MFC
Exascale flow solver
Loading...
Searching...
No Matches
m_amr Module Reference

Block-structured AMR: up to amr_max_blocks refined blocks (2:1 or 4:1 per amr_ref_ratio), optionally nested to amr_max_level, advanced with the shared solver via grid-state swap and conservatively coupled to each block's parent level (ghost prolongation, Berger-Colella flux reflux, restriction) and dynamic regrid. More...

Functions/Subroutines

impure subroutine s_initialize_amr_module ()
 Re-exports every AMR service module.
impure subroutine s_amr_init_extents ()
 Block caps and preallocation extents. buff_size is unknown at checker time, so the geometric aborts live here.
impure subroutine s_amr_init_report ()
 Rank-0 advisories and the memory demand. Collective (reduces over ranks).
impure subroutine s_amr_init_coarse_patch ()
 Coarse-patch gather buffer (see amr_cg): the largest block's coarse footprint + 2*amr_cpat_mar halo, device-mapped.
impure subroutine s_amr_init_first_blocks ()
 Place the initial block(s) and set their geometry. With dynamic regrid and bodies the block gets the same body-containment expansion regrid boxes get; non-IB blocks are tiled to <= amr_maxc_fit, IB keeps one block. Slot arrays are allocated lazily by s_amr_reconcile_slots.
impure subroutine s_amr_init_tags ()
 Tag bases: narrow-reduction tags sit above the per-box tags; keyed bands start at the next 65536 boundary above them.
impure subroutine s_finalize_amr_module ()

Detailed Description

Block-structured AMR: up to amr_max_blocks refined blocks (2:1 or 4:1 per amr_ref_ratio), optionally nested to amr_max_level, advanced with the shared solver via grid-state swap and conservatively coupled to each block's parent level (ghost prolongation, Berger-Colella flux reflux, restriction) and dynamic regrid.

Function/Subroutine Documentation

◆ s_amr_init_coarse_patch()

impure subroutine m_amr::s_amr_init_coarse_patch

Coarse-patch gather buffer (see amr_cg): the largest block's coarse footprint + 2*amr_cpat_mar halo, device-mapped.

Definition at line 538 of file m_amr.fpp.f90.

Here is the call graph for this function:
Here is the caller graph for this function:

◆ s_amr_init_extents()

impure subroutine m_amr::s_amr_init_extents

Block caps and preallocation extents. buff_size is unknown at checker time, so the geometric aborts live here.

Definition at line 440 of file m_amr.fpp.f90.

Here is the call graph for this function:
Here is the caller graph for this function:

◆ s_amr_init_first_blocks()

impure subroutine m_amr::s_amr_init_first_blocks

Place the initial block(s) and set their geometry. With dynamic regrid and bodies the block gets the same body-containment expansion regrid boxes get; non-IB blocks are tiled to <= amr_maxc_fit, IB keeps one block. Slot arrays are allocated lazily by s_amr_reconcile_slots.

Definition at line 762 of file m_amr.fpp.f90.

Here is the caller graph for this function:

◆ s_amr_init_report()

impure subroutine m_amr::s_amr_init_report

Rank-0 advisories and the memory demand. Collective (reduces over ranks).

Definition at line 490 of file m_amr.fpp.f90.

Here is the call graph for this function:
Here is the caller graph for this function:

◆ s_amr_init_tags()

impure subroutine m_amr::s_amr_init_tags

Tag bases: narrow-reduction tags sit above the per-box tags; keyed bands start at the next 65536 boundary above them.

Definition at line 793 of file m_amr.fpp.f90.

Here is the caller graph for this function:

◆ s_finalize_amr_module()

impure subroutine m_amr::s_finalize_amr_module

Definition at line 833 of file m_amr.fpp.f90.

◆ s_initialize_amr_module()

impure subroutine m_amr::s_initialize_amr_module

Re-exports every AMR service module.

Build the static refined level-1 block (no-op unless amr); runs after the level-0 grid and time-steppers are set up.

Definition at line 403 of file m_amr.fpp.f90.

Here is the call graph for this function:
Here is the caller graph for this function: