Actual source code: blockmat.c
1: #define PETSCMAT_DLL
3: /*
4: This provides a matrix that consists of Mats
5: */
7: #include private/matimpl.h
8: #include "../src/mat/impls/baij/seq/baij.h" /* use the common AIJ data-structure */
9: #include petscksp.h
11: #define CHUNKSIZE 15
13: typedef struct {
14: SEQAIJHEADER(Mat);
15: SEQBAIJHEADER;
16: Mat *diags;
18: Vec left,right,middle,workb; /* dummy vectors to perform local parts of product */
19: } Mat_BlockMat;
23: PetscErrorCode MatSOR_BlockMat_Symmetric(Mat A,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx)
24: {
25: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
26: PetscScalar *x;
27: const Mat *v = a->a;
28: const PetscScalar *b;
29: PetscErrorCode ierr;
30: PetscInt n = A->cmap->n,i,mbs = n/A->rmap->bs,j,bs = A->rmap->bs;
31: const PetscInt *idx;
32: IS row,col;
33: MatFactorInfo info;
34: Vec left = a->left,right = a->right, middle = a->middle;
35: Mat *diag;
38: its = its*lits;
39: if (its <= 0) SETERRQ2(PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D and local its %D both positive",its,lits);
40: if (flag & SOR_EISENSTAT) SETERRQ(PETSC_ERR_SUP,"No support yet for Eisenstat");
41: if (omega != 1.0) SETERRQ(PETSC_ERR_SUP,"No support yet for omega not equal to 1.0");
42: if (fshift) SETERRQ(PETSC_ERR_SUP,"No support yet for fshift");
43: if ((flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP) && !(flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP))
44: SETERRQ(PETSC_ERR_SUP,"Cannot do backward sweep without forward sweep");
46: if (!a->diags) {
47: PetscMalloc(mbs*sizeof(Mat),&a->diags);
48: MatFactorInfoInitialize(&info);
49: for (i=0; i<mbs; i++) {
50: MatGetOrdering(a->a[a->diag[i]], MATORDERING_ND,&row,&col);
51: MatCholeskyFactorSymbolic(a->diags[i],a->a[a->diag[i]],row,&info);
52: MatCholeskyFactorNumeric(a->diags[i],a->a[a->diag[i]],&info);
53: ISDestroy(row);
54: ISDestroy(col);
55: }
56: VecDuplicate(bb,&a->workb);
57: }
58: diag = a->diags;
60: VecSet(xx,0.0);
61: VecGetArray(xx,&x);
62: /* copy right hand side because it must be modified during iteration */
63: VecCopy(bb,a->workb);
64: VecGetArray(a->workb,(PetscScalar**)&b);
66: /* need to add code for when initial guess is zero, see MatSOR_SeqAIJ */
67: while (its--) {
68: if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP){
70: for (i=0; i<mbs; i++) {
71: n = a->i[i+1] - a->i[i] - 1;
72: idx = a->j + a->i[i] + 1;
73: v = a->a + a->i[i] + 1;
75: VecSet(left,0.0);
76: for (j=0; j<n; j++) {
77: VecPlaceArray(right,x + idx[j]*bs);
78: MatMultAdd(v[j],right,left,left);
79: VecResetArray(right);
80: }
81: VecPlaceArray(right,b + i*bs);
82: VecAYPX(left,-1.0,right);
83: VecResetArray(right);
85: VecPlaceArray(right,x + i*bs);
86: MatSolve(diag[i],left,right);
88: /* now adjust right hand side, see MatSOR_SeqSBAIJ */
89: for (j=0; j<n; j++) {
90: MatMultTranspose(v[j],right,left);
91: VecPlaceArray(middle,b + idx[j]*bs);
92: VecAXPY(middle,-1.0,left);
93: VecResetArray(middle);
94: }
95: VecResetArray(right);
97: }
98: }
99: if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP){
101: for (i=mbs-1; i>=0; i--) {
102: n = a->i[i+1] - a->i[i] - 1;
103: idx = a->j + a->i[i] + 1;
104: v = a->a + a->i[i] + 1;
106: VecSet(left,0.0);
107: for (j=0; j<n; j++) {
108: VecPlaceArray(right,x + idx[j]*bs);
109: MatMultAdd(v[j],right,left,left);
110: VecResetArray(right);
111: }
112: VecPlaceArray(right,b + i*bs);
113: VecAYPX(left,-1.0,right);
114: VecResetArray(right);
116: VecPlaceArray(right,x + i*bs);
117: MatSolve(diag[i],left,right);
