xref: /petsc/src/mat/impls/aij/seq/aijfact.c (revision c87ba875e4007ad659b117ea274f03d5f4cd5ea7)
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     = PetscArrayzero(done,n);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 && !A->symmetric && (ftype == MAT_FACTOR_CHOLESKY||ftype == MAT_FACTOR_ICC)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Hermitian CHOLESKY or ICC 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   if (n==1)   f = 1; /* prevent failure in corner case of 1x1 matrix with fill < 0.5 */
165   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
166   current_space = free_space;
167 
168   for (i=0; i<n; i++) {
169     /* copy previous fill into linked list */
170     nzi = 0;
171     nnz = ai[r[i]+1] - ai[r[i]];
172     ajtmp = aj + ai[r[i]];
173     ierr  = PetscLLAddPerm(nnz,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
174     nzi  += nlnk;
175 
176     /* add pivot rows into linked list */
177     row = lnk[n];
178     while (row < i) {
179       nzbd  = bdiag[row] - bi[row] + 1;   /* num of entries in the row with column index <= row */
180       ajtmp = bi_ptr[row] + nzbd;   /* points to the entry next to the diagonal */
181       ierr  = PetscLLAddSortedLU(ajtmp,row,nlnk,lnk,lnkbt,i,nzbd,im);CHKERRQ(ierr);
182       nzi  += nlnk;
183       row   = lnk[row];
184     }
185     bi[i+1] = bi[i] + nzi;
186     im[i]   = nzi;
187 
188     /* mark bdiag */
189     nzbd = 0;
190     nnz  = nzi;
191     k    = lnk[n];
192     while (nnz-- && k < i) {
193       nzbd++;
194       k = lnk[k];
195     }
196     bdiag[i] = bi[i] + nzbd;
197 
198     /* if free space is not available, make more free space */
199     if (current_space->local_remaining<nzi) {
200       nnz  = PetscIntMultTruncate(n - i,nzi); /* estimated and max additional space needed */
201       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
202       reallocs++;
203     }
204 
205     /* copy data into free space, then initialize lnk */
206     ierr = PetscLLClean(n,n,nzi,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
207 
208     bi_ptr[i]                       = current_space->array;
209     current_space->array           += nzi;
210     current_space->local_used      += nzi;
211     current_space->local_remaining -= nzi;
212   }
213 #if defined(PETSC_USE_INFO)
214   if (ai[n] != 0) {
215     PetscReal af = ((PetscReal)bi[n])/((PetscReal)ai[n]);
216     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
217     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
218     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g);\n",(double)af);CHKERRQ(ierr);
219     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
220   } else {
221     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
222   }
223 #endif
224 
225   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
226   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
227 
228   /* destroy list of free space and other temporary array(s) */
229   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
230   ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr);
231   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
232   ierr = PetscFree2(bi_ptr,im);CHKERRQ(ierr);
233 
234   /* put together the new matrix */
235   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
236   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
237   b    = (Mat_SeqAIJ*)(B)->data;
238 
239   b->free_a       = PETSC_TRUE;
240   b->free_ij      = PETSC_TRUE;
241   b->singlemalloc = PETSC_FALSE;
242 
243   ierr    = PetscMalloc1(bi[n]+1,&b->a);CHKERRQ(ierr);
244   b->j    = bj;
245   b->i    = bi;
246   b->diag = bdiag;
247   b->ilen = 0;
248   b->imax = 0;
249   b->row  = isrow;
250   b->col  = iscol;
251   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
252   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
253   b->icol = isicol;
254   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
255 
256   /* In b structure:  Free imax, ilen, old a, old j.  Allocate solve_work, new a, new j */
257   ierr     = PetscLogObjectMemory((PetscObject)B,(bi[n]-n)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
258   b->maxnz = b->nz = bi[n];
259 
260   (B)->factortype            = MAT_FACTOR_LU;
261   (B)->info.factor_mallocs   = reallocs;
262   (B)->info.fill_ratio_given = f;
263 
264   if (ai[n]) {
265     (B)->info.fill_ratio_needed = ((PetscReal)bi[n])/((PetscReal)ai[n]);
266   } else {
267     (B)->info.fill_ratio_needed = 0.0;
268   }
269   (B)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_inplace;
270   if (a->inode.size) {
271     (B)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
272   }
273   PetscFunctionReturn(0);
274 }
275 
276 PetscErrorCode MatLUFactorSymbolic_SeqAIJ(Mat B,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
277 {
278   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
279   IS                 isicol;
280   PetscErrorCode     ierr;
281   const PetscInt     *r,*ic,*ai=a->i,*aj=a->j,*ajtmp;
282   PetscInt           i,n=A->rmap->n;
283   PetscInt           *bi,*bj;
284   PetscInt           *bdiag,row,nnz,nzi,reallocs=0,nzbd,*im;
285   PetscReal          f;
286   PetscInt           nlnk,*lnk,k,**bi_ptr;
287   PetscFreeSpaceList free_space=NULL,current_space=NULL;
288   PetscBT            lnkbt;
289   PetscBool          missing;
290 
291   PetscFunctionBegin;
292   if (A->rmap->N != A->cmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"matrix must be square");
293   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
294   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
295 
296   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
297   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
298   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
299 
300   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
301   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
302   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
303   bi[0] = bdiag[0] = 0;
304 
305   /* linked list for storing column indices of the active row */
306   nlnk = n + 1;
307   ierr = PetscLLCreate(n,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
308 
309   ierr = PetscMalloc2(n+1,&bi_ptr,n+1,&im);CHKERRQ(ierr);
310 
311   /* initial FreeSpace size is f*(ai[n]+1) */
312   f             = info->fill;
313   if (n==1)   f = 1; /* prevent failure in corner case of 1x1 matrix with fill < 0.5 */
314   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
315   current_space = free_space;
316 
317   for (i=0; i<n; i++) {
318     /* copy previous fill into linked list */
319     nzi = 0;
320     nnz = ai[r[i]+1] - ai[r[i]];
321     ajtmp = aj + ai[r[i]];
322     ierr  = PetscLLAddPerm(nnz,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
323     nzi  += nlnk;
324 
325     /* add pivot rows into linked list */
326     row = lnk[n];
327     while (row < i) {
328       nzbd  = bdiag[row] + 1; /* num of entries in the row with column index <= row */
329       ajtmp = bi_ptr[row] + nzbd; /* points to the entry next to the diagonal */
330       ierr  = PetscLLAddSortedLU(ajtmp,row,nlnk,lnk,lnkbt,i,nzbd,im);CHKERRQ(ierr);
331       nzi  += nlnk;
332       row   = lnk[row];
333     }
334     bi[i+1] = bi[i] + nzi;
335     im[i]   = nzi;
336 
337     /* mark bdiag */
338     nzbd = 0;
339     nnz  = nzi;
340     k    = lnk[n];
341     while (nnz-- && k < i) {
342       nzbd++;
343       k = lnk[k];
344     }
345     bdiag[i] = nzbd; /* note: bdiag[i] = nnzL as input for PetscFreeSpaceContiguous_LU() */
346 
347     /* if free space is not available, make more free space */
348     if (current_space->local_remaining<nzi) {
349       /* estimated additional space needed */
350       nnz  = PetscIntMultTruncate(2,PetscIntMultTruncate(n-1,nzi));
351       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
352       reallocs++;
353     }
354 
355     /* copy data into free space, then initialize lnk */
356     ierr = PetscLLClean(n,n,nzi,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
357 
358     bi_ptr[i]                       = current_space->array;
359     current_space->array           += nzi;
360     current_space->local_used      += nzi;
361     current_space->local_remaining -= nzi;
362   }
363 
364   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
365   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
366 
367   /*   copy free_space into bj and free free_space; set bi, bj, bdiag in new datastructure; */
368   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
369   ierr = PetscFreeSpaceContiguous_LU(&free_space,bj,n,bi,bdiag);CHKERRQ(ierr);
370   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
371   ierr = PetscFree2(bi_ptr,im);CHKERRQ(ierr);
372 
373   /* put together the new matrix */
374   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
375   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
376   b    = (Mat_SeqAIJ*)(B)->data;
377 
378   b->free_a       = PETSC_TRUE;
379   b->free_ij      = PETSC_TRUE;
380   b->singlemalloc = PETSC_FALSE;
381 
382   ierr = PetscMalloc1(bdiag[0]+1,&b->a);CHKERRQ(ierr);
383 
384   b->j    = bj;
385   b->i    = bi;
386   b->diag = bdiag;
387   b->ilen = 0;
388   b->imax = 0;
389   b->row  = isrow;
390   b->col  = iscol;
391   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
392   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
393   b->icol = isicol;
394   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
395 
396   /* In b structure:  Free imax, ilen, old a, old j.  Allocate solve_work, new a, new j */
397   ierr     = PetscLogObjectMemory((PetscObject)B,(bdiag[0]+1)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
398   b->maxnz = b->nz = bdiag[0]+1;
399 
400   B->factortype            = MAT_FACTOR_LU;
401   B->info.factor_mallocs   = reallocs;
402   B->info.fill_ratio_given = f;
403 
404   if (ai[n]) {
405     B->info.fill_ratio_needed = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
406   } else {
407     B->info.fill_ratio_needed = 0.0;
408   }
409 #if defined(PETSC_USE_INFO)
410   if (ai[n] != 0) {
411     PetscReal af = B->info.fill_ratio_needed;
412     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
413     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
414     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g);\n",(double)af);CHKERRQ(ierr);
415     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
416   } else {
417     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
418   }
419 #endif
420   B->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ;
421   if (a->inode.size) {
422     B->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
423   }
424   ierr = MatSeqAIJCheckInode_FactorLU(B);CHKERRQ(ierr);
425   PetscFunctionReturn(0);
426 }
427 
428 /*
429     Trouble in factorization, should we dump the original matrix?
430 */
431 PetscErrorCode MatFactorDumpMatrix(Mat A)
432 {
433   PetscErrorCode ierr;
434   PetscBool      flg = PETSC_FALSE;
435 
436   PetscFunctionBegin;
437   ierr = PetscOptionsGetBool(((PetscObject)A)->options,NULL,"-mat_factor_dump_on_error",&flg,NULL);CHKERRQ(ierr);
438   if (flg) {
439     PetscViewer viewer;
440     char        filename[PETSC_MAX_PATH_LEN];
441 
442     ierr = PetscSNPrintf(filename,PETSC_MAX_PATH_LEN,"matrix_factor_error.%d",PetscGlobalRank);CHKERRQ(ierr);
443     ierr = PetscViewerBinaryOpen(PetscObjectComm((PetscObject)A),filename,FILE_MODE_WRITE,&viewer);CHKERRQ(ierr);
444     ierr = MatView(A,viewer);CHKERRQ(ierr);
445     ierr = PetscViewerDestroy(&viewer);CHKERRQ(ierr);
446   }
447   PetscFunctionReturn(0);
448 }
449 
450 PetscErrorCode MatLUFactorNumeric_SeqAIJ(Mat B,Mat A,const MatFactorInfo *info)
451 {
452   Mat             C     =B;
453   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
454   IS              isrow = b->row,isicol = b->icol;
455   PetscErrorCode  ierr;
456   const PetscInt  *r,*ic,*ics;
457   const PetscInt  n=A->rmap->n,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j,*bdiag=b->diag;
458   PetscInt        i,j,k,nz,nzL,row,*pj;
459   const PetscInt  *ajtmp,*bjtmp;
460   MatScalar       *rtmp,*pc,multiplier,*pv;
461   const MatScalar *aa=a->a,*v;
462   PetscBool       row_identity,col_identity;
463   FactorShiftCtx  sctx;
464   const PetscInt  *ddiag;
465   PetscReal       rs;
466   MatScalar       d;
467 
468   PetscFunctionBegin;
469   /* MatPivotSetUp(): initialize shift context sctx */
470   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
471 
472   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
473     ddiag          = a->diag;
474     sctx.shift_top = info->zeropivot;
475     for (i=0; i<n; i++) {
476       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
477       d  = (aa)[ddiag[i]];
478       rs = -PetscAbsScalar(d) - PetscRealPart(d);
479       v  = aa+ai[i];
480       nz = ai[i+1] - ai[i];
481       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
482       if (rs>sctx.shift_top) sctx.shift_top = rs;
483     }
484     sctx.shift_top *= 1.1;
485     sctx.nshift_max = 5;
486     sctx.shift_lo   = 0.;
487     sctx.shift_hi   = 1.;
488   }
489 
490   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
491   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
492   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
493   ics  = ic;
494 
495   do {
496     sctx.newshift = PETSC_FALSE;
497     for (i=0; i<n; i++) {
498       /* zero rtmp */
499       /* L part */
500       nz    = bi[i+1] - bi[i];
501       bjtmp = bj + bi[i];
502       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
503 
504       /* U part */
505       nz    = bdiag[i]-bdiag[i+1];
506       bjtmp = bj + bdiag[i+1]+1;
507       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
508 
509       /* load in initial (unfactored row) */
510       nz    = ai[r[i]+1] - ai[r[i]];
511       ajtmp = aj + ai[r[i]];
512       v     = aa + ai[r[i]];
513       for (j=0; j<nz; j++) {
514         rtmp[ics[ajtmp[j]]] = v[j];
515       }
516       /* ZeropivotApply() */
517       rtmp[i] += sctx.shift_amount;  /* shift the diagonal of the matrix */
518 
519       /* elimination */
520       bjtmp = bj + bi[i];
521       row   = *bjtmp++;
522       nzL   = bi[i+1] - bi[i];
523       for (k=0; k < nzL; k++) {
524         pc = rtmp + row;
525         if (*pc != 0.0) {
526           pv         = b->a + bdiag[row];
527           multiplier = *pc * (*pv);
528           *pc        = multiplier;
529 
530           pj = b->j + bdiag[row+1]+1; /* beginning of U(row,:) */
531           pv = b->a + bdiag[row+1]+1;
532           nz = bdiag[row]-bdiag[row+1]-1; /* num of entries in U(row,:) excluding diag */
533 
534           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
535           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
536         }
537         row = *bjtmp++;
538       }
539 
540       /* finished row so stick it into b->a */
541       rs = 0.0;
542       /* L part */
543       pv = b->a + bi[i];
544       pj = b->j + bi[i];
545       nz = bi[i+1] - bi[i];
546       for (j=0; j<nz; j++) {
547         pv[j] = rtmp[pj[j]]; rs += PetscAbsScalar(pv[j]);
548       }
549 
550       /* U part */
551       pv = b->a + bdiag[i+1]+1;
552       pj = b->j + bdiag[i+1]+1;
553       nz = bdiag[i] - bdiag[i+1]-1;
554       for (j=0; j<nz; j++) {
555         pv[j] = rtmp[pj[j]]; rs += PetscAbsScalar(pv[j]);
556       }
557 
558       sctx.rs = rs;
559       sctx.pv = rtmp[i];
560       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
561       if (sctx.newshift) break; /* break for-loop */
562       rtmp[i] = sctx.pv; /* sctx.pv might be updated in the case of MAT_SHIFT_INBLOCKS */
563 
564       /* Mark diagonal and invert diagonal for simplier triangular solves */
565       pv  = b->a + bdiag[i];
566       *pv = 1.0/rtmp[i];
567 
568     } /* endof for (i=0; i<n; i++) { */
569 
570     /* MatPivotRefine() */
571     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
572       /*
573        * if no shift in this attempt & shifting & started shifting & can refine,
574        * then try lower shift
575        */
576       sctx.shift_hi       = sctx.shift_fraction;
577       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
578       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
579       sctx.newshift       = PETSC_TRUE;
580       sctx.nshift++;
581     }
582   } while (sctx.newshift);
583 
584   ierr = PetscFree(rtmp);CHKERRQ(ierr);
585   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
586   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
587 
588   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
589   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
590   if (b->inode.size) {
591     C->ops->solve = MatSolve_SeqAIJ_Inode;
592   } else if (row_identity && col_identity) {
593     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
594   } else {
595     C->ops->solve = MatSolve_SeqAIJ;
596   }
597   C->ops->solveadd          = MatSolveAdd_SeqAIJ;
598   C->ops->solvetranspose    = MatSolveTranspose_SeqAIJ;
599   C->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ;
600   C->ops->matsolve          = MatMatSolve_SeqAIJ;
601   C->assembled              = PETSC_TRUE;
602   C->preallocated           = PETSC_TRUE;
603 
604   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
605 
606   /* MatShiftView(A,info,&sctx) */
607   if (sctx.nshift) {
608     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
609       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);
610     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
611       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
612     } else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) {
613       ierr = PetscInfo2(A,"number of shift_inblocks applied %D, each shift_amount %g\n",sctx.nshift,(double)info->shiftamount);CHKERRQ(ierr);
614     }
615   }
616   PetscFunctionReturn(0);
617 }
618 
619 PetscErrorCode MatLUFactorNumeric_SeqAIJ_inplace(Mat B,Mat A,const MatFactorInfo *info)
620 {
621   Mat             C     =B;
622   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
623   IS              isrow = b->row,isicol = b->icol;
624   PetscErrorCode  ierr;
625   const PetscInt  *r,*ic,*ics;
626   PetscInt        nz,row,i,j,n=A->rmap->n,diag;
627   const PetscInt  *ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j;
628   const PetscInt  *ajtmp,*bjtmp,*diag_offset = b->diag,*pj;
629   MatScalar       *pv,*rtmp,*pc,multiplier,d;
630   const MatScalar *v,*aa=a->a;
631   PetscReal       rs=0.0;
632   FactorShiftCtx  sctx;
633   const PetscInt  *ddiag;
634   PetscBool       row_identity, col_identity;
635 
636   PetscFunctionBegin;
637   /* MatPivotSetUp(): initialize shift context sctx */
638   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
639 
640   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
641     ddiag          = a->diag;
642     sctx.shift_top = info->zeropivot;
643     for (i=0; i<n; i++) {
644       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
645       d  = (aa)[ddiag[i]];
646       rs = -PetscAbsScalar(d) - PetscRealPart(d);
647       v  = aa+ai[i];
648       nz = ai[i+1] - ai[i];
649       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
650       if (rs>sctx.shift_top) sctx.shift_top = rs;
651     }
652     sctx.shift_top *= 1.1;
653     sctx.nshift_max = 5;
654     sctx.shift_lo   = 0.;
655     sctx.shift_hi   = 1.;
656   }
657 
658   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
659   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
660   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
661   ics  = ic;
662 
663   do {
664     sctx.newshift = PETSC_FALSE;
665     for (i=0; i<n; i++) {
666       nz    = bi[i+1] - bi[i];
667       bjtmp = bj + bi[i];
668       for  (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
669 
670       /* load in initial (unfactored row) */
671       nz    = ai[r[i]+1] - ai[r[i]];
672       ajtmp = aj + ai[r[i]];
673       v     = aa + ai[r[i]];
674       for (j=0; j<nz; j++) {
675         rtmp[ics[ajtmp[j]]] = v[j];
676       }
677       rtmp[ics[r[i]]] += sctx.shift_amount; /* shift the diagonal of the matrix */
678 
679       row = *bjtmp++;
680       while  (row < i) {
681         pc = rtmp + row;
682         if (*pc != 0.0) {
683           pv         = b->a + diag_offset[row];
684           pj         = b->j + diag_offset[row] + 1;
685           multiplier = *pc / *pv++;
686           *pc        = multiplier;
687           nz         = bi[row+1] - diag_offset[row] - 1;
688           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
689           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
690         }
691         row = *bjtmp++;
692       }
693       /* finished row so stick it into b->a */
694       pv   = b->a + bi[i];
695       pj   = b->j + bi[i];
696       nz   = bi[i+1] - bi[i];
697       diag = diag_offset[i] - bi[i];
698       rs   = 0.0;
699       for (j=0; j<nz; j++) {
700         pv[j] = rtmp[pj[j]];
701         rs   += PetscAbsScalar(pv[j]);
702       }
703       rs -= PetscAbsScalar(pv[diag]);
704 
705       sctx.rs = rs;
706       sctx.pv = pv[diag];
707       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
708       if (sctx.newshift) break;
709       pv[diag] = sctx.pv;
710     }
711 
712     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
713       /*
714        * if no shift in this attempt & shifting & started shifting & can refine,
715        * then try lower shift
716        */
717       sctx.shift_hi       = sctx.shift_fraction;
718       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
719       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
720       sctx.newshift       = PETSC_TRUE;
721       sctx.nshift++;
722     }
723   } while (sctx.newshift);
724 
725   /* invert diagonal entries for simplier triangular solves */
726   for (i=0; i<n; i++) {
727     b->a[diag_offset[i]] = 1.0/b->a[diag_offset[i]];
728   }
729   ierr = PetscFree(rtmp);CHKERRQ(ierr);
730   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
731   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
732 
733   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
734   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
735   if (row_identity && col_identity) {
736     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering_inplace;
737   } else {
738     C->ops->solve = MatSolve_SeqAIJ_inplace;
739   }
740   C->ops->solveadd          = MatSolveAdd_SeqAIJ_inplace;
741   C->ops->solvetranspose    = MatSolveTranspose_SeqAIJ_inplace;
742   C->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ_inplace;
743   C->ops->matsolve          = MatMatSolve_SeqAIJ_inplace;
744 
745   C->assembled    = PETSC_TRUE;
746   C->preallocated = PETSC_TRUE;
747 
748   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
749   if (sctx.nshift) {
750     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
751       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);
752     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
753       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
754     }
755   }
756   (C)->ops->solve          = MatSolve_SeqAIJ_inplace;
757   (C)->ops->solvetranspose = MatSolveTranspose_SeqAIJ_inplace;
758 
759   ierr = MatSeqAIJCheckInode(C);CHKERRQ(ierr);
