hypre/struct_ls/pfmg_setup.c
2011-03-01 17:10:23 +00:00

913 lines
27 KiB
C

/*BHEADER**********************************************************************
* Copyright (c) 2008, Lawrence Livermore National Security, LLC.
* Produced at the Lawrence Livermore National Laboratory.
* This file is part of HYPRE. See file COPYRIGHT for details.
*
* HYPRE is free software; you can redistribute it and/or modify it under the
* terms of the GNU Lesser General Public License (as published by the Free
* Software Foundation) version 2.1 dated February 1999.
*
* $Revision$
***********************************************************************EHEADER*/
#include "headers.h"
#include "pfmg.h"
#define DEBUG 0
#define hypre_PFMGSetCIndex(cdir, cindex) \
{ \
hypre_SetIndex(cindex, 0, 0, 0); \
hypre_IndexD(cindex, cdir) = 0; \
}
#define hypre_PFMGSetFIndex(cdir, findex) \
{ \
hypre_SetIndex(findex, 0, 0, 0); \
hypre_IndexD(findex, cdir) = 1; \
}
#define hypre_PFMGSetStride(cdir, stride) \
{ \
hypre_SetIndex(stride, 1, 1, 1); \
hypre_IndexD(stride, cdir) = 2; \
}
/*--------------------------------------------------------------------------
* hypre_PFMGSetup
*--------------------------------------------------------------------------*/
HYPRE_Int
hypre_PFMGSetup( void *pfmg_vdata,
hypre_StructMatrix *A,
hypre_StructVector *b,
hypre_StructVector *x )
{
hypre_PFMGData *pfmg_data = pfmg_vdata;
MPI_Comm comm = (pfmg_data -> comm);
HYPRE_Int relax_type = (pfmg_data -> relax_type);
HYPRE_Int usr_jacobi_weight= (pfmg_data -> usr_jacobi_weight);
double jacobi_weight = (pfmg_data -> jacobi_weight);
HYPRE_Int skip_relax = (pfmg_data -> skip_relax);
double *dxyz = (pfmg_data -> dxyz);
HYPRE_Int rap_type;
HYPRE_Int max_iter;
HYPRE_Int max_levels;
HYPRE_Int num_levels;
hypre_Index cindex;
hypre_Index findex;
hypre_Index stride;
hypre_Index coarsen;
HYPRE_Int *cdir_l;
HYPRE_Int *active_l;
hypre_StructGrid **grid_l;
hypre_StructGrid **P_grid_l;
double *data;
HYPRE_Int data_size = 0;
double *relax_weights;
double *mean, *deviation;
double alpha, beta;
hypre_StructMatrix **A_l;
hypre_StructMatrix **P_l;
hypre_StructMatrix **RT_l;
hypre_StructVector **b_l;
hypre_StructVector **x_l;
/* temp vectors */
hypre_StructVector **tx_l;
hypre_StructVector **r_l;
hypre_StructVector **e_l;
void **relax_data_l;
void **matvec_data_l;
void **restrict_data_l;
void **interp_data_l;
hypre_StructGrid *grid;
HYPRE_Int dim;
hypre_Box *cbox;
double min_dxyz;
HYPRE_Int cdir, periodic, cmaxsize;
HYPRE_Int d, l;
HYPRE_Int dxyz_flag;
HYPRE_Int b_num_ghost[] = {0, 0, 0, 0, 0, 0};
HYPRE_Int x_num_ghost[] = {1, 1, 1, 1, 1, 1};
HYPRE_Int ierr = 0;
#if DEBUG
char filename[255];
#endif
/*-----------------------------------------------------
* Set up coarse grids
*-----------------------------------------------------*/
grid = hypre_StructMatrixGrid(A);
dim = hypre_StructGridDim(grid);
/* Compute a new max_levels value based on the grid */
cbox = hypre_BoxDuplicate(hypre_StructGridBoundingBox(grid));
max_levels =
hypre_Log2(hypre_BoxSizeD(cbox, 0)) + 2 +
hypre_Log2(hypre_BoxSizeD(cbox, 1)) + 2 +
hypre_Log2(hypre_BoxSizeD(cbox, 2)) + 2;
if ((pfmg_data -> max_levels) > 0)
{
max_levels = hypre_min(max_levels, (pfmg_data -> max_levels));
}
(pfmg_data -> max_levels) = max_levels;
/* compute dxyz */
if ((dxyz[0] == 0) || (dxyz[1] == 0) || (dxyz[2] == 0))
{
mean = hypre_CTAlloc(double, 3);
deviation = hypre_CTAlloc(double, 3);