118: VecResetArray(right);
120: }
121: }
122: }
123: VecRestoreArray(xx,&x);
124: VecRestoreArray(a->workb,(PetscScalar**)&b);
125: return(0);
126: }
130: PetscErrorCode MatSOR_BlockMat(Mat A,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx)
131: {
132: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
133: PetscScalar *x;
134: const Mat *v = a->a;
135: const PetscScalar *b;
136: PetscErrorCode ierr;
137: PetscInt n = A->cmap->n,i,mbs = n/A->rmap->bs,j,bs = A->rmap->bs;
138: const PetscInt *idx;
139: IS row,col;
140: MatFactorInfo info;
141: Vec left = a->left,right = a->right;
142: Mat *diag;
145: its = its*lits;
146: if (its <= 0) SETERRQ2(PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D and local its %D both positive",its,lits);
147: if (flag & SOR_EISENSTAT) SETERRQ(PETSC_ERR_SUP,"No support yet for Eisenstat");
148: if (omega != 1.0) SETERRQ(PETSC_ERR_SUP,"No support yet for omega not equal to 1.0");
149: if (fshift) SETERRQ(PETSC_ERR_SUP,"No support yet for fshift");
151: if (!a->diags) {
152: PetscMalloc(mbs*sizeof(Mat),&a->diags);
153: MatFactorInfoInitialize(&info);
154: for (i=0; i<mbs; i++) {
155: MatGetOrdering(a->a[a->diag[i]], MATORDERING_ND,&row,&col);
156: MatLUFactorSymbolic(a->diags[i],a->a[a->diag[i]],row,col,&info);
157: MatLUFactorNumeric(a->diags[i],a->a[a->diag[i]],&info);
158: ISDestroy(row);
159: ISDestroy(col);
160: }
161: }
162: diag = a->diags;
164: VecSet(xx,0.0);
165: VecGetArray(xx,&x);
166: VecGetArray(bb,(PetscScalar**)&b);
168: /* need to add code for when initial guess is zero, see MatSOR_SeqAIJ */
169: while (its--) {
170: if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP){
172: for (i=0; i<mbs; i++) {
173: n = a->i[i+1] - a->i[i];
174: idx = a->j + a->i[i];
175: v = a->a + a->i[i];
177: VecSet(left,0.0);
178: for (j=0; j<n; j++) {
179: if (idx[j] != i) {
180: VecPlaceArray(right,x + idx[j]*bs);
181: MatMultAdd(v[j],right,left,left);
182: VecResetArray(right);
183: }
184: }
185: VecPlaceArray(right,b + i*bs);
186: VecAYPX(left,-1.0,right);
187: VecResetArray(right);
189: VecPlaceArray(right,x + i*bs);
190: MatSolve(diag[i],left,right);
191: VecResetArray(right);
192: }
193: }
194: if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP){
196: for (i=mbs-1; i>=0; i--) {
197: n = a->i[i+1] - a->i[i];
198: idx = a->j + a->i[i];
199: v = a->a + a->i[i];
201: VecSet(left,0.0);
202: for (j=0; j<n; j++) {
203: if (idx[j] != i) {
204: VecPlaceArray(right,x + idx[j]*bs);
205: MatMultAdd(v[j],right,left,left);
206: VecResetArray(right);
207: }
208: }
209: VecPlaceArray(right,b + i*bs);
210: VecAYPX(left,-1.0,right);
211: VecResetArray(right);
213: VecPlaceArray(right,x + i*bs);
214: MatSolve(diag[i],left,right);
215: VecResetArray(right);
217: }
218: }
219: }
220: VecRestoreArray(xx,&x);
221: VecRestoreArray(bb,(PetscScalar**)&b);
222: return(0);
223: }
227: PetscErrorCode MatSetValues_BlockMat(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode is)
228: {
229: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
230: PetscInt *rp,k,low,high,t,ii,row,nrow,i,col,l,rmax,N,lastcol = -1;
231: PetscInt *imax=a->imax,*ai=a->i,*ailen=a->ilen;
232: PetscInt *aj=a->j,nonew=a->nonew,bs=A->rmap->bs,brow,bcol;
234: PetscInt ridx,cidx;
235: PetscTruth roworiented=a->roworiented;
236: MatScalar value;
237: Mat *ap,*aa = a->a;
241: for (k=0; k<m; k++) { /* loop over added rows */
242: row = im[k];
243: brow = row/bs;
244: if (row < 0) continue;
245: #if defined(PETSC_USE_DEBUG)
246: if (row >= A->rmap->N) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",row,A->rmap->N-1);
