xref: /petsc/src/mat/impls/aij/seq/aijfact.c (revision 94ef8dde638caef1d0cd84a7dc8a2db65fcda8b6)
1 
2 #include <../src/mat/impls/aij/seq/aij.h>
3 #include <../src/mat/impls/sbaij/seq/sbaij.h>
4 #include <petscbt.h>
5 #include <../src/mat/utils/freespace.h>
6 
7 /*
8       Computes an ordering to get most of the large numerical values in the lower triangular part of the matrix
9 
10       This code does not work and is not called anywhere. It would be registered with MatOrderingRegisterAll()
11 */
12 PetscErrorCode MatGetOrdering_Flow_SeqAIJ(Mat mat,MatOrderingType type,IS *irow,IS *icol)
13 {
14   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)mat->data;
15   PetscErrorCode    ierr;
16   PetscInt          i,j,jj,k, kk,n = mat->rmap->n, current = 0, newcurrent = 0,*order;
17   const PetscInt    *ai = a->i, *aj = a->j;
18   const PetscScalar *aa = a->a;
19   PetscBool         *done;
20   PetscReal         best,past = 0,future;
21 
22   PetscFunctionBegin;
23   /* pick initial row */
24   best = -1;
25   for (i=0; i<n; i++) {
26     future = 0.0;
27     for (j=ai[i]; j<ai[i+1]; j++) {
28       if (aj[j] != i) future += PetscAbsScalar(aa[j]);
29       else              past  = PetscAbsScalar(aa[j]);
30     }
31     if (!future) future = 1.e-10; /* if there is zero in the upper diagonal part want to rank this row high */
32     if (past/future > best) {
33       best    = past/future;
34       current = i;
35     }
36   }
37 
38   ierr     = PetscMalloc1(n,&done);CHKERRQ(ierr);
39   ierr     = PetscMemzero(done,n*sizeof(PetscBool));CHKERRQ(ierr);
40   ierr     = PetscMalloc1(n,&order);CHKERRQ(ierr);
41   order[0] = current;
42   for (i=0; i<n-1; i++) {
43     done[current] = PETSC_TRUE;
44     best          = -1;
45     /* loop over all neighbors of current pivot */
46     for (j=ai[current]; j<ai[current+1]; j++) {
47       jj = aj[j];
48       if (done[jj]) continue;
49       /* loop over columns of potential next row computing weights for below and above diagonal */
50       past = future = 0.0;
51       for (k=ai[jj]; k<ai[jj+1]; k++) {
52         kk = aj[k];
53         if (done[kk]) past += PetscAbsScalar(aa[k]);
54         else if (kk != jj) future += PetscAbsScalar(aa[k]);
55       }
56       if (!future) future = 1.e-10; /* if there is zero in the upper diagonal part want to rank this row high */
57       if (past/future > best) {
58         best       = past/future;
59         newcurrent = jj;
60       }
61     }
62     if (best == -1) { /* no neighbors to select from so select best of all that remain */
63       best = -1;
64       for (k=0; k<n; k++) {
65         if (done[k]) continue;
66         future = 0.0;
67         past   = 0.0;
68         for (j=ai[k]; j<ai[k+1]; j++) {
69           kk = aj[j];
70           if (done[kk])       past += PetscAbsScalar(aa[j]);
71           else if (kk != k) future += PetscAbsScalar(aa[j]);
72         }
73         if (!future) future = 1.e-10; /* if there is zero in the upper diagonal part want to rank this row high */
74         if (past/future > best) {
75           best       = past/future;
76           newcurrent = k;
77         }
78       }
79     }
80     if (current == newcurrent) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"newcurrent cannot be current");
81     current    = newcurrent;
82     order[i+1] = current;
83   }
84   ierr  = ISCreateGeneral(PETSC_COMM_SELF,n,order,PETSC_COPY_VALUES,irow);CHKERRQ(ierr);
85   *icol = *irow;
86   ierr  = PetscObjectReference((PetscObject)*irow);CHKERRQ(ierr);
87   ierr  = PetscFree(done);CHKERRQ(ierr);
88   ierr  = PetscFree(order);CHKERRQ(ierr);
89   PetscFunctionReturn(0);
90 }
91 
92 PETSC_INTERN PetscErrorCode MatGetFactor_seqaij_petsc(Mat A,MatFactorType ftype,Mat *B)
93 {
94   PetscInt       n = A->rmap->n;
95   PetscErrorCode ierr;
96 
97   PetscFunctionBegin;
98 #if defined(PETSC_USE_COMPLEX)
99   if (A->hermitian && (ftype == MAT_FACTOR_CHOLESKY || ftype == MAT_FACTOR_ICC)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Hermitian Factor is not supported");
100 #endif
101   ierr = MatCreate(PetscObjectComm((PetscObject)A),B);CHKERRQ(ierr);
102   ierr = MatSetSizes(*B,n,n,n,n);CHKERRQ(ierr);
103   if (ftype == MAT_FACTOR_LU || ftype == MAT_FACTOR_ILU || ftype == MAT_FACTOR_ILUDT) {
104     ierr = MatSetType(*B,MATSEQAIJ);CHKERRQ(ierr);
105 
106     (*B)->ops->ilufactorsymbolic = MatILUFactorSymbolic_SeqAIJ;
107     (*B)->ops->lufactorsymbolic  = MatLUFactorSymbolic_SeqAIJ;
108 
109     ierr = MatSetBlockSizesFromMats(*B,A,A);CHKERRQ(ierr);
110   } else if (ftype == MAT_FACTOR_CHOLESKY || ftype == MAT_FACTOR_ICC) {
111     ierr = MatSetType(*B,MATSEQSBAIJ);CHKERRQ(ierr);
112     ierr = MatSeqSBAIJSetPreallocation(*B,1,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
113 
114     (*B)->ops->iccfactorsymbolic      = MatICCFactorSymbolic_SeqAIJ;
115     (*B)->ops->choleskyfactorsymbolic = MatCholeskyFactorSymbolic_SeqAIJ;
116   } else SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Factor type not supported");
117   (*B)->factortype = ftype;
118 
119   ierr = PetscFree((*B)->solvertype);CHKERRQ(ierr);
120   ierr = PetscStrallocpy(MATSOLVERPETSC,&(*B)->solvertype);CHKERRQ(ierr);
121   PetscFunctionReturn(0);
122 }
123 
124 PetscErrorCode MatLUFactorSymbolic_SeqAIJ_inplace(Mat B,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
125 {
126   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
127   IS                 isicol;
128   PetscErrorCode     ierr;
129   const PetscInt     *r,*ic;
130   PetscInt           i,n=A->rmap->n,*ai=a->i,*aj=a->j;
131   PetscInt           *bi,*bj,*ajtmp;
132   PetscInt           *bdiag,row,nnz,nzi,reallocs=0,nzbd,*im;
133   PetscReal          f;
134   PetscInt           nlnk,*lnk,k,**bi_ptr;
135   PetscFreeSpaceList free_space=NULL,current_space=NULL;
136   PetscBT            lnkbt;
137   PetscBool          missing;
138 
139   PetscFunctionBegin;
140   if (A->rmap->N != A->cmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"matrix must be square");
141   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
142   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
143 
144   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
145   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
146   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
147 
148   /* get new row pointers */
149   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
150   bi[0] = 0;
151 
152   /* bdiag is location of diagonal in factor */
153   ierr     = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
154   bdiag[0] = 0;
155 
156   /* linked list for storing column indices of the active row */
157   nlnk = n + 1;
158   ierr = PetscLLCreate(n,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
159 
160   ierr = PetscMalloc2(n+1,&bi_ptr,n+1,&im);CHKERRQ(ierr);
161 
162   /* initial FreeSpace size is f*(ai[n]+1) */
163   f             = info->fill;
164   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
165   current_space = free_space;
166 
167   for (i=0; i<n; i++) {
168     /* copy previous fill into linked list */
169     nzi = 0;
170     nnz = ai[r[i]+1] - ai[r[i]];
171     ajtmp = aj + ai[r[i]];
172     ierr  = PetscLLAddPerm(nnz,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
173     nzi  += nlnk;
174 
175     /* add pivot rows into linked list */
176     row = lnk[n];
177     while (row < i) {
178       nzbd  = bdiag[row] - bi[row] + 1;   /* num of entries in the row with column index <= row */
179       ajtmp = bi_ptr[row] + nzbd;   /* points to the entry next to the diagonal */
180       ierr  = PetscLLAddSortedLU(ajtmp,row,nlnk,lnk,lnkbt,i,nzbd,im);CHKERRQ(ierr);
181       nzi  += nlnk;
182       row   = lnk[row];
183     }
184     bi[i+1] = bi[i] + nzi;
185     im[i]   = nzi;
186 
187     /* mark bdiag */
188     nzbd = 0;
189     nnz  = nzi;
190     k    = lnk[n];
191     while (nnz-- && k < i) {
192       nzbd++;
193       k = lnk[k];
194     }
195     bdiag[i] = bi[i] + nzbd;
196 
197     /* if free space is not available, make more free space */
198     if (current_space->local_remaining<nzi) {
199       nnz  = PetscIntMultTruncate(n - i,nzi); /* estimated and max additional space needed */
200       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
201       reallocs++;
202     }
203 
204     /* copy data into free space, then initialize lnk */
205     ierr = PetscLLClean(n,n,nzi,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
206 
207     bi_ptr[i]                       = current_space->array;
208     current_space->array           += nzi;
209     current_space->local_used      += nzi;
210     current_space->local_remaining -= nzi;
211   }
212 #if defined(PETSC_USE_INFO)
213   if (ai[n] != 0) {
214     PetscReal af = ((PetscReal)bi[n])/((PetscReal)ai[n]);
215     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
216     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
217     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g);\n",(double)af);CHKERRQ(ierr);
218     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
219   } else {
220     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
221   }
222 #endif
223 
224   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
225   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
226 
227   /* destroy list of free space and other temporary array(s) */
228   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
229   ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr);
230   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
231   ierr = PetscFree2(bi_ptr,im);CHKERRQ(ierr);
232 
233   /* put together the new matrix */
234   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
235   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
236   b    = (Mat_SeqAIJ*)(B)->data;
237 
238   b->free_a       = PETSC_TRUE;
239   b->free_ij      = PETSC_TRUE;
240   b->singlemalloc = PETSC_FALSE;
241 
242   ierr    = PetscMalloc1(bi[n]+1,&b->a);CHKERRQ(ierr);
243   b->j    = bj;
244   b->i    = bi;
245   b->diag = bdiag;
246   b->ilen = 0;
247   b->imax = 0;
248   b->row  = isrow;
249   b->col  = iscol;
250   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
251   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
252   b->icol = isicol;
253   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
254 
255   /* In b structure:  Free imax, ilen, old a, old j.  Allocate solve_work, new a, new j */
256   ierr     = PetscLogObjectMemory((PetscObject)B,(bi[n]-n)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
257   b->maxnz = b->nz = bi[n];
258 
259   (B)->factortype            = MAT_FACTOR_LU;
260   (B)->info.factor_mallocs   = reallocs;
261   (B)->info.fill_ratio_given = f;
262 
263   if (ai[n]) {
264     (B)->info.fill_ratio_needed = ((PetscReal)bi[n])/((PetscReal)ai[n]);
265   } else {
266     (B)->info.fill_ratio_needed = 0.0;
267   }
268   (B)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_inplace;
269   if (a->inode.size) {
270     (B)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
271   }
272   PetscFunctionReturn(0);
273 }
274 
275 PetscErrorCode MatLUFactorSymbolic_SeqAIJ(Mat B,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
276 {
277   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
278   IS                 isicol;
279   PetscErrorCode     ierr;
280   const PetscInt     *r,*ic,*ai=a->i,*aj=a->j,*ajtmp;
281   PetscInt           i,n=A->rmap->n;
282   PetscInt           *bi,*bj;
283   PetscInt           *bdiag,row,nnz,nzi,reallocs=0,nzbd,*im;
284   PetscReal          f;
285   PetscInt           nlnk,*lnk,k,**bi_ptr;
286   PetscFreeSpaceList free_space=NULL,current_space=NULL;
287   PetscBT            lnkbt;
288   PetscBool          missing;
289 
290   PetscFunctionBegin;
291   if (A->rmap->N != A->cmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"matrix must be square");
292   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
293   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
294 
295   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
296   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
297   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
298 
299   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
300   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
301   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
302   bi[0] = bdiag[0] = 0;
303 
304   /* linked list for storing column indices of the active row */
305   nlnk = n + 1;
306   ierr = PetscLLCreate(n,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
307 
308   ierr = PetscMalloc2(n+1,&bi_ptr,n+1,&im);CHKERRQ(ierr);
309 
310   /* initial FreeSpace size is f*(ai[n]+1) */
311   f             = info->fill;
312   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
313   current_space = free_space;
314 
315   for (i=0; i<n; i++) {
316     /* copy previous fill into linked list */
317     nzi = 0;
318     nnz = ai[r[i]+1] - ai[r[i]];
319     ajtmp = aj + ai[r[i]];
320     ierr  = PetscLLAddPerm(nnz,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
321     nzi  += nlnk;
322 
323     /* add pivot rows into linked list */
324     row = lnk[n];
325     while (row < i) {
326       nzbd  = bdiag[row] + 1; /* num of entries in the row with column index <= row */
327       ajtmp = bi_ptr[row] + nzbd; /* points to the entry next to the diagonal */
328       ierr  = PetscLLAddSortedLU(ajtmp,row,nlnk,lnk,lnkbt,i,nzbd,im);CHKERRQ(ierr);
329       nzi  += nlnk;
330       row   = lnk[row];
331     }
332     bi[i+1] = bi[i] + nzi;
333     im[i]   = nzi;
334 
335     /* mark bdiag */
336     nzbd = 0;
337     nnz  = nzi;
338     k    = lnk[n];
339     while (nnz-- && k < i) {
340       nzbd++;
341       k = lnk[k];
342     }
343     bdiag[i] = nzbd; /* note: bdiag[i] = nnzL as input for PetscFreeSpaceContiguous_LU() */
344 
345     /* if free space is not available, make more free space */
346     if (current_space->local_remaining<nzi) {
347       /* estimated additional space needed */
348       nnz  = PetscIntMultTruncate(2,PetscIntMultTruncate(n-1,nzi));
349       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
350       reallocs++;
351     }
352 
353     /* copy data into free space, then initialize lnk */
354     ierr = PetscLLClean(n,n,nzi,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
355 
356     bi_ptr[i]                       = current_space->array;
357     current_space->array           += nzi;
358     current_space->local_used      += nzi;
359     current_space->local_remaining -= nzi;
360   }
361 
362   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
363   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
364 
365   /*   copy free_space into bj and free free_space; set bi, bj, bdiag in new datastructure; */
366   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
367   ierr = PetscFreeSpaceContiguous_LU(&free_space,bj,n,bi,bdiag);CHKERRQ(ierr);
368   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
369   ierr = PetscFree2(bi_ptr,im);CHKERRQ(ierr);
370 
371   /* put together the new matrix */
372   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
373   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
374   b    = (Mat_SeqAIJ*)(B)->data;
375 
376   b->free_a       = PETSC_TRUE;
377   b->free_ij      = PETSC_TRUE;
378   b->singlemalloc = PETSC_FALSE;
379 
380   ierr = PetscMalloc1(bdiag[0]+1,&b->a);CHKERRQ(ierr);
381 
382   b->j    = bj;
383   b->i    = bi;
384   b->diag = bdiag;
385   b->ilen = 0;
386   b->imax = 0;
387   b->row  = isrow;
388   b->col  = iscol;
389   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
390   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
391   b->icol = isicol;
392   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
393 
394   /* In b structure:  Free imax, ilen, old a, old j.  Allocate solve_work, new a, new j */
395   ierr     = PetscLogObjectMemory((PetscObject)B,(bdiag[0]+1)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
396   b->maxnz = b->nz = bdiag[0]+1;
397 
398   B->factortype            = MAT_FACTOR_LU;
399   B->info.factor_mallocs   = reallocs;
400   B->info.fill_ratio_given = f;
401 
402   if (ai[n]) {
403     B->info.fill_ratio_needed = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
404   } else {
405     B->info.fill_ratio_needed = 0.0;
406   }
407 #if defined(PETSC_USE_INFO)
408   if (ai[n] != 0) {
409     PetscReal af = B->info.fill_ratio_needed;
410     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
411     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
412     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g);\n",(double)af);CHKERRQ(ierr);
413     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
414   } else {
415     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
416   }
417 #endif
418   B->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ;
419   if (a->inode.size) {
420     B->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
421   }
422   ierr = MatSeqAIJCheckInode_FactorLU(B);CHKERRQ(ierr);
