Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

generalize interpolaters to not need or use extra values in a directi… #3415

Merged
merged 11 commits into from
Jul 16, 2023
Merged

generalize interpolaters to not need or use extra values in a directi… #3415

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
65e1558
generalize interpolaters to not need or use extra values in a directi…
asalmgren Jul 10, 2023
dd56177
Merge branch 'development' into fix_interp_for_ratio1
asalmgren Jul 11, 2023
6396ac1
more protection for when ratio == 1 in some direction
asalmgren Jul 12, 2023
578b471
Merge branch 'development' into fix_interp_for_ratio1
asalmgren Jul 12, 2023
7d24eb9
fix 1d interface for interpolater
asalmgren Jul 12, 2023
8981a6c
Merge branch 'fix_interp_for_ratio1' of github.com:asalmgren/amrex in…
asalmgren Jul 12, 2023
9fe3ff0
minor change
WeiqunZhang Jul 16, 2023
feb9687
Merge branch 'development' into fix_interp_for_ratio1
asalmgren Jul 16, 2023
86ff579
remove some restrictions on ratio for interpolation
asalmgren Jul 16, 2023
e182e35
merge with changes
asalmgren Jul 16, 2023
04ad7f3
Fix MFInterpolater
WeiqunZhang Jul 16, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
48 changes: 27 additions & 21 deletions Src/AmrCore/AMReX_Interpolater.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,8 @@ namespace amrex {
* PCInterp, NodeBilinear, FaceLinear, CellConservativeLinear, and
* CellBilinear are supported for all dimensions on cpu and gpu.
*
* CellConservativeProtected only works in 2D and 3D on cpu and gpu.
* CellConservativeProtected only works in 2D and 3D on cpu and gpu
* and assumes that ratio > 1 in all directions
*
* CellQuadratic only works in 2D and 3D on cpu and gpu.
*
Expand Down Expand Up @@ -411,7 +412,11 @@ CellConservativeLinear::CoarseBox (const Box& fine,
const IntVect& ratio)
{
Box crse = amrex::coarsen(fine,ratio);
crse.grow(1);
for (int dim = 0; dim < AMREX_SPACEDIM; dim++) {
if (ratio[dim] > 1) {
crse.grow(dim,1);
}
}
return crse;
}

Expand Down Expand Up @@ -440,7 +445,7 @@ CellConservativeLinear::interp (const FArrayBox& crse,
RunOn runon)
{
BL_PROFILE("CellConservativeLinear::interp()");
BL_ASSERT(bcr.size() >= ncomp);
AMREX_ASSERT(bcr.size() >= ncomp);

AMREX_ASSERT(fine.box().contains(fine_region));

Expand All @@ -454,7 +459,12 @@ CellConservativeLinear::interp (const FArrayBox& crse,
Array4<Real> const& finearr = fine.array();

const Box& crse_region = CoarseBox(fine_region,ratio);
const Box& cslope_bx = amrex::grow(crse_region,-1);
Box cslope_bx(crse_region);
for (int dim = 0; dim < AMREX_SPACEDIM; dim++) {
if (ratio[dim] > 1) {
cslope_bx.grow(dim,-1);
}
}

