xref: /petsc/src/mat/impls/baij/seq/baijfact2.c (revision 785e854f82a3c614b452fca2cf5ad4f2afe8bdde)

/*
    Factorization code for BAIJ format.
*/

#include <../src/mat/impls/baij/seq/baij.h>
#include <petsc-private/kernels/blockinvert.h>
#include <petscbt.h>
#include <../src/mat/utils/freespace.h>

/* ----------------------------------------------------------------*/
extern PetscErrorCode MatDuplicateNoCreate_SeqBAIJ(Mat,Mat,MatDuplicateOption,PetscBool);

#undef __FUNCT__
#define __FUNCT__ "MatLUFactorNumeric_SeqBAIJ_15_NaturalOrdering"
/*
   This is not much faster than MatLUFactorNumeric_SeqBAIJ_N() but the solve is faster at least sometimes
*/
PetscErrorCode MatLUFactorNumeric_SeqBAIJ_15_NaturalOrdering(Mat B,Mat A,const MatFactorInfo *info)
{
  Mat             C =B;
  Mat_SeqBAIJ     *a=(Mat_SeqBAIJ*)A->data,*b=(Mat_SeqBAIJ*)C->data;
  PetscErrorCode  ierr;
  PetscInt        i,j,k,ipvt[15];
  const PetscInt  n=a->mbs,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j,*ajtmp,*bjtmp,*bdiag=b->diag,*pj;
  PetscInt        nz,nzL,row;
  MatScalar       *rtmp,*pc,*mwork,*pv,*vv,work[225];
  const MatScalar *v,*aa=a->a;
  PetscInt        bs2 = a->bs2,bs=A->rmap->bs,flg;
  PetscInt        sol_ver;

  PetscFunctionBegin;
  ierr = PetscOptionsGetInt(((PetscObject)A)->prefix,"-sol_ver",&sol_ver,NULL);CHKERRQ(ierr);

  /* generate work space needed by the factorization */
  ierr = PetscMalloc2(bs2*n,&rtmp,bs2,&mwork);CHKERRQ(ierr);
  ierr = PetscMemzero(rtmp,bs2*n*sizeof(MatScalar));CHKERRQ(ierr);

  for (i=0; i<n; i++) {
    /* zero rtmp */
    /* L part */
    nz    = bi[i+1] - bi[i];
    bjtmp = bj + bi[i];
    for  (j=0; j<nz; j++) {
      ierr = PetscMemzero(rtmp+bs2*bjtmp[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* U part */
    nz    = bdiag[i] - bdiag[i+1];
    bjtmp = bj + bdiag[i+1]+1;
    for  (j=0; j<nz; j++) {
      ierr = PetscMemzero(rtmp+bs2*bjtmp[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* load in initial (unfactored row) */
    nz    = ai[i+1] - ai[i];
    ajtmp = aj + ai[i];
    v     = aa + bs2*ai[i];
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(rtmp+bs2*ajtmp[j],v+bs2*j,bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* elimination */
    bjtmp = bj + bi[i];
    nzL   = bi[i+1] - bi[i];
    for (k=0; k < nzL; k++) {
      row = bjtmp[k];
      pc  = rtmp + bs2*row;
      for (flg=0,j=0; j<bs2; j++) {
        if (pc[j]!=0.0) {
          flg = 1;
          break;
        }
      }
      if (flg) {
        pv = b->a + bs2*bdiag[row];
        PetscKernel_A_gets_A_times_B(bs,pc,pv,mwork);
        /*ierr = PetscKernel_A_gets_A_times_B_15(pc,pv,mwork);CHKERRQ(ierr);*/
        pj = b->j + bdiag[row+1]+1; /* begining of U(row,:) */
        pv = b->a + bs2*(bdiag[row+1]+1);
        nz = bdiag[row] - bdiag[row+1] - 1; /* num of entries inU(row,:), excluding diag */
        for (j=0; j<nz; j++) {
          vv = rtmp + bs2*pj[j];
          PetscKernel_A_gets_A_minus_B_times_C(bs,vv,pc,pv);
          /* ierr = PetscKernel_A_gets_A_minus_B_times_C_15(vv,pc,pv);CHKERRQ(ierr); */
          pv += bs2;
        }
        ierr = PetscLogFlops(2*bs2*bs*(nz+1)-bs2);CHKERRQ(ierr); /* flops = 2*bs^3*nz + 2*bs^3 - bs2) */
      }
    }

    /* finished row so stick it into b->a */
    /* L part */
    pv = b->a + bs2*bi[i];
    pj = b->j + bi[i];
    nz = bi[i+1] - bi[i];
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(pv+bs2*j,rtmp+bs2*pj[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* Mark diagonal and invert diagonal for simplier triangular solves */
    pv   = b->a + bs2*bdiag[i];
    pj   = b->j + bdiag[i];
    ierr = PetscMemcpy(pv,rtmp+bs2*pj[0],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    /* PetscKernel_A_gets_inverse_A(bs,pv,pivots,work); */
    ierr = PetscKernel_A_gets_inverse_A_15(pv,ipvt,work,info->shiftamount);CHKERRQ(ierr);

    /* U part */
    pv = b->a + bs2*(bdiag[i+1]+1);
    pj = b->j + bdiag[i+1]+1;
    nz = bdiag[i] - bdiag[i+1] - 1;
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(pv+bs2*j,rtmp+bs2*pj[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }
  }

  ierr = PetscFree2(rtmp,mwork);CHKERRQ(ierr);

  C->ops->solve          = MatSolve_SeqBAIJ_15_NaturalOrdering_ver1;
  C->ops->solvetranspose = MatSolve_SeqBAIJ_N_NaturalOrdering;
  C->assembled           = PETSC_TRUE;

  ierr = PetscLogFlops(1.333333333333*bs*bs2*b->mbs);CHKERRQ(ierr); /* from inverting diagonal blocks */
  PetscFunctionReturn(0);
}

#undef __FUNCT__
#define __FUNCT__ "MatLUFactorNumeric_SeqBAIJ_N"
PetscErrorCode MatLUFactorNumeric_SeqBAIJ_N(Mat B,Mat A,const MatFactorInfo *info)
{
  Mat            C     =B;
  Mat_SeqBAIJ    *a    =(Mat_SeqBAIJ*)A->data,*b=(Mat_SeqBAIJ*)C->data;
  IS             isrow = b->row,isicol = b->icol;
  PetscErrorCode ierr;
  const PetscInt *r,*ic;
  PetscInt       i,j,k,n=a->mbs,*ai=a->i,*aj=a->j,*bi=b->i,*bj=b->j;
  PetscInt       *ajtmp,*bjtmp,nz,nzL,row,*bdiag=b->diag,*pj;
  MatScalar      *rtmp,*pc,*mwork,*v,*pv,*aa=a->a;
  PetscInt       bs=A->rmap->bs,bs2 = a->bs2,*v_pivots,flg;
  MatScalar      *v_work;
  PetscBool      col_identity,row_identity,both_identity;

  PetscFunctionBegin;
  ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
  ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);

  ierr = PetscMalloc1(bs2*n,&rtmp);CHKERRQ(ierr);
  ierr = PetscMemzero(rtmp,bs2*n*sizeof(MatScalar));CHKERRQ(ierr);

  /* generate work space needed by dense LU factorization */
  ierr = PetscMalloc3(bs,&v_work,bs2,&mwork,bs,&v_pivots);CHKERRQ(ierr);

  for (i=0; i<n; i++) {
    /* zero rtmp */
    /* L part */
    nz    = bi[i+1] - bi[i];
    bjtmp = bj + bi[i];
    for  (j=0; j<nz; j++) {
      ierr = PetscMemzero(rtmp+bs2*bjtmp[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* U part */
    nz    = bdiag[i] - bdiag[i+1];
    bjtmp = bj + bdiag[i+1]+1;
    for  (j=0; j<nz; j++) {
      ierr = PetscMemzero(rtmp+bs2*bjtmp[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* load in initial (unfactored row) */
    nz    = ai[r[i]+1] - ai[r[i]];
    ajtmp = aj + ai[r[i]];
    v     = aa + bs2*ai[r[i]];
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(rtmp+bs2*ic[ajtmp[j]],v+bs2*j,bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* elimination */
    bjtmp = bj + bi[i];
    nzL   = bi[i+1] - bi[i];
    for (k=0; k < nzL; k++) {
      row = bjtmp[k];
      pc  = rtmp + bs2*row;
      for (flg=0,j=0; j<bs2; j++) {
        if (pc[j]!=0.0) {
          flg = 1;
          break;
        }
      }
      if (flg) {
        pv = b->a + bs2*bdiag[row];
        PetscKernel_A_gets_A_times_B(bs,pc,pv,mwork); /* *pc = *pc * (*pv); */
        pj = b->j + bdiag[row+1]+1;         /* begining of U(row,:) */
        pv = b->a + bs2*(bdiag[row+1]+1);
        nz = bdiag[row] - bdiag[row+1] - 1;         /* num of entries inU(row,:), excluding diag */
        for (j=0; j<nz; j++) {
          PetscKernel_A_gets_A_minus_B_times_C(bs,rtmp+bs2*pj[j],pc,pv+bs2*j);
        }
        ierr = PetscLogFlops(2*bs2*bs*(nz+1)-bs2);CHKERRQ(ierr); /* flops = 2*bs^3*nz + 2*bs^3 - bs2) */
      }
    }