760   PetscFunctionReturn(0);
761 }
762 
763 /*
764    This routine implements inplace ILU(0) with row or/and column permutations.
765    Input:
766      A - original matrix
767    Output;
768      A - a->i (rowptr) is same as original rowptr, but factored i-the row is stored in rowperm[i]
769          a->j (col index) is permuted by the inverse of colperm, then sorted
770          a->a reordered accordingly with a->j
771          a->diag (ptr to diagonal elements) is updated.
772 */
773 PetscErrorCode MatLUFactorNumeric_SeqAIJ_InplaceWithPerm(Mat B,Mat A,const MatFactorInfo *info)
774 {
775   Mat_SeqAIJ      *a    =(Mat_SeqAIJ*)A->data;
776   IS              isrow = a->row,isicol = a->icol;
777   PetscErrorCode  ierr;
778   const PetscInt  *r,*ic,*ics;
779   PetscInt        i,j,n=A->rmap->n,*ai=a->i,*aj=a->j;
780   PetscInt        *ajtmp,nz,row;
781   PetscInt        *diag = a->diag,nbdiag,*pj;
782   PetscScalar     *rtmp,*pc,multiplier,d;
783   MatScalar       *pv,*v;
784   PetscReal       rs;
785   FactorShiftCtx  sctx;
786   const MatScalar *aa=a->a,*vtmp;
787 
788   PetscFunctionBegin;
789   if (A != B) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"input and output matrix must have same address");
790 
791   /* MatPivotSetUp(): initialize shift context sctx */
792   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
793 
794   if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
795     const PetscInt *ddiag = a->diag;
796     sctx.shift_top = info->zeropivot;
797     for (i=0; i<n; i++) {
798       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
799       d    = (aa)[ddiag[i]];
800       rs   = -PetscAbsScalar(d) - PetscRealPart(d);
801       vtmp = aa+ai[i];
802       nz   = ai[i+1] - ai[i];
803       for (j=0; j<nz; j++) rs += PetscAbsScalar(vtmp[j]);
804       if (rs>sctx.shift_top) sctx.shift_top = rs;
805     }
806     sctx.shift_top *= 1.1;
807     sctx.nshift_max = 5;
808     sctx.shift_lo   = 0.;
809     sctx.shift_hi   = 1.;
810   }
811 
812   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
813   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
814   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
815   ierr = PetscArrayzero(rtmp,n+1);CHKERRQ(ierr);
816   ics  = ic;
817 
818 #if defined(MV)
819   sctx.shift_top      = 0.;
820   sctx.nshift_max     = 0;
821   sctx.shift_lo       = 0.;
822   sctx.shift_hi       = 0.;
823   sctx.shift_fraction = 0.;
824 
825   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
826     sctx.shift_top = 0.;
827     for (i=0; i<n; i++) {
828       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
829       d  = (a->a)[diag[i]];
830       rs = -PetscAbsScalar(d) - PetscRealPart(d);
831       v  = a->a+ai[i];
832       nz = ai[i+1] - ai[i];
833       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
834       if (rs>sctx.shift_top) sctx.shift_top = rs;
835     }
836     if (sctx.shift_top < info->zeropivot) sctx.shift_top = info->zeropivot;
837     sctx.shift_top *= 1.1;
838     sctx.nshift_max = 5;
839     sctx.shift_lo   = 0.;
840     sctx.shift_hi   = 1.;
841   }
842 
843   sctx.shift_amount = 0.;
844   sctx.nshift       = 0;
845 #endif
846 
847   do {
848     sctx.newshift = PETSC_FALSE;
849     for (i=0; i<n; i++) {
850       /* load in initial unfactored row */
851       nz    = ai[r[i]+1] - ai[r[i]];
852       ajtmp = aj + ai[r[i]];
853       v     = a->a + ai[r[i]];
854       /* sort permuted ajtmp and values v accordingly */
855       for (j=0; j<nz; j++) ajtmp[j] = ics[ajtmp[j]];
856       ierr = PetscSortIntWithScalarArray(nz,ajtmp,v);CHKERRQ(ierr);
857 
858       diag[r[i]] = ai[r[i]];
859       for (j=0; j<nz; j++) {
860         rtmp[ajtmp[j]] = v[j];
861         if (ajtmp[j] < i) diag[r[i]]++; /* update a->diag */
862       }
863       rtmp[r[i]] += sctx.shift_amount; /* shift the diagonal of the matrix */
864 
865       row = *ajtmp++;
866       while  (row < i) {
867         pc = rtmp + row;
868         if (*pc != 0.0) {
869           pv = a->a + diag[r[row]];
870           pj = aj + diag[r[row]] + 1;
871 
872           multiplier = *pc / *pv++;
873           *pc        = multiplier;
874           nz         = ai[r[row]+1] - diag[r[row]] - 1;
875           for (j=0; j<nz; j++) rtmp[pj[j]] -= multiplier * pv[j];
876           ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
877         }
878         row = *ajtmp++;
879       }
880       /* finished row so overwrite it onto a->a */
881       pv     = a->a + ai[r[i]];
882       pj     = aj + ai[r[i]];
883       nz     = ai[r[i]+1] - ai[r[i]];
884       nbdiag = diag[r[i]] - ai[r[i]]; /* num of entries before the diagonal */
885 
886       rs = 0.0;
887       for (j=0; j<nz; j++) {
888         pv[j] = rtmp[pj[j]];
889         if (j != nbdiag) rs += PetscAbsScalar(pv[j]);
890       }
891 
892       sctx.rs = rs;
893       sctx.pv = pv[nbdiag];
894       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
895       if (sctx.newshift) break;
896       pv[nbdiag] = sctx.pv;
897     }
898 
899     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE && !sctx.newshift && sctx.shift_fraction>0 && sctx.nshift<sctx.nshift_max) {
900       /*
901        * if no shift in this attempt & shifting & started shifting & can refine,
902        * then try lower shift
903        */
904       sctx.shift_hi       = sctx.shift_fraction;
905       sctx.shift_fraction = (sctx.shift_hi+sctx.shift_lo)/2.;
906       sctx.shift_amount   = sctx.shift_fraction * sctx.shift_top;
907       sctx.newshift       = PETSC_TRUE;
908       sctx.nshift++;
909     }
910   } while (sctx.newshift);
911 
912   /* invert diagonal entries for simplier triangular solves */
913   for (i=0; i<n; i++) {
914     a->a[diag[r[i]]] = 1.0/a->a[diag[r[i]]];
915   }
916 
917   ierr = PetscFree(rtmp);CHKERRQ(ierr);
918   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
919   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
920 
921   A->ops->solve             = MatSolve_SeqAIJ_InplaceWithPerm;
922   A->ops->solveadd          = MatSolveAdd_SeqAIJ_inplace;
923   A->ops->solvetranspose    = MatSolveTranspose_SeqAIJ_inplace;
924   A->ops->solvetransposeadd = MatSolveTransposeAdd_SeqAIJ_inplace;
925 
926   A->assembled    = PETSC_TRUE;
927   A->preallocated = PETSC_TRUE;
928 
929   ierr = PetscLogFlops(A->cmap->n);CHKERRQ(ierr);
930   if (sctx.nshift) {
931     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
932       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);
933     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
934       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
935     }
936   }
937   PetscFunctionReturn(0);
938 }
939 
940 /* ----------------------------------------------------------- */
941 PetscErrorCode MatLUFactor_SeqAIJ(Mat A,IS row,IS col,const MatFactorInfo *info)
942 {
943   PetscErrorCode ierr;
944   Mat            C;
945 
946   PetscFunctionBegin;
947   ierr = MatGetFactor(A,MATSOLVERPETSC,MAT_FACTOR_LU,&C);CHKERRQ(ierr);
948   ierr = MatLUFactorSymbolic(C,A,row,col,info);CHKERRQ(ierr);
949   ierr = MatLUFactorNumeric(C,A,info);CHKERRQ(ierr);
950 
951   A->ops->solve          = C->ops->solve;
952   A->ops->solvetranspose = C->ops->solvetranspose;
953 
954   ierr = MatHeaderMerge(A,&C);CHKERRQ(ierr);
955   ierr = PetscLogObjectParent((PetscObject)A,(PetscObject)((Mat_SeqAIJ*)(A->data))->icol);CHKERRQ(ierr);
956   PetscFunctionReturn(0);
957 }
958 /* ----------------------------------------------------------- */
959 
960 
961 PetscErrorCode MatSolve_SeqAIJ_inplace(Mat A,Vec bb,Vec xx)
962 {
963   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
964   IS                iscol = a->col,isrow = a->row;
965   PetscErrorCode    ierr;
966   PetscInt          i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
967   PetscInt          nz;
968   const PetscInt    *rout,*cout,*r,*c;
969   PetscScalar       *x,*tmp,*tmps,sum;
970   const PetscScalar *b;
971   const MatScalar   *aa = a->a,*v;
972 
973   PetscFunctionBegin;
974   if (!n) PetscFunctionReturn(0);
975 
976   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
977   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
978   tmp  = a->solve_work;
979 
980   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
981   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
982 
983   /* forward solve the lower triangular */
984   tmp[0] = b[*r++];
985   tmps   = tmp;
986   for (i=1; i<n; i++) {
987     v   = aa + ai[i];
988     vi  = aj + ai[i];
989     nz  = a->diag[i] - ai[i];
990     sum = b[*r++];
991     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
992     tmp[i] = sum;
993   }
994 
995   /* backward solve the upper triangular */
996   for (i=n-1; i>=0; i--) {
997     v   = aa + a->diag[i] + 1;
998     vi  = aj + a->diag[i] + 1;
999     nz  = ai[i+1] - a->diag[i] - 1;
1000     sum = tmp[i];
1001     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1002     x[*c--] = tmp[i] = sum*aa[a->diag[i]];
1003   }
1004 
1005   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1006   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1007   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1008   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
1009   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1010   PetscFunctionReturn(0);
1011 }
1012 
1013 PetscErrorCode MatMatSolve_SeqAIJ_inplace(Mat A,Mat B,Mat X)
1014 {
1015   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1016   IS                iscol = a->col,isrow = a->row;
1017   PetscErrorCode    ierr;
1018   PetscInt          i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
1019   PetscInt          nz,neq,ldb,ldx;
1020   const PetscInt    *rout,*cout,*r,*c;
1021   PetscScalar       *x,*tmp = a->solve_work,*tmps,sum;
1022   const PetscScalar *b,*aa = a->a,*v;
1023   PetscBool         isdense;
1024 
1025   PetscFunctionBegin;
1026   if (!n) PetscFunctionReturn(0);
1027   ierr = PetscObjectTypeCompare((PetscObject)B,MATSEQDENSE,&isdense);CHKERRQ(ierr);
1028   if (!isdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"B matrix must be a SeqDense matrix");
1029   if (X != B) {
1030     ierr = PetscObjectTypeCompare((PetscObject)X,MATSEQDENSE,&isdense);CHKERRQ(ierr);
1031     if (!isdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"X matrix must be a SeqDense matrix");
1032   }
1033   ierr = MatDenseGetArrayRead(B,&b);CHKERRQ(ierr);
1034   ierr = MatDenseGetLDA(B,&ldb);CHKERRQ(ierr);
1035   ierr = MatDenseGetArray(X,&x);CHKERRQ(ierr);
1036   ierr = MatDenseGetLDA(X,&ldx);CHKERRQ(ierr);
1037   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1038   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1039   for (neq=0; neq<B->cmap->n; neq++) {
1040     /* forward solve the lower triangular */
1041     tmp[0] = b[r[0]];
1042     tmps   = tmp;
1043     for (i=1; i<n; i++) {
1044       v   = aa + ai[i];
1045       vi  = aj + ai[i];
1046       nz  = a->diag[i] - ai[i];
1047       sum = b[r[i]];
1048       PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1049       tmp[i] = sum;
1050     }
1051     /* backward solve the upper triangular */
1052     for (i=n-1; i>=0; i--) {
1053       v   = aa + a->diag[i] + 1;
1054       vi  = aj + a->diag[i] + 1;
1055       nz  = ai[i+1] - a->diag[i] - 1;
1056       sum = tmp[i];
1057       PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1058       x[c[i]] = tmp[i] = sum*aa[a->diag[i]];
1059     }
1060     b += ldb;
1061     x += ldx;
1062   }
1063   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1064   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1065   ierr = MatDenseRestoreArrayRead(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,ldb,ldx;
1078   const PetscInt    *rout,*cout,*r,*c;
1079   PetscScalar       *x,*tmp = a->solve_work,sum;
1080   const PetscScalar *b,*aa = a->a,*v;
1081   PetscBool         isdense;
1082 
1083   PetscFunctionBegin;
1084   if (!n) PetscFunctionReturn(0);
1085   ierr = PetscObjectTypeCompare((PetscObject)B,MATSEQDENSE,&isdense);CHKERRQ(ierr);
1086   if (!isdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"B matrix must be a SeqDense matrix");
1087   if (X != B) {
1088     ierr = PetscObjectTypeCompare((PetscObject)X,MATSEQDENSE,&isdense);CHKERRQ(ierr);
1089     if (!isdense) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"X matrix must be a SeqDense matrix");
1090   }
1091   ierr = MatDenseGetArrayRead(B,&b);CHKERRQ(ierr);
1092   ierr = MatDenseGetLDA(B,&ldb);CHKERRQ(ierr);
1093   ierr = MatDenseGetArray(X,&x);CHKERRQ(ierr);
1094   ierr = MatDenseGetLDA(X,&ldx);CHKERRQ(ierr);
1095   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1096   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1097   for (neq=0; neq<B->cmap->n; neq++) {
1098     /* forward solve the lower triangular */
1099     tmp[0] = b[r[0]];
1100     v      = aa;
1101     vi     = aj;
1102     for (i=1; i<n; i++) {
1103       nz  = ai[i+1] - ai[i];
1104       sum = b[r[i]];
1105       PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
1106       tmp[i] = sum;
1107       v     += nz; vi += nz;
1108     }
1109     /* backward solve the upper triangular */
1110     for (i=n-1; i>=0; i--) {
1111       v   = aa + adiag[i+1]+1;
1112       vi  = aj + adiag[i+1]+1;
1113       nz  = adiag[i]-adiag[i+1]-1;
1114       sum = tmp[i];
1115       PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
1116       x[c[i]] = tmp[i] = sum*v[nz]; /* v[nz] = aa[adiag[i]] */
1117     }
1118     b += ldb;
1119     x += ldx;
1120   }
1121   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1122   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1123   ierr = MatDenseRestoreArrayRead(B,&b);CHKERRQ(ierr);
1124   ierr = MatDenseRestoreArray(X,&x);CHKERRQ(ierr);
1125   ierr = PetscLogFlops(B->cmap->n*(2.0*a->nz - n));CHKERRQ(ierr);
1126   PetscFunctionReturn(0);
1127 }
1128 
1129 PetscErrorCode MatSolve_SeqAIJ_InplaceWithPerm(Mat A,Vec bb,Vec xx)
1130 {
1131   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1132   IS                iscol = a->col,isrow = a->row;
1133   PetscErrorCode    ierr;
1134   const PetscInt    *r,*c,*rout,*cout;
1135   PetscInt          i, n = A->rmap->n,*vi,*ai = a->i,*aj = a->j;
1136   PetscInt          nz,row;
1137   PetscScalar       *x,*tmp,*tmps,sum;
1138   const PetscScalar *b;
1139   const MatScalar   *aa = a->a,*v;
1140 
1141   PetscFunctionBegin;
1142   if (!n) PetscFunctionReturn(0);
1143 
1144   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1145   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
1146   tmp  = a->solve_work;
1147 
1148   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1149   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
1150 
1151   /* forward solve the lower triangular */
1152   tmp[0] = b[*r++];
1153   tmps   = tmp;
1154   for (row=1; row<n; row++) {
1155     i   = rout[row]; /* permuted row */
1156     v   = aa + ai[i];
1157     vi  = aj + ai[i];