423   PetscFunctionReturn(0);
424 }
425 
426 /*
427     Trouble in factorization, should we dump the original matrix?
428 */
429 PetscErrorCode MatFactorDumpMatrix(Mat A)
430 {
431   PetscErrorCode ierr;
432   PetscBool      flg = PETSC_FALSE;
433 
434   PetscFunctionBegin;
435   ierr = PetscOptionsGetBool(((PetscObject)A)->options,NULL,"-mat_factor_dump_on_error",&flg,NULL);CHKERRQ(ierr);
436   if (flg) {
437     PetscViewer viewer;
438     char        filename[PETSC_MAX_PATH_LEN];
439 
440     ierr = PetscSNPrintf(filename,PETSC_MAX_PATH_LEN,"matrix_factor_error.%d",PetscGlobalRank);CHKERRQ(ierr);
441     ierr = PetscViewerBinaryOpen(PetscObjectComm((PetscObject)A),filename,FILE_MODE_WRITE,&viewer);CHKERRQ(ierr);
442     ierr = MatView(A,viewer);CHKERRQ(ierr);
443     ierr = PetscViewerDestroy(&viewer);CHKERRQ(ierr);
444   }
445   PetscFunctionReturn(0);
446 }
447 
448 PetscErrorCode MatLUFactorNumeric_SeqAIJ(Mat B,Mat A,const MatFactorInfo *info)
449 {
450   Mat             C     =B;
451   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
452   IS              isrow = b->row,isicol = b->icol;
453   PetscErrorCode  ierr;
454   const PetscInt  *r,*ic,*ics;
455   const PetscInt  n=A->rmap->n,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j,*bdiag=b->diag;
456   PetscInt        i,j,k,nz,nzL,row,*pj;
457   const PetscInt  *ajtmp,*bjtmp;
458   MatScalar       *rtmp,*pc,multiplier,*pv;
459   const MatScalar *aa=a->a,*v;
460   PetscBool       row_identity,col_identity;
461   FactorShiftCtx  sctx;
462   const PetscInt  *ddiag;
463   PetscReal       rs;
464   MatScalar       d;
465 
466   PetscFunctionBegin;
467   /* MatPivotSetUp(): initialize shift context sctx */
468   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
469 
470   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
471     ddiag          = a->diag;
472     sctx.shift_top = info->zeropivot;
473     for (i=0; i<n; i++) {
474       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
475       d  = (aa)[ddiag[i]];
476       rs = -PetscAbsScalar(d) - PetscRealPart(d);
477       v  = aa+ai[i];
478       nz = ai[i+1] - ai[i];
479       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
480       if (rs>sctx.shift_top) sctx.shift_top = rs;
481     }
482     sctx.shift_top *= 1.1;
483     sctx.nshift_max = 5;
484     sctx.shift_lo   = 0.;
485     sctx.shift_hi   = 1.;
486   }
487 
488   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
489   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
490   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
491   ics  = ic;
492 
493   do {
494     sctx.newshift = PETSC_FALSE;
495     for (i=0; i<n; i++) {
496       /* zero rtmp */
497       /* L part */
498       nz    = bi[i+1] - bi[i];
499       bjtmp = bj + bi[i];
500       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
501 
502       /* U part */
503       nz    = bdiag[i]-bdiag[i+1];
504       bjtmp = bj + bdiag[i+1]+1;
505       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
506 
507       /* load in initial (unfactored row) */
508       nz    = ai[r[i]+1] - ai[r[i]];
509       ajtmp = aj + ai[r[i]];
510       v     = aa + ai[r[i]];
511       for (j=0; j<nz; j++) {
512         rtmp[ics[ajtmp[j]]] = v[j];
513       }
514       /* ZeropivotApply() */
515       rtmp[i] += sctx.shift_amount;  /* shift the diagonal of the matrix */
516 
517       /* elimination */
518       bjtmp = bj + bi[i];
519       row   = *bjtmp++;
520       nzL   = bi[i+1] - bi[i];
521       for (k=0; k < nzL; k++) {
522         pc = rtmp + row;
523         if (*pc != 0.0) {
524           pv         = b->a + bdiag[row];
525           multiplier = *pc * (*pv);
526           *pc        = multiplier;
527 
528           pj = b->j + bdiag[row+1]+1; /* beginning of U(row,:) */
529           pv = b->a + bdiag[row+1]+1;
530           nz = bdiag[row]-bdiag[row+1]-1; /* num of entries in U(row,:) excluding diag */
531 
532           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
533           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
534         }
535         row = *bjtmp++;
536       }
537 
538       /* finished row so stick it into b->a */
539       rs = 0.0;
540       /* L part */
541       pv = b->a + bi[i];
542       pj = b->j + bi[i];
543       nz = bi[i+1] - bi[i];
544       for (j=0; j<nz; j++) {
545         pv[j] = rtmp[pj[j]]; rs += PetscAbsScalar(pv[j]);
546       }
547 
548       /* U part */
549       pv = b->a + bdiag[i+1]+1;
550       pj = b->j + bdiag[i+1]+1;
551       nz = bdiag[i] - bdiag[i+1]-1;
552       for (j=0; j<nz; j++) {
553         pv[j] = rtmp[pj[j]]; rs += PetscAbsScalar(pv[j]);
554       }
555 
556       sctx.rs = rs;
557       sctx.pv = rtmp[i];
558       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
559       if (sctx.newshift) break; /* break for-loop */
560       rtmp[i] = sctx.pv; /* sctx.pv might be updated in the case of MAT_SHIFT_INBLOCKS */
561 
562       /* Mark diagonal and invert diagonal for simplier triangular solves */
563       pv  = b->a + bdiag[i];
564       *pv = 1.0/rtmp[i];
565 
566     } /* endof for (i=0; i<n; i++) { */
567 
568     /* MatPivotRefine() */
569     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
570       /*
571        * if no shift in this attempt & shifting & started shifting & can refine,
572        * then try lower shift
573        */
574       sctx.shift_hi       = sctx.shift_fraction;
575       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
576       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
577       sctx.newshift       = PETSC_TRUE;
578       sctx.nshift++;
579     }
580   } while (sctx.newshift);
581 
582   ierr = PetscFree(rtmp);CHKERRQ(ierr);
583   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
584   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
585 
586   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
587   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
588   if (b->inode.size) {
589     C->ops->solve = MatSolve_SeqAIJ_Inode;
590   } else if (row_identity && col_identity) {
591     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
592   } else {
593     C->ops->solve = MatSolve_SeqAIJ;
594   }
595   C->ops->solveadd          = MatSolveAdd_SeqAIJ;
596   C->ops->solvetranspose    = MatSolveTranspose_SeqAIJ;
597   C->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ;
598   C->ops->matsolve          = MatMatSolve_SeqAIJ;
599   C->assembled              = PETSC_TRUE;
600   C->preallocated           = PETSC_TRUE;
601 
602   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
603 
604   /* MatShiftView(A,info,&sctx) */
605   if (sctx.nshift) {
606     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
607       ierr = PetscInfo4(A,"number of shift_pd tries %D, shift_amount %g, diagonal shifted up by %e fraction top_value %e\n",sctx.nshift,(double)sctx.shift_amount,(double)sctx.shift_fraction,(double)sctx.shift_top);CHKERRQ(ierr);
608     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
609       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
610     } else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) {
611       ierr = PetscInfo2(A,"number of shift_inblocks applied %D, each shift_amount %g\n",sctx.nshift,(double)info->shiftamount);CHKERRQ(ierr);
612     }
613   }
614   PetscFunctionReturn(0);
615 }
616 
617 PetscErrorCode MatLUFactorNumeric_SeqAIJ_inplace(Mat B,Mat A,const MatFactorInfo *info)
618 {
619   Mat             C     =B;
620   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
621   IS              isrow = b->row,isicol = b->icol;
622   PetscErrorCode  ierr;
623   const PetscInt  *r,*ic,*ics;
624   PetscInt        nz,row,i,j,n=A->rmap->n,diag;
625   const PetscInt  *ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j;
626   const PetscInt  *ajtmp,*bjtmp,*diag_offset = b->diag,*pj;
627   MatScalar       *pv,*rtmp,*pc,multiplier,d;
628   const MatScalar *v,*aa=a->a;
629   PetscReal       rs=0.0;
630   FactorShiftCtx  sctx;
631   const PetscInt  *ddiag;
632   PetscBool       row_identity, col_identity;
633 
634   PetscFunctionBegin;
635   /* MatPivotSetUp(): initialize shift context sctx */
636   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
637 
638   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
639     ddiag          = a->diag;
640     sctx.shift_top = info->zeropivot;
641     for (i=0; i<n; i++) {
642       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
643       d  = (aa)[ddiag[i]];
644       rs = -PetscAbsScalar(d) - PetscRealPart(d);
645       v  = aa+ai[i];
646       nz = ai[i+1] - ai[i];
647       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
648       if (rs>sctx.shift_top) sctx.shift_top = rs;
649     }
650     sctx.shift_top *= 1.1;
651     sctx.nshift_max = 5;
652     sctx.shift_lo   = 0.;
653     sctx.shift_hi   = 1.;
654   }
655 
656   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
657   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
658   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
659   ics  = ic;
660 
661   do {
662     sctx.newshift = PETSC_FALSE;
663     for (i=0; i<n; i++) {
664       nz    = bi[i+1] - bi[i];
665       bjtmp = bj + bi[i];
666       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
667 
668       /* load in initial (unfactored row) */
669       nz    = ai[r[i]+1] - ai[r[i]];
670       ajtmp = aj + ai[r[i]];
671       v     = aa + ai[r[i]];
672       for (j=0; j<nz; j++) {
673         rtmp[ics[ajtmp[j]]] = v[j];
674       }
675       rtmp[ics[r[i]]] += sctx.shift_amount; /* shift the diagonal of the matrix */
676 
677       row = *bjtmp++;
678       while  (row < i) {
679         pc = rtmp + row;
680         if (*pc != 0.0) {
681           pv         = b->a + diag_offset[row];
682           pj         = b->j + diag_offset[row] + 1;
683           multiplier = *pc / *pv++;
684           *pc        = multiplier;
685           nz         = bi[row+1] - diag_offset[row] - 1;
686           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
687           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
688         }
689         row = *bjtmp++;
690       }
691       /* finished row so stick it into b->a */
692       pv   = b->a + bi[i];
693       pj   = b->j + bi[i];
694       nz   = bi[i+1] - bi[i];
695       diag = diag_offset[i] - bi[i];
696       rs   = 0.0;
697       for (j=0; j<nz; j++) {
698         pv[j] = rtmp[pj[j]];
699         rs   += PetscAbsScalar(pv[j]);
700       }
701       rs -= PetscAbsScalar(pv[diag]);
702 
703       sctx.rs = rs;
704       sctx.pv = pv[diag];
705       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
706       if (sctx.newshift) break;
707       pv[diag] = sctx.pv;
708     }
709 
710     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
711       /*
712        * if no shift in this attempt & shifting & started shifting & can refine,
713        * then try lower shift
714        */
715       sctx.shift_hi       = sctx.shift_fraction;
716       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
717       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
718       sctx.newshift       = PETSC_TRUE;
719       sctx.nshift++;
720     }
721   } while (sctx.newshift);
722 
723   /* invert diagonal entries for simplier triangular solves */
724   for (i=0; i<n; i++) {
725     b->a[diag_offset[i]] = 1.0/b->a[diag_offset[i]];
726   }
727   ierr = PetscFree(rtmp);CHKERRQ(ierr);
728   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
729   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
730 
731   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
732   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
733   if (row_identity && col_identity) {
734     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering_inplace;
735   } else {
736     C->ops->solve = MatSolve_SeqAIJ_inplace;
737   }
738   C->ops->solveadd          = MatSolveAdd_SeqAIJ_inplace;
739   C->ops->solvetranspose    = MatSolveTranspose_SeqAIJ_inplace;
740   C->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ_inplace;
741   C->ops->matsolve          = MatMatSolve_SeqAIJ_inplace;
742 
743   C->assembled    = PETSC_TRUE;
744   C->preallocated = PETSC_TRUE;
745 
746   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
747   if (sctx.nshift) {
748     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
749       ierr = PetscInfo4(A,"number of shift_pd tries %D, shift_amount %g, diagonal shifted up by %e fraction top_value %e\n",sctx.nshift,(double)sctx.shift_amount,(double)sctx.shift_fraction,(double)sctx.shift_top);CHKERRQ(ierr);
750     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
751       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
752     }
753   }
754   (C)->ops->solve          = MatSolve_SeqAIJ_inplace;
755   (C)->ops->solvetranspose = MatSolveTranspose_SeqAIJ_inplace;
756 
757   ierr = MatSeqAIJCheckInode(C);CHKERRQ(ierr);
758   PetscFunctionReturn(0);
759 }
760 
761 /*
762    This routine implements inplace ILU(0) with row or/and column permutations.
763    Input:
764      A - original matrix
765    Output;
766      A - a->i (rowptr) is same as original rowptr, but factored i-the row is stored in rowperm[i]
767          a->j (col index) is permuted by the inverse of colperm, then sorted
768          a->a reordered accordingly with a->j
769          a->diag (ptr to diagonal elements) is updated.
770 */
771 PetscErrorCode MatLUFactorNumeric_SeqAIJ_InplaceWithPerm(Mat B,Mat A,const MatFactorInfo *info)
772 {
773   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data;
774   IS              isrow = a->row,isicol = a->icol;
775   PetscErrorCode  ierr;
776   const PetscInt  *r,*ic,*ics;
777   PetscInt        i,j,n=A->rmap->n,*ai=a->i,*aj=a->j;
778   PetscInt        *ajtmp,nz,row;
779   PetscInt        *diag = a->diag,nbdiag,*pj;
780   PetscScalar     *rtmp,*pc,multiplier,d;
781   MatScalar       *pv,*v;
782   PetscReal       rs;
783   FactorShiftCtx  sctx;
784   const MatScalar *aa=a->a,*vtmp;
785 
786   PetscFunctionBegin;
787   if (A != B) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"input and output matrix must have same address");
788 
789   /* MatPivotSetUp(): initialize shift context sctx */
790   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
791 
792   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
793     const PetscInt *ddiag = a->diag;
794     sctx.shift_top = info->zeropivot;
795     for (i=0; i<n; i++) {
796       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
797       d    = (aa)[ddiag[i]];
798       rs   = -PetscAbsScalar(d) - PetscRealPart(d);
799       vtmp = aa+ai[i];
800       nz   = ai[i+1] - ai[i];
801       for (j=0; j<nz; j++) rs += PetscAbsScalar(vtmp[j]);
802       if (rs>sctx.shift_top) sctx.shift_top = rs;
803     }
804     sctx.shift_top *= 1.1;
805     sctx.nshift_max = 5;
806     sctx.shift_lo   = 0.;
807     sctx.shift_hi   = 1.;
808   }
809 
810   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
811   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
812   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
813   ierr = PetscMemzero(rtmp,(n+1)*sizeof(PetscScalar));CHKERRQ(ierr);
814   ics  = ic;
815 
816 #if defined(MV)
817   sctx.shift_top      = 0.;
818   sctx.nshift_max     = 0;
819   sctx.shift_lo       = 0.;
820   sctx.shift_hi       = 0.;
821   sctx.shift_fraction = 0.;
822 
823   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
824     sctx.shift_top = 0.;
825     for (i=0; i<n; i++) {
826       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
827       d  = (a->a)[diag[i]];
828       rs = -PetscAbsScalar(d) - PetscRealPart(d);
829       v  = a->a+ai[i];
830       nz = ai[i+1] - ai[i];
831       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
832       if (rs>sctx.shift_top) sctx.shift_top = rs;
833     }
834     if (sctx.shift_top < info->zeropivot) sctx.shift_top = info->zeropivot;
835     sctx.shift_top *= 1.1;
836     sctx.nshift_max = 5;
837     sctx.shift_lo   = 0.;
838     sctx.shift_hi   = 1.;
839   }
840 
841   sctx.shift_amount = 0.;
842   sctx.nshift       = 0;
843 #endif
844 
845   do {
846     sctx.newshift = PETSC_FALSE;
847     for (i=0; i<n; i++) {
848       /* load in initial unfactored row */
849       nz    = ai[r[i]+1] - ai[r[i]];
850       ajtmp = aj + ai[r[i]];
851       v     = a->a + ai[r[i]];
852       /* sort permuted ajtmp and values v accordingly */
853       for (j=0; j<nz; j++) ajtmp[j] = ics[ajtmp[j]];
854       ierr = PetscSortIntWithScalarArray(nz,ajtmp,v);CHKERRQ(ierr);
855 
856       diag[r[i]] = ai[r[i]];
857       for (j=0; j<nz; j++) {
858         rtmp[ajtmp[j]] = v[j];
859         if (ajtmp[j] < i) diag[r[i]]++; /* update a->diag */
860       }
861       rtmp[r[i]] += sctx.shift_amount; /* shift the diagonal of the matrix */
862 
863       row = *ajtmp++;
864       while  (row < i) {
865         pc = rtmp + row;
866         if (*pc != 0.0) {
867           pv = a->a + diag[r[row]];
868           pj = aj + diag[r[row]] + 1;
869 
870           multiplier = *pc / *pv++;
871           *pc        = multiplier;
872           nz         = ai[r[row]+1] - diag[r[row]] - 1;
873           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
874           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
875         }
876         row = *ajtmp++;
877       }
878       /* finished row so overwrite it onto a->a */
879       pv     = a->a + ai[r[i]];
880       pj     = aj + ai[r[i]];
881       nz     = ai[r[i]+1] - ai[r[i]];
882       nbdiag = diag[r[i]] - ai[r[i]]; /* num of entries before the diagonal */
883 
884       rs = 0.0;
885       for (j=0; j<nz; j++) {
886         pv[j] = rtmp[pj[j]];
887         if (j != nbdiag) rs += PetscAbsScalar(pv[j]);
888       }
889 
890       sctx.rs = rs;
891       sctx.pv = pv[nbdiag];
892       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
