adam_realm_object

ADAM, realm class definition — common per-realm state and orchestrator contract.

A realm is the unit the forest_object orchestrator tends. The name is the crasi of two aspects this type entails simultaneously:

1. A physical-numerical model: the PDE system, numerics configuration, time integrator, boundary conditions, etc. (math layer).
2. A spatial region modelization: the grid, the AMR Morton tree, the field data, the communication maps (where layer).

A realm has both laws (1) and borders (2). Neither name equation_object (math-only) nor subdomain_object (region-only) captured both aspects — hence realm_object, evoking a bounded territory governed by its own rules.

Consumer apps (PRISM, NASTO, CHASE, ...) extend realm_object to define their specific PDE family (prism_common_object extends realm_object, etc.). Each extended type adds app-private data and overrides the orchestrator contract TBPs.

The orchestrator contract is the set of TBPs carrying the _forest suffix. These are the methods forest_object may call on a realm; grep "_forest" across the codebase reveals the entire orchestrator contract surface. Three architectural rules govern these TBPs:

• O1 — signature uses only ADAM-lib-visible types (intrinsic kinds plus class(realm_object), grid_object, etc.). Apps override the implementation; signatures are sacred.
• O2 — app-specific dispatch (which integrator, which physics) lives in self%'s components, not in orchestrator-supplied arguments.
• O3 — q is never public on realm_object. Halo exchange between realms goes through realm-side TBPs that operate on q internally; the orchestrator schedules but never touches q.

Source: src/lib/common/adam_realm_object.F90

Dependencies

Contents

• realm_object
• compute_block_total_variation_interface
• compute_curl_interface
• compute_derivative1_interface
• compute_derivative2_interface
• compute_derivative4_interface
• compute_divergence_interface
• compute_gradient_interface
• compute_laplacian_interface
• initialize
• load_fdv_from_file
• initialize_forest
• compute_local_dt_forest
• advance_one_step_forest
• coupling_descriptor_forest
• open_step_forest
• begin_stage_forest
• end_stage_forest
• close_step_forest
• post_step_forest
• is_done_forest
• finalize_forest
• finalize_mpi_forest
• fill_seam_from_peer_forest
• after_topology_build_forest
• apply_reflux_to_stage_forest
• close_block_xh5f
• close_file_xh5f
• open_block_xh5f
• open_file_xh5f
• save_q_xh5f
• compute_block_total_variation_fd
• compute_curl_fd
• compute_curl_fv
• compute_derivative1_fd
• compute_derivative1_fv
• compute_derivative2_fd
• compute_derivative2_fv
• compute_derivative4_fd
• compute_divergence_fd
• compute_divergence_fv
• compute_gradient_fd
• compute_gradient_fv
• compute_laplacian_fd
• compute_laplacian_fv
• stages_per_step_forest

Derived Types

realm_object

Realm system class definition, common data to all backends and applications.

@NOTE: the orchestrator contract methods carrying the _forest suffix are the surface forest_object may call on a realm. Default implementations here error-stop with a "not overridden" message; each consumer app must override.