1158     nz  = a->diag[i] - ai[i];
1159     sum = b[*r++];
1160     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1161     tmp[row] = sum;
1162   }
1163 
1164   /* backward solve the upper triangular */
1165   for (row=n-1; row>=0; row--) {
1166     i   = rout[row]; /* permuted row */
1167     v   = aa + a->diag[i] + 1;
1168     vi  = aj + a->diag[i] + 1;
1169     nz  = ai[i+1] - a->diag[i] - 1;
1170     sum = tmp[row];
1171     PetscSparseDenseMinusDot(sum,tmps,v,vi,nz);
1172     x[*c--] = tmp[row] = sum*aa[a->diag[i]];
1173   }
1174 
1175   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1176   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1177   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1178   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
1179   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1180   PetscFunctionReturn(0);
1181 }
1182 
1183 /* ----------------------------------------------------------- */
1184 #include <../src/mat/impls/aij/seq/ftn-kernels/fsolve.h>
1185 PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering_inplace(Mat A,Vec bb,Vec xx)
1186 {
1187   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
1188   PetscErrorCode    ierr;
1189   PetscInt          n   = A->rmap->n;
1190   const PetscInt    *ai = a->i,*aj = a->j,*adiag = a->diag;
1191   PetscScalar       *x;
1192   const PetscScalar *b;
1193   const MatScalar   *aa = a->a;
1194 #if !defined(PETSC_USE_FORTRAN_KERNEL_SOLVEAIJ)
1195   PetscInt        adiag_i,i,nz,ai_i;
1196   const PetscInt  *vi;
1197   const MatScalar *v;
1198   PetscScalar     sum;
1199 #endif
1200 
1201   PetscFunctionBegin;
1202   if (!n) PetscFunctionReturn(0);
1203 
1204   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1205   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
1206 
1207 #if defined(PETSC_USE_FORTRAN_KERNEL_SOLVEAIJ)
1208   fortransolveaij_(&n,x,ai,aj,adiag,aa,b);
1209 #else
1210   /* forward solve the lower triangular */
1211   x[0] = b[0];
1212   for (i=1; i<n; i++) {
1213     ai_i = ai[i];
1214     v    = aa + ai_i;
1215     vi   = aj + ai_i;
1216     nz   = adiag[i] - ai_i;
1217     sum  = b[i];
1218     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
1219     x[i] = sum;
1220   }
1221 
1222   /* backward solve the upper triangular */
1223   for (i=n-1; i>=0; i--) {
1224     adiag_i = adiag[i];
1225     v       = aa + adiag_i + 1;
1226     vi      = aj + adiag_i + 1;
1227     nz      = ai[i+1] - adiag_i - 1;
1228     sum     = x[i];
1229     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
1230     x[i] = sum*aa[adiag_i];
1231   }
1232 #endif
1233   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
1234   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1235   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
1236   PetscFunctionReturn(0);
1237 }
1238 
1239 PetscErrorCode MatSolveAdd_SeqAIJ_inplace(Mat A,Vec bb,Vec yy,Vec xx)
1240 {
1241   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1242   IS                iscol = a->col,isrow = a->row;
1243   PetscErrorCode    ierr;
1244   PetscInt          i, n = A->rmap->n,j;
1245   PetscInt          nz;
1246   const PetscInt    *rout,*cout,*r,*c,*vi,*ai = a->i,*aj = a->j;
1247   PetscScalar       *x,*tmp,sum;
1248   const PetscScalar *b;
1249   const MatScalar   *aa = a->a,*v;
1250 
1251   PetscFunctionBegin;
1252   if (yy != xx) {ierr = VecCopy(yy,xx);CHKERRQ(ierr);}
1253 
1254   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1255   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1256   tmp  = a->solve_work;
1257 
1258   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1259   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout + (n-1);
1260 
1261   /* forward solve the lower triangular */
1262   tmp[0] = b[*r++];
1263   for (i=1; i<n; i++) {
1264     v   = aa + ai[i];
1265     vi  = aj + ai[i];
1266     nz  = a->diag[i] - ai[i];
1267     sum = b[*r++];
1268     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1269     tmp[i] = sum;
1270   }
1271 
1272   /* backward solve the upper triangular */
1273   for (i=n-1; i>=0; i--) {
1274     v   = aa + a->diag[i] + 1;
1275     vi  = aj + a->diag[i] + 1;
1276     nz  = ai[i+1] - a->diag[i] - 1;
1277     sum = tmp[i];
1278     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1279     tmp[i]   = sum*aa[a->diag[i]];
1280     x[*c--] += tmp[i];
1281   }
1282 
1283   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1284   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1285   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1286   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1287   ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr);
1288   PetscFunctionReturn(0);
1289 }
1290 
1291 PetscErrorCode MatSolveAdd_SeqAIJ(Mat A,Vec bb,Vec yy,Vec xx)
1292 {
1293   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1294   IS                iscol = a->col,isrow = a->row;
1295   PetscErrorCode    ierr;
1296   PetscInt          i, n = A->rmap->n,j;
1297   PetscInt          nz;
1298   const PetscInt    *rout,*cout,*r,*c,*vi,*ai = a->i,*aj = a->j,*adiag = a->diag;
1299   PetscScalar       *x,*tmp,sum;
1300   const PetscScalar *b;
1301   const MatScalar   *aa = a->a,*v;
1302 
1303   PetscFunctionBegin;
1304   if (yy != xx) {ierr = VecCopy(yy,xx);CHKERRQ(ierr);}
1305 
1306   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1307   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1308   tmp  = a->solve_work;
1309 
1310   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1311   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1312 
1313   /* forward solve the lower triangular */
1314   tmp[0] = b[r[0]];
1315   v      = aa;
1316   vi     = aj;
1317   for (i=1; i<n; i++) {
1318     nz  = ai[i+1] - ai[i];
1319     sum = b[r[i]];
1320     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1321     tmp[i] = sum;
1322     v     += nz;
1323     vi    += nz;
1324   }
1325 
1326   /* backward solve the upper triangular */
1327   v  = aa + adiag[n-1];
1328   vi = aj + adiag[n-1];
1329   for (i=n-1; i>=0; i--) {
1330     nz  = adiag[i] - adiag[i+1] - 1;
1331     sum = tmp[i];
1332     for (j=0; j<nz; j++) sum -= v[j]*tmp[vi[j]];
1333     tmp[i]   = sum*v[nz];
1334     x[c[i]] += tmp[i];
1335     v       += nz+1; vi += nz+1;
1336   }
1337 
1338   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1339   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1340   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1341   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1342   ierr = PetscLogFlops(2.0*a->nz);CHKERRQ(ierr);
1343   PetscFunctionReturn(0);
1344 }
1345 
1346 PetscErrorCode MatSolveTranspose_SeqAIJ_inplace(Mat A,Vec bb,Vec xx)
1347 {
1348   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1349   IS                iscol = a->col,isrow = a->row;
1350   PetscErrorCode    ierr;
1351   const PetscInt    *rout,*cout,*r,*c,*diag = a->diag,*ai = a->i,*aj = a->j,*vi;
1352   PetscInt          i,n = A->rmap->n,j;
1353   PetscInt          nz;
1354   PetscScalar       *x,*tmp,s1;
1355   const MatScalar   *aa = a->a,*v;
1356   const PetscScalar *b;
1357 
1358   PetscFunctionBegin;
1359   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1360   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
1361   tmp  = a->solve_work;
1362 
1363   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1364   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1365 
1366   /* copy the b into temp work space according to permutation */
1367   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1368 
1369   /* forward solve the U^T */
1370   for (i=0; i<n; i++) {
1371     v   = aa + diag[i];
1372     vi  = aj + diag[i] + 1;
1373     nz  = ai[i+1] - diag[i] - 1;
1374     s1  = tmp[i];
1375     s1 *= (*v++);  /* multiply by inverse of diagonal entry */
1376     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1377     tmp[i] = s1;
1378   }
1379 
1380   /* backward solve the L^T */
1381   for (i=n-1; i>=0; i--) {
1382     v  = aa + diag[i] - 1;
1383     vi = aj + diag[i] - 1;
1384     nz = diag[i] - ai[i];
1385     s1 = tmp[i];
1386     for (j=0; j>-nz; j--) tmp[vi[j]] -= s1*v[j];
1387   }
1388 
1389   /* copy tmp into x according to permutation */
1390   for (i=0; i<n; i++) x[r[i]] = tmp[i];
1391 
1392   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1393   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1394   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1395   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
1396 
1397   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1398   PetscFunctionReturn(0);
1399 }
1400 
1401 PetscErrorCode MatSolveTranspose_SeqAIJ(Mat A,Vec bb,Vec xx)
1402 {
1403   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1404   IS                iscol = a->col,isrow = a->row;
1405   PetscErrorCode    ierr;
1406   const PetscInt    *rout,*cout,*r,*c,*adiag = a->diag,*ai = a->i,*aj = a->j,*vi;
1407   PetscInt          i,n = A->rmap->n,j;
1408   PetscInt          nz;
1409   PetscScalar       *x,*tmp,s1;
1410   const MatScalar   *aa = a->a,*v;
1411   const PetscScalar *b;
1412 
1413   PetscFunctionBegin;
1414   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1415   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
1416   tmp  = a->solve_work;
1417 
1418   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1419   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1420 
1421   /* copy the b into temp work space according to permutation */
1422   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1423 
1424   /* forward solve the U^T */
1425   for (i=0; i<n; i++) {
1426     v   = aa + adiag[i+1] + 1;
1427     vi  = aj + adiag[i+1] + 1;
1428     nz  = adiag[i] - adiag[i+1] - 1;
1429     s1  = tmp[i];
1430     s1 *= v[nz];  /* multiply by inverse of diagonal entry */
1431     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1432     tmp[i] = s1;
1433   }
1434 
1435   /* backward solve the L^T */
1436   for (i=n-1; i>=0; i--) {
1437     v  = aa + ai[i];
1438     vi = aj + ai[i];
1439     nz = ai[i+1] - ai[i];
1440     s1 = tmp[i];
1441     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1442   }
1443 
1444   /* copy tmp into x according to permutation */
1445   for (i=0; i<n; i++) x[r[i]] = tmp[i];
1446 
1447   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1448   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1449   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1450   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
1451 
1452   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1453   PetscFunctionReturn(0);
1454 }
1455 
1456 PetscErrorCode MatSolveTransposeAdd_SeqAIJ_inplace(Mat A,Vec bb,Vec zz,Vec xx)
1457 {
1458   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1459   IS                iscol = a->col,isrow = a->row;
1460   PetscErrorCode    ierr;
1461   const PetscInt    *rout,*cout,*r,*c,*diag = a->diag,*ai = a->i,*aj = a->j,*vi;
1462   PetscInt          i,n = A->rmap->n,j;
1463   PetscInt          nz;
1464   PetscScalar       *x,*tmp,s1;
1465   const MatScalar   *aa = a->a,*v;
1466   const PetscScalar *b;
1467 
1468   PetscFunctionBegin;
1469   if (zz != xx) {ierr = VecCopy(zz,xx);CHKERRQ(ierr);}
1470   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1471   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1472   tmp  = a->solve_work;
1473 
1474   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1475   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1476 
1477   /* copy the b into temp work space according to permutation */
1478   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1479 
1480   /* forward solve the U^T */
1481   for (i=0; i<n; i++) {
1482     v   = aa + diag[i];
1483     vi  = aj + diag[i] + 1;
1484     nz  = ai[i+1] - diag[i] - 1;
1485     s1  = tmp[i];
1486     s1 *= (*v++);  /* multiply by inverse of diagonal entry */
1487     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1488     tmp[i] = s1;
1489   }
1490 
1491   /* backward solve the L^T */
1492   for (i=n-1; i>=0; i--) {
1493     v  = aa + diag[i] - 1;
1494     vi = aj + diag[i] - 1;
1495     nz = diag[i] - ai[i];
1496     s1 = tmp[i];
1497     for (j=0; j>-nz; j--) tmp[vi[j]] -= s1*v[j];
1498   }
1499 
1500   /* copy tmp into x according to permutation */
1501   for (i=0; i<n; i++) x[r[i]] += tmp[i];
1502 
1503   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1504   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1505   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1506   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1507 
1508   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1509   PetscFunctionReturn(0);
1510 }
1511 
1512 PetscErrorCode MatSolveTransposeAdd_SeqAIJ(Mat A,Vec bb,Vec zz,Vec xx)
1513 {
1514   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
1515   IS                iscol = a->col,isrow = a->row;
1516   PetscErrorCode    ierr;
1517   const PetscInt    *rout,*cout,*r,*c,*adiag = a->diag,*ai = a->i,*aj = a->j,*vi;
1518   PetscInt          i,n = A->rmap->n,j;
1519   PetscInt          nz;
1520   PetscScalar       *x,*tmp,s1;
1521   const MatScalar   *aa = a->a,*v;
1522   const PetscScalar *b;
1523 
1524   PetscFunctionBegin;
1525   if (zz != xx) {ierr = VecCopy(zz,xx);CHKERRQ(ierr);}
1526   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
1527   ierr = VecGetArray(xx,&x);CHKERRQ(ierr);
1528   tmp  = a->solve_work;
1529 
1530   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
1531   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
1532 
1533   /* copy the b into temp work space according to permutation */
1534   for (i=0; i<n; i++) tmp[i] = b[c[i]];
1535 
1536   /* forward solve the U^T */
1537   for (i=0; i<n; i++) {
1538     v   = aa + adiag[i+1] + 1;
1539     vi  = aj + adiag[i+1] + 1;
1540     nz  = adiag[i] - adiag[i+1] - 1;
1541     s1  = tmp[i];
1542     s1 *= v[nz];  /* multiply by inverse of diagonal entry */
1543     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1544     tmp[i] = s1;
1545   }
1546 
1547 
1548   /* backward solve the L^T */
1549   for (i=n-1; i>=0; i--) {
1550     v  = aa + ai[i];
1551     vi = aj + ai[i];
1552     nz = ai[i+1] - ai[i];
1553     s1 = tmp[i];
1554     for (j=0; j<nz; j++) tmp[vi[j]] -= s1*v[j];
1555   }
1556 
1557   /* copy tmp into x according to permutation */
1558   for (i=0; i<n; i++) x[r[i]] += tmp[i];
1559 
1560   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
1561   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
1562   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
1563   ierr = VecRestoreArray(xx,&x);CHKERRQ(ierr);
1564 
1565   ierr = PetscLogFlops(2.0*a->nz-A->cmap->n);CHKERRQ(ierr);
1566   PetscFunctionReturn(0);
1567 }
1568 
1569 /* ----------------------------------------------------------------*/
1570 
1571 /*
1572    ilu() under revised new data structure.
