i++) {
408: svalues[starts[owner[i]]++] = rows[i];
409: }
411: starts[0] = 0;
412: for (i=1; i<size+1; i++) { starts[i] = starts[i-1] + nprocs[2*i-2];}
413: count = 0;
414: for (i=0; i<size; i++) {
415: if (nprocs[2*i+1]) {
416: MPI_Isend(svalues+starts[i],nprocs[2*i],MPIU_INT,i,tag,comm,send_waits+count++);
417: }
418: }
419: PetscFree(starts);
421: base = owners[rank];
423: /* wait on receives */
424: PetscMalloc(2*(nrecvs+1)*sizeof(PetscInt),&lens);
425: source = lens + nrecvs;
426: count = nrecvs; slen = 0;
427: while (count) {
428: MPI_Waitany(nrecvs,recv_waits,&imdex,&recv_status);
429: /* unpack receives into our local space */
430: MPI_Get_count(&recv_status,MPIU_INT,&n);
431: source[imdex] = recv_status.MPI_SOURCE;
432: lens[imdex] = n;
433: slen += n;
434: count--;
435: }
436: PetscFree(recv_waits);
437:
438: /* move the data into the send scatter */
439: PetscMalloc((slen+1)*sizeof(PetscInt),&lrows);
440: count = 0;
441: for (i=0; i<nrecvs; i++) {
442: values = rvalues + i*nmax;
443: for (j=0; j<lens[i]; j++) {
444: lrows[count++] = values[j] - base;
445: }
446: }
447: PetscFree(rvalues);
448: PetscFree(lens);
449: PetscFree(owner);
450: PetscFree(nprocs);
451:
452: /* actually zap the local rows */
453: MatZeroRows(l->A,slen,lrows,diag);
454: PetscFree(lrows);
456: /* wait on sends */
457: if (nsends) {
458: PetscMalloc(nsends*sizeof(MPI_Status),&send_status);
459: MPI_Waitall(nsends,send_waits,send_status);
460: PetscFree(send_status);
461: }
462: PetscFree(send_waits);
463: PetscFree(svalues);
465: return(0);
466: }
470: PetscErrorCode MatMult_MPIDense(Mat mat,Vec xx,Vec yy)
471: {
472: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
476: VecScatterBegin(xx,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
477: VecScatterEnd(xx,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
478: MatMult_SeqDense(mdn->A,mdn->lvec,yy);
479: return(0);
480: }
484: PetscErrorCode MatMultAdd_MPIDense(Mat mat,Vec xx,Vec yy,Vec zz)
485: {
486: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
490: VecScatterBegin(xx,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
491: VecScatterEnd(xx,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
492: MatMultAdd_SeqDense(mdn->A,mdn->lvec,yy,zz);
493: return(0);
494: }
498: PetscErrorCode MatMultTranspose_MPIDense(Mat A,Vec xx,Vec yy)
499: {
500: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
502: PetscScalar zero = 0.0;
505: VecSet(yy,zero);
506: MatMultTranspose_SeqDense(a->A,xx,a->lvec);
507: VecScatterBegin(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);
508: VecScatterEnd(a->lvec,yy,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);
509: return(0);
510: }
514: PetscErrorCode MatMultTransposeAdd_MPIDense(Mat A,Vec xx,Vec yy,Vec zz)
515: {
516: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
520: VecCopy(yy,zz);
521: MatMultTranspose_SeqDense(a->A,xx,a->lvec);
522: VecScatterBegin(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);
523: VecScatterEnd(a->lvec,zz,ADD_VALUES,SCATTER_REVERSE,a->Mvctx);
524: return(0);
525: }
529: PetscErrorCode MatGetDiagonal_MPIDense(Mat A,Vec v)
530: {
531: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
532: Mat_SeqDense *aloc = (Mat_SeqDense*)a->A->data;
534: PetscInt len,i,n,m = A->rmap.n,radd;
535: PetscScalar *x,zero = 0.0;
536:
538: VecSet(v,zero);
539: VecGetArray(v,&x);
540: VecGetSize(v,&n);
541: if (n != A->rmap.N) SETERRQ(PETSC_ERR_ARG_SIZ,"Nonconforming mat and vec");
542: len = PetscMin(a->A->rmap.n,a->A->cmap.n);
543: radd = A->rmap.rstart*m;
544: for (i=0; i<len; i++) {
545: x[i] = aloc->v[radd + i*m + i];
546: }
547: VecRestoreArray(v,&x);
548: return(0);
549: }
553: PetscErrorCode MatDestroy_MPIDense(Mat mat)
554: {
555: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
560: #if defined(PETSC_USE_LOG)
561: PetscLogObjectState((PetscObject)mat,"Rows=%D, Cols=%D",mat->rmap.N,mat->cmap.N);
562: #endif
563: MatStashDestroy_Private(&mat->stash);
564: MatDestroy(mdn->A);
565: if (mdn->lvec) {VecDestroy(mdn->lvec);}
566: if (mdn->Mvctx) {VecScatterDestroy(mdn->Mvctx);}
567: if (mdn->factor) {
568: PetscFree(mdn->factor->temp);
569: PetscFree(mdn->factor->tag);
570: PetscFree(mdn->factor->pivots);