FArrayBox ccfab(cslope_bx, ncomp*AMREX_SPACEDIM);
Array4<Real> const& tmp = ccfab.array();
Expand All @@ -478,7 +488,7 @@ CellConservativeLinear::interp (const FArrayBox& crse,
AMREX_HOST_DEVICE_PARALLEL_FOR_3D_FLAG(runon, cslope_bx, i, j, k,
{
mf_cell_cons_lin_interp_llslope(i,j,k, tmp, crsearr, crse_comp, ncomp,
cdomain, bcrp);
cdomain, ratio, bcrp);
});
} else {
AMREX_HOST_DEVICE_PARALLEL_FOR_4D_FLAG(runon, cslope_bx, ncomp, i, j, k, n,
Expand All @@ -504,7 +514,7 @@ CellConservativeLinear::interp (const FArrayBox& crse,
AMREX_HOST_DEVICE_PARALLEL_FOR_3D_FLAG(runon, cslope_bx, i, j, k,
{
mf_cell_cons_lin_interp_llslope(i,j,k, tmp, crsearr, crse_comp, ncomp,
cdomain, bcrp);
cdomain, ratio, bcrp);
});
} else {
AMREX_HOST_DEVICE_PARALLEL_FOR_4D_FLAG(runon, cslope_bx, ncomp, i, j, k, n,
Expand All @@ -528,7 +538,7 @@ CellConservativeLinear::interp (const FArrayBox& crse,
AMREX_HOST_DEVICE_PARALLEL_FOR_3D_FLAG(runon, cslope_bx, i, j, k,
{
mf_cell_cons_lin_interp_llslope(i,j,k, tmp, crsearr, crse_comp, ncomp,
cdomain, bcrp);
cdomain, ratio, bcrp);
});
} else {
AMREX_HOST_DEVICE_PARALLEL_FOR_4D_FLAG(runon, cslope_bx, ncomp, i, j, k, n,
Expand Down Expand Up @@ -586,7 +596,7 @@ CellQuadratic::interp (const FArrayBox& crse,
#else

BL_PROFILE("CellQuadratic::interp()");
BL_ASSERT(bcr.size() >= ncomp);
AMREX_ASSERT(bcr.size() >= ncomp);

//
// Make box which is intersection of fine_region and domain of fine.
Expand All @@ -595,7 +605,7 @@ CellQuadratic::interp (const FArrayBox& crse,

// Make Box for slopes.
Box cslope_bx = amrex::coarsen(target_fine_region,ratio);
BL_ASSERT(crse.box().contains(cslope_bx));
AMREX_ASSERT(crse.box().contains(cslope_bx));

// Are we running on GPU?
bool run_on_gpu = (runon == RunOn::Gpu && Gpu::inLaunchRegion());
Expand Down Expand Up @@ -739,6 +749,8 @@ CellConservativeProtected::protect (const FArrayBox& /*crse*/,
Vector<BCRec>& /*bcr*/,
RunOn runon)
{
AMREX_ALWAYS_ASSERT(ratio.allGT(IntVect(1)));

#if (AMREX_SPACEDIM == 1)
amrex::ignore_unused(fine,fine_state,
ncomp,fine_region,ratio,
Expand Down Expand Up @@ -838,13 +850,7 @@ CellConservativeQuartic::interp (const FArrayBox& crse,
RunOn runon)
{
BL_PROFILE("CellConservativeQuartic::interp()");
BL_ASSERT(ratio[0] == 2);
#if (AMREX_SPACEDIM >= 2)
BL_ASSERT(ratio[0] == ratio[1]);
#endif
#if (AMREX_SPACEDIM == 3)
BL_ASSERT(ratio[1] == ratio[2]);
#endif
AMREX_ASSERT(ratio == 2);
amrex::ignore_unused(ratio);

//
Expand All @@ -867,16 +873,16 @@ Box
FaceDivFree::CoarseBox (const Box& fine,
int ratio)
{
Box b = amrex::coarsen(fine,ratio).grow(1);
return b;
Box crse = amrex::coarsen(fine,ratio).grow(1);
return crse;
}

Box
FaceDivFree::CoarseBox (const Box& fine,
const IntVect& ratio)
{
Box b = amrex::coarsen(fine,ratio).grow(1);
return b;
Box crse = amrex::coarsen(fine,ratio).grow(1);
return crse;
}

void
Expand All @@ -894,7 +900,7 @@ FaceDivFree::interp (const FArrayBox& /*crse*/,
int /*actual_state*/,
RunOn /*runon*/)
{
amrex::Abort("FaceDivFree does not work on a single MultiFab. Call 'interp_arr' instead.");
amrex::Abort("FaceDivFree does not work on a single FArrayBox. Call 'interp_arr' instead.");
}

void
Expand Down
2 changes: 1 addition & 1 deletion Src/AmrCore/AMReX_MFInterp_1D_C.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@ void mf_cell_cons_lin_interp_limit_minmax_llslope (int i, int, int, Array4<Real>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
void mf_cell_cons_lin_interp_llslope (int i, int, int, Array4<Real> const& slope,
Array4<Real const> const& u, int scomp, int ncomp,
Box const& domain, BCRec const* bc) noexcept
Box const& domain, IntVect const& /*ratio*/, BCRec const* bc) noexcept
{
Real sfx = Real(1.0);