hypre_PFMGComputeDxyz(A, dxyz, mean, deviation);
dxyz_flag= 0;
for (d = 0; d < dim; d++)
{
deviation[d] -= mean[d]*mean[d];
/* square of coeff. of variation */
if (deviation[d]/(mean[d]*mean[d]) > .1)
{
dxyz_flag= 1;
break;
}
}
hypre_TFree(mean);
hypre_TFree(deviation);
}
grid_l = hypre_TAlloc(hypre_StructGrid *, max_levels);
hypre_StructGridRef(grid, &grid_l[0]);
P_grid_l = hypre_TAlloc(hypre_StructGrid *, max_levels);
P_grid_l[0] = NULL;
cdir_l = hypre_TAlloc(HYPRE_Int, max_levels);
active_l = hypre_TAlloc(HYPRE_Int, max_levels);
relax_weights = hypre_CTAlloc(double, max_levels);
hypre_SetIndex(coarsen, 1, 1, 1); /* forces relaxation on finest grid */
for (l = 0; ; l++)
{
/* determine cdir */
min_dxyz = dxyz[0] + dxyz[1] + dxyz[2] + 1;
cdir = -1;
alpha = 0.0;
for (d = 0; d < dim; d++)
{
if ((hypre_BoxIMaxD(cbox, d) > hypre_BoxIMinD(cbox, d)) &&
(dxyz[d] < min_dxyz))
{
min_dxyz = dxyz[d];
cdir = d;
}
alpha += 1.0/(dxyz[d]*dxyz[d]);
}
relax_weights[l] = 1.0;
/* If it's possible to coarsen, change relax_weights */
beta = 0.0;
if (cdir != -1)
{
if (dxyz_flag)
{
relax_weights[l] = 2.0/3.0;
}
else
{
for (d = 0; d < dim; d++)
{
if (d != cdir)
{
beta += 1.0/(dxyz[d]*dxyz[d]);
}
}
if (beta == alpha)
{
alpha = 0.0;
}
else
{
alpha = beta/alpha;
}
/* determine level Jacobi weights */
if (dim > 1)
{
relax_weights[l] = 2.0/(3.0 - alpha);
}
else
{
relax_weights[l] = 2.0/3.0; /* always 2/3 for 1-d */
}
}
}
if (cdir != -1)
{
/* don't coarsen if a periodic direction and not divisible by 2 */
periodic = hypre_IndexD(hypre_StructGridPeriodic(grid_l[l]), cdir);
if ((periodic) && (periodic % 2))
{
cdir = -1;
}
/* don't coarsen if we've reached max_levels */
if (l == (max_levels - 1))
{
cdir = -1;
}
}
/* stop coarsening */
if (cdir == -1)
{
active_l[l] = 1; /* forces relaxation on coarsest grid */
cmaxsize = 0;
for (d = 0; d < dim; d++)
{
cmaxsize = hypre_max(cmaxsize, hypre_BoxSizeD(cbox, d));
}
break;
}
cdir_l[l] = cdir;
if (hypre_IndexD(coarsen, cdir) != 0)
{
/* coarsened previously in this direction, relax level l */
active_l[l] = 1;
hypre_SetIndex(coarsen, 0, 0, 0);
hypre_IndexD(coarsen, cdir) = 1;
}
else
{
active_l[l] = 0;
hypre_IndexD(coarsen, cdir) = 1;
}
/* set cindex, findex, and stride */
hypre_PFMGSetCIndex(cdir, cindex);
hypre_PFMGSetFIndex(cdir, findex);
hypre_PFMGSetStride(cdir, stride);
/* update dxyz and coarsen cbox*/
dxyz[cdir] *= 2;
hypre_ProjectBox(cbox, cindex, stride);
hypre_StructMapFineToCoarse(hypre_BoxIMin(cbox), cindex, stride,
hypre_BoxIMin(cbox));
hypre_StructMapFineToCoarse(hypre_BoxIMax(cbox), cindex, stride,
hypre_BoxIMax(cbox));
/* build the interpolation grid */
hypre_StructCoarsen(grid_l[l], findex, stride, 0, &P_grid_l[l+1]);
/* build the coarse grid */
hypre_StructCoarsen(grid_l[l], cindex, stride, 1, &grid_l[l+1]);
}
num_levels = l + 1;
/* free up some things */
hypre_BoxDestroy(cbox);
/* set all levels active if skip_relax = 0 */
if (!skip_relax)
{
for (l = 0; l < num_levels; l++)
{
active_l[l] = 1;
}
}
(pfmg_data -> num_levels) = num_levels;
(pfmg_data -> cdir_l) = cdir_l;
(pfmg_data -> grid_l) = grid_l;
(pfmg_data -> P_grid_l) = P_grid_l;
/*-----------------------------------------------------
* Set up matrix and vector structures
*-----------------------------------------------------*/
/*-----------------------------------------------------
* Modify the rap_type if red-black Gauss-Seidel is
* used. Red-black gs is used only in the non-Galerkin
* case.