571: PetscFree(mdn->factor);
572: }
573: PetscFree(mdn);
574: PetscObjectChangeTypeName((PetscObject)mat,0);
575: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatGetDiagonalBlock_C","",PETSC_NULL);
576: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatMPIDenseSetPreallocation_C","",PETSC_NULL);
577: PetscObjectComposeFunction((PetscObject)mat,"MatMatMult_mpiaij_mpidense_C","",PETSC_NULL);
578: PetscObjectComposeFunction((PetscObject)mat,"MatMatMultSymbolic_mpiaij_mpidense_C","",PETSC_NULL);
579: PetscObjectComposeFunction((PetscObject)mat,"MatMatMultNumeric_mpiaij_mpidense_C","",PETSC_NULL);
580: return(0);
581: }
585: static PetscErrorCode MatView_MPIDense_Binary(Mat mat,PetscViewer viewer)
586: {
587: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
591: if (mdn->size == 1) {
592: MatView(mdn->A,viewer);
593: }
594: else SETERRQ(PETSC_ERR_SUP,"Only uniprocessor output supported");
595: return(0);
596: }
600: static PetscErrorCode MatView_MPIDense_ASCIIorDraworSocket(Mat mat,PetscViewer viewer)
601: {
602: Mat_MPIDense *mdn = (Mat_MPIDense*)mat->data;
603: PetscErrorCode ierr;
604: PetscMPIInt size = mdn->size,rank = mdn->rank;
605: PetscViewerType vtype;
606: PetscTruth iascii,isdraw;
607: PetscViewer sviewer;
608: PetscViewerFormat format;
611: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);
612: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);
613: if (iascii) {
614: PetscViewerGetType(viewer,&vtype);
615: PetscViewerGetFormat(viewer,&format);
616: if (format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
617: MatInfo info;
618: MatGetInfo(mat,MAT_LOCAL,&info);
619: PetscViewerASCIISynchronizedPrintf(viewer," [%d] local rows %D nz %D nz alloced %D mem %D \n",rank,mat->rmap.n,
620: (PetscInt)info.nz_used,(PetscInt)info.nz_allocated,(PetscInt)info.memory);
621: PetscViewerFlush(viewer);
622: VecScatterView(mdn->Mvctx,viewer);
623: return(0);
624: } else if (format == PETSC_VIEWER_ASCII_INFO) {
625: return(0);
626: }
627: } else if (isdraw) {
628: PetscDraw draw;
629: PetscTruth isnull;
631: PetscViewerDrawGetDraw(viewer,0,&draw);
632: PetscDrawIsNull(draw,&isnull);
633: if (isnull) return(0);
634: }
636: if (size == 1) {
637: MatView(mdn->A,viewer);
638: } else {
639: /* assemble the entire matrix onto first processor. */
640: Mat A;
641: PetscInt M = mat->rmap.N,N = mat->cmap.N,m,row,i,nz;
642: PetscInt *cols;
643: PetscScalar *vals;
645: MatCreate(mat->comm,&A);
646: if (!rank) {
647: MatSetSizes(A,M,N,M,N);
648: } else {
649: MatSetSizes(A,0,0,M,N);
650: }
651: /* Since this is a temporary matrix, MATMPIDENSE instead of A->type_name here is probably acceptable. */
652: MatSetType(A,MATMPIDENSE);
653: MatMPIDenseSetPreallocation(A,PETSC_NULL);
654: PetscLogObjectParent(mat,A);
656: /* Copy the matrix ... This isn't the most efficient means,
657: but it's quick for now */
658: A->insertmode = INSERT_VALUES;
659: row = mat->rmap.rstart; m = mdn->A->rmap.n;
660: for (i=0; i<m; i++) {
661: MatGetRow_MPIDense(mat,row,&nz,&cols,&vals);
662: MatSetValues_MPIDense(A,1,&row,nz,cols,vals,INSERT_VALUES);
663: MatRestoreRow_MPIDense(mat,row,&nz,&cols,&vals);
664: row++;
665: }
667: MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);
668: MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);
669: PetscViewerGetSingleton(viewer,&sviewer);
670: if (!rank) {
671: MatView(((Mat_MPIDense*)(A->data))->A,sviewer);
672: }
673: PetscViewerRestoreSingleton(viewer,&sviewer);
674: PetscViewerFlush(viewer);
675: MatDestroy(A);
676: }
677: return(0);
678: }
682: PetscErrorCode MatView_MPIDense(Mat mat,PetscViewer viewer)
683: {
685: PetscTruth iascii,isbinary,isdraw,issocket;
686:
688:
689: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_ASCII,&iascii);
690: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_BINARY,&isbinary);
691: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_SOCKET,&issocket);
692: PetscTypeCompare((PetscObject)viewer,PETSC_VIEWER_DRAW,&isdraw);
694: if (iascii || issocket || isdraw) {
695: MatView_MPIDense_ASCIIorDraworSocket(mat,viewer);
696: } else if (isbinary) {
697: MatView_MPIDense_Binary(mat,viewer);
698: } else {