1573    Factored arrays bj and ba are stored as
1574      L(0,:), L(1,:), ...,L(n-1,:),  U(n-1,:),...,U(i,:),U(i-1,:),...,U(0,:)
1575 
1576    bi=fact->i is an array of size n+1, in which
1577    bi+
1578      bi[i]:  points to 1st entry of L(i,:),i=0,...,n-1
1579      bi[n]:  points to L(n-1,n-1)+1
1580 
1581   bdiag=fact->diag is an array of size n+1,in which
1582      bdiag[i]: points to diagonal of U(i,:), i=0,...,n-1
1583      bdiag[n]: points to entry of U(n-1,0)-1
1584 
1585    U(i,:) contains bdiag[i] as its last entry, i.e.,
1586     U(i,:) = (u[i,i+1],...,u[i,n-1],diag[i])
1587 */
1588 PetscErrorCode MatILUFactorSymbolic_SeqAIJ_ilu0(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1589 {
1590   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data,*b;
1591   PetscErrorCode ierr;
1592   const PetscInt n=A->rmap->n,*ai=a->i,*aj,*adiag=a->diag;
1593   PetscInt       i,j,k=0,nz,*bi,*bj,*bdiag;
1594   IS             isicol;
1595 
1596   PetscFunctionBegin;
1597   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1598   ierr = MatDuplicateNoCreate_SeqAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_FALSE);CHKERRQ(ierr);
1599   b    = (Mat_SeqAIJ*)(fact)->data;
1600 
1601   /* allocate matrix arrays for new data structure */
1602   ierr = PetscMalloc3(ai[n]+1,&b->a,ai[n]+1,&b->j,n+1,&b->i);CHKERRQ(ierr);
1603   ierr = PetscLogObjectMemory((PetscObject)fact,ai[n]*(sizeof(PetscScalar)+sizeof(PetscInt))+(n+1)*sizeof(PetscInt));CHKERRQ(ierr);
1604 
1605   b->singlemalloc = PETSC_TRUE;
1606   if (!b->diag) {
1607     ierr = PetscMalloc1(n+1,&b->diag);CHKERRQ(ierr);
1608     ierr = PetscLogObjectMemory((PetscObject)fact,(n+1)*sizeof(PetscInt));CHKERRQ(ierr);
1609   }
1610   bdiag = b->diag;
1611 
1612   if (n > 0) {
1613     ierr = PetscArrayzero(b->a,ai[n]);CHKERRQ(ierr);
1614   }
1615 
1616   /* set bi and bj with new data structure */
1617   bi = b->i;
1618   bj = b->j;
1619 
1620   /* L part */
1621   bi[0] = 0;
1622   for (i=0; i<n; i++) {
1623     nz      = adiag[i] - ai[i];
1624     bi[i+1] = bi[i] + nz;
1625     aj      = a->j + ai[i];
1626     for (j=0; j<nz; j++) {
1627       /*   *bj = aj[j]; bj++; */
1628       bj[k++] = aj[j];
1629     }
1630   }
1631 
1632   /* U part */
1633   bdiag[n] = bi[n]-1;
1634   for (i=n-1; i>=0; i--) {
1635     nz = ai[i+1] - adiag[i] - 1;
1636     aj = a->j + adiag[i] + 1;
1637     for (j=0; j<nz; j++) {
1638       /*      *bj = aj[j]; bj++; */
1639       bj[k++] = aj[j];
1640     }
1641     /* diag[i] */
1642     /*    *bj = i; bj++; */
1643     bj[k++]  = i;
1644     bdiag[i] = bdiag[i+1] + nz + 1;
1645   }
1646 
1647   fact->factortype             = MAT_FACTOR_ILU;
1648   fact->info.factor_mallocs    = 0;
1649   fact->info.fill_ratio_given  = info->fill;
1650   fact->info.fill_ratio_needed = 1.0;
1651   fact->ops->lufactornumeric   = MatLUFactorNumeric_SeqAIJ;
1652   ierr = MatSeqAIJCheckInode_FactorLU(fact);CHKERRQ(ierr);
1653 
1654   b       = (Mat_SeqAIJ*)(fact)->data;
1655   b->row  = isrow;
1656   b->col  = iscol;
1657   b->icol = isicol;
1658   ierr    = PetscMalloc1(fact->rmap->n+1,&b->solve_work);CHKERRQ(ierr);
1659   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1660   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1661   PetscFunctionReturn(0);
1662 }
1663 
1664 PetscErrorCode MatILUFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1665 {
1666   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
1667   IS                 isicol;
1668   PetscErrorCode     ierr;
1669   const PetscInt     *r,*ic;
1670   PetscInt           n=A->rmap->n,*ai=a->i,*aj=a->j;
1671   PetscInt           *bi,*cols,nnz,*cols_lvl;
1672   PetscInt           *bdiag,prow,fm,nzbd,reallocs=0,dcount=0;
1673   PetscInt           i,levels,diagonal_fill;
1674   PetscBool          col_identity,row_identity,missing;
1675   PetscReal          f;
1676   PetscInt           nlnk,*lnk,*lnk_lvl=NULL;
1677   PetscBT            lnkbt;
1678   PetscInt           nzi,*bj,**bj_ptr,**bjlvl_ptr;
1679   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
1680   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
1681 
1682   PetscFunctionBegin;
1683   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);
1684   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
1685   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
1686 
1687   levels = (PetscInt)info->levels;
1688   ierr   = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
1689   ierr   = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);
1690   if (!levels && row_identity && col_identity) {
1691     /* special case: ilu(0) with natural ordering */
1692     ierr = MatILUFactorSymbolic_SeqAIJ_ilu0(fact,A,isrow,iscol,info);CHKERRQ(ierr);
1693     if (a->inode.size) {
1694       fact->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
1695     }
1696     PetscFunctionReturn(0);
1697   }
1698 
1699   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1700   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
1701   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
1702 
1703   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
1704   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
1705   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
1706   bi[0] = bdiag[0] = 0;
1707   ierr  = PetscMalloc2(n,&bj_ptr,n,&bjlvl_ptr);CHKERRQ(ierr);
1708 
1709   /* create a linked list for storing column indices of the active row */
1710   nlnk = n + 1;
1711   ierr = PetscIncompleteLLCreate(n,n,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1712 
1713   /* initial FreeSpace size is f*(ai[n]+1) */
1714   f                 = info->fill;
1715   diagonal_fill     = (PetscInt)info->diagonal_fill;
1716   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
1717   current_space     = free_space;
1718   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space_lvl);CHKERRQ(ierr);
1719   current_space_lvl = free_space_lvl;
1720   for (i=0; i<n; i++) {
1721     nzi = 0;
1722     /* copy current row into linked list */
1723     nnz = ai[r[i]+1] - ai[r[i]];
1724     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);
1725     cols   = aj + ai[r[i]];
1726     lnk[i] = -1; /* marker to indicate if diagonal exists */
1727     ierr   = PetscIncompleteLLInit(nnz,cols,n,ic,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1728     nzi   += nlnk;
1729 
1730     /* make sure diagonal entry is included */
1731     if (diagonal_fill && lnk[i] == -1) {
1732       fm = n;
1733       while (lnk[fm] < i) fm = lnk[fm];
1734       lnk[i]     = lnk[fm]; /* insert diagonal into linked list */
1735       lnk[fm]    = i;
1736       lnk_lvl[i] = 0;
1737       nzi++; dcount++;
1738     }
1739 
1740     /* add pivot rows into the active row */
1741     nzbd = 0;
1742     prow = lnk[n];
1743     while (prow < i) {
1744       nnz      = bdiag[prow];
1745       cols     = bj_ptr[prow] + nnz + 1;
1746       cols_lvl = bjlvl_ptr[prow] + nnz + 1;
1747       nnz      = bi[prow+1] - bi[prow] - nnz - 1;
1748       ierr     = PetscILULLAddSorted(nnz,cols,levels,cols_lvl,prow,nlnk,lnk,lnk_lvl,lnkbt,prow);CHKERRQ(ierr);
1749       nzi     += nlnk;
1750       prow     = lnk[prow];
1751       nzbd++;
1752     }
1753     bdiag[i] = nzbd;
1754     bi[i+1]  = bi[i] + nzi;
1755     /* if free space is not available, make more free space */
1756     if (current_space->local_remaining<nzi) {
1757       nnz  = PetscIntMultTruncate(2,PetscIntMultTruncate(nzi,n - i)); /* estimated and max additional space needed */
1758       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
1759       ierr = PetscFreeSpaceGet(nnz,&current_space_lvl);CHKERRQ(ierr);
1760       reallocs++;
1761     }
1762 
1763     /* copy data into free_space and free_space_lvl, then initialize lnk */
1764     ierr         = PetscIncompleteLLClean(n,n,nzi,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
1765     bj_ptr[i]    = current_space->array;
1766     bjlvl_ptr[i] = current_space_lvl->array;
1767 
1768     /* make sure the active row i has diagonal entry */
1769     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);
1770 
1771     current_space->array               += nzi;
1772     current_space->local_used          += nzi;
1773     current_space->local_remaining     -= nzi;
1774     current_space_lvl->array           += nzi;
1775     current_space_lvl->local_used      += nzi;
1776     current_space_lvl->local_remaining -= nzi;
1777   }
1778 
1779   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
1780   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
1781   /* copy free_space into bj and free free_space; set bi, bj, bdiag in new datastructure; */
1782   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
1783   ierr = PetscFreeSpaceContiguous_LU(&free_space,bj,n,bi,bdiag);CHKERRQ(ierr);
1784 
1785   ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
1786   ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
1787   ierr = PetscFree2(bj_ptr,bjlvl_ptr);CHKERRQ(ierr);
1788 
1789 #if defined(PETSC_USE_INFO)
1790   {
1791     PetscReal af = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
1792     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
1793     ierr = PetscInfo1(A,"Run with -[sub_]pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
1794     ierr = PetscInfo1(A,"PCFactorSetFill([sub]pc,%g);\n",(double)af);CHKERRQ(ierr);
1795     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
1796     if (diagonal_fill) {
1797       ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals\n",dcount);CHKERRQ(ierr);
1798     }
1799   }
1800 #endif
1801   /* put together the new matrix */
1802   ierr = MatSeqAIJSetPreallocation_SeqAIJ(fact,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
1803   ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);
1804   b    = (Mat_SeqAIJ*)(fact)->data;
1805 
1806   b->free_a       = PETSC_TRUE;
1807   b->free_ij      = PETSC_TRUE;
1808   b->singlemalloc = PETSC_FALSE;
1809 
1810   ierr = PetscMalloc1(bdiag[0]+1,&b->a);CHKERRQ(ierr);
1811 
1812   b->j    = bj;
1813   b->i    = bi;
1814   b->diag = bdiag;
1815   b->ilen = 0;
1816   b->imax = 0;
1817   b->row  = isrow;
1818   b->col  = iscol;
1819   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1820   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1821   b->icol = isicol;
1822 
1823   ierr = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
1824   /* In b structure:  Free imax, ilen, old a, old j.
1825      Allocate bdiag, solve_work, new a, new j */
1826   ierr     = PetscLogObjectMemory((PetscObject)fact,(bdiag[0]+1)*(sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
1827   b->maxnz = b->nz = bdiag[0]+1;
1828 
1829   (fact)->info.factor_mallocs    = reallocs;
1830   (fact)->info.fill_ratio_given  = f;
1831   (fact)->info.fill_ratio_needed = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
1832   (fact)->ops->lufactornumeric   = MatLUFactorNumeric_SeqAIJ;
1833   if (a->inode.size) {
1834     (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode;
1835   }
1836   ierr = MatSeqAIJCheckInode_FactorLU(fact);CHKERRQ(ierr);
1837   PetscFunctionReturn(0);
1838 }
1839 
1840 PetscErrorCode MatILUFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
1841 {
1842   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data,*b;
1843   IS                 isicol;
1844   PetscErrorCode     ierr;
1845   const PetscInt     *r,*ic;
1846   PetscInt           n=A->rmap->n,*ai=a->i,*aj=a->j;
1847   PetscInt           *bi,*cols,nnz,*cols_lvl;
1848   PetscInt           *bdiag,prow,fm,nzbd,reallocs=0,dcount=0;
1849   PetscInt           i,levels,diagonal_fill;
1850   PetscBool          col_identity,row_identity;
1851   PetscReal          f;
1852   PetscInt           nlnk,*lnk,*lnk_lvl=NULL;
1853   PetscBT            lnkbt;
1854   PetscInt           nzi,*bj,**bj_ptr,**bjlvl_ptr;
1855   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
1856   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
1857   PetscBool          missing;
1858 
1859   PetscFunctionBegin;
1860   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);
1861   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
1862   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
1863 
1864   f             = info->fill;
1865   levels        = (PetscInt)info->levels;
1866   diagonal_fill = (PetscInt)info->diagonal_fill;
1867 
1868   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
1869 
1870   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
1871   ierr = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);
1872   if (!levels && row_identity && col_identity) { /* special case: ilu(0) with natural ordering */
1873     ierr = MatDuplicateNoCreate_SeqAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_TRUE);CHKERRQ(ierr);
1874 
1875     (fact)->ops->lufactornumeric =  MatLUFactorNumeric_SeqAIJ_inplace;
1876     if (a->inode.size) {
1877       (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
1878     }
1879     fact->factortype               = MAT_FACTOR_ILU;
1880     (fact)->info.factor_mallocs    = 0;
1881     (fact)->info.fill_ratio_given  = info->fill;
1882     (fact)->info.fill_ratio_needed = 1.0;
1883 
1884     b    = (Mat_SeqAIJ*)(fact)->data;
1885     b->row  = isrow;
1886     b->col  = iscol;
1887     b->icol = isicol;
1888     ierr    = PetscMalloc1((fact)->rmap->n+1,&b->solve_work);CHKERRQ(ierr);
1889     ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
1890     ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
1891     PetscFunctionReturn(0);
1892   }
1893 
1894   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
1895   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
1896 
1897   /* get new row and diagonal pointers, must be allocated separately because they will be given to the Mat_SeqAIJ and freed separately */
1898   ierr  = PetscMalloc1(n+1,&bi);CHKERRQ(ierr);
1899   ierr  = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);
1900   bi[0] = bdiag[0] = 0;
1901 
1902   ierr = PetscMalloc2(n,&bj_ptr,n,&bjlvl_ptr);CHKERRQ(ierr);
1903 
1904   /* create a linked list for storing column indices of the active row */
1905   nlnk = n + 1;
1906   ierr = PetscIncompleteLLCreate(n,n,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1907 
1908   /* initial FreeSpace size is f*(ai[n]+1) */
1909   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space);CHKERRQ(ierr);
1910   current_space     = free_space;
1911   ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(f,ai[n]+1),&free_space_lvl);CHKERRQ(ierr);
1912   current_space_lvl = free_space_lvl;
1913 
1914   for (i=0; i<n; i++) {
1915     nzi = 0;
1916     /* copy current row into linked list */
1917     nnz = ai[r[i]+1] - ai[r[i]];
1918     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);
1919     cols   = aj + ai[r[i]];
1920     lnk[i] = -1; /* marker to indicate if diagonal exists */
1921     ierr   = PetscIncompleteLLInit(nnz,cols,n,ic,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
1922     nzi   += nlnk;
1923 
1924     /* make sure diagonal entry is included */
1925     if (diagonal_fill && lnk[i] == -1) {
1926       fm = n;
1927       while (lnk[fm] < i) fm = lnk[fm];
1928       lnk[i]     = lnk[fm]; /* insert diagonal into linked list */
1929       lnk[fm]    = i;
1930       lnk_lvl[i] = 0;
1931       nzi++; dcount++;
1932     }
1933 
1934     /* add pivot rows into the active row */
1935     nzbd = 0;
1936     prow = lnk[n];
1937     while (prow < i) {
1938       nnz      = bdiag[prow];
1939       cols     = bj_ptr[prow] + nnz + 1;
1940       cols_lvl = bjlvl_ptr[prow] + nnz + 1;
1941       nnz      = bi[prow+1] - bi[prow] - nnz - 1;
1942       ierr     = PetscILULLAddSorted(nnz,cols,levels,cols_lvl,prow,nlnk,lnk,lnk_lvl,lnkbt,prow);CHKERRQ(ierr);
1943       nzi     += nlnk;
1944       prow     = lnk[prow];
1945       nzbd++;
1946     }
1947     bdiag[i] = nzbd;
1948     bi[i+1]  = bi[i] + nzi;
1949 
1950     /* if free space is not available, make more free space */
1951     if (current_space->local_remaining<nzi) {
1952       nnz  = PetscIntMultTruncate(nzi,n - i); /* estimated and max additional space needed */
1953       ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
1954       ierr = PetscFreeSpaceGet(nnz,&current_space_lvl);CHKERRQ(ierr);
1955       reallocs++;
1956     }
1957 
1958     /* copy data into free_space and free_space_lvl, then initialize lnk */
1959     ierr         = PetscIncompleteLLClean(n,n,nzi,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
1960     bj_ptr[i]    = current_space->array;
1961     bjlvl_ptr[i] = current_space_lvl->array;
1962 
1963     /* make sure the active row i has diagonal entry */
1964     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);
1965 
1966     current_space->array               += nzi;
1967     current_space->local_used          += nzi;
1968     current_space->local_remaining     -= nzi;
1969     current_space_lvl->array           += nzi;
1970     current_space_lvl->local_used      += nzi;
1971     current_space_lvl->local_remaining -= nzi;
1972   }
1973 
1974   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
1975   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
1976 
1977   /* destroy list of free space and other temporary arrays */
1978   ierr = PetscMalloc1(bi[n]+1,&bj);CHKERRQ(ierr);
1979   ierr = PetscFreeSpaceContiguous(&free_space,bj);CHKERRQ(ierr); /* copy free_space -> bj */
1980   ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
1981   ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
1982   ierr = PetscFree2(bj_ptr,bjlvl_ptr);CHKERRQ(ierr);
1983 
1984 #if defined(PETSC_USE_INFO)
1985   {
1986     PetscReal af = ((PetscReal)bi[n])/((PetscReal)ai[n]);
1987     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)f,(double)af);CHKERRQ(ierr);
1988     ierr = PetscInfo1(A,"Run with -[sub_]pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
1989     ierr = PetscInfo1(A,"PCFactorSetFill([sub]pc,%g);\n",(double)af);CHKERRQ(ierr);
1990     ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
1991     if (diagonal_fill) {
1992       ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals\n",dcount);CHKERRQ(ierr);
1993     }
1994   }
1995 #endif
1996 
1997   /* put together the new matrix */
1998   ierr = MatSeqAIJSetPreallocation_SeqAIJ(fact,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
1999   ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);
2000   b    = (Mat_SeqAIJ*)(fact)->data;
2001 
2002   b->free_a       = PETSC_TRUE;
2003   b->free_ij      = PETSC_TRUE;
2004   b->singlemalloc = PETSC_FALSE;
2005 
2006   ierr = PetscMalloc1(bi[n],&b->a);CHKERRQ(ierr);
2007   b->j = bj;
2008   b->i = bi;
2009   for (i=0; i<n; i++) bdiag[i] += bi[i];
2010   b->diag = bdiag;
2011   b->ilen = 0;
2012   b->imax = 0;
2013   b->row  = isrow;
2014   b->col  = iscol;
2015   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
2016   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
2017   b->icol = isicol;
2018   ierr    = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
2019   /* In b structure:  Free imax, ilen, old a, old j.