*-----------------------------------------------------*/
if (relax_type == 2 || relax_type == 3) /* red-black gs */
{
(pfmg_data -> rap_type)= 1;
}
rap_type = (pfmg_data -> rap_type);
A_l = hypre_TAlloc(hypre_StructMatrix *, num_levels);
P_l = hypre_TAlloc(hypre_StructMatrix *, num_levels - 1);
RT_l = hypre_TAlloc(hypre_StructMatrix *, num_levels - 1);
b_l = hypre_TAlloc(hypre_StructVector *, num_levels);
x_l = hypre_TAlloc(hypre_StructVector *, num_levels);
tx_l = hypre_TAlloc(hypre_StructVector *, num_levels);
r_l = tx_l;
e_l = tx_l;
A_l[0] = hypre_StructMatrixRef(A);
b_l[0] = hypre_StructVectorRef(b);
x_l[0] = hypre_StructVectorRef(x);
tx_l[0] = hypre_StructVectorCreate(comm, grid_l[0]);
hypre_StructVectorSetNumGhost(tx_l[0], x_num_ghost);
hypre_StructVectorInitializeShell(tx_l[0]);
data_size += hypre_StructVectorDataSize(tx_l[0]);
for (l = 0; l < (num_levels - 1); l++)
{
cdir = cdir_l[l];
P_l[l] = hypre_PFMGCreateInterpOp(A_l[l], P_grid_l[l+1], cdir, rap_type);
hypre_StructMatrixInitializeShell(P_l[l]);
data_size += hypre_StructMatrixDataSize(P_l[l]);
if (hypre_StructMatrixSymmetric(A))
{
RT_l[l] = P_l[l];
}
else
{
RT_l[l] = P_l[l];
#if 0
/* Allow RT != P for non symmetric case */
/* NOTE: Need to create a non-pruned grid for this to work */
RT_l[l] = hypre_PFMGCreateRestrictOp(A_l[l], grid_l[l+1], cdir);
hypre_StructMatrixInitializeShell(RT_l[l]);
data_size += hypre_StructMatrixDataSize(RT_l[l]);
#endif
}
A_l[l+1] = hypre_PFMGCreateRAPOp(RT_l[l], A_l[l], P_l[l],
grid_l[l+1], cdir, rap_type);
hypre_StructMatrixInitializeShell(A_l[l+1]);
data_size += hypre_StructMatrixDataSize(A_l[l+1]);
b_l[l+1] = hypre_StructVectorCreate(comm, grid_l[l+1]);
hypre_StructVectorSetNumGhost(b_l[l+1], b_num_ghost);
hypre_StructVectorInitializeShell(b_l[l+1]);
data_size += hypre_StructVectorDataSize(b_l[l+1]);
x_l[l+1] = hypre_StructVectorCreate(comm, grid_l[l+1]);
hypre_StructVectorSetNumGhost(x_l[l+1], x_num_ghost);
hypre_StructVectorInitializeShell(x_l[l+1]);
data_size += hypre_StructVectorDataSize(x_l[l+1]);
tx_l[l+1] = hypre_StructVectorCreate(comm, grid_l[l+1]);
hypre_StructVectorSetNumGhost(tx_l[l+1], x_num_ghost);
hypre_StructVectorInitializeShell(tx_l[l+1]);
}
data = hypre_SharedCTAlloc(double, data_size);
(pfmg_data -> data) = data;
hypre_StructVectorInitializeData(tx_l[0], data);
hypre_StructVectorAssemble(tx_l[0]);
data += hypre_StructVectorDataSize(tx_l[0]);
for (l = 0; l < (num_levels - 1); l++)
{
hypre_StructMatrixInitializeData(P_l[l], data);
data += hypre_StructMatrixDataSize(P_l[l]);
#if 0
/* Allow R != PT for non symmetric case */
if (!hypre_StructMatrixSymmetric(A))
{
hypre_StructMatrixInitializeData(RT_l[l], data);
data += hypre_StructMatrixDataSize(RT_l[l]);
}
#endif
hypre_StructMatrixInitializeData(A_l[l+1], data);
data += hypre_StructMatrixDataSize(A_l[l+1]);
hypre_StructVectorInitializeData(b_l[l+1], data);
hypre_StructVectorAssemble(b_l[l+1]);
data += hypre_StructVectorDataSize(b_l[l+1]);
hypre_StructVectorInitializeData(x_l[l+1], data);
hypre_StructVectorAssemble(x_l[l+1]);
data += hypre_StructVectorDataSize(x_l[l+1]);
hypre_StructVectorInitializeData(tx_l[l+1],
hypre_StructVectorData(tx_l[0]));
hypre_StructVectorAssemble(tx_l[l+1]);
}
(pfmg_data -> A_l) = A_l;
(pfmg_data -> P_l) = P_l;
(pfmg_data -> RT_l) = RT_l;
(pfmg_data -> b_l) = b_l;
(pfmg_data -> x_l) = x_l;
(pfmg_data -> tx_l) = tx_l;
(pfmg_data -> r_l) = r_l;
(pfmg_data -> e_l) = e_l;
/*-----------------------------------------------------
* Set up multigrid operators and call setup routines
*-----------------------------------------------------*/
relax_data_l = hypre_TAlloc(void *, num_levels);
matvec_data_l = hypre_TAlloc(void *, num_levels);
restrict_data_l = hypre_TAlloc(void *, num_levels);
interp_data_l = hypre_TAlloc(void *, num_levels);
for (l = 0; l < (num_levels - 1); l++)
{
cdir = cdir_l[l];
hypre_PFMGSetCIndex(cdir, cindex);
hypre_PFMGSetFIndex(cdir, findex);
hypre_PFMGSetStride(cdir, stride);
/* set up interpolation operator */
hypre_PFMGSetupInterpOp(A_l[l], cdir, findex, stride, P_l[l], rap_type);
/* set up the restriction operator */
#if 0
/* Allow R != PT for non symmetric case */
if (!hypre_StructMatrixSymmetric(A))
hypre_PFMGSetupRestrictOp(A_l[l], tx_l[l],
cdir, cindex, stride, RT_l[l]);
#endif
/* set up the coarse grid operator */
hypre_PFMGSetupRAPOp(RT_l[l], A_l[l], P_l[l],
cdir, cindex, stride, rap_type, A_l[l+1]);
/* set up the interpolation routine */
interp_data_l[l] = hypre_SemiInterpCreate();
hypre_SemiInterpSetup(interp_data_l[l], P_l[l], 0, x_l[l+1], e_l[l],
cindex, findex, stride);
/* set up the restriction routine */
restrict_data_l[l] = hypre_SemiRestrictCreate();
hypre_SemiRestrictSetup(restrict_data_l[l], RT_l[l], 1, r_l[l], b_l[l+1],
cindex, findex, stride);
}
/*-----------------------------------------------------
* Check for zero diagonal on coarsest grid, occurs with
* singular problems like full Neumann or full periodic.
* Note that a processor with zero diagonal will set
* active_l =0, other processors will not. This is OK
* as we only want to avoid the division by zero on the
* one processor which owns the single coarse grid
* point.