Expand Down
152 changes: 97 additions & 55 deletions Src/AmrCore/AMReX_MFInterp_2D_C.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,24 +13,38 @@ void mf_cell_cons_lin_interp_limit_minmax_llslope (int i, int j, int, Array4<Rea
Real sfx = Real(1.0);
Real sfy = Real(1.0);

Real sx = Real(0.0);
Real sy = Real(0.0);

Real dcx = Real(0.0);
Real dcy = Real(0.0);

for (int ns = 0; ns < ncomp; ++ns) {
int nu = ns + scomp;

// x-direction
Real dcx = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
Real sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dcx)*amrex::min(sx,std::abs(dcx));
slope(i,j,0,ns ) = dcx;
if (ratio[0] > 1) {
dcx = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dcx)*amrex::min(sx,std::abs(dcx));
slope(i,j,0,ns ) = dcx;
} else {
slope(i,j,0,ns ) = Real(0.0);
}

// y-direction
Real dcy = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
Real sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dcy)*amrex::min(sy,std::abs(dcy));
slope(i,j,0,ns+ ncomp) = dcy;
if (ratio[1] > 1) {
dcy = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
Real db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dcy)*amrex::min(sy,std::abs(dcy));
slope(i,j,0,ns+ ncomp) = dcy;
} else {
slope(i,j,0,ns+ ncomp) = Real(0.0);
}

// adjust limited slopes to prevent new min/max for this component
Real alpha = Real(1.0);
Expand All @@ -39,8 +53,10 @@ void mf_cell_cons_lin_interp_limit_minmax_llslope (int i, int j, int, Array4<Rea
+ std::abs(sy) * Real(ratio[1]-1)/Real(2*ratio[1]);
Real umax = u(i,j,0,nu);
Real umin = u(i,j,0,nu);
for (int joff = -1; joff <= 1; ++joff) {
for (int ioff = -1; ioff <= 1; ++ioff) {
int ilim = ratio[0] > 1 ? 1 : 0;
int jlim = ratio[1] > 1 ? 1 : 0;
for (int joff = -jlim; joff <= jlim; ++joff) {
for (int ioff = -ilim; ioff <= ilim; ++ioff) {
umin = amrex::min(umin, u(i+ioff,j+joff,0,nu));
umax = amrex::max(umax, u(i+ioff,j+joff,0,nu));
}}
Expand Down Expand Up @@ -73,7 +89,7 @@ void mf_cell_cons_lin_interp_limit_minmax_llslope (int i, int j, int, Array4<Rea
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
void mf_cell_cons_lin_interp_llslope (int i, int j, int, Array4<Real> const& slope,
Array4<Real const> const& u, int scomp, int ncomp,
Box const& domain, BCRec const* bc) noexcept
Box const& domain, IntVect const& ratio, BCRec const* bc) noexcept
{
Real sfx = Real(1.0);
Real sfy = Real(1.0);
Expand All @@ -82,26 +98,34 @@ void mf_cell_cons_lin_interp_llslope (int i, int j, int, Array4<Real> const& slo
int nu = ns + scomp;

// x-direction
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
Real sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
if (dc != Real(0.0)) {
sfx = amrex::min(sfx, sx / dc);
if (ratio[0] > 1) {
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
Real sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
if (dc != Real(0.0)) {
sfx = amrex::min(sfx, sx / dc);
}
slope(i,j,0,ns ) = dc;
} else {
slope(i,j,0,ns ) = Real(0.0);
}
slope(i,j,0,ns ) = dc;