247: #endif
248: rp = aj + ai[brow];
249: ap = aa + ai[brow];
250: rmax = imax[brow];
251: nrow = ailen[brow];
252: low = 0;
253: high = nrow;
254: for (l=0; l<n; l++) { /* loop over added columns */
255: if (in[l] < 0) continue;
256: #if defined(PETSC_USE_DEBUG)
257: if (in[l] >= A->cmap->n) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[l],A->cmap->n-1);
258: #endif
259: col = in[l]; bcol = col/bs;
260: if (A->symmetric && brow > bcol) continue;
261: ridx = row % bs; cidx = col % bs;
262: if (roworiented) {
263: value = v[l + k*n];
264: } else {
265: value = v[k + l*m];
266: }
267: if (col <= lastcol) low = 0; else high = nrow;
268: lastcol = col;
269: while (high-low > 7) {
270: t = (low+high)/2;
271: if (rp[t] > bcol) high = t;
272: else low = t;
273: }
274: for (i=low; i<high; i++) {
275: if (rp[i] > bcol) break;
276: if (rp[i] == bcol) {
277: goto noinsert1;
278: }
279: }
280: if (nonew == 1) goto noinsert1;
281: if (nonew == -1) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) in the matrix", row, col);
282: MatSeqXAIJReallocateAIJ(A,a->mbs,1,nrow,brow,bcol,rmax,aa,ai,aj,rp,ap,imax,nonew,Mat);
283: N = nrow++ - 1; high++;
284: /* shift up all the later entries in this row */
285: for (ii=N; ii>=i; ii--) {
286: rp[ii+1] = rp[ii];
287: ap[ii+1] = ap[ii];
288: }
289: if (N>=i) ap[i] = 0;
290: rp[i] = bcol;
291: a->nz++;
292: noinsert1:;
293: if (!*(ap+i)) {
294: MatCreateSeqAIJ(PETSC_COMM_SELF,bs,bs,0,0,ap+i);
295: }
296: MatSetValues(ap[i],1,&ridx,1,&cidx,&value,is);
297: low = i;
298: }
299: ailen[brow] = nrow;
300: }
301: A->same_nonzero = PETSC_FALSE;
302: return(0);
303: }
307: PetscErrorCode MatLoad_BlockMat(PetscViewer viewer, const MatType type,Mat *A)
308: {
309: PetscErrorCode ierr;
310: Mat tmpA;
311: PetscInt i,j,m,n,bs = 1,ncols,*lens,currentcol,mbs,**ii,*ilens,nextcol,*llens,cnt = 0;
312: const PetscInt *cols;
313: const PetscScalar *values;
314: PetscTruth flg = PETSC_FALSE,notdone;
315: Mat_SeqAIJ *a;
316: Mat_BlockMat *amat;
319: MatLoad_SeqAIJ(viewer,MATSEQAIJ,&tmpA);
321: MatGetLocalSize(tmpA,&m,&n);
322: PetscOptionsBegin(PETSC_COMM_SELF,PETSC_NULL,"Options for loading BlockMat matrix 1","Mat");
323: PetscOptionsInt("-matload_block_size","Set the blocksize used to store the matrix","MatLoad",bs,&bs,PETSC_NULL);
324: PetscOptionsTruth("-matload_symmetric","Store the matrix as symmetric","MatLoad",flg,&flg,PETSC_NULL);
325: PetscOptionsEnd();
327: /* Determine number of nonzero blocks for each block row */
328: a = (Mat_SeqAIJ*) tmpA->data;
329: mbs = m/bs;
330: PetscMalloc3(mbs,PetscInt,&lens,bs,PetscInt*,&ii,bs,PetscInt,&ilens);
331: PetscMemzero(lens,mbs*sizeof(PetscInt));
333: for (i=0; i<mbs; i++) {
334: for (j=0; j<bs; j++) {
335: ii[j] = a->j + a->i[i*bs + j];
336: ilens[j] = a->i[i*bs + j + 1] - a->i[i*bs + j];
337: }
339: currentcol = -1;
340: notdone = PETSC_TRUE;
341: while (PETSC_TRUE) {
342: notdone = PETSC_FALSE;
343: nextcol = 1000000000;
344: for (j=0; j<bs; j++) {
345: while ((ilens[j] > 0 && ii[j][0]/bs <= currentcol)) {
346: ii[j]++;
347: ilens[j]--;
348: }
349: if (ilens[j] > 0) {
350: notdone = PETSC_TRUE;
351: nextcol = PetscMin(nextcol,ii[j][0]/bs);
352: }
353: }
354: if (!notdone) break;
355: if (!flg || (nextcol >= i)) lens[i]++;
356: currentcol = nextcol;
357: }
358: }
360: MatCreateBlockMat(PETSC_COMM_SELF,m,n,bs,0,lens,A);
361: if (flg) {
362: MatSetOption(*A,MAT_SYMMETRIC,PETSC_TRUE);
363: }
364: amat = (Mat_BlockMat*)(*A)->data;
366: /* preallocate the submatrices */
367: PetscMalloc(bs*sizeof(PetscInt),&llens);
368: for (i=0; i<mbs; i++) { /* loops for block rows */
369: for (j=0; j<bs; j++) {
370: ii[j] = a->j + a->i[i*bs + j];
371: ilens[j] = a->i[i*bs + j + 1] - a->i[i*bs + j];
372: }
374: currentcol = 1000000000;
375: for (j=0; j<bs; j++) { /* loop over rows in block finding first nonzero block */
376: if (ilens[j] > 0) {
377: currentcol = PetscMin(currentcol,ii[j][0]/bs);
378: }
379: }
381: notdone = PETSC_TRUE;
382: while (PETSC_TRUE) { /* loops over blocks in block row */
384: notdone = PETSC_FALSE;
385: nextcol = 1000000000;
386: PetscMemzero(llens,bs*sizeof(PetscInt));
387: for (j=0; j<bs; j++) { /* loop over rows in block */
388: while ((ilens[j] > 0 && ii[j][0]/bs <= currentcol)) { /* loop over columns in row */
389: ii[j]++;
390: ilens[j]--;
391: llens[j]++;
392: }
393: if (ilens[j] > 0) {
394: notdone = PETSC_TRUE;
395: nextcol = PetscMin(nextcol,ii[j][0]/bs);
396: }
397: }
398: if (cnt >= amat->maxnz) SETERRQ1(PETSC_ERR_PLIB,"Number of blocks found greater than expected %D",cnt);
399: if (!flg || currentcol >= i) {
400: amat->j[cnt] = currentcol;
401: MatCreateSeqAIJ(PETSC_COMM_SELF,bs,bs,0,llens,amat->a+cnt++);
402: }
404: if (!notdone) break;
405: currentcol = nextcol;
406: }
407: amat->ilen[i] = lens[i];
408: }
409: CHKMEMQ;
411: PetscFree3(lens,ii,ilens);
412: PetscFree(llens);
414: /* copy over the matrix, one row at a time */
415: for (i=0; i<m; i++) {
416: MatGetRow(tmpA,i,&ncols,&cols,&values);
417: MatSetValues(*A,1,&i,ncols,cols,values,INSERT_VALUES);
418: MatRestoreRow(tmpA,i,&ncols,&cols,&values);
419: }
420: MatAssemblyBegin(*A,MAT_FINAL_ASSEMBLY);
421: MatAssemblyEnd(*A,MAT_FINAL_ASSEMBLY);
422: return(0);
423: }
427: PetscErrorCode MatView_BlockMat(Mat A,PetscViewer viewer)
428: {
429: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
430: PetscErrorCode ierr;
431: const char *name;
432: PetscViewerFormat format;
435: PetscObjectGetName((PetscObject)A,&name);
436: PetscViewerGetFormat(viewer,&format);
437: if (format == PETSC_VIEWER_ASCII_FACTOR_INFO || format == PETSC_VIEWER_ASCII_INFO) {
438: PetscViewerASCIIPrintf(viewer,"Nonzero block matrices = %D \n",a->nz);
439: if (A->symmetric) {
440: PetscViewerASCIIPrintf(viewer,"Only upper triangular part of symmetric matrix is stored\n");
441: }
442: }
443: return(0);
444: }
448: PetscErrorCode MatDestroy_BlockMat(Mat mat)
449: {
451: Mat_BlockMat *bmat = (Mat_BlockMat*)mat->data;
452: PetscInt i;
455: if (bmat->right) {
456: VecDestroy(bmat->right);
457: }
458: if (bmat->left) {
459: VecDestroy(bmat->left);
460: }
461: if (bmat->middle) {
462: VecDestroy(bmat->middle);
463: }
464: if (bmat->workb) {
465: VecDestroy(bmat->workb);
466: }
467: if (bmat->diags) {
468: for (i=0; i<mat->rmap->n/mat->rmap->bs; i++) {
469: if (bmat->diags[i]) {MatDestroy(bmat->diags[i]);}
470: }
471: }
472: if (bmat->a) {
473: for (i=0; i<bmat->nz; i++) {
474: if (bmat->a[i]) {MatDestroy(bmat->a[i]);}
475: }
476: }
477: MatSeqXAIJFreeAIJ(mat,(PetscScalar**)&bmat->a,&bmat->j,&bmat->i);
478: PetscFree(bmat);
479: return(0);
480: }
484: PetscErrorCode MatMult_BlockMat(Mat A,Vec x,Vec y)
485: {
486: Mat_BlockMat *bmat = (Mat_BlockMat*)A->data;
488: PetscScalar *xx,*yy;
489: PetscInt *aj,i,*ii,jrow,m = A->rmap->n/A->rmap->bs,bs = A->rmap->bs,n,j;
490: Mat *aa;
493: CHKMEMQ;
494: /*
495: Standard CSR multiply except each entry is a Mat
496: */
497: VecGetArray(x,&xx);
499: VecSet(y,0.0);
500: VecGetArray(y,&yy);
501: aj = bmat->j;
502: aa = bmat->a;
503: ii = bmat->i;
504: for (i=0; i<m; i++) {