699: SETERRQ1(PETSC_ERR_SUP,"Viewer type %s not supported by MPI dense matrix",((PetscObject)viewer)->type_name);
700: }
701: return(0);
702: }
706: PetscErrorCode MatGetInfo_MPIDense(Mat A,MatInfoType flag,MatInfo *info)
707: {
708: Mat_MPIDense *mat = (Mat_MPIDense*)A->data;
709: Mat mdn = mat->A;
711: PetscReal isend[5],irecv[5];
714: info->rows_global = (double)A->rmap.N;
715: info->columns_global = (double)A->cmap.N;
716: info->rows_local = (double)A->rmap.n;
717: info->columns_local = (double)A->cmap.N;
718: info->block_size = 1.0;
719: MatGetInfo(mdn,MAT_LOCAL,info);
720: isend[0] = info->nz_used; isend[1] = info->nz_allocated; isend[2] = info->nz_unneeded;
721: isend[3] = info->memory; isend[4] = info->mallocs;
722: if (flag == MAT_LOCAL) {
723: info->nz_used = isend[0];
724: info->nz_allocated = isend[1];
725: info->nz_unneeded = isend[2];
726: info->memory = isend[3];
727: info->mallocs = isend[4];
728: } else if (flag == MAT_GLOBAL_MAX) {
729: MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_MAX,A->comm);
730: info->nz_used = irecv[0];
731: info->nz_allocated = irecv[1];
732: info->nz_unneeded = irecv[2];
733: info->memory = irecv[3];
734: info->mallocs = irecv[4];
735: } else if (flag == MAT_GLOBAL_SUM) {
736: MPI_Allreduce(isend,irecv,5,MPIU_REAL,MPI_SUM,A->comm);
737: info->nz_used = irecv[0];
738: info->nz_allocated = irecv[1];
739: info->nz_unneeded = irecv[2];
740: info->memory = irecv[3];
741: info->mallocs = irecv[4];
742: }
743: info->fill_ratio_given = 0; /* no parallel LU/ILU/Cholesky */
744: info->fill_ratio_needed = 0;
745: info->factor_mallocs = 0;
746: return(0);
747: }
751: PetscErrorCode MatSetOption_MPIDense(Mat A,MatOption op)
752: {
753: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
757: switch (op) {
758: case MAT_NO_NEW_NONZERO_LOCATIONS:
759: case MAT_YES_NEW_NONZERO_LOCATIONS:
760: case MAT_NEW_NONZERO_LOCATION_ERR:
761: case MAT_NEW_NONZERO_ALLOCATION_ERR:
762: case MAT_COLUMNS_SORTED:
763: case MAT_COLUMNS_UNSORTED:
764: MatSetOption(a->A,op);
765: break;
766: case MAT_ROW_ORIENTED:
767: a->roworiented = PETSC_TRUE;
768: MatSetOption(a->A,op);
769: break;
770: case MAT_ROWS_SORTED:
771: case MAT_ROWS_UNSORTED:
772: case MAT_YES_NEW_DIAGONALS:
773: case MAT_USE_HASH_TABLE:
774: PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);
775: break;
776: case MAT_COLUMN_ORIENTED:
777: a->roworiented = PETSC_FALSE;
778: MatSetOption(a->A,op);
779: break;
780: case MAT_IGNORE_OFF_PROC_ENTRIES:
781: a->donotstash = PETSC_TRUE;
782: break;
783: case MAT_NO_NEW_DIAGONALS:
784: SETERRQ(PETSC_ERR_SUP,"MAT_NO_NEW_DIAGONALS");
785: case MAT_SYMMETRIC:
786: case MAT_STRUCTURALLY_SYMMETRIC:
787: case MAT_NOT_SYMMETRIC:
788: case MAT_NOT_STRUCTURALLY_SYMMETRIC:
789: case MAT_HERMITIAN:
790: case MAT_NOT_HERMITIAN:
791: case MAT_SYMMETRY_ETERNAL:
792: case MAT_NOT_SYMMETRY_ETERNAL:
793: PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);
794: break;
795: default:
796: SETERRQ1(PETSC_ERR_SUP,"unknown option %d",op);
797: }
798: return(0);
799: }
804: PetscErrorCode MatDiagonalScale_MPIDense(Mat A,Vec ll,Vec rr)
805: {
806: Mat_MPIDense *mdn = (Mat_MPIDense*)A->data;
807: Mat_SeqDense *mat = (Mat_SeqDense*)mdn->A->data;
808: PetscScalar *l,*r,x,*v;
810: PetscInt i,j,s2a,s3a,s2,s3,m=mdn->A->rmap.n,n=mdn->A->cmap.n;
813: MatGetLocalSize(A,&s2,&s3);
814: if (ll) {
815: VecGetLocalSize(ll,&s2a);
816: if (s2a != s2) SETERRQ2(PETSC_ERR_ARG_SIZ,"Left scaling vector non-conforming local size, %d != %d.", s2a, s2);
817: VecGetArray(ll,&l);
818: for (i=0; i<m; i++) {
819: x = l[i];
820: v = mat->v + i;
821: for (j=0; j<n; j++) { (*v) *= x; v+= m;}
822: }
823: VecRestoreArray(ll,&l);
824: PetscLogFlops(n*m);
825: }
826: if (rr) {
827: VecGetLocalSize(rr,&s3a);
828: if (s3a != s3) SETERRQ2(PETSC_ERR_ARG_SIZ,"Right scaling vec non-conforming local size, %d != %d.", s3a, s3);
829: VecScatterBegin(rr,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
830: VecScatterEnd(rr,mdn->lvec,INSERT_VALUES,SCATTER_FORWARD,mdn->Mvctx);
831: VecGetArray(mdn->lvec,&r);
832: for (i=0; i<n; i++) {
833: x = r[i];
834: v = mat->v + i*m;
835: for (j=0; j<m; j++) { (*v++) *= x;}
836: }
837: VecRestoreArray(mdn->lvec,&r);
838: PetscLogFlops(n*m);
839: }
840: return(0);