Inheritance

Components

Name	Type	Attributes	Description
io	type(io_object)		IO handler.
amr	type(amr_object)		AMR marker handler.
slices	type(slices_object)		Slices handler.
blanesmoan	type(blanesmoan_object)		Blanes-Moan integrator.
cfm	type(cfm_object)		Commutator-Free Magnus integrator.
leapfrog	type(leapfrog_object)		Leapfrog integrator.
flail	type(flail_object)		Linear algebra methods handler.
weno	type(weno_object)		WENO reconstructor.
ib	type(ib_object)		Immersed boundary.
rk	type(rk_object)		Runge-Kutta integrator.
adam	type(adam_object)		ADAM, grid + tree + field + maps orchestrator.
fdv_scheme	character(len=:)	allocatable	FDV scheme, fd/fv.
fdv_order	integer(kind=I4P)		Order of finite difference/volume schemes, general order.
fdv_half_stencil	integer(kind=I4P)		Half stencil length of finite difference/volume schemes.
fdv_half_stencils	integer(kind=I4P)		Half stencil length of fdv schemes for each derivative up to 6.
ngc	integer(kind=I4P)	pointer	Number of ghost cells.
ni	integer(kind=I4P)	pointer	Number of cells in i direction.
nj	integer(kind=I4P)	pointer	Number of cells in j direction.
nk	integer(kind=I4P)	pointer	Number of cells in k direction.
nb	integer(kind=I4P)	pointer	Total blocks number for MPI.
blocks_number	integer(kind=I4P)	pointer	Actual blocks number.
nv	integer(kind=I4P)	pointer	Number of variables in q vector.
realm_index	integer(kind=I4P)		Inter-realm seam coupling — currently-published stage buffer key.
stage_active	integer(kind=I4P)		Procedure pointer TBPs for FDV operators (set at initialization by backend).
compute_block_total_variation	procedure(compute_block_total_variation_interface)	pass(self), pointer	Compute TV.
compute_curl	procedure(compute_curl_interface)	pass(self), pointer	Compute curl.
compute_derivative1	procedure(compute_derivative1_interface)	pass(self), pointer	Compute deriv1.
compute_derivative2	procedure(compute_derivative2_interface)	pass(self), pointer	Compute deriv2.
compute_derivative4	procedure(compute_derivative4_interface)	pass(self), pointer	Compute deriv4.
compute_divergence	procedure(compute_divergence_interface)	pass(self), pointer	Compute dive.
compute_gradient	procedure(compute_gradient_interface)	pass(self), pointer	Compute grad.
compute_laplacian	procedure(compute_laplacian_interface)	pass(self), pointer	Compute laplac.

Type-Bound Procedures

Name	Attributes	Description
initialize	pass(self)	Initialize common data.
load_fdv_from_file	pass(self)	Load FDV config from file.
initialize_forest	pass(self)	Invoked by forest%initialize per realm at startup.
compute_local_dt_forest	pass(self)	Invoked by forest%compute_global_dt during the min reduction.
advance_one_step_forest	pass(self)	Invoked by forest%evolve_one_step per realm per timestep.
stages_per_step_forest	pass(self)	Number of integrator stages this realm exposes per step.
coupling_descriptor_forest	pass(self)	Report (scheme_time, rk_scheme, nv) for β admissibility.
open_step_forest	pass(self)	Per-step prologue (multi-realm path).
begin_stage_forest	pass(self)	Begin one integrator stage on this realm (multi-realm path).
end_stage_forest	pass(self)	End the stage: residuals + assignment (multi-realm path).
close_step_forest	pass(self)	Per-step epilogue (multi-realm path).
post_step_forest	pass(self)	Invoked by forest%post_step per realm per timestep.
is_done_forest	pass(self)	Invoked by forest%is_done during the termination reduction.
finalize_forest	pass(self)	Invoked by forest%finalize per realm at shutdown.
finalize_mpi_forest	pass(self)	Process-global MPI finalize; forest calls it ONCE after all.
fill_seam_from_peer_forest	pass(self)	Receive-side roundtrip: copy peer's interior into self's ghosts.
after_topology_build_forest	pass(self)	Backend hook invoked once after the forest builds the seam maps.
apply_reflux_to_stage_forest	pass(self)	Apply Berger-Colella reflux to self's stage buffer.
close_block_xh5f	nopass	Close XH5F file block.
close_file_xh5f	nopass	Close XH5F file.
open_block_xh5f	pass(self)	Open block file XH5F.
open_file_xh5f	pass(self)	Open file XH5F.
save_q_xh5f	pass(self)	Save in XH5F (XDMF/HDF5) format.
compute_block_total_variation_fd	pass(self)	Return the max of block total variation for a given var.
compute_curl_fd	pass(self)	Compute curl of vector field, finite difference.
compute_curl_fv	pass(self)	Compute curl of vector field, finite volume.
compute_derivative1_fd	pass(self)	Compute derivative1 of scalar field, finite difference.
compute_derivative1_fv	pass(self)	Compute derivative1 of scalar field, finite volume.
compute_derivative2_fd	pass(self)	Compute derivative2 of scalar field, finite difference.
compute_derivative2_fv	pass(self)	Compute derivative2 of scalar field, finite volume.
compute_derivative4_fd	pass(self)	Compute derivative4 of scalar field, finite difference.
compute_divergence_fd	pass(self)	Compute divergence of vector field, finite difference.
compute_divergence_fv	pass(self)	Compute divergence of vector field, finite volume.
compute_gradient_fd	pass(self)	Compute gradient of scalar field, finite difference.
compute_gradient_fv	pass(self)	Compute gradient of scalar field, finite volume.
compute_laplacian_fd	pass(self)	Compute laplacian of scalar field, finite difference.
compute_laplacian_fv	pass(self)	Compute laplacian of scalar field, finite volume.

