DTypeFront.toml

# *****************************************************************
# Problem size and geometry
# *****************************************************************
geometry.prob_lo     =  [0.0, 0.0, 0.0]
geometry.prob_hi     =  [1.596e19, 1.596e19, 1.596e19]
quokka.bc = ["reflecting", "reflecting", "reflecting"]

# *****************************************************************
# VERBOSITY
# *****************************************************************
amr.v              = 0       # verbosity in Amr

# *****************************************************************
# Resolution and refinement
# *****************************************************************
amr.n_cell          = [64, 64, 64]
amr.max_level       = 0     # number of levels = max_level + 1
amr.blocking_factor = 32  # grid size must be divisible by this
amr.max_grid_size = 128  # at least 128 for GPUs
do_reflux = 0
do_subcycle = 0
suppress_output = 0

stop_time = 1.93e13 # = 20 t_s

stromgen.primary_species_1 = 1.0e-10
stromgen.primary_species_2 = 139.9
stromgen.primary_species_3 = 1.0e-10
stromgen.Q = 4e46
stromgen.temperature = 37.11

photochemistry.enabled = 1
integrator.use_number_densities = 1
integrator.do_species_clip = 0
integrator.jacobian = 1
integrator.renormalize_abundances = 0
integrator.subtract_internal_energy = 0
integrator.X_reject_buffer = 1e100
integrator.ode_max_steps = 50000

plotfile_interval = -1
plotfile_prefix = "DType"

network.allow_mixed_cell_cooling = 0


max_timesteps = 1000000  # stop_time governs; no artificial step cap

# VODE absolute tolerances — tuned for DTypeFront (n_H = 139.9 cm^-3).
# Erad_floor (M1 floor) is C::a_rad * 1e-8 (T~0.01 K) in testDTypeFront.cpp — see docs/markdown/photoionization.md.
#
# atol_rad_num: the performance plateau for this problem is at ~5e-2 cm^-3. Below this value VODE spends
# extra steps in dark cells; above it the step-rejection loop (N_gamma near -radiation_failure_tolerance)
# erodes the gain. radiation_failure_tolerance must exceed the maximum BDF predictor-corrector overshoot
# at the ionization front (~0.05–0.2 cm^-3 for this problem), so 0.5 provides a safe margin.
#
# species_failure_tolerance = atol_spec: the 1.5× final-state factor in vode_type.H
# (vode_final_state_species_failure_tolerance_factor) absorbs BDF interpolation overshoot
# without manual inflation — the two values can therefore be set equal.
integrator.atol_spec                     = 1.4e-3   # cm^-3; 1e-5 × n_H = physical negligibility floor
integrator.atol_enuc                     = 1.24e8   # erg/g; c_v × 1 K temperature accuracy
integrator.atol_rad_num                  = 5e-2     # cm^-3; performance plateau for this problem
integrator.species_failure_tolerance     = 1.4e-3   # cm^-3; = atol_spec (1.5× final-state factor in vode_type.H absorbs BDF overshoot)
integrator.radiation_failure_tolerance   = 0.5      # cm^-3; must exceed max N_gamma BDF overshoot at ionization front

integrator.rtol_spec     = 1e-2
integrator.rtol_rad_num  = 1e-2
integrator.rtol_enuc     = 1e-2

integrator.rosenbrock_tableau = 3