*-----------------------------------------------------*/
if ( hypre_ZeroDiagonal(A_l[l]))
{
active_l[l] = 0;
}
/* set up fine grid relaxation */
relax_data_l[0] = hypre_PFMGRelaxCreate(comm);
hypre_PFMGRelaxSetTol(relax_data_l[0], 0.0);
if (usr_jacobi_weight)
{
hypre_PFMGRelaxSetJacobiWeight(relax_data_l[0], jacobi_weight);
}
else
{
hypre_PFMGRelaxSetJacobiWeight(relax_data_l[0], relax_weights[0]);
}
hypre_PFMGRelaxSetType(relax_data_l[0], relax_type);
hypre_PFMGRelaxSetTempVec(relax_data_l[0], tx_l[0]);
hypre_PFMGRelaxSetup(relax_data_l[0], A_l[0], b_l[0], x_l[0]);
if (num_levels > 1)
{
for (l = 1; l < num_levels; l++)
{
/* set relaxation parameters */
if (active_l[l])
{
relax_data_l[l] = hypre_PFMGRelaxCreate(comm);
hypre_PFMGRelaxSetTol(relax_data_l[l], 0.0);
if (usr_jacobi_weight)
{
hypre_PFMGRelaxSetJacobiWeight(relax_data_l[l], jacobi_weight);
}
else
{
hypre_PFMGRelaxSetJacobiWeight(relax_data_l[l], relax_weights[l]);
}
hypre_PFMGRelaxSetType(relax_data_l[l], relax_type);
hypre_PFMGRelaxSetTempVec(relax_data_l[l], tx_l[l]);
}
}
/* change coarsest grid relaxation parameters */
l = num_levels - 1;
if (active_l[l])
{
HYPRE_Int maxwork, maxiter;
hypre_PFMGRelaxSetType(relax_data_l[l], 0);
/* do no more work on the coarsest grid than the cost of a V-cycle
* (estimating roughly 4 communications per V-cycle level) */
maxwork = 4*num_levels;
/* do sweeps proportional to the coarsest grid size */
maxiter = hypre_min(maxwork, cmaxsize);
#if 0
hypre_printf("maxwork = %d, cmaxsize = %d, maxiter = %d\n",
maxwork, cmaxsize, maxiter);
#endif
hypre_PFMGRelaxSetMaxIter(relax_data_l[l], maxiter);
}
/* call relax setup */
for (l = 1; l < num_levels; l++)
{
if (active_l[l])
{
hypre_PFMGRelaxSetup(relax_data_l[l], A_l[l], b_l[l], x_l[l]);
}
}
}
hypre_TFree(relax_weights);
for (l = 0; l < num_levels; l++)
{
/* set up the residual routine */
matvec_data_l[l] = hypre_StructMatvecCreate();
hypre_StructMatvecSetup(matvec_data_l[l], A_l[l], x_l[l]);
}
(pfmg_data -> active_l) = active_l;
(pfmg_data -> relax_data_l) = relax_data_l;
(pfmg_data -> matvec_data_l) = matvec_data_l;
(pfmg_data -> restrict_data_l) = restrict_data_l;
(pfmg_data -> interp_data_l) = interp_data_l;
/*-----------------------------------------------------
* Allocate space for log info
*-----------------------------------------------------*/
if ((pfmg_data -> logging) > 0)
{
max_iter = (pfmg_data -> max_iter);
(pfmg_data -> norms) = hypre_TAlloc(double, max_iter);
(pfmg_data -> rel_norms) = hypre_TAlloc(double, max_iter);
}
#if DEBUG
for (l = 0; l < (num_levels - 1); l++)
{
hypre_sprintf(filename, "zout_A.%02d", l);
hypre_StructMatrixPrint(filename, A_l[l], 0);
hypre_sprintf(filename, "zout_P.%02d", l);
hypre_StructMatrixPrint(filename, P_l[l], 0);
}
hypre_sprintf(filename, "zout_A.%02d", l);
hypre_StructMatrixPrint(filename, A_l[l], 0);
#endif
return ierr;
}
/*--------------------------------------------------------------------------
* hypre_PFMGComputeDxyz
*--------------------------------------------------------------------------*/
HYPRE_Int
hypre_PFMGComputeDxyz( hypre_StructMatrix *A,