    /* finished row so stick it into b->a */
    /* L part */
    pv = b->a + bs2*bi[i];
    pj = b->j + bi[i];
    nz = bi[i+1] - bi[i];
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(pv+bs2*j,rtmp+bs2*pj[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }

    /* Mark diagonal and invert diagonal for simplier triangular solves */
    pv = b->a + bs2*bdiag[i];
    pj = b->j + bdiag[i];
    /* if (*pj != i)SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_SUP,"row %d != *pj %d",i,*pj); */
    ierr = PetscMemcpy(pv,rtmp+bs2*pj[0],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    ierr = PetscKernel_A_gets_inverse_A(bs,pv,v_pivots,v_work);CHKERRQ(ierr);

    /* U part */
    pv = b->a + bs2*(bdiag[i+1]+1);
    pj = b->j + bdiag[i+1]+1;
    nz = bdiag[i] - bdiag[i+1] - 1;
    for (j=0; j<nz; j++) {
      ierr = PetscMemcpy(pv+bs2*j,rtmp+bs2*pj[j],bs2*sizeof(MatScalar));CHKERRQ(ierr);
    }
  }

  ierr = PetscFree(rtmp);CHKERRQ(ierr);
  ierr = PetscFree3(v_work,mwork,v_pivots);CHKERRQ(ierr);
  ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
  ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);

  ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
  ierr = ISIdentity(isicol,&col_identity);CHKERRQ(ierr);

  both_identity = (PetscBool) (row_identity && col_identity);
  if (both_identity) {
    C->ops->solve = MatSolve_SeqBAIJ_N_NaturalOrdering;
  } else {
    C->ops->solve = MatSolve_SeqBAIJ_N;
  }
  C->ops->solvetranspose = MatSolveTranspose_SeqBAIJ_N;

  C->assembled = PETSC_TRUE;

  ierr = PetscLogFlops(1.333333333333*bs*bs2*b->mbs);CHKERRQ(ierr); /* from inverting diagonal blocks */
  PetscFunctionReturn(0);
}

/*
   ilu(0) with natural ordering under new data structure.
   See MatILUFactorSymbolic_SeqAIJ_ilu0() for detailed description
   because this code is almost identical to MatILUFactorSymbolic_SeqAIJ_ilu0_inplace().
*/

#undef __FUNCT__
#define __FUNCT__ "MatILUFactorSymbolic_SeqBAIJ_ilu0"
PetscErrorCode MatILUFactorSymbolic_SeqBAIJ_ilu0(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
{

  Mat_SeqBAIJ    *a = (Mat_SeqBAIJ*)A->data,*b;
  PetscErrorCode ierr;
  PetscInt       n=a->mbs,*ai=a->i,*aj,*adiag=a->diag,bs2 = a->bs2;
  PetscInt       i,j,nz,*bi,*bj,*bdiag,bi_temp;

  PetscFunctionBegin;
  ierr = MatDuplicateNoCreate_SeqBAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_FALSE);CHKERRQ(ierr);
  b    = (Mat_SeqBAIJ*)(fact)->data;

  /* allocate matrix arrays for new data structure */
  ierr = PetscMalloc3(bs2*ai[n]+1,&b->a,ai[n]+1,&b->j,n+1,&b->i);CHKERRQ(ierr);
  ierr = PetscLogObjectMemory((PetscObject)fact,ai[n]*(bs2*sizeof(PetscScalar)+sizeof(PetscInt))+(n+1)*sizeof(PetscInt));CHKERRQ(ierr);

  b->singlemalloc    = PETSC_TRUE;
  b->free_a          = PETSC_TRUE;
  b->free_ij         = PETSC_TRUE;
  fact->preallocated = PETSC_TRUE;
  fact->assembled    = PETSC_TRUE;
  if (!b->diag) {
    ierr = PetscMalloc1((n+1),&b->diag);CHKERRQ(ierr);
    ierr = PetscLogObjectMemory((PetscObject)fact,(n+1)*sizeof(PetscInt));CHKERRQ(ierr);
  }
  bdiag = b->diag;

  if (n > 0) {
    ierr = PetscMemzero(b->a,bs2*ai[n]*sizeof(MatScalar));CHKERRQ(ierr);
  }

  /* set bi and bj with new data structure */
  bi = b->i;
  bj = b->j;