// y-direction
dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
Real sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
if (dc != Real(0.0)) {
sfy = amrex::min(sfy, sy / dc);
if (ratio[1] > 1) {
Real dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
Real db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
Real sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
if (dc != Real(0.0)) {
sfy = amrex::min(sfy, sy / dc);
}
slope(i,j,0,ns+ ncomp) = dc;
} else {
slope(i,j,0,ns+ ncomp) = Real(0.0);
}
slope(i,j,0,ns+ ncomp) = dc;
}

for (int ns = 0; ns < ncomp; ++ns) {
Expand All @@ -118,28 +142,37 @@ void mf_cell_cons_lin_interp_mcslope (int i, int j, int /*k*/, int ns, Array4<Re
{
int nu = ns + scomp;

Real sx = Real(0.);
Real sy = Real(0.);

// x-direction
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
Real sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
if (ratio[0] > 1) {
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
}

// y-direction
dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
Real sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
if (ratio[1] > 1) {
Real dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
Real db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
}

Real alpha = Real(1.0);
if (sx != Real(0.0) || sy != Real(0.0)) {
Real dumax = std::abs(sx) * Real(ratio[0]-1)/Real(2*ratio[0])
+ std::abs(sy) * Real(ratio[1]-1)/Real(2*ratio[1]);
Real umax = u(i,j,0,nu);
Real umin = u(i,j,0,nu);
for (int joff = -1; joff <= 1; ++joff) {
for (int ioff = -1; ioff <= 1; ++ioff) {
int ilim = ratio[0] > 1 ? 1 : 0;
int jlim = ratio[1] > 1 ? 1 : 0;
for (int joff = -jlim; joff <= jlim; ++joff) {
for (int ioff = -ilim; ioff <= ilim; ++ioff) {
umin = amrex::min(umin, u(i+ioff,j+joff,0,nu));
umax = amrex::max(umax, u(i+ioff,j+joff,0,nu));
}}
Expand Down Expand Up @@ -177,19 +210,26 @@ void mf_cell_cons_lin_interp_mcslope_rz (int i, int j, int ns, Array4<Real> cons
{
int nu = ns + scomp;

Real sx = Real(0.0);
Real sy = Real(0.0);

// x-direction
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
Real sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
if (ratio[0] > 1) {
Real dc = mf_compute_slopes_x(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i+1,j,0,nu) - u(i ,j,0,nu));
Real db = Real(2.0) * (u(i ,j,0,nu) - u(i-1,j,0,nu));
sx = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sx = std::copysign(Real(1.),dc)*amrex::min(sx,std::abs(dc));
}

// y-direction
dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
Real sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
if (ratio[1] > 1) {
Real dc = mf_compute_slopes_y(i, j, 0, u, nu, domain, bc[ns]);
Real df = Real(2.0) * (u(i,j+1,0,nu) - u(i,j ,0,nu));
Real db = Real(2.0) * (u(i,j ,0,nu) - u(i,j-1,0,nu));
sy = (df*db >= Real(0.0)) ? amrex::min(std::abs(df),std::abs(db)) : Real(0.);
sy = std::copysign(Real(1.),dc)*amrex::min(sy,std::abs(dc));
}

Real alpha = Real(1.0);
if (sx != Real(0.0) || sy != Real(0.0)) {
Expand All @@ -214,8 +254,10 @@ void mf_cell_cons_lin_interp_mcslope_rz (int i, int j, int ns, Array4<Real> cons
+ std::abs(sy) * Real(ratio[1]-1)/Real(2*ratio[1]);
Real umax = u(i,j,0,nu);
Real umin = u(i,j,0,nu);
for (int joff = -1; joff <= 1; ++joff) {
for (int ioff = -1; ioff <= 1; ++ioff) {
int ilim = ratio[0] > 1 ? 1 : 0;
int jlim = ratio[1] > 1 ? 1 : 0;
for (int joff = -jlim; joff <= jlim; ++joff) {
for (int ioff = -ilim; ioff <= ilim; ++ioff) {
umin = amrex::min(umin, u(i+ioff,j+joff,0,nu));
umax = amrex::max(umax, u(i+ioff,j+joff,0,nu));
}}
Expand Down
Loading
Loading