893       if (sctx.newshift) break;
894       pv[nbdiag] = sctx.pv;
895     }
896 
897     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
898       /*
899        * if no shift in this attempt & shifting & started shifting & can refine,
900        * then try lower shift
901        */
902       sctx.shift_hi       = sctx.shift_fraction;
903       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
904       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
905       sctx.newshift       = PETSC_TRUE;
906       sctx.nshift++;
907     }
908   } while (sctx.newshift);
909 
910   /* invert diagonal entries for simplier triangular solves */
911   for (i=0; i<n; i++) {
912     a->a[diag[r[i]]] = 1.0/a->a[diag[r[i]]];
913   }
914 
915   ierr = PetscFree(rtmp);CHKERRQ(ierr);
916   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
917   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
918 
919   A->ops->solve             = MatSolve_SeqAIJ_InplaceWithPerm;
920   A->ops->solveadd          = MatSolveAdd_SeqAIJ_inplace;
921   A->ops->solvetranspose    = MatSolveTranspose_SeqAIJ_inplace;
922   A->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ_inplace;
923 
924   A->assembled    = PETSC_TRUE;
925   A->preallocated = PETSC_TRUE;
926 
927   ierr = PetscLogFlops(A->cmap->n);CHKERRQ(ierr);
928   if (sctx.nshift) {
929     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
930       ierr = PetscInfo4(A,"number of shift_pd tries %D, shift_amount %g, diagonal shifted up by %e fraction top_value %e\n",sctx.nshift,(double)sctx.shift_amount,(double)sctx.shift_fraction,(double)sctx.shift_top);CHKERRQ(ierr);
931     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
932       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
933     }
934   }
935   PetscFunctionReturn(0);
936 }
937 
938 /* ----------------------------------------------------------- */
939 PetscErrorCode MatLUFactor_SeqAIJ(Mat A,IS row,IS col,const MatFactorInfo *info)
940 {
941   PetscErrorCode ierr;
942   Mat            C;
943 
944   PetscFunctionBegin;
945   ierr = MatGetFactor(A,MATSOLVERPETSC,MAT_FACTOR_LU,&C);CHKERRQ(ierr);
946   ierr = MatLUFactorSymbolic(C,A,row,col,info);CHKERRQ(ierr);
947   ierr = MatLUFactorNumeric(C,A,info);CHKERRQ(ierr);
948 
949   A->ops->solve          = C->ops->solve;
950   A->ops->solvetranspose = C->ops->solvetranspose;
951 
952   ierr = MatHeaderMerge(A,&C);CHKERRQ(ierr);
953   ierr = PetscLogObjectParent((PetscObject)A,(PetscObject)((Mat_SeqAIJ*)(A->data))->icol);CHKERRQ(ierr);
954   PetscFunctionReturn(0);
955 }
956 /* ----------------------------------------------------------- */
957 
958 
959 PetscErrorCode MatSolve_SeqAIJ_inplace(Mat A,Vec bb,Vec xx)
960 {
961   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
962   IS                iscol = a->col,isrow = a->row;
963   PetscErrorCode    ierr;
964   PetscInt          i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
965   PetscInt          nz;
966   const PetscInt    *rout,*cout,*r,*c;
967   PetscScalar       *x,*tmp,*tmps,sum;
968   const PetscScalar *b;
969   const MatScalar   *aa = a->a,*v;
970 
971   PetscFunctionBegin;
972   if (!n) PetscFunctionReturn(0);
973 
974   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
975   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
976   tmp  = a->solve_work;
977 
978   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
979   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
980 
981   /* forward solve the lower triangular */
982   tmp[0] = b[*r++];
983   tmps   = tmp;
984   for (i=1; i<n; i++) {
985     v   = aa + ai[i];
986     vi  = aj + ai[i];
987     nz  = a->diag[i] - ai[i];
988     sum = b[*r++];
989     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
990     tmp[i] = sum;
991   }
992 
993   /* backward solve the upper triangular */
994   for (i=n-1; i>=0; i--) {
995     v   = aa + a->diag[i] + 1;
996     vi  = aj + a->diag[i] + 1;
997     nz  = ai[i+1] - a->diag[i] - 1;
998     sum = tmp[i];
999     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1000     x[*c--] = tmp[i] = sum*aa[a->diag[i]];
1001   }
1002 
1003   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1004   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1005   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1006   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1007   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1008   PetscFunctionReturn(0);
1009 }
1010 
1011 PetscErrorCode MatMatSolve_SeqAIJ_inplace(Mat A,Mat B,Mat X)
1012 {
1013   Mat_SeqAIJ      *a    = (Mat_SeqAIJ*)A->data;
1014   IS              iscol = a->col,isrow = a->row;
1015   PetscErrorCode  ierr;
1016   PetscInt        i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
1017   PetscInt        nz,neq;
1018   const PetscInt  *rout,*cout,*r,*c;
1019   PetscScalar     *x,*b,*tmp,*tmps,sum;
1020   const MatScalar *aa = a->a,*v;
1021   PetscBool       bisdense,xisdense;
1022 
1023   PetscFunctionBegin;
1024   if (!n) PetscFunctionReturn(0);
1025 
1026   ierr = PetscObjectTypeCompare((PetscObject)B,MATSEQDENSE,&bisdense);CHKERRQ(ierr);
1027   if (!bisdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"B matrix must be a SeqDense matrix");
1028   ierr = PetscObjectTypeCompare((PetscObject)X,MATSEQDENSE,&xisdense);CHKERRQ(ierr);
1029   if (!xisdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"X matrix must be a SeqDense matrix");
1030 
1031   ierr = MatDenseGetArray(B,&b);CHKERRQ(ierr);
1032   ierr = MatDenseGetArray(X,&x);CHKERRQ(ierr);
1033 
1034   tmp  = a->solve_work;
1035   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1036   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1037 
1038   for (neq=0; neq<B->cmap->n; neq++) {
1039     /* forward solve the lower triangular */
1040     tmp[0] = b[r[0]];
1041     tmps   = tmp;
1042     for (i=1; i<n; i++) {
1043       v   = aa + ai[i];
1044       vi  = aj + ai[i];
1045       nz  = a->diag[i] - ai[i];
1046       sum = b[r[i]];
1047       PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1048       tmp[i] = sum;
1049     }
1050     /* backward solve the upper triangular */
1051     for (i=n-1; i>=0; i--) {
1052       v   = aa + a->diag[i] + 1;
1053       vi  = aj + a->diag[i] + 1;
1054       nz  = ai[i+1] - a->diag[i] - 1;
1055       sum = tmp[i];
1056       PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1057       x[c[i]] = tmp[i] = sum*aa[a->diag[i]];
1058     }
1059 
1060     b += n;
1061     x += n;
1062   }
1063   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1064   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1065   ierr = MatDenseRestoreArray(B,&b);CHKERRQ(ierr);
1066   ierr = MatDenseRestoreArray(X,&x);CHKERRQ(ierr);
1067   ierr = PetscLogFlops(B->cmap->n*(2.0*a->nz - n));CHKERRQ(ierr);
1068   PetscFunctionReturn(0);
1069 }
1070 
1071 PetscErrorCode MatMatSolve_SeqAIJ(Mat A,Mat B,Mat X)
1072 {
1073   Mat_SeqAIJ      *a    = (Mat_SeqAIJ*)A->data;
1074   IS              iscol = a->col,isrow = a->row;
1075   PetscErrorCode  ierr;
1076   PetscInt        i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j,*adiag = a->diag;
1077   PetscInt        nz,neq;
1078   const PetscInt  *rout,*cout,*r,*c;
1079   PetscScalar     *x,*b,*tmp,sum;
1080   const MatScalar *aa = a->a,*v;
1081   PetscBool       bisdense,xisdense;
1082 
1083   PetscFunctionBegin;
1084   if (!n) PetscFunctionReturn(0);
1085 
1086   ierr = PetscObjectTypeCompare((PetscObject)B,MATSEQDENSE,&bisdense);CHKERRQ(ierr);
1087   if (!bisdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"B matrix must be a SeqDense matrix");
1088   ierr = PetscObjectTypeCompare((PetscObject)X,MATSEQDENSE,&xisdense);CHKERRQ(ierr);
1089   if (!xisdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"X matrix must be a SeqDense matrix");
1090 
1091   ierr = MatDenseGetArray(B,&b);CHKERRQ(ierr);
1092   ierr = MatDenseGetArray(X,&x);CHKERRQ(ierr);
1093 
1094   tmp  = a->solve_work;
1095   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1096   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1097 
1098   for (neq=0; neq<B->cmap->n; neq++) {
1099     /* forward solve the lower triangular */
1100     tmp[0] = b[r[0]];
1101     v      = aa;
1102     vi     = aj;
1103     for (i=1; i<n; i++) {
1104       nz  = ai[i+1] - ai[i];
1105       sum = b[r[i]];
1106       PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
1107       tmp[i] = sum;
1108       v     += nz; vi += nz;
1109     }
1110 
1111     /* backward solve the upper triangular */
1112     for (i=n-1; i>=0; i--) {
1113       v   = aa + adiag[i+1]+1;
1114       vi  = aj + adiag[i+1]+1;
1115       nz  = adiag[i]-adiag[i+1]-1;
1116       sum = tmp[i];
1117       PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
1118       x[c[i]] = tmp[i] = sum*v[nz]; /* v[nz] = aa[adiag[i]] */
1119     }
1120 
1121     b += n;
1122     x += n;
1123   }
1124   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1125   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1126   ierr = MatDenseRestoreArray(B,&b);CHKERRQ(ierr);
1127   ierr = MatDenseRestoreArray(X,&x);CHKERRQ(ierr);
1128   ierr = PetscLogFlops(B->cmap->n*(2.0*a->nz - n));CHKERRQ(ierr);
1129   PetscFunctionReturn(0);
1130 }
1131 
1132 PetscErrorCode MatSolve_SeqAIJ_InplaceWithPerm(Mat A,Vec bb,Vec xx)
1133 {
1134   Mat_SeqAIJ      *a    = (Mat_SeqAIJ*)A->data;
1135   IS              iscol = a->col,isrow = a->row;
1136   PetscErrorCode  ierr;
1137   const PetscInt  *r,*c,*rout,*cout;
1138   PetscInt        i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
1139   PetscInt        nz,row;
1140   PetscScalar     *x,*b,*tmp,*tmps,sum;
1141   const MatScalar *aa = a->a,*v;
1142 
1143   PetscFunctionBegin;
1144   if (!n) PetscFunctionReturn(0);
1145 
1146   ierr = VecGetArray(bb,&b);CHKERRQ(ierr);
1147   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1148   tmp  = a->solve_work;
1149 
1150   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1151   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
1152 
1153   /* forward solve the lower triangular */
1154   tmp[0] = b[*r++];
1155   tmps   = tmp;
1156   for (row=1; row<n; row++) {
1157     i   = rout[row]; /* permuted row */
1158     v   = aa + ai[i];
1159     vi  = aj + ai[i];
1160     nz  = a->diag[i] - ai[i];
1161     sum = b[*r++];
1162     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1163     tmp[row] = sum;
1164   }
1165 
1166   /* backward solve the upper triangular */
1167   for (row=n-1; row>=0; row--) {
1168     i   = rout[row]; /* permuted row */
1169     v   = aa + a->diag[i] + 1;
1170     vi  = aj + a->diag[i] + 1;
1171     nz  = ai[i+1] - a->diag[i] - 1;
1172     sum = tmp[row];
1173     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1174     x[*c--] = tmp[row] = sum*aa[a->diag[i]];
1175   }
1176 
1177   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1178   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1179   ierr = VecRestoreArray(bb,&b);CHKERRQ(ierr);
1180   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1181   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1182   PetscFunctionReturn(0);
1183 }
1184 
1185 /* ----------------------------------------------------------- */
1186 #include <../src/mat/impls/aij/seq/ftn-kernels/fsolve.h>
1187 PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering_inplace(Mat A,Vec bb,Vec xx)
1188 {
1189   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
1190   PetscErrorCode    ierr;
1191   PetscInt          n   = A->rmap->n;
1192   const PetscInt    *ai = a->i,*aj = a->j,*adiag = a->diag;
1193   PetscScalar       *x;
1194   const PetscScalar *b;
1195   const MatScalar   *aa = a->a;
1196 #if !defined(PETSC_USE_FORTRAN_KERNEL_SOLVEAIJ)
1197   PetscInt        adiag_i,i,nz,ai_i;
1198   const PetscInt  *vi;
1199   const MatScalar *v;
1200   PetscScalar     sum;
1201 #endif
1202 
1203   PetscFunctionBegin;
1204   if (!n) PetscFunctionReturn(0);
1205 
1206   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1207   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1208 
1209 #if defined(PETSC_USE_FORTRAN_KERNEL_SOLVEAIJ)
1210   fortransolveaij_(&n,x,ai,aj,adiag,aa,b);
1211 #else
1212   /* forward solve the lower triangular */
1213   x[0] = b[0];
1214   for (i=1; i<n; i++) {
1215     ai_i = ai[i];
1216     v    = aa + ai_i;
1217     vi   = aj + ai_i;
1218     nz   = adiag[i] - ai_i;
1219     sum  = b[i];
1220     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
1221     x[i] = sum;
1222   }
1223 
1224   /* backward solve the upper triangular */
1225   for (i=n-1; i>=0; i--) {
1226     adiag_i = adiag[i];
1227     v       = aa + adiag_i + 1;
1228     vi      = aj + adiag_i + 1;
1229     nz      = ai[i+1] - adiag_i - 1;
1230     sum     = x[i];
1231     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
1232     x[i] = sum*aa[adiag_i];
1233   }
1234 #endif
1235   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1236   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1237   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1238   PetscFunctionReturn(0);
1239 }
1240 
1241 PetscErrorCode MatSolveAdd_SeqAIJ_inplace(Mat A,Vec bb,Vec yy,Vec xx)
1242 {
1243   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1244   IS                iscol = a->col,isrow = a->row;
1245   PetscErrorCode    ierr;
1246   PetscInt          i, n = A->rmap->n,j;
1247   PetscInt          nz;
1248   const PetscInt    *rout,*cout,*r,*c,*vi,*ai = a->i,*aj = a->j;
1249   PetscScalar       *x,*tmp,sum;
1250   const PetscScalar *b;
1251   const MatScalar   *aa = a->a,*v;
1252 
1253   PetscFunctionBegin;
1254   if (yy != xx) {ierr = VecCopy(yy,xx);CHKERRQ(ierr);}
1255 
1256   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1257   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1258   tmp  = a->solve_work;
1259 
1260   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1261   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
1262 
1263   /* forward solve the lower triangular */
1264   tmp[0] = b[*r++];
1265   for (i=1; i<n; i++) {
1266     v   = aa + ai[i];
1267     vi  = aj + ai[i];
1268     nz  = a->diag[i] - ai[i];
1269     sum = b[*r++];
1270     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1271     tmp[i] = sum;
1272   }
1273 
1274   /* backward solve the upper triangular */
1275   for (i=n-1; i>=0; i--) {
1276     v   = aa + a->diag[i] + 1;
1277     vi  = aj + a->diag[i] + 1;
1278     nz  = ai[i+1] - a->diag[i] - 1;
1279     sum = tmp[i];
1280     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1281     tmp[i]   = sum*aa[a->diag[i]];
1282     x[*c--] += tmp[i];
1283   }
1284 
1285   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1286   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1287   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1288   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1289   ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr);
1290   PetscFunctionReturn(0);
1291 }
1292 
1293 PetscErrorCode MatSolveAdd_SeqAIJ(Mat A,Vec bb,Vec yy,Vec xx)
1294 {
1295   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1296   IS                iscol = a->col,isrow = a->row;
1297   PetscErrorCode    ierr;
1298   PetscInt          i, n = A->rmap->n,j;
1299   PetscInt          nz;
1300   const PetscInt    *rout,*cout,*r,*c,*vi,*ai = a->i,*aj = a->j,*adiag = a->diag;
1301   PetscScalar       *x,*tmp,sum;
1302   const PetscScalar *b;
1303   const MatScalar   *aa = a->a,*v;
1304 
1305   PetscFunctionBegin;
1306   if (yy != xx) {ierr = VecCopy(yy,xx);CHKERRQ(ierr);}
1307 
1308   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1309   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1310   tmp  = a->solve_work;
1311 
1312   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1313   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1314 
1315   /* forward solve the lower triangular */
1316   tmp[0] = b[r[0]];
1317   v      = aa;
1318   vi     = aj;
1319   for (i=1; i<n; i++) {
1320     nz  = ai[i+1] - ai[i];
1321     sum = b[r[i]];
1322     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1323     tmp[i] = sum;
1324     v     += nz;
1325     vi    += nz;
1326   }
1327 
1328   /* backward solve the upper triangular */
1329   v  = aa + adiag[n-1];
1330   vi = aj + adiag[n-1];
1331   for (i=n-1; i>=0; i--) {
1332     nz  = adiag[i] - adiag[i+1] - 1;
1333     sum = tmp[i];
1334     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1335     tmp[i]   = sum*v[nz];
1336     x[c[i]] += tmp[i];
1337     v       += nz+1; vi += nz+1;
1338   }
1339 
1340   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1341   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1342   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1343   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1344   ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr);
1345   PetscFunctionReturn(0);
1346 }
1347 
1348 PetscErrorCode MatSolveTranspose_SeqAIJ_inplace(Mat A,Vec bb,Vec xx)
1349 {
1350   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1351   IS                iscol = a->col,isrow = a->row;
1352   PetscErrorCode    ierr;
1353   const PetscInt    *rout,*cout,*r,*c,*diag = a->diag,*ai = a->i,*aj = a->j,*vi;
1354   PetscInt          i,n = A->rmap->n,j;
1355   PetscInt          nz;
1356   PetscScalar       *x,*tmp,s1;
1357   const MatScalar   *aa = a->a,*v;
1358   const PetscScalar *b;
1359 
1360   PetscFunctionBegin;
1361   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1362   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1363   tmp  = a->solve_work;
1364 
1365   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1366   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1367 
1368   /* copy the b into temp work space according to permutation */
1369   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1370 
1371   /* forward solve the U^T */
1372   for (i=0; i<n; i++) {