  /* L part */
  bi[0] = 0;
  for (i=0; i<n; i++) {
    nz      = adiag[i] - ai[i];
    bi[i+1] = bi[i] + nz;
    aj      = a->j + ai[i];
    for (j=0; j<nz; j++) {
      *bj = aj[j]; bj++;
    }
  }

  /* U part */
  bi_temp  = bi[n];
  bdiag[n] = bi[n]-1;
  for (i=n-1; i>=0; i--) {
    nz      = ai[i+1] - adiag[i] - 1;
    bi_temp = bi_temp + nz + 1;
    aj      = a->j + adiag[i] + 1;
    for (j=0; j<nz; j++) {
      *bj = aj[j]; bj++;
    }
    /* diag[i] */
    *bj      = i; bj++;
    bdiag[i] = bi_temp - 1;
  }
  PetscFunctionReturn(0);
}

#undef __FUNCT__
#define __FUNCT__ "MatILUFactorSymbolic_SeqBAIJ"
PetscErrorCode MatILUFactorSymbolic_SeqBAIJ(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
{
  Mat_SeqBAIJ        *a = (Mat_SeqBAIJ*)A->data,*b;
  IS                 isicol;
  PetscErrorCode     ierr;
  const PetscInt     *r,*ic;
  PetscInt           n=a->mbs,*ai=a->i,*aj=a->j,d;
  PetscInt           *bi,*cols,nnz,*cols_lvl;
  PetscInt           *bdiag,prow,fm,nzbd,reallocs=0,dcount=0;
  PetscInt           i,levels,diagonal_fill;
  PetscBool          col_identity,row_identity,both_identity;
  PetscReal          f;
  PetscInt           nlnk,*lnk,*lnk_lvl=NULL;
  PetscBT            lnkbt;
  PetscInt           nzi,*bj,**bj_ptr,**bjlvl_ptr;
  PetscFreeSpaceList free_space    =NULL,current_space=NULL;
  PetscFreeSpaceList free_space_lvl=NULL,current_space_lvl=NULL;
  PetscBool          missing;
  PetscInt           bs=A->rmap->bs,bs2=a->bs2;

  PetscFunctionBegin;
  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);
  if (bs>1) {  /* check shifttype */
    if (info->shifttype == MAT_SHIFT_NONZERO || info->shifttype == MAT_SHIFT_POSITIVE_DEFINITE)
      SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only MAT_SHIFT_NONE and MAT_SHIFT_INBLOCKS are supported for BAIJ matrix");
  }

  ierr = MatMissingDiagonal(A,&missing,&d);CHKERRQ(ierr);
  if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry %D",d);

  f             = info->fill;
  levels        = (PetscInt)info->levels;
  diagonal_fill = (PetscInt)info->diagonal_fill;

  ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);

  ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
  ierr = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);

  both_identity = (PetscBool) (row_identity && col_identity);

  if (!levels && both_identity) {
    /* special case: ilu(0) with natural ordering */
    ierr = MatILUFactorSymbolic_SeqBAIJ_ilu0(fact,A,isrow,iscol,info);CHKERRQ(ierr);
    ierr = MatSeqBAIJSetNumericFactorization(fact,both_identity);CHKERRQ(ierr);

    fact->factortype               = MAT_FACTOR_ILU;
    (fact)->info.factor_mallocs    = 0;
    (fact)->info.fill_ratio_given  = info->fill;
    (fact)->info.fill_ratio_needed = 1.0;

    b                = (Mat_SeqBAIJ*)(fact)->data;
    b->row           = isrow;
    b->col           = iscol;
    b->icol          = isicol;
    ierr             = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
    ierr             = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
    b->pivotinblocks = (info->pivotinblocks) ? PETSC_TRUE : PETSC_FALSE;

    ierr = PetscMalloc1((n+1)*bs,&b->solve_work);CHKERRQ(ierr);
    PetscFunctionReturn(0);
  }

  ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
  ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);

  /* get new row pointers */
  ierr  = PetscMalloc1((n+1),&bi);CHKERRQ(ierr);
  bi[0] = 0;
  /* bdiag is location of diagonal in factor */
  ierr     = PetscMalloc1((n+1),&bdiag);CHKERRQ(ierr);
  bdiag[0] = 0;

  ierr = PetscMalloc2(n,&bj_ptr,n,&bjlvl_ptr);CHKERRQ(ierr);

  /* create a linked list for storing column indices of the active row */
  nlnk = n + 1;
  ierr = PetscIncompleteLLCreate(n,n,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);

  /* initial FreeSpace size is f*(ai[n]+1) */
  ierr              = PetscFreeSpaceGet((PetscInt)(f*(ai[n]+1)),&free_space);CHKERRQ(ierr);
  current_space     = free_space;
  ierr              = PetscFreeSpaceGet((PetscInt)(f*(ai[n]+1)),&free_space_lvl);CHKERRQ(ierr);
  current_space_lvl = free_space_lvl;

  for (i=0; i<n; i++) {
    nzi = 0;
    /* copy current row into linked list */
    nnz = ai[r[i]+1] - ai[r[i]];
    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);
    cols   = aj + ai[r[i]];
    lnk[i] = -1; /* marker to indicate if diagonal exists */
    ierr   = PetscIncompleteLLInit(nnz,cols,n,ic,nlnk,lnk,lnk_lvl,lnkbt);CHKERRQ(ierr);
    nzi   += nlnk;

    /* make sure diagonal entry is included */
    if (diagonal_fill && lnk[i] == -1) {
      fm = n;
      while (lnk[fm] < i) fm = lnk[fm];
      lnk[i]     = lnk[fm]; /* insert diagonal into linked list */
      lnk[fm]    = i;
      lnk_lvl[i] = 0;
      nzi++; dcount++;
    }

    /* add pivot rows into the active row */
    nzbd = 0;
    prow = lnk[n];
    while (prow < i) {
      nnz      = bdiag[prow];
      cols     = bj_ptr[prow] + nnz + 1;
      cols_lvl = bjlvl_ptr[prow] + nnz + 1;
      nnz      = bi[prow+1] - bi[prow] - nnz - 1;

      ierr = PetscILULLAddSorted(nnz,cols,levels,cols_lvl,prow,nlnk,lnk,lnk_lvl,lnkbt,prow);CHKERRQ(ierr);
      nzi += nlnk;
      prow = lnk[prow];
      nzbd++;
    }
    bdiag[i] = nzbd;
    bi[i+1]  = bi[i] + nzi;

    /* if free space is not available, make more free space */
    if (current_space->local_remaining<nzi) {
      nnz  = 2*nzi*(n - i); /* estimated and max additional space needed */
      ierr = PetscFreeSpaceGet(nnz,&current_space);CHKERRQ(ierr);
      ierr = PetscFreeSpaceGet(nnz,&current_space_lvl);CHKERRQ(ierr);
      reallocs++;
    }

    /* copy data into free_space and free_space_lvl, then initialize lnk */
    ierr = PetscIncompleteLLClean(n,n,nzi,lnk,lnk_lvl,current_space->array,current_space_lvl->array,lnkbt);CHKERRQ(ierr);

    bj_ptr[i]    = current_space->array;
    bjlvl_ptr[i] = current_space_lvl->array;

    /* make sure the active row i has diagonal entry */
    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);

    current_space->array           += nzi;
    current_space->local_used      += nzi;
    current_space->local_remaining -= nzi;

    current_space_lvl->array           += nzi;
    current_space_lvl->local_used      += nzi;
    current_space_lvl->local_remaining -= nzi;
  }

  ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
  ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);

  /* copy free_space into bj and free free_space; set bi, bj, bdiag in new datastructure; */
  ierr = PetscMalloc1((bi[n]+1),&bj);CHKERRQ(ierr);
  ierr = PetscFreeSpaceContiguous_LU(&free_space,bj,n,bi,bdiag);CHKERRQ(ierr);

  ierr = PetscIncompleteLLDestroy(lnk,lnkbt);CHKERRQ(ierr);
  ierr = PetscFreeSpaceDestroy(free_space_lvl);CHKERRQ(ierr);
  ierr = PetscFree2(bj_ptr,bjlvl_ptr);CHKERRQ(ierr);

#if defined(PETSC_USE_INFO)
  {
    PetscReal af = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);
    ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %G needed %G\n",reallocs,f,af);CHKERRQ(ierr);
    ierr = PetscInfo1(A,"Run with -[sub_]pc_factor_fill %G or use \n",af);CHKERRQ(ierr);
    ierr = PetscInfo1(A,"PCFactorSetFill([sub]pc,%G);\n",af);CHKERRQ(ierr);
    ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
    if (diagonal_fill) {
      ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals",dcount);CHKERRQ(ierr);
    }
  }
#endif