double *dxyz,
double *mean,
double *deviation)
{
hypre_BoxArray *compute_boxes;
hypre_Box *compute_box;
hypre_Box *A_dbox;
HYPRE_Int Ai;
double *Ap;
double cxyz[3], sqcxyz[3], tcxyz[3];
double cxyz_max;
HYPRE_Int tot_size;
hypre_StructStencil *stencil;
hypre_Index *stencil_shape;
HYPRE_Int stencil_size;
HYPRE_Int constant_coefficient;
HYPRE_Int Astenc;
hypre_Index loop_size;
hypre_IndexRef start;
hypre_Index stride;
HYPRE_Int i, si, d;
HYPRE_Int loopi, loopj, loopk;
HYPRE_Int ierr = 0;
double cx, cy, cz, sqcx, sqcy, sqcz, tcx, tcy, tcz;
/*----------------------------------------------------------
* Initialize some things
*----------------------------------------------------------*/
stencil = hypre_StructMatrixStencil(A);
stencil_shape = hypre_StructStencilShape(stencil);
stencil_size = hypre_StructStencilSize(stencil);
hypre_SetIndex(stride, 1, 1, 1);
/*----------------------------------------------------------
* Compute cxyz (use arithmetic mean)
*----------------------------------------------------------*/
cx = 0.0;
cy = 0.0;
cz = 0.0;
sqcx = 0.0;
sqcy = 0.0;
sqcz = 0.0;
constant_coefficient = hypre_StructMatrixConstantCoefficient(A);
compute_boxes = hypre_StructGridBoxes(hypre_StructMatrixGrid(A));
tot_size= hypre_StructGridGlobalSize(hypre_StructMatrixGrid(A));
hypre_ForBoxI(i, compute_boxes)
{
compute_box = hypre_BoxArrayBox(compute_boxes, i);
A_dbox = hypre_BoxArrayBox(hypre_StructMatrixDataSpace(A), i);
start = hypre_BoxIMin(compute_box);
hypre_BoxGetStrideSize(compute_box, stride, loop_size);
/* all coefficients constant or variable diagonal */
if ( constant_coefficient )
{
Ai = hypre_CCBoxIndexRank( A_dbox, start );
tcx = 0.0;
tcy = 0.0;
tcz = 0.0;
for (si = 0; si < stencil_size; si++)
{
Ap = hypre_StructMatrixBoxData(A, i, si);
/* x-direction */
Astenc = hypre_IndexD(stencil_shape[si], 0);
if (Astenc)
{
tcx -= Ap[Ai];
}
/* y-direction */
Astenc = hypre_IndexD(stencil_shape[si], 1);
if (Astenc)
{
tcy -= Ap[Ai];
}
/* z-direction */
Astenc = hypre_IndexD(stencil_shape[si], 2);
if (Astenc)
{
tcz -= Ap[Ai];
}
}
cx += tcx;
cy += tcy;
cz += tcz;
sqcx += (tcx*tcx);
sqcy += (tcy*tcy);
sqcz += (tcz*tcz);
}
/* constant_coefficient==0, all coefficients vary with space */
else
{
hypre_BoxLoop1Begin(loop_size, A_dbox, start, stride, Ai);
#define HYPRE_BOX_SMP_PRIVATE loopk,loopi,loopj,Ai,si,Astenc,tcx,tcy,tcz
#define HYPRE_SMP_REDUCTION_OP +
#define HYPRE_SMP_REDUCTION_VARS cx,cy,cz,sqcx,sqcy,sqcz
#include "hypre_box_smp_forloop.h"
hypre_BoxLoop1For(loopi, loopj, loopk, Ai)
{
tcx = 0.0;
tcy = 0.0;
tcz = 0.0;
for (si = 0; si < stencil_size; si++)
{
Ap = hypre_StructMatrixBoxData(A, i, si);
/* x-direction */
Astenc = hypre_IndexD(stencil_shape[si], 0);
if (Astenc)
{
tcx -= Ap[Ai];
}
/* y-direction */
Astenc = hypre_IndexD(stencil_shape[si], 1);
if (Astenc)
{
tcy -= Ap[Ai];
}
/* z-direction */
Astenc = hypre_IndexD(stencil_shape[si], 2);
if (Astenc)
{
tcz -= Ap[Ai];
}
}
cx += tcx;
cy += tcy;
cz += tcz;
sqcx += (tcx*tcx);
sqcy += (tcy*tcy);
sqcz += (tcz*tcz);
}
hypre_BoxLoop1End(Ai);
}
}
cxyz[0] = cx;
cxyz[1] = cy;