1373     v   = aa + diag[i];
1374     vi  = aj + diag[i] + 1;
1375     nz  = ai[i+1] - diag[i] - 1;
1376     s1  = tmp[i];
1377     s1 *= (*v++);  /* multiply by inverse of diagonal entry */
1378     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1379     tmp[i] = s1;
1380   }
1381 
1382   /* backward solve the L^T */
1383   for (i=n-1; i>=0; i--) {
1384     v  = aa + diag[i] - 1;
1385     vi = aj + diag[i] - 1;
1386     nz = diag[i] - ai[i];
1387     s1 = tmp[i];
1388     for (j=0; j>-nz; j--) tmp[vi[j]] -= s1*v[j];
1389   }
1390 
1391   /* copy tmp into x according to permutation */
1392   for (i=0; i<n; i++) x[r[i]] = tmp[i];
1393 
1394   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1395   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1396   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1397   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1398 
1399   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1400   PetscFunctionReturn(0);
1401 }
1402 
1403 PetscErrorCode MatSolveTranspose_SeqAIJ(Mat A,Vec bb,Vec xx)
1404 {
1405   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1406   IS                iscol = a->col,isrow = a->row;
1407   PetscErrorCode    ierr;
1408   const PetscInt    *rout,*cout,*r,*c,*adiag = a->diag,*ai = a->i,*aj = a->j,*vi;
1409   PetscInt          i,n = A->rmap->n,j;
1410   PetscInt          nz;
1411   PetscScalar       *x,*tmp,s1;
1412   const MatScalar   *aa = a->a,*v;
1413   const PetscScalar *b;
1414 
1415   PetscFunctionBegin;
1416   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1417   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1418   tmp  = a->solve_work;
1419 
1420   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1421   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1422 
1423   /* copy the b into temp work space according to permutation */
1424   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1425 
1426   /* forward solve the U^T */
1427   for (i=0; i<n; i++) {
1428     v   = aa + adiag[i+1] + 1;
1429     vi  = aj + adiag[i+1] + 1;
1430     nz  = adiag[i] - adiag[i+1] - 1;
1431     s1  = tmp[i];
1432     s1 *= v[nz];  /* multiply by inverse of diagonal entry */
1433     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1434     tmp[i] = s1;
1435   }
1436 
1437   /* backward solve the L^T */
1438   for (i=n-1; i>=0; i--) {
1439     v  = aa + ai[i];
1440     vi = aj + ai[i];
1441     nz = ai[i+1] - ai[i];
1442     s1 = tmp[i];
1443     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1444   }
1445 
1446   /* copy tmp into x according to permutation */
1447   for (i=0; i<n; i++) x[r[i]] = tmp[i];
1448 
1449   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1450   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1451   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1452   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1453 
1454   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1455   PetscFunctionReturn(0);
1456 }
1457 
1458 PetscErrorCode MatSolveTransposeAdd_SeqAIJ_inplace(Mat A,Vec bb,Vec zz,Vec xx)
1459 {
1460   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1461   IS                iscol = a->col,isrow = a->row;
1462   PetscErrorCode    ierr;
1463   const PetscInt    *rout,*cout,*r,*c,*diag = a->diag,*ai = a->i,*aj = a->j,*vi;
1464   PetscInt          i,n = A->rmap->n,j;
1465   PetscInt          nz;
1466   PetscScalar       *x,*tmp,s1;
1467   const MatScalar   *aa = a->a,*v;
1468   const PetscScalar *b;
1469 
1470   PetscFunctionBegin;
1471   if (zz != xx) {ierr = VecCopy(zz,xx);CHKERRQ(ierr);}
1472   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1473   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1474   tmp  = a->solve_work;
1475 
1476   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1477   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1478 
1479   /* copy the b into temp work space according to permutation */
1480   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1481 
1482   /* forward solve the U^T */
1483   for (i=0; i<n; i++) {
1484     v   = aa + diag[i];
1485     vi  = aj + diag[i] + 1;
1486     nz  = ai[i+1] - diag[i] - 1;
1487     s1  = tmp[i];
1488     s1 *= (*v++);  /* multiply by inverse of diagonal entry */
1489     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1490     tmp[i] = s1;
1491   }
1492 
1493   /* backward solve the L^T */
1494   for (i=n-1; i>=0; i--) {
1495     v  = aa + diag[i] - 1;
1496     vi = aj + diag[i] - 1;
1497     nz = diag[i] - ai[i];
1498     s1 = tmp[i];
1499     for (j=0; j>-nz; j--) tmp[vi[j]] -= s1*v[j];
1500   }
1501 
1502   /* copy tmp into x according to permutation */
1503   for (i=0; i<n; i++) x[r[i]] += tmp[i];
1504 
1505   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1506   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1507   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1508   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1509 
1510   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1511   PetscFunctionReturn(0);
1512 }
1513 
1514 PetscErrorCode MatSolveTransposeAdd_SeqAIJ(Mat A,Vec bb,Vec zz,Vec xx)
1515 {
1516   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1517   IS                iscol = a->col,isrow = a->row;
1518   PetscErrorCode    ierr;
1519   const PetscInt    *rout,*cout,*r,*c,*adiag = a->diag,*ai = a->i,*aj = a->j,*vi;
1520   PetscInt          i,n = A->rmap->n,j;
1521   PetscInt          nz;
1522   PetscScalar       *x,*tmp,s1;
1523   const MatScalar   *aa = a->a,*v;
1524   const PetscScalar *b;
1525 
1526   PetscFunctionBegin;
1527   if (zz != xx) {ierr = VecCopy(zz,xx);CHKERRQ(ierr);}
1528   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1529   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1530   tmp  = a->solve_work;
1531 
1532   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1533   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1534 
1535   /* copy the b into temp work space according to permutation */
1536   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1537 
1538   /* forward solve the U^T */
1539   for (i=0; i<n; i++) {
1540     v   = aa + adiag[i+1] + 1;
1541     vi  = aj + adiag[i+1] + 1;
1542     nz  = adiag[i] - adiag[i+1] - 1;
1543     s1  = tmp[i];
1544     s1 *= v[nz];  /* multiply by inverse of diagonal entry */
1545     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1546     tmp[i] = s1;
1547   }
1548 
1549 
1550   /* backward solve the L^T */
1551   for (i=n-1; i>=0; i--) {
1552     v  = aa + ai[i];
1553     vi = aj + ai[i];
1554     nz = ai[i+1] - ai[i];
1555     s1 = tmp[i];
1556     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1557   }
1558 
1559   /* copy tmp into x according to permutation */
1560   for (i=0; i<n; i++) x[r[i]] += tmp[i];
1561 
1562   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1563   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1564   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1565   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1566 
1567   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1568   PetscFunctionReturn(0);
1569 }
1570 
1571 /* ----------------------------------------------------------------*/
1572 
1573 /*
1574    ilu() under revised new data structure.
1575    Factored arrays bj and ba are stored as
1576      L(0,:), L(1,:), ...,L(n-1,:),  U(n-1,:),...,U(i,:),U(i-1,:),...,U(0,:)
1577 
1578    bi=fact->i is an array of size n+1, in which
1579    bi+
1580      bi[i]:  points to 1st entry of L(i,:),i=0,...,n-1
1581      bi[n]:  points to L(n-1,n-1)+1
1582 
1583   bdiag=fact->diag is an array of size n+1,in which
1584      bdiag[i]: points to diagonal of U(i,:), i=0,...,n-1
1585      bdiag[n]: points to entry of U(n-1,0)-1
1586 
1587    U(i,:) contains bdiag[i] as its last entry, i.e.,
1588     U(i,:) = (u[i,i+1],...,u[i,n-1],diag[i])
1589 */
1590 PetscErrorCode MatILUFactorSymbolic_SeqAIJ_ilu0(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1591 {
1592   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data,*b;
1593   PetscErrorCode ierr;
1594   const PetscInt n=A->rmap->n,*ai=a->i,*aj,*adiag=a->diag;
1595   PetscInt       i,j,k=0,nz,*bi,*bj,*bdiag;
1596   IS             isicol;
1597 
1598   PetscFunctionBegin;
1599   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1600   ierr = MatDuplicateNoCreate_SeqAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_FALSE);CHKERRQ(ierr);
1601   b    = (Mat_SeqAIJ*)(fact)->data;
1602 
1603   /* allocate matrix arrays for new data structure */
1604   ierr = PetscMalloc3(ai[n]+1,&b->a,ai[n]+1,&b->j,n+1,&b->i);CHKERRQ(ierr);
1605   ierr = PetscLogObjectMemory((PetscObject)fact,ai[n]*(sizeof(PetscScalar)+sizeof(PetscInt))+(n+1)*sizeof(PetscInt));CHKERRQ(ierr);
1606 
1607   b->singlemalloc = PETSC_TRUE;
1608   if (!b->diag) {
1609     ierr = PetscMalloc1(n+1,&b->diag);CHKERRQ(ierr);
1610     ierr = PetscLogObjectMemory((PetscObject)fact,(n+1)*sizeof(PetscInt));CHKERRQ(ierr);
1611   }
1612   bdiag = b->diag;
1613 
1614   if (n > 0) {
1615     ierr = PetscMemzero(b->a,(ai[n])*sizeof(MatScalar));CHKERRQ(ierr);
1616   }
1617 
1618   /* set bi and bj with new data structure */
1619   bi = b->i;
1620   bj = b->j;
1621 
1622   /* L part */
1623   bi[0] = 0;
1624   for (i=0; i<n; i++) {
1625     nz      = adiag[i] - ai[i];
1626     bi[i+1] = bi[i] + nz;
1627     aj      = a->j + ai[i];
1628     for (j=0; j<nz; j++) {
1629       /*   *bj = aj[j]; bj++; */
1630       bj[k++] = aj[j];
1631     }
1632   }
1633 
1634   /* U part */
1635   bdiag[n] = bi[n]-1;
1636   for (i=n-1; i>=0; i--) {
1637     nz = ai[i+1] - adiag[i] - 1;
1638     aj = a->j + adiag[i] + 1;
1639     for (j=0; j<nz; j++) {
1640       /*      *bj = aj[j]; bj++; */
1641       bj[k++] = aj[j];
1642     }
1643     /* diag[i] */
1644     /*    *bj = i; bj++; */
1645     bj[k++]  = i;
1646     bdiag[i] = bdiag[i+1] + nz + 1;
1647   }
1648 
1649   fact->factortype             = MAT_FACTOR_ILU;
1650   fact->info.factor_mallocs    = 0;
1651   fact->info.fill_ratio_given  = info->fill;
1652   fact->info.fill_ratio_needed = 1.0;
1653   fact->ops->lufactornumeric   = MatLUFactorNumeric_SeqAIJ;
1654   ierr = MatSeqAIJCheckInode_FactorLU(fact);CHKERRQ(ierr);
1655 
1656   b       = (Mat_SeqAIJ*)(fact)->data;
1657   b->row  = isrow;
1658   b->col  = iscol;
1659   b->icol = isicol;
1660   ierr    = PetscMalloc1(fact->rmap->n+1,&b->solve_work);CHKERRQ(ierr);
1661   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1662   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1663   PetscFunctionReturn(0);
1664 }
1665 
1666 PetscErrorCode MatILUFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1667 {
1668   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
1669   IS                 isicol;
1670   PetscErrorCode     ierr;
1671   const PetscInt     *r,*ic;
1672   PetscInt           n=A->rmap->n,*ai=a->i,*aj=a->j;
1673   PetscInt           *bi,*cols,nnz,*cols_lvl;
1674   PetscInt           *bdiag,prow,fm,nzbd,reallocs=0,dcount=0;
1675   PetscInt           i,levels,diagonal_fill;
1676   PetscBool          col_identity,row_identity,missing;
1677   PetscReal          f;
1678   PetscInt           nlnk,*lnk,*lnk_lvl=NULL;
1679   PetscBT            lnkbt;
1680   PetscInt           nzi,*bj,**bj_ptr,**bjlvl_ptr;
1681   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
1682   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
1683 
1684   PetscFunctionBegin;
1685   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
1686   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
1687   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
1688 
1689   levels = (PetscInt)info->levels;
1690   ierr   = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
1691   ierr   = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);
1692   if (!levels && row_identity && col_identity) {
1693     /* special case: ilu(0) with natural ordering */
1694     ierr = MatILUFactorSymbolic_SeqAIJ_ilu0(fact,A,isrow,iscol,info);CHKERRQ(ierr);
1695     if (a->inode.size) {
1696       fact->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
1697     }
1698     PetscFunctionReturn(0);
1699   }
1700 
1701   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1702   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
1703   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
1704 
1705   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
1706   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
1707   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
1708   bi[0] = bdiag[0] = 0;
1709   ierr  = PetscMalloc2(n,&bj_ptr,n,&bjlvl_ptr);CHKERRQ(ierr);
1710 
1711   /* create a linked list for storing column indices of the active row */
1712   nlnk = n + 1;
1713   ierr = PetscIncompleteLLCreate(n,n,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1714 
1715   /* initial FreeSpace size is f*(ai[n]+1) */
1716   f                 = info->fill;
1717   diagonal_fill     = (PetscInt)info->diagonal_fill;
1718   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
1719   current_space     = free_space;
1720   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space_lvl);CHKERRQ(ierr);
1721   current_space_lvl = free_space_lvl;
1722   for (i=0; i<n; i++) {
1723     nzi = 0;
1724     /* copy current row into linked list */
1725     nnz = ai[r[i]+1] - ai[r[i]];
1726     if (!nnz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",r[i],i);
1727     cols   = aj + ai[r[i]];
1728     lnk[i] = -1; /* marker to indicate if diagonal exists */
1729     ierr   = PetscIncompleteLLInit(nnz,cols,n,ic,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1730     nzi   += nlnk;
1731 
1732     /* make sure diagonal entry is included */
1733     if (diagonal_fill && lnk[i] == -1) {
1734       fm = n;
1735       while (lnk[fm] < i) fm = lnk[fm];
1736       lnk[i]     = lnk[fm]; /* insert diagonal into linked list */
1737       lnk[fm]    = i;
1738       lnk_lvl[i] = 0;
1739       nzi++; dcount++;
1740     }
1741 
1742     /* add pivot rows into the active row */
1743     nzbd = 0;
1744     prow = lnk[n];
1745     while (prow < i) {
1746       nnz      = bdiag[prow];
1747       cols     = bj_ptr[prow] + nnz + 1;
1748       cols_lvl = bjlvl_ptr[prow] + nnz + 1;
1749       nnz      = bi[prow+1] - bi[prow] - nnz - 1;
1750       ierr     = PetscILULLAddSorted(nnz,cols,levels,cols_lvl,prow,nlnk,lnk,lnk_lvl,lnkbt,prow);CHKERRQ(ierr);
1751       nzi     += nlnk;
1752       prow     = lnk[prow];
1753       nzbd++;
1754     }
1755     bdiag[i] = nzbd;
1756     bi[i+1]  = bi[i] + nzi;
1757     /* if free space is not available, make more free space */
1758     if (current_space->local_remaining<nzi) {
1759       nnz  = PetscIntMultTruncate(2,PetscIntMultTruncate(nzi,n - i)); /* estimated and max additional space needed */
1760       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
1761       ierr = PetscFreeSpaceGet(nnz,&current_space_lvl);CHKERRQ(ierr);
1762       reallocs++;
1763     }
1764 
1765     /* copy data into free_space and free_space_lvl, then initialize lnk */
1766     ierr         = PetscIncompleteLLClean(n,n,nzi,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
1767     bj_ptr[i]    = current_space->array;
1768     bjlvl_ptr[i] = current_space_lvl->array;
1769 
1770     /* make sure the active row i has diagonal entry */
1771     if (*(bj_ptr[i]+bdiag[i]) != i) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Row %D has missing diagonal in factored matrix\ntry running with -pc_factor_nonzeros_along_diagonal or -pc_factor_diagonal_fill",i);
1772 
1773     current_space->array               += nzi;
1774     current_space->local_used          += nzi;
1775     current_space->local_remaining     -= nzi;
1776     current_space_lvl->array           += nzi;
1777     current_space_lvl->local_used      += nzi;
1778     current_space_lvl->local_remaining -= nzi;
1779   }
1780 
1781   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
1782   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
1783   /* copy free_space into bj and free free_space; set bi, bj, bdiag in new datastructure; */
1784   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
1785   ierr = PetscFreeSpaceContiguous_LU(&free_space,bj,n,bi,bdiag);CHKERRQ(ierr);
1786 
1787   ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
1788   ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
1789   ierr = PetscFree2(bj_ptr,bjlvl_ptr);CHKERRQ(ierr);
1790 
1791 #if defined(PETSC_USE_INFO)
1792   {
1793     PetscReal af = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
1794     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
1795     ierr = PetscInfo1(A,"Run with -[sub_]pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
1796     ierr = PetscInfo1(A,"PCFactorSetFill([sub]pc,%g);\n",(double)af);CHKERRQ(ierr);
1797     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
1798     if (diagonal_fill) {
1799       ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals\n",dcount);CHKERRQ(ierr);
1800     }
1801   }
1802 #endif
1803   /* put together the new matrix */
1804   ierr = MatSeqAIJSetPreallocation_SeqAIJ(fact,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
1805   ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);
1806   b    = (Mat_SeqAIJ*)(fact)->data;
1807 
1808   b->free_a       = PETSC_TRUE;
1809   b->free_ij      = PETSC_TRUE;
1810   b->singlemalloc = PETSC_FALSE;
1811 
1812   ierr = PetscMalloc1(bdiag[0]+1,&b->a);CHKERRQ(ierr);
1813 
1814   b->j    = bj;
1815   b->i    = bi;
1816   b->diag = bdiag;
1817   b->ilen = 0;
1818   b->imax = 0;
1819   b->row  = isrow;
1820   b->col  = iscol;
1821   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1822   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1823   b->icol = isicol;
1824 
1825   ierr = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
1826   /* In b structure:  Free imax, ilen, old a, old j.