Interfaces

compute_block_total_variation_interface

compute_curl_interface

compute_derivative1_interface

compute_derivative2_interface

compute_derivative4_interface

compute_divergence_interface

compute_gradient_interface

compute_laplacian_interface

Subroutines

initialize

Initialize common data: MPI, ADAM, grid, field and data replica pointers.

subroutine initialize(self, filename, memory_avail, nv, verbose, L0)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout	target	The equation.
filename	character(len=*)	in		Input parameters file name.
memory_avail	real(kind=R8P)	in	value	Memory available for single MPI process.
nv	integer(kind=I4P)	in	optional	Number of field variables.
verbose	logical	in	optional	Trigger verbose output.
L0	real(kind=R8P)	in	optional	Adimensionalization parameter.

Call graph

load_fdv_from_file

Load FDV config from file.

subroutine load_fdv_from_file(self, file_parameters, go_on_fail)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The equation.
file_parameters	type(file_ini)	in		Simulation parameters ini file handler.
go_on_fail	logical	in	optional	Go on if load fails.

Call graph

initialize_forest

Initialize this realm from scratch: app-level initialize, IC injection (or restart load), initial AMR, IO file open, initial diagnostics dump.

Invoked by forest%initialize once per realm at startup, before the time loop begins. This is the single per-realm setup entry point the orchestrator uses: it both runs the basic per-realm initialize (which the legacy simulate did via realm%initialize_prism(filename) or equivalent) and performs the post-init / pre-loop setup that has to happen before forest%evolve_one_step can be called.

Realm-position context: the forest sets self%realm_index = is BEFORE invoking this TBP, so the override body can already read its own 1-based forest position from self%realm_index for per-realm IO filenames, log prefixes, etc.

Default implementation error-stops: every consumer app MUST override this method (the initial-condition catalog, AMR strategy, and which IO files to open are app-specific).

subroutine initialize_forest(self, filename, realms_number, memory_avail, nv, verbose)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
filename	character(len=*)	in		Input parameters file name.
realms_number	integer(kind=I4P)	in	optional	Realm count in the forest.
memory_avail	real(kind=R8P)	in	optional	Per-process memory budget override.
nv	integer(kind=I4P)	in	optional	Number of field variables override.
verbose	logical	in	optional	Trigger verbose output.

Call graph

compute_local_dt_forest

Compute this realm's local stability-limited dt (no MPI reduction).

Invoked by forest%compute_global_dt during the min reduction across all realms in the forest. The reduction itself is the orchestrator's job; this method computes only the value local to self.

Default implementation error-stops: every consumer app MUST override this method (the CFL criterion and umax are app-specific). PRISM's override lives on prism_cpu_object/prism_fnl_object.

subroutine compute_local_dt_forest(self, dt_local)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The realm.
dt_local	real(kind=R8P)	out		Local stability-limited dt.

Call graph

advance_one_step_forest

Advance this realm by one full timestep of size dt.

Invoked by forest%evolve_one_step once per realm per timestep when the forest contains a single realm (N=1 fast path). For N>1 the forest drives the stage loop itself via open_step_forest, begin_stage_forest, end_stage_forest, close_step_forest, and this TBP is NOT invoked — see forest_object%evolve_one_step for the multi-realm path.

