POLYCFD / Design notebook

AMReX alignment / design proposal / 26 September 2026

From one grid
to a hierarchy.

Give today's solver a clearer structure. Then teach it to connect patches. Then add resolution where the physics needs it.

Three separate gates. The first changes the representation of existing simulations. Multiblock and refinement are later numerical features. Nothing on this page is evidence that those features already work.

01 / Vocabulary

A domain is a place.
A patch is a piece of storage.

The domain is the physical region being simulated. A level gives it a particular grid spacing. A patch stores a rectangular part of that level.

One coarse patch can cover the entire domain. Fine levels can use several smaller patches. Splitting a level into patches does not change its resolution.

AMReX uses boxes and field collections to express this separation. Official data-model reference.

Physical domainbounds · coordinates · periodicity
Level 0 · spacing h
Patch Aindex box + fields
Patch Bindex box + fields
Proposed structure. During alignment, Level 0 has only one full-domain patch.

Box

Where a patch lives in the level's integer index space.

[lower, upper]

Placement

Where the quantity is sampled relative to a cell.

cell · x/y/z face

Field view

How a kernel reaches values without owning their memory.

bounds + strides + data

Level layout

Which boxes share a resolution and cover which region.

spacing + patch boxes

02 / Follow the same physical domain

The representation changes first.
The mesh changes later.

Choose a stage to inspect the difference. These are schematic 2D grids, not simulation results.

Alignment: one level with one full-domain patch A sixteen by ten cell grid. The alignment effort changes descriptors and ownership while leaving the resolution and physics unchanged. Physical domain · 16 × 10 coarse cells
Level 0 / patch A covers the whole domain. No added resolution.

AMR-02 / same capability

One patch, expressed clearly.

The solver still advances the same grid. Domain coordinates, patch bounds, field placement and memory ownership become explicit and consistently used.

What changes
Data contracts and their consumers.
What must be proved
Existing scientific results, selective capture and performance are preserved.

160 active cells · 1 physical level · 1 patch

Why does keeping one coarse grid not solve all the interface work?

It guarantees base coverage. Fine patches still need values from their neighbors, and the coarse and fine regions must agree on normal volume flux and pressure coupling. The coarse solution must respond to the fine region as part of one physical solution.

Why are allocated and active cells different in the refined illustration?

The fine patch covers 48 of the 160 coarse cells and contains 192 fine cells. We retain 160 + 192 = 352 cell slots, but the physical composite uses 160 − 48 + 192 = 304 active cells. A uniformly fine 2D grid would contain 640 cells. These counts exclude face arrays, ghosts, geometry and solver buffers; they are not a GPU memory or speed prediction. In 3D, a 2:1 refinement divides a coarse cell into eight fine cells.

03 / What already fits

Our MAC fields have a natural home.

Pressure at cell centers, velocity at facesA 2D cell has pressure P in its center, U at the centers of its left and right faces, and V at its top and bottom faces. W is on the out-of-plane faces in three dimensions. P UU VV
2D section of a MAC cell. In 3D, W occupies the Z-normal faces.

Pressure and velocity should not be forced into one identically shaped array. A patch with Nx × Ny × Nz cells has different valid extents for each field.

Pressure
Nx × Ny × Nz
U velocity
(Nx + 1) × Ny × Nz
V velocity
Nx × (Ny + 1) × Nz
W velocity
Nx × Ny × (Nz + 1)

We keep device-resident arrays, float fields, Half geometry and double weighted reductions. Alignment makes their meaning and ownership explicit.

04 / Where the hard work lives

Three boundaries.
Three different responsibilities.

01

The physical boundary

A wall, inlet, outlet or periodic domain edge. Apply the problem's physical conditions here.

Exists today
02

A same-resolution patch edge

An internal storage split. Obtain neighboring samples and agree on the shared face value; this edge is not a wall.

S2
03

A coarse/fine interface

Different resolutions describe the same fluid. Couple pressure, transfer fields and make weighted fluxes consistent.

S3

Current advection clamps or wraps at the current grid's edges. S2 must replace that domain-wide assumption with a policy for sampling across patches, including departure points and corner data.

See AMReX boundary filling and AMReX-Hydro projection methods.

05 / Our current code

Adaptable foundations.
Specific assumptions to untangle.

What alignment changes, and what it leaves for later
AreaAMR-02: representationS2 / S3: behavior
Grid & coordinatesSeparate domain geometry, global boxes and local strides.Support several patches and new refinement layouts.
Session & fieldsOne level with one patch; explicit owners and borrowed views.Coordinate operators and reductions across patches.
Boundary samplingName and guard the current patch-equals-domain restriction.Exchange same-level data, then fill coarse/fine data.
Solid geometrySeparate coordinates from cut-cell metrics; preserve sampling.Establish consistent geometry and fluxes across interfaces.
Results & viewersDerive current metadata from the canonical layout. Preserve wire meanings.Carry native patch/level coverage or label resampled previews.

Physical refinement levels

Where the solution lives

Coarse background plus finer local regions. These levels carry physical simulation state and affect resolution.

Coarse domain → fine region

Multigrid solver levels

How an equation is solved

Auxiliary coarse problems accelerate pressure convergence. They do not add locally refined simulation state.

Pressure problem → coarser corrections

Both can use box descriptors. They need separate identities and ownership. Existing GpuMgHierarchy is the second kind.

06 / Delivery gates

Prove each step before adding the next.

  1. AMR-02

    Align today's structures

    Inventory consumers, introduce canonical descriptors, migrate ownership and callers, then remove superseded representations.

    Gate: existing inputs, fields, measurements, outputs and scientific acceptance remain valid. No new mesh capability and no abstraction-driven performance regression.

  2. S2

    Connect patches at one resolution

    Begin with the same problem split in two. Add same-level sampling, shared-face consistency, coupled pressure/diffusion, geometry, measurements and output.

    Gate: partitioned and monolithic solutions agree within declared tolerances, including periodic and cut interfaces. A simple two-patch test alone does not finish S2.

  3. S3 / fixed

    Add resolution in one region

    Start with one coarse base, one aligned 2:1 fine patch and synchronized timesteps. Add coverage masks, transfers and composite equations.

    Gate: accuracy and conservation pass against uniform-fine references, with less measured memory or time at comparable measurement accuracy.

  4. S3 / adaptive

    Let refined regions change

    Extend fixed refinement to multiple patches/levels. Then add tagging, regridding, safe state transfer, layout revisions and geometry/solver rebuilding.

    Gate: refine/derefine cycles preserve the required budgets and constraints. Thermal and moving-geometry combinations need their own evidence. Time subcycling is a later decision.

Continue reading

The design behind the diagrams.

The Markdown plans define scope and acceptance. The work board owns current status. Proposed type names and diagrams are explanatory; they are not executable API or file-format specifications.