cxyz[2] = cz;
sqcxyz[0] = sqcx;
sqcxyz[1] = sqcy;
sqcxyz[2] = sqcz;
/*----------------------------------------------------------
* Compute dxyz
*----------------------------------------------------------*/
/* all coefficients constant or variable diagonal */
if ( constant_coefficient )
{
for (d= 0; d< 3; d++)
{
mean[d]= cxyz[d];
deviation[d]= sqcxyz[d];
}
}
/* constant_coefficient==0, all coefficients vary with space */
else
{
tcxyz[0] = cxyz[0];
tcxyz[1] = cxyz[1];
tcxyz[2] = cxyz[2];
hypre_MPI_Allreduce(tcxyz, cxyz, 3, hypre_MPI_DOUBLE, hypre_MPI_SUM,
hypre_StructMatrixComm(A));
tcxyz[0] = sqcxyz[0];
tcxyz[1] = sqcxyz[1];
tcxyz[2] = sqcxyz[2];
hypre_MPI_Allreduce(tcxyz, sqcxyz, 3, hypre_MPI_DOUBLE, hypre_MPI_SUM,
hypre_StructMatrixComm(A));
for (d= 0; d< 3; d++)
{
mean[d]= cxyz[d]/tot_size;
deviation[d]= sqcxyz[d]/tot_size;
}
}
cxyz_max = 0.0;
for (d = 0; d < 3; d++)
{
cxyz_max = hypre_max(cxyz_max, cxyz[d]);
}
if (cxyz_max == 0.0)
{
cxyz_max = 1.0;
}
for (d = 0; d < 3; d++)
{
if (cxyz[d] > 0)
{
cxyz[d] /= cxyz_max;
dxyz[d] = sqrt(1.0 / cxyz[d]);
}
else
{
dxyz[d] = 1.0e+123;
}
}
/*-----------------------------------------------------------------------
* Return
*-----------------------------------------------------------------------*/
return ierr;
}
/*--------------------------------------------------------------------------
* hypre_ZeroDiagonal
*
* Returns 1 if there is a diagonal coefficient that is zero,
* otherwise returns 0.
*--------------------------------------------------------------------------*/
HYPRE_Int
hypre_ZeroDiagonal( hypre_StructMatrix *A )
{
hypre_BoxArray *compute_boxes;
hypre_Box *compute_box;
hypre_Index loop_size;
hypre_IndexRef start;
hypre_Index stride;
double *Ap;
hypre_Box *A_dbox;
HYPRE_Int Ai;
HYPRE_Int i;
HYPRE_Int loopi, loopj, loopk;
hypre_Index diag_index;
double diag_product = 1.0;
HYPRE_Int zero_diag = 0;
HYPRE_Int constant_coefficient;
/*----------------------------------------------------------
* Initialize some things
*----------------------------------------------------------*/
hypre_SetIndex(stride, 1, 1, 1);
hypre_SetIndex(diag_index, 0, 0, 0);
/* Need to modify here */
constant_coefficient = hypre_StructMatrixConstantCoefficient(A);
compute_boxes = hypre_StructGridBoxes(hypre_StructMatrixGrid(A));
hypre_ForBoxI(i, compute_boxes)
{
compute_box = hypre_BoxArrayBox(compute_boxes, i);
start = hypre_BoxIMin(compute_box);
A_dbox = hypre_BoxArrayBox(hypre_StructMatrixDataSpace(A), i);
Ap = hypre_StructMatrixExtractPointerByIndex(A, i, diag_index);
hypre_BoxGetStrideSize(compute_box, stride, loop_size);
if ( constant_coefficient )
{
Ai = hypre_CCBoxIndexRank( A_dbox, start );
diag_product *= Ap[Ai];
}
else
{
hypre_BoxLoop1Begin(loop_size, A_dbox, start, stride, Ai);
#define HYPRE_BOX_SMP_PRIVATE loopk,loopi,loopj,Ai
#define HYPRE_SMP_REDUCTION_OP *
#define HYPRE_SMP_REDUCTION_VARS diag_product
#include "hypre_box_smp_forloop.h"
hypre_BoxLoop1For(loopi, loopj, loopk, Ai)
{
diag_product *= Ap[Ai];
}
hypre_BoxLoop1End(Ai);
}
}
if (diag_product == 0)
{
zero_diag = 1;
}
return zero_diag;
}