2020      Allocate bdiag, solve_work, new a, new j */
2021   ierr     = PetscLogObjectMemory((PetscObject)fact,(bi[n]-n) * (sizeof(PetscInt)+sizeof(PetscScalar)));CHKERRQ(ierr);
2022   b->maxnz = b->nz = bi[n];
2023 
2024   (fact)->info.factor_mallocs    = reallocs;
2025   (fact)->info.fill_ratio_given  = f;
2026   (fact)->info.fill_ratio_needed = ((PetscReal)bi[n])/((PetscReal)ai[n]);
2027   (fact)->ops->lufactornumeric   =  MatLUFactorNumeric_SeqAIJ_inplace;
2028   if (a->inode.size) {
2029     (fact)->ops->lufactornumeric = MatLUFactorNumeric_SeqAIJ_Inode_inplace;
2030   }
2031   PetscFunctionReturn(0);
2032 }
2033 
2034 PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ(Mat B,Mat A,const MatFactorInfo *info)
2035 {
2036   Mat            C = B;
2037   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data;
2038   Mat_SeqSBAIJ   *b=(Mat_SeqSBAIJ*)C->data;
2039   IS             ip=b->row,iip = b->icol;
2040   PetscErrorCode ierr;
2041   const PetscInt *rip,*riip;
2042   PetscInt       i,j,mbs=A->rmap->n,*bi=b->i,*bj=b->j,*bdiag=b->diag,*bjtmp;
2043   PetscInt       *ai=a->i,*aj=a->j;
2044   PetscInt       k,jmin,jmax,*c2r,*il,col,nexti,ili,nz;
2045   MatScalar      *rtmp,*ba=b->a,*bval,*aa=a->a,dk,uikdi;
2046   PetscBool      perm_identity;
2047   FactorShiftCtx sctx;
2048   PetscReal      rs;
2049   MatScalar      d,*v;
2050 
2051   PetscFunctionBegin;
2052   /* MatPivotSetUp(): initialize shift context sctx */
2053   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
2054 
2055   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
2056     sctx.shift_top = info->zeropivot;
2057     for (i=0; i<mbs; i++) {
2058       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
2059       d  = (aa)[a->diag[i]];
2060       rs = -PetscAbsScalar(d) - PetscRealPart(d);
2061       v  = aa+ai[i];
2062       nz = ai[i+1] - ai[i];
2063       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
2064       if (rs>sctx.shift_top) sctx.shift_top = rs;
2065     }
2066     sctx.shift_top *= 1.1;
2067     sctx.nshift_max = 5;
2068     sctx.shift_lo   = 0.;
2069     sctx.shift_hi   = 1.;
2070   }
2071 
2072   ierr = ISGetIndices(ip,&rip);CHKERRQ(ierr);
2073   ierr = ISGetIndices(iip,&riip);CHKERRQ(ierr);
2074 
2075   /* allocate working arrays
2076      c2r: linked list, keep track of pivot rows for a given column. c2r[col]: head of the list for a given col
2077      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
2078   */
2079   ierr = PetscMalloc3(mbs,&rtmp,mbs,&il,mbs,&c2r);CHKERRQ(ierr);
2080 
2081   do {
2082     sctx.newshift = PETSC_FALSE;
2083 
2084     for (i=0; i<mbs; i++) c2r[i] = mbs;
2085     if (mbs) il[0] = 0;
2086 
2087     for (k = 0; k<mbs; k++) {
2088       /* zero rtmp */
2089       nz    = bi[k+1] - bi[k];
2090       bjtmp = bj + bi[k];
2091       for (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
2092 
2093       /* load in initial unfactored row */
2094       bval = ba + bi[k];
2095       jmin = ai[rip[k]]; jmax = ai[rip[k]+1];
2096       for (j = jmin; j < jmax; j++) {
2097         col = riip[aj[j]];
2098         if (col >= k) { /* only take upper triangular entry */
2099           rtmp[col] = aa[j];
2100           *bval++   = 0.0; /* for in-place factorization */
2101         }
2102       }
2103       /* shift the diagonal of the matrix: ZeropivotApply() */
2104       rtmp[k] += sctx.shift_amount;  /* shift the diagonal of the matrix */
2105 
2106       /* modify k-th row by adding in those rows i with U(i,k)!=0 */
2107       dk = rtmp[k];
2108       i  = c2r[k]; /* first row to be added to k_th row  */
2109 
2110       while (i < k) {
2111         nexti = c2r[i]; /* next row to be added to k_th row */
2112 
2113         /* compute multiplier, update diag(k) and U(i,k) */
2114         ili     = il[i]; /* index of first nonzero element in U(i,k:bms-1) */
2115         uikdi   = -ba[ili]*ba[bdiag[i]]; /* diagonal(k) */
2116         dk     += uikdi*ba[ili]; /* update diag[k] */
2117         ba[ili] = uikdi; /* -U(i,k) */
2118 
2119         /* add multiple of row i to k-th row */
2120         jmin = ili + 1; jmax = bi[i+1];
2121         if (jmin < jmax) {
2122           for (j=jmin; j<jmax; j++) rtmp[bj[j]] += uikdi*ba[j];
2123           /* update il and c2r for row i */
2124           il[i] = jmin;
2125           j     = bj[jmin]; c2r[i] = c2r[j]; c2r[j] = i;
2126         }
2127         i = nexti;
2128       }
2129 
2130       /* copy data into U(k,:) */
2131       rs   = 0.0;
2132       jmin = bi[k]; jmax = bi[k+1]-1;
2133       if (jmin < jmax) {
2134         for (j=jmin; j<jmax; j++) {
2135           col = bj[j]; ba[j] = rtmp[col]; rs += PetscAbsScalar(ba[j]);
2136         }
2137         /* add the k-th row into il and c2r */
2138         il[k] = jmin;
2139         i     = bj[jmin]; c2r[k] = c2r[i]; c2r[i] = k;
2140       }
2141 
2142       /* MatPivotCheck() */
2143       sctx.rs = rs;
2144       sctx.pv = dk;
2145       ierr    = MatPivotCheck(B,A,info,&sctx,i);CHKERRQ(ierr);
2146       if (sctx.newshift) break;
2147       dk = sctx.pv;
2148 
2149       ba[bdiag[k]] = 1.0/dk; /* U(k,k) */
2150     }
2151   } while (sctx.newshift);
2152 
2153   ierr = PetscFree3(rtmp,il,c2r);CHKERRQ(ierr);
2154   ierr = ISRestoreIndices(ip,&rip);CHKERRQ(ierr);
2155   ierr = ISRestoreIndices(iip,&riip);CHKERRQ(ierr);
2156 
2157   ierr = ISIdentity(ip,&perm_identity);CHKERRQ(ierr);
2158   if (perm_identity) {
2159     B->ops->solve          = MatSolve_SeqSBAIJ_1_NaturalOrdering;
2160     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_NaturalOrdering;
2161     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_NaturalOrdering;
2162     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_NaturalOrdering;
2163   } else {
2164     B->ops->solve          = MatSolve_SeqSBAIJ_1;
2165     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1;
2166     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1;
2167     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1;
2168   }
2169 
2170   C->assembled    = PETSC_TRUE;
2171   C->preallocated = PETSC_TRUE;
2172 
2173   ierr = PetscLogFlops(C->rmap->n);CHKERRQ(ierr);
2174 
2175   /* MatPivotView() */
2176   if (sctx.nshift) {
2177     if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
2178       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);
2179     } else if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
2180       ierr = PetscInfo2(A,"number of shift_nz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2181     } else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) {
2182       ierr = PetscInfo2(A,"number of shift_inblocks applied %D, each shift_amount %g\n",sctx.nshift,(double)info->shiftamount);CHKERRQ(ierr);
2183     }
2184   }
2185   PetscFunctionReturn(0);
2186 }
2187 
2188 PetscErrorCode MatCholeskyFactorNumeric_SeqAIJ_inplace(Mat B,Mat A,const MatFactorInfo *info)
2189 {
2190   Mat            C = B;
2191   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data;
2192   Mat_SeqSBAIJ   *b=(Mat_SeqSBAIJ*)C->data;
2193   IS             ip=b->row,iip = b->icol;
2194   PetscErrorCode ierr;
2195   const PetscInt *rip,*riip;
2196   PetscInt       i,j,mbs=A->rmap->n,*bi=b->i,*bj=b->j,*bcol,*bjtmp;
2197   PetscInt       *ai=a->i,*aj=a->j;
2198   PetscInt       k,jmin,jmax,*jl,*il,col,nexti,ili,nz;
2199   MatScalar      *rtmp,*ba=b->a,*bval,*aa=a->a,dk,uikdi;
2200   PetscBool      perm_identity;
2201   FactorShiftCtx sctx;
2202   PetscReal      rs;
2203   MatScalar      d,*v;
2204 
2205   PetscFunctionBegin;
2206   /* MatPivotSetUp(): initialize shift context sctx */
2207   ierr = PetscMemzero(&sctx,sizeof(FactorShiftCtx));CHKERRQ(ierr);
2208 
2209   if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) { /* set sctx.shift_top=max{rs} */
2210     sctx.shift_top = info->zeropivot;
2211     for (i=0; i<mbs; i++) {
2212       /* calculate sum(|aij|)-RealPart(aii), amt of shift needed for this row */
2213       d  = (aa)[a->diag[i]];
2214       rs = -PetscAbsScalar(d) - PetscRealPart(d);
2215       v  = aa+ai[i];
2216       nz = ai[i+1] - ai[i];
2217       for (j=0; j<nz; j++) rs += PetscAbsScalar(v[j]);
2218       if (rs>sctx.shift_top) sctx.shift_top = rs;
2219     }
2220     sctx.shift_top *= 1.1;
2221     sctx.nshift_max = 5;
2222     sctx.shift_lo   = 0.;
2223     sctx.shift_hi   = 1.;
2224   }
2225 
2226   ierr = ISGetIndices(ip,&rip);CHKERRQ(ierr);
2227   ierr = ISGetIndices(iip,&riip);CHKERRQ(ierr);
2228 
2229   /* initialization */
2230   ierr = PetscMalloc3(mbs,&rtmp,mbs,&il,mbs,&jl);CHKERRQ(ierr);
2231 
2232   do {
2233     sctx.newshift = PETSC_FALSE;
2234 
2235     for (i=0; i<mbs; i++) jl[i] = mbs;
2236     il[0] = 0;
2237 
2238     for (k = 0; k<mbs; k++) {
2239       /* zero rtmp */
2240       nz    = bi[k+1] - bi[k];
2241       bjtmp = bj + bi[k];
2242       for (j=0; j<nz; j++) rtmp[bjtmp[j]] = 0.0;
2243 
2244       bval = ba + bi[k];
2245       /* initialize k-th row by the perm[k]-th row of A */
2246       jmin = ai[rip[k]]; jmax = ai[rip[k]+1];
2247       for (j = jmin; j < jmax; j++) {
2248         col = riip[aj[j]];
2249         if (col >= k) { /* only take upper triangular entry */
2250           rtmp[col] = aa[j];
2251           *bval++   = 0.0; /* for in-place factorization */
2252         }
2253       }
2254       /* shift the diagonal of the matrix */
2255       if (sctx.nshift) rtmp[k] += sctx.shift_amount;
2256 
2257       /* modify k-th row by adding in those rows i with U(i,k)!=0 */
2258       dk = rtmp[k];
2259       i  = jl[k]; /* first row to be added to k_th row  */
2260 
2261       while (i < k) {
2262         nexti = jl[i]; /* next row to be added to k_th row */
2263 
2264         /* compute multiplier, update diag(k) and U(i,k) */
2265         ili     = il[i]; /* index of first nonzero element in U(i,k:bms-1) */
2266         uikdi   = -ba[ili]*ba[bi[i]]; /* diagonal(k) */
2267         dk     += uikdi*ba[ili];
2268         ba[ili] = uikdi; /* -U(i,k) */
2269 
2270         /* add multiple of row i to k-th row */
2271         jmin = ili + 1; jmax = bi[i+1];
2272         if (jmin < jmax) {
2273           for (j=jmin; j<jmax; j++) rtmp[bj[j]] += uikdi*ba[j];
2274           /* update il and jl for row i */
2275           il[i] = jmin;
2276           j     = bj[jmin]; jl[i] = jl[j]; jl[j] = i;
2277         }
2278         i = nexti;
2279       }
2280 
2281       /* shift the diagonals when zero pivot is detected */
2282       /* compute rs=sum of abs(off-diagonal) */
2283       rs   = 0.0;
2284       jmin = bi[k]+1;
2285       nz   = bi[k+1] - jmin;
2286       bcol = bj + jmin;
2287       for (j=0; j<nz; j++) {
2288         rs += PetscAbsScalar(rtmp[bcol[j]]);
2289       }
2290 
2291       sctx.rs = rs;
2292       sctx.pv = dk;
2293       ierr    = MatPivotCheck(B,A,info,&sctx,k);CHKERRQ(ierr);
2294       if (sctx.newshift) break;
2295       dk = sctx.pv;
2296 
2297       /* copy data into U(k,:) */
2298       ba[bi[k]] = 1.0/dk; /* U(k,k) */
2299       jmin      = bi[k]+1; jmax = bi[k+1];
2300       if (jmin < jmax) {
2301         for (j=jmin; j<jmax; j++) {
2302           col = bj[j]; ba[j] = rtmp[col];
2303         }
2304         /* add the k-th row into il and jl */
2305         il[k] = jmin;
2306         i     = bj[jmin]; jl[k] = jl[i]; jl[i] = k;
2307       }
2308     }
2309   } while (sctx.newshift);
2310 
2311   ierr = PetscFree3(rtmp,il,jl);CHKERRQ(ierr);
2312   ierr = ISRestoreIndices(ip,&rip);CHKERRQ(ierr);
2313   ierr = ISRestoreIndices(iip,&riip);CHKERRQ(ierr);
2314 
2315   ierr = ISIdentity(ip,&perm_identity);CHKERRQ(ierr);
2316   if (perm_identity) {
2317     B->ops->solve          = MatSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2318     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2319     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2320     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_NaturalOrdering_inplace;
2321   } else {
2322     B->ops->solve          = MatSolve_SeqSBAIJ_1_inplace;
2323     B->ops->solvetranspose = MatSolve_SeqSBAIJ_1_inplace;
2324     B->ops->forwardsolve   = MatForwardSolve_SeqSBAIJ_1_inplace;
2325     B->ops->backwardsolve  = MatBackwardSolve_SeqSBAIJ_1_inplace;
2326   }
2327 
2328   C->assembled    = PETSC_TRUE;
2329   C->preallocated = PETSC_TRUE;
2330 
2331   ierr = PetscLogFlops(C->rmap->n);CHKERRQ(ierr);
2332   if (sctx.nshift) {
2333     if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) {
2334       ierr = PetscInfo2(A,"number of shiftnz tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2335     } else if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) {
2336       ierr = PetscInfo2(A,"number of shiftpd tries %D, shift_amount %g\n",sctx.nshift,(double)sctx.shift_amount);CHKERRQ(ierr);
2337     }
2338   }
2339   PetscFunctionReturn(0);
2340 }
2341 
2342 /*
2343    icc() under revised new data structure.