1827      Allocate bdiag, solve_work, new a, new j */
1828   ierr     = PetscLogObjectMemory((PetscObject)fact,(bdiag[0]+1)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
1829   b->maxnz = b->nz = bdiag[0]+1;
1830 
1831   (fact)->info.factor_mallocs    = reallocs;
1832   (fact)->info.fill_ratio_given  = f;
1833   (fact)->info.fill_ratio_needed = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
1834   (fact)->ops->lufactornumeric   = MatLUFactorNumeric_SeqAIJ;
1835   if (a->inode.size) {
1836     (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
1837   }
1838   ierr = MatSeqAIJCheckInode_FactorLU(fact);CHKERRQ(ierr);
1839   PetscFunctionReturn(0);
1840 }
1841 
1842 PetscErrorCode MatILUFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1843 {
1844   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
1845   IS                 isicol;
1846   PetscErrorCode     ierr;
1847   const PetscInt     *r,*ic;
1848   PetscInt           n=A->rmap->n,*ai=a->i,*aj=a->j;
1849   PetscInt           *bi,*cols,nnz,*cols_lvl;
1850   PetscInt           *bdiag,prow,fm,nzbd,reallocs=0,dcount=0;
1851   PetscInt           i,levels,diagonal_fill;
1852   PetscBool          col_identity,row_identity;
1853   PetscReal          f;
1854   PetscInt           nlnk,*lnk,*lnk_lvl=NULL;
1855   PetscBT            lnkbt;
1856   PetscInt           nzi,*bj,**bj_ptr,**bjlvl_ptr;
1857   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
1858   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
1859   PetscBool          missing;
1860 
1861   PetscFunctionBegin;
1862   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
1863   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
1864   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
1865 
1866   f             = info->fill;
1867   levels        = (PetscInt)info->levels;
1868   diagonal_fill = (PetscInt)info->diagonal_fill;
1869 
1870   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1871 
1872   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
1873   ierr = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);
1874   if (!levels && row_identity && col_identity) { /* special case: ilu(0) with natural ordering */
1875     ierr = MatDuplicateNoCreate_SeqAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_TRUE);CHKERRQ(ierr);
1876 
1877     (fact)->ops->lufactornumeric =  MatLUFactorNumeric_SeqAIJ_inplace;
1878     if (a->inode.size) {
1879       (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
1880     }
1881     fact->factortype               = MAT_FACTOR_ILU;
1882     (fact)->info.factor_mallocs    = 0;
1883     (fact)->info.fill_ratio_given  = info->fill;
1884     (fact)->info.fill_ratio_needed = 1.0;
1885 
1886     b    = (Mat_SeqAIJ*)(fact)->data;
1887     b->row  = isrow;
1888     b->col  = iscol;
1889     b->icol = isicol;
1890     ierr    = PetscMalloc1((fact)->rmap->n+1,&b->solve_work);CHKERRQ(ierr);
1891     ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1892     ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1893     PetscFunctionReturn(0);
1894   }
1895 
1896   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
1897   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
1898 
1899   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
1900   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
1901   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
1902   bi[0] = bdiag[0] = 0;
1903 
1904   ierr = PetscMalloc2(n,&bj_ptr,n,&bjlvl_ptr);CHKERRQ(ierr);
1905 
1906   /* create a linked list for storing column indices of the active row */
1907   nlnk = n + 1;
1908   ierr = PetscIncompleteLLCreate(n,n,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1909 
1910   /* initial FreeSpace size is f*(ai[n]+1) */
1911   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
1912   current_space     = free_space;
1913   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space_lvl);CHKERRQ(ierr);
1914   current_space_lvl = free_space_lvl;
1915 
1916   for (i=0; i<n; i++) {
1917     nzi = 0;
1918     /* copy current row into linked list */
1919     nnz = ai[r[i]+1] - ai[r[i]];
1920     if (!nnz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",r[i],i);
1921     cols   = aj + ai[r[i]];
1922     lnk[i] = -1; /* marker to indicate if diagonal exists */
1923     ierr   = PetscIncompleteLLInit(nnz,cols,n,ic,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1924     nzi   += nlnk;
1925 
1926     /* make sure diagonal entry is included */
1927     if (diagonal_fill && lnk[i] == -1) {
1928       fm = n;
1929       while (lnk[fm] < i) fm = lnk[fm];
1930       lnk[i]     = lnk[fm]; /* insert diagonal into linked list */
1931       lnk[fm]    = i;
1932       lnk_lvl[i] = 0;
1933       nzi++; dcount++;
1934     }
1935 
1936     /* add pivot rows into the active row */
1937     nzbd = 0;
1938     prow = lnk[n];
1939     while (prow < i) {
1940       nnz      = bdiag[prow];
1941       cols     = bj_ptr[prow] + nnz + 1;
1942       cols_lvl = bjlvl_ptr[prow] + nnz + 1;
1943       nnz      = bi[prow+1] - bi[prow] - nnz - 1;
1944       ierr     = PetscILULLAddSorted(nnz,cols,levels,cols_lvl,prow,nlnk,lnk,lnk_lvl,lnkbt,prow);CHKERRQ(ierr);
1945       nzi     += nlnk;
1946       prow     = lnk[prow];
1947       nzbd++;
1948     }
1949     bdiag[i] = nzbd;
1950     bi[i+1]  = bi[i] + nzi;
1951 
1952     /* if free space is not available, make more free space */
1953     if (current_space->local_remaining<nzi) {
1954       nnz  = PetscIntMultTruncate(nzi,n - i); /* estimated and max additional space needed */
1955       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
1956       ierr = PetscFreeSpaceGet(nnz,&current_space_lvl);CHKERRQ(ierr);
1957       reallocs++;
1958     }
1959 
1960     /* copy data into free_space and free_space_lvl, then initialize lnk */
1961     ierr         = PetscIncompleteLLClean(n,n,nzi,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
1962     bj_ptr[i]    = current_space->array;
1963     bjlvl_ptr[i] = current_space_lvl->array;
1964 
1965     /* make sure the active row i has diagonal entry */
1966     if (*(bj_ptr[i]+bdiag[i]) != i) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Row %D has missing diagonal in factored matrix\ntry running with -pc_factor_nonzeros_along_diagonal or -pc_factor_diagonal_fill",i);
1967 
1968     current_space->array               += nzi;
1969     current_space->local_used          += nzi;
1970     current_space->local_remaining     -= nzi;
1971     current_space_lvl->array           += nzi;
1972     current_space_lvl->local_used      += nzi;
1973     current_space_lvl->local_remaining -= nzi;
1974   }
1975 
1976   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
1977   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
1978 
1979   /* destroy list of free space and other temporary arrays */
1980   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
1981   ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr); /* copy free_space -> bj */
1982   ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
1983   ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
1984   ierr = PetscFree2(bj_ptr,bjlvl_ptr);CHKERRQ(ierr);
1985 
1986 #if defined(PETSC_USE_INFO)
1987   {
1988     PetscReal af = ((PetscReal)bi[n])/((PetscReal)ai[n]);
1989     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
1990     ierr = PetscInfo1(A,"Run with -[sub_]pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
1991     ierr = PetscInfo1(A,"PCFactorSetFill([sub]pc,%g);\n",(double)af);CHKERRQ(ierr);
1992     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
1993     if (diagonal_fill) {
1994       ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals\n",dcount);CHKERRQ(ierr);
1995     }
1996   }
1997 #endif
1998 
1999   /* put together the new matrix */
2000   ierr = MatSeqAIJSetPreallocation_SeqAIJ(fact,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
2001   ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);
2002   b    = (Mat_SeqAIJ*)(fact)->data;
2003 
2004   b->free_a       = PETSC_TRUE;
2005   b->free_ij      = PETSC_TRUE;
2006   b->singlemalloc = PETSC_FALSE;
2007 
2008   ierr = PetscMalloc1(bi[n],&b->a);CHKERRQ(ierr);
2009   b->j = bj;
2010   b->i = bi;
2011   for (i=0; i<n; i++) bdiag[i] += bi[i];
2012   b->diag = bdiag;
2013   b->ilen = 0;
2014   b->imax = 0;
2015   b->row  = isrow;
2016   b->col  = iscol;
2017   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
2018   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
2019   b->icol = isicol;
2020   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
2021   /* In b structure:  Free imax, ilen, old a, old j.
2022      Allocate bdiag, solve_work, new a, new j */
2023   ierr     = PetscLogObjectMemory((PetscObject)fact,(bi[n]-n) * (sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
2024   b->maxnz = b->nz = bi[n];
2025 
2026   (fact)->info.factor_mallocs    = reallocs;
2027   (fact)->info.fill_ratio_given  = f;
2028   (fact)->info.fill_ratio_needed = ((PetscReal)bi[n])/((PetscReal)ai[n]);
2029   (fact)->ops->lufactornumeric   =  MatLUFactorNumeric_SeqAIJ_inplace;
2030   if (a->inode.size) {
2031     (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
2032   }
2033   PetscFunctionReturn(0);
2034 }
2035 
2036 PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ(Mat B,Mat A,const MatFactorInfo *info)
2037 {
2038   Mat            C = B;
2039   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data;
2040   Mat_SeqSBAIJ   *b=(Mat_SeqSBAIJ*)C->data;
2041   IS             ip=b->row,iip = b->icol;
2042   PetscErrorCode ierr;
2043   const PetscInt *rip,*riip;
2044   PetscInt       i,j,mbs=A->rmap->n,*bi=b->i,*bj=b->j,*bdiag=b->diag,*bjtmp;
2045   PetscInt       *ai=a->i,*aj=a->j;
2046   PetscInt       k,jmin,jmax,*c2r,*il,col,nexti,ili,nz;
2047   MatScalar      *rtmp,*ba=b->a,*bval,*aa=a->a,dk,uikdi;
2048   PetscBool      perm_identity;
2049   FactorShiftCtx sctx;
2050   PetscReal      rs;
2051   MatScalar      d,*v;
2052 
2053   PetscFunctionBegin;
2054   /* MatPivotSetUp(): initialize shift context sctx */
2055   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
2056 
2057   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
2058     sctx.shift_top = info->zeropivot;
2059     for (i=0; i<mbs; i++) {
2060       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
2061       d  = (aa)[a->diag[i]];
2062       rs = -PetscAbsScalar(d) - PetscRealPart(d);
2063       v  = aa+ai[i];
2064       nz = ai[i+1] - ai[i];
2065       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
2066       if (rs>sctx.shift_top) sctx.shift_top = rs;
2067     }
2068     sctx.shift_top *= 1.1;
2069     sctx.nshift_max = 5;
2070     sctx.shift_lo   = 0.;
2071     sctx.shift_hi   = 1.;
2072   }
2073 
2074   ierr = ISGetIndices(ip,&rip);CHKERRQ(ierr);
2075   ierr = ISGetIndices(iip,&riip);CHKERRQ(ierr);
2076 
2077   /* allocate working arrays
2078      c2r: linked list, keep track of pivot rows for a given column. c2r[col]: head of the list for a given col
2079      il:  for active k row, il[i] gives the index of the 1st nonzero entry in U[i,k:n-1] in bj and ba arrays
2080   */
2081   ierr = PetscMalloc3(mbs,&rtmp,mbs,&il,mbs,&c2r);CHKERRQ(ierr);
2082 
2083   do {
2084     sctx.newshift = PETSC_FALSE;
2085 
2086     for (i=0; i<mbs; i++) c2r[i] = mbs;
2087     if (mbs) il[0] = 0;
2088 
2089     for (k = 0; k<mbs; k++) {
2090       /* zero rtmp */
2091       nz    = bi[k+1] - bi[k];
2092       bjtmp = bj + bi[k];
2093       for (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
2094 
2095       /* load in initial unfactored row */
2096       bval = ba + bi[k];
2097       jmin = ai[rip[k]]; jmax = ai[rip[k]+1];
2098       for (j = jmin; j < jmax; j++) {
2099         col = riip[aj[j]];
2100         if (col >= k) { /* only take upper triangular entry */
2101           rtmp[col] = aa[j];
2102           *bval++   = 0.0; /* for in-place factorization */
2103         }
2104       }
2105       /* shift the diagonal of the matrix: ZeropivotApply() */
2106       rtmp[k] += sctx.shift_amount;  /* shift the diagonal of the matrix */
2107 
2108       /* modify k-th row by adding in those rows i with U(i,k)!=0 */
2109       dk = rtmp[k];
2110       i  = c2r[k]; /* first row to be added to k_th row  */
2111 
2112       while (i < k) {
2113         nexti = c2r[i]; /* next row to be added to k_th row */
2114 
2115         /* compute multiplier, update diag(k) and U(i,k) */
2116         ili     = il[i]; /* index of first nonzero element in U(i,k:bms-1) */
2117         uikdi   = -ba[ili]*ba[bdiag[i]]; /* diagonal(k) */
2118         dk     += uikdi*ba[ili]; /* update diag[k] */
2119         ba[ili] = uikdi; /* -U(i,k) */
2120 
2121         /* add multiple of row i to k-th row */
2122         jmin = ili + 1; jmax = bi[i+1];
2123         if (jmin < jmax) {
2124           for (j=jmin; j<jmax; j++) rtmp[bj[j]] += uikdi*ba[j];
2125           /* update il and c2r for row i */
2126           il[i] = jmin;
2127           j     = bj[jmin]; c2r[i] = c2r[j]; c2r[j] = i;
2128         }
2129         i = nexti;
2130       }
2131 
2132       /* copy data into U(k,:) */
2133       rs   = 0.0;
2134       jmin = bi[k]; jmax = bi[k+1]-1;
2135       if (jmin < jmax) {
2136         for (j=jmin; j<jmax; j++) {
2137           col = bj[j]; ba[j] = rtmp[col]; rs += PetscAbsScalar(ba[j]);
2138         }
2139         /* add the k-th row into il and c2r */
2140         il[k] = jmin;
2141         i     = bj[jmin]; c2r[k] = c2r[i]; c2r[i] = k;
2142       }
2143 
2144       /* MatPivotCheck() */
2145       sctx.rs = rs;
2146       sctx.pv = dk;
2147       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
2148       if (sctx.newshift) break;
2149       dk = sctx.pv;
2150 
2151       ba[bdiag[k]] = 1.0/dk; /* U(k,k) */
2152     }
2153   } while (sctx.newshift);
2154 
2155   ierr = PetscFree3(rtmp,il,c2r);CHKERRQ(ierr);
2156   ierr = ISRestoreIndices(ip,&rip);CHKERRQ(ierr);
2157   ierr = ISRestoreIndices(iip,&riip);CHKERRQ(ierr);
2158 
2159   ierr = ISIdentity(ip,&perm_identity);CHKERRQ(ierr);
2160   if (perm_identity) {
2161     B->ops->solve          = MatSolve_SeqSBAIJ_1_NaturalOrdering;
2162     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_NaturalOrdering;
2163     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_NaturalOrdering;
2164     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_NaturalOrdering;
2165   } else {
2166     B->ops->solve          = MatSolve_SeqSBAIJ_1;
2167     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1;
2168     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1;
2169     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1;
2170   }
2171 
2172   C->assembled    = PETSC_TRUE;
2173   C->preallocated = PETSC_TRUE;
2174 
2175   ierr = PetscLogFlops(C->rmap->n);CHKERRQ(ierr);
2176 
2177   /* MatPivotView() */
2178   if (sctx.nshift) {
2179     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
2180       ierr = PetscInfo4(A,"number of shift_pd tries %D, shift_amount %g, diagonal shifted up by %e fraction top_value %e\n",sctx.nshift,(double)sctx.shift_amount,(double)sctx.shift_fraction,(double)sctx.shift_top);CHKERRQ(ierr);
2181     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
2182       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2183     } else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) {
2184       ierr = PetscInfo2(A,"number of shift_inblocks applied %D, each shift_amount %g\n",sctx.nshift,(double)info->shiftamount);CHKERRQ(ierr);
2185     }
2186   }
2187   PetscFunctionReturn(0);
2188 }
2189 
2190 PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ_inplace(Mat B,Mat A,const MatFactorInfo *info)
2191 {
2192   Mat            C = B;
2193   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data;
2194   Mat_SeqSBAIJ   *b=(Mat_SeqSBAIJ*)C->data;
2195   IS             ip=b->row,iip = b->icol;
2196   PetscErrorCode ierr;
2197   const PetscInt *rip,*riip;
2198   PetscInt       i,j,mbs=A->rmap->n,*bi=b->i,*bj=b->j,*bcol,*bjtmp;
2199   PetscInt       *ai=a->i,*aj=a->j;
2200   PetscInt       k,jmin,jmax,*jl,*il,col,nexti,ili,nz;
2201   MatScalar      *rtmp,*ba=b->a,*bval,*aa=a->a,dk,uikdi;
2202   PetscBool      perm_identity;
2203   FactorShiftCtx sctx;
2204   PetscReal      rs;
2205   MatScalar      d,*v;
2206 
2207   PetscFunctionBegin;
2208   /* MatPivotSetUp(): initialize shift context sctx */
2209   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
2210 
2211   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
2212     sctx.shift_top = info->zeropivot;
2213     for (i=0; i<mbs; i++) {
2214       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
2215       d  = (aa)[a->diag[i]];
2216       rs = -PetscAbsScalar(d) - PetscRealPart(d);
2217       v  = aa+ai[i];
2218       nz = ai[i+1] - ai[i];
2219       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
2220       if (rs>sctx.shift_top) sctx.shift_top = rs;
2221     }
2222     sctx.shift_top *= 1.1;
2223     sctx.nshift_max = 5;
2224     sctx.shift_lo   = 0.;
2225     sctx.shift_hi   = 1.;
2226   }
2227 
2228   ierr = ISGetIndices(ip,&rip);CHKERRQ(ierr);
2229   ierr = ISGetIndices(iip,&riip);CHKERRQ(ierr);
2230 
2231   /* initialization */
2232   ierr = PetscMalloc3(mbs,&rtmp,mbs,&il,mbs,&jl);CHKERRQ(ierr);
2233 
2234   do {
2235     sctx.newshift = PETSC_FALSE;
2236 
2237     for (i=0; i<mbs; i++) jl[i] = mbs;
2238     il[0] = 0;
2239 
2240     for (k = 0; k<mbs; k++) {
2241       /* zero rtmp */
2242       nz    = bi[k+1] - bi[k];
2243       bjtmp = bj + bi[k];
2244       for (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
2245 
2246       bval = ba + bi[k];
2247       /* initialize k-th row by the perm[k]-th row of A */
2248       jmin = ai[rip[k]]; jmax = ai[rip[k]+1];
2249       for (j = jmin; j < jmax; j++) {
2250         col = riip[aj[j]];
2251         if (col >= k) { /* only take upper triangular entry */
2252           rtmp[col] = aa[j];
2253           *bval++   = 0.0; /* for in-place factorization */
2254         }
2255       }
2256       /* shift the diagonal of the matrix */
2257       if (sctx.nshift) rtmp[k] += sctx.shift_amount;
2258 
2259       /* modify k-th row by adding in those rows i with U(i,k)!=0 */
2260       dk = rtmp[k];
2261       i  = jl[k]; /* first row to be added to k_th row  */
2262 
2263       while (i < k) {
2264         nexti = jl[i]; /* next row to be added to k_th row */
2265 
2266         /* compute multiplier, update diag(k) and U(i,k) */
2267         ili     = il[i]; /* index of first nonzero element in U(i,k:bms-1) */
2268         uikdi   = -ba[ili]*ba[bi[i]]; /* diagonal(k) */
2269         dk     += uikdi*ba[ili];
2270         ba[ili] = uikdi; /* -U(i,k) */
2271 
2272         /* add multiple of row i to k-th row */
2273         jmin = ili + 1; jmax = bi[i+1];
2274         if (jmin < jmax) {
2275           for (j=jmin; j<jmax; j++) rtmp[bj[j]] += uikdi*ba[j];
2276           /* update il and jl for row i */
2277           il[i] = jmin;
2278           j     = bj[jmin]; jl[i] = jl[j]; jl[j] = i;
2279         }
2280         i = nexti;
2281       }
2282 
2283       /* shift the diagonals when zero pivot is detected */
2284       /* compute rs=sum of abs(off-diagonal) */
2285       rs   = 0.0;
2286       jmin = bi[k]+1;
2287       nz   = bi[k+1] - jmin;
2288       bcol = bj + jmin;
2289       for (j=0; j<nz; j++) {
2290         rs += PetscAbsScalar(rtmp[bcol[j]]);
2291       }
2292 
2293       sctx.rs = rs;
2294       sctx.pv = dk;
2295       ierr    = MatPivotCheck(B,A,info,&sctx,k);CHKERRQ(ierr);
2296       if (sctx.newshift) break;
2297       dk = sctx.pv;
2298 
2299       /* copy data into U(k,:) */
2300       ba[bi[k]] = 1.0/dk; /* U(k,k) */
2301       jmin      = bi[k]+1; jmax = bi[k+1];
2302       if (jmin < jmax) {
2303         for (j=jmin; j<jmax; j++) {
2304           col = bj[j]; ba[j] = rtmp[col];
2305         }
2306         /* add the k-th row into il and jl */
2307         il[k] = jmin;
2308         i     = bj[jmin]; jl[k] = jl[i]; jl[i] = k;
2309       }
2310     }
2311   } while (sctx.newshift);
2312 
2313   ierr = PetscFree3(rtmp,il,jl);CHKERRQ(ierr);
2314   ierr = ISRestoreIndices(ip,&rip);CHKERRQ(ierr);
2315   ierr = ISRestoreIndices(iip,&riip);CHKERRQ(ierr);
2316 
2317   ierr = ISIdentity(ip,&perm_identity);CHKERRQ(ierr);
2318   if (perm_identity) {
2319     B->ops->solve          = MatSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2320     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2321     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2322     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2323   } else {
2324     B->ops->solve          = MatSolve_SeqSBAIJ_1_inplace;
2325     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_inplace;
2326     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_inplace;
2327     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_inplace;
2328   }
2329 
2330   C->assembled    = PETSC_TRUE;
2331   C->preallocated = PETSC_TRUE;
2332 
2333   ierr = PetscLogFlops(C->rmap->n);CHKERRQ(ierr);
2334   if (sctx.nshift) {
2335     if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
2336       ierr = PetscInfo2(A,"number of shiftnz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2337     } else if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
2338       ierr = PetscInfo2(A,"number of shiftpd tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2339     }
2340   }
2341   PetscFunctionReturn(0);
2342 }
2343 
2344 /*
2345    icc() under revised new data structure.
2346    Factored arrays bj and ba are stored as
2347      U(0,:),...,U(i,:),U(n-1,:)
2348 
2349    ui=fact->i is an array of size n+1, in which
2350    ui+
2351      ui[i]:  points to 1st entry of U(i,:),i=0,...,n-1
2352      ui[n]:  points to U(n-1,n-1)+1
2353 
2354   udiag=fact->diag is an array of size n,in which
2355      udiag[i]: points to diagonal of U(i,:), i=0,...,n-1
2356 
2357    U(i,:) contains udiag[i] as its last entry, i.e.,
2358     U(i,:) = (u[i,i+1],...,u[i,n-1],diag[i])
2359 */
2360 
2361 PetscErrorCode MatICCFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2362 {
2363   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2364   Mat_SeqSBAIJ       *b;
2365   PetscErrorCode     ierr;
2366   PetscBool          perm_identity,missing;
2367   PetscInt           reallocs=0,i,*ai=a->i,*aj=a->j,am=A->rmap->n,*ui,*udiag;
2368   const PetscInt     *rip,*riip;
2369   PetscInt           jmin,jmax,nzk,k,j,*jl,prow,*il,nextprow;
2370   PetscInt           nlnk,*lnk,*lnk_lvl=NULL,d;
2371   PetscInt           ncols,ncols_upper,*cols,*ajtmp,*uj,**uj_ptr,**uj_lvl_ptr;
2372   PetscReal          fill          =info->fill,levels=info->levels;
2373   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
2374   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
2375   PetscBT            lnkbt;
2376   IS                 iperm;
2377 
2378   PetscFunctionBegin;
2379   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
2380   ierr = MatMissingDiagonal(A,&missing,&d);CHKERRQ(ierr);
2381   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",d);
2382   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2383   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2384 
2385   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2386   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2387   ui[0] = 0;
2388 
2389   /* ICC(0) without matrix ordering: simply rearrange column indices */
2390   if (!levels && perm_identity) {
2391     for (i=0; i<am; i++) {
2392       ncols    = ai[i+1] - a->diag[i];
2393       ui[i+1]  = ui[i] + ncols;
2394       udiag[i] = ui[i+1] - 1; /* points to the last entry of U(i,:) */
2395     }
2396     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2397     cols = uj;
2398     for (i=0; i<am; i++) {
2399       aj    = a->j + a->diag[i] + 1; /* 1st entry of U(i,:) without diagonal */
2400       ncols = ai[i+1] - a->diag[i] -1;
2401       for (j=0; j<ncols; j++) *cols++ = aj[j];
2402       *cols++ = i; /* diagoanl is located as the last entry of U(i,:) */
2403     }
2404   } else { /* case: levels>0 || (levels=0 && !perm_identity) */
2405     ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2406     ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2407 
2408     /* initialization */
2409     ierr = PetscMalloc1(am+1,&ajtmp);CHKERRQ(ierr);
2410 
2411     /* jl: linked list for storing indices of the pivot rows
2412        il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2413     ierr = PetscMalloc4(am,&uj_ptr,am,&uj_lvl_ptr,am,&jl,am,&il);CHKERRQ(ierr);
2414     for (i=0; i<am; i++) {
2415       jl[i] = am; il[i] = 0;
2416     }
2417 
2418     /* create and initialize a linked list for storing column indices of the active row k */
2419     nlnk = am + 1;
2420     ierr = PetscIncompleteLLCreate(am,am,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2421 
2422     /* initial FreeSpace size is fill*(ai[am]+am)/2 */
2423     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space);CHKERRQ(ierr);
2424     current_space     = free_space;
2425     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space_lvl);CHKERRQ(ierr);
2426     current_space_lvl = free_space_lvl;
2427 
2428     for (k=0; k<am; k++) {  /* for each active row k */
2429       /* initialize lnk by the column indices of row rip[k] of A */
2430       nzk   = 0;
2431       ncols = ai[rip[k]+1] - ai[rip[k]];
2432       if (!ncols) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_CH_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",rip[k],k);
2433       ncols_upper = 0;
2434       for (j=0; j<ncols; j++) {
2435         i = *(aj + ai[rip[k]] + j); /* unpermuted column index */
2436         if (riip[i] >= k) { /* only take upper triangular entry */
2437           ajtmp[ncols_upper] = i;
2438           ncols_upper++;
2439         }
2440       }
2441       ierr = PetscIncompleteLLInit(ncols_upper,ajtmp,am,riip,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2442       nzk += nlnk;
2443 
2444       /* update lnk by computing fill-in for each pivot row to be merged in */
2445       prow = jl[k]; /* 1st pivot row */
2446 
2447       while (prow < k) {
2448         nextprow = jl[prow];
2449 
2450         /* merge prow into k-th row */
2451         jmin  = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2452         jmax  = ui[prow+1];
2453         ncols = jmax-jmin;
2454         i     = jmin - ui[prow];
2455         cols  = uj_ptr[prow] + i; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2456         uj    = uj_lvl_ptr[prow] + i; /* levels of cols */
2457         j     = *(uj - 1);
2458         ierr  = PetscICCLLAddSorted(ncols,cols,levels,uj,am,nlnk,lnk,lnk_lvl,lnkbt,j);CHKERRQ(ierr);
2459         nzk  += nlnk;
2460 
2461         /* update il and jl for prow */
2462         if (jmin < jmax) {
2463           il[prow] = jmin;
2464           j        = *cols; jl[prow] = jl[j]; jl[j] = prow;
2465         }
2466         prow = nextprow;
2467       }
2468 
2469       /* if free space is not available, make more free space */
2470       if (current_space->local_remaining<nzk) {
2471         i    = am - k + 1; /* num of unfactored rows */
2472         i    = PetscIntMultTruncate(i,PetscMin(nzk, i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2473         ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2474         ierr = PetscFreeSpaceGet(i,&current_space_lvl);CHKERRQ(ierr);
2475         reallocs++;
2476       }
2477 
2478       /* copy data into free_space and free_space_lvl, then initialize lnk */
2479       if (nzk == 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Empty row %D in ICC matrix factor",k);
2480       ierr = PetscIncompleteLLClean(am,am,nzk,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
2481 
2482       /* add the k-th row into il and jl */
2483       if (nzk > 1) {
2484         i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2485         jl[k] = jl[i]; jl[i] = k;
2486         il[k] = ui[k] + 1;
2487       }
2488       uj_ptr[k]     = current_space->array;
2489       uj_lvl_ptr[k] = current_space_lvl->array;
2490 
2491       current_space->array           += nzk;
2492       current_space->local_used      += nzk;
2493       current_space->local_remaining -= nzk;
2494 
2495       current_space_lvl->array           += nzk;
2496       current_space_lvl->local_used      += nzk;
2497       current_space_lvl->local_remaining -= nzk;
2498 
2499       ui[k+1] = ui[k] + nzk;
2500     }
2501 
2502     ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2503     ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2504     ierr = PetscFree4(uj_ptr,uj_lvl_ptr,jl,il);CHKERRQ(ierr);
2505     ierr = PetscFree(ajtmp);CHKERRQ(ierr);
2506 
2507     /* copy free_space into uj and free free_space; set ui, uj, udiag in new datastructure; */
2508     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2509     ierr = PetscFreeSpaceContiguous_Cholesky(&free_space,uj,am,ui,udiag);CHKERRQ(ierr); /* store matrix factor  */
2510     ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2511     ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
2512 
2513   } /* end of case: levels>0 || (levels=0 && !perm_identity) */
2514 
2515   /* put together the new matrix in MATSEQSBAIJ format */
2516   b               = (Mat_SeqSBAIJ*)(fact)->data;
2517   b->singlemalloc = PETSC_FALSE;
2518 
2519   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2520 
2521   b->j             = uj;
2522   b->i             = ui;
2523   b->diag          = udiag;
2524   b->free_diag     = PETSC_TRUE;
2525   b->ilen          = 0;
2526   b->imax          = 0;
2527   b->row           = perm;
2528   b->col           = perm;
2529   ierr             = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2530   ierr             = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2531   b->icol          = iperm;
2532   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2533 
2534   ierr = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2535   ierr = PetscLogObjectMemory((PetscObject)fact,ui[am]*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2536 
2537   b->maxnz   = b->nz = ui[am];
2538   b->free_a  = PETSC_TRUE;
2539   b->free_ij = PETSC_TRUE;
2540 
2541   fact->info.factor_mallocs   = reallocs;
2542   fact->info.fill_ratio_given = fill;
2543   if (ai[am] != 0) {
2544     /* nonzeros in lower triangular part of A (including diagonals) = (ai[am]+am)/2 */
2545     fact->info.fill_ratio_needed = ((PetscReal)2*ui[am])/(ai[am]+am);
2546   } else {
2547     fact->info.fill_ratio_needed = 0.0;
2548   }
2549 #if defined(PETSC_USE_INFO)
2550   if (ai[am] != 0) {
2551     PetscReal af = fact->info.fill_ratio_needed;
2552     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2553     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2554     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2555   } else {
2556     ierr = PetscInfo(A,"Empty matrix.\n");CHKERRQ(ierr);
2557   }
2558 #endif
2559   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ;
2560   PetscFunctionReturn(0);
2561 }
2562 
2563 PetscErrorCode MatICCFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2564 {
2565   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2566   Mat_SeqSBAIJ       *b;
2567   PetscErrorCode     ierr;
2568   PetscBool          perm_identity,missing;
2569   PetscInt           reallocs=0,i,*ai=a->i,*aj=a->j,am=A->rmap->n,*ui,*udiag;
2570   const PetscInt     *rip,*riip;
2571   PetscInt           jmin,jmax,nzk,k,j,*jl,prow,*il,nextprow;
2572   PetscInt           nlnk,*lnk,*lnk_lvl=NULL,d;
2573   PetscInt           ncols,ncols_upper,*cols,*ajtmp,*uj,**uj_ptr,**uj_lvl_ptr;
2574   PetscReal          fill          =info->fill,levels=info->levels;
2575   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
2576   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
2577   PetscBT            lnkbt;
2578   IS                 iperm;
2579 
2580   PetscFunctionBegin;
2581   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
2582   ierr = MatMissingDiagonal(A,&missing,&d);CHKERRQ(ierr);
2583   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",d);
2584   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2585   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2586 
2587   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2588   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2589   ui[0] = 0;
2590 
2591   /* ICC(0) without matrix ordering: simply copies fill pattern */
2592   if (!levels && perm_identity) {
2593 
2594     for (i=0; i<am; i++) {
2595       ui[i+1]  = ui[i] + ai[i+1] - a->diag[i];
2596       udiag[i] = ui[i];
2597     }
2598     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2599     cols = uj;
2600     for (i=0; i<am; i++) {
2601       aj    = a->j + a->diag[i];
2602       ncols = ui[i+1] - ui[i];
2603       for (j=0; j<ncols; j++) *cols++ = *aj++;
2604     }
2605   } else { /* case: levels>0 || (levels=0 && !perm_identity) */
2606     ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2607     ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2608 
2609     /* initialization */
2610     ierr = PetscMalloc1(am+1,&ajtmp);CHKERRQ(ierr);
2611 
2612     /* jl: linked list for storing indices of the pivot rows
2613        il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2614     ierr = PetscMalloc4(am,&uj_ptr,am,&uj_lvl_ptr,am,&jl,am,&il);CHKERRQ(ierr);
2615     for (i=0; i<am; i++) {
2616       jl[i] = am; il[i] = 0;
2617     }
2618 
2619     /* create and initialize a linked list for storing column indices of the active row k */
2620     nlnk = am + 1;
2621     ierr = PetscIncompleteLLCreate(am,am,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2622 
2623     /* initial FreeSpace size is fill*(ai[am]+1) */
2624     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space);CHKERRQ(ierr);
2625     current_space     = free_space;
2626     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space_lvl);CHKERRQ(ierr);
2627     current_space_lvl = free_space_lvl;
2628 
2629     for (k=0; k<am; k++) {  /* for each active row k */
2630       /* initialize lnk by the column indices of row rip[k] of A */
2631       nzk   = 0;
2632       ncols = ai[rip[k]+1] - ai[rip[k]];
2633       if (!ncols) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_CH_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",rip[k],k);
2634       ncols_upper = 0;
2635       for (j=0; j<ncols; j++) {
2636         i = *(aj + ai[rip[k]] + j); /* unpermuted column index */
2637         if (riip[i] >= k) { /* only take upper triangular entry */
2638           ajtmp[ncols_upper] = i;
2639           ncols_upper++;
2640         }
2641       }
2642       ierr = PetscIncompleteLLInit(ncols_upper,ajtmp,am,riip,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2643       nzk += nlnk;
2644 
2645       /* update lnk by computing fill-in for each pivot row to be merged in */
2646       prow = jl[k]; /* 1st pivot row */
2647 
2648       while (prow < k) {
2649         nextprow = jl[prow];
2650 
2651         /* merge prow into k-th row */
2652         jmin  = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2653         jmax  = ui[prow+1];
2654         ncols = jmax-jmin;
2655         i     = jmin - ui[prow];
2656         cols  = uj_ptr[prow] + i; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2657         uj    = uj_lvl_ptr[prow] + i; /* levels of cols */
2658         j     = *(uj - 1);
2659         ierr  = PetscICCLLAddSorted(ncols,cols,levels,uj,am,nlnk,lnk,lnk_lvl,lnkbt,j);CHKERRQ(ierr);
2660         nzk  += nlnk;
2661 
2662         /* update il and jl for prow */
2663         if (jmin < jmax) {
2664           il[prow] = jmin;
2665           j        = *cols; jl[prow] = jl[j]; jl[j] = prow;
2666         }
2667         prow = nextprow;
2668       }
2669 
2670       /* if free space is not available, make more free space */
2671       if (current_space->local_remaining<nzk) {
2672         i    = am - k + 1; /* num of unfactored rows */
2673         i    = PetscIntMultTruncate(i,PetscMin(nzk, i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2674         ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2675         ierr = PetscFreeSpaceGet(i,&current_space_lvl);CHKERRQ(ierr);
2676         reallocs++;
2677       }
2678 
2679       /* copy data into free_space and free_space_lvl, then initialize lnk */
2680       if (!nzk) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Empty row %D in ICC matrix factor",k);
2681       ierr = PetscIncompleteLLClean(am,am,nzk,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
2682 
2683       /* add the k-th row into il and jl */
2684       if (nzk > 1) {
2685         i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2686         jl[k] = jl[i]; jl[i] = k;
2687         il[k] = ui[k] + 1;
2688       }
2689       uj_ptr[k]     = current_space->array;
2690       uj_lvl_ptr[k] = current_space_lvl->array;
2691 
2692       current_space->array           += nzk;
2693       current_space->local_used      += nzk;
2694       current_space->local_remaining -= nzk;
2695 
2696       current_space_lvl->array           += nzk;
2697       current_space_lvl->local_used      += nzk;
2698       current_space_lvl->local_remaining -= nzk;
2699 
2700       ui[k+1] = ui[k] + nzk;
2701     }
2702 
2703 #if defined(PETSC_USE_INFO)
2704     if (ai[am] != 0) {
2705       PetscReal af = (PetscReal)ui[am]/((PetscReal)ai[am]);
2706       ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2707       ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2708       ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2709     } else {
2710       ierr = PetscInfo(A,"Empty matrix.\n");CHKERRQ(ierr);
2711     }
2712 #endif
2713 
2714     ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2715     ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2716     ierr = PetscFree4(uj_ptr,uj_lvl_ptr,jl,il);CHKERRQ(ierr);
2717     ierr = PetscFree(ajtmp);CHKERRQ(ierr);
2718 
2719     /* destroy list of free space and other temporary array(s) */
2720     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2721     ierr = PetscFreeSpaceContiguous(&free_space,uj);CHKERRQ(ierr);
2722     ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2723     ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
2724 
2725   } /* end of case: levels>0 || (levels=0 && !perm_identity) */
2726 