841: }
845: PetscErrorCode MatNorm_MPIDense(Mat A,NormType type,PetscReal *nrm)
846: {
847: Mat_MPIDense *mdn = (Mat_MPIDense*)A->data;
848: Mat_SeqDense *mat = (Mat_SeqDense*)mdn->A->data;
850: PetscInt i,j;
851: PetscReal sum = 0.0;
852: PetscScalar *v = mat->v;
855: if (mdn->size == 1) {
856: MatNorm(mdn->A,type,nrm);
857: } else {
858: if (type == NORM_FROBENIUS) {
859: for (i=0; i<mdn->A->cmap.n*mdn->A->rmap.n; i++) {
860: #if defined(PETSC_USE_COMPLEX)
861: sum += PetscRealPart(PetscConj(*v)*(*v)); v++;
862: #else
863: sum += (*v)*(*v); v++;
864: #endif
865: }
866: MPI_Allreduce(&sum,nrm,1,MPIU_REAL,MPI_SUM,A->comm);
867: *nrm = sqrt(*nrm);
868: PetscLogFlops(2*mdn->A->cmap.n*mdn->A->rmap.n);
869: } else if (type == NORM_1) {
870: PetscReal *tmp,*tmp2;
871: PetscMalloc(2*A->cmap.N*sizeof(PetscReal),&tmp);
872: tmp2 = tmp + A->cmap.N;
873: PetscMemzero(tmp,2*A->cmap.N*sizeof(PetscReal));
874: *nrm = 0.0;
875: v = mat->v;
876: for (j=0; j<mdn->A->cmap.n; j++) {
877: for (i=0; i<mdn->A->rmap.n; i++) {
878: tmp[j] += PetscAbsScalar(*v); v++;
879: }
880: }
881: MPI_Allreduce(tmp,tmp2,A->cmap.N,MPIU_REAL,MPI_SUM,A->comm);
882: for (j=0; j<A->cmap.N; j++) {
883: if (tmp2[j] > *nrm) *nrm = tmp2[j];
884: }
885: PetscFree(tmp);
886: PetscLogFlops(A->cmap.n*A->rmap.n);
887: } else if (type == NORM_INFINITY) { /* max row norm */
888: PetscReal ntemp;
889: MatNorm(mdn->A,type,&ntemp);
890: MPI_Allreduce(&ntemp,nrm,1,MPIU_REAL,MPI_MAX,A->comm);
891: } else {
892: SETERRQ(PETSC_ERR_SUP,"No support for two norm");
893: }
894: }
895: return(0);
896: }
900: PetscErrorCode MatTranspose_MPIDense(Mat A,Mat *matout)
901: {
902: Mat_MPIDense *a = (Mat_MPIDense*)A->data;
903: Mat_SeqDense *Aloc = (Mat_SeqDense*)a->A->data;
904: Mat B;
905: PetscInt M = A->rmap.N,N = A->cmap.N,m,n,*rwork,rstart = A->rmap.rstart;
907: PetscInt j,i;
908: PetscScalar *v;
911: if (!matout && M != N) {
912: SETERRQ(PETSC_ERR_SUP,"Supports square matrix only in-place");
913: }
914: MatCreate(A->comm,&B);
915: MatSetSizes(B,PETSC_DECIDE,PETSC_DECIDE,N,M);
916: MatSetType(B,A->type_name);
917: MatMPIDenseSetPreallocation(B,PETSC_NULL);
919: m = a->A->rmap.n; n = a->A->cmap.n; v = Aloc->v;
920: PetscMalloc(n*sizeof(PetscInt),&rwork);
921: for (j=0; j<n; j++) {
922: for (i=0; i<m; i++) rwork[i] = rstart + i;
923: MatSetValues(B,1,&j,m,rwork,v,INSERT_VALUES);
924: v += m;
925: }
926: PetscFree(rwork);
927: MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
928: MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);
929: if (matout) {
930: *matout = B;
931: } else {
932: MatHeaderCopy(A,B);
933: }
934: return(0);
935: }
937: #include petscblaslapack.h
940: PetscErrorCode MatScale_MPIDense(Mat inA,PetscScalar alpha)
941: {
942: Mat_MPIDense *A = (Mat_MPIDense*)inA->data;
943: Mat_SeqDense *a = (Mat_SeqDense*)A->A->data;
944: PetscScalar oalpha = alpha;
945: PetscBLASInt one = 1,nz = (PetscBLASInt)inA->rmap.n*inA->cmap.N;
949: BLASscal_(&nz,&oalpha,a->v,&one);
950: PetscLogFlops(nz);
951: return(0);
952: }
954: static PetscErrorCode MatDuplicate_MPIDense(Mat,MatDuplicateOption,Mat *);
958: PetscErrorCode MatSetUpPreallocation_MPIDense(Mat A)
959: {
963: MatMPIDenseSetPreallocation(A,0);
964: return(0);
965: }
969: PetscErrorCode MatMatMultSymbolic_MPIDense_MPIDense(Mat A,Mat B,PetscReal fill,Mat *C)
970: {
972: PetscInt m=A->rmap.n,n=B->cmap.n;
973: Mat Cmat;
976: if (A->cmap.n != B->rmap.n) SETERRQ2(PETSC_ERR_ARG_SIZ,"A->cmap.n %d != B->rmap.n %d\n",A->cmap.n,B->rmap.n);
977: MatCreate(B->comm,&Cmat);
978: MatSetSizes(Cmat,m,n,A->rmap.N,B->cmap.N);
979: MatSetType(Cmat,MATMPIDENSE);
980: MatMPIDenseSetPreallocation(Cmat,PETSC_NULL);
981: MatAssemblyBegin(Cmat,MAT_FINAL_ASSEMBLY);
982: MatAssemblyEnd(Cmat,MAT_FINAL_ASSEMBLY);
983: *C = Cmat;
984: return(0);
985: }
987: /* -------------------------------------------------------------------*/
988: static struct _MatOps MatOps_Values = {MatSetValues_MPIDense,
989: MatGetRow_MPIDense,
990: MatRestoreRow_MPIDense,
991: MatMult_MPIDense,
992: /* 4*/ MatMultAdd_MPIDense,
993: MatMultTranspose_MPIDense,
994: MatMultTransposeAdd_MPIDense,
995: 0,
996: 0,
997: 0,
998: /*10*/ 0,
999: 0,
1000: 0,
1001: 0,
1002: MatTranspose_MPIDense,