  /* put together the new matrix */
  ierr = MatSeqBAIJSetPreallocation_SeqBAIJ(fact,bs,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
  ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);

  b               = (Mat_SeqBAIJ*)(fact)->data;
  b->free_a       = PETSC_TRUE;
  b->free_ij      = PETSC_TRUE;
  b->singlemalloc = PETSC_FALSE;

  ierr = PetscMalloc1((bs2*(bdiag[0]+1)),&b->a);CHKERRQ(ierr);

  b->j          = bj;
  b->i          = bi;
  b->diag       = bdiag;
  b->free_diag  = PETSC_TRUE;
  b->ilen       = 0;
  b->imax       = 0;
  b->row        = isrow;
  b->col        = iscol;
  ierr          = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
  ierr          = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
  b->icol       = isicol;

  ierr = PetscMalloc1((bs*n+bs),&b->solve_work);CHKERRQ(ierr);
  /* In b structure:  Free imax, ilen, old a, old j.
     Allocate bdiag, solve_work, new a, new j */
  ierr     = PetscLogObjectMemory((PetscObject)fact,(bdiag[0]+1) * (sizeof(PetscInt)+bs2*sizeof(PetscScalar)));CHKERRQ(ierr);
  b->maxnz = b->nz = bdiag[0]+1;

  fact->info.factor_mallocs    = reallocs;
  fact->info.fill_ratio_given  = f;
  fact->info.fill_ratio_needed = ((PetscReal)(bdiag[0]+1))/((PetscReal)ai[n]);

  ierr = MatSeqBAIJSetNumericFactorization(fact,both_identity);CHKERRQ(ierr);
  PetscFunctionReturn(0);
}

/*
     This code is virtually identical to MatILUFactorSymbolic_SeqAIJ
   except that the data structure of Mat_SeqAIJ is slightly different.
   Not a good example of code reuse.
*/
#undef __FUNCT__
#define __FUNCT__ "MatILUFactorSymbolic_SeqBAIJ_inplace"
PetscErrorCode MatILUFactorSymbolic_SeqBAIJ_inplace(Mat fact,Mat A,IS isrow,IS iscol,const MatFactorInfo *info)
{
  Mat_SeqBAIJ    *a = (Mat_SeqBAIJ*)A->data,*b;
  IS             isicol;
  PetscErrorCode ierr;
  const PetscInt *r,*ic,*ai = a->i,*aj = a->j,*xi;
  PetscInt       prow,n = a->mbs,*ainew,*ajnew,jmax,*fill,nz,*im,*ajfill,*flev,*xitmp;
  PetscInt       *dloc,idx,row,m,fm,nzf,nzi,reallocate = 0,dcount = 0;
  PetscInt       incrlev,nnz,i,bs = A->rmap->bs,bs2 = a->bs2,levels,diagonal_fill,dd;
  PetscBool      col_identity,row_identity,both_identity,flg;
  PetscReal      f;

  PetscFunctionBegin;
  ierr = MatMissingDiagonal_SeqBAIJ(A,&flg,&dd);CHKERRQ(ierr);
  if (flg) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix A is missing diagonal entry in row %D",dd);

  f             = info->fill;
  levels        = (PetscInt)info->levels;
  diagonal_fill = (PetscInt)info->diagonal_fill;

  ierr = ISInvertPermutation(iscol,PETSC_DECIDE,&isicol);CHKERRQ(ierr);

  ierr = ISIdentity(isrow,&row_identity);CHKERRQ(ierr);
  ierr = ISIdentity(iscol,&col_identity);CHKERRQ(ierr);
  both_identity = (PetscBool) (row_identity && col_identity);

  if (!levels && both_identity) {  /* special case copy the nonzero structure */
    ierr = MatDuplicateNoCreate_SeqBAIJ(fact,A,MAT_DO_NOT_COPY_VALUES,PETSC_TRUE);CHKERRQ(ierr);
    ierr = MatSeqBAIJSetNumericFactorization_inplace(fact,both_identity);CHKERRQ(ierr);

    fact->factortype = MAT_FACTOR_ILU;
    b                = (Mat_SeqBAIJ*)fact->data;
    b->row           = isrow;
    b->col           = iscol;
    ierr             = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
    ierr             = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
    b->icol          = isicol;
    b->pivotinblocks = (info->pivotinblocks) ? PETSC_TRUE : PETSC_FALSE;

    ierr         = PetscMalloc1((n+1)*bs,&b->solve_work);CHKERRQ(ierr);
    PetscFunctionReturn(0);
  }

  /* general case perform the symbolic factorization */
  ierr = ISGetIndices(isrow,&r);CHKERRQ(ierr);
  ierr = ISGetIndices(isicol,&ic);CHKERRQ(ierr);