2727   /* put together the new matrix in MATSEQSBAIJ format */
2728 
2729   b               = (Mat_SeqSBAIJ*)fact->data;
2730   b->singlemalloc = PETSC_FALSE;
2731 
2732   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2733 
2734   b->j         = uj;
2735   b->i         = ui;
2736   b->diag      = udiag;
2737   b->free_diag = PETSC_TRUE;
2738   b->ilen      = 0;
2739   b->imax      = 0;
2740   b->row       = perm;
2741   b->col       = perm;
2742 
2743   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2744   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2745 
2746   b->icol          = iperm;
2747   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2748   ierr             = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2749   ierr             = PetscLogObjectMemory((PetscObject)fact,(ui[am]-am)*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2750   b->maxnz         = b->nz = ui[am];
2751   b->free_a        = PETSC_TRUE;
2752   b->free_ij       = PETSC_TRUE;
2753 
2754   fact->info.factor_mallocs   = reallocs;
2755   fact->info.fill_ratio_given = fill;
2756   if (ai[am] != 0) {
2757     fact->info.fill_ratio_needed = ((PetscReal)ui[am])/((PetscReal)ai[am]);
2758   } else {
2759     fact->info.fill_ratio_needed = 0.0;
2760   }
2761   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ_inplace;
2762   PetscFunctionReturn(0);
2763 }
2764 
2765 PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2766 {
2767   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2768   Mat_SeqSBAIJ       *b;
2769   PetscErrorCode     ierr;
2770   PetscBool          perm_identity,missing;
2771   PetscReal          fill = info->fill;
2772   const PetscInt     *rip,*riip;
2773   PetscInt           i,am=A->rmap->n,*ai=a->i,*aj=a->j,reallocs=0,prow;
2774   PetscInt           *jl,jmin,jmax,nzk,*ui,k,j,*il,nextprow;
2775   PetscInt           nlnk,*lnk,ncols,ncols_upper,*cols,*uj,**ui_ptr,*uj_ptr,*udiag;
2776   PetscFreeSpaceList free_space=NULL,current_space=NULL;
2777   PetscBT            lnkbt;
2778   IS                 iperm;
2779 
2780   PetscFunctionBegin;
2781   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
2782   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
2783   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
2784 
2785   /* check whether perm is the identity mapping */
2786   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2787   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2788   ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2789   ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2790 
2791   /* initialization */
2792   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2793   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2794   ui[0] = 0;
2795 
2796   /* jl: linked list for storing indices of the pivot rows
2797      il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2798   ierr = PetscMalloc4(am,&ui_ptr,am,&jl,am,&il,am,&cols);CHKERRQ(ierr);
2799   for (i=0; i<am; i++) {
2800     jl[i] = am; il[i] = 0;
2801   }
2802 
2803   /* create and initialize a linked list for storing column indices of the active row k */
2804   nlnk = am + 1;
2805   ierr = PetscLLCreate(am,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2806 
2807   /* initial FreeSpace size is fill*(ai[am]+am)/2 */
2808   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space);CHKERRQ(ierr);
2809   current_space = free_space;
2810 
2811   for (k=0; k<am; k++) {  /* for each active row k */
2812     /* initialize lnk by the column indices of row rip[k] of A */
2813     nzk   = 0;
2814     ncols = ai[rip[k]+1] - ai[rip[k]];
2815     if (!ncols) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_CH_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",rip[k],k);
2816     ncols_upper = 0;
2817     for (j=0; j<ncols; j++) {
2818       i = riip[*(aj + ai[rip[k]] + j)];
2819       if (i >= k) { /* only take upper triangular entry */
2820         cols[ncols_upper] = i;
2821         ncols_upper++;
2822       }
2823     }
2824     ierr = PetscLLAdd(ncols_upper,cols,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2825     nzk += nlnk;
2826 
2827     /* update lnk by computing fill-in for each pivot row to be merged in */
2828     prow = jl[k]; /* 1st pivot row */
2829 
2830     while (prow < k) {
2831       nextprow = jl[prow];
2832       /* merge prow into k-th row */
2833       jmin   = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2834       jmax   = ui[prow+1];
2835       ncols  = jmax-jmin;
2836       uj_ptr = ui_ptr[prow] + jmin - ui[prow]; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2837       ierr   = PetscLLAddSorted(ncols,uj_ptr,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2838       nzk   += nlnk;
2839 
2840       /* update il and jl for prow */
2841       if (jmin < jmax) {
2842         il[prow] = jmin;
2843         j        = *uj_ptr;
2844         jl[prow] = jl[j];
2845         jl[j]    = prow;
2846       }
2847       prow = nextprow;
2848     }
2849 
2850     /* if free space is not available, make more free space */
2851     if (current_space->local_remaining<nzk) {
2852       i    = am - k + 1; /* num of unfactored rows */
2853       i    = PetscIntMultTruncate(i,PetscMin(nzk,i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2854       ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2855       reallocs++;
2856     }
2857 
2858     /* copy data into free space, then initialize lnk */
2859     ierr = PetscLLClean(am,am,nzk,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
2860 
2861     /* add the k-th row into il and jl */
2862     if (nzk > 1) {
2863       i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2864       jl[k] = jl[i]; jl[i] = k;
2865       il[k] = ui[k] + 1;
2866     }
2867     ui_ptr[k] = current_space->array;
2868 
2869     current_space->array           += nzk;
2870     current_space->local_used      += nzk;
2871     current_space->local_remaining -= nzk;
2872 
2873     ui[k+1] = ui[k] + nzk;
2874   }
2875 
2876   ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2877   ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2878   ierr = PetscFree4(ui_ptr,jl,il,cols);CHKERRQ(ierr);
2879 
2880   /* copy free_space into uj and free free_space; set ui, uj, udiag in new datastructure; */
2881   ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2882   ierr = PetscFreeSpaceContiguous_Cholesky(&free_space,uj,am,ui,udiag);CHKERRQ(ierr); /* store matrix factor */
2883   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2884 
2885   /* put together the new matrix in MATSEQSBAIJ format */
2886 
2887   b               = (Mat_SeqSBAIJ*)fact->data;
2888   b->singlemalloc = PETSC_FALSE;
2889   b->free_a       = PETSC_TRUE;
2890   b->free_ij      = PETSC_TRUE;
2891 
2892   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2893 
2894   b->j         = uj;
2895   b->i         = ui;
2896   b->diag      = udiag;
2897   b->free_diag = PETSC_TRUE;
2898   b->ilen      = 0;
2899   b->imax      = 0;
2900   b->row       = perm;
2901   b->col       = perm;
2902 
2903   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2904   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2905 
2906   b->icol          = iperm;
2907   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2908 
2909   ierr = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2910   ierr = PetscLogObjectMemory((PetscObject)fact,ui[am]*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2911 
2912   b->maxnz = b->nz = ui[am];
2913 
2914   fact->info.factor_mallocs   = reallocs;
2915   fact->info.fill_ratio_given = fill;
2916   if (ai[am] != 0) {
2917     /* nonzeros in lower triangular part of A (including diagonals) = (ai[am]+am)/2 */
2918     fact->info.fill_ratio_needed = ((PetscReal)2*ui[am])/(ai[am]+am);
2919   } else {
2920     fact->info.fill_ratio_needed = 0.0;
2921   }
2922 #if defined(PETSC_USE_INFO)
2923   if (ai[am] != 0) {
2924     PetscReal af = fact->info.fill_ratio_needed;
2925     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2926     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2927     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2928   } else {
2929     ierr = PetscInfo(A,"Empty matrix.\n");CHKERRQ(ierr);
2930   }
2931 #endif
2932   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ;
2933   PetscFunctionReturn(0);
2934 }
2935 
2936 PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2937 {
2938   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2939   Mat_SeqSBAIJ       *b;
2940   PetscErrorCode     ierr;
2941   PetscBool          perm_identity,missing;
2942   PetscReal          fill = info->fill;
2943   const PetscInt     *rip,*riip;
2944   PetscInt           i,am=A->rmap->n,*ai=a->i,*aj=a->j,reallocs=0,prow;
2945   PetscInt           *jl,jmin,jmax,nzk,*ui,k,j,*il,nextprow;
2946   PetscInt           nlnk,*lnk,ncols,ncols_upper,*cols,*uj,**ui_ptr,*uj_ptr;
2947   PetscFreeSpaceList free_space=NULL,current_space=NULL;
2948   PetscBT            lnkbt;
2949   IS                 iperm;
2950 
2951   PetscFunctionBegin;
2952   if (A->rmap->n != A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Must be square matrix, rows %D columns %D",A->rmap->n,A->cmap->n);
2953   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
2954   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
2955 
2956   /* check whether perm is the identity mapping */
2957   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2958   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2959   ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2960   ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2961 
2962   /* initialization */
2963   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2964   ui[0] = 0;
2965 
2966   /* jl: linked list for storing indices of the pivot rows
2967      il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2968   ierr = PetscMalloc4(am,&ui_ptr,am,&jl,am,&il,am,&cols);CHKERRQ(ierr);
2969   for (i=0; i<am; i++) {
2970     jl[i] = am; il[i] = 0;
2971   }
2972 
2973   /* create and initialize a linked list for storing column indices of the active row k */
2974   nlnk = am + 1;
2975   ierr = PetscLLCreate(am,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2976 
2977   /* initial FreeSpace size is fill*(ai[am]+1) */
2978   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space);CHKERRQ(ierr);
2979   current_space = free_space;
2980 
2981   for (k=0; k<am; k++) {  /* for each active row k */
2982     /* initialize lnk by the column indices of row rip[k] of A */
2983     nzk   = 0;
2984     ncols = ai[rip[k]+1] - ai[rip[k]];
2985     if (!ncols) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_CH_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",rip[k],k);
2986     ncols_upper = 0;
2987     for (j=0; j<ncols; j++) {
2988       i = riip[*(aj + ai[rip[k]] + j)];
2989       if (i >= k) { /* only take upper triangular entry */
2990         cols[ncols_upper] = i;
2991         ncols_upper++;
2992       }
2993     }
2994     ierr = PetscLLAdd(ncols_upper,cols,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2995     nzk += nlnk;
2996 
2997     /* update lnk by computing fill-in for each pivot row to be merged in */
2998     prow = jl[k]; /* 1st pivot row */
2999 
3000     while (prow < k) {
3001       nextprow = jl[prow];
3002       /* merge prow into k-th row */
3003       jmin   = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
3004       jmax   = ui[prow+1];
3005       ncols  = jmax-jmin;
3006       uj_ptr = ui_ptr[prow] + jmin - ui[prow]; /* points to the 2nd nzero entry in U(prow,k:am-1) */
3007       ierr   = PetscLLAddSorted(ncols,uj_ptr,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3008       nzk   += nlnk;
3009 
3010       /* update il and jl for prow */
3011       if (jmin < jmax) {
3012         il[prow] = jmin;
3013         j        = *uj_ptr; jl[prow] = jl[j]; jl[j] = prow;
3014       }
3015       prow = nextprow;
3016     }
3017 
3018     /* if free space is not available, make more free space */
3019     if (current_space->local_remaining<nzk) {
3020       i    = am - k + 1; /* num of unfactored rows */
3021       i    = PetscMin(i*nzk, i*(i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
3022       ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
3023       reallocs++;
3024     }
3025 
3026     /* copy data into free space, then initialize lnk */
3027     ierr = PetscLLClean(am,am,nzk,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
3028 
3029     /* add the k-th row into il and jl */
3030     if (nzk-1 > 0) {
3031       i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
3032       jl[k] = jl[i]; jl[i] = k;
3033       il[k] = ui[k] + 1;
3034     }
3035     ui_ptr[k] = current_space->array;
3036 
3037     current_space->array           += nzk;
3038     current_space->local_used      += nzk;
3039     current_space->local_remaining -= nzk;
3040 
3041     ui[k+1] = ui[k] + nzk;
3042   }
3043 
3044 #if defined(PETSC_USE_INFO)
3045   if (ai[am] != 0) {
3046     PetscReal af = (PetscReal)(ui[am])/((PetscReal)ai[am]);
3047     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
3048     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
3049     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
3050   } else {
3051     ierr = PetscInfo(A,"Empty matrix.\n");CHKERRQ(ierr);
3052   }
3053 #endif
3054 
3055   ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
3056   ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
3057   ierr = PetscFree4(ui_ptr,jl,il,cols);CHKERRQ(ierr);
3058 
3059   /* destroy list of free space and other temporary array(s) */
3060   ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
3061   ierr = PetscFreeSpaceContiguous(&free_space,uj);CHKERRQ(ierr);
3062   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
3063 
3064   /* put together the new matrix in MATSEQSBAIJ format */
3065 
3066   b               = (Mat_SeqSBAIJ*)fact->data;
3067   b->singlemalloc = PETSC_FALSE;
3068   b->free_a       = PETSC_TRUE;
3069   b->free_ij      = PETSC_TRUE;
3070 
3071   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
3072 
3073   b->j    = uj;
3074   b->i    = ui;
3075   b->diag = 0;
3076   b->ilen = 0;
3077   b->imax = 0;
3078   b->row  = perm;
3079   b->col  = perm;
3080 
3081   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
3082   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
3083 
3084   b->icol          = iperm;
3085   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
3086 
3087   ierr     = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
3088   ierr     = PetscLogObjectMemory((PetscObject)fact,(ui[am]-am)*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
3089   b->maxnz = b->nz = ui[am];
3090 
3091   fact->info.factor_mallocs   = reallocs;
3092   fact->info.fill_ratio_given = fill;
3093   if (ai[am] != 0) {
3094     fact->info.fill_ratio_needed = ((PetscReal)ui[am])/((PetscReal)ai[am]);
3095   } else {
3096     fact->info.fill_ratio_needed = 0.0;
3097   }
3098   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ_inplace;
3099   PetscFunctionReturn(0);
3100 }
3101 
3102 PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering(Mat A,Vec bb,Vec xx)
3103 {
3104   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
3105   PetscErrorCode    ierr;
3106   PetscInt          n   = A->rmap->n;
3107   const PetscInt    *ai = a->i,*aj = a->j,*adiag = a->diag,*vi;
3108   PetscScalar       *x,sum;
3109   const PetscScalar *b;
3110   const MatScalar   *aa = a->a,*v;
3111   PetscInt          i,nz;
3112 
3113   PetscFunctionBegin;
3114   if (!n) PetscFunctionReturn(0);
3115 
3116   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
3117   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
3118 
3119   /* forward solve the lower triangular */
3120   x[0] = b[0];
3121   v    = aa;
3122   vi   = aj;
3123   for (i=1; i<n; i++) {
3124     nz  = ai[i+1] - ai[i];
3125     sum = b[i];
3126     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
3127     v   += nz;
3128     vi  += nz;
3129     x[i] = sum;
3130   }
3131 
3132   /* backward solve the upper triangular */
3133   for (i=n-1; i>=0; i--) {
3134     v   = aa + adiag[i+1] + 1;
3135     vi  = aj + adiag[i+1] + 1;
3136     nz  = adiag[i] - adiag[i+1]-1;
3137     sum = x[i];
3138     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
3139     x[i] = sum*v[nz]; /* x[i]=aa[adiag[i]]*sum; v++; */
3140   }
3141 
3142   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
3143   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
3144   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
3145   PetscFunctionReturn(0);
3146 }
3147 
3148 PetscErrorCode MatSolve_SeqAIJ(Mat A,Vec bb,Vec xx)
3149 {
3150   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
3151   IS                iscol = a->col,isrow = a->row;
3152   PetscErrorCode    ierr;
3153   PetscInt          i,n=A->rmap->n,*vi,*ai=a->i,*aj=a->j,*adiag = a->diag,nz;