1003: /*15*/ MatGetInfo_MPIDense,
1004: MatEqual_MPIDense,
1005: MatGetDiagonal_MPIDense,
1006: MatDiagonalScale_MPIDense,
1007: MatNorm_MPIDense,
1008: /*20*/ MatAssemblyBegin_MPIDense,
1009: MatAssemblyEnd_MPIDense,
1010: 0,
1011: MatSetOption_MPIDense,
1012: MatZeroEntries_MPIDense,
1013: /*25*/ MatZeroRows_MPIDense,
1014: MatLUFactorSymbolic_MPIDense,
1015: 0,
1016: MatCholeskyFactorSymbolic_MPIDense,
1017: 0,
1018: /*30*/ MatSetUpPreallocation_MPIDense,
1019: 0,
1020: 0,
1021: MatGetArray_MPIDense,
1022: MatRestoreArray_MPIDense,
1023: /*35*/ MatDuplicate_MPIDense,
1024: 0,
1025: 0,
1026: 0,
1027: 0,
1028: /*40*/ 0,
1029: MatGetSubMatrices_MPIDense,
1030: 0,
1031: MatGetValues_MPIDense,
1032: 0,
1033: /*45*/ 0,
1034: MatScale_MPIDense,
1035: 0,
1036: 0,
1037: 0,
1038: /*50*/ 0,
1039: 0,
1040: 0,
1041: 0,
1042: 0,
1043: /*55*/ 0,
1044: 0,
1045: 0,
1046: 0,
1047: 0,
1048: /*60*/ MatGetSubMatrix_MPIDense,
1049: MatDestroy_MPIDense,
1050: MatView_MPIDense,
1051: 0,
1052: 0,
1053: /*65*/ 0,
1054: 0,
1055: 0,
1056: 0,
1057: 0,
1058: /*70*/ 0,
1059: 0,
1060: 0,
1061: 0,
1062: 0,
1063: /*75*/ 0,
1064: 0,
1065: 0,
1066: 0,
1067: 0,
1068: /*80*/ 0,
1069: 0,
1070: 0,
1071: 0,
1072: /*84*/ MatLoad_MPIDense,
1073: 0,
1074: 0,
1075: 0,
1076: 0,
1077: 0,
1078: /*90*/ 0,
1079: MatMatMultSymbolic_MPIDense_MPIDense,
1080: 0,
1081: 0,
1082: 0,
1083: /*95*/ 0,
1084: 0,
1085: 0,
1086: 0};
1091: PetscErrorCode MatMPIDenseSetPreallocation_MPIDense(Mat mat,PetscScalar *data)
1092: {
1093: Mat_MPIDense *a;
1097: mat->preallocated = PETSC_TRUE;
1098: /* Note: For now, when data is specified above, this assumes the user correctly
1099: allocates the local dense storage space. We should add error checking. */
1101: a = (Mat_MPIDense*)mat->data;
1102: MatCreate(PETSC_COMM_SELF,&a->A);
1103: MatSetSizes(a->A,mat->rmap.n,mat->cmap.N,mat->rmap.n,mat->cmap.N);
1104: MatSetType(a->A,MATSEQDENSE);
1105: MatSeqDenseSetPreallocation(a->A,data);
1106: PetscLogObjectParent(mat,a->A);
1107: return(0);
1108: }
1111: /*MC
1112: MATMPIDENSE - MATMPIDENSE = "mpidense" - A matrix type to be used for distributed dense matrices.
1114: Options Database Keys:
1115: . -mat_type mpidense - sets the matrix type to "mpidense" during a call to MatSetFromOptions()
1117: Level: beginner
1119: .seealso: MatCreateMPIDense
1120: M*/
1125: PetscErrorCode MatCreate_MPIDense(Mat mat)
1126: {
1127: Mat_MPIDense *a;
1131: PetscNew(Mat_MPIDense,&a);
1132: mat->data = (void*)a;
1133: PetscMemcpy(mat->ops,&MatOps_Values,sizeof(struct _MatOps));
1134: mat->factor = 0;
1135: mat->mapping = 0;
1137: a->factor = 0;
1138: mat->insertmode = NOT_SET_VALUES;
1139: MPI_Comm_rank(mat->comm,&a->rank);
1140: MPI_Comm_size(mat->comm,&a->size);
1142: mat->rmap.bs = mat->cmap.bs = 1;
1143: PetscMapInitialize(mat->comm,&mat->rmap);
1144: PetscMapInitialize(mat->comm,&mat->cmap);
1145: a->nvec = mat->cmap.n;
1147: /* build cache for off array entries formed */
1148: a->donotstash = PETSC_FALSE;
1149: MatStashCreate_Private(mat->comm,1,&mat->stash);
1151: /* stuff used for matrix vector multiply */
1152: a->lvec = 0;
1153: a->Mvctx = 0;
1154: a->roworiented = PETSC_TRUE;
1156: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatGetDiagonalBlock_C",
1157: "MatGetDiagonalBlock_MPIDense",
1158: MatGetDiagonalBlock_MPIDense);
1159: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatMPIDenseSetPreallocation_C",
1160: "MatMPIDenseSetPreallocation_MPIDense",
1161: MatMPIDenseSetPreallocation_MPIDense);
1162: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatMatMult_mpiaij_mpidense_C",
1163: "MatMatMult_MPIAIJ_MPIDense",
1164: MatMatMult_MPIAIJ_MPIDense);
1165: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatMatMultSymbolic_mpiaij_mpidense_C",
1166: "MatMatMultSymbolic_MPIAIJ_MPIDense",
1167: MatMatMultSymbolic_MPIAIJ_MPIDense);
1168: PetscObjectComposeFunctionDynamic((PetscObject)mat,"MatMatMultNumeric_mpiaij_mpidense_C",
1169: "MatMatMultNumeric_MPIAIJ_MPIDense",
1170: MatMatMultNumeric_MPIAIJ_MPIDense);
1171: PetscObjectChangeTypeName((PetscObject)mat,MATMPIDENSE);
1172: return(0);
1173: }
1176: /*MC
1177: MATDENSE - MATDENSE = "dense" - A matrix type to be used for dense matrices.