505: jrow = ii[i];
506: VecPlaceArray(bmat->left,yy + bs*i);
507: n = ii[i+1] - jrow;
508: for (j=0; j<n; j++) {
509: VecPlaceArray(bmat->right,xx + bs*aj[jrow]);
510: MatMultAdd(aa[jrow],bmat->right,bmat->left,bmat->left);
511: VecResetArray(bmat->right);
512: jrow++;
513: }
514: VecResetArray(bmat->left);
515: }
516: VecRestoreArray(x,&xx);
517: VecRestoreArray(y,&yy);
518: CHKMEMQ;
519: return(0);
520: }
524: PetscErrorCode MatMult_BlockMat_Symmetric(Mat A,Vec x,Vec y)
525: {
526: Mat_BlockMat *bmat = (Mat_BlockMat*)A->data;
528: PetscScalar *xx,*yy;
529: PetscInt *aj,i,*ii,jrow,m = A->rmap->n/A->rmap->bs,bs = A->rmap->bs,n,j;
530: Mat *aa;
533: CHKMEMQ;
534: /*
535: Standard CSR multiply except each entry is a Mat
536: */
537: VecGetArray(x,&xx);
539: VecSet(y,0.0);
540: VecGetArray(y,&yy);
541: aj = bmat->j;
542: aa = bmat->a;
543: ii = bmat->i;
544: for (i=0; i<m; i++) {
545: jrow = ii[i];
546: n = ii[i+1] - jrow;
547: VecPlaceArray(bmat->left,yy + bs*i);
548: VecPlaceArray(bmat->middle,xx + bs*i);
549: /* if we ALWAYS required a diagonal entry then could remove this if test */
550: if (aj[jrow] == i) {
551: VecPlaceArray(bmat->right,xx + bs*aj[jrow]);
552: MatMultAdd(aa[jrow],bmat->right,bmat->left,bmat->left);
553: VecResetArray(bmat->right);
554: jrow++;
555: n--;
556: }
557: for (j=0; j<n; j++) {
558: VecPlaceArray(bmat->right,xx + bs*aj[jrow]); /* upper triangular part */
559: MatMultAdd(aa[jrow],bmat->right,bmat->left,bmat->left);
560: VecResetArray(bmat->right);
562: VecPlaceArray(bmat->right,yy + bs*aj[jrow]); /* lower triangular part */
563: MatMultTransposeAdd(aa[jrow],bmat->middle,bmat->right,bmat->right);
564: VecResetArray(bmat->right);
565: jrow++;
566: }
567: VecResetArray(bmat->left);
568: VecResetArray(bmat->middle);
569: }
570: VecRestoreArray(x,&xx);
571: VecRestoreArray(y,&yy);
572: CHKMEMQ;
573: return(0);
574: }
578: PetscErrorCode MatMultAdd_BlockMat(Mat A,Vec x,Vec y,Vec z)
579: {
581: return(0);
582: }
586: PetscErrorCode MatMultTranspose_BlockMat(Mat A,Vec x,Vec y)
587: {
589: return(0);
590: }
594: PetscErrorCode MatMultTransposeAdd_BlockMat(Mat A,Vec x,Vec y,Vec z)
595: {
597: return(0);
598: }
600: /*
601: Adds diagonal pointers to sparse matrix structure.
602: */
605: PetscErrorCode MatMarkDiagonal_BlockMat(Mat A)
606: {
607: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
609: PetscInt i,j,mbs = A->rmap->n/A->rmap->bs;
612: if (!a->diag) {
613: PetscMalloc(mbs*sizeof(PetscInt),&a->diag);
614: }
615: for (i=0; i<mbs; i++) {
616: a->diag[i] = a->i[i+1];
617: for (j=a->i[i]; j<a->i[i+1]; j++) {
618: if (a->j[j] == i) {
619: a->diag[i] = j;
620: break;
621: }
622: }
623: }
624: return(0);
625: }
629: PetscErrorCode MatGetSubMatrix_BlockMat(Mat A,IS isrow,IS iscol,MatReuse scall,Mat *B)
630: {
631: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
632: Mat_SeqAIJ *c;
634: PetscInt i,k,first,step,lensi,nrows,ncols;
635: PetscInt *j_new,*i_new,*aj = a->j,*ailen = a->ilen;
636: PetscScalar *a_new;
637: Mat C,*aa = a->a;
638: PetscTruth stride,equal;
641: ISEqual(isrow,iscol,&equal);
642: if (!equal) SETERRQ(PETSC_ERR_SUP,"Only for idential column and row indices");
643: ISStride(iscol,&stride);
644: if (!stride) SETERRQ(PETSC_ERR_SUP,"Only for stride indices");
645: ISStrideGetInfo(iscol,&first,&step);
646: if (step != A->rmap->bs) SETERRQ(PETSC_ERR_SUP,"Can only select one entry from each block");
648: ISGetLocalSize(isrow,&nrows);
649: ncols = nrows;
651: /* create submatrix */
652: if (scall == MAT_REUSE_MATRIX) {
653: PetscInt n_cols,n_rows;
