callbacks_step.jl

# By default, Julia/LLVM does not use fused multiply-add operations (FMAs).
# Since these FMAs can increase the performance of many numerical algorithms,
# we need to opt-in explicitly.
# See https://ranocha.de/blog/Optimizing_EC_Trixi for further details.
@muladd begin
#! format: noindent

# overload this function for specific callbacks which use element element variables
# that should be saved
function get_element_variables!(element_variables, u, mesh, equations, solver, cache,
                                callback; kwargs...)
    return nothing
end

@inline function get_element_variables!(element_variables, u_ode,
                                        semi::AbstractSemidiscretization,
                                        cb::DiscreteCallback;
                                        kwargs...)
    mesh, equations, solver, cache = mesh_equations_solver_cache(semi)
    u = wrap_array(u_ode, mesh, equations, solver, cache)
    get_element_variables!(element_variables, u, mesh, equations, solver, cache,
                           cb.affect!; kwargs...)
end

@inline function isfinished(integrator)
    # Checking for floating point equality is OK here as `DifferentialEquations.jl`
    # sets the time exactly to the final time in the last iteration
    return integrator.t == last(integrator.sol.prob.tspan) ||
           isempty(integrator.opts.tstops) ||
           integrator.iter == integrator.opts.maxiters
end

# `include` callback definitions in the order that we currently prefer
# when combining them into a `CallbackSet` which is called *after* a complete step
# The motivation is as follows: The first callbacks belong to the current time step iteration:
# * `SummaryCallback` controls, among other things, timers and should thus be first
# * `SteadyStateCallback` may mark a time step as the last step, which is needed by other callbacks
# * `AnalysisCallback` may also do some checks that mark a step as the last one
# * `AliveCallback` belongs to `AnalysisCallback` and should thus be nearby
# * `SaveRestartCallback`, `SaveSolutionCallback`, and `TimeSeriesCallback` should save the current
#    solution state before it is potentially degraded by AMR
# * `VisualizationCallback` similarly should be called before the mesh is adapted
#
# From here on, the remaining callbacks essentially already belong to the next time step iteration:
# * `AMRCallback` really belongs to the next time step already, as it should be the "first" callback
#   in a time step loop (however, callbacks are always executed *after* a step, thus it comes near
#   the end here)
# * `StepsizeCallback` must come after AMR to accommodate potential changes in the minimum cell size
# * `GlmSpeedCallback` must come after computing time step size because it affects the value of c_h
# * `LBMCollisionCallback` must come after computing time step size because it is already part of
#    the next time step calculation
include("summary.jl")
include("steady_state.jl")
include("analysis.jl")
include("alive.jl")
include("save_restart.jl")
include("save_solution.jl")
include("time_series.jl")
include("visualization.jl")
include("averaging.jl")

include("amr.jl")
include("stepsize.jl")
include("glm_speed.jl")
include("lbm_collision.jl")
include("euler_acoustics_coupling.jl")

# The `TrivialCallback` purposely does nothing: It allows to quickly disable specific callbacks
# when using `trixi_include` or `test_trixi_include`
include("trivial.jl")

# DGMulti callbacks
include("analysis_dgmulti.jl")
end # @muladd