Skip to content

Speed up calculation of subcell bounds #2727

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

Merged
DanielDoehring merged 2 commits into from
Jan 8, 2026
Merged

Speed up calculation of subcell bounds #2727

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
dcfed80
Revise calculation of two-sided bounds
bennibolm Jan 8, 2026
437ed1b
typo
bennibolm Jan 8, 2026
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
47 changes: 27 additions & 20 deletions src/solvers/dgsem_tree/subcell_limiters_2d.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,30 +17,33 @@
mesh, _, dg, cache = mesh_equations_solver_cache(semi)
# Calc bounds inside elements
@threaded for element in eachelement(dg, cache)
var_min[:, :, element] .= typemax(eltype(var_min))
var_max[:, :, element] .= typemin(eltype(var_max))
# Calculate bounds at Gauss-Lobatto nodes using u
# Calculate bounds at Gauss-Lobatto nodes
for j in eachnode(dg), i in eachnode(dg)
var = u[variable, i, j, element]
var_min[i, j, element] = var
var_max[i, j, element] = var
end

# Apply values in x direction
for j in eachnode(dg), i in 2:nnodes(dg)
var = u[variable, i - 1, j, element]
var_min[i, j, element] = min(var_min[i, j, element], var)
var_max[i, j, element] = max(var_max[i, j, element], var)

if i > 1
var_min[i - 1, j, element] = min(var_min[i - 1, j, element], var)
var_max[i - 1, j, element] = max(var_max[i - 1, j, element], var)
end
if i < nnodes(dg)
var_min[i + 1, j, element] = min(var_min[i + 1, j, element], var)
var_max[i + 1, j, element] = max(var_max[i + 1, j, element], var)
end
if j > 1
var_min[i, j - 1, element] = min(var_min[i, j - 1, element], var)
var_max[i, j - 1, element] = max(var_max[i, j - 1, element], var)
end
if j < nnodes(dg)
var_min[i, j + 1, element] = min(var_min[i, j + 1, element], var)
var_max[i, j + 1, element] = max(var_max[i, j + 1, element], var)
end
var = u[variable, i, j, element]
var_min[i - 1, j, element] = min(var_min[i - 1, j, element], var)
var_max[i - 1, j, element] = max(var_max[i - 1, j, element], var)
end

# Apply values in y direction
for j in 2:nnodes(dg), i in eachnode(dg)
var = u[variable, i, j - 1, element]
var_min[i, j, element] = min(var_min[i, j, element], var)
var_max[i, j, element] = max(var_max[i, j, element], var)

var = u[variable, i, j, element]
var_min[i, j - 1, element] = min(var_min[i, j - 1, element], var)
var_max[i, j - 1, element] = max(var_max[i, j - 1, element], var)
end
end

Expand Down Expand Up @@ -120,6 +123,10 @@ end
@inline function calc_bounds_onesided!(var_minmax, min_or_max, variable,
u::AbstractArray{<:Any, 4}, t, semi)
mesh, equations, dg, cache = mesh_equations_solver_cache(semi)

# The approach used in `calc_bounds_twosided!` is not used here because it requires more
# evaluations of the variable and is therefore slower.

# Calc bounds inside elements
@threaded for element in eachelement(dg, cache)
# Reset bounds
Expand All @@ -131,7 +138,7 @@ end
end
end

# Calculate bounds at Gauss-Lobatto nodes using u
# Calculate bounds at Gauss-Lobatto nodes
for j in eachnode(dg), i in eachnode(dg)
var = variable(get_node_vars(u, equations, dg, i, j, element), equations)
var_minmax[i, j, element] = min_or_max(var_minmax[i, j, element], var)
Expand Down
Loading