654: C = *B;
655: MatGetSize(C,&n_rows,&n_cols);
656: if (n_rows != nrows || n_cols != ncols) SETERRQ(PETSC_ERR_ARG_SIZ,"Reused submatrix wrong size");
657: MatZeroEntries(C);
658: } else {
659: MatCreate(((PetscObject)A)->comm,&C);
660: MatSetSizes(C,nrows,ncols,PETSC_DETERMINE,PETSC_DETERMINE);
661: if (A->symmetric) {
662: MatSetType(C,MATSEQSBAIJ);
663: } else {
664: MatSetType(C,MATSEQAIJ);
665: }
666: MatSeqAIJSetPreallocation(C,0,ailen);
667: MatSeqSBAIJSetPreallocation(C,1,0,ailen);
668: }
669: c = (Mat_SeqAIJ*)C->data;
670:
671: /* loop over rows inserting into submatrix */
672: a_new = c->a;
673: j_new = c->j;
674: i_new = c->i;
675:
676: for (i=0; i<nrows; i++) {
677: lensi = ailen[i];
678: for (k=0; k<lensi; k++) {
679: *j_new++ = *aj++;
680: MatGetValue(*aa++,first,first,a_new++);
681: }
682: i_new[i+1] = i_new[i] + lensi;
683: c->ilen[i] = lensi;
684: }
686: MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);
687: MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);
688: *B = C;
689: return(0);
690: }
694: PetscErrorCode MatAssemblyEnd_BlockMat(Mat A,MatAssemblyType mode)
695: {
696: Mat_BlockMat *a = (Mat_BlockMat*)A->data;
698: PetscInt fshift = 0,i,j,*ai = a->i,*aj = a->j,*imax = a->imax;
699: PetscInt m = a->mbs,*ip,N,*ailen = a->ilen,rmax = 0;
700: Mat *aa = a->a,*ap;
703: if (mode == MAT_FLUSH_ASSEMBLY) return(0);
705: if (m) rmax = ailen[0]; /* determine row with most nonzeros */
706: for (i=1; i<m; i++) {
707: /* move each row back by the amount of empty slots (fshift) before it*/
708: fshift += imax[i-1] - ailen[i-1];
709: rmax = PetscMax(rmax,ailen[i]);
710: if (fshift) {
711: ip = aj + ai[i] ;
712: ap = aa + ai[i] ;
713: N = ailen[i];
714: for (j=0; j<N; j++) {
715: ip[j-fshift] = ip[j];
716: ap[j-fshift] = ap[j];
717: }
718: }
719: ai[i] = ai[i-1] + ailen[i-1];
720: }
721: if (m) {
722: fshift += imax[m-1] - ailen[m-1];
723: ai[m] = ai[m-1] + ailen[m-1];
724: }
725: /* reset ilen and imax for each row */
726: for (i=0; i<m; i++) {
727: ailen[i] = imax[i] = ai[i+1] - ai[i];
728: }
729: a->nz = ai[m];
730: for (i=0; i<a->nz; i++) {
731: #if defined(PETSC_USE_DEBUG)
732: if (!aa[i]) SETERRQ3(PETSC_ERR_PLIB,"Null matrix at location %D column %D nz %D",i,aj[i],a->nz);
733: #endif
734: MatAssemblyBegin(aa[i],MAT_FINAL_ASSEMBLY);
735: MatAssemblyEnd(aa[i],MAT_FINAL_ASSEMBLY);
736: }
737: CHKMEMQ;
738: PetscInfo4(A,"Matrix size: %D X %D; storage space: %D unneeded,%D used\n",m,A->cmap->n/A->cmap->bs,fshift,a->nz);
739: PetscInfo1(A,"Number of mallocs during MatSetValues() is %D\n",a->reallocs);
740: PetscInfo1(A,"Maximum nonzeros in any row is %D\n",rmax);
741: a->reallocs = 0;
742: A->info.nz_unneeded = (double)fshift;
743: a->rmax = rmax;
745: A->same_nonzero = PETSC_TRUE;
746: MatMarkDiagonal_BlockMat(A);
747: return(0);
748: }
752: PetscErrorCode MatSetOption_BlockMat(Mat A,MatOption opt,PetscTruth flg)
753: {
755: if (opt == MAT_SYMMETRIC && flg) {
756: A->ops->sor = MatSOR_BlockMat_Symmetric;
757: A->ops->mult = MatMult_BlockMat_Symmetric;
758: } else {
759: PetscInfo1(A,"Unused matrix option %s\n",MatOptions[opt]);
760: }
761: return(0);
762: }
765: static struct _MatOps MatOps_Values = {MatSetValues_BlockMat,
766: 0,
767: 0,
768: MatMult_BlockMat,
769: /* 4*/ MatMultAdd_BlockMat,
770: MatMultTranspose_BlockMat,
771: MatMultTransposeAdd_BlockMat,
772: 0,
773: 0,
774: 0,
775: /*10*/ 0,
776: 0,
777: 0,
778: MatSOR_BlockMat,
779: 0,
780: /*15*/ 0,
781: 0,
782: 0,
783: 0,
784: 0,
785: /*20*/ 0,
786: MatAssemblyEnd_BlockMat,
787: MatSetOption_BlockMat,
788: 0,
789: /*24*/ 0,
790: 0,
791: 0,
792: 0,
793: 0,
794: /*29*/ 0,
795: 0,
796: 0,
797: 0,
798: 0,
799: /*34*/ 0,
800: 0,
801: 0,
802: 0,
803: 0,
804: /*39*/ 0,
805: 0,
806: 0,
807: 0,
808: 0,
809: /*44*/ 0,
810: 0,
811: 0,
812: 0,
813: 0,
814: /*49*/ 0,
815: 0,
816: 0,
817: 0,
818: 0,
819: /*54*/ 0,
820: 0,
821: 0,
822: 0,
823: 0,
824: /*59*/ MatGetSubMatrix_BlockMat,
825: MatDestroy_BlockMat,
826: MatView_BlockMat,
827: 0,
828: 0,
829: /*64*/ 0,
830: 0,
831: 0,
832: 0,
833: 0,
834: /*69*/ 0,
835: 0,
836: 0,
837: 0,
838: 0,
839: /*74*/ 0,
840: 0,
841: 0,
842: 0,
843: 0,
844: /*79*/ 0,
845: 0,
846: 0,
847: 0,
848: MatLoad_BlockMat,
849: /*84*/ 0,
850: 0,
851: 0,
852: 0,
853: 0,
854: /*89*/ 0,
855: 0,
856: 0,
857: 0,
858: 0,
859: /*94*/ 0,
860: 0,
861: 0,
862: 0};
866: /*@C
867: MatBlockMatSetPreallocation - For good matrix assembly performance
868: the user should preallocate the matrix storage by setting the parameter nz
869: (or the array nnz). By setting these parameters accurately, performance
870: during matrix assembly can be increased by more than a factor of 50.
872: Collective on MPI_Comm
874: Input Parameters:
875: + B - The matrix
876: . bs - size of each block in matrix
877: . nz - number of nonzeros per block row (same for all rows)
878: - nnz - array containing the number of nonzeros in the various block rows
879: (possibly different for each row) or PETSC_NULL
881: Notes:
882: If nnz is given then nz is ignored
884: Specify the preallocated storage with either nz or nnz (not both).
885: Set nz=PETSC_DEFAULT and nnz=PETSC_NULL for PETSc to control dynamic memory
886: allocation. For large problems you MUST preallocate memory or you
887: will get TERRIBLE performance, see the users' manual chapter on matrices.
889: Level: intermediate
891: .seealso: MatCreate(), MatCreateBlockMat(), MatSetValues()
893: @*/
894: PetscErrorCode MatBlockMatSetPreallocation(Mat B,PetscInt bs,PetscInt nz,const PetscInt nnz[])
895: {
896: PetscErrorCode ierr,(*f)(Mat,PetscInt,PetscInt,const PetscInt[]);
899: PetscObjectQueryFunction((PetscObject)B,"MatBlockMatSetPreallocation_C",(void (**)(void))&f);
900: if (f) {
901: (*f)(B,bs,nz,nnz);
902: }
903: return(0);
904: }
909: PetscErrorCode MatBlockMatSetPreallocation_BlockMat(Mat A,PetscInt bs,PetscInt nz,PetscInt *nnz)
910: {
911: Mat_BlockMat *bmat = (Mat_BlockMat*)A->data;
913: PetscInt i;
916: if (bs < 1) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"Block size given %D must be great than zero",bs);
917: if (A->rmap->n % bs) SETERRQ2(PETSC_ERR_ARG_INCOMP,"Blocksize %D does not divide number of rows %D",bs,A->rmap->n);
918: if (A->cmap->n % bs) SETERRQ2(PETSC_ERR_ARG_INCOMP,"Blocksize %D does not divide number of columns %D",bs,A->cmap->n);
919: if (nz == PETSC_DEFAULT || nz == PETSC_DECIDE) nz = 5;
920: if (nz < 0) SETERRQ1(PETSC_ERR_ARG_OUTOFRANGE,"nz cannot be less than 0: value %d",nz);
921: if (nnz) {
922: for (i=0; i<A->rmap->n/bs; i++) {
923: if (nnz[i] < 0) SETERRQ2(PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be less than 0: local row %d value %d",i,nnz[i]);
924: if (nnz[i] > A->cmap->n/bs) SETERRQ3(PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be greater than row length: local row %d value %d rowlength %d",i,nnz[i],A->cmap->n/bs);
925: }
926: }
927: A->rmap->bs = A->cmap->bs = bs;
928: bmat->mbs = A->rmap->n/bs;
930: VecCreateSeqWithArray(PETSC_COMM_SELF,bs,PETSC_NULL,&bmat->right);
931: VecCreateSeqWithArray(PETSC_COMM_SELF,bs,PETSC_NULL,&bmat->middle);
932: VecCreateSeq(PETSC_COMM_SELF,bs,&bmat->left);