The orchestrator owns global dt selection (compute_global_dt) and the termination check; this method owns the integration itself — RK substages, BC application, intra-realm ghost exchange, divergence cleaning — i.e. everything that turns q at time t into q at time t + dt.

Default implementation error-stops: every consumer app MUST override this method (the integrator catalog and the substage layout are app-specific). PRISM's override lives on prism_cpu_object and prism_fnl_object as a thin wrapper around the legacy integrate / integrate_dev dispatch.

subroutine advance_one_step_forest(self, dt)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
dt	real(kind=R8P)	in		Timestep size.

Call graph

coupling_descriptor_forest

Report this realm's coupling descriptor for β admissibility.

Returns the identity of the realm's ODE solver + physics layout. Two realms may share a seam with CADENCE_STAGE_COINCIDENT only if all three values match between them; otherwise the forest rejects the manifest. Spatial operator (scheme_space) and grid resolution are NOT in the descriptor — those are precisely the dimensions β allows to differ between realms.

Default implementation returns empty strings and sentinel nv = -1: a realm that does not override this TBP cannot participate in β. App-side overrides (e.g. prism_common_object) report numerics%scheme_time, rk%scheme, physics%nv.

subroutine coupling_descriptor_forest(self, scheme_time, rk_scheme, nv)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The realm.
scheme_time	character(len=:)	out	allocatable	Time-integration family tag ("runge_kutta", "leapfrog", ...).
rk_scheme	character(len=:)	out	allocatable	Within-family scheme tag ("runge-kutta-ssp-54", ...).
nv	integer(kind=I4P)	out		Number of conserved variables on this realm.

Call graph

open_step_forest

Per-step prologue on the multi-realm path: pre-stage-loop setup.

Invoked once per realm per timestep, BEFORE the forest's stage loop starts. Body owns whatever the integrator does once per step before its first stage: external-field prelude, integrator-private state initialization (for RK realms: rk%initialize_stages), time bookkeeping (it increment, dt capping for time_max), progress prints.

Default implementation error-stops: every multi-realm participant MUST override this. N=1 apps can leave it.

subroutine open_step_forest(self, dt)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
dt	real(kind=R8P)	in		Timestep size from the forest.

Call graph

begin_stage_forest

Open one integrator stage on this realm — multi-realm path.

Invoked once per (realm, stage) by the forest, with k iterating from 1 to K_total. Body publishes the realm's stage-k buffer to peers: it runs whatever integrator-private work is needed BEFORE residuals (for RK realms: rk%compute_stage(s=k) populating rk%q_rk(:,...,k) from prior stages and self%q), and assigns self%stage_active = k so peer-realm reads pick up the right buffer. Crucially it does NOT touch ghost cells from peer realms; peer realms may not yet have opened their own stage-k buffer when this fires.

The forest invokes begin_stage_forest on EVERY realm before proceeding to the seam fill and end_stage_forest, so by the time residuals evaluate, every peer realm has its stage-k buffer populated and the inter-realm ghost exchange returns coherent values.

Default implementation error-stops: every multi-realm participant MUST override this.

subroutine begin_stage_forest(self, k, K_total, dt, realm)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
k	integer(kind=I4P)	in		Stage index (1..K_total).
K_total	integer(kind=I4P)	in		Forest-wide stage count for this step.
dt	real(kind=R8P)	in		Timestep size from the forest.
realm	class(realm_object)	inout	target, optional	Sibling realms (contract parity).

Call graph

end_stage_forest

Close one integrator stage on this realm — multi-realm path.

Invoked once per (realm, stage) by the forest AFTER its Phase 2 inter-realm seam fill has populated the seam ghost cells. The body runs the integrator's per-stage work that needs valid peer ghosts: residual computation and stage assignment in a single sweep (for RK realms: compute_residuals + rk%assign_stage). The read-after- overwrite race that originally required a three-phase split is eliminated by the upstream seam fill, which decouples peer state reads from the in-place overwrite of the stage buffer's interior.

flux_register is an optional dummy used by FV residuals (compute_residuals_fv_centered) to accumulate seam-face flux contributions for Phase-A Berger–Colella reflux. Backends that don't implement FV reflux (e.g. FNL's FD-centered path) ignore it.