1179: This matrix type is identical to MATSEQDENSE when constructed with a single process communicator,
1180: and MATMPIDENSE otherwise.
1182: Options Database Keys:
1183: . -mat_type dense - sets the matrix type to "dense" during a call to MatSetFromOptions()
1185: Level: beginner
1187: .seealso: MatCreateMPIDense,MATSEQDENSE,MATMPIDENSE
1188: M*/
1193: PetscErrorCode MatCreate_Dense(Mat A)
1194: {
1196: PetscMPIInt size;
1199: MPI_Comm_size(A->comm,&size);
1200: if (size == 1) {
1201: MatSetType(A,MATSEQDENSE);
1202: } else {
1203: MatSetType(A,MATMPIDENSE);
1204: }
1205: return(0);
1206: }
1211: /*@C
1212: MatMPIDenseSetPreallocation - Sets the array used to store the matrix entries
1214: Not collective
1216: Input Parameters:
1217: . A - the matrix
1218: - data - optional location of matrix data. Set data=PETSC_NULL for PETSc
1219: to control all matrix memory allocation.
1221: Notes:
1222: The dense format is fully compatible with standard Fortran 77
1223: storage by columns.
1225: The data input variable is intended primarily for Fortran programmers
1226: who wish to allocate their own matrix memory space. Most users should
1227: set data=PETSC_NULL.
1229: Level: intermediate
1231: .keywords: matrix,dense, parallel
1233: .seealso: MatCreate(), MatCreateSeqDense(), MatSetValues()
1234: @*/
1235: PetscErrorCode MatMPIDenseSetPreallocation(Mat mat,PetscScalar *data)
1236: {
1237: PetscErrorCode ierr,(*f)(Mat,PetscScalar *);
1240: PetscObjectQueryFunction((PetscObject)mat,"MatMPIDenseSetPreallocation_C",(void (**)(void))&f);
1241: if (f) {
1242: (*f)(mat,data);
1243: }
1244: return(0);
1245: }
1249: /*@C
1250: MatCreateMPIDense - Creates a sparse parallel matrix in dense format.
1252: Collective on MPI_Comm
1254: Input Parameters:
1255: + comm - MPI communicator
1256: . m - number of local rows (or PETSC_DECIDE to have calculated if M is given)
1257: . n - number of local columns (or PETSC_DECIDE to have calculated if N is given)
1258: . M - number of global rows (or PETSC_DECIDE to have calculated if m is given)
1259: . N - number of global columns (or PETSC_DECIDE to have calculated if n is given)
1260: - data - optional location of matrix data. Set data=PETSC_NULL (PETSC_NULL_SCALAR for Fortran users) for PETSc
1261: to control all matrix memory allocation.
1263: Output Parameter:
1264: . A - the matrix
1266: Notes:
1267: The dense format is fully compatible with standard Fortran 77
1268: storage by columns.
1270: The data input variable is intended primarily for Fortran programmers
1271: who wish to allocate their own matrix memory space. Most users should
1272: set data=PETSC_NULL (PETSC_NULL_SCALAR for Fortran users).
1274: The user MUST specify either the local or global matrix dimensions
1275: (possibly both).
1277: Level: intermediate
1279: .keywords: matrix,dense, parallel
1281: .seealso: MatCreate(), MatCreateSeqDense(), MatSetValues()
1282: @*/
1283: PetscErrorCode MatCreateMPIDense(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt M,PetscInt N,PetscScalar *data,Mat *A)
1284: {
1286: PetscMPIInt size;
1289: MatCreate(comm,A);
1290: MatSetSizes(*A,m,n,M,N);
1291: MPI_Comm_size(comm,&size);
1292: if (size > 1) {
1293: MatSetType(*A,MATMPIDENSE);
1294: MatMPIDenseSetPreallocation(*A,data);
1295: } else {
1296: MatSetType(*A,MATSEQDENSE);
1297: MatSeqDenseSetPreallocation(*A,data);
1298: }
1299: return(0);
1300: }
1304: static PetscErrorCode MatDuplicate_MPIDense(Mat A,MatDuplicateOption cpvalues,Mat *newmat)
1305: {
1306: Mat mat;
1307: Mat_MPIDense *a,*oldmat = (Mat_MPIDense*)A->data;
1311: *newmat = 0;
1312: MatCreate(A->comm,&mat);
1313: MatSetSizes(mat,A->rmap.n,A->cmap.n,A->rmap.N,A->cmap.N);
1314: MatSetType(mat,A->type_name);
1315: a = (Mat_MPIDense*)mat->data;
1316: PetscMemcpy(mat->ops,A->ops,sizeof(struct _MatOps));
1317: mat->factor = A->factor;
1318: mat->assembled = PETSC_TRUE;
1319: mat->preallocated = PETSC_TRUE;
1321: mat->rmap.rstart = A->rmap.rstart;
1322: mat->rmap.rend = A->rmap.rend;
1323: a->size = oldmat->size;
1324: a->rank = oldmat->rank;
1325: mat->insertmode = NOT_SET_VALUES;
1326: a->nvec = oldmat->nvec;
1327: a->donotstash = oldmat->donotstash;
1328:
1329: PetscMemcpy(mat->rmap.range,A->rmap.range,(a->size+1)*sizeof(PetscInt));
1330: PetscMemcpy(mat->cmap.range,A->cmap.range,(a->size+1)*sizeof(PetscInt));
1331: MatStashCreate_Private(A->comm,1,&mat->stash);
1333: MatSetUpMultiply_MPIDense(mat);
1334: MatDuplicate(oldmat->A,cpvalues,&a->A);
1335: PetscLogObjectParent(mat,a->A);
1336: *newmat = mat;
1337: return(0);
1338: }
1340: #include petscsys.h
1344: PetscErrorCode MatLoad_MPIDense_DenseInFile(MPI_Comm comm,PetscInt fd,PetscInt M,PetscInt N, MatType type,Mat *newmat)
1345: {
1347: PetscMPIInt rank,size;
1348: PetscInt *rowners,i,m,nz,j;
1349: PetscScalar *array,*vals,*vals_ptr;
1350: MPI_Status status;
1353: MPI_Comm_rank(comm,&rank);
1354: MPI_Comm_size(comm,&size);
1356: /* determine ownership of all rows */
1357: m = M/size + ((M % size) > rank);
1358: PetscMalloc((size+2)*sizeof(PetscInt),&rowners);
1359: MPI_Allgather(&m,1,MPIU_INT,rowners+1,1,MPIU_INT,comm);
1360: rowners[0] = 0;
1361: for (i=2; i<=size; i++) {
1362: rowners[i] += rowners[i-1];
1363: }
1365: MatCreate(comm,newmat);
1366: MatSetSizes(*newmat,m,PETSC_DECIDE,M,N);
1367: MatSetType(*newmat,type);
1368: MatMPIDenseSetPreallocation(*newmat,PETSC_NULL);
1369: MatGetArray(*newmat,&array);
1371: if (!rank) {
1372: PetscMalloc(m*N*sizeof(PetscScalar),&vals);
1374: /* read in my part of the matrix numerical values */
1375: PetscBinaryRead(fd,vals,m*N,PETSC_SCALAR);
1376:
1377: /* insert into matrix-by row (this is why cannot directly read into array */
1378: vals_ptr = vals;
1379: for (i=0; i<m; i++) {
1380: for (j=0; j<N; j++) {
1381: array[i + j*m] = *vals_ptr++;
1382: }
1383: }
1385: /* read in other processors and ship out */
1386: for (i=1; i<size; i++) {
1387: nz = (rowners[i+1] - rowners[i])*N;
1388: PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);
1389: MPI_Send(vals,nz,MPIU_SCALAR,i,(*newmat)->tag,comm);
1390: }
1391: } else {
1392: /* receive numeric values */
1393: PetscMalloc(m*N*sizeof(PetscScalar),&vals);
1395: /* receive message of values*/
1396: MPI_Recv(vals,m*N,MPIU_SCALAR,0,(*newmat)->tag,comm,&status);
1398: /* insert into matrix-by row (this is why cannot directly read into array */
1399: vals_ptr = vals;
1400: for (i=0; i<m; i++) {
1401: for (j=0; j<N; j++) {
1402: array[i + j*m] = *vals_ptr++;
1403: }
1404: }
1405: }
1406: PetscFree(rowners);
1407: PetscFree(vals);
1408: MatAssemblyBegin(*newmat,MAT_FINAL_ASSEMBLY);
1409: MatAssemblyEnd(*newmat,MAT_FINAL_ASSEMBLY);
1410: return(0);
1411: }
1415: PetscErrorCode MatLoad_MPIDense(PetscViewer viewer, MatType type,Mat *newmat)
1416: {
1417: Mat A;
1418: PetscScalar *vals,*svals;
1419: MPI_Comm comm = ((PetscObject)viewer)->comm;
1420: MPI_Status status;
1421: PetscMPIInt rank,size,tag = ((PetscObject)viewer)->tag,*rowners,*sndcounts,m,maxnz;
1422: PetscInt header[4],*rowlengths = 0,M,N,*cols;
1423: PetscInt *ourlens,*procsnz = 0,*offlens,jj,*mycols,*smycols;
1424: PetscInt i,nz,j,rstart,rend;
1425: int fd;
1429: MPI_Comm_size(comm,&size);
1430: MPI_Comm_rank(comm,&rank);
1431: if (!rank) {
1432: PetscViewerBinaryGetDescriptor(viewer,&fd);
1433: PetscBinaryRead(fd,(char *)header,4,PETSC_INT);
1434: if (header[0] != MAT_FILE_COOKIE) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"not matrix object");
1435: }
1437: MPI_Bcast(header+1,3,MPIU_INT,0,comm);
1438: M = header[1]; N = header[2]; nz = header[3];
1440: /*
1441: Handle case where matrix is stored on disk as a dense matrix
1442: */
1443: if (nz == MATRIX_BINARY_FORMAT_DENSE) {
1444: MatLoad_MPIDense_DenseInFile(comm,fd,M,N,type,newmat);
1445: return(0);
1446: }
1448: /* determine ownership of all rows */
1449: m = M/size + ((M % size) > rank);
1450: PetscMalloc((size+2)*sizeof(PetscInt),&rowners);
1451: MPI_Allgather(&m,1,MPI_INT,rowners+1,1,MPI_INT,comm);
1452: rowners[0] = 0;
1453: for (i=2; i<=size; i++) {
1454: rowners[i] += rowners[i-1];
1455: }
1456: rstart = rowners[rank];
1457: rend = rowners[rank+1];
1459: /* distribute row lengths to all processors */
1460: PetscMalloc(2*(rend-rstart+1)*sizeof(PetscInt),&ourlens);
1461: offlens = ourlens + (rend-rstart);
1462: if (!rank) {
1463: PetscMalloc(M*sizeof(PetscInt),&rowlengths);
1464: PetscBinaryRead(fd,rowlengths,M,PETSC_INT);
1465: PetscMalloc(size*sizeof(PetscMPIInt),&sndcounts);