934: if (!bmat->imax) {
935: PetscMalloc2(A->rmap->n,PetscInt,&bmat->imax,A->rmap->n,PetscInt,&bmat->ilen);
936: PetscLogObjectMemory(A,2*A->rmap->n*sizeof(PetscInt));
937: }
938: if (nnz) {
939: nz = 0;
940: for (i=0; i<A->rmap->n/A->rmap->bs; i++) {
941: bmat->imax[i] = nnz[i];
942: nz += nnz[i];
943: }
944: } else {
945: SETERRQ(PETSC_ERR_SUP,"Currently requires block row by row preallocation");
946: }
948: /* bmat->ilen will count nonzeros in each row so far. */
949: for (i=0; i<bmat->mbs; i++) { bmat->ilen[i] = 0;}
951: /* allocate the matrix space */
952: MatSeqXAIJFreeAIJ(A,(PetscScalar**)&bmat->a,&bmat->j,&bmat->i);
953: PetscMalloc3(nz,Mat,&bmat->a,nz,PetscInt,&bmat->j,A->rmap->n+1,PetscInt,&bmat->i);
954: PetscLogObjectMemory(A,(A->rmap->n+1)*sizeof(PetscInt)+nz*(sizeof(PetscScalar)+sizeof(PetscInt)));
955: bmat->i[0] = 0;
956: for (i=1; i<bmat->mbs+1; i++) {
957: bmat->i[i] = bmat->i[i-1] + bmat->imax[i-1];
958: }
959: bmat->singlemalloc = PETSC_TRUE;
960: bmat->free_a = PETSC_TRUE;
961: bmat->free_ij = PETSC_TRUE;
963: bmat->nz = 0;
964: bmat->maxnz = nz;
965: A->info.nz_unneeded = (double)bmat->maxnz;
967: return(0);
968: }
971: /*MC
972: MATBLOCKMAT - A matrix that is defined by a set of Mat's that represents a sparse block matrix
973: consisting of (usually) sparse blocks.
975: Level: advanced
977: .seealso: MatCreateBlockMat()
979: M*/
984: PetscErrorCode MatCreate_BlockMat(Mat A)
985: {
986: Mat_BlockMat *b;
990: PetscNewLog(A,Mat_BlockMat,&b);
991: A->data = (void*)b;
992: PetscMemcpy(A->ops,&MatOps_Values,sizeof(struct _MatOps));
994: PetscLayoutSetBlockSize(A->rmap,1);
995: PetscLayoutSetBlockSize(A->cmap,1);
996: PetscLayoutSetUp(A->rmap);
997: PetscLayoutSetUp(A->cmap);
999: A->assembled = PETSC_TRUE;
1000: A->preallocated = PETSC_FALSE;
1001: PetscObjectChangeTypeName((PetscObject)A,MATBLOCKMAT);
1003: PetscObjectComposeFunctionDynamic((PetscObject)A,"MatBlockMatSetPreallocation_C",
1004: "MatBlockMatSetPreallocation_BlockMat",
1005: MatBlockMatSetPreallocation_BlockMat);
1007: return(0);
1008: }
1013: /*@C
1014: MatCreateBlockMat - Creates a new matrix based sparse Mat storage
1016: Collective on MPI_Comm
1018: Input Parameters:
1019: + comm - MPI communicator
1020: . m - number of rows
1021: . n - number of columns
1022: . bs - size of each submatrix
1023: . nz - expected maximum number of nonzero blocks in row (use PETSC_DEFAULT if not known)
1024: - nnz - expected number of nonzers per block row if known (use PETSC_NULL otherwise)
1027: Output Parameter:
1028: . A - the matrix
1030: Level: intermediate
1032: PETSc requires that matrices and vectors being used for certain
1033: operations are partitioned accordingly. For example, when
1034: creating a bmat matrix, A, that supports parallel matrix-vector
1035: products using MatMult(A,x,y) the user should set the number
1036: of local matrix rows to be the number of local elements of the
1037: corresponding result vector, y. Note that this is information is
1038: required for use of the matrix interface routines, even though
1039: the bmat matrix may not actually be physically partitioned.
1040: For example,
1042: .keywords: matrix, bmat, create
1044: .seealso: MATBLOCKMAT
1045: @*/
1046: PetscErrorCode MatCreateBlockMat(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt bs,PetscInt nz,PetscInt *nnz, Mat *A)
1047: {
1051: MatCreate(comm,A);
1052: MatSetSizes(*A,m,n,PETSC_DETERMINE,PETSC_DETERMINE);
1053: MatSetType(*A,MATBLOCKMAT);
1054: MatBlockMatSetPreallocation(*A,bs,nz,nnz);
1055: return(0);
1056: }