2344    Factored arrays bj and ba are stored as
2345      U(0,:),...,U(i,:),U(n-1,:)
2346 
2347    ui=fact->i is an array of size n+1, in which
2348    ui+
2349      ui[i]:  points to 1st entry of U(i,:),i=0,...,n-1
2350      ui[n]:  points to U(n-1,n-1)+1
2351 
2352   udiag=fact->diag is an array of size n,in which
2353      udiag[i]: points to diagonal of U(i,:), i=0,...,n-1
2354 
2355    U(i,:) contains udiag[i] as its last entry, i.e.,
2356     U(i,:) = (u[i,i+1],...,u[i,n-1],diag[i])
2357 */
2358 
2359 PetscErrorCode MatICCFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2360 {
2361   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2362   Mat_SeqSBAIJ       *b;
2363   PetscErrorCode     ierr;
2364   PetscBool          perm_identity,missing;
2365   PetscInt           reallocs=0,i,*ai=a->i,*aj=a->j,am=A->rmap->n,*ui,*udiag;
2366   const PetscInt     *rip,*riip;
2367   PetscInt           jmin,jmax,nzk,k,j,*jl,prow,*il,nextprow;
2368   PetscInt           nlnk,*lnk,*lnk_lvl=NULL,d;
2369   PetscInt           ncols,ncols_upper,*cols,*ajtmp,*uj,**uj_ptr,**uj_lvl_ptr;
2370   PetscReal          fill          =info->fill,levels=info->levels;
2371   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
2372   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
2373   PetscBT            lnkbt;
2374   IS                 iperm;
2375 
2376   PetscFunctionBegin;
2377   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);
2378   ierr = MatMissingDiagonal(A,&missing,&d);CHKERRQ(ierr);
2379   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",d);
2380   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2381   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2382 
2383   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2384   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2385   ui[0] = 0;
2386 
2387   /* ICC(0) without matrix ordering: simply rearrange column indices */
2388   if (!levels && perm_identity) {
2389     for (i=0; i<am; i++) {
2390       ncols    = ai[i+1] - a->diag[i];
2391       ui[i+1]  = ui[i] + ncols;
2392       udiag[i] = ui[i+1] - 1; /* points to the last entry of U(i,:) */
2393     }
2394     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2395     cols = uj;
2396     for (i=0; i<am; i++) {
2397       aj    = a->j + a->diag[i] + 1; /* 1st entry of U(i,:) without diagonal */
2398       ncols = ai[i+1] - a->diag[i] -1;
2399       for (j=0; j<ncols; j++) *cols++ = aj[j];
2400       *cols++ = i; /* diagonal is located as the last entry of U(i,:) */
2401     }
2402   } else { /* case: levels>0 || (levels=0 && !perm_identity) */
2403     ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2404     ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2405 
2406     /* initialization */
2407     ierr = PetscMalloc1(am+1,&ajtmp);CHKERRQ(ierr);
2408 
2409     /* jl: linked list for storing indices of the pivot rows
2410        il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2411     ierr = PetscMalloc4(am,&uj_ptr,am,&uj_lvl_ptr,am,&jl,am,&il);CHKERRQ(ierr);
2412     for (i=0; i<am; i++) {
2413       jl[i] = am; il[i] = 0;
2414     }
2415 
2416     /* create and initialize a linked list for storing column indices of the active row k */
2417     nlnk = am + 1;
2418     ierr = PetscIncompleteLLCreate(am,am,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2419 
2420     /* initial FreeSpace size is fill*(ai[am]+am)/2 */
2421     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space);CHKERRQ(ierr);
2422     current_space     = free_space;
2423     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space_lvl);CHKERRQ(ierr);
2424     current_space_lvl = free_space_lvl;
2425 
2426     for (k=0; k<am; k++) {  /* for each active row k */
2427       /* initialize lnk by the column indices of row rip[k] of A */
2428       nzk   = 0;
2429       ncols = ai[rip[k]+1] - ai[rip[k]];
2430       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);
2431       ncols_upper = 0;
2432       for (j=0; j<ncols; j++) {
2433         i = *(aj + ai[rip[k]] + j); /* unpermuted column index */
2434         if (riip[i] >= k) { /* only take upper triangular entry */
2435           ajtmp[ncols_upper] = i;
2436           ncols_upper++;
2437         }
2438       }
2439       ierr = PetscIncompleteLLInit(ncols_upper,ajtmp,am,riip,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2440       nzk += nlnk;
2441 
2442       /* update lnk by computing fill-in for each pivot row to be merged in */
2443       prow = jl[k]; /* 1st pivot row */
2444 
2445       while (prow < k) {
2446         nextprow = jl[prow];
2447 
2448         /* merge prow into k-th row */
2449         jmin  = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2450         jmax  = ui[prow+1];
2451         ncols = jmax-jmin;
2452         i     = jmin - ui[prow];
2453         cols  = uj_ptr[prow] + i; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2454         uj    = uj_lvl_ptr[prow] + i; /* levels of cols */
2455         j     = *(uj - 1);
2456         ierr  = PetscICCLLAddSorted(ncols,cols,levels,uj,am,nlnk,lnk,lnk_lvl,lnkbt,j);CHKERRQ(ierr);
2457         nzk  += nlnk;
2458 
2459         /* update il and jl for prow */
2460         if (jmin < jmax) {
2461           il[prow] = jmin;
2462           j        = *cols; jl[prow] = jl[j]; jl[j] = prow;
2463         }
2464         prow = nextprow;
2465       }
2466 
2467       /* if free space is not available, make more free space */
2468       if (current_space->local_remaining<nzk) {
2469         i    = am - k + 1; /* num of unfactored rows */
2470         i    = PetscIntMultTruncate(i,PetscMin(nzk, i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2471         ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2472         ierr = PetscFreeSpaceGet(i,&current_space_lvl);CHKERRQ(ierr);
2473         reallocs++;
2474       }
2475 
2476       /* copy data into free_space and free_space_lvl, then initialize lnk */
2477       if (nzk == 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Empty row %D in ICC matrix factor",k);
2478       ierr = PetscIncompleteLLClean(am,am,nzk,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
2479 
2480       /* add the k-th row into il and jl */
2481       if (nzk > 1) {
2482         i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2483         jl[k] = jl[i]; jl[i] = k;
2484         il[k] = ui[k] + 1;
2485       }
2486       uj_ptr[k]     = current_space->array;
2487       uj_lvl_ptr[k] = current_space_lvl->array;
2488 
2489       current_space->array           += nzk;
2490       current_space->local_used      += nzk;
2491       current_space->local_remaining -= nzk;
2492 
2493       current_space_lvl->array           += nzk;
2494       current_space_lvl->local_used      += nzk;
2495       current_space_lvl->local_remaining -= nzk;
2496 
2497       ui[k+1] = ui[k] + nzk;
2498     }
2499 
2500     ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2501     ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2502     ierr = PetscFree4(uj_ptr,uj_lvl_ptr,jl,il);CHKERRQ(ierr);
2503     ierr = PetscFree(ajtmp);CHKERRQ(ierr);
2504 
2505     /* copy free_space into uj and free free_space; set ui, uj, udiag in new datastructure; */
2506     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2507     ierr = PetscFreeSpaceContiguous_Cholesky(&free_space,uj,am,ui,udiag);CHKERRQ(ierr); /* store matrix factor  */
2508     ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2509     ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
2510 
2511   } /* end of case: levels>0 || (levels=0 && !perm_identity) */
2512 
2513   /* put together the new matrix in MATSEQSBAIJ format */
2514   b               = (Mat_SeqSBAIJ*)(fact)->data;
2515   b->singlemalloc = PETSC_FALSE;
2516 
2517   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2518 
2519   b->j             = uj;
2520   b->i             = ui;
2521   b->diag          = udiag;
2522   b->free_diag     = PETSC_TRUE;
2523   b->ilen          = 0;
2524   b->imax          = 0;
2525   b->row           = perm;
2526   b->col           = perm;
2527   ierr             = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2528   ierr             = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2529   b->icol          = iperm;
2530   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2531 
2532   ierr = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2533   ierr = PetscLogObjectMemory((PetscObject)fact,ui[am]*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2534 
2535   b->maxnz   = b->nz = ui[am];
2536   b->free_a  = PETSC_TRUE;
2537   b->free_ij = PETSC_TRUE;
2538 
2539   fact->info.factor_mallocs   = reallocs;
2540   fact->info.fill_ratio_given = fill;
2541   if (ai[am] != 0) {
2542     /* nonzeros in lower triangular part of A (including diagonals) = (ai[am]+am)/2 */
2543     fact->info.fill_ratio_needed = ((PetscReal)2*ui[am])/(ai[am]+am);
2544   } else {
2545     fact->info.fill_ratio_needed = 0.0;
2546   }
2547 #if defined(PETSC_USE_INFO)
2548   if (ai[am] != 0) {
2549     PetscReal af = fact->info.fill_ratio_needed;
2550     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2551     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2552     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2553   } else {
2554     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
2555   }
2556 #endif
2557   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ;
2558   PetscFunctionReturn(0);
2559 }
2560 
2561 PetscErrorCode MatICCFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2562 {
2563   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2564   Mat_SeqSBAIJ       *b;
2565   PetscErrorCode     ierr;
2566   PetscBool          perm_identity,missing;
2567   PetscInt           reallocs=0,i,*ai=a->i,*aj=a->j,am=A->rmap->n,*ui,*udiag;
2568   const PetscInt     *rip,*riip;
2569   PetscInt           jmin,jmax,nzk,k,j,*jl,prow,*il,nextprow;
2570   PetscInt           nlnk,*lnk,*lnk_lvl=NULL,d;
2571   PetscInt           ncols,ncols_upper,*cols,*ajtmp,*uj,**uj_ptr,**uj_lvl_ptr;
2572   PetscReal          fill          =info->fill,levels=info->levels;
2573   PetscFreeSpaceList free_space    =NULL,current_space=NULL;
2574   PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
2575   PetscBT            lnkbt;
2576   IS                 iperm;
2577 
2578   PetscFunctionBegin;
2579   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);
2580   ierr = MatMissingDiagonal(A,&missing,&d);CHKERRQ(ierr);
2581   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",d);
2582   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2583   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2584 
2585   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2586   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2587   ui[0] = 0;
2588 
2589   /* ICC(0) without matrix ordering: simply copies fill pattern */
2590   if (!levels && perm_identity) {
2591 
2592     for (i=0; i<am; i++) {
2593       ui[i+1]  = ui[i] + ai[i+1] - a->diag[i];
2594       udiag[i] = ui[i];
2595     }
2596     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2597     cols = uj;
2598     for (i=0; i<am; i++) {
2599       aj    = a->j + a->diag[i];
2600       ncols = ui[i+1] - ui[i];
2601       for (j=0; j<ncols; j++) *cols++ = *aj++;
2602     }
2603   } else { /* case: levels>0 || (levels=0 && !perm_identity) */
2604     ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2605     ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2606 
2607     /* initialization */
2608     ierr = PetscMalloc1(am+1,&ajtmp);CHKERRQ(ierr);
2609 
2610     /* jl: linked list for storing indices of the pivot rows
2611        il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2612     ierr = PetscMalloc4(am,&uj_ptr,am,&uj_lvl_ptr,am,&jl,am,&il);CHKERRQ(ierr);
2613     for (i=0; i<am; i++) {
2614       jl[i] = am; il[i] = 0;
2615     }
2616 
2617     /* create and initialize a linked list for storing column indices of the active row k */
2618     nlnk = am + 1;
2619     ierr = PetscIncompleteLLCreate(am,am,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2620 
2621     /* initial FreeSpace size is fill*(ai[am]+1) */
2622     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space);CHKERRQ(ierr);
2623     current_space     = free_space;
2624     ierr              = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space_lvl);CHKERRQ(ierr);
2625     current_space_lvl = free_space_lvl;
2626 
2627     for (k=0; k<am; k++) {  /* for each active row k */
2628       /* initialize lnk by the column indices of row rip[k] of A */
2629       nzk   = 0;
2630       ncols = ai[rip[k]+1] - ai[rip[k]];
2631       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);
2632       ncols_upper = 0;
2633       for (j=0; j<ncols; j++) {
2634         i = *(aj + ai[rip[k]] + j); /* unpermuted column index */
2635         if (riip[i] >= k) { /* only take upper triangular entry */
2636           ajtmp[ncols_upper] = i;
2637           ncols_upper++;
2638         }
2639       }
2640       ierr = PetscIncompleteLLInit(ncols_upper,ajtmp,am,riip,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
2641       nzk += nlnk;
2642 
2643       /* update lnk by computing fill-in for each pivot row to be merged in */
2644       prow = jl[k]; /* 1st pivot row */
2645 
2646       while (prow < k) {
2647         nextprow = jl[prow];
2648 
2649         /* merge prow into k-th row */
2650         jmin  = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2651         jmax  = ui[prow+1];
2652         ncols = jmax-jmin;
2653         i     = jmin - ui[prow];
2654         cols  = uj_ptr[prow] + i; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2655         uj    = uj_lvl_ptr[prow] + i; /* levels of cols */
2656         j     = *(uj - 1);
2657         ierr  = PetscICCLLAddSorted(ncols,cols,levels,uj,am,nlnk,lnk,lnk_lvl,lnkbt,j);CHKERRQ(ierr);
2658         nzk  += nlnk;
2659 
2660         /* update il and jl for prow */
2661         if (jmin < jmax) {
2662           il[prow] = jmin;
2663           j        = *cols; jl[prow] = jl[j]; jl[j] = prow;
2664         }
2665         prow = nextprow;
2666       }
2667 
2668       /* if free space is not available, make more free space */
2669       if (current_space->local_remaining<nzk) {
2670         i    = am - k + 1; /* num of unfactored rows */
2671         i    = PetscIntMultTruncate(i,PetscMin(nzk, i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2672         ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2673         ierr = PetscFreeSpaceGet(i,&current_space_lvl);CHKERRQ(ierr);
2674         reallocs++;
2675       }
2676 
2677       /* copy data into free_space and free_space_lvl, then initialize lnk */
2678       if (!nzk) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Empty row %D in ICC matrix factor",k);
2679       ierr = PetscIncompleteLLClean(am,am,nzk,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);
2680 
2681       /* add the k-th row into il and jl */
2682       if (nzk > 1) {
2683         i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2684         jl[k] = jl[i]; jl[i] = k;
2685         il[k] = ui[k] + 1;
2686       }
2687       uj_ptr[k]     = current_space->array;
2688       uj_lvl_ptr[k] = current_space_lvl->array;
2689 
2690       current_space->array           += nzk;
2691       current_space->local_used      += nzk;
2692       current_space->local_remaining -= nzk;
2693 
2694       current_space_lvl->array           += nzk;
2695       current_space_lvl->local_used      += nzk;
2696       current_space_lvl->local_remaining -= nzk;
2697 
2698       ui[k+1] = ui[k] + nzk;
2699     }
2700 
2701 #if defined(PETSC_USE_INFO)
2702     if (ai[am] != 0) {
2703       PetscReal af = (PetscReal)ui[am]/((PetscReal)ai[am]);
2704       ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2705       ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2706       ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2707     } else {
2708       ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
2709     }
2710 #endif
2711 
2712     ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2713     ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2714     ierr = PetscFree4(uj_ptr,uj_lvl_ptr,jl,il);CHKERRQ(ierr);
2715     ierr = PetscFree(ajtmp);CHKERRQ(ierr);
2716 
2717     /* destroy list of free space and other temporary array(s) */