Default implementation error-stops: every multi-realm participant MUST override this. realm accepted on contract for parity but is not needed: peer state is read via the seam fill, not here.

subroutine end_stage_forest(self, k, K_total, dt, realm, flux_register)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
k	integer(kind=I4P)	in		Stage index (1..K_total).
K_total	integer(kind=I4P)	in		Forest-wide stage count for this step.
dt	real(kind=R8P)	in		Timestep size from the forest.
realm	class(realm_object)	inout	target, optional	Sibling realms (parity only).
flux_register	class(flux_register_object)	inout	optional	Forest's flux register for FV reflux.

Call graph

close_step_forest

Per-step epilogue on the multi-realm path: post-stage-loop finalization.

Invoked once per realm per timestep, AFTER the forest's stage loop completes. Body owns whatever the integrator does once per step at the end: assembly of the committed state from stage buffers (for RK realms: rk%update_q), BC update, divergence cleaning, coil current refresh, residual save, time advance. Should also clear self%stage_active = 0 so any inter-step fill_seam_from_peer_forest query reads from the committed state rather than a stale stage buffer.

Default implementation error-stops: every multi-realm participant MUST override this.

subroutine close_step_forest(self, dt)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
dt	real(kind=R8P)	in		Timestep size from the forest.

Call graph

post_step_forest

Run this realm's per-timestep post-step work: diagnostics, IO, AMR.

Invoked by forest%post_step once per realm per timestep, after the advance. The orchestrator owns cadence decisions (when to save state, when to refine the mesh, etc.) and conveys them via the optional do_* flags; this method executes the work that those flags enable. Cadence flags absent ⇒ caller wants the default cadence (today: do everything every step, since save_*_data routines internally respect the realm's own it_save settings).

Default implementation error-stops: every consumer app MUST override this method (the diagnostic catalog and IO layout are app-specific). PRISM's override lives on prism_cpu_object and prism_fnl_object.

Optional realm(:): forest realm array threaded for inter-realm halo refresh inside save_simulation_data (which calls update_ghost). FNL backend forwards it through; CPU backend ignores it.

subroutine post_step_forest(self, dt, t, it, do_save_state, do_save_residuals, do_save_restart, do_amr, realm)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
dt	real(kind=R8P)	in		Timestep size just advanced.
t	real(kind=R8P)	in		Simulation time after the advance.
it	integer(kind=I4P)	in		Iteration index after the advance.
do_save_state	logical	in	optional	Save state output this step.
do_save_residuals	logical	in	optional	Save residuals output this step.
do_save_restart	logical	in	optional	Save restart dump this step.
do_amr	logical	in	optional	Run AMR update this step.
realm	class(realm_object)	inout	target, optional	Sibling realms for inter-realm halo refresh.

Call graph

is_done_forest

Decide whether this realm's local termination criterion is met.

Invoked by forest%is_done during the termination reduction across all realms in the forest. The reduction itself (typically a logical AND or OR depending on coupling semantics) is the orchestrator's job; this method computes only the predicate local to self.

Default implementation error-stops: every consumer app MUST override this method (the termination criterion mixes iteration count, simulated time, and app-specific physics signals). PRISM's override lives on prism_cpu_object/prism_fnl_object.

subroutine is_done_forest(self, done)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The realm.
done	logical	out		True if this realm is done evolving.

Call graph

finalize_forest

Shut this realm down: final state dump, final diagnostics, close output files, finalize the realm's MPI handler.

Invoked by forest%finalize once per realm at shutdown, after the time loop exits. Mirror of initialize_forest: every IO file opened in initialize_forest is closed here; every per-realm resource the orchestrator did not acquire is released here.

Default implementation error-stops: every consumer app MUST override this method (the close-order and final-dump catalog are app-specific). PRISM's override lives on prism_cpu_object/prism_fnl_object.

subroutine finalize_forest(self)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.

Call graph

finalize_mpi_forest

Finalize the process-global MPI handler.

MPI_FINALIZE is process-global and must run EXACTLY ONCE, after every realm has completed all MPI-using teardown. The forest therefore calls this on a SINGLE realm after the per-realm finalize_forest loop — not inside finalize_forest (which runs per realm; doing the finalize there tore MPI down while later realms still had IO/ghost work pending — the next realm's update_ghost would hit MPI_Type_f2c on a finalized communicator). The default targets the CPU mpih singleton; the FNL backend overrides to finalize mpih_fnl.

subroutine finalize_mpi_forest(self)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm (carries no MPI state; dispatch picks the backend handler).

Call graph

fill_seam_from_peer_forest

Fill THIS realm's seam ghost cells for peer slot p_idx from peer's interior. Default: error_stops; every multi-realm participant MUST override (the receiver is the dispatch target).

subroutine fill_seam_from_peer_forest(self, peer, p_idx)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout
peer	class(realm_object)	in	target
p_idx	integer(kind=I4P)	in

Call graph

after_topology_build_forest

Backend hook invoked by forest%populate_inter_realm_topology AFTER all host-side seam maps have been built. CPU default is a no-op (host maps are the production state). GPU backends override to propagate the freshly-built host seam maps to device-resident copies (e.g. via maps_fnl%copy_cpu_gpu).

subroutine after_topology_build_forest(self)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout

Call graph

apply_reflux_to_stage_forest

Apply the Berger-Colella reflux correction to THIS realm's stage-stage buffer for every flux-register face where self is the coarse side.

Invoked by forest_object%evolve_one_step once per realm per stage, AFTER flux_register%reduce_fine_sums has globalised the fine-side accumulators. The realm filters the register's face array by face%coarse_realm == self%realm_index, picks up its OWN integrator- specific stage weight (for RK realms: self%rk%ark(stage)), and writes the correction delta into its OWN stage buffer (for RK realms: self%rk%q_rk(:, i, j, k, b, stage)).

Architectural role: this TBP keeps the per-realm reflux write integrator-private. The forest never names q_rk, ark, or any other RK-specific identifier — those live realm-side, so a future Leapfrog / Yoshida / CFM realm can implement reflux on its own buffer naming without touching the orchestrator.

The realm's 1-based forest position is read from self%realm_index, which the forest writes once at initialize-time; no per-call plumbing of is through the dispatch is needed.

Default implementation error-stops: every multi-realm participant that uses FV reflux MUST override this. Multi-realm participants that DO NOT accumulate flux contributions (e.g. an FD-only realm) should override with a no-op body.

subroutine apply_reflux_to_stage_forest(self, stage, dt, flux_register)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The realm.
stage	integer(kind=I4P)	in		Integrator stage 1..K_total.
dt	real(kind=R8P)	in		Time step.
flux_register	class(flux_register_object)	in		Forest's flux register (read-only).

Call graph

close_block_xh5f

Close XH5F file block.

subroutine close_block_xh5f(xh5f)

Arguments

Name	Type	Intent	Attributes	Description
xh5f	type(xh5f_file_object)	inout		XH5F file handler.

Call graph

close_file_xh5f

Close XH5F file.

subroutine close_file_xh5f(xh5f)

Arguments

Name	Type	Intent	Attributes	Description
xh5f	type(xh5f_file_object)	inout		XH5F file handler.

Call graph

open_block_xh5f

Open XH5F file block.

subroutine open_block_xh5f(self, xh5f, b, nijk, t, time)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The equation.
xh5f	type(xh5f_file_object)	inout		XH5F file handler.
b	integer(kind=I4P)	in		Block index.
nijk	integer(kind=I8P)	in		Blocks dimensions.
t	integer(kind=I4P)	in	optional	Time iteration.
time	real(kind=R8P)	in	optional	Time.