  /* get new row pointers */
  ierr     = PetscMalloc1((n+1),&ainew);CHKERRQ(ierr);
  ainew[0] = 0;
  /* don't know how many column pointers are needed so estimate */
  jmax = (PetscInt)(f*ai[n] + 1);
  ierr = PetscMalloc1((jmax),&ajnew);CHKERRQ(ierr);
  /* ajfill is level of fill for each fill entry */
  ierr = PetscMalloc1((jmax),&ajfill);CHKERRQ(ierr);
  /* fill is a linked list of nonzeros in active row */
  ierr = PetscMalloc1((n+1),&fill);CHKERRQ(ierr);
  /* im is level for each filled value */
  ierr = PetscMalloc1((n+1),&im);CHKERRQ(ierr);
  /* dloc is location of diagonal in factor */
  ierr    = PetscMalloc1((n+1),&dloc);CHKERRQ(ierr);
  dloc[0] = 0;
  for (prow=0; prow<n; prow++) {

    /* copy prow into linked list */
    nzf = nz = ai[r[prow]+1] - ai[r[prow]];
    if (!nz) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Empty row in matrix: row in original ordering %D in permuted ordering %D",r[prow],prow);
    xi         = aj + ai[r[prow]];
    fill[n]    = n;
    fill[prow] = -1;   /* marker for diagonal entry */
    while (nz--) {
      fm  = n;
      idx = ic[*xi++];
      do {
        m  = fm;
        fm = fill[m];
      } while (fm < idx);
      fill[m]   = idx;
      fill[idx] = fm;
      im[idx]   = 0;
    }

    /* make sure diagonal entry is included */
    if (diagonal_fill && fill[prow] == -1) {
      fm = n;
      while (fill[fm] < prow) fm = fill[fm];
      fill[prow] = fill[fm];    /* insert diagonal into linked list */
      fill[fm]   = prow;
      im[prow]   = 0;
      nzf++;
      dcount++;
    }

    nzi = 0;
    row = fill[n];
    while (row < prow) {
      incrlev = im[row] + 1;
      nz      = dloc[row];
      xi      = ajnew  + ainew[row] + nz + 1;
      flev    = ajfill + ainew[row] + nz + 1;
      nnz     = ainew[row+1] - ainew[row] - nz - 1;
      fm      = row;
      while (nnz-- > 0) {
        idx = *xi++;
        if (*flev + incrlev > levels) {
          flev++;
          continue;
        }
        do {
          m  = fm;
          fm = fill[m];
        } while (fm < idx);
        if (fm != idx) {
          im[idx]   = *flev + incrlev;
          fill[m]   = idx;
          fill[idx] = fm;
          fm        = idx;
          nzf++;
        } else if (im[idx] > *flev + incrlev) im[idx] = *flev+incrlev;
        flev++;
      }
      row = fill[row];
      nzi++;
    }
    /* copy new filled row into permanent storage */
    ainew[prow+1] = ainew[prow] + nzf;
    if (ainew[prow+1] > jmax) {

      /* estimate how much additional space we will need */
      /* use the strategy suggested by David Hysom <hysom@perch-t.icase.edu> */
      /* just double the memory each time */
      PetscInt maxadd = jmax;
      /* maxadd = (int)(((f*ai[n]+1)*(n-prow+5))/n); */
      if (maxadd < nzf) maxadd = (n-prow)*(nzf+1);
      jmax += maxadd;

      /* allocate a longer ajnew and ajfill */
      ierr   = PetscMalloc1(jmax,&xitmp);CHKERRQ(ierr);
      ierr   = PetscMemcpy(xitmp,ajnew,ainew[prow]*sizeof(PetscInt));CHKERRQ(ierr);
      ierr   = PetscFree(ajnew);CHKERRQ(ierr);
      ajnew  = xitmp;
      ierr   = PetscMalloc1(jmax,&xitmp);CHKERRQ(ierr);
      ierr   = PetscMemcpy(xitmp,ajfill,ainew[prow]*sizeof(PetscInt));CHKERRQ(ierr);
      ierr   = PetscFree(ajfill);CHKERRQ(ierr);
      ajfill = xitmp;
      reallocate++;   /* count how many reallocations are needed */
    }
    xitmp      = ajnew + ainew[prow];
    flev       = ajfill + ainew[prow];
    dloc[prow] = nzi;
    fm         = fill[n];
    while (nzf--) {
      *xitmp++ = fm;
      *flev++  = im[fm];
      fm       = fill[fm];
    }
    /* make sure row has diagonal entry */
    if (ajnew[ainew[prow]+dloc[prow]] != prow) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Row %D has missing diagonal in factored matrix\n\
                                                        try running with -pc_factor_nonzeros_along_diagonal or -pc_factor_diagonal_fill",prow);
  }
  ierr = PetscFree(ajfill);CHKERRQ(ierr);
  ierr = ISRestoreIndices(isrow,&r);CHKERRQ(ierr);
  ierr = ISRestoreIndices(isicol,&ic);CHKERRQ(ierr);
  ierr = PetscFree(fill);CHKERRQ(ierr);
  ierr = PetscFree(im);CHKERRQ(ierr);