2718     ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2719     ierr = PetscFreeSpaceContiguous(&free_space,uj);CHKERRQ(ierr);
2720     ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2721     ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
2722 
2723   } /* end of case: levels>0 || (levels=0 && !perm_identity) */
2724 
2725   /* put together the new matrix in MATSEQSBAIJ format */
2726 
2727   b               = (Mat_SeqSBAIJ*)fact->data;
2728   b->singlemalloc = PETSC_FALSE;
2729 
2730   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2731 
2732   b->j         = uj;
2733   b->i         = ui;
2734   b->diag      = udiag;
2735   b->free_diag = PETSC_TRUE;
2736   b->ilen      = 0;
2737   b->imax      = 0;
2738   b->row       = perm;
2739   b->col       = perm;
2740 
2741   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2742   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2743 
2744   b->icol          = iperm;
2745   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2746   ierr             = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2747   ierr             = PetscLogObjectMemory((PetscObject)fact,(ui[am]-am)*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2748   b->maxnz         = b->nz = ui[am];
2749   b->free_a        = PETSC_TRUE;
2750   b->free_ij       = PETSC_TRUE;
2751 
2752   fact->info.factor_mallocs   = reallocs;
2753   fact->info.fill_ratio_given = fill;
2754   if (ai[am] != 0) {
2755     fact->info.fill_ratio_needed = ((PetscReal)ui[am])/((PetscReal)ai[am]);
2756   } else {
2757     fact->info.fill_ratio_needed = 0.0;
2758   }
2759   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ_inplace;
2760   PetscFunctionReturn(0);
2761 }
2762 
2763 PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2764 {
2765   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2766   Mat_SeqSBAIJ       *b;
2767   PetscErrorCode     ierr;
2768   PetscBool          perm_identity,missing;
2769   PetscReal          fill = info->fill;
2770   const PetscInt     *rip,*riip;
2771   PetscInt           i,am=A->rmap->n,*ai=a->i,*aj=a->j,reallocs=0,prow;
2772   PetscInt           *jl,jmin,jmax,nzk,*ui,k,j,*il,nextprow;
2773   PetscInt           nlnk,*lnk,ncols,ncols_upper,*cols,*uj,**ui_ptr,*uj_ptr,*udiag;
2774   PetscFreeSpaceList free_space=NULL,current_space=NULL;
2775   PetscBT            lnkbt;
2776   IS                 iperm;
2777 
2778   PetscFunctionBegin;
2779   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);
2780   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
2781   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
2782 
2783   /* check whether perm is the identity mapping */
2784   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2785   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2786   ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2787   ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2788 
2789   /* initialization */
2790   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2791   ierr  = PetscMalloc1(am+1,&udiag);CHKERRQ(ierr);
2792   ui[0] = 0;
2793 
2794   /* jl: linked list for storing indices of the pivot rows
2795      il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2796   ierr = PetscMalloc4(am,&ui_ptr,am,&jl,am,&il,am,&cols);CHKERRQ(ierr);
2797   for (i=0; i<am; i++) {
2798     jl[i] = am; il[i] = 0;
2799   }
2800 
2801   /* create and initialize a linked list for storing column indices of the active row k */
2802   nlnk = am + 1;
2803   ierr = PetscLLCreate(am,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2804 
2805   /* initial FreeSpace size is fill*(ai[am]+am)/2 */
2806   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,(ai[am]+am)/2),&free_space);CHKERRQ(ierr);
2807   current_space = free_space;
2808 
2809   for (k=0; k<am; k++) {  /* for each active row k */
2810     /* initialize lnk by the column indices of row rip[k] of A */
2811     nzk   = 0;
2812     ncols = ai[rip[k]+1] - ai[rip[k]];
2813     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);
2814     ncols_upper = 0;
2815     for (j=0; j<ncols; j++) {
2816       i = riip[*(aj + ai[rip[k]] + j)];
2817       if (i >= k) { /* only take upper triangular entry */
2818         cols[ncols_upper] = i;
2819         ncols_upper++;
2820       }
2821     }
2822     ierr = PetscLLAdd(ncols_upper,cols,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2823     nzk += nlnk;
2824 
2825     /* update lnk by computing fill-in for each pivot row to be merged in */
2826     prow = jl[k]; /* 1st pivot row */
2827 
2828     while (prow < k) {
2829       nextprow = jl[prow];
2830       /* merge prow into k-th row */
2831       jmin   = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
2832       jmax   = ui[prow+1];
2833       ncols  = jmax-jmin;
2834       uj_ptr = ui_ptr[prow] + jmin - ui[prow]; /* points to the 2nd nzero entry in U(prow,k:am-1) */
2835       ierr   = PetscLLAddSorted(ncols,uj_ptr,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2836       nzk   += nlnk;
2837 
2838       /* update il and jl for prow */
2839       if (jmin < jmax) {
2840         il[prow] = jmin;
2841         j        = *uj_ptr;
2842         jl[prow] = jl[j];
2843         jl[j]    = prow;
2844       }
2845       prow = nextprow;
2846     }
2847 
2848     /* if free space is not available, make more free space */
2849     if (current_space->local_remaining<nzk) {
2850       i    = am - k + 1; /* num of unfactored rows */
2851       i    = PetscIntMultTruncate(i,PetscMin(nzk,i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
2852       ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
2853       reallocs++;
2854     }
2855 
2856     /* copy data into free space, then initialize lnk */
2857     ierr = PetscLLClean(am,am,nzk,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
2858 
2859     /* add the k-th row into il and jl */
2860     if (nzk > 1) {
2861       i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
2862       jl[k] = jl[i]; jl[i] = k;
2863       il[k] = ui[k] + 1;
2864     }
2865     ui_ptr[k] = current_space->array;
2866 
2867     current_space->array           += nzk;
2868     current_space->local_used      += nzk;
2869     current_space->local_remaining -= nzk;
2870 
2871     ui[k+1] = ui[k] + nzk;
2872   }
2873 
2874   ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
2875   ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
2876   ierr = PetscFree4(ui_ptr,jl,il,cols);CHKERRQ(ierr);
2877 
2878   /* copy free_space into uj and free free_space; set ui, uj, udiag in new datastructure; */
2879   ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
2880   ierr = PetscFreeSpaceContiguous_Cholesky(&free_space,uj,am,ui,udiag);CHKERRQ(ierr); /* store matrix factor */
2881   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
2882 
2883   /* put together the new matrix in MATSEQSBAIJ format */
2884 
2885   b               = (Mat_SeqSBAIJ*)fact->data;
2886   b->singlemalloc = PETSC_FALSE;
2887   b->free_a       = PETSC_TRUE;
2888   b->free_ij      = PETSC_TRUE;
2889 
2890   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
2891 
2892   b->j         = uj;
2893   b->i         = ui;
2894   b->diag      = udiag;
2895   b->free_diag = PETSC_TRUE;
2896   b->ilen      = 0;
2897   b->imax      = 0;
2898   b->row       = perm;
2899   b->col       = perm;
2900 
2901   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2902   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
2903 
2904   b->icol          = iperm;
2905   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
2906 
2907   ierr = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
2908   ierr = PetscLogObjectMemory((PetscObject)fact,ui[am]*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
2909 
2910   b->maxnz = b->nz = ui[am];
2911 
2912   fact->info.factor_mallocs   = reallocs;
2913   fact->info.fill_ratio_given = fill;
2914   if (ai[am] != 0) {
2915     /* nonzeros in lower triangular part of A (including diagonals) = (ai[am]+am)/2 */
2916     fact->info.fill_ratio_needed = ((PetscReal)2*ui[am])/(ai[am]+am);
2917   } else {
2918     fact->info.fill_ratio_needed = 0.0;
2919   }
2920 #if defined(PETSC_USE_INFO)
2921   if (ai[am] != 0) {
2922     PetscReal af = fact->info.fill_ratio_needed;
2923     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
2924     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
2925     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
2926   } else {
2927     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
2928   }
2929 #endif
2930   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ;
2931   PetscFunctionReturn(0);
2932 }
2933 
2934 PetscErrorCode MatCholeskyFactorSymbolic_SeqAIJ_inplace(Mat fact,Mat A,IS perm,const MatFactorInfo *info)
2935 {
2936   Mat_SeqAIJ         *a = (Mat_SeqAIJ*)A->data;
2937   Mat_SeqSBAIJ       *b;
2938   PetscErrorCode     ierr;
2939   PetscBool          perm_identity,missing;
2940   PetscReal          fill = info->fill;
2941   const PetscInt     *rip,*riip;
2942   PetscInt           i,am=A->rmap->n,*ai=a->i,*aj=a->j,reallocs=0,prow;
2943   PetscInt           *jl,jmin,jmax,nzk,*ui,k,j,*il,nextprow;
2944   PetscInt           nlnk,*lnk,ncols,ncols_upper,*cols,*uj,**ui_ptr,*uj_ptr;
2945   PetscFreeSpaceList free_space=NULL,current_space=NULL;
2946   PetscBT            lnkbt;
2947   IS                 iperm;
2948 
2949   PetscFunctionBegin;
2950   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);
2951   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
2952   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
2953 
2954   /* check whether perm is the identity mapping */
2955   ierr = ISIdentity(perm,&perm_identity);CHKERRQ(ierr);
2956   ierr = ISInvertPermutation(perm,PETSC_DECIDE,&iperm);CHKERRQ(ierr);
2957   ierr = ISGetIndices(iperm,&riip);CHKERRQ(ierr);
2958   ierr = ISGetIndices(perm,&rip);CHKERRQ(ierr);
2959 
2960   /* initialization */
2961   ierr  = PetscMalloc1(am+1,&ui);CHKERRQ(ierr);
2962   ui[0] = 0;
2963 
2964   /* jl: linked list for storing indices of the pivot rows
2965      il: il[i] points to the 1st nonzero entry of U(i,k:am-1) */
2966   ierr = PetscMalloc4(am,&ui_ptr,am,&jl,am,&il,am,&cols);CHKERRQ(ierr);
2967   for (i=0; i<am; i++) {
2968     jl[i] = am; il[i] = 0;
2969   }
2970 
2971   /* create and initialize a linked list for storing column indices of the active row k */
2972   nlnk = am + 1;
2973   ierr = PetscLLCreate(am,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2974 
2975   /* initial FreeSpace size is fill*(ai[am]+1) */
2976   ierr          = PetscFreeSpaceGet(PetscRealIntMultTruncate(fill,ai[am]+1),&free_space);CHKERRQ(ierr);
2977   current_space = free_space;
2978 
2979   for (k=0; k<am; k++) {  /* for each active row k */
2980     /* initialize lnk by the column indices of row rip[k] of A */
2981     nzk   = 0;
2982     ncols = ai[rip[k]+1] - ai[rip[k]];
2983     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);
2984     ncols_upper = 0;
2985     for (j=0; j<ncols; j++) {
2986       i = riip[*(aj + ai[rip[k]] + j)];
2987       if (i >= k) { /* only take upper triangular entry */
2988         cols[ncols_upper] = i;
2989         ncols_upper++;
2990       }
2991     }
2992     ierr = PetscLLAdd(ncols_upper,cols,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
2993     nzk += nlnk;
2994 
2995     /* update lnk by computing fill-in for each pivot row to be merged in */
2996     prow = jl[k]; /* 1st pivot row */
2997 
2998     while (prow < k) {
2999       nextprow = jl[prow];
3000       /* merge prow into k-th row */
3001       jmin   = il[prow] + 1; /* index of the 2nd nzero entry in U(prow,k:am-1) */
3002       jmax   = ui[prow+1];
3003       ncols  = jmax-jmin;
3004       uj_ptr = ui_ptr[prow] + jmin - ui[prow]; /* points to the 2nd nzero entry in U(prow,k:am-1) */
3005       ierr   = PetscLLAddSorted(ncols,uj_ptr,am,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3006       nzk   += nlnk;
3007 
3008       /* update il and jl for prow */
3009       if (jmin < jmax) {
3010         il[prow] = jmin;
3011         j        = *uj_ptr; jl[prow] = jl[j]; jl[j] = prow;
3012       }
3013       prow = nextprow;
3014     }
3015 
3016     /* if free space is not available, make more free space */
3017     if (current_space->local_remaining<nzk) {
3018       i    = am - k + 1; /* num of unfactored rows */
3019       i    = PetscMin(i*nzk, i*(i-1)); /* i*nzk, i*(i-1): estimated and max additional space needed */
3020       ierr = PetscFreeSpaceGet(i,&current_space);CHKERRQ(ierr);
3021       reallocs++;
3022     }
3023 
3024     /* copy data into free space, then initialize lnk */
3025     ierr = PetscLLClean(am,am,nzk,lnk,current_space->array,lnkbt);CHKERRQ(ierr);
3026 
3027     /* add the k-th row into il and jl */
3028     if (nzk-1 > 0) {
3029       i     = current_space->array[1]; /* col value of the first nonzero element in U(k, k+1:am-1) */
3030       jl[k] = jl[i]; jl[i] = k;
3031       il[k] = ui[k] + 1;
3032     }
3033     ui_ptr[k] = current_space->array;
3034 
3035     current_space->array           += nzk;
3036     current_space->local_used      += nzk;
3037     current_space->local_remaining -= nzk;
3038 
3039     ui[k+1] = ui[k] + nzk;
3040   }
3041 
3042 #if defined(PETSC_USE_INFO)
3043   if (ai[am] != 0) {
3044     PetscReal af = (PetscReal)(ui[am])/((PetscReal)ai[am]);
3045     ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %g needed %g\n",reallocs,(double)fill,(double)af);CHKERRQ(ierr);
3046     ierr = PetscInfo1(A,"Run with -pc_factor_fill %g or use \n",(double)af);CHKERRQ(ierr);
3047     ierr = PetscInfo1(A,"PCFactorSetFill(pc,%g) for best performance.\n",(double)af);CHKERRQ(ierr);
3048   } else {
3049     ierr = PetscInfo(A,"Empty matrix\n");CHKERRQ(ierr);
3050   }
3051 #endif
3052 
3053   ierr = ISRestoreIndices(perm,&rip);CHKERRQ(ierr);
3054   ierr = ISRestoreIndices(iperm,&riip);CHKERRQ(ierr);
3055   ierr = PetscFree4(ui_ptr,jl,il,cols);CHKERRQ(ierr);
3056 
3057   /* destroy list of free space and other temporary array(s) */
3058   ierr = PetscMalloc1(ui[am]+1,&uj);CHKERRQ(ierr);
3059   ierr = PetscFreeSpaceContiguous(&free_space,uj);CHKERRQ(ierr);
3060   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
3061 
3062   /* put together the new matrix in MATSEQSBAIJ format */
3063 
3064   b               = (Mat_SeqSBAIJ*)fact->data;
3065   b->singlemalloc = PETSC_FALSE;
3066   b->free_a       = PETSC_TRUE;
3067   b->free_ij      = PETSC_TRUE;
3068 
3069   ierr = PetscMalloc1(ui[am]+1,&b->a);CHKERRQ(ierr);
3070 
3071   b->j    = uj;
3072   b->i    = ui;
3073   b->diag = 0;
3074   b->ilen = 0;
3075   b->imax = 0;
3076   b->row  = perm;
3077   b->col  = perm;
3078 
3079   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
3080   ierr = PetscObjectReference((PetscObject)perm);CHKERRQ(ierr);
3081 
3082   b->icol          = iperm;
3083   b->pivotinblocks = PETSC_FALSE; /* need to get from MatFactorInfo */
3084 
3085   ierr     = PetscMalloc1(am+1,&b->solve_work);CHKERRQ(ierr);
3086   ierr     = PetscLogObjectMemory((PetscObject)fact,(ui[am]-am)*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
3087   b->maxnz = b->nz = ui[am];
3088 
3089   fact->info.factor_mallocs   = reallocs;
3090   fact->info.fill_ratio_given = fill;
3091   if (ai[am] != 0) {
3092     fact->info.fill_ratio_needed = ((PetscReal)ui[am])/((PetscReal)ai[am]);
3093   } else {
3094     fact->info.fill_ratio_needed = 0.0;
3095   }
3096   fact->ops->choleskyfactornumeric = MatCholeskyFactorNumeric_SeqAIJ_inplace;
3097   PetscFunctionReturn(0);
3098 }
3099 
3100 PetscErrorCode MatSolve_SeqAIJ_NaturalOrdering(Mat A,Vec bb,Vec xx)
3101 {
3102   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
3103   PetscErrorCode    ierr;
3104   PetscInt          n   = A->rmap->n;
3105   const PetscInt    *ai = a->i,*aj = a->j,*adiag = a->diag,*vi;
3106   PetscScalar       *x,sum;
3107   const PetscScalar *b;
3108   const MatScalar   *aa = a->a,*v;
3109   PetscInt          i,nz;
3110 
3111   PetscFunctionBegin;
3112   if (!n) PetscFunctionReturn(0);
3113 
3114   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
3115   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
3116 
3117   /* forward solve the lower triangular */
3118   x[0] = b[0];
3119   v    = aa;
3120   vi   = aj;
3121   for (i=1; i<n; i++) {
3122     nz  = ai[i+1] - ai[i];
3123     sum = b[i];
3124     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