Call graph

open_file_xh5f

Open XH5F file.

subroutine open_file_xh5f(self, basename, xh5f, directory)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The equation.
basename	character(len=*)	in		Base name of output files.
xh5f	type(xh5f_file_object)	out		XH5F file handler.
directory	character(len=*)	in	optional	Directory name of output files.

Call graph

save_q_xh5f

Save ADAM in XH5F format.

subroutine save_q_xh5f(self, basename, q, q_name, directory, with_ghost, with_cell_morton, t, time)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	inout		The equation.
basename	character(len=*)	in		Base name of output files.
q	real(kind=R8P)	in		Q-vector variables [nv,ni,nj,nk,nb].
q_name	character(len=*)	in	optional	Q-vector variables names [nv].
directory	character(len=*)	in	optional	Directory name of output files.
with_ghost	logical	in	optional	Flag to save ghost cells.
with_cell_morton	logical	in	optional	Flag to save Morton code also in cells.
t	integer(kind=I4P)	in	optional	Time iteration.
time	real(kind=R8P)	in	optional	Time.

Call graph

compute_block_total_variation_fd

Return the max of block total variation for a given var.

subroutine compute_block_total_variation_fd(self, hs, dxyz, ivar, q, tot_var_field, total_variation)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dxyz	real(kind=R8P)	in		Space steps.
ivar	integer(kind=I4P)	in		Index of first component of vec field.
q	real(kind=R8P)	in		Field variables.
tot_var_field	real(kind=R8P)	inout		Total variation field on blocks.
total_variation	real(kind=R8P)	out		Max total variation on given block.

Call graph

compute_curl_fd

Compute curl of vector fields, div(q(ivar:ivar+2), using finite difference schemes.

subroutine compute_curl_fd(self, hs, ivar, q, curl)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Start index of variable of q.
q	real(kind=R8P)	in		Field variables.
curl	real(kind=R8P)	inout		Curl.

Call graph

compute_curl_fv

Compute curl of vector fields, div(q(ivar:ivar+2), using finite volume schemes.

subroutine compute_curl_fv(self, hs, ivar, q, curl)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Start index of variable of q.
q	real(kind=R8P)	in		Field variables.
curl	real(kind=R8P)	inout		Curl.

Call graph

compute_derivative1_fd

Compute derivative1 of scalar fields, dq(ivar)/ds, using finite difference schemes.

subroutine compute_derivative1_fd(self, hs, dir, ivar, q, dq_ds)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dir	integer(kind=I4P)	in		Direction, 1=X, 2=Y, 3=Z.
ivar	integer(kind=I4P)	in		Start index of (vec.) variable of q.
q	real(kind=R8P)	in		Field variables.
dq_ds	real(kind=R8P)	inout		Derivative1, dq/ds.

Call graph

compute_derivative1_fv

Compute derivative1 of scalar fields, dq(ivar)/ds, using finite volume schemes.

subroutine compute_derivative1_fv(self, hs, dir, ivar, q, dq_ds)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dir	integer(kind=I4P)	in		Direction, 1=X, 2=Y, 3=Z.
ivar	integer(kind=I4P)	in		Start index of (vec.) variable of q.
q	real(kind=R8P)	in		Field variables.
dq_ds	real(kind=R8P)	inout		Derivative1, dq/ds.

Call graph

compute_derivative2_fd

Compute derivative2 of scalar fields, d2q(ivar)/ds2, using finite difference schemes.

subroutine compute_derivative2_fd(self, hs, dir, ivar, q, d2q_ds2)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dir	integer(kind=I4P)	in		Direction, 1=X, 2=Y, 3=Z.
ivar	integer(kind=I4P)	in		Start index of vec variable of q.
q	real(kind=R8P)	in		Field variables.
d2q_ds2	real(kind=R8P)	inout		Derivative2, d2q/ds2.