3154   const PetscInt    *rout,*cout,*r,*c;
3155   PetscScalar       *x,*tmp,sum;
3156   const PetscScalar *b;
3157   const MatScalar   *aa = a->a,*v;
3158 
3159   PetscFunctionBegin;
3160   if (!n) PetscFunctionReturn(0);
3161 
3162   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
3163   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
3164   tmp  = a->solve_work;
3165 
3166   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
3167   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
3168 
3169   /* forward solve the lower triangular */
3170   tmp[0] = b[r[0]];
3171   v      = aa;
3172   vi     = aj;
3173   for (i=1; i<n; i++) {
3174     nz  = ai[i+1] - ai[i];
3175     sum = b[r[i]];
3176     PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
3177     tmp[i] = sum;
3178     v     += nz; vi += nz;
3179   }
3180 
3181   /* backward solve the upper triangular */
3182   for (i=n-1; i>=0; i--) {
3183     v   = aa + adiag[i+1]+1;
3184     vi  = aj + adiag[i+1]+1;
3185     nz  = adiag[i]-adiag[i+1]-1;
3186     sum = tmp[i];
3187     PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
3188     x[c[i]] = tmp[i] = sum*v[nz]; /* v[nz] = aa[adiag[i]] */
3189   }
3190 
3191   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
3192   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
3193   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
3194   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
3195   ierr = PetscLogFlops(2*a->nz - A->cmap->n);CHKERRQ(ierr);
3196   PetscFunctionReturn(0);
3197 }
3198 
3199 /*
3200     This will get a new name and become a varient of MatILUFactor_SeqAIJ() there is no longer separate functions in the matrix function table for dt factors
3201 */
3202 PetscErrorCode MatILUDTFactor_SeqAIJ(Mat A,IS isrow,IS iscol,const MatFactorInfo *info,Mat *fact)
3203 {
3204   Mat            B = *fact;
3205   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data,*b;
3206   IS             isicol;
3207   PetscErrorCode ierr;
3208   const PetscInt *r,*ic;
3209   PetscInt       i,n=A->rmap->n,*ai=a->i,*aj=a->j,*ajtmp,*adiag;
3210   PetscInt       *bi,*bj,*bdiag,*bdiag_rev;
3211   PetscInt       row,nzi,nzi_bl,nzi_bu,*im,nzi_al,nzi_au;
3212   PetscInt       nlnk,*lnk;
3213   PetscBT        lnkbt;
3214   PetscBool      row_identity,icol_identity;
3215   MatScalar      *aatmp,*pv,*batmp,*ba,*rtmp,*pc,multiplier,*vtmp,diag_tmp;
3216   const PetscInt *ics;
3217   PetscInt       j,nz,*pj,*bjtmp,k,ncut,*jtmp;
3218   PetscReal      dt     =info->dt,shift=info->shiftamount;
3219   PetscInt       dtcount=(PetscInt)info->dtcount,nnz_max;
3220   PetscBool      missing;
3221 
3222   PetscFunctionBegin;
3223   if (dt      == PETSC_DEFAULT) dt = 0.005;
3224   if (dtcount == PETSC_DEFAULT) dtcount = (PetscInt)(1.5*a->rmax);
3225 
3226   /* ------- symbolic factorization, can be reused ---------*/
3227   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
3228   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
3229   adiag=a->diag;
3230 
3231   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
3232 
3233   /* bdiag is location of diagonal in factor */
3234   ierr = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);     /* becomes b->diag */
3235   ierr = PetscMalloc1(n+1,&bdiag_rev);CHKERRQ(ierr); /* temporary */
3236 
3237   /* allocate row pointers bi */
3238   ierr = PetscMalloc1(2*n+2,&bi);CHKERRQ(ierr);
3239 
3240   /* allocate bj and ba; max num of nonzero entries is (ai[n]+2*n*dtcount+2) */
3241   if (dtcount > n-1) dtcount = n-1; /* diagonal is excluded */
3242   nnz_max = ai[n]+2*n*dtcount+2;
3243 
3244   ierr = PetscMalloc1(nnz_max+1,&bj);CHKERRQ(ierr);
3245   ierr = PetscMalloc1(nnz_max+1,&ba);CHKERRQ(ierr);
3246 
3247   /* put together the new matrix */
3248   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
3249   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
3250   b    = (Mat_SeqAIJ*)B->data;
3251 
3252   b->free_a       = PETSC_TRUE;
3253   b->free_ij      = PETSC_TRUE;
3254   b->singlemalloc = PETSC_FALSE;
3255 
3256   b->a    = ba;
3257   b->j    = bj;
3258   b->i    = bi;
3259   b->diag = bdiag;
3260   b->ilen = 0;
3261   b->imax = 0;
3262   b->row  = isrow;
3263   b->col  = iscol;
3264   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
3265   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
3266   b->icol = isicol;
3267 
3268   ierr     = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
3269   ierr     = PetscLogObjectMemory((PetscObject)B,nnz_max*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
3270   b->maxnz = nnz_max;
3271 
3272   B->factortype            = MAT_FACTOR_ILUDT;
3273   B->info.factor_mallocs   = 0;
3274   B->info.fill_ratio_given = ((PetscReal)nnz_max)/((PetscReal)ai[n]);
3275   /* ------- end of symbolic factorization ---------*/
3276 
3277   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
3278   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
3279   ics  = ic;
3280 
3281   /* linked list for storing column indices of the active row */
3282   nlnk = n + 1;
3283   ierr = PetscLLCreate(n,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3284 
3285   /* im: used by PetscLLAddSortedLU(); jtmp: working array for column indices of active row */
3286   ierr = PetscMalloc2(n,&im,n,&jtmp);CHKERRQ(ierr);
3287   /* rtmp, vtmp: working arrays for sparse and contiguous row entries of active row */
3288   ierr = PetscMalloc2(n,&rtmp,n,&vtmp);CHKERRQ(ierr);
3289   ierr = PetscMemzero(rtmp,n*sizeof(MatScalar));CHKERRQ(ierr);
3290 
3291   bi[0]        = 0;
3292   bdiag[0]     = nnz_max-1; /* location of diag[0] in factor B */
3293   bdiag_rev[n] = bdiag[0];
3294   bi[2*n+1]    = bdiag[0]+1; /* endof bj and ba array */
3295   for (i=0; i<n; i++) {
3296     /* copy initial fill into linked list */
3297     nzi = ai[r[i]+1] - ai[r[i]];
3298     if (!nzi) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",r[i],i);
3299     nzi_al = adiag[r[i]] - ai[r[i]];
3300     nzi_au = ai[r[i]+1] - adiag[r[i]] -1;
3301     ajtmp  = aj + ai[r[i]];
3302     ierr   = PetscLLAddPerm(nzi,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3303 
3304     /* load in initial (unfactored row) */
3305     aatmp = a->a + ai[r[i]];
3306     for (j=0; j<nzi; j++) {
3307       rtmp[ics[*ajtmp++]] = *aatmp++;
3308     }
3309 
3310     /* add pivot rows into linked list */
3311     row = lnk[n];
3312     while (row < i) {
3313       nzi_bl = bi[row+1] - bi[row] + 1;
3314       bjtmp  = bj + bdiag[row+1]+1; /* points to 1st column next to the diagonal in U */
3315       ierr   = PetscLLAddSortedLU(bjtmp,row,nlnk,lnk,lnkbt,i,nzi_bl,im);CHKERRQ(ierr);
3316       nzi   += nlnk;
3317       row    = lnk[row];
3318     }
3319 
3320     /* copy data from lnk into jtmp, then initialize lnk */
3321     ierr = PetscLLClean(n,n,nzi,lnk,jtmp,lnkbt);CHKERRQ(ierr);
3322 
3323     /* numerical factorization */
3324     bjtmp = jtmp;
3325     row   = *bjtmp++; /* 1st pivot row */
3326     while (row < i) {
3327       pc         = rtmp + row;
3328       pv         = ba + bdiag[row]; /* 1./(diag of the pivot row) */
3329       multiplier = (*pc) * (*pv);
3330       *pc        = multiplier;
3331       if (PetscAbsScalar(*pc) > dt) { /* apply tolerance dropping rule */
3332         pj = bj + bdiag[row+1] + 1;         /* point to 1st entry of U(row,:) */
3333         pv = ba + bdiag[row+1] + 1;
3334         /* if (multiplier < -1.0 or multiplier >1.0) printf("row/prow %d, %d, multiplier %g\n",i,row,multiplier); */
3335         nz = bdiag[row] - bdiag[row+1] - 1;         /* num of entries in U(row,:), excluding diagonal */
3336         for (j=0; j<nz; j++) rtmp[*pj++] -= multiplier * (*pv++);
3337         ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
3338       }
3339       row = *bjtmp++;
3340     }
3341 
3342     /* copy sparse rtmp into contiguous vtmp; separate L and U part */
3343     diag_tmp = rtmp[i];  /* save diagonal value - may not needed?? */
3344     nzi_bl   = 0; j = 0;
3345     while (jtmp[j] < i) { /* Note: jtmp is sorted */
3346       vtmp[j] = rtmp[jtmp[j]]; rtmp[jtmp[j]]=0.0;
3347       nzi_bl++; j++;
3348     }
3349     nzi_bu = nzi - nzi_bl -1;
3350     while (j < nzi) {
3351       vtmp[j] = rtmp[jtmp[j]]; rtmp[jtmp[j]]=0.0;
3352       j++;
3353     }
3354 
3355     bjtmp = bj + bi[i];
3356     batmp = ba + bi[i];
3357     /* apply level dropping rule to L part */
3358     ncut = nzi_al + dtcount;
3359     if (ncut < nzi_bl) {
3360       ierr = PetscSortSplit(ncut,nzi_bl,vtmp,jtmp);CHKERRQ(ierr);
3361       ierr = PetscSortIntWithScalarArray(ncut,jtmp,vtmp);CHKERRQ(ierr);
3362     } else {
3363       ncut = nzi_bl;
3364     }
3365     for (j=0; j<ncut; j++) {
3366       bjtmp[j] = jtmp[j];
3367       batmp[j] = vtmp[j];
3368       /* printf(" (%d,%g),",bjtmp[j],batmp[j]); */
3369     }
3370     bi[i+1] = bi[i] + ncut;
3371     nzi     = ncut + 1;
3372 
3373     /* apply level dropping rule to U part */
3374     ncut = nzi_au + dtcount;
3375     if (ncut < nzi_bu) {
3376       ierr = PetscSortSplit(ncut,nzi_bu,vtmp+nzi_bl+1,jtmp+nzi_bl+1);CHKERRQ(ierr);
3377       ierr = PetscSortIntWithScalarArray(ncut,jtmp+nzi_bl+1,vtmp+nzi_bl+1);CHKERRQ(ierr);
3378     } else {
3379       ncut = nzi_bu;
3380     }
3381     nzi += ncut;
3382 
3383     /* mark bdiagonal */
3384     bdiag[i+1]       = bdiag[i] - (ncut + 1);
3385     bdiag_rev[n-i-1] = bdiag[i+1];
3386     bi[2*n - i]      = bi[2*n - i +1] - (ncut + 1);
3387     bjtmp            = bj + bdiag[i];
3388     batmp            = ba + bdiag[i];
3389     *bjtmp           = i;
3390     *batmp           = diag_tmp; /* rtmp[i]; */
3391     if (*batmp == 0.0) {
3392       *batmp = dt+shift;
3393       /* printf(" row %d add shift %g\n",i,shift); */
3394     }
3395     *batmp = 1.0/(*batmp); /* invert diagonal entries for simplier triangular solves */
3396     /* printf(" (%d,%g),",*bjtmp,*batmp); */
3397 
3398     bjtmp = bj + bdiag[i+1]+1;
3399     batmp = ba + bdiag[i+1]+1;
3400     for (k=0; k<ncut; k++) {
3401       bjtmp[k] = jtmp[nzi_bl+1+k];
3402       batmp[k] = vtmp[nzi_bl+1+k];
3403       /* printf(" (%d,%g),",bjtmp[k],batmp[k]); */
3404     }
3405     /* printf("\n"); */
3406 
3407     im[i] = nzi;   /* used by PetscLLAddSortedLU() */
3408     /*
3409     printf("row %d: bi %d, bdiag %d\n",i,bi[i],bdiag[i]);
3410     printf(" ----------------------------\n");
3411     */
3412   } /* for (i=0; i<n; i++) */
3413     /* printf("end of L %d, beginning of U %d\n",bi[n],bdiag[n]); */
3414   if (bi[n] >= bdiag[n]) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"end of L array %d cannot >= the beginning of U array %d",bi[n],bdiag[n]);
3415 
3416   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
3417   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
3418 
3419   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
3420   ierr = PetscFree2(im,jtmp);CHKERRQ(ierr);
3421   ierr = PetscFree2(rtmp,vtmp);CHKERRQ(ierr);
3422   ierr = PetscFree(bdiag_rev);CHKERRQ(ierr);
3423 
3424   ierr     = PetscLogFlops(B->cmap->n);CHKERRQ(ierr);
3425   b->maxnz = b->nz = bi[n] + bdiag[0] - bdiag[n];
3426 
3427   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
3428   ierr = ISIdentity(isicol,&icol_identity);CHKERRQ(ierr);
3429   if (row_identity && icol_identity) {
3430     B->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
3431   } else {
3432     B->ops->solve = MatSolve_SeqAIJ;
3433   }
3434 
3435   B->ops->solveadd          = 0;
3436   B->ops->solvetranspose    = 0;
3437   B->ops->solvetransposeadd = 0;
3438   B->ops->matsolve          = 0;
3439   B->assembled              = PETSC_TRUE;
3440   B->preallocated           = PETSC_TRUE;
3441   PetscFunctionReturn(0);
3442 }
3443 
3444 /* a wraper of MatILUDTFactor_SeqAIJ() */
3445 /*
3446     This will get a new name and become a varient of MatILUFactor_SeqAIJ() there is no longer separate functions in the matrix function table for dt factors
3447 */
3448 
3449 PetscErrorCode  MatILUDTFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS row,IS col,const MatFactorInfo *info)
3450 {
3451   PetscErrorCode ierr;
3452 
3453   PetscFunctionBegin;
3454   ierr = MatILUDTFactor_SeqAIJ(A,row,col,info,&fact);CHKERRQ(ierr);
3455   PetscFunctionReturn(0);
3456 }
3457 
3458 /*
3459    same as MatLUFactorNumeric_SeqAIJ(), except using contiguous array matrix factors
3460    - intend to replace existing MatLUFactorNumeric_SeqAIJ()
3461 */
3462 /*
3463     This will get a new name and become a varient of MatILUFactor_SeqAIJ() there is no longer separate functions in the matrix function table for dt factors
3464 */
3465 
3466 PetscErrorCode  MatILUDTFactorNumeric_SeqAIJ(Mat fact,Mat A,const MatFactorInfo *info)
3467 {
3468   Mat            C     =fact;
3469   Mat_SeqAIJ     *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
3470   IS             isrow = b->row,isicol = b->icol;
3471   PetscErrorCode ierr;
3472   const PetscInt *r,*ic,*ics;
3473   PetscInt       i,j,k,n=A->rmap->n,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j;
3474   PetscInt       *ajtmp,*bjtmp,nz,nzl,nzu,row,*bdiag = b->diag,*pj;
3475   MatScalar      *rtmp,*pc,multiplier,*v,*pv,*aa=a->a;
3476   PetscReal      dt=info->dt,shift=info->shiftamount;
3477   PetscBool      row_identity, col_identity;
3478 
3479   PetscFunctionBegin;
3480   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
3481   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
3482   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
3483   ics  = ic;
3484 
3485   for (i=0; i<n; i++) {
3486     /* initialize rtmp array */
3487     nzl   = bi[i+1] - bi[i];       /* num of nozeros in L(i,:) */
3488     bjtmp = bj + bi[i];
3489     for  (j=0; j<nzl; j++) rtmp[*bjtmp++] = 0.0;
3490     rtmp[i] = 0.0;
3491     nzu     = bdiag[i] - bdiag[i+1]; /* num of nozeros in U(i,:) */
3492     bjtmp   = bj + bdiag[i+1] + 1;
3493     for  (j=0; j<nzu; j++) rtmp[*bjtmp++] = 0.0;
3494 
3495     /* load in initial unfactored row of A */
3496     /* printf("row %d\n",i); */
3497     nz    = ai[r[i]+1] - ai[r[i]];
3498     ajtmp = aj + ai[r[i]];
3499     v     = aa + ai[r[i]];
3500     for (j=0; j<nz; j++) {
3501       rtmp[ics[*ajtmp++]] = v[j];
3502       /* printf(" (%d,%g),",ics[ajtmp[j]],rtmp[ics[ajtmp[j]]]); */
3503     }
3504     /* printf("\n"); */
3505 
3506     /* numerical factorization */
3507     bjtmp = bj + bi[i]; /* point to 1st entry of L(i,:) */
3508     nzl   = bi[i+1] - bi[i]; /* num of entries in L(i,:) */
3509     k     = 0;
3510     while (k < nzl) {
3511       row = *bjtmp++;
3512       /* printf("  prow %d\n",row); */
3513       pc         = rtmp + row;
3514       pv         = b->a + bdiag[row]; /* 1./(diag of the pivot row) */
3515       multiplier = (*pc) * (*pv);
3516       *pc        = multiplier;
3517       if (PetscAbsScalar(multiplier) > dt) {
3518         pj = bj + bdiag[row+1] + 1;         /* point to 1st entry of U(row,:) */
3519         pv = b->a + bdiag[row+1] + 1;
3520         nz = bdiag[row] - bdiag[row+1] - 1;         /* num of entries in U(row,:), excluding diagonal */
3521         for (j=0; j<nz; j++) rtmp[*pj++] -= multiplier * (*pv++);
3522         ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
3523       }
3524       k++;
3525     }
3526 
3527     /* finished row so stick it into b->a */
3528     /* L-part */
3529     pv  = b->a + bi[i];
3530     pj  = bj + bi[i];
3531     nzl = bi[i+1] - bi[i];
3532     for (j=0; j<nzl; j++) {
3533       pv[j] = rtmp[pj[j]];
3534       /* printf(" (%d,%g),",pj[j],pv[j]); */
3535     }
3536 
3537     /* diagonal: invert diagonal entries for simplier triangular solves */
3538     if (rtmp[i] == 0.0) rtmp[i] = dt+shift;
3539     b->a[bdiag[i]] = 1.0/rtmp[i];
3540     /* printf(" (%d,%g),",i,b->a[bdiag[i]]); */
3541 
3542     /* U-part */
3543     pv  = b->a + bdiag[i+1] + 1;
3544     pj  = bj + bdiag[i+1] + 1;
3545     nzu = bdiag[i] - bdiag[i+1] - 1;
3546     for (j=0; j<nzu; j++) {
3547       pv[j] = rtmp[pj[j]];
3548       /* printf(" (%d,%g),",pj[j],pv[j]); */
3549     }
3550     /* printf("\n"); */
3551   }
3552 
3553   ierr = PetscFree(rtmp);CHKERRQ(ierr);
3554   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
3555   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
3556 
3557   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
3558   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
3559   if (row_identity && col_identity) {
3560     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
3561   } else {
3562     C->ops->solve = MatSolve_SeqAIJ;
3563   }
3564   C->ops->solveadd          = 0;
3565   C->ops->solvetranspose    = 0;
3566   C->ops->solvetransposeadd = 0;
3567   C->ops->matsolve          = 0;
3568   C->assembled              = PETSC_TRUE;
3569   C->preallocated           = PETSC_TRUE;
3570 
3571   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
3572   PetscFunctionReturn(0);
3573 }
3574