3125     v   += nz;
3126     vi  += nz;
3127     x[i] = sum;
3128   }
3129 
3130   /* backward solve the upper triangular */
3131   for (i=n-1; i>=0; i--) {
3132     v   = aa + adiag[i+1] + 1;
3133     vi  = aj + adiag[i+1] + 1;
3134     nz  = adiag[i] - adiag[i+1]-1;
3135     sum = x[i];
3136     PetscSparseDenseMinusDot(sum,x,v,vi,nz);
3137     x[i] = sum*v[nz]; /* x[i]=aa[adiag[i]]*sum; v++; */
3138   }
3139 
3140   ierr = PetscLogFlops(2.0*a->nz - A->cmap->n);CHKERRQ(ierr);
3141   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
3142   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
3143   PetscFunctionReturn(0);
3144 }
3145 
3146 PetscErrorCode MatSolve_SeqAIJ(Mat A,Vec bb,Vec xx)
3147 {
3148   Mat_SeqAIJ        *a    = (Mat_SeqAIJ*)A->data;
3149   IS                iscol = a->col,isrow = a->row;
3150   PetscErrorCode    ierr;
3151   PetscInt          i,n=A->rmap->n,*vi,*ai=a->i,*aj=a->j,*adiag = a->diag,nz;
3152   const PetscInt    *rout,*cout,*r,*c;
3153   PetscScalar       *x,*tmp,sum;
3154   const PetscScalar *b;
3155   const MatScalar   *aa = a->a,*v;
3156 
3157   PetscFunctionBegin;
3158   if (!n) PetscFunctionReturn(0);
3159 
3160   ierr = VecGetArrayRead(bb,&b);CHKERRQ(ierr);
3161   ierr = VecGetArrayWrite(xx,&x);CHKERRQ(ierr);
3162   tmp  = a->solve_work;
3163 
3164   ierr = ISGetIndices(isrow,&rout);CHKERRQ(ierr); r = rout;
3165   ierr = ISGetIndices(iscol,&cout);CHKERRQ(ierr); c = cout;
3166 
3167   /* forward solve the lower triangular */
3168   tmp[0] = b[r[0]];
3169   v      = aa;
3170   vi     = aj;
3171   for (i=1; i<n; i++) {
3172     nz  = ai[i+1] - ai[i];
3173     sum = b[r[i]];
3174     PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
3175     tmp[i] = sum;
3176     v     += nz; vi += nz;
3177   }
3178 
3179   /* backward solve the upper triangular */
3180   for (i=n-1; i>=0; i--) {
3181     v   = aa + adiag[i+1]+1;
3182     vi  = aj + adiag[i+1]+1;
3183     nz  = adiag[i]-adiag[i+1]-1;
3184     sum = tmp[i];
3185     PetscSparseDenseMinusDot(sum,tmp,v,vi,nz);
3186     x[c[i]] = tmp[i] = sum*v[nz]; /* v[nz] = aa[adiag[i]] */
3187   }
3188 
3189   ierr = ISRestoreIndices(isrow,&rout);CHKERRQ(ierr);
3190   ierr = ISRestoreIndices(iscol,&cout);CHKERRQ(ierr);
3191   ierr = VecRestoreArrayRead(bb,&b);CHKERRQ(ierr);
3192   ierr = VecRestoreArrayWrite(xx,&x);CHKERRQ(ierr);
3193   ierr = PetscLogFlops(2*a->nz - A->cmap->n);CHKERRQ(ierr);
3194   PetscFunctionReturn(0);
3195 }
3196 
3197 /*
3198     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
3199 */
3200 PetscErrorCode MatILUDTFactor_SeqAIJ(Mat A,IS isrow,IS iscol,const MatFactorInfo *info,Mat *fact)
3201 {
3202   Mat            B = *fact;
3203   Mat_SeqAIJ     *a=(Mat_SeqAIJ*)A->data,*b;
3204   IS             isicol;
3205   PetscErrorCode ierr;
3206   const PetscInt *r,*ic;
3207   PetscInt       i,n=A->rmap->n,*ai=a->i,*aj=a->j,*ajtmp,*adiag;
3208   PetscInt       *bi,*bj,*bdiag,*bdiag_rev;
3209   PetscInt       row,nzi,nzi_bl,nzi_bu,*im,nzi_al,nzi_au;
3210   PetscInt       nlnk,*lnk;
3211   PetscBT        lnkbt;
3212   PetscBool      row_identity,icol_identity;
3213   MatScalar      *aatmp,*pv,*batmp,*ba,*rtmp,*pc,multiplier,*vtmp,diag_tmp;
3214   const PetscInt *ics;
3215   PetscInt       j,nz,*pj,*bjtmp,k,ncut,*jtmp;
3216   PetscReal      dt     =info->dt,shift=info->shiftamount;
3217   PetscInt       dtcount=(PetscInt)info->dtcount,nnz_max;
3218   PetscBool      missing;
3219 
3220   PetscFunctionBegin;
3221   if (dt      == PETSC_DEFAULT) dt = 0.005;
3222   if (dtcount == PETSC_DEFAULT) dtcount = (PetscInt)(1.5*a->rmax);
3223 
3224   /* ------- symbolic factorization, can be reused ---------*/
3225   ierr = MatMissingDiagonal(A,&missing,&i);CHKERRQ(ierr);
3226   if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",i);
3227   adiag=a->diag;
3228 
3229   ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);
3230 
3231   /* bdiag is location of diagonal in factor */
3232   ierr = PetscMalloc1(n+1,&bdiag);CHKERRQ(ierr);     /* becomes b->diag */
3233   ierr = PetscMalloc1(n+1,&bdiag_rev);CHKERRQ(ierr); /* temporary */
3234 
3235   /* allocate row pointers bi */
3236   ierr = PetscMalloc1(2*n+2,&bi);CHKERRQ(ierr);
3237 
3238   /* allocate bj and ba; max num of nonzero entries is (ai[n]+2*n*dtcount+2) */
3239   if (dtcount > n-1) dtcount = n-1; /* diagonal is excluded */
3240   nnz_max = ai[n]+2*n*dtcount+2;
3241 
3242   ierr = PetscMalloc1(nnz_max+1,&bj);CHKERRQ(ierr);
3243   ierr = PetscMalloc1(nnz_max+1,&ba);CHKERRQ(ierr);
3244 
3245   /* put together the new matrix */
3246   ierr = MatSeqAIJSetPreallocation_SeqAIJ(B,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
3247   ierr = PetscLogObjectParent((PetscObject)B,(PetscObject)isicol);CHKERRQ(ierr);
3248   b    = (Mat_SeqAIJ*)B->data;
3249 
3250   b->free_a       = PETSC_TRUE;
3251   b->free_ij      = PETSC_TRUE;
3252   b->singlemalloc = PETSC_FALSE;
3253 
3254   b->a    = ba;
3255   b->j    = bj;
3256   b->i    = bi;
3257   b->diag = bdiag;
3258   b->ilen = 0;
3259   b->imax = 0;
3260   b->row  = isrow;
3261   b->col  = iscol;
3262   ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
3263   ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
3264   b->icol = isicol;
3265 
3266   ierr     = PetscMalloc1(n+1,&b->solve_work);CHKERRQ(ierr);
3267   ierr     = PetscLogObjectMemory((PetscObject)B,nnz_max*(sizeof(PetscInt)+sizeof(MatScalar)));CHKERRQ(ierr);
3268   b->maxnz = nnz_max;
3269 
3270   B->factortype            = MAT_FACTOR_ILUDT;
3271   B->info.factor_mallocs   = 0;
3272   B->info.fill_ratio_given = ((PetscReal)nnz_max)/((PetscReal)ai[n]);
3273   /* ------- end of symbolic factorization ---------*/
3274 
3275   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
3276   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
3277   ics  = ic;
3278 
3279   /* linked list for storing column indices of the active row */
3280   nlnk = n + 1;
3281   ierr = PetscLLCreate(n,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3282 
3283   /* im: used by PetscLLAddSortedLU(); jtmp: working array for column indices of active row */
3284   ierr = PetscMalloc2(n,&im,n,&jtmp);CHKERRQ(ierr);
3285   /* rtmp, vtmp: working arrays for sparse and contiguous row entries of active row */
3286   ierr = PetscMalloc2(n,&rtmp,n,&vtmp);CHKERRQ(ierr);
3287   ierr = PetscArrayzero(rtmp,n);CHKERRQ(ierr);
3288 
3289   bi[0]        = 0;
3290   bdiag[0]     = nnz_max-1; /* location of diag[0] in factor B */
3291   bdiag_rev[n] = bdiag[0];
3292   bi[2*n+1]    = bdiag[0]+1; /* endof bj and ba array */
3293   for (i=0; i<n; i++) {
3294     /* copy initial fill into linked list */
3295     nzi = ai[r[i]+1] - ai[r[i]];
3296     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);
3297     nzi_al = adiag[r[i]] - ai[r[i]];
3298     nzi_au = ai[r[i]+1] - adiag[r[i]] -1;
3299     ajtmp  = aj + ai[r[i]];
3300     ierr   = PetscLLAddPerm(nzi,ajtmp,ic,n,nlnk,lnk,lnkbt);CHKERRQ(ierr);
3301 
3302     /* load in initial (unfactored row) */
3303     aatmp = a->a + ai[r[i]];
3304     for (j=0; j<nzi; j++) {
3305       rtmp[ics[*ajtmp++]] = *aatmp++;
3306     }
3307 
3308     /* add pivot rows into linked list */
3309     row = lnk[n];
3310     while (row < i) {
3311       nzi_bl = bi[row+1] - bi[row] + 1;
3312       bjtmp  = bj + bdiag[row+1]+1; /* points to 1st column next to the diagonal in U */
3313       ierr   = PetscLLAddSortedLU(bjtmp,row,nlnk,lnk,lnkbt,i,nzi_bl,im);CHKERRQ(ierr);
3314       nzi   += nlnk;
3315       row    = lnk[row];
3316     }
3317 
3318     /* copy data from lnk into jtmp, then initialize lnk */
3319     ierr = PetscLLClean(n,n,nzi,lnk,jtmp,lnkbt);CHKERRQ(ierr);
3320 
3321     /* numerical factorization */
3322     bjtmp = jtmp;
3323     row   = *bjtmp++; /* 1st pivot row */
3324     while (row < i) {
3325       pc         = rtmp + row;
3326       pv         = ba + bdiag[row]; /* 1./(diag of the pivot row) */
3327       multiplier = (*pc) * (*pv);
3328       *pc        = multiplier;
3329       if (PetscAbsScalar(*pc) > dt) { /* apply tolerance dropping rule */
3330         pj = bj + bdiag[row+1] + 1;         /* point to 1st entry of U(row,:) */
3331         pv = ba + bdiag[row+1] + 1;
3332         nz = bdiag[row] - bdiag[row+1] - 1;         /* num of entries in U(row,:), excluding diagonal */
3333         for (j=0; j<nz; j++) rtmp[*pj++] -= multiplier * (*pv++);
3334         ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
3335       }
3336       row = *bjtmp++;
3337     }
3338 
3339     /* copy sparse rtmp into contiguous vtmp; separate L and U part */
3340     diag_tmp = rtmp[i];  /* save diagonal value - may not needed?? */
3341     nzi_bl   = 0; j = 0;
3342     while (jtmp[j] < i) { /* Note: jtmp is sorted */
3343       vtmp[j] = rtmp[jtmp[j]]; rtmp[jtmp[j]]=0.0;
3344       nzi_bl++; j++;
3345     }
3346     nzi_bu = nzi - nzi_bl -1;
3347     while (j < nzi) {
3348       vtmp[j] = rtmp[jtmp[j]]; rtmp[jtmp[j]]=0.0;
3349       j++;
3350     }
3351 
3352     bjtmp = bj + bi[i];
3353     batmp = ba + bi[i];
3354     /* apply level dropping rule to L part */
3355     ncut = nzi_al + dtcount;
3356     if (ncut < nzi_bl) {
3357       ierr = PetscSortSplit(ncut,nzi_bl,vtmp,jtmp);CHKERRQ(ierr);
3358       ierr = PetscSortIntWithScalarArray(ncut,jtmp,vtmp);CHKERRQ(ierr);
3359     } else {
3360       ncut = nzi_bl;
3361     }
3362     for (j=0; j<ncut; j++) {
3363       bjtmp[j] = jtmp[j];
3364       batmp[j] = vtmp[j];
3365     }
3366     bi[i+1] = bi[i] + ncut;
3367     nzi     = ncut + 1;
3368 
3369     /* apply level dropping rule to U part */
3370     ncut = nzi_au + dtcount;
3371     if (ncut < nzi_bu) {
3372       ierr = PetscSortSplit(ncut,nzi_bu,vtmp+nzi_bl+1,jtmp+nzi_bl+1);CHKERRQ(ierr);
3373       ierr = PetscSortIntWithScalarArray(ncut,jtmp+nzi_bl+1,vtmp+nzi_bl+1);CHKERRQ(ierr);
3374     } else {
3375       ncut = nzi_bu;
3376     }
3377     nzi += ncut;
3378 
3379     /* mark bdiagonal */
3380     bdiag[i+1]       = bdiag[i] - (ncut + 1);
3381     bdiag_rev[n-i-1] = bdiag[i+1];
3382     bi[2*n - i]      = bi[2*n - i +1] - (ncut + 1);
3383     bjtmp            = bj + bdiag[i];
3384     batmp            = ba + bdiag[i];
3385     *bjtmp           = i;
3386     *batmp           = diag_tmp; /* rtmp[i]; */
3387     if (*batmp == 0.0) {
3388       *batmp = dt+shift;
3389     }
3390     *batmp = 1.0/(*batmp); /* invert diagonal entries for simplier triangular solves */
3391 
3392     bjtmp = bj + bdiag[i+1]+1;
3393     batmp = ba + bdiag[i+1]+1;
3394     for (k=0; k<ncut; k++) {
3395       bjtmp[k] = jtmp[nzi_bl+1+k];
3396       batmp[k] = vtmp[nzi_bl+1+k];
3397     }
3398 
3399     im[i] = nzi;   /* used by PetscLLAddSortedLU() */
3400   } /* for (i=0; i<n; i++) */
3401   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]);
3402 
3403   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
3404   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
3405 
3406   ierr = PetscLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
3407   ierr = PetscFree2(im,jtmp);CHKERRQ(ierr);
3408   ierr = PetscFree2(rtmp,vtmp);CHKERRQ(ierr);
3409   ierr = PetscFree(bdiag_rev);CHKERRQ(ierr);
3410 
3411   ierr     = PetscLogFlops(B->cmap->n);CHKERRQ(ierr);
3412   b->maxnz = b->nz = bi[n] + bdiag[0] - bdiag[n];
3413 
3414   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
3415   ierr = ISIdentity(isicol,&icol_identity);CHKERRQ(ierr);
3416   if (row_identity && icol_identity) {
3417     B->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
3418   } else {
3419     B->ops->solve = MatSolve_SeqAIJ;
3420   }
3421 
3422   B->ops->solveadd          = 0;
3423   B->ops->solvetranspose    = 0;
3424   B->ops->solvetransposeadd = 0;
3425   B->ops->matsolve          = 0;
3426   B->assembled              = PETSC_TRUE;
3427   B->preallocated           = PETSC_TRUE;
3428   PetscFunctionReturn(0);
3429 }
3430 
3431 /* a wraper of MatILUDTFactor_SeqAIJ() */
3432 /*
3433     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
3434 */
3435 
3436 PetscErrorCode  MatILUDTFactorSymbolic_SeqAIJ(Mat fact,Mat A,IS row,IS col,const MatFactorInfo *info)
3437 {
3438   PetscErrorCode ierr;
3439 
3440   PetscFunctionBegin;
3441   ierr = MatILUDTFactor_SeqAIJ(A,row,col,info,&fact);CHKERRQ(ierr);
3442   PetscFunctionReturn(0);
3443 }
3444 
3445 /*
3446    same as MatLUFactorNumeric_SeqAIJ(), except using contiguous array matrix factors
3447    - intend to replace existing MatLUFactorNumeric_SeqAIJ()
3448 */
3449 /*
3450     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
3451 */
3452 
3453 PetscErrorCode  MatILUDTFactorNumeric_SeqAIJ(Mat fact,Mat A,const MatFactorInfo *info)
3454 {
3455   Mat            C     =fact;
3456   Mat_SeqAIJ     *a    =(Mat_SeqAIJ*)A->data,*b=(Mat_SeqAIJ*)C->data;
3457   IS             isrow = b->row,isicol = b->icol;
3458   PetscErrorCode ierr;
3459   const PetscInt *r,*ic,*ics;
3460   PetscInt       i,j,k,n=A->rmap->n,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j;
3461   PetscInt       *ajtmp,*bjtmp,nz,nzl,nzu,row,*bdiag = b->diag,*pj;
3462   MatScalar      *rtmp,*pc,multiplier,*v,*pv,*aa=a->a;
3463   PetscReal      dt=info->dt,shift=info->shiftamount;
3464   PetscBool      row_identity, col_identity;
3465 
3466   PetscFunctionBegin;
3467   ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
3468   ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);
3469   ierr = PetscMalloc1(n+1,&rtmp);CHKERRQ(ierr);
3470   ics  = ic;
3471 
3472   for (i=0; i<n; i++) {
3473     /* initialize rtmp array */
3474     nzl   = bi[i+1] - bi[i];       /* num of nozeros in L(i,:) */
3475     bjtmp = bj + bi[i];
3476     for  (j=0; j<nzl; j++) rtmp[*bjtmp++] = 0.0;
3477     rtmp[i] = 0.0;
3478     nzu     = bdiag[i] - bdiag[i+1]; /* num of nozeros in U(i,:) */
3479     bjtmp   = bj + bdiag[i+1] + 1;
3480     for  (j=0; j<nzu; j++) rtmp[*bjtmp++] = 0.0;
3481 
3482     /* load in initial unfactored row of A */
3483     nz    = ai[r[i]+1] - ai[r[i]];
3484     ajtmp = aj + ai[r[i]];
3485     v     = aa + ai[r[i]];
3486     for (j=0; j<nz; j++) {
3487       rtmp[ics[*ajtmp++]] = v[j];
3488     }
3489 
3490     /* numerical factorization */
3491     bjtmp = bj + bi[i]; /* point to 1st entry of L(i,:) */
3492     nzl   = bi[i+1] - bi[i]; /* num of entries in L(i,:) */
3493     k     = 0;
3494     while (k < nzl) {
3495       row        = *bjtmp++;
3496       pc         = rtmp + row;
3497       pv         = b->a + bdiag[row]; /* 1./(diag of the pivot row) */
3498       multiplier = (*pc) * (*pv);
3499       *pc        = multiplier;
3500       if (PetscAbsScalar(multiplier) > dt) {
3501         pj = bj + bdiag[row+1] + 1;         /* point to 1st entry of U(row,:) */
3502         pv = b->a + bdiag[row+1] + 1;
3503         nz = bdiag[row] - bdiag[row+1] - 1;         /* num of entries in U(row,:), excluding diagonal */
3504         for (j=0; j<nz; j++) rtmp[*pj++] -= multiplier * (*pv++);
3505         ierr = PetscLogFlops(1+2*nz);CHKERRQ(ierr);
3506       }
3507       k++;
3508     }
3509 
3510     /* finished row so stick it into b->a */
3511     /* L-part */
3512     pv  = b->a + bi[i];
3513     pj  = bj + bi[i];
3514     nzl = bi[i+1] - bi[i];
3515     for (j=0; j<nzl; j++) {
3516       pv[j] = rtmp[pj[j]];
3517     }
3518 
3519     /* diagonal: invert diagonal entries for simplier triangular solves */
3520     if (rtmp[i] == 0.0) rtmp[i] = dt+shift;
3521     b->a[bdiag[i]] = 1.0/rtmp[i];
3522 
3523     /* U-part */
3524     pv  = b->a + bdiag[i+1] + 1;
3525     pj  = bj + bdiag[i+1] + 1;
3526     nzu = bdiag[i] - bdiag[i+1] - 1;
3527     for (j=0; j<nzu; j++) {
3528       pv[j] = rtmp[pj[j]];
3529     }
3530   }
3531 
3532   ierr = PetscFree(rtmp);CHKERRQ(ierr);
3533   ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
3534   ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
3535 
3536   ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
3537   ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);
3538   if (row_identity && col_identity) {
3539     C->ops->solve = MatSolve_SeqAIJ_NaturalOrdering;
3540   } else {
3541     C->ops->solve = MatSolve_SeqAIJ;
3542   }
3543   C->ops->solveadd          = 0;
3544   C->ops->solvetranspose    = 0;
3545   C->ops->solvetransposeadd = 0;
3546   C->ops->matsolve          = 0;
3547   C->assembled              = PETSC_TRUE;
3548   C->preallocated           = PETSC_TRUE;
3549 
3550   ierr = PetscLogFlops(C->cmap->n);CHKERRQ(ierr);
3551   PetscFunctionReturn(0);
3552 }
3553