1466: for (i=0; i<size; i++) sndcounts[i] = rowners[i+1] - rowners[i];
1467: MPI_Scatterv(rowlengths,sndcounts,rowners,MPIU_INT,ourlens,rend-rstart,MPIU_INT,0,comm);
1468: PetscFree(sndcounts);
1469: } else {
1470: MPI_Scatterv(0,0,0,MPIU_INT,ourlens,rend-rstart,MPIU_INT,0,comm);
1471: }
1473: if (!rank) {
1474: /* calculate the number of nonzeros on each processor */
1475: PetscMalloc(size*sizeof(PetscInt),&procsnz);
1476: PetscMemzero(procsnz,size*sizeof(PetscInt));
1477: for (i=0; i<size; i++) {
1478: for (j=rowners[i]; j< rowners[i+1]; j++) {
1479: procsnz[i] += rowlengths[j];
1480: }
1481: }
1482: PetscFree(rowlengths);
1484: /* determine max buffer needed and allocate it */
1485: maxnz = 0;
1486: for (i=0; i<size; i++) {
1487: maxnz = PetscMax(maxnz,procsnz[i]);
1488: }
1489: PetscMalloc(maxnz*sizeof(PetscInt),&cols);
1491: /* read in my part of the matrix column indices */
1492: nz = procsnz[0];
1493: PetscMalloc(nz*sizeof(PetscInt),&mycols);
1494: PetscBinaryRead(fd,mycols,nz,PETSC_INT);
1496: /* read in every one elses and ship off */
1497: for (i=1; i<size; i++) {
1498: nz = procsnz[i];
1499: PetscBinaryRead(fd,cols,nz,PETSC_INT);
1500: MPI_Send(cols,nz,MPIU_INT,i,tag,comm);
1501: }
1502: PetscFree(cols);
1503: } else {
1504: /* determine buffer space needed for message */
1505: nz = 0;
1506: for (i=0; i<m; i++) {
1507: nz += ourlens[i];
1508: }
1509: PetscMalloc((nz+1)*sizeof(PetscInt),&mycols);
1511: /* receive message of column indices*/
1512: MPI_Recv(mycols,nz,MPIU_INT,0,tag,comm,&status);
1513: MPI_Get_count(&status,MPIU_INT,&maxnz);
1514: if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file");
1515: }
1517: /* loop over local rows, determining number of off diagonal entries */
1518: PetscMemzero(offlens,m*sizeof(PetscInt));
1519: jj = 0;
1520: for (i=0; i<m; i++) {
1521: for (j=0; j<ourlens[i]; j++) {
1522: if (mycols[jj] < rstart || mycols[jj] >= rend) offlens[i]++;
1523: jj++;
1524: }
1525: }
1527: /* create our matrix */
1528: for (i=0; i<m; i++) {
1529: ourlens[i] -= offlens[i];
1530: }
1531: MatCreate(comm,newmat);
1532: MatSetSizes(*newmat,m,PETSC_DECIDE,M,N);
1533: MatSetType(*newmat,type);
1534: MatMPIDenseSetPreallocation(*newmat,PETSC_NULL);
1535: A = *newmat;
1536: for (i=0; i<m; i++) {
1537: ourlens[i] += offlens[i];
1538: }
1540: if (!rank) {
1541: PetscMalloc(maxnz*sizeof(PetscScalar),&vals);
1543: /* read in my part of the matrix numerical values */
1544: nz = procsnz[0];
1545: PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);
1546:
1547: /* insert into matrix */
1548: jj = rstart;
1549: smycols = mycols;
1550: svals = vals;
1551: for (i=0; i<m; i++) {
1552: MatSetValues(A,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);
1553: smycols += ourlens[i];
1554: svals += ourlens[i];
1555: jj++;
1556: }
1558: /* read in other processors and ship out */
1559: for (i=1; i<size; i++) {
1560: nz = procsnz[i];
1561: PetscBinaryRead(fd,vals,nz,PETSC_SCALAR);
1562: MPI_Send(vals,nz,MPIU_SCALAR,i,A->tag,comm);
1563: }
1564: PetscFree(procsnz);
1565: } else {
1566: /* receive numeric values */
1567: PetscMalloc((nz+1)*sizeof(PetscScalar),&vals);
1569: /* receive message of values*/
1570: MPI_Recv(vals,nz,MPIU_SCALAR,0,A->tag,comm,&status);
1571: MPI_Get_count(&status,MPIU_SCALAR,&maxnz);
1572: if (maxnz != nz) SETERRQ(PETSC_ERR_FILE_UNEXPECTED,"something is wrong with file");
1574: /* insert into matrix */
1575: jj = rstart;
1576: smycols = mycols;
1577: svals = vals;
1578: for (i=0; i<m; i++) {
1579: MatSetValues(A,1,&jj,ourlens[i],smycols,svals,INSERT_VALUES);
1580: smycols += ourlens[i];
1581: svals += ourlens[i];
1582: jj++;
1583: }
1584: }
1585: PetscFree(ourlens);
1586: PetscFree(vals);
1587: PetscFree(mycols);
1588: PetscFree(rowners);
1590: MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);
1591: MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);
1592: return(0);
1593: }
1597: PetscErrorCode MatEqual_MPIDense(Mat A,Mat B,PetscTruth *flag)
1598: {
1599: Mat_MPIDense *matB = (Mat_MPIDense*)B->data,*matA = (Mat_MPIDense*)A->data;
1600: Mat a,b;
1601: PetscTruth flg;
1605: a = matA->A;
1606: b = matB->A;
1607: MatEqual(a,b,&flg);
1608: return(0);
1609: }