Call graph

compute_derivative2_fv

Compute derivative2 of scalar fields, d2q(ivar)/ds2, using finite volume schemes.

subroutine compute_derivative2_fv(self, hs, dir, ivar, q, d2q_ds2)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dir	integer(kind=I4P)	in		Direction, 1=X, 2=Y, 3=Z.
ivar	integer(kind=I4P)	in		Start index of variable of q.
q	real(kind=R8P)	in		Field variables.
d2q_ds2	real(kind=R8P)	inout		Derivative2, d2q/ds2.

compute_derivative4_fd

Compute derivative2 of scalar fields, d4q(ivar)/ds4, using finite difference schemes.

subroutine compute_derivative4_fd(self, hs, dir, ivar, q, d4q_ds4)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
dir	integer(kind=I4P)	in		Direction, 1=X, 2=Y, 3=Z.
ivar	integer(kind=I4P)	in		Start index of variable of q.
q	real(kind=R8P)	in		Field variables.
d4q_ds4	real(kind=R8P)	inout		Derivative4, d4q/ds4.

Call graph

compute_divergence_fd

Compute divergence of vector fields, div(q(ivar:ivar+2), using finite difference schemes.

subroutine compute_divergence_fd(self, hs, ivar, q, divergence)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Start index of field of q.
q	real(kind=R8P)	in		Field variables.
divergence	real(kind=R8P)	inout		Divergence.

Call graph

compute_divergence_fv

Compute divergence of vector fields, div(q(ivar:ivar+2), using finite volume schemes.

subroutine compute_divergence_fv(self, hs, ivar, q, divergence)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Start index of field of q.
q	real(kind=R8P)	in		Field variables.
divergence	real(kind=R8P)	inout		Divergence.

Call graph

compute_gradient_fd

Compute gradient of scalar variable q(ivar), finite difference schemes.

subroutine compute_gradient_fd(self, hs, ivar, q, gradient)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Index of scalar variable of q.
q	real(kind=R8P)	in		Field variables.
gradient	real(kind=R8P)	inout		Gradient.

Call graph

compute_gradient_fv

Compute gradient of scalar variable q(ivar), finite volume schemes.

subroutine compute_gradient_fv(self, hs, ivar, q, gradient)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Index of scalar variable of q.
q	real(kind=R8P)	in		Field variables.
gradient	real(kind=R8P)	inout		Gradient.

Call graph

compute_laplacian_fd

Compute laplacian of scalar variable q(ivar), finite difference schemes.

subroutine compute_laplacian_fd(self, hs, ivar, q, laplacian)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Index of scalar variable of q.
q	real(kind=R8P)	in		Field variables.
laplacian	real(kind=R8P)	inout		Gradient.

Call graph

compute_laplacian_fv

Compute laplacian of scalar variable q(ivar), finite volume schemes.

subroutine compute_laplacian_fv(self, hs, ivar, q, laplacian)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The equation.
hs	integer(kind=I4P)	in		FDV half stencil length.
ivar	integer(kind=I4P)	in		Index of scalar variable of q.
q	real(kind=R8P)	in		Field variables.
laplacian	real(kind=R8P)	inout		Gradient.

Functions

stages_per_step_forest

Return the number of integrator stages this realm exposes per step.

Integrator-agnostic: K is the number of inter-realm seam-coherence boundaries the realm wants visible to the forest. For SSP-RK realms K = rk%nrk; for Leapfrog realms K = 2 (kick, drift); for Yoshida composition K is the number of sub-operators; for an operator-split realm that does all its work atomically and only needs one peer-sync per step, K = 1.

Invoked by forest%evolve_one_step on the multi-realm path to size the stage loop. Under α (end-of-step seam coupling) realms may report different K values: the orchestrator iterates k = 1..max(K_realm) and gates each per-stage TBP on k <= K_realm(is). Under β (stage-coincident) admissibility requires equal K on both endpoints of every β seam (enforced at init).

Default implementation error-stops: every realm that participates in a multi-realm forest MUST override this. Apps that only ever run as N=1 may leave it (the N=1 fast path doesn't call this).

Returns: integer(kind=I4P)

function stages_per_step_forest(self) result(K)

Arguments

Name	Type	Intent	Attributes	Description
self	class(realm_object)	in		The realm.

Call graph