#if defined(PETSC_USE_INFO)
  {
    PetscReal af = ((PetscReal)ainew[n])/((PetscReal)ai[n]);
    ierr = PetscInfo3(A,"Reallocs %D Fill ratio:given %G needed %G\n",reallocate,f,af);CHKERRQ(ierr);
    ierr = PetscInfo1(A,"Run with -pc_factor_fill %G or use \n",af);CHKERRQ(ierr);
    ierr = PetscInfo1(A,"PCFactorSetFill(pc,%G);\n",af);CHKERRQ(ierr);
    ierr = PetscInfo(A,"for best performance.\n");CHKERRQ(ierr);
    if (diagonal_fill) {
      ierr = PetscInfo1(A,"Detected and replaced %D missing diagonals\n",dcount);CHKERRQ(ierr);
    }
  }
#endif

  /* put together the new matrix */
  ierr = MatSeqBAIJSetPreallocation_SeqBAIJ(fact,bs,MAT_SKIP_ALLOCATION,NULL);CHKERRQ(ierr);
  ierr = PetscLogObjectParent((PetscObject)fact,(PetscObject)isicol);CHKERRQ(ierr);
  b    = (Mat_SeqBAIJ*)fact->data;

  b->free_a       = PETSC_TRUE;
  b->free_ij      = PETSC_TRUE;
  b->singlemalloc = PETSC_FALSE;

  ierr = PetscMalloc1(bs2*ainew[n],&b->a);CHKERRQ(ierr);

  b->j          = ajnew;
  b->i          = ainew;
  for (i=0; i<n; i++) dloc[i] += ainew[i];
  b->diag          = dloc;
  b->free_diag     = PETSC_TRUE;
  b->ilen          = 0;
  b->imax          = 0;
  b->row           = isrow;
  b->col           = iscol;
  b->pivotinblocks = (info->pivotinblocks) ? PETSC_TRUE : PETSC_FALSE;

  ierr    = PetscObjectReference((PetscObject)isrow);CHKERRQ(ierr);
  ierr    = PetscObjectReference((PetscObject)iscol);CHKERRQ(ierr);
  b->icol = isicol;
  ierr    = PetscMalloc1((bs*n+bs),&b->solve_work);CHKERRQ(ierr);
  /* In b structure:  Free imax, ilen, old a, old j.
     Allocate dloc, solve_work, new a, new j */
  ierr     = PetscLogObjectMemory((PetscObject)fact,(ainew[n]-n)*(sizeof(PetscInt))+bs2*ainew[n]*sizeof(PetscScalar));CHKERRQ(ierr);
  b->maxnz = b->nz = ainew[n];

  fact->info.factor_mallocs    = reallocate;
  fact->info.fill_ratio_given  = f;
  fact->info.fill_ratio_needed = ((PetscReal)ainew[n])/((PetscReal)ai[prow]);

  ierr = MatSeqBAIJSetNumericFactorization_inplace(fact,both_identity);CHKERRQ(ierr);
  PetscFunctionReturn(0);
}

#undef __FUNCT__
#define __FUNCT__ "MatSetUnfactored_SeqBAIJ_4_NaturalOrdering_SSE"
PetscErrorCode MatSetUnfactored_SeqBAIJ_4_NaturalOrdering_SSE(Mat A)
{
  /* Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data; */
  /* int i,*AJ=a->j,nz=a->nz; */

  PetscFunctionBegin;
  /* Undo Column scaling */
  /*    while (nz--) { */
  /*      AJ[i] = AJ[i]/4; */
  /*    } */
  /* This should really invoke a push/pop logic, but we don't have that yet. */
  A->ops->setunfactored = NULL;
  PetscFunctionReturn(0);
}

#undef __FUNCT__
#define __FUNCT__ "MatSetUnfactored_SeqBAIJ_4_NaturalOrdering_SSE_usj"
PetscErrorCode MatSetUnfactored_SeqBAIJ_4_NaturalOrdering_SSE_usj(Mat A)
{
  Mat_SeqBAIJ    *a = (Mat_SeqBAIJ*)A->data;
  PetscInt       *AJ=a->j,nz=a->nz;
  unsigned short *aj=(unsigned short*)AJ;

  PetscFunctionBegin;
  /* Is this really necessary? */
  while (nz--) {
    AJ[nz] = (int)((unsigned int)aj[nz]); /* First extend, then convert to signed. */
  }
  A->ops->setunfactored = NULL;
  PetscFunctionReturn(0);
}